JP2010283665A - Image reading method and image reading system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image reading method and image reading system for correcting uneven luminance of a captured image and a change in color tone without using blank paper. <P>SOLUTION: A non-contact type image reading system includes: a document platen for placing an object to capture an image thereon; an image sensor for outputting a signal corresponding to the quantity of incident light; and a camera for imaging, in color, the document platen and a document placed on the document platen using the image sensor. A document platen image 14 and a blank paper image 13, which are imaged under the same condition, are used to calculate a blank paper luminance value with respect to a document platen luminance value, and hold it as a blank paper inference table 107. The document platen image 14 obtained by imaging the document platen and the blank paper inference table 107 are used for white balance data calculation 105 and shading data generation 106. The white balance data 108 are used for white balance control 110. The shading data 109 are used for shading correction 111. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an image reading method and an image reading system for reading, correcting, and processing office or home documents, forms such as slips, postal items, banknotes, and other three-dimensional objects with a camera using an area image sensor. .

Non-contact scanners that read documents placed upward from an image sensor fixed to a column from above are scanned when papers of various sizes and thicknesses are mixed compared to a flatbed scanner. Since it can reduce the time and effort required to do so, it is used in the window of financial institutions that handle a variety of forms.
In the non-contact type scanner device, brightness unevenness (Shade) such as unnecessary distortion or shadow is generated in the image data read according to the illumination condition or the position / shape of the document. Also, the color of the read image data changes due to the difference in the color of the ambient light. Therefore, it is necessary to remove the influence of ambient light by performing calibration. As a general calibration method, white balance data and shading data are acquired by placing a blank sheet on a platen and reading it. The white balance data is corrected using the white balance data, and the shading data is read. Is used to perform shading correction of the document image. However, it is time-consuming to put a blank paper every time the lighting environment changes, and there is a need to make it unnecessary because the blank paper has a problem of dirt and loss.
As a conventional technique, as disclosed in Patent Document 1, a shade of blank paper is estimated from read image data of a document table using the reflectivity of the document table to generate shading data.
In addition, as disclosed in Patent Document 2 and Patent Document 3, there is a device in which a display unit is provided on a reading table, white is displayed during calibration, and this is read to obtain correction data.

JP 2008-199556 A JP 2008-61170 A Japanese Patent Laid-Open No. 2004-200842

When white balance data is acquired using only the document table without using blank paper, white balance can be obtained with the luminance value of the document table, but white balance cannot be acquired in a high-luminance region such as blank paper.
Further, the shading data has a problem that the shadow generated on the document table has a lower luminance value than the shadow generated on the blank paper, and cannot be directly used for the shading correction of the blank paper.
The method of Patent Document 1 has a problem that it cannot be applied when the output value of the image sensor is nonlinear because the document table luminance value is multiplied by a uniform count.
In addition, the methods disclosed in Patent Document 2 and Patent Document 3 require a display unit and are costly, and when displaying white on the display unit, display it uniformly at a level that is read and uniform by a high-definition camera. There is a problem that is difficult.

  In order to solve these problems, an object of the present invention is to provide an image reading method and an image reading system that perform white balance adjustment and shading data acquisition without blank paper.

In order to solve the above problems, an image reading method according to the present invention includes a document table on which an imaging target is placed, an image sensor that outputs a signal corresponding to the amount of incident light, and the image sensor placed on the document table and the document table. In an image reading method using a camera for color photographing of a document, a blank sheet brightness value corresponding to each value of the document table brightness value is stored as a blank sheet estimation table, and the document table is captured to obtain a document table image. A blank sheet equivalent image obtained by converting the original table image using the blank sheet estimation table is generated, and white balance data is generated using the blank sheet equivalent image and the photographed original table image.
For example, the blank sheet estimation table is generated by calculating using a document table image and a blank sheet image captured under the same shooting conditions.
Further, for example, the white balance data is generated by calculating a coefficient for matching each of the R level, G level, and B level signal levels when shooting an achromatic color with a reference signal level.
In the image reading method of the present invention, the document table on which the object to be imaged is provided is provided with an out-of-frame black portion and a reference mark outside the document reading area, so that the blank sheet equivalent image and the photographed document table image are displayed. In addition, white balance data may be generated using the luminance value of the black portion outside the frame and the luminance value of the reference mark.
In the image reading method of the present invention, shading data is further generated using the white balance data and a blank sheet equivalent image obtained by converting the photographed document table image using the blank sheet estimation table.
Further, a document placed on the document table is photographed to obtain a document image, white balance correction is performed on the photographed document image using the white balance data, and the photographed document image is captured using the shading data. Shading correction is performed.

