JP5132259B2 - Information processing apparatus and information processing method - Google Patents

Description

  The present invention relates to an information processing apparatus and an information processing method for detecting streak noise included in an image obtained by scanning the surface of a paper sheet with a line sensor or the like.

  Conventionally, a method for detecting or reducing streak noise included in an image obtained by scanning the surface of a paper sheet with a line sensor or the like has been proposed.

  For example, in Patent Document 1, the image data is smoothed by a predetermined length in the streak direction of the streak noise to obtain a first signal in which signal fluctuations in the streak direction are suppressed, and the first signal is obtained in a direction different from the streak direction. Smoothing the first signal, obtaining a second signal that suppresses the streak noise, taking a difference between the first signal and the second signal, obtaining a difference signal mainly including the streak noise, It describes that the streak noise in the image data is reduced by subtracting the difference signal from the image data.

Further, in Patent Document 2, an average value per pixel obtained from all pixel data output from a specific detector i in a line sensor and pixel data output from a detector in the vicinity of the detector i are obtained. It is described that the generation of streak noise is prevented by correcting the pixel data output from the detector i using the difference from the average value.
JP 2003-153004 A JP-A-7-75633

  However, in the method of Patent Document 1, when a streak-like object having a relatively short length and not noise appears in the image, this is erroneously removed.

  Further, the method of Patent Document 2 is not effective for an image having a small flat portion (an image having a large drop in values between individual pixels). It is also not effective when the streak noise component changes in a direction parallel to the streak direction.

  The present invention has been made in view of the above circumstances, and an information processing apparatus and information capable of accurately detecting streak noise included in an image obtained by scanning the surface of a paper sheet with a line sensor or the like. An object is to provide a processing method.

  An information processing apparatus according to the present invention is an arithmetic unit that calculates an evaluation value of noise for each pixel in an image for each color component in an information processing apparatus that obtains an image by scanning the surface of a sheet with a scanner. When the value of the pixel to be evaluated is larger than any of the two pixels located on both sides of the pixel to be evaluated in a predetermined direction, the value of the pixel to be evaluated and the two pixels The difference from the larger value of the two values is used as the evaluation value of the pixel to be evaluated. On the other hand, when the value is smaller than either of the two pixels, the smaller value of the two pixels is selected. An operation for performing, for each color component, processing for setting the difference between the value and the value of the pixel to be evaluated as the evaluation value of the pixel to be evaluated, and in other cases, setting 0 as the evaluation value of the pixel to be evaluated Means and individual images obtained by the computing means. For each color component, an adder that calculates a norm of a vector having the addition result of each color component as an element, and an individual obtained by the adder And determining means for determining that streak noise is generated in a pixel group in which a norm equal to or greater than a certain threshold is generated.

  The information processing apparatus according to the present invention is an arithmetic unit that calculates an evaluation value of noise for each pixel in an image in the information processing apparatus that obtains an image by scanning the surface of a paper sheet with a scanner, When the value of the pixel to be evaluated is larger than any of the two pixels located on both sides of the pixel to be evaluated in a predetermined direction, the value of the pixel to be evaluated and the value of the two pixels The difference between the larger value of the two pixels is used as the evaluation value of the pixel to be evaluated. On the other hand, when the value is smaller than either of the two pixels, the smaller value of the two pixels is The difference between the value of the pixel to be evaluated is the evaluation value of the pixel to be evaluated, and in other cases, 0 is used as the evaluation value of the pixel to be evaluated; The evaluation value of the pixel of the predetermined direction Addition means for adding in the vertical direction, and determination means for determining that streak noise has occurred in a pixel group that has produced an addition result equal to or greater than a certain threshold among the individual addition results obtained by the addition means. It is characterized by comprising.

  In addition, an information processing method according to the present invention is an information processing method for detecting streak noise included in an image obtained by scanning the surface of a paper sheet with a scanner, with respect to individual pixels in the image. A calculation step for calculating an evaluation value of noise for each color component, wherein the value of the pixel to be evaluated is larger than any of two pixels located on both sides of the pixel to be evaluated in a predetermined direction The difference between the value of the pixel to be evaluated and the larger value of the two pixels is set as the evaluation value of the pixel to be evaluated, and is smaller than any value of the two pixels. In this case, the difference between the smaller value of the two pixels and the value of the pixel to be evaluated is used as the evaluation value of the pixel to be evaluated. Processing to evaluate the pixel value A calculation step performed for each component, and a process of adding the evaluation value of each pixel obtained in the calculation step in a direction perpendicular to the predetermined direction for each color component, and a vector having the addition result of each color component as an element An addition step for calculating the norm of each of the two, and a determination step for determining that streak noise is generated in a pixel group that has generated a norm equal to or greater than a certain threshold among the individual norms obtained by the addition step. It is characterized by comprising.

