JP2844622B2 - Image reading device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image reading apparatus that separates a color image into three primary colors by an image sensor, reads the image, and outputs an image signal corresponding to each color. Related to balance correction.

[Conventional technology]

Conventionally, a still image such as a document is read by an image sensor as an image input unit of a computer and a document image reading unit of a digital copying machine.
2. Description of the Related Art An image reading apparatus that performs various types of image processing on obtained image data and then outputs an image signal is used.

As an optical system of such an image reading apparatus, a light source for illumination is provided below a document glass on which a document is placed,
A rod lens that collects the reflected light from the original, and a CCD (charge-coupled device) arranged in the main scanning direction.
An equal-magnification type optical system for moving a two-dimensional image sensor (line sensor) in the sub-scanning direction is generally used. In a system for reading a color image, three primary colors, ie, three primary colors, R (red), G
(Green) and B (blue) color separation filters are provided.

The photoelectric conversion output of each color from the line sensor, which is obtained by decomposing the original image into three primary colors and read, is appropriately amplified, quantized by analog / digital (A / D) conversion means, and the reflected light intensity of each color at each pixel. Is generated according to the image data.

These image data are sent to an image forming apparatus such as a printer as image signals through various image processing.

Generally, in image formation, in order to correctly reproduce the tone of an original color image, image data called white balance correction is standardized between an image reading apparatus and an image forming apparatus. That is, the reference color (usually white) is determined, and the ratio of the image data of each color (R, G, B) when the original image of the uniform reference color is read is kept constant. Such standardization increases compatibility with various image forming apparatuses, and when the line sensor is replaced or when the light receiving sensitivity of the line sensor varies in mass production of the image reading apparatus, the optical system changes with time. , Even when the color shift occurs, correct color tone can be reproduced.

[Problems to be solved by the invention]

In a conventional image reading apparatus, white balance correction is performed at the stage of quantization of a photoelectric conversion output. That is,
The analog reference potential applied to the A / D converter at the time of quantization is adjusted for each color so that the image data of each color becomes equal when the reference color image is read.

Therefore, with a high-speed image reading device, it is almost impossible to adjust the reference potential for each color according to the reading scan of each color, which is performed almost simultaneously, so it is necessary to provide an A / D converter for each of the three primary colors. In addition, since analog processing is included, there is a problem that correction accuracy is easily affected by external factors such as temperature.

CPU (Central Processing Unit) especially for white balance correction
When using the A / D converter,
A digital-to-analog (D / A) converter that generates a reference voltage based on PU operation result data is indispensable, and the configuration is complicated.

Furthermore, since the size of a solid-state imaging device such as a CCD is limited by the size of a semiconductor wafer, in an image reading device that reads an A4 or A3 size image, a line sensor is configured by a plurality of CCD chips. Thus, for example, 5
In the case of a chip configuration, three for each chip, a total of 15 (3 ×
5) The A / D converter and the D / A converter are required, and there is a problem that the device becomes large and expensive.

Japanese Patent Application Laid-Open No. 60-87569 proposes performing white balance correction on digital data after A / D conversion. However, according to this, the correction coefficient M is obtained as M = (2C / V _in ).
The correction coefficient M is determined by a constant C which is the lower limit of the white balance correction range. Since data processing is usually performed with a certain number of bits,
The correction range and the correction accuracy have a trade-off relationship. Therefore,
According to this conventional example, when a certain correction accuracy is ensured, a constant dynamic range (correction range) is always obtained, so that an optimal adjustment is made for a change in the performance of the image sensor due to a change in the use environment and a change over time. Is not done.

The present invention has been made in view of the above-described problems, and can perform high-accuracy white balance correction with a simple configuration, and can also have a wide dynamic range, so that the use environment changes and changes over time. It is an object of the present invention to provide an image reading device capable of performing optimal adjustment for the performance fluctuation of the image sensor due to the above.

[Means for solving the problem]

In order to solve the above-mentioned problems, the present invention separates a color image into three primary colors using an image sensor, reads the color image, and reads R,
In an image reading device that outputs image signals corresponding to G and B colors, an A / D conversion unit that quantizes the output of the image sensor to generate image data corresponding to each color, and each color when a reference color image is read Of the image data of each color obtained from a predetermined pixel, the image data of the color indicating the maximum is extracted, and the extracted image data is compared with the image data of other colors based on the extracted image data. A coefficient calculating means for calculating correction coefficient data for each color as a value based on the relative difference; and a corrected image in which a predetermined operation is performed on image data corresponding to each color based on the correction coefficient data to correct the balance of each color. Calculating means for outputting data, and matching the extracted color image data with the other color image data to correct the image data obtained by reading the reference color image. Each color ratio of the corrected image data is configured to be a predetermined value.

