JP7146526B2 - Document reader - Google Patents

Description

The present invention relates to a document reading device for reading an image of a document.

For example, an image reading apparatus for reading an image of a document is applied to an image forming apparatus such as a copier or a facsimile machine. 2. Description of the Related Art In an image reading apparatus, shading correction is performed to correct luminance unevenness when reading an image of a document. Japanese Unexamined Patent Application Publication No. 2002-200003 proposes a technique related to an image reading device that performs shading correction. In this technology, images on both the front and back sides of a document can be read at once, and a front side reading section and a back side reading section are arranged to face each other with a scanning glass interposed therebetween. A white reference plate for correcting the optical system of the back surface reading unit is provided on the opposite surface of the glass from the document conveying surface at a position that does not interfere with the optical path of the front surface reading unit.

JP-A-2002-335380

Since the position of the shading correction plate (white reference plate) for correcting optical reading unevenness of the image sensor and lighting is different from the reading position of the document, there is a difference in the illumination luminance distribution between the two positions. For this reason, if shading correction is performed using a shading coefficient generated by reading the shading correction plate, unevenness or the like may occur in the original image. Therefore, by generating a shading coefficient based on the correlation between the data obtained by sampling the white reference plate and the data obtained by sampling a reference member different from the white reference plate placed on the document conveying surface, the unevenness of the document image can be reduced. It is possible to perform shading correction with reduced

However, if foreign matter such as dust adheres to the white reference plate or reference member, the data sampled when the white reference plate or reference member is sampled becomes abnormal data including the influence of dust or the like. If shading correction is performed based on the abnormal data, image streaks occur in the shading-corrected image, degrading the image quality of the read document image data.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a document reading apparatus that suppresses the influence of foreign matter when performing shading correction based on the result of reading a white reference plate at a position different from the document reading position.

In order to achieve the above object, the document reading apparatus of the present invention comprises: a document placing portion on which a document is placed; a document table glass; and a document placed on the document placing portion on the document table glass. Conveying means for conveying, reading means for reading the document conveyed by the conveying means, a reference plate provided on the opposite side of the platen glass to the side on which the original is conveyed, and storage means for storing correction data. shading correction means for performing shading correction using shading data on image data obtained by the reading means reading a document conveyed by the conveying means; and calculating the correction data and the shading data. and calculating means for calculating shading data from data obtained by reading the reference plate by the reading means and the correction data, and reading means for reading the reference plate. obtaining a plurality of ratios corresponding to each of a plurality of pixel positions based on the data obtained by reading the data obtained by reading the reference member installed on the side of the platen glass on which the document is conveyed, Detecting a pixel position that is a singular point from the plurality of ratios, determining the presence or absence of a pixel position that is the detected singular point in a target pixel position and surrounding pixel positions of the target pixel position, Based on the result of determination of presence/absence of a pixel position, the correction data is obtained from each ratio corresponding to the pixel position of interest and pixel positions surrounding the position of interest, and the pixel position which is the singular point when obtaining the correction data. In order not to use the ratio corresponding to The ratio is less than the number of each ratio used when determining the correction data when it is determined that there is no pixel position that is the detected singular point .

According to the present invention, it is possible to suppress the influence of foreign matter when performing shading correction based on the result of reading a white reference plate at a position different from the document reading position.

1 is a diagram illustrating a configuration example of an image reading apparatus according to an embodiment; FIG. 3 is a configuration diagram of a control system for controlling operations of the image reading apparatus; FIG. FIG. 4 is a diagram showing a reading state of a document reading unit; 4 is a flow chart showing the flow of a method for generating data to be stored in a backup memory; 4 is a graph showing luminance distributions in the main scanning direction of a point light source and a line light source; 4 is a graph showing ratio data and moving average values in the main scanning direction; 7 is a graph showing average value data and foreign matter determination data in the main scanning direction; 8A and 8B are a graph showing the relationship between the foreign matter determination data of each pixel and the foreign matter determination threshold value in the main scanning direction, and a graph showing the foreign matter determination data of the pixels near the foreign matter; 7 is a flow chart showing the flow of processing for generating a moving average flag; 7 is a flowchart showing the flow of moving average processing; 3A and 3B are a graph showing the ratio of each pixel in the main scanning direction and an enlarged graph of the graph; 4 is a flow chart showing the flow of operations for reading a document by an image reading apparatus; 4 is a graph of various data read by the CPU;

Hereinafter, each embodiment of the present invention will be described in detail with reference to the drawings. However, the configurations described in each embodiment below are merely examples, and the scope of the present invention is not limited by the configurations described in each embodiment.