  In addition, an information processing method according to the present invention is an information processing method for detecting streak noise included in an image obtained by scanning the surface of a paper sheet with a scanner, with respect to individual pixels in the image. In the calculation step of calculating the noise evaluation value, when the value of the pixel to be evaluated is larger than any of the two pixels located on both sides of the pixel to be evaluated in a predetermined direction, When the difference between the value of the pixel to be evaluated and the larger value of the two pixels is set as the evaluation value of the pixel to be evaluated, on the other hand, when the value is smaller than any of the two pixels, The difference between the smaller value of the two pixels and the value of the pixel to be evaluated is used as the evaluation value of the pixel to be evaluated. In other cases, 0 is the evaluation value of the pixel to be evaluated. And the calculation step An addition step of adding the evaluation values of the individual pixels obtained by the calculation step in a direction perpendicular to the predetermined direction, and an addition result of a certain threshold value or more is generated among the individual addition results obtained by the addition step. And a determination step for determining that streak noise is generated in the pixel group.

  According to the present invention, it is possible to accurately detect streak noise included in an image obtained by scanning the surface of a paper sheet with a line sensor or the like.

  Embodiments of the present invention will be described below with reference to the drawings.

  FIG. 1 is a diagram for explaining streak noise to be detected that is common to the embodiments of the present invention.

  A two-dimensional image obtained by scanning the surface of a paper sheet with a scanner such as a line sensor may contain streak noise caused by sensor sensitivity shift or dust. In many cases, streak noise appears as a line having a width of one pixel or several pixels when displayed on a screen of a display device. In the example of FIG. 1, streak noise appears as a line with a width of several pixels parallel to the sub-scanning direction (j-axis direction in FIG. 1) perpendicular to the main scanning direction (i-axis direction in FIG. 1). ing. In addition to the example of FIG. 1, streak noise may appear as a line parallel to the main scanning direction (i-axis direction in FIG. 1). In each embodiment, such streak noise is detected.

(First embodiment)
In the first embodiment of the present invention, RGB color system color images are handled. The color image signal has a value representing the degree of gradation (lightness or density) of each color component of the red component (R), green component (G), and blue component (B) for each pixel. This value is referred to as a “pixel value” in the present embodiment.

  FIG. 2 is a block diagram showing an example of the configuration of the information processing apparatus according to the first embodiment of the present invention.

  A line sensor 2 and an output device 3 are connected to the information processing apparatus 1. Note that the line sensor 2 and the output device 3 may be included in the information processing apparatus 1.

  The information processing apparatus 1 is realized as a personal computer (PC), for example, and detects streak noise included in a color image obtained by scanning the surface of the paper sheet S with the line sensor 2 and detects the noise. The result or the like can be displayed on the output device 3 or printed.

  The line sensor 2 is a scanner that optically scans the surface of the paper sheet S and performs photoelectric conversion to generate information described on the paper sheet S as a color image (or a single color grayscale image). There are, for example, a light source that irradiates light onto the paper sheet S and a self-scanning CCD image sensor that receives the reflected light and converts it into an electrical signal. An image signal generated by the line sensor 2 is supplied to the information processing apparatus 1.

  The output device 3 displays an image of the paper sheet S output from the information processing device 1, a detection result of streak noise, and the like.

  The above-described information processing apparatus 1 has, as various processing functions, a streak noise feature vector computation unit (calculation unit) 11, a streak noise evaluation value addition unit (addition unit) 12, and a streak noise determination unit (determination unit). 13 These functions are realized as a computer program that can be executed by a processor such as a CPU (Central Processing Unit). When the computer program is executed, a storage medium such as a memory (not shown) is used.

  The streak noise feature vector calculation unit 11 has a function of calculating an evaluation value of noise for each pixel in the image obtained from the line sensor 2 for each color component.