[Action]

The image sensor separates a color image into three primary colors and reads it.

The A / D converter quantizes the output of the image sensor to generate image data corresponding to the densities of R, G, and B colors.

The coefficient calculating means calculates correction coefficient data for each color, and inputs this to the calculating means. In calculating the correction coefficient data, the image data of the color having the largest size is extracted from the image data of each color when the reference color image is read, and the image of the other color is extracted based on the extracted image data. The data is compared with the data, and the correction coefficient data is calculated as a value based on the relative difference between the two. As a result, the width of the correction coefficient data can be widened, and thus the dynamic range of the correction can be widened.

The arithmetic means corrects the image data corresponding to each color generated by the A / D converter based on the corresponding correction coefficient data so that each color is balanced, and outputs corrected image data.

The correction coefficient data is calculated such that the ratio of each color of the corrected image data obtained by correcting the image data when the reference color image is read becomes a predetermined value.

〔Example〕

Hereinafter, an image reader section IR incorporated in a digital copying machine will be described as an embodiment of the present invention with reference to the drawings.

The digital copier includes an image reader unit IR as an image reading device, and a laser printer unit LP that forms a color image by electrophotography based on an image signal sent from the image reader unit IR.
Performs various kinds of signal processing on the pixel signal obtained by reading the image of the document, and outputs an image signal.

FIG. 4 is a perspective view showing an optical system of the image reader unit IR,
FIG. 5 is a plan view of the image sensor 11, and FIG. 6 is an enlarged view showing the CCD sensor chip 11a of FIG.

In FIG. 4, a document D placed on a document table glass (not shown) is line-scanned in a sub-scanning direction by a slider 14 having an image sensor 11, and an exposure lamp 17, a rod lens array 15, and an image sensor 11 are provided. Is read out after being separated into three primary colors of R (red), G (green), and B (blue) by an equal magnification type optical system having The R, G, and B photoelectric conversion output signals are output by a color correction circuit 105, which will be described later, into three colors of Y (yellow), M (magenta), and C (cyan), or four colors obtained by adding Bk (black) thereto. After undergoing various signal processing, the signal is sent as an image signal to a laser printer unit that forms a color image by deflection scanning of a laser beam.

As shown in FIG. 5, the image sensor 11 includes five contact type CCD sensor chips 11a, 11a,... Which are continuous in the horizontal direction (main scanning direction) and in the vertical direction (sub scanning direction). Are arranged in a zigzag pattern alternately at a constant pitch. Since there is a constant pitch in the sub-scanning direction, a delay occurs in the output signal from the CCD sensor chip 11a at the rear in the sub-scanning direction.
Is compensated by delaying the output signal from.

Each CCD sensor chip 11a has one size of 62.5 μm (d = 1/16) as shown in an enlarged view in FIG.
mm) A number of square elements are arranged in one row.

Each element is divided into three, and a spectral filter is provided so that one separation area receives light of one of the three primary colors RGB.

Such one element corresponds to one pixel obtained by subdividing the original image, and the photoelectric conversion output of one element represents the reflected light intensity of one color of one pixel.

FIG. 3 is a block diagram of an electric circuit of the image reader unit IR.

In the image sensor 11, in order to increase the reading speed in the main scanning direction, five CCD sensor chips 11a, 11a,... Are driven simultaneously, and valid reading pixel signals of 2928 pixels in total of RGB are sequentially output from each of them.

The photoelectric conversion outputs serially output simultaneously (in parallel) from the five CCD sensor chips 11a are quantized by a digitization processing circuit 101 having a sample-and-hold circuit and an A / D converter, and are 8-bit (256 gradations). The data is converted into digital data, separated into image data of each color by a latch circuit, and then input to the 5-channel combining circuit 102.

The five-channel synthesizing circuit 102 temporarily stores image data for each of two lines in a total of 15 (3 × 5) first-in first-out memories for each chip and each color, and sequentially selects image data from each chip in a one-line cycle. Then, a serial image signal corresponding to the pixel arrangement (reading scan order) is generated.

The image data of each color transmitted as a serial image signal is normalized by adjusting the relative ratio between the colors by a white balance correction circuit 103 so that an image of a correct color tone can be formed in the laser printer unit.