FIG. 1 is a diagram showing a configuration example of an image reading apparatus 120 according to an embodiment. The image reading device 120 in FIG. 1 is equipped with an automatic document feed mechanism. Image reading device 120 has document tray 101 , document reading unit 106 , paper discharge tray 108 , document reading glass 119 , document pick roller 121 , document separation roller 122 , document transport roller 123 and document offset roller 124 . The image reading device 120 also has a white reference plate 125, a platen glass 126, and a reference member 500, which will be described later. A document tray 101 is a tray on which documents 102 to be read are stacked. Hereinafter, the manuscript 102 is assumed to be a blank manuscript.

The document reading unit 106 is reading means for reading the document 102 conveyed between the document reading glass 119 and the document platen glass 126 . The document reading unit 106 discharges the document 102 to the discharge tray 108 after reading the image of the document 102 is completed. A document transport motor 105 , which will be described later, drives a document pick roller 121 , a document separation roller 122 , a document transport roller 123 and a document offset roller 124 . As a result, the document 102 placed on the document tray 101 is fed, and the fed document 102 is conveyed on the platen glass 126 . A white reference plate 125 is arranged at a position facing (on the opposite side of) the document reading position P of the document 102 conveyed on the platen glass 126 as a transport path. The white reference plate 125 is a member used when generating shading coefficients.

The document reading unit 106 has a light guide 202 for irradiating the document 102 with light output from an LED light source 201 (to be described later), a CIS (Contact Image Sensor) 203, and a lens 204. . The CIS 203 is a line sensor that photoelectrically converts the reflected light from the document 102 guided by the lens 204 with a light receiving element and outputs an electric signal corresponding to the amount of incident light. With the above configuration, the image reading device 120 reads the image of the document 102 at the document reading position P while the document 102 is being transported.

FIG. 2 is a configuration diagram of a control system for controlling the operation of the image reading apparatus 120 according to this embodiment. The CPU 401 is control means for controlling the image reading device 120 , controls the LED light source 201 , the CIS 203 , the document conveying motor 105 and the backup memory 402 to read the image of the document 102 . An electric signal corresponding to the density of the document 102 read by the CIS 203 is converted from an analog electric signal to a digital image signal by the A/D conversion circuit 403 . A shading correction circuit 404 as correction means corrects the influence of non-uniformity in the amount of light of the LED light source 201 and the influence of variations in sensitivity of each pixel of the CIS 203 with respect to the digital image signal output from the A/D conversion circuit 403. . The backup memory 402 stores correlation data between illumination luminance distribution data of the white reference plate 125 and illumination luminance distribution data read from the reference member 500 at the document reading position P, and correction data (moving average value J(n)) described later. is a non-volatile memory for storing A CPU 401 as control means can read/write a backup memory 402 , and can input/output data to/from a shading correction circuit 404 via the CPU 401 . The reference member 500 is a member having a uniform density in the main scanning direction, and there is no limit to the brightness. The reference member 500 is, for example, a white member. A RAM 405 is a RAM used by the CPU 401 when temporarily storing data of the backup memory 402 and the shading correction circuit 404, and is an SRAM, for example. The CPU 401 can input/output data to/from the RAM 405 .

Next, shading correction performed by the shading correction circuit 404 will be described. The shading correction circuit 404 corrects (shading correction) the influence of non-uniformity of the light amount of the LED light source 201 and the influence of variations in sensitivity of each pixel of the CIS 203 . The shading correction circuit 404 of the embodiment performs the following calculation for each pixel of the CIS 203 on the digital image signal output from the A/D conversion circuit 403 .

"Shading correction output value (n)=document reading value (n)/shading coefficient (n)×reading target value" (Formula 1)

The value “n” in Equation 1 above indicates the pixel position of the CIS 203 . The read target value is the target value of the read value when the white reference document on the white reference plate 125 is read, and can be set to a predetermined value. A shading coefficient is a coefficient used for shading correction.

