JPH0591312A - Picture processor - Google Patents

Abstract

PURPOSE:To quickly and accurately discriminate an image area by comparing an adjacency number indicating the adjacency state with a prescribed threshold and discriminating a pseudo halftone and a non-halftone and changing the threshold based on the resolution of binary, picture data. CONSTITUTION:One of a first threshold for ultrafine and a second threshold for fine is used as the threshold of the 4-direction adjacent number for detecting a distributed halftone picture in accordance with a set resolution by an adjacent state discriminating part 105. A systematic half tone discriminating part 106 detects whether the picture in the area having a noticed picture element as the center is a systematic halftone picture or not with respect to each picture element. The adjacent number indicating the adjacent state 18 compared with the prescribed threshold to perform discrimination, and the threshold is changed based on the resolution of binary picture data. For example, when the fine mode and the ultrafine mode are provided as the mode for picture read, the threshold is changed in accordance with each mode to discriminate the image area.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus having an image restoration circuit for restoring pseudo halftone binary image data into multi-tone image data (hereinafter referred to as multi-valued image data).

[0002]

2. Description of the Related Art Conventionally, in a facsimile machine, an image signal is transmitted through a public telephone line. Therefore, for a non-halftone image such as a character, the non-halftone image has a predetermined threshold value. Is converted to non-halftone binary image data by being binarized using
With regard to halftone images such as photographs, the multi-valued image data of the above halftone images is binarized by a dither method or the like to be converted into pseudo halftone binary image data and transmitted to the receiving side. On the other hand, on the receiving side, different processing is performed depending on the received non-halftone binary image data and the pseudo-halftone binary image data, for example, extraction of binary image data and switching of pixel density conversion method. , It is necessary to switch encoding methods such as compression of image data.

Therefore, it is necessary to automatically determine whether the received binary image data is the non-halftone binary image data or the pseudo-halftone binary image data. An image area signal generation method (hereinafter referred to as a first conventional example method) for performing this determination and outputting a determination result signal is disclosed in, for example, Japanese Patent Publication No. Sho 63-11832. The method of the first conventional example is
A method for generating an image area signal indicating which area each pixel is included in, for an image signal having both a halftone picture area and a text area, and a mask surrounding a plurality of pixels including a target pixel. And a pixel pattern formed from the values of the plurality of pixels in the mask is extracted, and all possible pixel patterns are divided into a halftone dot picture area and a text area and stored in advance. According to the area signal table, an image area signal corresponding to the pixel pattern is extracted, and this image area signal is used as an image area signal for the target pixel. Specifically, the halftone dot photograph area and the text area are divided, and each pixel pattern is stored in advance in the table ROM, and the image signal to be processed is input to the address terminal of the ROM, so that the ROM is processed. The image area signal is obtained from the data terminal of.

A device for automatically determining whether the input binary image data is the non-halftone binary image data or the pseudo-halftone binary image data (hereinafter , Referred to as the second conventional device), Yoshinobu Mita et al., "High-definition multilevel reconstruction of binary image by neural network", Japan Hardcopy '90, NI.
P-24, pp 233-236, proposed in 1990. In the device of the second conventional example, the backpropagation of a neural network, which has recently been spotlighted, is used to perform multi-valued conversion of a binary image by the neural network, and the binary image is divided into regions by the neural network. Separately, multi-value conversion is performed. The neural network used in this device has an input layer,
Using a three-layer structure of an intermediate layer and an output layer, the input layer is provided with as many units as the number of pixels included in the window provided around the pixel of interest, and the output layer is used for outputting multi-valued image data. It has one unit.

However, since the method of the first conventional example is provided with the ROM for the pixel pattern table in order to discriminate the image area, if the area for discriminating the area is made large, a large capacity ROM is required. There was a problem that the cost increased. Further, the second conventional apparatus has a problem that the apparatus configuration is complicated and the processing speed is very slow.

In order to solve the above problems, the present inventor has proposed in the patent application of Japanese Patent Application No. 3-100992 an image processing apparatus equipped with an image area discriminating means (hereinafter referred to as a third conventional example apparatus). .) Was proposed. The proposed third conventional example apparatus uses binary image data binarized in pseudo-halftone,
Based on the input binary image data including the binary image data that has been binarized in a non-halftone using a predetermined threshold value, the same pixel in the block including the target pixel to be processed is included. A pseudo intermediate for each pixel of the input binary image data based on the adjacency number indicating the adjacency state of one kind of pixel and the adjacency number indicating the adjacency state calculated by the arithmetic means. It is characterized in that it is provided with a discriminating means for discriminating whether it is an area of binary image data of a tonality or an area of binary image data of a non-halftone.
The device of the third conventional example configured as described above is a ROM
It has an advantage that it can be configured by a simple circuit without using, and the image area can be determined at high speed.

By the way, in a conventional facsimile apparatus, since an image signal is transmitted through a public telephone line,
For halftone images such as photographs, the transmitting side converts the multivalued image data of the halftone image into binary image data of pseudo halftone by binarizing it by a dither method or the like, and transmits it to the receiving side. On the other hand, on the receiving side, a method of restoring the received pseudo-halftone binary image data into multi-valued image data is used.

In recent years, a color laser printer for recording multi-valued image data at high speed and high resolution has been put into practical use, but generally, a binary printer for recording binary image data is often used. When storing multi-valued image data in a storage device, a large-capacity storage device is required. To solve this problem, however, multi-valued image data is once converted into binary image data and stored in the storage device. For image processing or recording, a method has been proposed in which the binary image data is read from the storage device and then restored to multivalued image data.

A method and an apparatus of this kind are disclosed in, for example, Japanese Patent Laid-Open Nos. 62-114378 and 62-10.
It is disclosed in Japanese Patent No. 7573.

The former fourth image processing method of the prior art uses the multi-valued printer while using the conventional binary image data to bring out the performance of these devices sufficiently and restores the character image to the restored character. In order to improve the quality, a halftone image is restored from the binary image data, and the restored halftone image is subjected to predetermined image processing such as enlargement / reduction processing or image enhancement processing. In the method of the fourth conventional example, in order to restore the binary image data into the multi-valued image data, a square window of a predetermined size is provided around a pixel to be restored (hereinafter referred to as a pixel of interest). It is provided and the smoothing process is performed within the window.

The latter fifth image processing apparatus of the prior art prevents deterioration of image quality when a simple binary circuit such as the dither method is used, and is constituted by a simple circuit. A unit that divides the binarized image information into predetermined blocks, an identifying unit that identifies the image tone for each block, and the image information in the block for each pixel according to the identification result of the identifying unit. It is characterized in that it is provided with a converting means for converting into a multi-valued level. In the device,
When the image is transmitted and stored, it is treated as binary image data to improve the efficiency of display and editing of the image, and at the time of reproducing the image, a multi-valued representation similar to an analog image is performed. Specifically, the identifying means and the converting means are respectively composed of an image-tone determining ROM for performing image-tone determination by pattern matching in a block corresponding to the size of the dither matrix, and a converting ROM.

[0012]

However, in the third conventional apparatus, for example, when the image is read in the fine mode, it is converted into image data having sharp edges, but the resolution is higher than that in the fine mode. When read by, the image data may be image data with a gentle edge. Therefore, there is a problem that the image area may be erroneously discriminated when the resolution at the time of reading the image is changed. Furthermore, when the binary image data is restored to multi-valued image data using the restoration device of the fourth or fifth conventional example using the erroneous image area discrimination result, the original image cannot be restored. However, there was a problem that the reproducibility of was decreased.

Further, using the method of the fourth conventional example or the apparatus of the fifth conventional example, binary image data binarized in pseudo halftone is converted into multi-valued image data such as a photographic image. When restoring to, the spatial filter of a predetermined size is used as described above. Here, in general, the size of the spatial filter is determined depending on in which spatial frequency band the pseudo-halftone texture is excluded and in which spatial frequency band the original component of the image is stored or emphasized. However, when restoring to multi-valued image data using a spatial filter of a predetermined fixed size,
If the resolution of the binary image data changes, the latter spatial frequency band moves, which causes a problem that the original image cannot be restored and the reproducibility of the image deteriorates.

Further, in the binary image data binarized by the pseudo halftone, the spatial frequency component of the pseudo halftone texture and the spatial frequency component of the original image are mixed. Generally, the former spatial frequency is higher than the latter spatial frequency, but in particular, an image of binary image data binarized by pseudo-halftone by the dot-concentrated systematic dither method is In comparison, the former spatial frequency exists in a lower spatial frequency band.

Therefore, when there is no difference between the spatial frequency bands of the two, the image area discriminating circuit, which only discriminates whether the image data is pseudo-halftone binary image data or non-halftone binary image data, There is a possibility of making an incorrect image area discrimination. Using the erroneous image area discrimination result, for example, the fourth or fifth
When the binary image data is restored to the multi-valued image data by using the restoration device of the related art, there is a problem that the original image cannot be restored and the reproducibility of the image is deteriorated.

The first object of the present invention is to solve the above problems and to construct a simple circuit as compared with the conventional example, to obtain binary image data binarized in pseudo halftone and a predetermined threshold. Even if the resolution of the binary image data is changed from the binary image data that has been binarized in the non-halftone by using the value, it is possible to discriminate at high speed and more accurately as compared with the conventional example. An object of the present invention is to provide an image processing device capable of performing the same. The second aspect of the present invention
The object of solving the above problems includes binary image data binarized in pseudo-halftone and binary image data binarized in non-halftone using a predetermined threshold value. Entered 2
When the value image data is restored to the multi-valued image data, even if the resolution of the binary image data changes, the reproducibility of the image is improved as compared with the conventional example, and the deterioration of the image is prevented. It is to provide an image processing device capable of

[0017]

An image processing apparatus according to a first aspect of the present invention uses binary image data binarized in pseudo-halftone and non-halftone in a predetermined threshold value. Based on the input binary image data including the binarized binary image data, the adjacency number indicating the adjacency state of pixels of the same type in a block composed of a plurality of pixels including the target pixel to be processed is calculated. Based on the adjoining number indicating the adjacency state computed by the computing means and the computing means
Whether the value image data is pseudo halftone binary image data,
Alternatively, in the image processing apparatus provided with a discriminating means for discriminating whether the image data is non-halftone binary image data, the discriminating means compares the number of adjacencies indicating the adjacency state with a predetermined threshold value. The above-mentioned determination is performed, and the threshold value is changed based on the resolution of the input binary image data.

According to a second aspect of the present invention, in the image processing apparatus according to the first aspect, the input binary image data is multi-graded based on the result of the discrimination by the discrimination means. It is characterized in that it is provided with a restoring means for restoring the image data.

Further, the image processing apparatus according to claim 3 is
In the image processing device, which is provided with a restoring unit that restores the input binary image data into multi-tone image data by using a spatial filter having a predetermined window based on the input binary image data. When the resolution of the input binary image data is higher than a predetermined threshold value, the restoring means uses the spatial filter having a window having a predetermined first size to extract the input binary image data. When the resolution of the input binary image data is lower than a predetermined threshold,
The input binary image data is restored to multi-tone image data by using a spatial filter having a window smaller than the first size.

Furthermore, the image processing apparatus according to the fourth aspect is binary image data obtained by binarizing the binary image data based on the input binary image data by the dot-concentrated systematic dither method. A discriminating means for discriminating whether or not there is any, and the inputted binary image data is subjected to multi-gradation by performing image processing on the inputted binary image data according to the discrimination result by the discriminating means. And a restoration unit for restoring the image data, wherein the determination unit determines that the resolution of the input binary image data is less than or equal to a predetermined value.
It is characterized in that it is determined not to be binarized by the dot-concentrated systematic dither method.

[0021]

According to the image processing apparatus of the present invention configured as described above, the discrimination means performs the discrimination by comparing the number of adjacencies indicating the adjacency state with a predetermined threshold value, and The threshold value is changed based on the resolution of the binary image data. For example, when there are a fine mode and an ultra fine mode as a mode at the time of reading an image, the threshold value is changed according to each mode to determine the image area. Therefore, a ROM can be used and a simpler circuit can be used as compared with the conventional method or apparatus, and even when the resolution of the binary image data is changed, the operation can be performed at high speed and more accurately. Each data can be discriminated.