Next, the shading correction circuit 104 controls the exposure lamp 17
The light distribution in the main scanning direction (light intensity unevenness) and the sensitivity difference between the CCD sensor chips 11a are corrected, and the data signal that is proportional to the reflection intensity is based on the visual characteristics. Logarithmically converted to a density data signal proportional to the density of the document D.

In the color correction circuit 105, as described above, masking processing for generating density data corresponding to the three primary colors Y, M, and C of the printing toner from density data for each of the RGB colors, and density data corresponding to Bk (black) are generated. UCR processing and the like are performed, and the gamma correction circuit 106 performs gamma correction based on the background color and density gradient of the document D.

The color editing circuit 107 performs three types of color image editing processes: negative / positive inversion, color change (color change), and paint (filling).

In addition, the scaling / movement processing circuit 108 uses a thinning method, an arithmetic method, an interpolation method, or the like to form a scaled image that is enlarged or reduced, and an edit image such as movement or mirror inversion.
A process for changing the output timing and output order of the density data signal or the scanning speed in the sub-scanning direction is performed, and the MTF correction circuit 109
Performs smoothing for preventing the occurrence of moire fringes and edge enhancement for eliminating edge loss.

As described above, the density data D97 to 9
The 0 signal is binarized by an area gradation method in a gradation reproduction circuit 110, and is converted into image signals VIDEO4 to 0 in a laser printer section LP.
Sent to In addition, 111 is a line memory for storing image data at a specific processing stage, and 112 is a CPU for controlling each circuit.
(Central processing unit) 113 is a ROM in which programs and various data are stored.

FIG. 1 is a block diagram of the white balance correction circuit 103.

The white balance correction circuit 103 includes image data RD27 to RD20 and GD27 to GD2 of each color from the 5-channel synthesis circuit 102.
0, BD27-20 (8 bits each), correction coefficient data Wr7-0, Wg7-0, Wb7-0 (8 each)
Bit), the corrected image data RD37-3
0, GD37-30, and BD37-30 (8 bits each) The white balance correction circuit 103 includes multipliers 21-23 to which image data (R, G, B) D27-20 are input as multiplicands, respectively. , Correction coefficient data Wr7-0, Wg7-0,
Adders 31 to 33 which add each of Wb7 to 0 and auxiliary data ND from two's complement circuit 115 which is given in common to each color, and gives the arithmetic sum to multipliers 21 to 23 as multiplier data MDr, MDg, MDb. Adders 41 to 43 for adding the image data (R, G, B) D27 to D20 and the output data of the multipliers 21 to 23 to generate corrected image data (R, G, B) D37 to D30 have.

The correction coefficient data Wr7-0, Wg7-0, Wb7-0, and auxiliary data ND all represent an 8-digit decimal number with a decimal point between the most significant bit and the next bit. That is, 20 ⁰ , 2 ⁻¹ , 2 ⁻² … 2 ^-7 are assigned in order from the most significant bit, and ^1/1
It is treated as a small number of data in increments of 28 ( ^2-7 ).

The 80's (1000 0000B) bit signal (integer 1) is always supplied from the CPU 112 to the two's complement circuit 115 as the converted data CD, and the correction coefficient data Wr7-0, Wg7-0, Wb7-0
When the most significant bit (sign bit) is "0", the converted data CD is directly output as the positive auxiliary data ND, and when the sign bit is "1", the two's complement of the converted data CD is output.
That is, "-1" is output as the negative auxiliary data ND.

Therefore, the multiplier data MD (r, g, b) is a value obtained by adding "1" to the correction coefficient data Wr7-0, Wg7-0, Wb7-0, or subtracting "-1".

FIG. 2 is a flowchart of the white balance correction process controlled by the CPU 112.

This processing can be executed at any time in consideration of the aging of the optical system, but is usually performed before the scanning of the document D is started.

First, in step # 101, the A /
The reference potential of the D converter is set, and laser adjustment between the CCD sensor chips 11a is performed.

In step # 102, the correction coefficient data Wr7-0, Wg7-
Set 80H (integer 1) as 0, Wb7-0. As a result, the multiplier data MD (r, g, b) becomes 0, and the multiplier 21
Since the multiplication of 0 times is performed in 〜23, the image data (R, G, B) D27DD20 are output from the adders 41〜43 as corrected image data (R, G, B) D37〜30 as they are. Will be.

Next, in step # 103, the exposure lamp 17 is turned on at the standby position of the slider 14, and the reference white plate 16 of uniform density provided at the end of the platen glass (see FIG. 4)
Read. Ideally, when a reference white image is read, the image data of the three primary colors are equal, but in practice, CC
Differences occur between RGB due to the spectral sensitivity of the D sensor chip 11a and the like.

