JP6107493B2 - Image reading apparatus, image processing apparatus, and program - Google Patents

Description

  The present invention relates to an image reading device, an image processing device, and a program.

  As a conventional technique described in the publication, after reading an image for one page of a document, the read image data for one page of the document is binarized, and the generated binary image data is subjected to a Hough transform. There is an image processing apparatus that calculates a document skew angle of read image data from Hough space data obtained by Hough transform (see Patent Document 1).

JP 2002-84420 A

  An object of the present invention is to reduce the capacity of a memory for storing data obtained by the Hough transformation when the inclination angle of the original is obtained using the Hough transformation.

According to a first aspect of the present invention, there is provided a reading unit that reads an image of a conveyed document, a detection unit that performs edge detection processing on read image data read by the reading unit, and creates edge image data, and the edge image data By performing Hough transform, conversion means for creating Hough transform data expressed in θρ parameter space using angle θ and distance ρ, prediction means for predicting the range of the distance ρ in the Hough transform data, From the Hough transform data, extraction means for extracting the range of the distance ρ predicted by the prediction means to create extracted Hough transform data, storage means for storing the extracted Hough transform data, and the storage means Based on the extracted extracted Hough transform data, a determination means for determining an inclination angle of the document, and the read image data An image reading apparatus including a correction means for performing tilt correction using the tilt angle.
According to a second aspect of the present invention, the predicting unit sets a reference value of the end based on the position of the end of the document acquired from the edge image data created by the detecting unit, and the reference The image reading apparatus according to claim 1, wherein the range of the distance ρ is predicted based on the value.
The invention according to claim 3 is characterized in that the predicting means limits the range of the angle θ and predicts the range of the distance ρ based on the reference value and the range of the angle θ. 2. The image reading apparatus according to 2.
According to a fourth aspect of the present invention, the image forming apparatus further includes a detecting unit that detects a size of the document to be read, and the predicting unit detects the size of the document based on the size of the document detected by the detecting unit. The image reading apparatus according to claim 1, wherein a reference position of an end is set, and the range of the distance ρ is predicted based on the reference position.

  According to a fifth aspect of the present invention, there is provided detection means for generating edge image data by performing edge detection processing on read image data obtained by reading an image of a conveyed document, and performing Hough transform on the edge image data. A conversion unit that creates Hough transform data expressed in a θρ parameter space using an angle θ and a distance ρ, a prediction unit that predicts a range of the distance ρ in the Hough transform data, and a prediction from the Hough transform data Extraction means for extracting the extracted range of the distance ρ to create extracted Hough conversion data; storage means for storing the extracted Hough conversion data; and based on the extracted Hough conversion data read from the storage means And an deciding means for deciding the inclination angle of the original.

  According to a sixth aspect of the present invention, there is provided a function of generating edge image data by performing edge detection processing on read image data obtained by reading an image of a conveyed document to a computer, and performing Hough transform on the edge image data. Thus, a function to create Hough transform data expressed in a θρ parameter space using an angle θ and a distance ρ, a function to predict the range of the distance ρ in the Hough transform data, and a prediction from the Hough transform data A function of extracting the extracted range of the distance ρ to create extracted Hough transform data, a function of storing the extracted Hough transform data in a memory, and the extracted Hough transform data read from the memory , A program for realizing the function of determining the inclination angle of the original.

According to the first aspect of the present invention, the memory for storing the data obtained by the Hough conversion when the inclination angle of the original is obtained by using the Hough conversion as compared with the case without this configuration. The capacity can be reduced.
According to the second aspect of the present invention, it is possible to increase the possibility that the extracted Hough transform data includes the solution of the inclination angle as compared with the case where the present configuration is not provided.
According to invention of Claim 3, compared with the case where it does not have this structure, possibility that the solution of an inclination angle will be contained in the extracted Hough conversion data can be raised further.
According to the fourth aspect of the present invention, it is possible to increase the possibility that the extracted Hough transform data includes the solution of the inclination angle as compared with the case where the present configuration is not provided.
According to the fifth aspect of the present invention, the memory for storing the data obtained by the Hough transformation when the inclination angle of the document is obtained by using the Hough transformation as compared with the case where the present configuration is not provided. The capacity can be reduced.
According to the sixth aspect of the present invention, the memory for storing the data obtained by the Hough transformation when the inclination angle of the document is obtained by using the Hough transformation as compared with the case without this configuration. The capacity can be reduced.

It is a figure which shows an example of the whole structure of the image reading apparatus with which this Embodiment is applied. FIG. 3 is a diagram for explaining a configuration of a document to be read. It is a functional block diagram of a control and image processing unit. It is a functional block diagram of a signal processing part. (A) is a figure for demonstrating read image data, (b) is a figure for demonstrating the image data for a detection. It is a functional block diagram of a skew amount calculation unit. It is a functional block diagram of an offset setting part. (A) and (b) are figures which show an example of the edge detection data input into an offset setting part. It is a figure which shows an example of the preparation method of the offset setting data which an offset setting data preparation part produces. (A)-(c) is a figure for demonstrating the content of the process in an offset process part.