The image processing apparatus according to claim 2 further includes the image processing apparatus according to claim 1.
Since the restoration means for restoring the input binary image data to the multi-gradation image data based on the result determined by the determination means is provided, the restoration processing is performed based on the more accurate image area determination result. be able to. Therefore, when restoring input binary image data to multi-valued image data,
Even when the resolution of the binary image data changes, the image reproducibility can be improved and image deterioration can be prevented as compared with the conventional example.

Further, in the image processing apparatus according to the third aspect of the invention, the restoring means has a predetermined first size when the resolution of the input binary image data is higher than a predetermined threshold value. When the input binary image data is restored to multi-tone image data using a spatial filter having a window, while the resolution of the input binary image data is lower than a predetermined threshold value, The input binary image data is restored to multi-tone image data by using a spatial filter having a window smaller than the first size. Therefore, when the input binary image data is restored to multi-valued image data, the size of the spatial filter window should be changed as described above even if the resolution of the binary image data changes. As a result, the spatial frequency component originally possessed by the image can be preserved, the reproducibility of the image can be improved and deterioration of the image can be prevented as compared with the conventional example.

Further, in the image processing apparatus according to the fourth aspect, the discriminating means binarizes the binary image data based on the input binary image data by the dot-concentrated systematic dither method. If the resolution of the input binary image data is less than a predetermined value, it is determined that it is not binarized by the dot-concentrated systematic dither method. To do. Therefore, even if the resolution of the binary image data changes, the spatial frequency component originally possessed by the image can be preserved, the reproducibility of the image is improved as compared with the conventional example, and the deterioration of the image is prevented. can do.

[0025]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A facsimile apparatus according to an embodiment of the present invention will be described below with reference to the drawings. Here, the facsimile apparatus according to the present embodiment is characterized by including an image restoration processing unit 62 for restoring the received binary image data into multi-valued image data, as shown in FIG.

In the following description of the embodiments, the "halftone image" and the "halftone region" respectively refer to multivalued image data of a halftone image such as a photograph and a pseudo halftone method such as a dither method. Means a pseudo halftone image and an area of the image, which are binarized in the above. On the other hand, the “non-halftone image” and the “non-halftone area” are, for example, a non-halftone image such as a character and the area of the image. Say. As an intermediate image between the “halftone image” and the “non-halftone image”, an image in which a character image is binarized in pseudo-halftone, an image formed artificially such as a character font, etc. can be considered. Image is treated as a "non-halftone image" in this embodiment. Furthermore, since a photographic image is simply binarized in a non-halftone image using a predetermined threshold value, image information is almost lost, and thus it is not considered in this embodiment.

Generally, the binarization method of image data can be classified as follows. That is, 2 of the image data
The binarization method can be roughly classified into a simple binarization method that binarizes using a predetermined threshold and a pseudo halftone binarization method that binarizes by pseudo halftone. .. Further, the pseudo halftone binarization method is a random dither method (also referred to as an error diffusion method) (hereinafter, in the present embodiment, the first halftone 2).
It is called a binarization method or a binarization method in the first halftone mode. )When,
Can be divided into two types of systematic dithering,
The systematic dither method is a dot dispersion type dither method and a dot concentrated type dither method (hereinafter, in the present embodiment, it is referred to as a second halftone binarization method or a second halftone mode binarization method). It can be classified into two. Now, a binary image binarized by pseudo-halftone using the random dither method and the systematic dither method is used as a random halftone image and a systematic halftone image, respectively. A binary image binarized by pseudo halftone is used as a concentrated halftone image. On the other hand, since the random dither method is basically a dot dispersion type binarization method, other pseudo halftone images can be called a dispersion type halftone image. In the present embodiment, by invalidating the systematic halftone discrimination result for the random halftone image, while invalidating the distributed halftone discrimination result for the concentrated type halftone image, The discrimination accuracy regarding image area discrimination is improved.

A facsimile machine according to an embodiment of the present invention will be described in the order of the following items. (1) Features of this embodiment (2) Facsimile apparatus configuration and operation (3) Image restoration processing section (4) 15 × 18 matrix memory circuit (5) Image area determination section (5-1) Configuration and operation of each section (5-2) Adjacent state determination unit (5-3) Systematic halftone determination unit (5-4) 9 × 17 matrix memory circuit (5-5) Determination data generation unit (5-6) Determination data signal generation unit (6) Halftone image restoration unit (6-1) Configuration and operation of each unit (6-2) Smoothing amount calculation unit (6-3) Edge emphasis amount calculation unit (6-4) Edge discrimination amount calculation unit (6- 5) Restored data calculation unit

(1) Features of this Embodiment As shown in FIG. 7, the facsimile apparatus of this embodiment has the following features.
(A) In the received binary image data including the binary image data binarized in pseudo halftone and the binary image data binarized in non-halftone using a predetermined threshold value A predetermined edge enhancement amount, a predetermined smoothing amount, and a predetermined edge discrimination amount are calculated based on the discrimination data signal generation unit 1 described later.
And a centralized type halftone discrimination signal and a systematic halftone discrimination signal which are respectively output from 14 and indicate the determination results of the predetermined concentrated type halftone image and the systematic halftone image, respectively.
A halftone image restoration unit 101 that restores the binary image data input based on the ultrafine signal output from U50 and each of the above calculated amounts into multivalued halftone data,
(B) Based on the received binary image data, the concentrated halftone signal, the ultrafine signal, and the random halftone signal output from the MPU 50, for each pixel in a predetermined region centered on the pixel of interest In addition, the centralized halftone discrimination signal indicating the discrimination result of the concentrated halftone image and the systematic halftone discrimination signal indicating the discrimination result of the systematic halftone image are generated, and An image area discriminating unit 102 that outputs image area discrimination data indicating the result of discriminating whether there is a certain area or a non-halftone area, and (c) binarization with a non-halftone using a predetermined threshold value. A simple multi-value conversion unit 103 for simply converting binary image data into multi-valued non-halftone image data showing white or black; and (d) MPU 50.
When the restoration execution signal output from H is high, the multi-valued halftone image data output from the halftone image restoration unit 101 and the simple multi-value quantization unit 103 according to the mixing ratio indicated by the image area discrimination data. Multi-valued non-halftone image data output from the multi-valued image data is generated and output to the printer controller 55 through the interpolation processor 64, while the restoration execution signal is at the L level. And a data mixing unit 104 that outputs the multi-valued non-halftone image data output from the simple multi-value conversion unit 103 to the printer control unit 55 via the interpolation processing unit 64 as it is. To do.

Here, the halftone image restoration unit 101
(A) A smoothing amount calculator 109 that calculates and outputs three first to third smoothing amount data for restoring halftone image data based on each pixel data of the received binary image data.
And (b) an edge enhancement amount calculation unit 110 that calculates and outputs first to third edge enhancement amount data for performing edge enhancement processing based on each pixel data of the received binary image data, ( c) A first used to determine an edge area based on each pixel data of the received binary image data
And an edge discrimination amount calculation unit 111 that calculates and outputs the second edge discrimination amount, (d) data output from each of the units 109 to 111, and a centralized halftone discrimination signal output from the discrimination data signal generation unit 114. A systematic halftone discrimination signal, and a restored data calculation unit 112 that restores and outputs multivalued halftone image data based on the ultrafine signal output from the MPU 50.

Further, the image area discriminating section 102 (a) adjoins the same kind of pixels, which is the smallest in a predetermined area based on the input pixel data, in four main and sub-scanning directions. In addition to calculating the number of adjacent in the main and sub scanning directions,
Data of the number of black pixels in a predetermined 7 × 7 window is calculated, and attention is paid to each pixel based on the calculated data, the ultrafine signal and the concentrated halftone signal output from the MPU 50, respectively. A non-dispersive halftone image detection signal indicating that the image in a predetermined area centered on a pixel is a non-dispersive halftone image, and a variance indicating that the image in the predetermined area is a dispersive halftone image Type halftone image detection signal and an all-white all-black image detection signal that indicates whether the 7 × 7 window is an all-white image or all-black image, and outputs the adjacent-state determination unit 105. And (b)
A systematic halftone detection signal indicating the detection result by detecting whether or not the image of a predetermined area centered on the pixel of interest is a systematic halftone image for each pixel based on the input pixel data. Of the four detection signals (total of 4 in total) output for each pixel from the systematic halftone judging section 106 that outputs the (c) adjacent state judging section 105 and the systematic halftone judging section 106.
9x17 matrix memory circuit 107 for simultaneously outputting (bit)) within a predetermined 9x17 window centered on the pixel of interest, and (d) matrix memory circuit 107
A determination data generation unit 108 that generates and outputs determination data obtained by adding the detection signals in the 9 × 17 window for each detection signal based on the detection signals output from e) Based on each determination data output from the determination data generation unit 108, it is determined whether or not the image in the predetermined 9 × 17 window area is a concentrated halftone image, and a concentration indicating the determination result A type halftone discrimination signal is generated and output, and it is determined whether or not the image in the predetermined 9 × 17 window area is a systematic halftone image, and a systematic halftone determination signal indicating the determination result is output. A discriminant data signal generation unit 114 is provided for generating and outputting, and for outputting image region discrimination data indicating a result of discriminating whether the region is a halftone region or a non-halftone region.

In particular, in the present embodiment, the adjacent state judging section 10
5, as a threshold of the number of adjacencies in four directions, which will be described later in detail for detecting the dispersed halftone image, according to the resolution to be set, a first threshold for ultrafine, and
One of two second threshold values for fine is used.

(2) Configuration and Operation of Facsimile Device FIG. 1 is a vertical cross-sectional view of a mechanical portion of a facsimile device according to an embodiment of the present invention, and FIG. 3 shows signals of the facsimile device shown in FIG. It is a block diagram which shows the structure of a process part. As shown in FIG. 1, this facsimile apparatus is roughly divided into a printer unit 1 and an image reading unit 20 installed above the printer unit 1, an operation panel 40 is provided on the printer unit 1, and the printer unit 1 includes A telephone 42 is provided on the side surface.

In FIG. 1, a printer unit 1 is an electrophotographic laser beam printer having the same structure as a conventional apparatus, and its operation will be briefly described below. First, the photoconductor on the photoconductor drum 2 which is rotationally driven is uniformly charged by the charger 3. Next, the optical system 4 irradiates a laser beam according to the image data to form an electrostatic latent image on the photosensitive drum 2. The toner of the developing device 5 adheres to this electrostatic latent image. On the other hand, cut sheets are placed in the paper feed cassette 11, and the cut sheets are picked up one by one by the pickup roller 12, and then fed by the paper feed roller 13 toward the transfer portion of the photosensitive drum 2. The toner attached to the photoconductor drum 2 is transferred to the cut paper by the transfer charger 6 and fixed by the fixing device 12. The cut sheet after the fixing process is the sheet discharge rollers 14,
The sheet 16 is discharged to the sheet discharge tray 13 via the sheet discharge passage 15. The toner that has not adhered to the cut sheet is collected by the cleaner 8, and one printing is completed.

Next, the operation of the image reading section 20 will be described. The transmission document is read in the same manner as the conventional device. That is, the document placed on the document tray 21 is detected by the document sensor 22, and the document is fed one by one to the position of the sensor 25 by the roller 23. Next, the original is read by the contact type linear image sensor 26 in synchronization with the rotation of the roller 24 by a motor (not shown) and the reading of the contact type linear image sensor 26.
The original image is converted into digital image data and then output to the buffer memory 59 shown in FIG. After the image reading is completed, the document is ejected by the ejection roller 27 to the ejection tray 2
It is discharged to 8.