Under the control of the CPU 401, the shading correction circuit 404 can also output the document reading value (n) as it is without multiplying it by the shading coefficient. in this case,
"Shading correction output value (n)=original read value (n)". The operation of outputting the document reading value (n) as it is is hereinafter referred to as "shading correction OFF". The shading correction circuit 404 sets a shading coefficient generated based on the correlation data stored in the backup memory 402 from the CPU 401 to the shading correction circuit 404 . The correlation data corresponds to ratio data D(n), which will be described later.

FIG. 3 is a diagram showing a reading state of the document reading unit 106. As shown in FIG. FIG. 3A is a diagram showing that nothing is installed at the document reading position P, and that the document reading unit 106 is arranged to read the white reference plate 125 . FIG. 3B is a diagram showing that the reference member 500 is installed at the document reading position P, and the document reading unit 106 is arranged to read the reference member 500 . The reference member 500 is installed at the position shown in FIG. 5B by, for example, an operator in generating backup data, which will be described later.

FIG. 4 is a flow chart showing the flow of a method for generating data to be stored in the backup memory 402. As shown in FIG. When the CPU 401 detects a reading start signal indicating the start of reading the document 102, the CPU 401 performs reading settings (S1000). Specifically, the CPU 401 controls the lighting of the LED light source 201 and controls the reading of the CIS 203 . At this time, the CPU 401 performs control to turn off the shading correction. As a result, the shading correction circuit 404 outputs the document reading value (n) as it is without multiplying it by the shading coefficient. Then, the CPU 401 samples the data of the white reference plate 125 while the reference member 500 is not placed on the platen glass 126 (S1001). Specifically, read data A(n) of the white reference plate 125 is stored in the RAM 405 . The read data A(n) is illumination luminance distribution data (first data) of the white reference plate 125 .

Next, while the reference member 500 is placed on the platen glass 126, the CPU 401 samples the data of the reference member 500 (S1002). For example, after the data of the white reference plate 125 is sampled, the operator places the reference member 500 on the platen glass 126 and performs a reading start operation. Detecting this reading start operation as a trigger, the CPU 401 can start sampling the data of the reference member 500 . By the processing of S1002, the CPU 401 stores the read data C(n) of the reference member 500 in a different area of the RAM 405 from the read data A(n). Read data C(n) is illumination luminance distribution data (second data) obtained by reading the reference member 500 .

FIGS. 5A and 5B are graphs representing the luminance distribution in the main scanning direction, the broken line representing the read data A(n), and the solid line representing the read data C(n). . FIG. 5A is a graph showing luminance distribution in the main scanning direction (hereinafter referred to as main scanning luminance distribution) at different reading positions when the LED light source 201 is assumed to be a point light source. FIG. 5(b) is a graph showing the luminance distribution at different reading positions by the LED 201 as the linear light source and the light guide 202. FIG. As described above, read data A(n) is read data of the white reference plate 125 . The white reference plate 125 is farther from the document reading unit 106 than the reading position P by the thickness of the platen glass 126 . Therefore, the reading data A(n) indicated by the dashed line in FIG. 5B has a low absolute luminance and a long optical path length, so that the light emitted from the light guide 202 is sufficiently diffused in the main scanning direction. Due to this, it has the characteristic of becoming broad data in the main scanning direction. In addition, as shown in FIG. 5B, the read data A(n) has the characteristic that the luminance is lower at the edges than at the center in the main scanning direction. Considering the illumination luminance distribution by the light guide 202 indicated by the dashed line in FIG. 5B as an array of point light sources as indicated by the dashed line in FIG. n) is data in which the end portion is darker than the main scanning central portion.

On the other hand, read data C(n) indicated by a solid line is obtained by reading the reference member 500 at a position closer to the document reading unit 106 than the white reference plate 125 (reading position P). For this reason, the reading data C(n) has a high absolute luminance and a short optical path length, as indicated by the dashed line in FIG. 5(b). It has the characteristic of becoming data having a peak in the main scanning direction without being completely diffused. Further, the read data C(n) has the characteristic that the read data at the end portion is substantially flat with respect to the central portion in the main scanning direction, as shown by the solid line in FIG. 5(b). Considering the illumination luminance distribution by the light guide 202 as an array of point light sources indicated by the solid line in FIG. The read data becomes almost flat.

As shown in FIG. 4, CPU 401 divides read data A(n) by read data C(n) to generate ratio data D(n) (S1003). Therefore, the CPU 401 acquires the read data A(n) and the read data C(n) from the RAM 405 and performs the calculation "D(n)=C(n)/A(n)" in S1003. The CPU 401 stores the calculated ratio data D(n) in an area of the RAM 405 different from that of the read data A(n) and the read data C(n).