An image reading system according to the present invention includes an original platen on which an object to be imaged is placed, an image sensor that outputs a signal corresponding to the amount of incident light, and a camera that performs color photographing of the original platen placed on the original platen and the original platen by the image sensor. In the image reading system, a blank sheet estimation table that stores blank sheet brightness values corresponding to respective values of the document table brightness values, and a document sheet image obtained by photographing the document table using the blank sheet estimation table is converted into a blank sheet equivalent image. A white balance data generation unit for generating white balance data using the blank sheet equivalent image and the document table image, and a white of the document image obtained by photographing the document placed on the document table using the white balance data A white balance correction unit for performing balance correction.
In the image reading system of the present invention, the blank sheet estimation table is further calculated by calculating a blank sheet brightness value corresponding to each value of the document table brightness value using the document table image and the blank sheet image captured under the same shooting conditions. You may provide the blank paper estimation table production | generation part to produce | generate.
In the image reading system of the present invention, a shading data generating unit that generates shading data using the white balance data and a blank sheet equivalent image obtained by converting the photographed document table image using the blank sheet estimation table; And a shading correction unit that performs shading correction of the document image using the shading data.
In the image reading system of the present invention, the document table on which the imaging object is placed includes an out-of-frame black portion and a reference mark outside the document reading area, and the white balance data generation unit The white balance data may be generated using the luminance value of the black portion outside the frame and the luminance value of the reference mark in addition to the blank sheet equivalent image and the photographed original table image.

According to the present invention, it is possible to approximate the shading data using the blank read image data by performing the calculation so that the document table image becomes a blank image, and thus the shading data generated in this way is used for correction. Thus, an output equivalent to the shading correction using the blank paper read image data can be obtained. As a result, it is possible to save the trouble of placing a blank sheet at the time of proofreading and to improve the usability.

Functional block diagram of the image reading system of the present invention Schematic diagram of image reading device Hardware block diagram of image processing device Functional block diagram of the blank sheet estimation table generation unit of the present invention Diagram showing data retention format of blank sheet estimation table Blank sheet estimation table generation flowchart Functional block diagram of the white balance data generation unit of the present invention Diagram showing the data retention format of white balance data Blank paper equivalent image generation flowchart White balance data calculation flowchart White balance correction flowchart The figure which shows the output value before and after white balance adjustment at the time of achromatic photography Functional block diagram of the shading data generation unit of the present invention Shading data generation flowchart Shading correction flowchart Diagram showing output values before and after shading correction

Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.
FIG. 1 is a diagram illustrating functions of the image reading system. The image reading system includes an image reading device 11 and an image processing device 12.
The image reading device 11 includes an image acquisition unit 101, an image acquisition command issue unit 102, and a communication unit 103. The image acquisition unit 101 acquires an image using a camera. The image acquisition command issuing unit 102 receives an instruction from the image processing device 12 and issues an image acquisition command to the image acquisition unit 101. The communication unit 103 receives an image acquisition command from the image processing apparatus 12 and delivers an image to the image processing apparatus 12.
The image processing apparatus 12 includes a communication unit 302, a blank sheet estimation table generation unit 104, a white balance data generation unit 105, a shading data generation unit 106, a white balance correction unit 110, and a shading correction unit 111.
Processing stages in the image processing apparatus 12 include a process performed once at the time of initial registration, a process 17 performed every time ambient light changes, and a process 18 performed every time a document is read. One process performed at the time of initial registration includes generation of a blank sheet estimation table. Processing 17 performed every time ambient light changes includes white balance data generation and shading data generation. Processing 18 performed each time a document is read includes white balance correction and shading correction.
The communication unit 302 issues an image acquisition command to the image reading device 11, and acquires an image captured by the image reading device 11. The blank sheet estimation table generation unit 104 generates a blank sheet estimation table 107 using the blank sheet image 13 and the document table image 14 obtained through the communication unit 302. Details of the blank sheet estimation table generation unit 104 will be described with reference to FIGS. The white balance data generation unit 105 generates white balance data 108 using the blank sheet estimation table 107 and the document table image 14. The detailed description of the white balance data generation unit will be described with reference to FIGS. The shading data generation unit 106 generates shading data 109 using the document table image 14, the blank sheet estimation table 107, and the white balance data 108. A detailed description of the shading data generation unit 106 will be given with reference to FIGS. The white balance correction unit 110 performs white balance correction on the document image 15 using the white balance data 108. A detailed description of the white balance correction unit 110 will be given with reference to FIG. The shading correction unit 111 uses the shading data 109 to perform shading correction of the original image 15 after white balance correction. Details of the shading correction unit 111 will be described with reference to FIG. The document image 15 is subjected to white balance correction in the white balance correction unit 110 and is subjected to shading correction in the shading correction unit 111, whereby a corrected image 16 having no luminance unevenness and no change in hue is obtained.