FIG. 2 is a front view of the operation panel 40 of this facsimile apparatus. As shown in FIG. 2, the operation panel 40 includes a telephone numeric keypad 540 including numeric keys “0” to “9”, an asterisk key “*”, and a sharp key “#”; a liquid crystal display panel 541; Key 54
2,543,546 and display LEDs 544,545,5
47 and 548. The liquid crystal panel 541 displays the operating state of the facsimile device or displays instructions to the operator. The key 542 is a transmission key for instructing the start of the transmission operation of the facsimile device. The key 543 is a key for setting the halftone mode at the time of transmission, and every time the key 543 is pressed, the “first halftone mode”, the “second halftone mode”, and the “simple binary value” are selected. The three modes of the "ization mode" are selectively set in order. These setting states are displayed by the display LEDs 544 and 545, and display L is displayed when the "first halftone mode" is set.
When only the ED 544 is turned on and the “second halftone mode” is set, only the display LED 545 is turned on and when the “simple binary mode” is set, the display LEDs 544, 5
Both 45 are turned off. A key 546 is a key for setting the resolution at the time of reading an image, and the key 54
Every time 6 is pressed once, three modes of “fine mode”, “ultra fine mode”, and “standard mode” are selectively set in order. These setting states are displayed by the display LEDs 547 and 548. When the “fine mode” is set, only the display LED 547 is turned on, when the “ultra fine mode” is set, only the display LED 548 is turned on, and the “standard mode” is displayed. When is set, both the display LEDs 547 and 548 are turned off. In this embodiment, the resolution of the standard mode is
The main scanning direction is 8 dots / mm and the sub-scanning direction is 3.85 dots / mm, and the fine mode resolution is 8 dots / mm in the main scanning direction and 7.7 dots / mm in the sub-scanning direction. The resolution is 16 dots / mm in the main scanning direction and 7.7 dots / mm in the sub scanning direction.

As shown in FIG. 3, in this facsimile apparatus, an MPU 50 that controls the entire facsimile apparatus, an HDLC analysis unit 52 that performs facsimile signal processing and communication processing, and a modem 53, respectively.
And NCU 54, an image compression image memory 51 for temporarily storing a facsimile image signal, a buffer memory 59, and a page memory 61, and a compression / expansion unit 60 and an image restoration processing unit 62 for processing predetermined image signals. And each processing unit 20, 51, 52, 5
3, 54, 59, 60, 61 are MPUs via the bus 63
Connected to 50. In addition, the operation panel 40 directly
A printer control unit 55 that is connected to the U50 and that controls the multi-valued laser printer 70 that is provided in the printer unit 1 and that prints an image of the image data based on the multi-valued image data is connected to the MPU 50.

When receiving a facsimile signal from the facsimile machine of the other party, the MPU 50, as will be described later in detail with reference to FIG. 4 and FIG.
A random halftone signal, a concentrated halftone signal, based on each data signal received in phase B defined in 30 and the restoration ON signal output from the image processing unit 20a in the image reading unit 20, A restoration execution signal, an ultrafine signal, and a fine signal are generated, and a total of five control signals are generated. The former four control signals are output to the image restoration processing unit 62, and the latter two control signals are output to the interpolation processing unit 64. To do.

The image restoration processing section 62 processes the binary image data outputted from the page memory 61 based on the random halftone signal, the concentrated halftone signal, the restoration execution signal and the ultra fine signal outputted from the MPU 50. Then, after performing image restoration processing as described later in detail, the restored multivalued image data is output to the interpolation processing unit 64. Next, the interpolation processing unit 64 is configured by a known circuit, and based on the ultrafine signal and the fine signal output from the MPU 50 and indicating the resolution at the time of recording, the interpolation processing unit 64 determines a predetermined value with respect to the restored multi-valued image data. Output to the printer control unit 55. In this embodiment, the laser printer 70 has an ultra fine resolution, and the interpolation processing unit 64 operates in the main scanning direction when both the ultra fine signal and the fine signal are at the L level and the resolution is the standard mode. The interpolation processing of 2 times is performed, and the interpolation processing of 4 times is performed in the sub-scanning direction. When the ultrafine signal is at the L level, the fine signal is at the H level, and the resolution is in the fine mode, the interpolation processing unit 64 performs the double interpolation processing in the main scanning direction and the subscanning direction. Then, double interpolation processing is performed. Further, when the ultra fine signal is at the H level, the fine signal is at the L level, and the resolution is in the ultra fine mode, the interpolation processing unit 64 does not perform the interpolation processing and outputs the input multi-valued image data. It is directly output to the printer control unit 55.

First, the transmission operation of the facsimile will be described. When all the above-mentioned image reading operations by the image reading unit 20 are completed, the NCU 54 calls the facsimile machine of the other party, and the line connection processing with the facsimile machine of the other party is executed according to the call setting procedure prescribed in the CCITT recommendation. (Defined in CCITT Recommendation T.30 (hereinafter the same applies) Phase A). After the line connection process is completed, the MPU 50 determines the resolution and compression method according to the capabilities of the mutual facsimile machines, and transmits information and adjusts the phase in the unique mode permitted between the machines of the same manufacturer. Processing such as training is performed by the NCU 54, the modem 53, and the HDLC analysis unit 52 (phase B). The resolution information transmitted from the transmission side facsimile apparatus to the reception side facsimile apparatus is set by the operation key 546, and in the facsimile communication with the other company, the resolution information in the fine mode or the standard mode is transmitted. It In the present embodiment, in the unique mode, the following information is transmitted from the facsimile apparatus on the transmitting side to the facsimile apparatus on the receiving side by using, for example, a non-standard function setting signal (NSS) prescribed in CCITT recommendation in the communication of the phase B. Is transmitted.
(A) Information indicating which resolution is selected from among ultra fine mode, fine mode, and standard mode, (b) any binary value among the first halftone mode, second halftone mode, and simple binary mode Information indicating that the image data is to be restored, and (c) information of a restoration on signal indicating whether the transmitted image data is image data to be restored.

Next, the compressed image data stored in the compression image memory 51 is once expanded by the compression / expansion unit 60 into the page memory 61, and then recompressed according to the capability of the facsimile machine of the other party. It is stored in the compression image memory 51. The stored image data is HDL
After the predetermined HDLC frame processing is executed by the C analysis unit 52, it is modulated into a predetermined facsimile signal by the modem 53. The facsimile signal modulated with the image data is transmitted to the facsimile machine of the other party through the NCU 54 and the telephone line (phase C). When the completion of the transmission of the image data is confirmed (Phase D), the line disconnection process with the destination facsimile machine is executed according to a predetermined line disconnection procedure (Phase E), and the transmission operation is completed.

Next, the receiving operation of the facsimile will be described. When an incoming call is received from the facsimile machine of the other party via the telephone line, an incoming call signal is input to the MPU 50 via the NCU 54 and the modem 53 and detected, and then according to a predetermined facsimile line connection procedure, The line connection processing with the facsimile is executed (phase A). After the line connection processing is completed, the resolution, compression method, etc. are decided according to the capabilities of the mutual facsimile machines.
The NCU 54 and the modem 53 perform processing such as information transmission, phase adjustment, and training in a unique mode that is performed by the U50 and is permitted between devices manufactured by the same manufacturer.
And the HDLC analysis unit 52 (phase B). Here, in facsimile communication with a facsimile machine of another company, the resolution information indicating the fine mode or the standard mode is transmitted from the facsimile machine on the transmission side to the facsimile machine on the reception side, and then received by the MP machine.
Input to U50. Further, in the unique mode, the information of (a) to (c), which is transmitted by using the non-standard function setting signal (NSS), is transmitted from the facsimile apparatus on the transmitting side to the facsimile apparatus on the receiving side and received. M
It is input to the PU 50.

Then, the compressed image signal transmitted from the facsimile machine of the other party is sent to the modem 5 via the NCU 54.
3 is input to and demodulated, and the compressed image data after demodulation is H
The DLC analysis unit 52 performs a predetermined HDLC reverse processing process for extracting only the compressed image data from the HDLC frame, and then stores it in the compression image memory 51 (phase C). When the compressed image signals of all pages are received and the reception of the image data is confirmed (Phase D), the line disconnection process with the destination facsimile machine is executed according to the predetermined facsimile line disconnection procedure (Phase D). E). The image data stored in the image memory for compression 51 is processed by the compression / expansion unit 60 page by page in the page memory 61.
Is used to decompress and expand the actual image data. Here, the MPU 50 monitors the data amount of the image data stored in the compression image memory 51 on a page-by-page basis, and checks whether the data amount of each page exceeds a predetermined data amount M. This check is performed to determine whether the image data received from the facsimile machine of the other party is the halftone image data in the halftone mode in the facsimile communication with the facsimile machine of another company, not in the original mode. Be seen. In the case of facsimile communication with a facsimile device manufactured by another company, it is not possible to determine whether the binarization method is the halftone mode because there is no transmission of the halftone mode information in the unique mode. Therefore, in this embodiment, the data amount in page units is checked, and when the data amount exceeds a predetermined data amount, it is determined that the halftone mode is set.

In the present embodiment, the halftone mode is determined by a simple comparison of the data amount in page units with a predetermined data amount, but the present invention is not limited to this and compression of image data is performed. Depending on the method, the threshold data amount M may be changed, or the fill bit may be deleted from the received image data.

Further, the image data expanded in the page memory 61 is input to the image restoration processing unit 62, converted into high-density binary image data by the processing described later in detail, and then interpolation processing is performed. It is output to the printer control unit 55 via the processing unit 64. A recording start signal is output from the MPU 50 to the printer control unit 55 in synchronization with the transfer of the image data to the printer control unit 55, and the printer control unit 55 transmits the control signal and the image data to the laser printer 70 to transfer the image. Cause data recording.

The MPU 50 performs predetermined processing based on an operator's instruction input using the operation panel 40, and outputs instruction information to the operator and status information of the facsimile apparatus to the operation panel 40. indicate.

FIGS. 4 and 5 show the MPU 5 shown in FIG.
9 is a flowchart showing a control flow of a control signal setting process executed in a reception process of facsimile communication at 0. The MPU 50 receives the following information on the basis of the information about the unique mode, the resolution information, the binarization method information, and the restoration on signal information received in the phase B of the reception process with the destination facsimile machine. The control signal setting process is performed.

As shown in FIG. 4, in step S101, it is determined whether or not it is the unique mode with the facsimile machine manufactured by the company. If it is the unique mode (YES in step S101), the first step It is determined whether or not it is the 1 halftone mode, and then
In step S103, it is determined whether the second halftone mode is set. On the other hand, when it is not the unique mode (NO in step S101), step S107 in FIG.
Proceed to.

In the first halftone mode (step S
After the random halftone signal is set to the H level and the concentrated type halftone signal is set to the L level in S104, the process proceeds to step S108. In the second halftone mode (YES in step S103), the random halftone signal is set to the L level and the concentrated halftone signal is set to the H level in S105, and then the process proceeds to step S108. Furthermore, when neither the first halftone mode nor the second halftone mode (NO in steps S102 and S103), S1
After the random halftone signal is set to the L level and the concentrated type halftone signal is set to the L level in 06, the process proceeds to step S108.

Next, in step S108, it is determined whether or not the restoration ON signal is at the H level, and then in step S109, it is determined whether or not the resolution is the standard mode. Here, when the restoration ON signal is at the H level (YES in step S108) and the resolution is not the standard mode (N in step S109).
O), after setting the restoration execution signal to the H level in S110, the process proceeds to step S116. On the other hand, when the restoration ON signal is at the L level (NO in step S108) or the resolution is the standard mode (step S10).
(YES in 9), after setting the restoration execution signal to the L level in S111, the process proceeds to step S116.

Furthermore, it is determined in step S116 whether the resolution is the standard mode, and then in step S117 it is determined whether the resolution is the fine mode. Here, when the resolution is the standard mode (YES in step S116), step S
After setting the fine signal to the L level and the ultra fine signal to the L level at 118, the setting process is ended. When the resolution is the fine mode (YES in step S117), the fine signal is set to the H level and the ultra fine signal is set to the L level in step S119, and then the setting process is ended. Further, when the resolution is not the standard mode (NO in step S116) and not the fine mode (NO in step S117), the fine signal is set to the L level and the ultrafine signal is set to the H level in step S120, and The setting process ends.