Embodiments of the present invention will be described below in detail with reference to the accompanying drawings.
FIG. 1 is a diagram illustrating an example of the overall configuration of an image reading apparatus to which the exemplary embodiment is applied.
The image reading apparatus includes a document feeding device 10 that sequentially conveys a document M from a stacked document stack, and a scanner device 40 that reads a document image by scanning.

  The document feeder 10 includes a document storage unit 11 on which a bundle of documents composed of a plurality of documents M is stacked, and a paper discharge storage unit 12 that is provided below the document storage unit 11 and stacks the document M that has been read. . In addition, the document feeder 10 includes a take-out roll 13 that takes out and transports the document M in the document storage unit 11. Further, on the downstream side of the take-out roll 13 in the document conveyance direction, a mechanism 14 for separating the documents M one by one is provided. A first conveyance roll 16, a second conveyance roll 17, a third conveyance roll 18 and a fourth conveyance roll 19 are provided in order from the upstream side in the document conveyance direction in the conveyance path 15 where the document M is conveyed. Among these, the first transport roll 16 and the second transport roll 17 send out the document M toward the reading position by the scanner device 40. The third transport roll 18 transports the document M read by the scanner device 40 further downstream by passing through the reading position. The fourth transport roll 19 further transports the read original M and discharges it to the paper discharge storage unit 12.

  Further, at the reading position, there is provided a reflecting plate 20 that is a white plate-like member that extends from the front side to the back side in the drawing and that also serves as the background of the document M that passes through the reading position. The reflector 20 is also used as a white reference for shading correction.

  The document feeder 10 also includes a document width detection sensor 21 as an example of a detection unit that detects the width (length in the main scanning direction) of a document stored in the document storage unit 11.

  On the other hand, the scanner device 40 supports the document feeder 10 described above so as to be openable and closable, supports the document feeder 10 by an apparatus frame 41, and reads an image of the document M. The scanner device 40 is provided below the device frame 41 forming the housing, the first platen glass 42A on which the document M to be read in an image is placed in a stationary state, and the reflection plate 20, and is conveyed by the document feeder 10. 2nd platen glass 42B which comprises the opening part of the light for reading the original M to be read is provided.

  Further, the scanner device 40 is stationary under the second platen glass 42B, or scans the entire first platen glass 42A to read an image, and the light obtained from the full rate carriage 43 is input to the imaging unit. A half-rate carriage 45 is provided. Here, the full rate carriage 43 includes a light source device 44A (white light source) that irradiates light on the original M, a light source mirror 44B that reflects light from the light source device 44A toward the original M, and reflected light obtained from the original M. And a first mirror 46A for receiving light. On the other hand, the half-rate carriage 45 includes a second mirror 46B and a third mirror 46C that provide the light obtained from the first mirror 46A to the imaging unit.

  Furthermore, the scanner device 40 includes an imaging lens 47 and a light receiving unit 48. Among these, the imaging lens 47 optically reduces the optical image obtained from the third mirror 46C. The light receiving unit 48 photoelectrically converts the optical image formed by the imaging lens 47. That is, in the scanner device 40, an image is formed on the light receiving unit 48 using a so-called reduction optical system. In the present embodiment, as the light receiving unit 48, a red, green, and blue image pickup element array provided along the main scanning direction is arranged in the sub scanning direction. Thereby, the image formed on the original M is read as a full-color image using the light receiving unit 48 as an example of a reading unit.

  The scanner device 40 further includes a control / image processing unit 49. The control / image processing unit 49 performs various types of image processing on the image data of the document M input from the light receiving unit 48. The control / image processing unit 49 controls the operation of each unit in the reading operation of the reading device.

  For example, in the fixed reading mode in which the image of the document M placed on the first platen glass 42A is read, the full rate carriage 43 and the half rate carriage 45 move in the arrow direction at a ratio of 2: 1. At this time, light from the light source device 44 </ b> A provided in the full rate carriage 43 is irradiated on the read surface of the document M. Then, the reflected light from the document M is reflected in the order of the first mirror 46 A, the second mirror 46 B, and the third mirror 46 C and guided to the imaging lens 47. The light guided to the imaging lens 47 is imaged on the light receiving surface of the light receiving unit 48. The image pickup element arrays for each color constituting the light receiving unit 48 are each constituted by a one-dimensional sensor and process one line at the same time. The full-rate carriage 43 and the half-rate carriage 45 are moved in the direction (sub-scanning direction of reading) intersecting the line direction (reading main-scanning direction), and the next line of the document M is read. By executing this over the entire document M, reading of one page of the document is completed.