FIG. 2 is a schematic diagram of the image reading apparatus 11 of the image reading system. The image reading apparatus 11 includes a camera unit 201, a column 202 that fixes the camera unit 201, a document table 203 on which an object to be imaged is placed, a reading region 205 within a camera imaging range 204 on the document table, and a reference mark 206. The document is placed in the reading area 205 and photographed. The read image data includes an image in the camera imaging range. That is, when the image is read, not only the reading area 205 but also the image data of the reference mark 206 outside thereof can be acquired.
The document table 203 is preferably a color that can be easily distinguished from the document, and is preferably black (reflectance R is 0.05 or less). The reference mark 206 is used as a reference when calculating white balance data. Here, for simplicity of explanation, the reference mark 206 is white (reflectance 1.0). However, if the reflectance is known, the reference mark 206 can be calculated by converting it to white based on the reflectance.
Conventionally, image data obtained by reading a blank sheet has been used as shading data. However, distortion such as unevenness of illumination such as shading also occurs in a non-white color. That is, distortion such as illumination unevenness can be removed even by using shading generated in a non-white color.
However, if the reflectance is low, the SN ratio is poor and only noise is generated. Therefore, it is conceivable that the reflectance R of the reading area 205 of the document table is 0.05 (5%) or more. If the platen reflectance R is 1.0, it is difficult to find the boundary between the document and the platen. Since the reflectivity of the original is not actually 1.0 (100%) and is considered to be about 90 to 75%, the reflectivity R of the reading area 205 of the original table is considered to be 75% or less.
By using the non-white (reflectance 5% or more and 75% or less) platen reading area 205 as a shading reading medium, blank paper can be eliminated.

FIG. 3 is a diagram illustrating a hardware configuration of the image processing apparatus 12 of the image reading system. The image processing apparatus includes an arithmetic device CPU 301, a communication unit 302, a power source 303, an external storage device 304 such as a hard disk drive, a keyboard 305, a display 306, and a memory 307.
The image processing device 12 issues a read command to the image reading device 11 and acquires an image through the communication unit 302. The CPU 301 interprets an image acquisition request from the user input through the keyboard 305 and issues it to the image reading apparatus 11 through the communication unit 302. The image acquisition request can be issued with a mouse using a graphical interface (GUI), or can be issued by providing an image acquisition switch in the image reading apparatus and pressing the switch. Further, the CPU stores the program stored in the external storage device 304 in the memory 307 and executes it, thereby correcting the image acquired through the communication unit 302 in accordance with the instruction of the program, and the corrected image 16 is stored in the external storage device. The information is stored in 306 or displayed on the display 306.