On the other hand, when the facsimile communication is not with the unique mode but with a facsimile machine of another company, the flow proceeds from step S101 to step S107 of FIG. After setting the control signal to the L level, step S112
Proceed to. In step S112, it is determined whether or not the page-based data amount of the received image data exceeds a predetermined threshold data amount M, and then step S1.
At 13, it is determined whether the resolution is the standard mode. Here, the data amount of the received image data in page units exceeds the data amount M of a predetermined threshold value (step S
If YES in 112) and the resolution is not the standard mode (NO in step S113), step S1
After setting the restoration execution signal to the H level in 14, the process proceeds to step S121. On the other hand, when the data amount of the received image data in page units does not exceed the data amount M of the predetermined threshold value (NO in step S112) or the resolution is in the standard mode (Y in step S113).
ES), after setting the restoration execution signal to the L level in step S115, the process proceeds to step S121.

Next, in step S121, it is determined whether or not the resolution is the standard mode, and if the resolution is the standard mode (YE in step S121).
S), after setting the fine signal to the L level and the ultra fine signal to the L level, the setting process ends. On the other hand, when the resolution is not the standard mode (NO in step S121), the fine signal is set to H level and the ultra fine signal is set to L level, and then the setting process is ended.

In the control signal setting process, the restoration execution signal is set to the L level when the resolution is not the original mode and the resolution is the standard mode. This is because the spatial frequency of the texture approaches the spatial frequency of the original image, and a sufficient restoration result cannot be obtained. In addition,
Since the image restoration is performed only in the fine mode or the ultra fine mode, only the ultra fine signal is sent to the image restoration processing unit 62.

FIG. 6 is a block diagram of the image processing section 20a in the image reading section 20 shown in FIG. As shown in FIG. 6, when the first halftone mode is set using the key 543 of the operation panel 40, the first halftone mode signal of H level is sent from the MPU 50 to the first half of the OR gate 78 in the image processing unit 20a. Input to the input terminal of. Also, when the second halftone mode is set using the key 543 of the operation panel 40, the H level second halftone mode signal is output from the MPU 50.
From the second input terminal of the OR gate 78 in the image processing unit 20a and the selection signal input terminal SEL of the data selector 75.
Entered in. The signal output from the OR gate 78 is input to the selection signal input terminal SEL of the data selector 77 and the first input terminal of the AND gate 83.

The analog image signal output from the image sensor 26 is converted into a multivalued digital image signal by the A / D converter 71, and then the simple binarization unit 72 and the pseudo halftone binarization unit 73, 74 and the area separation unit 79. The simple binarization unit 72 performs simple binarization processing on the input multivalued digital image signal, and then outputs the processed binarized image data to the input terminals A of the data selectors 76 and 77. Output. Also, the pseudo halftone binarization unit 73
Performs binarization processing on the input multi-valued digital image signal using the error diffusion method or the random dither method, and then outputs the processed binarized image data to the input terminal A of the data selector 75. To do. Further, the pseudo halftone binarization unit 7
Numeral 4 performs binarization processing on the input multi-valued digital image signal by using the dot-concentrated systematic dither method, and then the processed binarized image data is input to the input terminal B of the data selector 75. Output. When the L level selection signal is input, the data selector 75 selects the binary image data input to the input terminal A and selects the data selector 76.
When the H level selection signal is input, the binarized image data input to the input terminal B is selected and output to the input terminal B of the data selector 76.

Further, the area separating section 79 discriminates, for each pixel, whether it is a photographic image or a character image based on the input multi-valued digital image signal, and when it is discriminated as a photographic image, an H level discrimination signal is sent to the data selector 76. Select signal input terminal SE
It outputs to the L and halftone pixel counting section 80, and when it determines that it is a character image, it also outputs an L level determination signal.
The data selector 76 is a binarized binary image when a L level selection signal is input in the character image area.
When the binarized image data is selected and output to the input terminal B of the data selector 77, while the H-level selection signal is input in the photographic image area, the binarization is performed in pseudo halftone. Image data is selected and similarly output. further,
When the L level selection signal is input, the data selector 77 selects simple binarized binarized image data and outputs it to the buffer memory 59, while the H level selection signal is input. If so, the binarized image data output from the data selector 76 is similarly output.

The halftone pixel counting section 80 counts the input H-level discrimination signal for each page, and outputs the count data to the input terminal A of the comparator 82 via the multiplier 81 having a multiplier of 8. To do. The comparator 82 compares the data input to the input terminal A with the data of the number of pixels of one page input to the input terminal B, and outputs an H level signal only when A> B. Output to the input terminal. Then, the restoration ON signal is output from the AND gate 83.

The image processing section 20a configured as described above.
When the first halftone mode or the second halftone mode is set using the key 543 of the operation panel 40 and the halftone area which is a photographic image exceeds 1/8 of the area of one page, the H level The restoration ON signal of is output to the MPU 50. The restoration on signal is generated in this manner because it is preferable not to perform image restoration when the halftone region is sufficiently small even if binarization is performed in the halftone mode.

(3) Image Restoration Processing Unit As shown in FIG. 7, the image restoration processing unit 62 includes a halftone image restoration unit 101 which restores multivalued halftone data from the received binary image data. However, this halftone image restoration processing has the following effects. That is, halftone image data such as a photographic image is generally represented by multi-valued image data having a plurality of bits per pixel. However, during image data communication such as facsimile communication or during image data storage such as filing, the multi-valued image data is binarized in pseudo-halftone and converted into binary image data so that communication should be performed or stored. It is possible to significantly reduce the amount of data to be processed.

This halftone image restoration processing is effective, for example, when halftone data binarized by pseudo halftone is binary-coded or displayed with different pixel densities. That is, it is possible to prevent the occurrence of moire due to the periodicity of the original pseudo-halftone binary image data by restoring the multi-valued image data and then performing the scaling process without simply performing the scaling process. .. The restored multi-valued image data is binarized in pseudo halftone and is output to an output system such as a display or a printer. At this time, if the output system is a device capable of processing input data at high density, the characteristics of the device can be fully utilized. Further, for example, the halftone image restoration process is performed when the binary image data binarized in pseudo halftone is restored to multivalued image data and is output to an output system such as a multivalued display or printer. It is valid.

FIG. 7 shows the image restoration processing unit 6 shown in FIG.
2 is a block diagram of FIG. As shown in FIG. 7, the binary image data serially read from the page memory 61 is
It is input to the 15 × 18 matrix memory circuit 100. The 15 × 18 matrix memory circuit 100 is shown in FIG.
As shown in, each pixel data D000 to D14 located at each position of the matrix in the 15 × 18 window
17 is generated, and the smoothing amount calculating unit 109, the edge emphasis amount calculating unit 110, the edge discriminating amount calculating unit 111 in the halftone image restoring unit 101, the adjacent state detecting unit 105 in the image region discriminating unit 102, and the systematic It is output to the halftone determination unit 106 and the simple multi-value quantization unit 103.

The halftone image restoration unit 101 includes a smoothing amount calculation unit 109, an edge enhancement amount calculation unit 110, an edge discrimination amount calculation unit 111, and a restored data calculation unit 112.
Here, the smoothing amount calculation unit 109 calculates and outputs three first to third smoothing amount data for restoring the halftone image data based on each pixel data of the received binary image data. The edge enhancement amount calculation unit 110 also calculates and outputs first to third edge enhancement amount data for performing edge enhancement processing based on each pixel data of the received binary image data. Further, the edge discrimination amount calculation unit 111
Calculates and outputs the first and second edge discrimination amounts used for discriminating the edge region based on each pixel data of the received binary image data. Restored data calculation unit 112
Is based on the data output from each unit 109 to 111, the concentrated type halftone determination signal and the systematic halftone determination signal output from the determination data signal generation unit 114, and the ultra fine signal output from the MPU 50. The halftone image data of the value is restored and output.

The image area discriminating unit 102 includes an adjacent state detecting unit 10
5, a systematic halftone judgment unit 106, a 9 × 17 matrix memory circuit 107, a judgment data generation unit 108,
And a discrimination data signal generator 114. Here, the adjacency state determination unit 105 is a main / sub-scan indicating the adjacency state in the four main / sub-scanning directions of the same type of pixels, which is the smaller number in a predetermined area, based on the input pixel data. The number of adjacent pixels in the direction is calculated, and the number of black pixels in a predetermined 7 × 7 window is calculated. Then, for each pixel, a non-dispersive halftone image detection signal indicating that the image of a predetermined area centered on the pixel of interest is a non-dispersive halftone image,
A distributed halftone image detection signal indicating that the image in the predetermined area is a distributed halftone image, and a total indicating whether the 7 × 7 window is an all-white image or an all-black image A black and white image detection signal is generated and output. On the other hand, the systematic halftone determination unit 106 detects, for each pixel, whether or not the image of a predetermined area centered on the pixel of interest is a systematic halftone image based on the input pixel data. And outputs a systematic halftone detection signal indicating the detection result. Further, the 9 × 17 matrix memory circuit 107 outputs four detection signals (total of 4 bits) serially output for each pixel from the adjacent state determination unit 105 and the systematic halftone determination unit 106 to the target pixel. Simultaneously output within a predetermined 9 × 17 window as the center.

The determination data generation unit 108 obtains each determination data obtained by adding the above-mentioned detection signals within the above 9 × 17 window for each detection signal based on each detection signal output from the matrix memory circuit 107. Is generated and output. Finally, the discrimination data signal generation unit 114 determines whether or not the image of the predetermined 9 × 17 window area is a concentrated halftone image based on each determination data output from the determination data generation unit 108. Is generated and a centralized halftone discrimination signal indicating the discrimination result is generated and output.
It determines whether or not the image in the area of the window of × 17 is a systematic halftone image, generates and outputs a systematic halftone determination signal indicating the determination result, and determines whether the image is a halftone region or a non-halftone region. Image area determination data indicating the result of determining whether the area is a tonal area is output. Here, the image area discrimination data becomes 0 when the image in the area is a completely halftone image,
On the other hand, it has a range of 0 to 1 which is 1 when the image is a non-halftone image.

Further, the simple multi-value quantization unit 103 binarizes the pixel data output from the matrix memory circuit 100 by binarizing in non-halftone using a predetermined threshold value.
The value image data is simply converted into multi-valued non-halftone image data indicating white or black and output to the data mixing unit 104 as non-halftone data.

Further, the data mixing unit 104, as shown in FIG. 19, includes a multiplier 281, a subtractor 282, a multiplier 283, an adder 284, and a data selector 285. When the restoration execution signal is at the H level, the data mixing unit 104 determines that the multi-valued halftone image data output from the halftone image restoration unit 101 and the multi-valued non-halftone output from the simple multi-value quantization unit 103. The following "Equation 1" is calculated based on the image data and the image area discrimination data, that is, these data are mixed according to the mixing ratio indicated by the image area discrimination data to obtain the multivalued image data. It is generated and output to the printer control unit 55 via the interpolation processing unit 64.

## EQU00001 ## Multivalued image data = (halftone image data) × {1- (image area discrimination data)} + (non-halftone image data) × (image area discrimination data)

On the other hand, when the restoration execution signal is at the L level, the data mixing unit 104 uses the data selector 285 to select the multivalued non-halftone image data output from the simple multivalue conversion unit 103. The multi-valued image data is output to the printer control unit 55 via the interpolation processing unit 64.

In the present embodiment, in order to make erroneous discrimination of the image area discrimination in a region which can be halftone or non-halftone inconspicuous,
As described above, the data mixing unit 104 mixes the halftone image data and the non-halftone image data according to the mixing ratio of the image area discrimination data indicating the halftone image-likeness and the non-halftone image-likeness and mixes them. The value image data is being restored.

(4) 15 × 18 Matrix Memory Circuit FIG. 8 is a block diagram of the 15 × 18 matrix memory circuit 100 shown in FIG. As shown in FIG.
The 5 × 18 matrix memory circuit 100 inputs an image based on a clock CLK having the same cycle as the transfer clock cycle of the binary image data input from the page memory 61, that is, a cycle of one dot of the image data. Fourteen FIFO memories DM1 to DM14 for delaying data by one horizontal period, which is one scanning time in the main scanning direction, and image data input respectively in synchronization with the clock CLK, for one cycle period of the clock CLK. (15 × 17) delay-type flip-flops DF001 to DF017, DF101 to DF117, DF201 to DF217 ,. ． ． , DF
1401 to DF1417.