FIG. 6A is a graph showing pixel brightness in the main scanning direction. A solid line represents read data A(n), and a dotted line represents read data C(n). As described above, the read data A(n) is data obtained by reading the white reference plate 125 at a position farther than the reading position P with the document reading unit 106 as a reference. On the other hand, read data C(n) is data obtained by reading the reference member 500 placed at the reading position P. FIG. Therefore, the luminance of the read data A(n) is lower than the luminance of the read data C(n). Here, as shown in FIG. 6A, it is assumed that foreign matter such as dust adheres to the reference member 500 . The read data A(n) reflects the influence of the foreign matter. If foreign matter adheres to the white reference plate 125, the influence of the foreign matter is reflected in the read data C(n).

FIG. 6B is a graph showing ratio data D(n) in the main scanning direction. Since the read data A(n) reflects the influence of the foreign matter, the ratio data D(n) has a singular point reflecting the influence of the foreign matter. Further, read data A(n) and read data C(n) have the characteristics of the above-described waveform shape. Therefore, the ratio data D(n) has high brightness at the edges in the main scanning direction. Also, the ratio data D(n) is almost flat in the central portion in the main scanning direction, but has unevenness. As described above, the reference member 500 is described as a white member, but even if it is a halftone member or the like, the waveform shape itself in the main scanning direction does not change.

The CPU 401 divides the ratio data D(n) of each pixel of the CIS 203 into pixels in a predetermined range, and calculates an average value for each divided area (hereinafter referred to as main scanning direction area) (S1004). . Specifically, the calculation of Equation 2 below is performed for all pixel regions in the main scanning direction of the CIS 203 .

"X(k+1)=(D(kp+1)+D(kp+2)+...D((k+1)xp))/p"...(Formula 2)

In the above formula 2, "k" represents a constant from "0" to "m-1", "X(k)" represents the average value (average value data) for each main scanning area, and "m ” represents the number of main scanning area divisions, “n” represents the number of main scanning pixels, and “p” represents “p=n/m”.

In Equation 2 above, for example, if m=24 and n=5184, then p=216. In this case, X(1) to X(24) are as follows.
X(1)=(D(1)+D(2)+...D(216))/216
X(2)=(D(217)+D(218)+...D(432))/216
・・・
X(24)=(D(4969)+D(4970)+...D(5184))/216

Although X(3) to X(23) are omitted in the above, X(3) to X(23) are obtained by the above formula (2).

FIG. 7A is a graph showing average value data X(k) for each main scanning region in the main scanning direction. In the graph example of FIG. 7A, average value data X(1) to X(24) are represented by dashed lines. As described above, D(n) is not uniform data in the main scanning direction, so the CPU 401 finely divides the main scanning region and calculates average value data corresponding to the waveform shape in each region. .

Next, the CPU 401 subtracts the average value data X(k+1) from the ratio data D(n) for each pixel of the CIS 203 to generate the foreign matter determination data Y(n) as shown in the following equation 3 ( S1005).

"Y(n)=D(n)-X(k+1)" (Formula 3)

However, "n=1 to 5184", and the relationship between "n" and "p" is defined as follows. "k=0" when "n≤p", "k=1" when "p+1≤n≤2p", and finally "k=23" when "23p+1≤n≤24p". The same applies to "k=2 to 22". FIG. 7B shows the waveform of the foreign object determination data Y(n). The CPU 401 calculates, for each pixel, the difference between the average value data X(n) and the ratio data D(n) obtained for each predetermined main scanning area. Therefore, the CPU 401 can calculate the singular point corresponding to the foreign matter without depending on the waveform shape originally included in the ratio data D(n).

As described above, the ratio data D(n) is data that is almost flat in the central portion in the main scanning direction, but has unevenness. Since the foreign matter determination data Y(n) is the result of calculating the difference between the average value data X(n) and the ratio data D(n) for each pixel, the influence of the unevenness of the ratio data D(n) is reduced, and it becomes possible to calculate a singular point corresponding to the foreign matter.