  On the other hand, in the transport reading mode in which the image of the document M transported by the document feeder 10 is read, the document M transported in the sub-scanning direction passes over the second platen glass 42B. At this time, the full rate carriage 43 and the half rate carriage 45 are stopped at the position indicated by the solid line in FIG. Then, the reflected light of the first line of the conveyed document M is imaged by the imaging lens 47 through the first mirror 46A, the second mirror 46B, and the third mirror 46C, and is imaged by the light receiving unit 48. Is read. That is, after one line in the main scanning direction is simultaneously processed by the light receiving unit 48, one line in the next main scanning direction of the document M conveyed by the document feeder 10 is read. Then, after the leading edge of the document M reaches the reading position of the second platen glass 42B, the trailing edge of the document M passes the reading position on the second platen glass 42B, so that the document M is moved by one line in the main scanning direction. By sequentially reading each page, reading of one page of the document is completed in the sub-scanning direction.

  In the image reading apparatus according to the present embodiment, in the conveyance reading mode, the document M supplied into the conveyance path 15 is directly stopped in the conveyance path 15 without being temporarily stopped in the discharge path 12. It is designed to be transported towards. That is, this image reading apparatus does not particularly align the original M while the original M is being conveyed.

  In the image reading apparatus of the present embodiment, the standard reading resolution in each of the fixed reading mode and the conveyance reading mode is set to 600 dpi (main scanning direction) × 600 dpi (sub-scanning direction).

FIG. 2 is a diagram for explaining the configuration of the document M to be read in the image reading apparatus shown in FIG.
In this example, the document M has a rectangular shape, and the short side of the document M is along the main scanning direction FS, and the long side of the document M is along the sub-scanning direction SS. However, the long side of the document M may be along the main scanning direction FS, and the short side of the document M may be along the sub-scanning direction SS. In the following description, the side of the document M that is upstream in the sub-scanning direction SS is referred to as the document front end Mt, and the side that is downstream of the sub-scanning direction SS is referred to as the document trailing end Mb. The side on the left side of the surface to be read (the surface facing the front side in the figure) is called the document left end Ml, and the side on the right side of the surface to be read of the document M is called the document right end Mr.

  FIG. 3 is a functional block diagram of the control / image processing unit 49 shown in FIG. The control / image processing unit 49 processes read image data input from a red image sensor array, a green image sensor array, and a blue image sensor array (all not shown) provided in the light receiving unit 48. A signal processing unit 50 to be applied, and a control unit 70 for controlling the operations of the document feeder 10 and the scanner device 40. Details of the signal processing unit 50 will be described later.

The control unit 70 includes a reading controller 71, a light receiving unit driver 72, a light source driver 73, a scan driver 74, and a transport mechanism driver 75.
The reading controller 71 controls the entire document feeding device 10 and the scanner device 40 shown in FIG.
The light receiving unit driver 72 controls the image data capturing operation by the light receiving unit 48 (red image pickup device array, green image pickup device array, and blue image pickup device array) shown in FIG.
The light source driver 73 outputs a lighting signal, and controls turning on / off of the light source device 44A shown in FIG. 1 in accordance with the reading timing of the document.
The scan driver 74 turns on / off a motor in the scanner device 40 and controls the scanning operation by the full rate carriage 43 and the half rate carriage 45 shown in FIG.
The transport mechanism driver 75 controls motor control, various rolls, clutches, and gate switching operations in the document feeder 10 shown in FIG.

  From these various drivers, control signals are output to the document feeder 10 and the scanner device 40, and these operations can be controlled based on the control signals. The reading controller 71 sets a reading mode based on a control signal from the host system, a sensor output detected in, for example, an automatic selection reading function, a selection from a user received via a UI (not shown), and the like. The document feeder 10 and the scanner device 40 are controlled. Examples of the reading mode include the above-described fixed reading mode and conveyance reading mode.

FIG. 4 is a functional block diagram of the signal processing unit 50 shown in FIG.
The signal processing unit 50 according to the present embodiment includes an image preprocessing unit 51, an output destination setting unit 52, a read image data storage unit 53, a skew amount calculation unit 54, a skew correction processing unit 55, and an image postprocessing. Part 56.

  The image preprocessing unit 51 performs preprocessing such as gain adjustment, A (analog) / D (digital) conversion, shading correction, and gap correction on the analog read image data input from the light receiving unit 48, and digitizes the data. Is output as read image data Dr.

  The output destination setting unit 52 reads the read image data Dr input from the image preprocessing unit 51 and outputs the read image data Dr to the read image data storage unit 53. At the same time, a part of the read image data Dr is converted into a document. The detected image data Dd for detecting the M skew amount is output to the skew amount calculation unit 54.

  Here, FIG. 5A is a diagram for explaining the read image data Dr output from the output destination setting unit 52 to the read image data storage unit 53, and FIG. FIG. 6 is a diagram for explaining detection image data Dd output to a quantity calculation unit 54; The read image data Dr and the detection image data Dd shown in FIG. 5 exemplify those obtained based on the result of reading the same document M.