The binary image data serially output from the page memory 61 in the direction from the first pixel to the last pixel of the image of each page is input to the flip-flop DF001, and is then cascade-connected. The data is output via the flip-flops DF001 to DF017, input to the FIFO memory DM1, and then output via the fourteen cascade-connected FIFO memories DM1 to DM14. The image data output from the FIFO memory DM1 is input to the flip-flop DF101 and then cascade-connected flip-flops DF101 to D
It is output via F117. In addition, the FIFO memory D
The image data output from M2 is the flip-flop D
After being input to F201, it is output via flip-flops DF201 to DF217 that are connected in cascade. Hereinafter, similarly, each of the FIFO memories DM3 to DM14 is similarly performed.
The image data output from the flip-flops DF301 to DF1401 are input to the flip-flops DF301 to DF1401, respectively, and are then cascaded.
317, DF401 to DF417 ,. ． ． , DF14
01 to DF1417 are output.

In the 15 × 18 matrix memory circuit 100 configured as described above, when the 1-dot pixel data first input to the circuit 100 is output from the flip-flop DF1417, the image input at that time is output. The data is output as pixel data D000, and the flip-flops DF001 to DF01 are also output.
7 outputs pixel data D001 to D017 on the main scanning line of i = 0 in each 15 × 18 window, and the FIFO memory DM1 and each flip-flop DF
101 to DF117 output pixel data D100 to D117 on the main scanning line of i = 1 in a 15 × 18 window, respectively, and FIFO memory DM2 and flip-flops DF201 to DF217 output i in a 15 × 18 window. The pixel data D200 to D217 on the main scanning line of = 2 are output, and in the same manner, the pixel data D300 to D1417 are output from the FIFO memories DM3 to DM14 and the flip-flops DF301 to DF1417, respectively.

(5) Image Area Discriminating Section (5-1) Configuration and Operation of Each Section FIGS. 9 to 14 are block diagrams of the image area discriminating section 102 shown in FIG. The adjacent state determination unit 105, the systematic halftone determination unit 106, the 9 × 17 matrix memory circuit 107, and the determination data generation unit 10
8 and the discrimination data signal generator 114. Less than,
The characteristics of the processing in the image area discrimination unit 102 will be described.

FIG. 20 shows an example of a non-halftone image obtained when the character image is read and then binarized by using a predetermined threshold value. FIG. 21 shows the uniform density chart. An example of a pseudo halftone binarized image obtained when binarized by the error diffusion method is shown. Further, FIG. 22 shows an example of a systematic halftone image obtained when a photographic image is read and then binarized by a dot-concentrated systematic dither method with a screen angle of 0 degree.

In this embodiment, the adjacency state determining unit 105 performs a process for determining whether the image of the input image data is a dispersed halftone image, while the input image data of the input image data is processed. The systematic halftone determination unit 106 performs a process for determining whether the image is a systematic halftone image. Further, the adjacent state determination unit 105 determines whether the image of FIG. 20 and the image of FIG. 21 are distinguished. In each window W7 in FIGS. 20 and 21, there are the same number of black pixels, and it can be said that the image densities of the images in the window W7 are the same. A major difference between the images in the window W7 is the adjacency state of a small number of pixels in the main scanning direction and the sub scanning direction (hereinafter referred to as the main sub scanning direction). Here, the minority pixel means a pixel having a smaller number of white pixels and black pixels in a predetermined window, and in the example of FIGS. 20 and 21,
The minority pixels are black pixels.

Further, the total number of the minority pixels connected by the same kind of pixel as the minority pixel in any of the four directions from the minority pixel to the main / sub scanning direction is hereinafter referred to as the number of adjacencies in the four directions.
In the present embodiment, as shown in FIG. 23, the number of adjacencies in the main scanning direction is counted, and as shown in FIG. 24, the number of adjacencies in the sub scanning direction is counted. Generally, in a graph of the number of adjacencies in four main and sub-scanning directions with respect to the number of black pixels in a 7 × 7 window, a halftone image area that is a dispersed halftone image and a non-halftone image that is a non-dispersed halftone image. The toned image area is divided and shown as shown in FIG. Figure 25
As is apparent from the above, when the number of black pixels and the number of white pixels are equal in a predetermined window, the number of adjacent pixels in the four directions, which indicates the threshold value on the boundary line of each image area, becomes larger than the threshold value. Also has a non-halftone image when the number of adjacencies in four directions is large, and a non-halftone image exists when the number of adjacencies in four directions is smaller than the threshold value.

However, when the resolution at the time of reading an image becomes higher, the threshold value of the number of adjacent pixels in the four directions also increases as shown in FIG. 26 for the following reason. (A) A steep edge portion of the dispersed halftone region may be erroneously determined as a non-dispersed halftone image. However, for example, an image having a sharp edge when read in the fine mode becomes an image with a gentle edge when read in the ultra fine mode. Therefore, it can be said that the higher the resolution is, the smaller the sharp edge portion is. (B) A thin line image is likely to be a broken line image, and the broken line is discriminated as a dispersed halftone image. However, a thin line image read at a relatively low resolution also becomes a straight line having a sufficiently large width when read at a relatively high resolution. Therefore, the higher the resolution is read, the less such broken line images are.

Therefore, in the present embodiment, two thresholds, that is, the first threshold for fine mode and the second threshold for ultra fine are used as the thresholds for the number of adjacencies in the four directions. Table ROM for which the former is prepared and the former is shown in FIG.
123 and stores the latter in the table ROM 1
It stores in 24. Then, the comparator 127 sets the number of adjacencies in four directions, that is, the number of adjacencies in the main and sub-scanning directions, and 0 when the threshold value or the concentrated halftone signal output from the table ROM 123 or 124 is at the H level. By comparing, a region determination of a dispersed halftone image or a non-dispersed halftone image is performed.

Next, the method of judgment in the systematic halftone judging section 106 for judging the systematic halftone image shown in FIG. 11 will be described. The systematic halftone image can be discriminated from the periodicity of the image, and in this embodiment, the pattern matching method using the five windows W4a to W4d shown in FIG. 27 is used.

In FIG. 27, 4 × 4 including the pixel of interest *
Window W4a and four four windows adjacent to the window W4a in the first and second diagonal directions.
By matching the image patterns with the × 4 windows W4b to W4d and counting the number of pixels in which the pixel data do not match, the halftone discrimination value for the concentrated halftone image can be obtained. Here, the first diagonal direction means a diagonal direction from the upper right to the lower left which is inclined by 45 degrees with respect to both the main scanning direction and the sub scanning direction, and the second diagonal direction means the main scanning. Direction from the upper left to the lower right, which is inclined by 45 degrees with respect to both the scanning direction and the sub-scanning direction. The above four windows W
4b to W4d are used to simultaneously support two types of dot-concentrated systematic dither images with screen angles of 0 degree and 45 degrees.

Now, for example, the image pattern PA of FIG.
Table 1 shows the results of counting the number of non-matching pixels for T1, the image pattern PAT2 in FIG. 29, and the image pattern PAT3 in FIG. 30 using the above pattern matching method. Here, the image pattern PAT1 is a non-halftone image, and the image pattern PAT2 is binarized in pseudo-halftone using the dot-concentrated systematic dither method in which the screen angle is 0 degree and one period is 4 pixels. The image pattern PAT3 is a systematic halftone image obtained by binarizing the image pattern PAT3 with pseudo halftones using a dot-concentrated systematic dither method with a screen angle of 45 degrees and one period of 4 pixels. It is an image.

[0082]

[Table 1]

As is clear from Table 1, in the case of the image pattern PAT1 of FIG. 28, which is a non-halftone image pattern, the number of non-matching pixels has a large value. It is clear that it is possible to distinguish systematic halftone images such as 30.

(5-2) Adjacent State Determination Unit FIG. 9 is a block diagram of the adjacent state determination unit 105. In this embodiment, the number of adjoining minority pixels in the four main- and sub-scanning directions is obtained by counting the number of adjoining portions between the pixels indicated by the arrows in FIGS. Here, in order to count the number of adjacencies on each one scanning line in the main scanning direction or the sub scanning direction, the adjacency number counting circuit 120 shown in FIG. 10 is used.

As shown in FIG. 9, the 7 × 7 black pixel number counting circuit 121 counts the number of black pixels in the 7 × 7 window centered on the pixel of interest based on the input binary pixel data. Then, the count data is output to the input terminal A of the comparator 122, the address terminals of the table ROMs 123 and 124, and the input terminals A of the comparators 128 and 129. In response to this, the table ROM 123 outputs to the input terminal A of the data selector 125 the first threshold value data for fine mode corresponding to the input count data of the number of black pixels, while the table ROM 124. Outputs to the input terminal B of the data selector 125 the second threshold value data for the ultra fine mode corresponding to the input count data of the number of black pixels. Further, the data selector 125
When the ultrafine signal input as the selection signal is at the L level, the first threshold value data is selected and output to the input terminal B of the comparator 127 via the clear circuit 126, while the ultrafine signal is at the H level. When it is at the level, the second threshold value data is selected and similarly output.

On the other hand, the comparator 122 determines whether the black pixel or the white pixel is the minority pixel by comparing the count data of the number of black pixels and the data “24”, and the minority pixel is When the pixel is a white pixel, the H-level discrimination signal C122 is output to the adjacent number counting circuit 120, while when the minority pixel is a black pixel, the L-level discrimination signal C122 is similarly output.

As shown in FIG. 10, the adjacency number counting circuit 120 includes a black pixel adjacency number counting circuit 3 for each black pixel and white pixel.
01 and the white pixel adjacent number counting circuit 302,
The number of adjacencies in the four directions shown in FIGS. 23 and 24 is counted, and the data selector 303 obtains the adjacency number data in the four directions of the minority pixels using the discrimination signal C122 indicating the minority pixels. The adjacent number data is the input terminal A of the comparator 127.
Entered in. The comparator 127 compares the number-of-adjacent data with the threshold data, and when A> B, the non-dispersive halftone detection signal J-A of H level and the dispersion halftone detection signal J of L level. -B and 9X17 matrix memory circuit 107, and when A <B, L-level non-dispersive halftone detection signal JA and H-level dispersed halftone detection signal J-B. Is output in the same manner, and A =
In the case of B, the L-level non-dispersive halftone detection signal JA and the L-level dispersed halftone detection signal J-B are similarly output. When the received binary image data is the image data binarized by pseudo halftone using the dot-concentrated systematic dither method, that is, the concentrated-type halftone signal is H. When at level
The clear circuit 126 clears the threshold value data to 0, which causes the distributed halftone detection signal J-B to go to L level.
It becomes a level.

On the other hand, the comparator 128 compares the count data of the number of black pixels with the data "0", and when they are equal, outputs an H level signal to the OR gate 130, and the comparator 129 compares the count data of the number of black pixels with the count data. Data "49" is compared, and when they are equal, an H level signal is output to the OR gate 130.
The OR gate 130 outputs an H level all-black / all-white detection signal J indicating that the 7 × 7 window is all black pixels or white pixels when any of the input signals is at H level.
-C is output to the 9 × 17 matrix memory circuit 107.

(5-3) Systematic Halftone Determining Section FIG. 11 is a block diagram of the systematic halftone determining section 106 shown in FIG. As shown in FIG. 11, the non-matching pixel counting circuit 140 has the window W4 shown in FIG.
The number of non-matching pixels between a and the window W4b is counted, and the non-matching pixel counting circuit 141 counts the number of non-matching pixels between the window W4a and the window W4c shown in FIG. Further, the non-matching pixel counting circuit 142 uses the window W4a shown in FIG.
The number of non-matching pixels between the window W4d and the window W4d is counted, and the non-matching pixel counting circuit 143 counts the number of non-matching pixels between the window W4a and the window W4e shown in FIG. Each counting circuit 140 to 1
The count data counted in 43 is added by the adders 144 to 14
After being added by 6, it is input to the input terminal A of the comparator 147. The comparator 147 compares the input addition data with the predetermined threshold value data J2, and A <B
At this time, an H level comparison result signal is output to the first input terminal of the AND gate 149. In this embodiment, the threshold value data J2 input to the comparator 147 is preferably 15.