The CPU 401 sets the foreign matter flag B(n) based on whether the absolute value of the foreign matter data Y(n) is equal to or greater than a predetermined threshold value (foreign matter determination threshold value F) for all pixels of the CIS 203 in the main scanning direction. Generate (S1006). That is, when "│Y(n)│<F", "B(n)=0" is obtained, and when "│Y(n)│≥F", "B(n)=1" is obtained. A foreign object determination threshold value F is a foreign object determination threshold value for detecting a foreign object. The CPU 401 stores the foreign substance flag B(n) generated for each pixel in the RAM 405 . FIG. 8(a) is a graph showing the relationship between the foreign matter determination data Y(n) of each pixel in the main scanning direction and the foreign matter determination threshold value F, and FIG. 8(b) is the graph of FIG. 8(a). Among them, foreign matter determination data Y(n) for pixels in the vicinity of the foreign matter is shown. In the graph example of FIG. 8B, Y(3000) to Y(3007) are below the foreign object determination threshold "-F". Therefore, the CPU 401 determines that the foreign matter flag B(n) of the pixels Y(3000) to Y(3007) is "1". That is, it is detected that there is a foreign object in the pixel.

Next, the CPU 401 generates a moving average flag H(n) according to the foreign matter flag B(n) (S1007). The figure is a flow chart showing the flow of processing for generating the moving average flag H(n). The CPU 401 sets the value "1" to "n" indicating the position of each pixel (S2000). The CPU 401 refers to the foreign substance flag B(n) of each pixel stored in the RAM 405 (S2001). The CPU 401 determines whether or not the value of the foreign object flag B(n) is "0" (S2002). If the determination in S2002 is NO, the foreign object flag B(n) has a value of "1", and the CPU 401 sets the moving average flag H(n) to a value of "1" (S2003). If the determination in S2002 is YES, the value of the foreign object flag B(n) of the target pixel "n" is "0", so the CPU 401 can recognize that there is no foreign object in the target pixel "n".

In the embodiment, the CPU 401 determines whether or not there is a foreign substance in predetermined pixels around the target pixel "n" (hereinafter, the predetermined pixels are 10 pixels) (S2004). The CPU 401 detects that the value of the foreign substance flag B(n) is "1" in the range of 10 pixels before and after the pixel of interest "n" even if the value of the foreign substance flag B(n) of the pixel of interest "n" is "0". It is searched whether there is a pixel with The range of pixels to be searched may be any number of pixels instead of 10 pixels. If the determination in S2004 is NO, the CPU 401 assumes that there is a foreign substance in the predetermined pixels surrounding the pixel of interest, and sets the moving average flag H(n) to "1" by performing the processing in S2003. . “n” is an integer from “1 to 5174”. Therefore, in the case of "n<10", the CPU 401 performs determination (search) in S2004 within the range of "1 pixel to n pixel", and in the case of "n≧5173", the CPU 401 searches "5173 pixel to n pixel". ” (search is performed) in S2004.

If the determination in step S2004 is YES, the CPU 401 sets the value "0" to the moving average flag H(n), assuming that there is no foreign matter in the pixel of interest itself or in the vicinity of the pixel of interest. As described above, the value "1" or "0" is set to the moving average flag H(n). The CPU 401 increments "n" (S2007) and moves the target pixel to the next pixel. The CPU 401 determines whether "n" has reached "5184" (S2008). If the determination in S2008 is NO, the flow moves to S2001. If it is determined YES in S2008, the number of pixels of interest has reached "5184 pixels", so the moving average flag generation process ends.

Next, the CPU 401 performs moving average processing on each pixel based on the moving average flag H(n). FIG. 10 is a flowchart showing the flow of moving average processing. The CPU 401 sets the value "1" to "n" (S3000). The CPU 401 refers to the moving average flag H(n) of each pixel stored in the RAM 405 (S3001). The CPU 401 determines whether or not the value of the moving average flag H(n) is "0" (S3002). If the determination in S3002 is NO, the value of the moving average flag H(n) is "0", and the CPU 401 assumes that there is no foreign matter in the pixel of interest and the surrounding pixels, and moves 10 pixels before and after the pixel of interest. An average calculation is performed (S3003).