  In the streak noise feature vector calculation unit 11, the value of the pixel to be evaluated (the pixel in the i-th column) is positioned on both sides of the pixel to be evaluated in a predetermined direction (for example, the i-axis direction in FIG. 1). If the value is larger than any of the two pixels (i-1th column pixel, i + 1th column pixel), the evaluation target pixel (ith column pixel) as shown in FIG. And the difference between the larger value of the two pixels (pixels in the (i-1) th column and pixels in the (i + 1) th column) as the evaluation value of the evaluation target pixel, When the value of (pixel in i column) is smaller than the value of any of the two pixels (pixel in i-1 column, pixel in i + 1 column), as shown in FIG. The smaller one of the values of two pixels (the pixel in the (i−1) th column and the pixel in the (i + 1) th column) and the pixel to be evaluated (i The difference between the value of the eye pixels) as the evaluation value of the pixel of the evaluation, carried out 0 if other than the above processing to the evaluation value of the pixel of the evaluation, for each color component.

  The streak noise evaluation value addition unit 12 adds the evaluation value of each pixel obtained by the streak noise feature vector calculation unit 11 in a direction perpendicular to the predetermined direction (for example, the j-axis direction in FIG. 1). This is a function for performing processing for each color component and calculating a norm (length) of a vector having the addition result of each color component as an element.

  The streak noise determining unit 13 has a function of determining that streak noise is generated in a pixel group that has generated a norm equal to or greater than a certain threshold among the individual norms obtained by the streak noise evaluation value adding unit 12. It is.

  The streak noise feature vector computing unit 11 includes a streak noise evaluation value calculation unit 111, a norm calculation unit (norm calculation unit) 112, and a streak noise evaluation value adjustment unit (evaluation value adjustment unit) 113.

  The streak noise evaluation value calculation unit 111 has a function of calculating the noise evaluation value for each pixel in the image for each color component.

  The norm calculation unit 112 has a function of calculating a norm of a vector whose elements are evaluation values of each color component in each pixel obtained by the streak noise evaluation value calculation unit 111.

  The streak noise evaluation value adjustment unit 113 has a function of performing, for each color component, processing for reducing the evaluation value of a corresponding pixel when there is a norm exceeding a certain threshold among individual norms obtained by the norm calculation unit 112. is there.

  Here, the operation of the information processing apparatus 1 according to the embodiment will be described with reference to FIGS.

  First, the overall operation will be described with reference to FIG.

  When the information processing apparatus 1 inputs a color image captured by the line sensor 2 (step S11), the streak noise feature vector calculation unit 11 calculates an evaluation value of noise for each pixel in the image for each color component. (Step S12). Next, the streak noise evaluation value adding unit 12 performs, for each color component, a process of adding (projecting) the evaluation value of each pixel existing in each i column in the j-axis direction, and adds the addition result of each color component to an element. Is a streak noise evaluation value (step S13). Next, it is determined whether or not there is a streak noise evaluation value equal to or greater than the threshold (step S14). If it exists, a warning message indicating that streak noise has been detected and information indicating the streak noise area are output to the display device 3 (step S15). If not, streak noise has not been detected. The information shown is output to the display device 3 (step S16).

  Next, the details of step S12 in FIG. 5 will be described with reference to FIG.

One pixel in the image is selected, the streak noise evaluation value v _R for the red component of the pixel, the streak noise evaluation value v _G for the green component of the pixel, and the streak noise evaluation value v _B for the blue component of the pixel, Are respectively calculated (steps S21 to S23).

Next, the norm d of the vector (v _R , v _G , v _B ) is calculated (step S24). For example, the norm d can be calculated by the equation d = 0.30 v _R +0.59 v _G +0.11 v _B.

Next, it is determined whether or not the norm d is larger than the threshold value v _th (step S25). When the norm d is larger than the threshold value v _th , v _R , v _G and v _B are reduced so that the norm d becomes v _th . That is, a value obtained by calculating v _R × v _th / d is substituted for v _R (i, j), a value obtained by calculating v _G × v _th / d is substituted for v _G (i, j), and v _B A value obtained by calculating v _B × v _th / d is substituted for (i, j) (step S26).

On the other hand, when the norm d is not larger than the threshold value v _th , the norm d is not changed. That is, v _R , v _G , and v _B are directly substituted for v _R (i, j), v _G (i, j), and v _B (i, j) (step S27).

  Then, it is determined whether or not the processing for all the pixels has been completed (step S28). If not completed, the next pixel is selected (step S29), and the processing from step S21 is repeated. On the other hand, if the processing for all the pixels has been completed, the processing ends.

  Next, with reference to FIG. 7, the detail of the process common to step S21, step S22, and step S23 in FIG. 6 is demonstrated.

  In step S31, the pixel (i, j) is compared with the pixel (i-1, j). If the pixel (i, j) is large, the process proceeds to step S32. Otherwise, the process proceeds to step S33.