On the other hand, the random halftone signal is NOR gate 1.
The ultrafine signal is input to the first input terminal of the NOR gate 48, and the ultrafine signal is input to the second inverting input terminal of the NOR gate 148. The output signal of the NOR gate 148 is the AND gate 14
9 is input to the second input terminal, and the systematic halftone detection signal J-D is output from the AND gate 149.

When the number of non-matching pixels counted and added by each of the counting circuits 140 to 143 is smaller than the threshold value data J2 in the systematic halftone judging section 106 configured as described above, systematic halftone determination is performed. It is determined that the image is a halftone image, and the H-level systematic halftone detection signal J-D is output. However, in the unique mode, when the received binary image data is binary image data binarized by pseudo halftone using the random dither method, the NOR gate 148 and the AND gate 149 detect the systematic halftone. The signal J-D is set to L level. This is for the following reason.
The systematic halftone discrimination method using the pattern matching method is
It is assumed that there is no large density change between two windows that are adjacent and perform matching processing. However, if the resolution at the time of reading is reduced, the assumption is no longer satisfied, and in this case, the systematic halftone determination unit 106 erroneously determines a systematic halftone image to be a non-systematic halftone. This is because the possibility increases.

(5-4) 9 × 17 Matrix Memory Circuit FIG. 12 is a block diagram of the 9 × 17 matrix memory circuit 107 shown in FIG.

As shown in FIG. 12, the 9 × 17 matrix memory circuit 107 has the same cycle as the cycle of the transfer clock of the binary image data input from the page memory 61, that is, the cycle of one dot of the image data. The 4-bit determination data composed of the following four detection signals detected and input for each pixel in synchronization with the clock CLK is delayed by one horizontal period which is one scanning time in the main scanning direction 8 FIFO memories DM11 to DM1
8 and four (9 × 16) delay type flip-flops DG001 to DG016, DG101 to DG116 which delay and output 4-bit determination data input in synchronization with the clock CLK by one cycle period of the clock CLK. , DG201 to DG216 ,. ． ． , DG
801 to DG816. In each circuit of the matrix memory circuit 107, the following 4-bit determination data is processed in parallel. (A) A non-dispersive halftone detection signal JA (hereinafter referred to as determination data JA) output from the adjacent state determination unit 105. (B) Distributed halftone detection signal J-B (hereinafter referred to as judgment data J-B) output from the adjacent state judgment unit 105. (C) All-black / all-white detection signal J-C output from the adjacent state determination unit 105 (hereinafter, referred to as determination data J-C). (D) Systematic halftone detection signal J-D output from the systematic halftone determination unit 106 (hereinafter referred to as determination data J-D).

The 4-bit determination data serially output from the determination units 105 and 106 in the direction from the first pixel to the last pixel of the image on each page is cascaded after being input to the flip-flop DG001. The eight flip-flops DM11 to DG016, which are connected to each other, are input to the FIFO memory DM11 and then cascade-connected to the eight FIFO memories DM11.
Through DM18. FIFO memory DM
The image data output from 11 is the flip-flop D.
After being input to G101, it is output via the flip-flops DG101 to DG116 connected in cascade. The image data output from the FIFO memory DM12 is input to the flip-flop DG201 and then cascade-connected flip-flops DG201 to DG210.
Is output via. Hereinafter, in the same manner, the determination data output from each of the FIFO memories DM13 to DM18 are respectively flip-flops DG301 to DG801.
To the flip-flops DG301 to DG316 ,. ． ． , DG801 to DG816.

In the 9 × 17 matrix memory circuit 107 configured as described above, 4-bit determination data corresponding to pixel data for 1 dot first input to the circuit 107 is output from the flip-flop DG816. At this time, the determination data input at that time is output as 4-bit determination data J000, and each of the flip-flops DG001 to DG016 outputs pixel data on the main scanning line of i = 0 in the 9 × 17 window. Corresponding determination data J001 to J016 are output, and each determination data J1 on the main scanning line of i = 1 in the 9 × 17 window from the FIFO memory DM11 and each flip-flop DG101 to DG116.
00 to J116 are output, and the FIFO memory DM12 is output.
And the respective flip-flops DG201 to DG216 output the respective determination data J200 to J216 on the main scanning line of i = 2 in the 9 × 17 window, respectively. To DG816 output determination data J300 to J816, respectively.

Therefore, as shown in FIG. 13, each pixel (i = 0, 1, 2, ..., 8; j =
0, 1, 2, ..., 16) corresponding to 4 bits of judgment data Jij-A to Jij-D per pixel from the matrix memory circuit 107 at the same time.
It is output to 08.

(5-5) Judgment Data Generation Unit FIG. 13 is a block diagram of the judgment data counting unit 108 shown in FIG. As shown in FIG. 13, the determination data Jij− output from the matrix memory circuit 107.
A to Jij-D are input to the counting circuits 160 to 163 for each determination data, and the counting circuits 160 to 163 have the H level (“1”) of the input determination data.
Count the number of data in the 9 × 17 window,
The count data of the determination data JA to JD below is output to the determination data signal generation unit 114. (A) Count data of determination data JA: number of non-dispersive halftone determination pixels. (B) Count data of the judgment data J-B: number of dispersed halftone judgment pixels. (C) Count data of determination data J-C: the number of all black and all white pixels. (D) Count data of determination data J-D: number of systematic halftone determination pixels.

(5-6) Discrimination Data Signal Generation Unit FIG. 14 shows the discrimination data signal generation unit 114 shown in FIG.
It is a block diagram of.

As shown in FIG. 14, the number of non-dispersive halftone discrimination pixels JS-A is the first address terminal of the table ROM 172 for outputting the non-halftone index for the dispersed halftone image, and the adder. 170 input terminal A and comparator 1
71 is input to the input terminal A of 71, and the number of dispersed halftone discrimination pixels JS-B is input to the input terminal B of the adder 170 and the comparator 17.
1 is input to the input terminal B. The adder 170 adds the two input data and then adds the data to the table ROM 172.
To the second address terminal of. The comparator 171 is A
When <B, the H-level distributed halftone discrimination signal is output to the first inverting input terminal of the AND gate 179.

Also, the data "153" of the total number of pixels in the 9 × 17 window is input to the input terminal A of the subtractor 173, and the total black and white pixel number JS-C is input to the input terminal B of the subtractor 173. Input the systematic halftone discrimination pixel number JS-D,
It is input to the second address terminal of the table ROM 174 that outputs a non-halftone index indicating the likelihood of non-halftone with respect to the systematic halftone image, and is also input to the comparator 177 via the multiplier 176 having a multiplier of 3. Input to terminal B. The subtractor 173 subtracts the number of all-black / all-white pixels JS-C from the data “153” to obtain the subtraction result data as a table R.
While outputting to the first address terminal of OM174,
It outputs to the input terminal A of the comparator 177 via the multiplier 175 whose multiplier is 2. When A> B, the comparator 177 outputs the H-level systematic halftone discrimination signal to the restored data calculation unit 112 and also to the second input terminal of the AND gate 179. Systematic halftone discrimination signal is H
When the level is the level and the distributed halftone determination signal is at the L level, the AND gate 179 outputs the H level concentrated type halftone determination signal to the restored data calculation unit 112.

FIG. 31 is a graph of a non-halftone index stored in the table ROM 172, showing a non-halftone index for a distributed halftone image, and FIG. 32 is a systematic table stored in the table ROM 174. 3 is a graph of non-halftone index for a halftone image. Here, the data x1 and x2 on the horizontal axis of each graph are represented by the following "Equation 2" and "Equation 3".

## EQU00002 ## x1 = (number of non-dispersive halftone discrimination pixels) / {(number of non-dispersion halftone discrimination pixels) + (number of dispersed halftone discrimination pixels)}

## EQU00003 ## x2 = (systematic halftone discrimination pixel number) / {(total pixel number)-
(Number of detected pixels for all white and all black images)}

The denominator of the data x1 is calculated by the adder 170, and the denominator of the data x2 is calculated by the subtractor 173. As is apparent from FIG. 31, the non-halftone index y indicating the non-halftone likelihood for the dispersed halftone image.
1 has the following values for the data x1.
(A) When 0 ≦ x1 ≦ 0.5, y1 = 0, (b) 0.
When 5 <x1 ≦ 0.8, y1 = 2 × x1-1, (c)
When x1> 0.8, y1 = 1.

Further, as is apparent from FIG. 32, the non-halftone index y2, which indicates the likelihood of non-halftone with respect to the systematic halftone image, has the following values for the data x2.
(A) When 0 ≦ x2 ≦ 2/3, y2 = 0, (b) x2
When> 2/3, y2 = 2 × x2-1.

For convenience of explanation, the non-halftone index y1,
Although the value of y2 is a value from 0 to 1, the non-halftone indexes y1 and y2 are represented by 4-bit data in the circuit of FIG.

The table ROM 172 obtains the non-halftone index for the dispersed halftone image from the stored table based on the data JS-A input to the address terminal and the output data of the adder 170, and thereafter, Comparison selector 1
It outputs to the 1st input terminal of 78. Also, R for table
The OM 174 obtains the non-halftone index for the systematic halftone image from the stored table based on the data JS-D input to the address terminal and the output data of the subtractor 173, and then the comparison selector 178 outputs the non-halftone index. Output to the second input terminal. The comparison / selection unit 178 selects the maximum data from the respective inputted halftone indexes, and then outputs the selected data to the data mixing unit 104 as image area discrimination data.

In the discrimination data signal generator 114 having the above-mentioned configuration, the comparator 171 is used to execute the JS-
If A <JS-B, that is, x1 <0.5, the dispersed halftone discrimination signal is set to the H level. In addition, the multiplier 17
5, 176 and the comparator 177, 3 (JS-D)
<2 {total number of pixels- (JS-C)}, that is, x2 <2/3
Then, the systematic halftone discrimination signal is set to the H level. Further, since the concentrated type halftone image is an image that is not a random halftone image among the systematic halftone images, the concentrated type halftone discrimination signal is obtained using the AND gate 179.

(6) Halftone Image Restoring Unit (6-1) Configuration and Operation of Each Unit The halftone image restoring unit 101 shown in FIG. The discriminant amount calculation unit 111 and the restored data calculation unit 112 are provided.
The features of the processing in each of the calculation units 109 to 112 will be described below.

The halftone image restoration section 101 carries out a process of restoring binary image data binarized by pseudo halftone into multivalued image data close to the original photographic image. First, restoration of a dispersed halftone image to multi-valued image data will be described. In order to restore multi-valued image data, it is necessary to refer to pixel values around the pixel of interest. Taking into account that most of the pseudo halftone binarization methods are area gradation methods, the smoothing amount calculation unit 109 uses the smooth spatial filters F1 to F3 shown in FIGS. It is possible to restore binary image data to multi-valued image data.

However, the above smooth spatial filter F
If the image restoration processing is performed using only 1 to F3, there is a high possibility that the high-frequency spatial component of the original binary image data will disappear. Therefore, in the present embodiment, in the edge emphasis amount calculation unit 110, the edge amount detection spatial filter F which is the second derivative filter shown in FIGS. 36 to 47.
11 to F14, F21 to F24, F31 to F34
Is used to detect the disappearing high-frequency spatial frequency component and include it in the multi-valued image data.

However, these second-order differential filters may detect a pseudo halftone texture as an edge amount. That is, the texture is buried in the original edge in the so-called edge area where the edge is steep, but in the area where the edge is gentle, the texture may appear as an edge amount larger than the original edge amount. In order to solve this, in the edge discriminant amount calculation unit 111, the edge amount detection spatial filters F41 to F which are the first-order differential filters shown in FIGS.
Using 44 and F51 to F54, it is determined whether or not it is the edge region. Further, in the restored data calculation unit 112, the edge enhancement using the above-mentioned second derivative filter is performed only for the pixels which are determined to be the edge region based on the edge determination amount calculated by the edge determination amount calculation unit 111.