Here, the calculation of the moving average in S3003 will be described. The CPU 401 performs the following calculation using ratio data D(n) of a total of 21 pixels, ie, the pixel of interest and 10 pixels before and after the pixel of interest.
k: target pixel (1 to 5184)
When J(k) is the moving average value of the k-th pixel and D(k) is the ratio data of the k-th pixel, the following three calculations are performed depending on the value of k. First, when "k=1 to 10", the following calculation (i) is performed.
(i) When k=1 to 10 When k=1 "J(1)=(D1×11+D1+D2+D3+D4+D5+D6+D7+D8+D9+D10)/21"
When k=2, "J(2)=(D1×10+D1+D2+D3+D4+D5+D6+D7+D8+D9+D10+D11)/21"
When k=3, "J(3)=(D1×9+D1+D2+D3+D4+D5+D6+D7+D8+D9+D10+D11+D12)/21"
・・・
When k=10, "J(10)=(D1×2+D1+D2+D3+D4+D5+D6+D7+D8+D9+D10+D11+D12+D13+...+D19)/21"
Also for k = 4 to 9, as the value of k increases, the value to be multiplied by D1 in the numerator decreases, and the total number of ratio data D(n) is 21, which is the same as above. .

Next, when "k=11 to 5173", the following calculation (ii) is performed.
(ii) when k = 11 to 5173

（ｉｉｉ）ｋ＝５１７４～５１８４の場合ｋ＝５１７４のときＪ（５１７４）＝（Ｄ５１８４×２＋Ｄ５１８３＋Ｄ５１８２＋Ｄ５１８１＋Ｄ５１８０＋Ｄ５１７９＋Ｄ５１７８＋Ｄ５１７７＋Ｄ５１７６＋Ｄ５１７５＋Ｄ５１７４＋Ｄ５１７３＋Ｄ５１７２＋Ｄ５１７１＋Ｄ５１７０＋Ｄ５１６９＋Ｄ５１６８＋Ｄ５１６７＋Ｄ５１６６＋Ｄ５１６５）/２１
When k=5175, J(5175)=(D5184×3+D5183+D5182+D5181+D5180+D5179+D5178+D5177+D5176+D5175+D5174+D5173+D5172+D5171+D5170+D5169/D5168+D5168+D5167)
When k=5184, J(5184)=(D5184×11+D5183+D5182+D5181+D5180+D5179+D5178+D5177+D5176+D5175+D5174)/21
For k=5176 to 5173, as the value of k increases, the value by which the numerator D5184 is multiplied increases, and the total number of ratio data D is 21, which is the same as above. The CPU 401 stores the moving average value J(n) obtained as described above in the RAM 405 as correction data.

When determined as NO in S3002, the CPU 401 determines that the moving average flag H(n) is "1". In other words, there is a foreign substance in the pixel of interest or in the range of 10 pixels before and after the pixel of interest. In this case, the CPU 401 selects a predetermined number of pixels from the pixel with the maximum ratio data D(n) value (maximum pixel) and the value of the ratio data D(n) among the 21 pixels searched in step S3004. A predetermined number of pixels are searched from the minimum pixel (minimum pixel). In the embodiment, the predetermined number of pixels to be searched is described as 8 pixels, but the predetermined number of pixels is not limited to 8 pixels. FIG. 11A is a graph showing the ratio data D(n) of each pixel in the main scanning direction. FIG. 11(b) is a graph showing an enlarged portion of FIG. 11(a) containing foreign matter. For example, when the target pixel is "n=3004", the eight pixels from the minimum pixel are D(3000) to D(3007). On the other hand, eight pixels from the maximum pixel are D (2999), D (3008) to D (3014).

When the target pixel is "n=3004", the range searched in S3004 is D(2994) to D(3014). The CPU 401 removes 8 pixels from the minimum pixel and 8 pixels from the maximum pixel from this range, and calculates the moving average using the remaining pixels (5 pixels) (S3005). The calculations performed by the CPU 401 are shown below.
k: target pixel (1 to 5184)
J(k): Moving average value of k-th pixel D(k): Ratio data of k-th pixel When "k=3004" is the target pixel described above, the CPU 401 calculates the moving average value J(3004) as follows. is calculated by Equation 4.

"J(3004)=(D2994+D2995+D2996+D2997+D2998)/5"...(Formula 4)

The CPU 401 stores the moving average value J(n) calculated in S3005 in the RAM405.

As described above, the moving average value J(n) is generated. The CPU 401 increments "n" (S3007) and moves the target pixel to the next pixel. The CPU 401 determines whether "n" has reached "5184" (S3008). If the determination in S2008 is NO, the flow moves to S3001. If it is determined YES in S3008, the number of pixels of interest has reached "5184 pixels", so the moving average flag generation process ends.