  In step S32, pixel (i, j) is compared with pixel (i + 1, j). If the pixel (i, j) is large, the process proceeds to step S34. Otherwise, the process proceeds to step S35.

  In step S33, the pixel (i, j) and the pixel (i + 1, j) are compared. If the pixel (i, j) is large, the process proceeds to step S35. Otherwise, the process proceeds to step S36.

  In step S34, the pixel (i-1, j) is compared with the pixel (i + 1, j). If the pixel (i-1, j) is large, the process proceeds to step S37. Otherwise, the process proceeds to step S38.

  In step S35, the streak noise evaluation value v is set to 0, and the process ends.

  In step S36, the pixel (i-1, j) is compared with the pixel (i + 1, j). If the pixel (i-1, j) is large, the process proceeds to step S38. Otherwise, the process proceeds to step S37.

  In step S37, the value obtained by subtracting the value of the pixel (i-1, j) from the value of the pixel (i, j) is substituted for the streak noise evaluation value v, and the process is terminated.

  In step S38, the value obtained by subtracting the value of the pixel (i + 1, j) from the value of the pixel (i, j) is substituted for the streak noise evaluation value v, and the process is terminated.

  Next, the details of step S13 in FIG. 5 will be described with reference to FIG.

  First, 1 is substituted for i (step S41).

Next, the position i _v R (i, j) the sum of all the j direction is substituted into _{P R} (step S42).

Then, the sum of all _v G (i, j) to the j-direction for position i, and substitutes the _{P G} (step S43).

Next, v _B (i, j) is added together in the j direction for position i, and assigned to P _B (step S44).

Next, the norm of the vector (P _R , P _G , P _B ) is calculated (step S45). For example, the norm is calculated by 0.30P _R + 0.59P _G + 0.11P _B. This norm becomes the streak noise evaluation value at the position i.

  Then, it is determined whether or not the processing for all i has been completed (step S46). If not completed, 1 is added to i (step S47), and the processing from step S42 is repeated. On the other hand, if the processing for all i has been completed, the processing ends.

  The above description shows an example in which streak noise having a width of one pixel is detected. When detecting streak noise having a width of 2 pixels or more, by replacing i−1 and i + 1 in steps S31 to S38 in FIG. 7 with i−n and i + n, respectively, the width 2n−1. It becomes possible to detect streak noise.

  As described above, according to the first embodiment, it is possible to accurately detect streak noise included in a color image and output a highly accurate detection result to a display device or the like.

(Second Embodiment)
In the second embodiment of the present invention, a monochrome image (gray image or the like) is handled. The signal of the monochromatic image has a value representing the degree of gradation (lightness or density) based on, for example, a gray scale for each pixel. This value is referred to as a “pixel value” in the present embodiment.

  FIG. 9 is a block diagram showing an example of the configuration of the information processing apparatus according to the second embodiment of the present invention. Elements that are the same as those in FIG. 2 are given the same reference numerals, and descriptions of their functions and operations are omitted.

  The information processing apparatus 1 detects streak noise included in a monochromatic image obtained by scanning the surface of the paper sheet S with the line sensor 2 and causes the output device 3 to display or print the detection result. be able to.

  The information processing apparatus 1 described above has, as various processing functions, a streak noise feature value calculation unit (calculation unit) 21, a streak noise evaluation value addition unit (addition unit) 22, and a streak noise determination unit (determination unit). 23. These functions are realized as a computer program that can be executed by a processor such as a CPU (Central Processing Unit). When the computer program is executed, a storage medium such as a memory (not shown) is used.

  The streak noise feature value calculation unit 21 has a function of calculating a noise evaluation value for each pixel in the image obtained from the line sensor 2.

  In the streak noise feature value calculation unit 21, the value of the pixel to be evaluated (the pixel in the i-th column) is positioned on both sides of the pixel to be evaluated in a predetermined direction (for example, the i-axis direction in FIG. 1). If the value is larger than any of the two pixels (i-1th column pixel, i + 1th column pixel), the evaluation target pixel (ith column pixel) as shown in FIG. And the difference between the larger value of the two pixels (pixels in the (i-1) th column and pixels in the (i + 1) th column) as the evaluation value of the evaluation target pixel, When the value of (pixel in i column) is smaller than the value of any of the two pixels (pixel in i-1 column, pixel in i + 1 column), as shown in FIG. The smaller value of the two pixels (i-1th column pixel, i + 1th column pixel) and the evaluation target pixel (ith column pixel) The difference between the value of the element) as the evaluation value of the pixel of the evaluation, the 0 in other cases it is the evaluation values of the pixels of the evaluation.