FIG. 4 is a block diagram showing the function of the blank sheet estimation table generation unit 104 that calculates a blank sheet estimation table 107 for converting the document table image 14 into an image equivalent to a blank sheet.
The blank sheet estimation table generation unit 104 includes a document table image acquisition unit 401, a reading area image extraction unit 402, a blank page image acquisition unit 403, a blank page area image extraction unit 404, a blank page estimation table calculation unit 407, and a blank page estimation table writing unit 408. . The document table image acquisition unit 401 acquires the document table image 14 from the image reading device 11 through the communication unit 302. A blank image acquisition unit 403 acquires the blank image 13 from the image reading device 11 through the communication unit 302. The reading area image extraction unit 402 extracts only the area of the reading area 205 as the reading area image 405 from the image of the camera imaging range 204 included in the document table image 14. The blank area image extraction unit 404 extracts only a blank area as a blank area image 406 from the image of the camera imaging range 204 included in the blank image 13. Although it is desirable to place the blank paper on the entire reading area, it may be placed on only a part of the reading area. When a blank sheet is placed in part, it is necessary to extract the same area as the blank area image 406 extracted from the blank image 13 as the reading area image 405 from the document table image 14. The blank page estimation table calculation unit 407 calculates the blank page estimation table 107 from the reading area image 405 and the blank page area image 406 and stores it in the memory 307. The blank sheet estimation table writing unit 408 stores the blank sheet estimation table 107 on the memory 307 in the external storage device 304.

  FIG. 5 is a diagram illustrating a holding format of the blank sheet estimation table 107, and the numerical values represent an example. The blank sheet estimation table holds the document table luminance value Ig501, the blank sheet luminance value Iw502 corresponding to the document table luminance value, and the data number Dg503 of data having the same document table luminance value.

Next, a procedure for generating the blank sheet estimation table 107 in the blank sheet estimation table calculation unit 407 will be described with reference to FIG. The blank sheet estimation table 107 is a table indicating input / output characteristics of the image reading apparatus 11, and is device-specific. Accordingly, the blank sheet estimation table may be calculated once at the time of factory shipment or at the time of initial calibration. In the following explanation, an example of obtaining the blank sheet estimation table by luminance is shown for the sake of simplicity, but it may be calculated for each output color (R, G, B, CMY, R, Gr, Gb, B, etc.) of the image reading device 11. good.
When starting the program 601, the blank sheet estimation table calculation unit 407 acquires 602 the reading area image 405 from the reading area image extraction unit 402, and acquires 603 the blank area image 406 from the blank area image extraction unit 404. Next, taking the coordinates where the upper left corner of the image is the origin, the main scanning direction is x, and the sub scanning direction is y, the luminance value Ig at the position (x, y) on the reading area image is acquired 604. As for the position (x, y), the origin (0, 0) is assigned as the initial value, and (0, 0) is assumed to be assigned as the initial value unless otherwise specified in the following description. Next, with reference to the blank sheet estimation table 407, a blank sheet brightness value Iw502 and the number of data Dg503 corresponding to Ig are acquired. For example, in the example of FIG. 5, when the document table luminance value Ig is 2, the blank paper luminance value Iw is 6 and the data number Dg is 25. The obtained blank sheet brightness value Iw is multiplied by the number of data Dg, and the result is held 606 as IwTemp. Next, in the blank area image 406, the brightness value IwNew at the position (x, y) is acquired 607, added to IwTemp 608, and 1 is added 609 to the number of data Dg. Then, IwTemp is divided by Dg, the result is stored in Iw 610, Iw is registered in the blank sheet brightness value 502 of the blank sheet estimation table 107, and Dg is registered 611 in the number of data 503, respectively. 1 is added to x 612, and it is determined whether x does not exceed the image width 613, and if not, steps 604 to 612 are repeated. When x exceeds the main scanning direction width (image width) of the image, 0 is substituted for x, and 1 is added 615 to y. It is determined whether y does not exceed the width of the image in the sub-scanning direction (image vertical width), and if not, steps 604 to 615 are repeated. When y exceeds the image height, the program ends 617.

Next, white balance adjustment and shading correction using the blank paper estimation table 107 will be described. White balance adjustment and shading data generation must be performed each time the ambient light changes. For example, it needs to be performed when an object is placed around the image reading device 11 or when a new illumination is installed.
White balance adjustment is a process of adjusting the output of each color (R, G, B) at the achromatic color portion to be the same. If white balance adjustment is performed using only the image on the platen, it is possible to achieve white balance at the platen's reflectivity (0.05 to 0.75), but to achieve white balance at the white plate's reflectivity (0.75 or more). In order to solve this problem, it is possible to adjust the white balance of the blank area by using an image equivalent to a blank sheet.