In S3003 described above, the CPU 401 calculates the moving average value J(n) using 10 pixels before and after the pixel of interest. It is not limited to 10 pixels before and after. As the number of pixels used to calculate the moving average value J(n) increases, the effect of foreign matter on the moving average value J(n) decreases. However, when the number of pixels used to calculate the moving average value J(n) increases, it becomes difficult to obtain the difference in illumination components (illumination luminance difference) due to different reading positions to be corrected. Therefore, it is preferable to set the number of pixels used for calculating the moving average value J(n) based on both the influence of the foreign matter and the image quality.

Also, in S3004 described above, the CPU 401 searches 10 pixels before and after the pixel of interest, but the number of pixels to be searched is not limited to 10 pixels before and after the pixel of interest. For example, the CPU 401 may detect the size of the foreign matter based on the foreign matter flag B(n), and set the number of pixels to be searched based on the size of the detected foreign matter. A portion of pixels in which the value of the foreign matter flag B(n) is continuously "1" corresponds to the size of the foreign matter. The CPU 401 may set the number of pixels to be searched according to the size of the portion of pixels in which the value of the foreign substance flag B(n) is continuously "1". In S3005, the CPU 401 calculates a moving average using the number of pixels remaining after removing a predetermined number of pixels from the number of pixels in the range searched in S3004. As a result, specific pixels are removed from the number of searched pixels, so that the CPU 401 can calculate the moving average using a plurality of pixels in which the effects of foreign matter are suppressed. The number of pixels searched in S3004 is preferably set to a small number that can reduce the influence of foreign matter.

As described above, the CPU 401 calculates the moving average value J(n) for each pixel. As shown in FIG. 4, the CPU 401 writes the calculated moving average value J(n) of each pixel to the backup memory 402 (S1009), and the process ends. As shown in FIG. 6C, the moving average value J(n) in the case where the influence of the foreign matter is not removed by the above processing has a width of It becomes a thick and expanded distribution. On the other hand, as shown in FIG. 6D, the moving average value J(n) for which various calculations are performed under the influence of the foreign matter, as in the embodiment, has a reduced influence of the foreign matter on the surrounding pixels. distribution. That is, when the data for correcting the illumination luminance difference due to the different reading positions is stored (backed up) in the backup memory 402, even if there is an influence of the foreign matter, the correction data that reduces the influence of the foreign matter is backed up. It becomes possible to

That is, by storing the correction data obtained by reducing singular points (foreign matter) from the ratio data D(n) in the backup memory 402, the correction data can be applied in the shading correction described later. By applying the above correction data during shading correction, it is possible to correct differences in illumination luminance and to suppress the influence of foreign matter during image reading of the document 102, which will be described later. In the above example, the moving average value J(n) for each pixel is stored in backup memory 402 . The moving average value J(n) is correction data for reducing singular points included in the ratio data D(n). The moving average value J(n) is stored in the backup memory 402 at the time of shipment from the factory, when the document reading unit 106 is replaced, or when the control board itself on which the backup memory 402 is mounted fails or is replaced. It may be implemented in time.

Next, the operation of reading the document 102 by the image reading device 120 will be described with reference to the flowchart of FIG. For example, when the image reading device 120 receives an operator's operation, a reading start signal is input to the CPU 401 . When the reading start signal is input to the CPU 401, the CPU 401 turns on the LED light source 201, controls the reading of the CIS 203, and turns off the shading correction (S4000). Then, the CPU 401 performs data sampling of the white reference plate 125 (S4001). Specifically, the read data P(n) (where “n” is the pixel position) of the white reference plate 125 is stored in the RAM 405 .

Then, the CPU 401 performs data calculation at the reading position P of the document 102 (S4002). In this case, the CPU 401 multiplies the reading data P(n) of the white reference plate 125 stored in the RAM 405 by the moving average value J(n) as the correction data stored in the backup memory 402 to obtain the brightness. Obtain distribution data Q(n). That is, the CPU 401 performs a calculation of “Q(n)=J(n)×P(n)” for each pixel and stores the luminance distribution data Q(n) in the RAM 405 . Then, the CPU 401 calculates a shading coefficient R(n) (n is a pixel position) based on Q(n) stored in the RAM 405 (S4003). Specifically, the CPU 401 calculates the reciprocal of the luminance distribution data Q(n) as the shading coefficient R(n). After that, the CPU 401 sets the shading coefficient R(n) in the shading correction circuit 404 (S4004). The shading correction circuit 404 performs the following calculations on the read values of the document 102 .