First, the read image data Dr shown in FIG. 5A will be described.
The read image data Dr of the present embodiment is wider than the original M to be read and includes the read surface of the original M, that is, the front end Mt, the rear end Mb, the left end Ml of the original, and the right end Mr of the original. The target area is set. Note that the outside of the document M in the read image data Dr is an area where the reflection plate 20 shown in FIG. 1 is read. In this example, the position of each edge in the document M can be detected based on the difference in reflectance between the reflector 20 and the document M. In this example, the left side in FIG. 5A is the upstream side in the main scanning direction FS, and the upper side in FIG. 5A is the upstream side in the sub-scanning direction SS. Accordingly, the read image data Dr is sequentially acquired for each line in the main scanning direction FS from the upstream side in the sub-scanning direction SS to the downstream side (from the upper side to the lower side in the figure).

  FIG. 5A illustrates a case where the document M is read in a state inclined (skewed) to the right side in the drawing with respect to the main scanning direction FS and the sub-scanning direction SS. Therefore, the document front end Mt and the document rear end Mb of the document M are inclined with respect to the main scanning direction FS, and the document left end Ml and the document right end Mr of the document M are inclined with respect to the sub-scanning direction SS.

Next, the detection image data Dd shown in FIG. 5B will be described.
In the detection image data Dd, the target area is set so as to include a part of the read surface of the document M on the upstream side in the sub-scanning direction SS in the read image data Dr shown in FIG. (Indicated by the solid line in the figure). Here, in the present embodiment, the detection image data Dd includes all of the document front end Mt of the document M, a part of the document left end Ml upstream of the sub-scanning direction SS, and the document right end Mr upstream of the sub-scanning direction SS upstream. A part is included, and a part of the document left end Ml downstream of the sub-scanning direction SS, a part of the document right end Mr downstream of the sub-scanning direction SS, and the document trailing end Mb are not included. (Indicated by broken lines in the figure).

Returning to FIG. 4, the description will be continued.
The read image data storage unit 53 stores the read image data Dr input from the output destination setting unit 52. Note that the read image data storage unit 53 is configured by, for example, a RAM (Random Access Memory).

  In addition, the skew amount calculation unit 54 performs various calculation processes on the detection image data Dd input from the output destination setting unit 52, so that the skew amount of the document M that is the source of the detection image data Dd is obtained. A certain document skew angle θm is output. The details of the skew amount calculation unit 54 will be described later.

  The skew correction processing unit 55 as an example of a correction unit inputs the read image data Dr of one page of the document M read from the read image data storage unit 53 from the skew amount calculation unit 54 and inputs the document M. The skew correction (rotation of image data) using the document skew angle θm corresponding to is performed, and the obtained skew-corrected image data Dj is output.

  The image post-processing unit 56 subjects the skew-corrected image data Dj input from the skew correction processing unit 55 to color space conversion processing, enlargement / reduction processing, and the like, as necessary, and outputs the obtained output. The image data Dk is output to a device (such as a printer or a computer device) provided outside the image reading apparatus shown in FIG.

  Note that the function of each unit constituting the signal processing unit 50 is realized by cooperation of software and hardware resources. That is, a CPU (Central Processing Unit) (not shown) provided in the signal processing unit 50 reads a program that realizes each function in the signal processing unit 50 from an external storage device such as a hard disk into the main memory, for example. Realize the function. However, the present invention is not limited to this. For example, the function of each unit constituting the signal processing unit 50 may be realized using an application specific integrated circuit (ASIC) or a field-programmable gate array (FPGA). Absent.

FIG. 6 is a functional block diagram of the skew amount calculation unit 54 shown in FIG.
The skew amount calculation unit 54 of the present embodiment includes an edge detection processing unit 541, an edge detection data storage unit 542, a Hough conversion processing unit 543, an offset setting unit 544, an offset processing unit 545, and a Hough conversion data storage. A section 546 and a document skew angle determination section 547.

  An edge detection processing unit 541 as an example of a detection unit is configured to apply an edge (edge in the document M) to the detection image data Dd obtained from the read image data Dr input from the output destination setting unit 52 (see FIG. 4). Part) is detected, and edge detection data De as an example of the obtained edge image data is output. When the detection image data Dd includes red image data, green image data, and blue image data, a process for monochromaticizing these three color image data is performed before the edge detection process is performed. The Alternatively, image data of one color (for example, green) may be selected from red image data, green image data, and blue image data. Further, before performing the edge detection process, a process for reducing the resolution of the detection image data Dd may be performed.

  The edge detection data storage unit 542 stores the edge detection data De input from the edge detection processing unit 541. Note that the edge detection data storage unit 542 is constituted by a RAM, for example.

  The Hough conversion processing unit 543 as an example of conversion means performs Hough conversion on the edge detection data De read from the edge detection data storage unit 542, and outputs the obtained Hough conversion data Dh. In this example, the Hough transform processing unit 543 obtains the Hough transform data Dh by transforming the edge detection data De defined in the xy image space into the θρ parameter space defined by the distance ρ and the perpendicular θ. ing.