Therefore, in this embodiment, in order to perform the normalization so that the ratio between RGB is 1: 1: 1, the following steps # 104 to # 108
, The calculation processing of the correction coefficient data Wr7-0, Wg7-0, Wb7-0 is executed.

In step # 104, the corrected image data (R, G, B) D37 to D3 output without being substantially corrected as described above.
0 is stored in the line memory 111 one line at a time via the shading correction circuit 104. At this time, the shading correction circuit 104 does not perform the correction processing, and performs control to pass the input corrected image data (R, G, B) D37 to D30 as they are.

Next, in step # 105, an average value of one line is obtained for each color, and in step # 106, the maximum value among the three average values is set to "1" to calculate relative data of each color. In step # 107, Then, it is determined whether the ratio between the relative data is 1: 1: 1. If yes in step # 107, it means that the standardization has been completed, so the process returns to the main routine for controlling other image processing and operations of each section of the digital copying machine. move on.

In step # 108, the correction coefficient data Wr7-0, Wg7-
The reciprocals of the relative data corresponding to each color are set as 0 and Wb7-0, respectively, and at the same time, the value of the sign bit is transmitted to the two's complement circuit 115 as the control signal 2SC, and the process returns to step # 103.

For example, if the relative data values of R, G, B are 1, 0.95, 0.
In the case of 65, 1/1 as the correction coefficient data Wr7-0, 1 / 0.95 (= 1 + 5/95) as the data Wg7-0, and the data Wb7
1 / 0.65 (= 1 + 35/65) is set as ０0.

Therefore, in this case, the multiplier data MDr, MDg, and MDb applied to each of the multipliers 21 to 23 respectively include the correction coefficient data W
Since "1" is subtracted from r7-0, Wg7-0, Wb7-0,
0, 5/95 and 35/65.

Here, when the reference white plate 16 is read again in step # 103, when the same image data as the previous time, that is, relative data is obtained, the white balance circuit 103 outputs 1, 95/100, 65/100 (R, G, B) D37-30
Is entered.

The multiplier 21 corresponding to G performs multiplication of 5/95 times, and as a result, the adder 42 outputs 5/100 of the output of the multiplier 21 and GD 27 to 20.
100/100 corrected image data GD37 to GD30 obtained by adding 95/100 which is the value of

Similarly, the multiplication and addition operations are performed for B, and the adder 43 outputs corrected image data BD37 to BD30 equal to the corrected image data RD37 to RD30 corresponding to R.

Therefore, the mutual ratio of the corrected image data (R, G, B) D37-30 of each color corresponding to the image data (R, G, B) D27-20 read from the reference white plate is 1: 1: 1, This means that the white balance correction has been completed.

Thereafter, when the document D is read, the white balance correction circuit 103 outputs the set correction coefficient data Wr7 to Wr0, Wg
By performing an operation based on 7-0 and Wb7-0, the image data input from the preceding stage is corrected and transmitted to the subsequent image processing circuit.

According to the above-described embodiment, the correction coefficient data Wr7
Multiplier data obtained by subtracting "1" from ~ 0, Wg7-0, Wb7-0
MDr, MDg, MDb and input image data (R, G, B) D27-20
Multiplied by the original image data (R, G, B) D27-20
Correction image data (R, G, B) D37 to D30 whose color tone has been standardized is generated by the added operation, so that the correction range is expanded as compared with the correction using only multiplication.

That is, when the 8-bit multipliers 21 to 23 are used in accordance with the input / output image data, the maximum number of bits of the multiplier data MDr, MDg, MDb added to the multipliers 21 to 23 is also 8 bits. When the number of bits is limited in this way, the correction range and the correction accuracy that are in conflict with each other are determined by the position where the decimal point is set. For example, if a decimal point is placed between the high-order second and third bits, an integer is represented by the high-order 2 bits, and a numerical value after the decimal point is represented by the low-order 6 bits, 0 to 4 (more precisely, 0 to 3 + 63/6)
Since the numerical value of 4) can be represented, the correction range is 0 to 4 times. On the other hand, the correction accuracy is in 1/64 steps since 2 ^-6 is assigned to the least significant bit.

As described above, in this embodiment, the multiplier data MDr, MDg, MDb
Since 2 ^-7 is assigned to the least significant bit of, a correction accuracy of 1/128 is assured. However, since the integer is 1 bit, if the correction is performed only by multiplication in this bit configuration, the correction range becomes 0 to 2 times, and correct correction can be performed when the minimum input image data is smaller than 1/2 of the maximum input image data. Disappears.