FIG. 7 is a block diagram illustrating functions of the white balance data generation unit 105 that generates white balance data. The white balance data generation unit 105 includes a platen image acquisition unit 801, a blank page estimation table acquisition unit 802, a black portion luminance value acquisition unit 803, a blank page equivalent image generation unit 804, a reference mark luminance value acquisition unit 805, and a reading area luminance value acquisition. A unit 806, an interpolation unit 807, and a white balance data writing unit 808.
The document table image acquisition unit 801 acquires the document table image 14 from the image reading device 11 through the communication unit 302. The blank sheet estimation table acquisition unit 802 acquires the blank sheet estimation table 107 stored in the external storage device 304. The blank sheet equivalent image generation unit 804 generates a blank sheet equivalent image 19 using the document table image 14 and the blank sheet estimation table 107. A black luminance value acquisition unit 803 acquires each color luminance value of the outer peripheral portion excluding the reading area 205 and the reference mark 206 in the camera imaging range 204. A reference mark luminance value acquisition unit 805 acquires each color luminance value of the reference mark 206. When a plurality of reference marks 206 are installed, each color luminance value may be handled for each reference mark or may be an average value. The reading area luminance value acquisition unit 806 acquires the color luminance values of the reading area 205 of the document table image 14 and the blank sheet equivalent image 19. The brightness value of the reading area may be obtained by using four corners, maximum value, minimum value, and if there is uneven brightness and the brightness value is changing, other brightness values for the reference color (for example, green) The number of data can be increased by acquiring luminance values of colors (for example, red and blue). The interpolation unit 807 interpolates the luminance values that could not be acquired with the luminance values acquired by the black portion luminance value acquisition unit 803, the reference mark luminance value acquisition unit 805, and the reading area luminance value acquisition unit 806. Interpolation methods include cubic spline interpolation, linear interpolation, Newton interpolation, and the like, and an arbitrary interpolation method can be selected in accordance with sensor characteristics. The white balance data writing unit 808 writes the generated white balance data 108 to the external storage device 304.

  FIG. 8 shows how the white balance data 108 is stored on the memory 307 and the external storage device 304. The white balance data 108 holds an input value and an output value for each color. There are three primary colors (R, G, B), Bayer type (R, Gr, Gb, B), subtractive color mixing (CMY), etc., but the expression method is the same. Here, the case of three primary colors (R, G, B) will be described. The white balance data 108 holds R white balance data 92, G white balance data 93, and B white balance data 94, respectively. The R input value 921 and the output values 922 and G corresponding to the input values are input. 931 and the output value 932 corresponding to the input value, the B input value 941 and the output value 942 corresponding to the input value, respectively.

First, the flow of processing performed by the blank paper equivalent image generation unit 804 will be described with reference to FIG. The processing in the blank page equivalent image generation unit 804 converts the document table image 14 into a blank page equivalent image using the blank page estimation table 107.
When the blank page equivalent image generation unit 804 starts 701, the original table image 14 is acquired 702 through the original table image acquisition unit 801, and the blank page estimation table 107 is acquired 703 through the blank page estimation table acquisition unit 802. Next, the brightness value Ig at the position (x, y) of the platen image 14 is acquired 704, the blank sheet brightness value Iw corresponding to Ig of the blank sheet estimation table 107 is acquired 705, and the brightness at the position (x, y) is acquired. Substitute 706 for Iw. For example, in the example of the blank page estimation table in FIG. 5, when the document table brightness value Ig is “2”, “6” is acquired as the blank page brightness value Iw, and “6” is set as the brightness value at the position (x, y). substitute. Next, 1 is added to x707, and it is determined 708 whether x does not exceed the image width, and if not, steps 704 to 707 are repeated. When x exceeds the image width, 0 is substituted 709 for x, 1 is added 710 to y, and it is determined 711 whether y does not exceed the image vertical width. If y does not exceed the image height, 704 to 710 are repeated. When y exceeds the image height, the program ends 712.
As described above, the document table image 14 can be converted into the blank sheet equivalent image 19.