  The offset setting unit 544 as an example of a predicting unit is based on the edge detection data De input from the edge detection processing unit 541, and the range of the distance ρ in the Hough transform data Dh obtained by Hough transforming the edge detection data De And offset setting data Do as a prediction result is output. Details of the offset setting unit 544 will be described later.

  An offset processing unit 545 as an example of an extracting unit performs offset processing using the offset setting data Do input from the offset setting unit 544 on the Hough conversion data Dh input from the Hough conversion processing unit 543, Offset Hough conversion data Di as an example of the extracted Hough conversion data obtained is output.

  The Hough conversion data storage unit 546 as an example of a storage unit stores the offset Hough conversion data Di input from the offset processing unit 545. The Hough conversion data storage unit 546 is constituted by, for example, a RAM.

  The document skew angle determination unit 547 as an example of a determination unit is based on the offset Hough conversion data Di read from the Hough conversion data storage unit 546, and the document skew of the document M that is the source of the offset Hough conversion data Di. The angle θm is determined and output to the skew correction processing unit 55 (see FIG. 4). Further, the document skew angle determination unit 547 uses the offset setting data Do input from the offset setting unit 544 as necessary when determining the document skew angle θm as an example of the tilt angle.

FIG. 7 is a functional block diagram of the offset setting unit 544 shown in FIG.
The offset setting unit 544 of the present embodiment is offset reference value data indicating a reference value for setting an offset, which will be described later, based on the edge detection data De input from the edge detection processing unit 541 (see FIG. 6). Based on the offset reference value determination unit 5441 for determining Ds and the offset reference value data Ds input from the offset reference value determination unit 5441, the offset processing unit 545 (see FIG. 6) performs an offset process on the Hough transform data Dh. And an offset setting data creation unit 5442 for creating offset setting data Do used for application.

Next, a procedure of processing executed by the skew amount calculation unit 54 in the image reading operation will be described with reference to FIGS. 6 and 7.
In this process, first, the edge detection processing unit 541 creates edge detection data De based on the detection image data Dd input from the output destination setting unit 52 (see FIG. 4). The obtained edge detection data De is output to the edge detection data storage unit 542 and the offset setting unit 544.

  Next, the edge detection data storage unit 542 stores the edge detection data De input from the edge detection processing unit 541. Next, the Hough conversion processing unit 543 performs Hough conversion on the edge detection data De read from the edge detection data storage unit 542 to obtain Hough conversion data Dh. Then, the obtained Hough transform data Dh is output to the offset processing unit 545.

  On the other hand, in the offset setting unit 544, the offset reference value determination unit 5441 creates the offset reference value data Ds based on the edge detection data De input from the edge detection processing unit 541, and the offset setting data creation unit 5442 Based on the offset reference value data Ds input from the offset reference value determination unit 5441, the offset setting data Do is created. The obtained offset setting data Do is output to the offset processing unit 545 and the document skew angle determination unit 547.

  Subsequently, the offset processing unit 545 applies the offset setting input from the offset setting unit 544 (more specifically, the offset setting data creation unit 5442) to the Hough conversion data Dh input from the Hough conversion processing unit 543. Offset processing using the data Do is performed to obtain offset Hough transform data Di. At this time, the offset processing unit 545 creates the offset Hough conversion data Di using the Hough conversion data Dh obtained based on the same edge detection data De and the offset setting data Do. The obtained offset Hough conversion data Di is output to the Hough conversion data storage unit 546. Next, the Hough transform data storage unit 546 stores the offset Hough transform data Di input from the offset processing unit 545.

  Then, the document skew angle determination unit 547 determines the document skew angle θm of the document M based on the offset Hough conversion data Di read from the Hough conversion data storage unit 546. At this time, the document skew angle determination unit 547 refers to the offset setting data Do input from the offset setting unit 544 when determining the document skew angle θm. Then, the obtained document skew angle θm is output to the skew correction processing unit 55 (see FIG. 4), whereby the processing for one document M is completed.

Now, various processes executed by the offset setting unit 544 in the above-described image reading operation will be described with specific examples.
FIG. 8 is a diagram illustrating an example of the edge detection data De input to the offset setting unit 544 illustrated in FIG. Here, FIG. 8A shows the edge detection data De obtained by reading the document M tilted to the left with respect to the transport direction, and FIG. 8B shows the document M tilted to the right with respect to the transport direction. The edge detection data De obtained by reading is shown.

As described above, the offset reference value determination unit 5441 provided in the offset setting unit 544 creates the offset reference value data Ds based on the input edge detection data De.
At this time, the offset reference value determination unit 5441 uses the edge detection data De to determine the front end reference position St as a reference for the document front end Mt of the document M, the left end reference position S1 as the reference for the document left end Ml of the document M, and the document. A right end reference position Sr serving as a reference for the M document right end Mr is determined. Then, the offset reference value determination unit 5441 outputs the determined front end reference position St, left end reference position Sl, and right end reference position Sr to the offset setting data creation unit 5442 as offset reference value data Ds.