Therefore, when an operation combining multiplication and addition is performed as described above, the doubled input image data and the original input image data are added. As a result, the original input image data is tripled. Corrected image data is obtained. Therefore, the minimum input image data is
Even in the case of 1/3, the mutual ratio of the corrected image data (R, G, B) D37 to D30 of each color can be set to 1: 1: 1. For example, in the characteristic surface of the CCD chip sensor 11a, The degree of freedom of choice can be increased.

The circuit configuration intended to expand the correction range is not limited to the illustrated one.For example, the adder 31 to 33 of the present embodiment is omitted and the correction coefficient data Wr7 to 0, Wg7 to 0, and Wb7 to 0 are used. When the most significant bits (sign bits) of the correction coefficient data Wr7-0, Wg7-0, and Wb7-0 are "0", the multipliers 21-23 are directly added to the multipliers 21-23 as multipliers. Is output to the adders 41 to 43 as it is, and when it is "1", it is converted to a two's complement and added to the adders 41 to 43, and the same result as in the present embodiment can be obtained.

According to the above-described embodiment, since the other corrected image data (R, G, B) D37-30 of each color is matched with the largest one, the dynamic range set in the preceding stage Is not compromised.

In the above-described embodiment, in order to enhance the reliability of the correction, the correction coefficient data Wr7-0, Wg7- are based on the average value of one line of the image data (R, G, B) D27-20 of each color.
0 and Wb7-0 are calculated. However, when the variation in the characteristics between the CCD chip sensors 11a or between the elements is very small, the image data for 1/50 line or 1 pixel is used. R, G, B) Correction coefficient data Wr7- based on D27-20
0, Wg7-0, Wb7-0 may be calculated.

In the above-described embodiment, normalization is performed so that the mutual ratio of the corrected image data (R, G, B) D37 to D37 of each color is 1: 1: 1. The mutual ratio can be appropriately determined according to the convenience of processing in the image forming apparatus or the color of the reference image.

〔The invention's effect〕

According to the present invention, after the output of the image sensor read by separating the original image into three primary colors is quantized, the color tone is normalized for the image data corresponding to each color by digital operation. There is no need to provide A / D conversion means, the configuration is simplified, and the correction accuracy is stable because of digital processing.

Also, in calculating the correction coefficient data, the image data of the color indicating the maximum is extracted from the image data of each color when the reference color image is read, and corrected as a relative value determined based on the extracted image data. Since the coefficient data is calculated, the width of the correction coefficient data can be widened while securing the correction accuracy, and thus the dynamic range of the correction can be widened. This makes it possible to perform optimal adjustment for performance fluctuations of the image sensor due to fluctuations in the use environment and changes over time.

[Brief description of the drawings]

1 is a block diagram of a white balance correction circuit, FIG. 2 is a flowchart of a white balance correction process, FIG. 3 is a block diagram of an electric circuit of an image reader unit, FIG. 4 is a perspective view showing the optical system of the image reader unit, FIG. 5 is a plan view of the image sensor, and FIG. 6 is an enlarged view showing the CCD sensor chip of FIG. 11 image sensor 101 digitization processing circuit (A / D conversion means) 112 CPU (coefficient calculation means) 103
… White balance correction circuit (arithmetic means), IR …… Image reader section (image reading device), RD27-20, RG27
~ 20, RB27 ~ 20 ... Image data, RD37 ~ 30, RG37 ~ 30,
RB37-30: Corrected image data, Wr7-0, Wg7-0, Wb7
0 to... Correction coefficient data.

Claims (1)

(57) [Claims] 1. An image reading apparatus which separates a color image into three primary colors by an image sensor and reads the image signals and outputs image signals corresponding to each of R, G, and B colors. A / D conversion means for generating image data, and among the image data of each color when the reference color image is read,
The image data of each color obtained from a predetermined pixel is compared to extract the image data of the color showing the maximum, and the extracted image data is compared with the image data of other colors based on the extracted image data, and based on the relative difference between the two. Coefficient calculating means for calculating correction coefficient data for each color as a value; calculating means for performing a predetermined operation on image data corresponding to each color based on the correction coefficient data and outputting corrected image data in which the balance of each color has been corrected And adjusting the ratio of each color of the corrected image data obtained by correcting the image data obtained by reading the reference color image to a predetermined value by matching the image data of the other color with the image data of the extracted color. An image reading device characterized by the above-mentioned.