Next, FIG. 10 shows a processing flow of the white balance data generation unit 105 using the blank sheet estimation table 107. White balance is a process that makes the output value of each color the same when shooting an achromatic color, and the same output for luminance Y and the method of matching other colors (red and blue) to the reference color (for example, green) There is a way to make it. Here, a method of matching other colors with green will be described. When starting the program 1001, the white balance data generation unit 105 acquires 1002 the document table image 14 through the document table image acquisition unit 801. Next, each black color luminance value (Rb, Gb, Bb) outside the frame is acquired 1003 in the black portion luminance value acquisition unit 803, and each color luminance value (Rw, Gw) of the reference mark 206 is acquired in the reference mark luminance value acquisition unit 805. , Bw) is acquired 1004, and the reading area luminance value acquisition unit 806 acquires the luminance value (Rg, Gg, Bg) of the reading area 1005. Next, the blank sheet estimation table 107 is acquired 703 through the blank sheet estimation table acquisition unit 802 to convert 1007 the original table image into a blank sheet equivalent image 19. Then, the reading area luminance value acquisition unit 806 acquires 1008 the luminance values (Rwe, Gwe, Bwe) of the reading area of the blank paper equivalent image 19. Next, a graph plotting red luminance values (0, Rb, Rg, Rw, Rwe, 255) on the X-axis and green luminance values (0, Gb, Gg, Gw, Gwe, 255) on the Y-axis. Generate 1009 and interpolate between plots by the interpolation unit 807 to obtain 1010 the output value 922 for the red input value 921 from 0 to 255. Similarly for blue, the blue brightness values (0, Bb, Bg, Bw, Bwe, 255) are plotted on the X axis, and the green brightness values (0, Gb, Gg, Gw, Gwe, 255) are plotted on the Y axis. The generated graph 1011 is generated, and the interpolation unit 807 performs interpolation between plots, thereby obtaining 1012 an output value 942 for an input value 941 of blue from 0 to 255. Since green is a reference signal, an output value 932 from 0 to 255 is obtained 1013 linearly with respect to an input value 931 from 0 to 255.
In this embodiment, interpolation is performed between plots including each color luminance value (Rb, Gb, Bb) of the black portion outside the frame and each color luminance value (Rw, Gw, Bw) of the reference mark. Further, it is characterized in that interpolation is performed between plots using the luminance value (Rg, Gg, Bg) of the reading area of the platen image 14 and the luminance value (Rwe, Gwe, Bwe) of the reading area of the blank sheet equivalent image 19. There is something. By using each color luminance value (Rb, Gb, Bb) of the black part outside the frame and each color luminance value (Rw, Gw, Bw) of the reference mark, more accurate interpolation can be performed.

  FIG. 10 shows the flow of processing when the reference signal is green, but the reference signal can be red. In that case, it is possible to calculate the output value 932 with respect to the green input value 931 by plotting and interpolating the green luminance value on the X axis in 1009 and the red luminance value on the Y axis, and in 1011 the X axis It is possible to calculate the output value 942 with respect to the blue input value 941 by plotting and interpolating the blue luminance value on the Y axis and the red luminance value on the Y axis. Even when the reference signal is blue, it is possible to obtain output values for the input values of the respective colors in the same procedure.

  FIG. 11 shows a flow of white balance correction processing in the white balance correction unit 110. White balance correction must be performed each time an image is read. When starting the program 1101, the white balance correction unit 110 acquires 1102 an arbitrary image 15 through the communication unit 302 and acquires 1103 the white balance data 108. Next, obtain 1104 each color luminance value (Rin, Gin, Bin) of the position (x, y) on the image, and output values (Rout, Gin, Bin) corresponding to the input values (Rin, Gin, Bin) of the white balance data 108 Gout, Bout) is acquired 1105. For example, in the example of FIG. 8, when the input value of each color is (1, 3, 2), the output value is (3, 3, 3). Thereafter, (Rout, Gout, Bout) is substituted 1106 as the luminance value at the position (x, y). Next, 1 is added to x 1107, and it is determined 1108 whether x does not exceed the image width. If not, 1104 to 1107 are repeated. When x exceeds the image width, 0 is substituted 1109 for x, 1 is added 1110, and it is determined 1111 whether y exceeds the image vertical width. If y does not exceed the image height, 1104 to 1110 are repeated. When y exceeds the image height, the program ends 1112.