The offset reference value determination unit 5441 determines the front end reference position St, the left end reference position Sl, and the right end reference position Sr according to the following procedure.
Here, in the edge detection data De, the image density differs between the background portion where the document M does not exist and the document portion where the document M exists (particularly, the edge portion where the edge of the document M exists). For example, if the background of the document M at the time of reading is set to white, the color of the document M shows a slightly darker color than white, so that the background portion and the document portion can be distinguished. At this time, by focusing on the pixels before and after the image density greatly changes in the edge detection data De, the edge pixel corresponding to the document front end Mt, the edge pixel corresponding to the document left end Ml, and the edge pixel corresponding to the document right end Mr. Can be detected.

  Then, the offset reference value determination unit 5441 determines, as the front end reference position St, the position of the edge pixel on the most upstream side (front side) in the sub-scanning direction SS among the edge pixels constituting the edge detection data De. Further, the offset reference value determination unit 5441 determines the position of the edge pixel on the most upstream side (left side) in the main scanning direction FS among the edge pixels constituting the edge detection data De as the left end reference position Sl. Further, the offset reference value determination unit 5441 determines, as the right end reference position Sr, the position of the edge pixel on the most downstream side (right side) in the main scanning direction FS among the edges constituting the edge detection data De.

  Thus, for example, when the document M is tilted to the left as shown in FIG. 8A, the most upstream side in the sub-scanning direction SS at the document right end Mr (the intersection of the document front end Mt and the document right end Mr). Becomes the front end reference position St, the most upstream side in the sub-scanning direction SS at the document left end Ml (the intersection of the document front end Mt and the document left end Ml) becomes the left end reference position Sl, and the most downstream side in the sub-scanning direction SS at the document right end Mr. It becomes the right end reference position Sr.

  On the other hand, for example, as shown in FIG. 8B, when the document M is tilted to the right side, the most upstream side in the sub-scanning direction SS at the document left end Ml (the intersection of the document front end Mt and the document left end Ml). It becomes the front end reference position St, the most downstream side in the sub-scanning direction SS at the document left end Ml becomes the left end reference position Sl, and the most upstream side in the sub-scanning direction SS at the document right end Mr (the intersection of the document front end Mt and the document right end Mr) is the right end. It becomes the reference position Sr.

  FIG. 9 is a diagram illustrating an example of a method of creating the offset setting data Do created by the offset setting data creating unit 5442 shown in FIG. 7 based on the offset reference value data Ds determined by the offset reference value determining unit 5441. FIG. 9 exemplifies edge detection data De (corresponding to FIG. 8B) obtained by reading the document M tilted to the left with respect to the transport direction.

In FIG. 9, the sub-scanning direction SS is shown as the x coordinate in the xy image space, and the main scanning direction FS is shown as the y coordinate in the xy image space.
FIG. 9 illustrates each element related to the document left end Ml.

  Based on the offset reference value data Ds (the front end reference position St, the left end reference position S1 and the right end reference position Sr), the offset setting data creation unit 5442 uses the Hough conversion data Dh output from the Hough conversion processing unit 543 to convert the document front end Mt. The range of the distance ρ that can be taken by the document left end Ml and the document right end Mr is predicted.

  Here, in the image reading apparatus of the present embodiment, for example, the tolerance of each component is set so that the range of the document skew angle θm of the document M in the transport reading mode is a predetermined range (for example, ± 3 °). The accuracy of each mechanism is determined. For this reason, the size (size) and orientation of the document M to be read and the positions of the respective ends of the document M (here, the document front edge Mt, the document left edge Ml, and the document right edge Mr) in the edge detection data De are grasped. If possible, it is possible to predict the range of the distance ρ where each end exists in the Hough transform data Dh.

  The offset setting data creation unit 5442 sets the position of the document left end Ml when the document skew angle θm is 0 ° among the documents M that can pass through the left end reference position Sl as the first left end position Ml1. . Next, the offset setting data creation unit 5442 determines the position of the document left end Ml when the document M is inclined to the leftmost side (θa = −3 °) among the documents M that can pass through the left end reference position S1. The left end second position Ml2 is set, and the position of the document left end Ml when the document M is tilted to the rightmost (θb = + 3 °) is set to the left end third position Ml3.

  Then, the offset setting data creation unit 5442 has a left end reference distance ρlstd obtained by Hough transform of the left end first position Ml1, and a left end minimum distance (left end offset distance) ρlmin obtained by Hough transform of the left end second position Ml2. The left end maximum distance ρlmax obtained by the Hough transform of the left end third position Ml3 is obtained.