"Shading correction output value (n)=document reading value (n)/shading coefficient R(n)×reading target value" (Formula 5)

Then, the CPU 401 reads the document 102 (S4005). Specifically, the CPU 401 turns on the LED light source 201 , performs control to read the CIS 203 , and drives the document conveying motor 105 . Thereby, the document 102 is conveyed so as to pass the reading position P, and the document 102 is read by the document reading unit 106 .

FIG. 13 is a graph of various data read by the CPU 401. FIG. FIG. 13A shows a graph of the main scanning luminance distribution of the read data P(n) of the white reference plate 125 acquired in S1502. The horizontal axis is the main scanning position, and the vertical axis is the reading value. Since the shading correction is turned off, it is synonymous with sampling the illumination luminance distribution of the document reading unit 106 itself. FIG. 13B shows a graph of the main scanning luminance distribution of the luminance distribution data Q(n) acquired in S1503. The graph in FIG. 13B corresponds to data obtained by reading the reference member 500 at the document reading position P by the document reading unit 106 . Since "Q(n)=J(n)×P(n)", the graph of Q(n) in FIG. 13(b) compares with the graph of P(n) in FIG. The edges are corrected so that the read data is larger, and the center has an uneven luminance distribution. Since J(n) is moving-average ratio data in the main scanning direction after reducing the influence of foreign matter, the luminance distribution data Q(n) does not show any significant change in data.

FIG. 13C shows the waveform shape in the main scanning direction of the shading coefficient R(n) acquired in S1504. The shading coefficient R(n) is a coefficient (shading coefficient) for flatly correcting the main scanning luminance distribution of the luminance distribution data Q(n), so it is the reciprocal of Q(n). FIG. 13D shows the luminance distribution in the main scanning direction of the shading correction output values (output values) of the document 102 read in S1506. Since the reading value of the document 102 is multiplied by the shading coefficient R(n) for each pixel, an image that is flat in the main scanning direction can be read.

Although preferred embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments, and various modifications and changes are possible within the scope of the gist of the present invention. The present invention supplies a program that implements one or more functions of each of the above-described embodiments to a system or device via a network or a storage medium, and one or more processors of the computer of the system or device executes the program. It can also be realized by reading and executing processing. The invention can also be implemented by a circuit (eg, an ASIC) that implements one or more functions.

102 Document 106 Document reading unit 125 White reference plate 126 Document platen glass 201 LED light source 203 CIS
401 CPUs
402 backup memory 404 shading correction circuit

Claims (2)

a document placement section on which a document is placed;
platen glass,
a conveying means for conveying a document placed on the document placing portion to the document platen glass;
reading means for reading the document being conveyed by the conveying means;
a reference plate provided on the side of the platen glass opposite to the side on which the document is conveyed;
a storage means for storing correction data;
shading correction means for performing shading correction using shading data on image data obtained by the reading means reading a document conveyed by the conveying means;
computing means for computing the correction data and the shading data;
The computing means is
calculating shading data from the data obtained by reading the reference plate by the reading means and the correction data;
Based on the data obtained by reading the reference plate by the reading means and the data obtained by reading the reference member installed on the side of the platen glass on which the document is conveyed, each of a plurality of pixel positions is determined. find multiple ratios corresponding to
Detecting a pixel position that is a singular point from the plurality of ratios,
Determining whether or not there is a pixel position that is the detected singular point in the target pixel position and surrounding pixel positions of the target pixel position,
obtaining the correction data from each ratio corresponding to the pixel position of interest and the pixel positions surrounding the pixel position of interest based on the result of determining whether or not there is a pixel position that is a singular point;
The correction data is obtained when it is determined in the determination that there is a pixel position that is the detected singular point so that the ratio corresponding to the pixel position that is the singular point is not used when obtaining the correction data. that the number of ratios used in the determination is smaller than the number of ratios used in obtaining the correction data when it is determined that there is no pixel position that is the detected singular point in the determination . A document reader characterized by: The computing means is
If it is determined in the determination that there is a pixel position that is the detected singular point, a ratio having a maximum value and a minimum value in each ratio corresponding to the pixel position of interest and surrounding pixel positions of the pixel position of interest. 2. The document reading apparatus according to claim 1 , wherein said correction data is obtained without using a ratio .

Images