  As described above, an image with white balance at each point on the image can be obtained. The concept of white balance correction is shown in FIG. FIG. 12 shows a graph in which each color luminance value at the center in the sub-scanning direction when a blank sheet is photographed is output along the main scanning direction. Each color output 1201 before white balance adjustment is different for each color R, G, and B, whereas each color output 1202 after white balance adjustment can obtain the same value.

Next, shading data generation and shading correction using the blank sheet estimation table 107 will be described with reference to FIGS.
FIG. 13 is a block diagram illustrating functions of the shading data generation unit 106. The shading data generation unit 106 includes a platen image acquisition unit 801, a blank page estimation table acquisition unit 802, a blank page equivalent image generation unit 804, a white balance correction unit 110, a white balance data acquisition unit 1301, a division unit 1302, and a shading data writing unit 1303. Consists of. The white balance data acquisition unit 1301 acquires the white balance data 108 stored in the external storage device 304. The dividing unit 1302 divides the luminance value from the reference value for each pixel of the white paper after the white balance correction. The shading data writing unit 1303 writes the generated shading data 109 to the external storage device 304.

  The processing flow of the shading data generation unit 106 will be described with reference to FIG. When the program is started 1401, the shading data generation unit 106 reads the document table image 14 through the document table image acquisition unit 801 1402, and acquires 703 the blank sheet estimation table 107 through the blank page estimation table acquisition unit 802. The document table image 14 is converted 1007 into a blank paper equivalent image 19 by the blank paper equivalent image generation unit 804, and then the white balance correction unit 110 performs white balance correction 1403. The procedure for white balance correction is the same as in FIG. Thereafter, the blank equivalent image 19 divides 1404 the luminance value from the reference value (for example, 255) in the division unit 1302. For all pixels, a value divided from the reference value for each color is calculated for each pixel, the result image is saved as shading data 109, and the program ends 1405. The shading data 109 may be generated for each color or for the luminance Y.

Next, the correction processing of the document image 15 will be described with reference to FIG. By performing the white balance correction 110 and the shading correction 111, the original image 15 can be obtained an image having no luminance unevenness without the influence of the illumination environment. The correction process needs to be performed every time the document image 15 is acquired.
The document image 15 obtained through the communication unit 302 is subjected to white balance correction by the white balance correction unit 110 through the process shown in FIG. The document image 15 after the white balance correction is subjected to a shading correction 1501 by multiplying the shading data 109 by each pixel. Thereby, it is possible to obtain a corrected image 16 without luminance unevenness.

  FIG. 16 is a conceptual diagram showing output values before and after shading correction when a blank sheet is read. The output value 1601 before shading correction is non-linear due to luminance unevenness, whereas the output value 1602 after shading correction can be made uniform by performing correction.

According to the present invention, white balance data and shading data can be acquired without using white paper, and white balance correction and shading correction of a document image can be performed. Thereby, the troublesome trouble of placing a blank sheet can be saved and the usability can be improved.
As a result, the present invention can read office and home documents, forms such as slips, postal items, banknotes, and other three-dimensional objects with a camera using an image sensor.

DESCRIPTION OF SYMBOLS 11 ... Image reading apparatus, 12 ... Image processing apparatus, 13 ... Blank paper image, 14 ... Original plate image, 15 ... Original image, 16 ... Corrected image, 17 ... Processing performed whenever ambient light changes, 18... Processing performed each time a document is read, 19.
DESCRIPTION OF SYMBOLS 101 ... Image acquisition part, 102 ... Image acquisition command issue part, 103 ... Communication part, 104 ... Blank paper estimation table generation part, 105 ... White balance data generation part, 106 ... Shading Data generation unit 107 ... Blank sheet estimation table 108 ... White balance data 109 ... Shading data 110 ... White balance correction unit 111 ... Shading correction unit
201 ... Camera unit, 202 ... Post, 203 ... Document table, 204 ... Camera imaging range, 205 ... Reading area, 206 ... Reference mark,
301 ... CPU, 302 ... communication unit, 303 ... power supply, 304 ... external storage device, 305 ... keyboard, 306 ... display, 307 ... memory,
401... Document platen image acquisition unit, 402... Reading region image acquisition unit, 403... Blank image acquisition unit, 404... Blank page image acquisition unit, 405.・ Blank area image, 407... Blank sheet estimation table calculation unit, 408... Blank sheet estimation table writing unit,
501... Original plate luminance value, 502... Blank paper luminance value, 503.
801... Document platen image acquisition unit, 802... Blank sheet estimation table acquisition unit, 803... Black portion luminance value acquisition unit, 804... White equivalent image generation unit, 805. 806 ... Reading area luminance value acquisition unit, 807 ... Interpolation unit, 808 ... White balance data writing unit,
92 ... R white balance data, 921 ... R input value, 922 ... R output value, 93 ... G white balance data, 931 ... G input value, 932. -G output value, 94 ... B white balance data, 941 ... B input value, 942 ... B output value,
1201 ... Each color output before white balance adjustment, 1202 ... Each color output after white balance adjustment,
1301 ... White balance data acquisition unit, 1302 ... Division unit, 1303 ... Shading data writing unit,
1601 ... Output value before shading correction, 1602 ... Output value after shading correction.