Further, the offset setting data creation unit 5442 obtains the front end reference distance ρtstd, the front end minimum distance (front end offset distance) ρtmin, and the front end maximum distance ρtmax based on the front end reference position St, and further, the right end reference position. A right end reference distance ρrstd, a right end minimum distance (right end offset distance) ρrmin, and a right end maximum distance ρrmax are obtained based on Sr.
Thereafter, the offset setting data creation unit 5442 obtains the obtained front end reference distance ρtstd, front end minimum distance ρtmin, front end maximum distance ρtmax, left end reference distance ρlstd, left end minimum distance plmin, left end maximum distance ρlmax, right end reference distance ρrstd, right end minimum The distance ρrmin and the right end maximum distance ρrmax are output to the offset processing unit 545 as the offset setting data Do.

  FIG. 10 is a diagram for explaining the contents of processing in the offset processing unit 545 shown in FIG. Here, FIG. 10A shows the processing contents corresponding to the document front edge Mt, FIG. 10B shows the processing contents corresponding to the document left edge Ml, and FIG. 10C shows the document right edge Mr. The processing content corresponding to is shown. 10A to 10C, the Hough transform data Dh input from the Hough transform processing unit 543 to the offset processing unit 545, and the input from the offset setting unit 544 to the offset processing unit 545, respectively. Offset setting data Do (front end reference distance ρtstd, front end minimum distance ρtmin, front end maximum distance ρtmax, left end reference distance ρlstd, left end minimum distance ρlmin, left end maximum distance ρlmax, right end reference distance ρrstd, right end minimum distance ρrmin and right end maximum distance ρrmax) Shows the relationship. 10A to 10C show coordinate systems obtained by Hough transforming the xy image space into the θρ parameter space, in which the horizontal axis is the angle θ (°), and the vertical axis The axis is the distance ρ (au).

  In this embodiment, voting is performed in the Hough transform, and a plurality of sinusoidal patterns appear in the θρ coordinates shown in FIGS. 10 (a) to 10 (c). Then, the angle θ corresponding to the position where the number of votes is the largest in each is determined as the document skew angle θm in the document skew angle determination unit 547 (see FIG. 6).

  In the present embodiment, the offset processing unit 545 sets an offset for the Hough conversion data Dh input from the Hough conversion processing unit 543 using the offset setting data Do input from the offset setting unit 544. The obtained offset Hough transform data Di is output.

For example, in the example shown in FIG. 10A, the offset processing unit 545 has a range in which the distance ρ is not less than the front end minimum distance ρtmin and not more than the front end maximum distance ρtmax in the Hough transform data Dh corresponding to the document front end Mt (ρtmin ≦ Only ρ ≦ ρtmax) is extracted to create offset Hough transform data Di.
For example, in the example shown in FIG. 10B, the offset processing unit 545 has a range (within the Hough transform data Dh corresponding to the document left end Ml, the distance ρ is not less than the left end minimum distance ρlmin and not more than the left end maximum distance ρlmax. Only ρlmin ≦ ρ ≦ ρlmax) is extracted to create offset Hough transform data Di.
Further, for example, in the example shown in FIG. 10C, the offset processing unit 545 has a range (within the Hough transformed data Dh corresponding to the document right end Mr, a range in which the distance ρ is not less than the right end minimum distance ρrmin and the right end maximum distance ρrmax Only ρrmin ≦ ρ ≦ ρrmax) is extracted to create offset Hough transform data Di.

  Here, in the present embodiment, the position of each edge in the document M is predicted based on the edge detection data De obtained based on the reading result of the document M, and the offset setting data Do is determined based on the prediction result. Looking for. For this reason, for example, in the example shown in FIG. 10A, coordinates having the largest number of votes are present in the range of the Hough transform data Dh that is not less than the front end minimum distance ρtmin and not more than the front end maximum distance ρtmax. Yes. Further, for example, in the example shown in FIG. 10B, coordinates having the largest number of votes are present in a range of the Hough transform data Dh that is not less than the left end minimum distance ρlmin and not more than the left end maximum distance ρlmax. . Further, for example, in the example shown in FIG. 10C, the position with the largest number of votes is coordinated in the range of the Hough transform data Dh that is not less than the right end minimum distance ρrmin and not more than the right end maximum distance ρrmax. .

  In the skew amount calculation unit 54 of the present embodiment, the Hough conversion data storage unit 546 stores the offset Hough conversion data Di output from the offset processing unit 545, and the document skew angle determination unit 547 performs the Hough conversion. A document skew angle θm of the document M is determined based on the offset Hough conversion data Di read from the data storage unit 546.

  Here, the Hough conversion data Dh is stored in the Hough conversion data storage unit 546 as it is, and the original skew angle determination unit 547 reads the original of the original M based on the Hough conversion data Dh read from the Hough conversion data storage unit 546. It is also possible to determine the skew angle θm. However, in this case, in addition to data necessary for determining the document skew angle θm (data relating to coordinates with a large number of votes), data that is actually unnecessary for determining the document skew angle θm (the distance ρ is too small). In other words, the memory capacity required for the Hough transform data storage unit 546 is increased. As the data capacity increases, the time required for determining the document skew angle θm in the document skew angle determination unit 547 also increases.