The edge amount detection spatial filter is provided for the four main and sub scanning directions in order to detect the edge amounts in a plurality of directions and obtain the largest edge amount among them. Further, in the edge amount detection spatial filter, the edge amount is calculated by using windows of plural kinds of sizes in order to detect the edge amounts of a plurality of spatial frequency components and obtain the largest edge amount from them. Is. However, the size of each window used in the smoothing spatial filter and the edge amount detecting spatial filter depends on the resolution at the time of reading a document in the facsimile apparatus on the transmission side. That is, even if the same original is read, if the resolution at the time of reading is high, the low frequency component becomes large in the binary image data, and if the resolution is low, the high frequency component becomes large. Therefore, in this embodiment, the spatial filter used is changed as follows in the fine mode and the ultra fine mode. (A) Spatial filter used in fine mode: F
1, F11 to F14, F21 to F24, F41 to F44. (B) Spatial filters used in ultra fine mode: F2, F11 to F14, F21 to F24, F
31 to F34, F41 to F44, F51 to F5
4.

Image restoration for a systematic halftone image will be described below. The size of one side of each window of the smooth spatial filters F1 to F3 used in the smoothing amount calculation unit 109 is set to an integral multiple of the size of one side of the threshold matrix of systematic dither so that moire does not occur. Must be set. Therefore, in this embodiment, since the target systematic dither uses a threshold value of 4 × 4 or 8 × 8, the 8 × 8 smooth spatial filter F3 of FIG. 35 is further used. Further, in the edge enhancement processing, the processing method is different between the distributed halftone image and the concentrated halftone image among the systematic halftone images. Since the spatial frequency of the texture in the concentrated halftone image is low, the size of the window for performing the edge enhancement process becomes too large. Therefore, in this embodiment, the edge enhancement is not performed on the concentrated type halftone image. The distributed halftone image of the systematic halftone image is subjected to the same edge enhancement processing as that of the dispersed halftone image.

(6-2) Smoothing Quantity Calculating Section FIG. 15 is a block diagram of the smoothing quantity calculating section 109 shown in FIG.

As shown in FIG. 15, the smooth spatial filter F
1 is output from the matrix memory circuit 100 2
The smoothed amount is calculated based on the value pixel data, and the calculated smoothed amount data is used as the first smoothed amount data through the multiplier 311 having a multiplier of 63/49 to be the restored data calculation unit 112.
Output to. The smoothing space filter F2 calculates a smoothing amount based on the binary pixel data output from the matrix memory circuit 100, and the calculated smoothing amount data is
The second smoothed amount data is output to the restored data calculation unit 112 via the multiplier 312 having a multiplier of 3/81. Further, the smooth spatial filter F3 calculates a smoothing amount based on the binary pixel data output from the matrix memory circuit 100, and the calculated smoothing amount data is passed through the multiplier 313 having a multiplier of 63/64 to the first smoothing amount data. 3 smoothed amount data is output to the restored data calculation unit 112. The multiplier 3
11 to 313 are provided to correct the difference in smoothing amount caused by the difference in window size between the smooth spatial filters F1 to F3.

(6-3) Edge Enhancement Amount Calculation Unit FIG. 16 is a block diagram of the edge enhancement amount calculation unit 110 shown in FIG. As shown in FIG. 16, each of the edge amount detection spatial filters F11 and F12 calculates the edge emphasis amount based on the binary pixel data output from the matrix memory circuit 100, and the absolute value comparison selector 321.
Output to. The absolute value comparison selector 321 compares the absolute values of the two input edge enhancement amounts and selects the maximum edge enhancement amount, and then outputs it to the absolute value comparison selector 323. In addition, each of the edge amount detection spatial filters F13 and F14 calculates the edge emphasis amount based on the binary pixel data output from the matrix memory circuit 100, and outputs the edge emphasis amount to the absolute value comparison selector 322. Absolute value comparison selector 322
Compares the absolute values of the two input edge enhancement amounts and selects the maximum edge enhancement amount, and then the absolute value comparison selector 32
Output to 3. The absolute value comparison selector 323 compares the absolute values of the two input edge enhancement amounts and selects the maximum edge enhancement amount, and then outputs the first edge enhancement amount data via the multiplier 324 whose multiplier is 3. As the restored data calculation unit 11
Output to 2.

Edge amount detection spatial filters F21, F2
2 is 2 output from the matrix memory circuit 100.
The edge emphasis amount is calculated based on the value pixel data, and is output to the absolute value comparison selector 331. The absolute value comparison selector 331 compares the absolute values of the two input edge enhancement amounts and selects the maximum edge enhancement amount, and then outputs it to the absolute value comparison selector 333. The edge amount detection spatial filters F23 and F24 are arranged in the matrix memory circuit 10
The edge enhancement amount is calculated based on the binary pixel data output from 0, and is output to the absolute value comparison selector 332. The absolute value comparison selector 332 compares the absolute values of the two input edge emphasis amounts, selects the maximum edge emphasis amount, and then outputs it to the absolute value comparison selector 333. The absolute value comparison selector 333 compares the absolute values of the two input edge enhancement amounts, selects the maximum edge enhancement amount, and then outputs the second edge enhancement amount via the multiplier 334 whose multiplier is 3/2. The quantity data is output to the restored data calculation unit 112.

Edge amount detection spatial filters F31, F3
2 is 2 output from the matrix memory circuit 100.
The edge emphasis amount is calculated based on the value pixel data, and is output to the absolute value comparison selector 341. The absolute value comparison selector 341 compares the absolute values of the two input edge enhancement amounts and selects the maximum edge enhancement amount, and then outputs it to the absolute value comparison selector 343. The edge amount detection spatial filters F33 and F34 are provided in the matrix memory circuit 10 respectively.
The edge enhancement amount is calculated based on the binary pixel data output from 0, and is output to the absolute value comparison selector 342. The absolute value comparison selector 342 compares the absolute values of the two input edge enhancement amounts and selects the maximum edge enhancement amount, and then outputs it to the absolute value comparison selector 343. The absolute value comparison selector 343 compares the absolute values of the two input edge enhancement amounts, selects the maximum edge enhancement amount, and then outputs it as the third edge enhancement amount data to the restored data calculation unit 112.

The multipliers 324 and 334 have the edge amount detection spatial filters F11 to F14 and F21 to F2, respectively.
It is provided to correct the difference in edge enhancement amount caused by the difference in window size between F4 and F31 to F34.

(6-4) Edge Discrimination Amount Calculation Unit FIG. 17 is a block diagram of the edge discrimination amount calculation unit 111 shown in FIG. As shown in FIG. 17, each of the edge amount detection spatial filters F41 and F42 calculates the edge determination amount based on the binary pixel data output from the matrix memory circuit 100, and calculates the absolute value of the input data. It outputs to the comparison selector 353 via each absolute value circuit 351 and 352 which outputs. The comparison selector 353 compares the two input edge discrimination amounts and selects the maximum edge discrimination amount, and then outputs it to the comparison selector 357. The edge amount detection spatial filters F43 and F44 respectively calculate the edge determination amount based on the binary pixel data output from the matrix memory circuit 100, and the comparison selector 356 via the absolute value circuits 354 and 355. Output to. The comparison selector 356 compares the two input edge discrimination amounts and selects the maximum edge discrimination amount, and then outputs it to the comparison selector 357. The comparison selector 357 compares two input edge discrimination amounts and selects the maximum edge discrimination amount, and then outputs the first edge discrimination amount via the multiplier 358 having a multiplier of 6/4.
The data is output to the restored data calculation unit 112 as edge discrimination amount data.

Edge amount detection spatial filters F51, F5
2 is 2 output from the matrix memory circuit 100.
The edge discrimination amount is calculated based on the value pixel data, and the absolute value of the input data is calculated and output to the comparison selector 363 via the absolute value circuits 361 and 362. The comparison selector 363 compares the two input edge discrimination amounts and selects the maximum edge discrimination amount, and then outputs it to the comparison selector 367. The edge amount detection spatial filters F53 and F54 are arranged in the matrix memory circuit 100.
The edge discrimination amount is calculated on the basis of the binary pixel data output from, and is output to the comparison selector 366 via the absolute value circuits 364 and 365. The comparison selector 366 compares the two input edge discrimination amounts and selects the maximum edge discrimination amount, and then outputs it to the comparison selector 367. The comparison selector 367 compares the two input edge discrimination amounts and selects the maximum edge discrimination amount, and then outputs the second edge discrimination amount data to the restored data calculation unit 112.

The multiplier 358 is provided to correct the difference in edge discrimination amount caused by the difference in window size between the edge amount detection spatial filters F41 to F44 and F51 to F54.

(6-5) Restored Data Calculation Unit FIG. 18 is a block diagram of the restored data calculation unit 112 shown in FIG. As shown in FIG. 18, the first and second smoothed amount data are input to the data selector 251, and the data selector 251 selects the first smoothed amount data when the ultrafine signal is at the L level and selects the data selector 252.
On the other hand, the second smoothing amount data is selected and output to the input terminal A of the data selector 252 when the ultrafine signal is at H level. Data selector 2
52 selects the first or second smoothing amount data input to the input terminal A when the systematic halftone discrimination signal is at the L level and outputs it as the smoothing component amount to the adder 261; When the key discrimination signal is at the H level, the third smoothing amount data input to the input terminal B is selected and output to the adder 261 as the smoothing component amount.

The absolute comparison selector 253 receives the first input
The edge emphasis amount data having a larger absolute value is selected from the second edge emphasis amount data and output to the absolute value comparison selector 256. The ultrafine signal is input to the clear circuit 254 via the inverter 255, and when the inversion signal of the ultrafine signal is at the L level, the clear circuit 254 inputs the third edge emphasis amount data that is input to the absolute value comparison selector 256 as it is. On the other hand, when the inverted signal of the ultrafine signal is at the H level, the input third edge emphasis amount data is cleared and data "0" is output to the absolute value comparison selector 256. The absolute value comparison selector 256 selects the edge emphasis amount data having a larger absolute value from the two pieces of input edge emphasis amount data, and outputs it to the adder 261 via the clear circuit 257.

The first and second edge discrimination amount data are input to the data selector 259, and the data selector 259
When the ultrafine signal is at the L level, the first edge discrimination amount data is selected and output to the input terminal A of the comparator 260, while when the ultrafine signal is at the H level, the second edge determination amount data is output.
The edge discrimination amount data is selected and output to the input terminal A of the comparator 260. The predetermined threshold value data J3 is input to the input terminal B of the comparator 260, and the comparator 260 outputs A
When <B, the H level comparison result signal is sent to the OR gate 258.
It outputs to the clear circuit 257 via. In the present embodiment, the threshold value data J3 is preferably 4. On the other hand, the centralized halftone discrimination signal is input to the clear circuit 257 via the OR gate 258, and the clear circuit 257
When the signal input from the OR gate 258 is L level, the data input from the absolute value comparison selector 256 is output to the adder 261 as it is, while the OR gate 258 outputs the data.
When the signal input from is at H level, the data input from the absolute value comparison selector 256 is cleared and the data “0” is output to the adder 261.

Further, the adder 261 adds the input smoothing component amount and edge enhancement component amount to add the limiter circuit 26.
2, the limiter circuit 262 rounds the input data into data having values 0 to 63, and then outputs the data as halftone image data to the data mixing unit 104.

[0126]

As described above in detail, according to the image processing apparatus of the first aspect of the present invention, binary image data binarized by pseudo halftone and a predetermined threshold value are used. Input 2 including non-halftone binary image data
An arithmetic unit that calculates the number of adjacencies indicating the adjacency state of pixels of the same type in a block composed of a plurality of pixels including the target pixel to be processed based on the value image data, and the adjacency state calculated by the arithmetic unit are shown. An image provided with a discriminating means for discriminating whether the inputted binary image data is pseudo-halftone binary image data or non-halftone binary image data based on the number of adjacent pixels. In the processing device,
The discrimination means makes the discrimination by comparing the number of adjacencies indicating the adjacency state with a predetermined threshold value, and changes the threshold value based on the resolution of the binary image data. Therefore, a ROM can be used and a simple circuit can be used as compared with the conventional method or apparatus.
Even when the resolution of the binary image data is changed, each data can be discriminated at high speed and more accurately.