Claims (11)

In an image reading method using an original platen on which an object to be imaged is placed, an image sensor that outputs a signal corresponding to the amount of incident light, and a camera that shoots an original placed on the original platen and the original platen by the image sensor.
A blank sheet brightness value corresponding to each value of the document table brightness value is held as a blank sheet estimation table, and
Take an image of the platen, get an image of the platen,
Generating a blank sheet equivalent image obtained by converting the document table image using the blank sheet estimation table;
An image reading method, wherein white balance data is generated using the blank paper equivalent image. The image reading method according to claim 1.
An image reading method, wherein white balance data is generated using the blank sheet equivalent image and the photographed original platen image. The image reading method according to claim 2.
2. The image reading method according to claim 1, wherein the blank sheet estimation table is calculated using a document table image and a blank sheet image captured under the same shooting conditions. The image reading method according to claim 2.
2. The image reading method according to claim 1, wherein the white balance data is generated by calculating a coefficient for matching each of the R level, G level, and B level signal levels when shooting an achromatic color with a reference signal level. The image reading method according to claim 2.
The document table on which the imaging object is placed includes a black portion outside the frame and a reference mark on the outside of the reading area of the document,
An image reading method, wherein white balance data is generated using a luminance value of the black portion outside the frame and a luminance value of the reference mark in addition to the blank sheet equivalent image and the photographed original platen image. The image reading method according to claim 2, further comprising:
An image reading method, wherein shading data is generated using the white balance data and a blank sheet equivalent image obtained by converting the photographed document table image using the blank sheet estimation table. The image reading method according to claim 6.
Capture a document on the platen to obtain a document image,
Using the white balance data, correct the white balance of the photographed document image,
An image reading method, wherein shading correction of the photographed document image is performed using the shading data. In an image reading system including an original platen on which an imaging target is placed, an image sensor that outputs a signal corresponding to the amount of incident light, and a camera that performs color photographing of the original placed on the original platen and the original platen by the image sensor,
A blank sheet estimation table for storing blank sheet brightness values corresponding to the respective values of the document table brightness values;
A white balance data generating unit that converts a document table image obtained by photographing a document table using the blank sheet estimation table to generate a blank sheet equivalent image, and generates white balance data using the blank sheet equivalent image and the document table image; ,
An image reading system comprising: a white balance correction unit that performs white balance correction of an original image obtained by photographing an original placed on an original table using the white balance data. 9. The image reading system according to claim 8, further comprising:
A blank sheet estimation table generating unit that generates a blank sheet estimation table by calculating a blank sheet brightness value corresponding to each value of the document table brightness value using a document table image and a blank sheet image captured under the same shooting conditions; An image reading system characterized by the above. 9. The image reading system according to claim 8, further comprising:
A shading data generation unit that generates shading data using the white balance data and a blank sheet equivalent image obtained by converting the photographed document table image using the blank sheet estimation table;
An image reading system comprising: a shading correction unit that performs shading correction of the document image using the shading data. The image reading system according to claim 8.
The document table on which the imaging object is placed is provided with an out-of-frame black portion and a reference mark on the outside of the document reading area,
The white balance data generation unit generates white balance data using a luminance value of the black portion outside the frame and a luminance value of the reference mark in addition to the blank sheet equivalent image and the photographed original platen image. Image reading system.

Images