  On the other hand, in the present embodiment, based on the reading result of the original M, a range in which each end of the original M can exist in the xy image space is predicted, and the xy image space is In the ρθ parameter space obtained by the conversion, a range in which the distance ρ corresponding to the document M can exist is predicted, and the offset Hough transform data obtained by extracting the range in which ρ can exist is extracted from the Hough transform data Dh. Since Di is stored in the Hough conversion data storage unit 546, an increase in memory capacity required for the Hough conversion data storage unit 546 can be suppressed.

  Further, since the offset Hough transform data Di is obtained based on the prediction result of the range that the distance ρ can take as described above, the offset Hough transform data whose capacity is reduced by extraction with respect to the Hough transform data Dh. The document skew angle θm can be obtained from the conversion data Di.

  Here, the front end reference position St, the left end reference position S1 and the right end reference position Sr corresponding to the document M are determined based on the edge detection data De obtained along with the reading of the target document M. However, it is not limited to this. For example, the positions of the front end reference position St, the left end reference position Sl, and the right end reference position Sr may be predicted based on the detection result by the document width detection sensor 21 provided in the document feeder 10.

  In the present embodiment, offset processing is performed on the Hough transform data Dh obtained by performing the Hough transform on the edge detection data De, thereby obtaining the offset Hough transform data Di with a reduced capacity. However, the method of reducing the capacity of the Hough conversion data Dh is not limited to this. For example, the edge detection data De may be subjected to offset processing before being subjected to Hough transform.

DESCRIPTION OF SYMBOLS 10 ... Document feeder, 40 ... Scanner device, 48 ... Light-receiving part, 49 ... Control / image processing unit, 50 ... Signal processing part, 51 ... Image pre-processing part, 52 ... Output destination setting part, 53 ... Reading image data storage 54: Skew amount calculation unit, 55 ... Skew correction processing unit, 56 ... Image post-processing unit, 70 ... Control unit, 541 ... Edge detection processing unit, 542 ... Edge detection data storage unit, 543 ... Hough transform processing unit, 544 ... Offset setting unit, 545 ... Offset processing unit, 546 ... Hough conversion data storage unit, 547 ... Document skew angle determination unit

Claims (6)

Reading means for reading an image of a conveyed document;
Detection means for performing edge detection processing on read image data read by the reading means to create edge image data;
Conversion means for creating Hough transform data expressed in a θρ parameter space using an angle θ and a distance ρ by performing a Hough transform on the edge image data;
Prediction means for predicting the range of the distance ρ in the Hough transform data;
Extraction means for extracting the range of the distance ρ predicted by the prediction means from the Hough transform data and creating extracted Hough transform data;
Storage means for storing the extracted Hough transform data;
Determining means for determining an inclination angle of the document based on the extracted Hough transform data read from the storage means;
An image reading apparatus comprising: correction means for performing inclination correction using the inclination angle on the read image data.   The predicting unit sets a reference value of the edge based on the position of the edge of the document acquired from the edge image data created by the detecting unit, and the distance ρ is set based on the reference value. The image reading apparatus according to claim 1, wherein the range is predicted. The prediction means limits the range of the angle θ;
The image reading apparatus according to claim 2, wherein the range of the distance ρ is predicted based on the reference value and the range of the angle θ. A detection unit for detecting the size of the document to be read;
The predicting unit sets a reference position of an edge of the document based on the size of the document detected by the detecting unit, and predicts the range of the distance ρ based on the reference position. The image reading apparatus according to claim 1. Detection means for performing edge detection processing on read image data obtained by reading an image of a conveyed document and creating edge image data;
Conversion means for creating Hough transform data expressed in a θρ parameter space using an angle θ and a distance ρ by performing a Hough transform on the edge image data;
Prediction means for predicting the range of the distance ρ in the Hough transform data;
Extraction means for extracting the range of the distance ρ predicted by the prediction means from the Hough transform data and creating extracted Hough transform data;
Storage means for storing the extracted Hough transform data;
An image processing apparatus including: a determination unit configured to determine an inclination angle of the document based on the extracted Hough conversion data read from the storage unit; On the computer,
A function of performing edge detection processing on read image data obtained by reading an image of a conveyed document and creating edge image data;
A function of creating Hough transform data expressed in a θρ parameter space using an angle θ and a distance ρ by performing a Hough transform on the edge image data;
A function of predicting the range of the distance ρ in the Hough transform data;
A function of extracting the predicted range of the distance ρ from the Hough transform data and creating extracted Hough transform data;
A function of storing the extracted Hough transform data in a memory;
A program for realizing a function of determining an inclination angle of the document based on the extracted Hough conversion data read from the memory.

Images