The image processing apparatus according to claim 2 is the image processing apparatus according to claim 1, further comprising:
Since the restoration means for restoring the input binary image data to the multi-gradation image data based on the result determined by the determination means is provided, the restoration processing is performed based on the more accurate image area determination result. be able to. Therefore, when restoring input binary image data to multi-valued image data,
Even when the resolution of the binary image data changes, the image reproducibility can be improved and image deterioration can be prevented as compared with the conventional example.

Further, in the image processing apparatus according to the third aspect, when the input binary image data is restored to the multi-valued image data, even if the resolution of the binary image data changes, the above-mentioned By changing the size of the window of the spatial filter as described above, the spatial frequency component originally possessed by the image can be preserved, the reproducibility of the image is improved compared with the conventional example, and the deterioration of the image is prevented. be able to.

Further, in the image processing apparatus according to the fourth aspect, the binary image data is binarized by the dot-concentrated systematic dither method based on the input binary image data. If the resolution of the input binary image data is less than or equal to a predetermined value, it is determined that the binary image data has not been binarized by the dot-concentrated systematic dither method. Therefore, even if the resolution of the binary image data changes, the spatial frequency component originally possessed by the image can be preserved, the reproducibility of the image is improved as compared with the conventional example, and the deterioration of the image is prevented. can do.

[Brief description of drawings]

FIG. 1 is a vertical cross-sectional view of a mechanical portion of a facsimile apparatus according to an embodiment of the present invention.

FIG. 2 is a front view of an operation panel of the facsimile device shown in FIG.

FIG. 3 is a block diagram of a signal processing unit of the facsimile apparatus shown in FIG.

4 is a first flowchart of a control flow of a control signal setting process executed by an MPU in the signal processing unit of the facsimile device shown in FIG.

5 is a second flowchart of a control flow of a control signal setting process executed by the MPU in the signal processing unit of the facsimile device shown in FIG.

6 is a block diagram of an image processing unit in the signal processing unit of the facsimile apparatus shown in FIG.

7 is a block diagram of an image restoration processing unit illustrated in FIG.

FIG. 8 is a block diagram of the 15 × 18 matrix memory circuit shown in FIG. 7.

9 is a block diagram of an adjacent state determination unit illustrated in FIG.

10 is a block diagram of the adjacent number counting circuit shown in FIG. 9. FIG.

11 is a block diagram of a systematic halftone judging section shown in FIG. 7. FIG.

12 is a block diagram of the 9 × 17 matrix memory circuit shown in FIG. 7. FIG.

13 is a block diagram of a determination data generation unit illustrated in FIG.

14 is a block diagram of a discrimination data signal generator shown in FIG. 7. FIG.

FIG. 15 is a block diagram of a smoothing amount calculation unit illustrated in FIG. 7.

16 is a block diagram of an edge enhancement amount calculation unit shown in FIG. 7. FIG.

17 is a block diagram of an edge discrimination amount calculation unit illustrated in FIG. 7. FIG.

18 is a block diagram of a restored data calculator shown in FIG. 7. FIG.

FIG. 19 is a block diagram of a data mixing unit illustrated in FIG. 7.

FIG. 20 is a diagram showing an example of a non-halftone image obtained when a character image is read and then binarized using a predetermined threshold value.

FIG. 21 is a diagram showing an example of a pseudo halftone binarized image obtained when the uniform density chart is read and then binarized by the error diffusion method.

FIG. 22 is a diagram showing an example of a pseudo-halftone binarized image obtained when a photographic image is read and then binarized by a dot-concentrated systematic dither method with a screen angle of 0 °.

FIG. 23 is a diagram showing adjacency in the main scanning direction at each pixel in a 7 × 7 window.

FIG. 24 is a diagram showing adjacency in the sub-scanning direction of each pixel in a 7 × 7 window.

FIG. 25 is a threshold value of the number of adjacencies in each direction of the halftone image and the non-halftone image in the graph showing the number of adjacencies in the four main scanning directions with respect to the number of black pixels in the 7 × 7 window. FIG.

FIG. 26 is a graph showing the number of black pixels in a 7 × 7 window and the number of adjacencies in the four main- and sub-scanning directions. It is a figure which shows each threshold value of the adjacent number of.

FIG. 27 is a fifth diagram for explaining a pattern matching method for calculating a halftone index for a concentrated halftone image.
Pattern matching windows W4a to W4e
FIG.

28 is a first image pattern PAT1 used to show an example of calculation by the pattern matching method performed using the pattern matching window shown in FIG.
FIG.

FIG. 29 is a second image pattern PAT2 used for showing an example of calculation by the pattern matching method performed using the pattern matching window shown in FIG.
FIG.

30 is a third image pattern PAT3 used to show an example of calculation by the pattern matching method performed using the pattern matching window shown in FIG. 27.
FIG.

31 is a graph of a non-halftone index for a dispersed halftone image stored in the table ROM of the discrimination data signal generation unit shown in FIG. 7. FIG.

32 is a graph of a non-halftone index for a systematic halftone image, which is stored in the table ROM of the discrimination data signal generation unit illustrated in FIG. 7. FIG.

33 is a diagram showing a smooth spatial filter F1 used in the smoothing amount calculator shown in FIG. 15 for counting the number of black pixels in a 7 × 7 window.

FIG. 34 is a diagram showing a smooth spatial filter F2 used in the smoothing amount calculator shown in FIG. 15 for counting the number of black pixels in a 9 × 9 window.

FIG. 35 is a diagram showing a smooth spatial filter F3 used in the smoothing amount calculator shown in FIG. 15 for counting the number of black pixels in an 8 × 8 window.

36 is a diagram showing an edge amount detection spatial filter F11 for calculating an edge enhancement amount, which is used in the edge enhancement amount calculation unit shown in FIG.

37 is a diagram showing an edge amount detection spatial filter F12 used in the edge enhancement amount calculation unit shown in FIG. 16 for calculating an edge enhancement amount.

FIG. 38 is a diagram showing an edge amount detection spatial filter F13 for calculating an edge enhancement amount, which is used in the edge enhancement amount calculation unit shown in FIG.

39 is a diagram showing an edge amount detection spatial filter F14 for calculating an edge enhancement amount used in the edge enhancement amount calculation unit shown in FIG.

FIG. 40 is a diagram showing an edge amount detection spatial filter F21 used in the edge enhancement amount calculation unit shown in FIG. 16 for calculating an edge enhancement amount.

41 is a diagram showing an edge amount detection spatial filter F22 for calculating an edge enhancement amount, which is used in the edge enhancement amount calculation unit shown in FIG.

42 is a diagram showing an edge amount detection spatial filter F23 for calculating an edge enhancement amount used in the edge enhancement amount calculation unit shown in FIG.

43 is a diagram showing an edge amount detection spatial filter F24 for calculating an edge enhancement amount used in the edge enhancement amount calculation unit shown in FIG.

FIG. 44 is a diagram showing an edge amount detection spatial filter F31 for calculating an edge enhancement amount, which is used in the edge enhancement amount calculation unit shown in FIG.

45 is a diagram showing an edge amount detection spatial filter F32 for calculating an edge enhancement amount, which is used in the edge enhancement amount calculation unit shown in FIG.

46 is a diagram showing an edge amount detection spatial filter F33 for calculating an edge enhancement amount used in the edge enhancement amount calculation unit shown in FIG.

FIG. 47 is a diagram showing an edge amount detection spatial filter F34 for calculating an edge enhancement amount, which is used in the edge enhancement amount calculation unit shown in FIG.

48 is a diagram showing an edge amount detection spatial filter F41 used in the edge discrimination amount calculation unit shown in FIG. 17 for calculating an edge discrimination amount.

49 is a diagram showing an edge amount detection spatial filter F42 for calculating an edge discrimination amount, which is used in the edge discrimination amount calculation unit shown in FIG.

50 is a diagram showing an edge amount detection spatial filter F43 for calculating an edge discrimination amount used in the edge discrimination amount calculation unit shown in FIG.

51 is a diagram showing an edge amount detection spatial filter F44 for calculating an edge discrimination amount used in the edge discrimination amount calculation unit shown in FIG.

FIG. 52 is a diagram showing an edge amount detection spatial filter F51 used in the edge determination amount calculation unit shown in FIG. 17 for calculating an edge determination amount.

53 is a diagram showing an edge amount detection spatial filter F52 used in the edge determination amount calculation unit shown in FIG. 17 for calculating an edge determination amount.

FIG. 54 is a diagram showing an edge amount detection spatial filter F53 used in the edge determination amount calculation unit shown in FIG. 17 for calculating an edge determination amount.

55 is a diagram showing an edge amount detection spatial filter F54 used in the edge discrimination amount calculation unit shown in FIG. 17, for calculating an edge discrimination amount.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Printer part, 20 ... Image reading part, 20a ... Image processing part, 40 ... Operation panel, 50 ... MPU, 53 ... Modem, 54 ... NCU, 55 ... Printer control part, 59 ... Buffer memory, 61 ... Page memory, 62 ... Image restoration processing unit, 64 ... Interpolation processing unit, 70 ... Laser printer, 100 ... 15 × 18 matrix memory circuit, 101 ... Halftone image restoration unit, 102 ... Image area discrimination unit, 103 ... Simple multi-valued unit, 104 ... Data mixing unit, 105 ... Adjacent state determination unit, 106 ... Systematic halftone determination unit, 107 ... 9 × 17 matrix memory circuit, 108 ... Judgment data generation unit, 109 ... Smoothing amount calculation unit, 110 ... Edge enhancement amount Calculation unit, 111 ... Edge discrimination amount calculation unit, 112 ... Restoration data calculation unit, 114 ... Discrimination data signal generation unit, 121 ... 7 × 7 black pixel count Road, for 123 and 124 ... table ROM, 125 ... data selector, 126,254 ... clear circuit, 251,259 ... data selector, 543 ... operation key.

Claims (4)

[Claims] 1. An input binary image including binary image data binarized in pseudo-halftone and binary image data binarized in non-halftone using a predetermined threshold value. A computing unit that computes an adjacency number indicating the adjacency state of pixels of the same type in a block composed of a plurality of pixels including the target pixel to be processed, and an adjacency number that indicates the adjacency state calculated by the arithmetic unit. Based on the above, the input binary image data is a pseudo-halftone binary image data or a non-halftone binary image data. In the above, the determination means performs the determination by comparing the number of adjacencies indicating the adjacency state with a predetermined threshold, and changes the threshold based on the resolution of the input binary image data. Characterized by That image processing apparatus. 2. The image processing apparatus is further provided with the input 2 based on a result determined by the determining means.
The image processing apparatus according to claim 1, further comprising a restoring unit that restores the value image data to multi-tone image data. 3. Based on the input binary image data,
In an image processing apparatus provided with a restoring unit that restores the input binary image data into multi-tone image data using a spatial filter having a predetermined window, the restoring unit is the input binary image. When the resolution of the data is higher than a predetermined threshold value, the input binary image data is restored to multi-tone image data using a spatial filter having a window of a predetermined first size,
On the other hand, when the resolution of the input binary image data is lower than a predetermined threshold value, the input binary image data is multi-ordered using a spatial filter having a window smaller than the first size. An image processing apparatus which restores tonal image data. 4. A discriminating means for discriminating whether or not the binary image data is binary image data binarized by a dot-concentrated systematic dither method based on the inputted binary image data. And a restoration unit that restores the input binary image data into multi-tone image data by performing image processing on the input binary image data according to the determination result of the determination unit, The image processing apparatus, wherein the determining means determines that the input binary image data is not binarized by the dot-concentrated systematic dither method when the resolution of the input binary image data is equal to or less than a predetermined value.