  The streak noise evaluation value addition unit 22 adds the evaluation values of the individual pixels obtained by the streak noise feature value calculation unit 21 in a direction perpendicular to the predetermined direction (for example, the j-axis direction in FIG. 1). It is a function.

  The streak noise determination unit 23 determines that streak noise is generated in a pixel group that has generated an addition result equal to or greater than a certain threshold among the individual addition results obtained by the streak noise evaluation value addition unit 22. It is a function to do.

  The streak noise feature value calculation unit 21 includes a streak noise evaluation value calculation unit 211 and a streak noise evaluation value adjustment unit (evaluation value adjustment unit) 212.

  The streak noise evaluation value calculation unit 211 has a function of calculating a noise evaluation value for each pixel in the image.

  The streak noise evaluation value adjustment unit 113 is a function of reducing the evaluation value of a corresponding pixel when there is an evaluation value exceeding a certain threshold among the individual evaluation values obtained by the streak noise evaluation value calculation unit 211. .

  Here, with reference to FIGS. 10 to 12, the operation of the information processing apparatus 1 according to the embodiment will be described.

  First, the overall operation will be described with reference to FIG.

  When the information processing apparatus 1 inputs a monochromatic image captured by the line sensor 2 (step S51), the streak noise feature value calculation unit 21 calculates an evaluation value of noise for each pixel in the image (step S52). . Next, the streak noise evaluation value adding unit 22 performs a process of adding (projecting) the evaluation value of each pixel existing in each i column in the j-axis direction, and the addition result of each color component is used as the streak noise evaluation value. (Step S53). Next, it is determined whether or not there is a streak noise evaluation value equal to or greater than the threshold (step S54). If it exists, a warning message indicating that streak noise has been detected and information indicating the streak noise region are output to the display device 3 (step S55). If not, streak noise has not been detected. The information shown is output to the display device 3 (step S56).

  Next, the details of step S52 in FIG. 10 will be described with reference to FIG.

  One pixel in the image is selected, and the streak noise evaluation value v of that pixel is calculated (step S61).

Next, it is determined whether or not the streak noise evaluation value v is larger than the threshold value v _th (step S62). When the streak noise evaluation value v is larger than the threshold value v _th , v _th is substituted for v (i, j) so that the streak noise evaluation value v becomes v _th (step S63). On the other hand, when the streak noise evaluation value v is not larger than the threshold value v _th , the streak noise evaluation value v is not changed. That is, v is substituted for v (i, j) as it is (step S64).

  Then, it is determined whether or not the processing for all the pixels has been completed (step S65). If not completed, the next pixel is selected (step S66), and the processing from step S61 is repeated. On the other hand, if the processing for all the pixels has been completed, the processing ends.

  The details of step S61 described above are the same as the processing described in FIG.

  Next, the details of step S53 in FIG. 10 will be described with reference to FIG.

  First, 1 is substituted for i (step S71).

  Next, all v (i, j) are added in the j direction for the position i (step S72). The added value becomes the streak noise evaluation value at the position i.

  Then, it is determined whether or not the processing for all i has been completed (step S73). If not completed, 1 is added to i (step S74), and the processing from step S72 is repeated. On the other hand, if the processing for all i has been completed, the processing ends.

  The above description shows an example in which streak noise having a width of one pixel is detected. When detecting streak noise having a width of 2 pixels or more, by replacing i−1 and i + 1 in steps S31 to S38 in FIG. 7 with i−n and i + n, respectively, the width 2n−1. It becomes possible to detect streak noise.

  As described above, according to the second embodiment, it is possible to accurately detect streak noise included in a single-color image and output a highly accurate detection result to a display device or the like.

  In the third and fourth embodiments described below, when streak noise having a width of w pixels in the i-axis direction is detected as shown in FIG. 13, the streak noise is detected. A method for effectively reducing the above will be described. For convenience, the position coordinates of the pixel in the upper left corner of the pixel group related to the streak noise in FIG. 13 will be described as (0, 0). In order to reduce streak noise, the value of each pixel included in the pixel group in which the streak noise is generated is corrected. In this correction, an image of a streak pattern erroneously detected at the time of erroneous detection is obtained. The amount of correction is limited so as not to erase the image, and it does not give a sense of incongruity to a person viewing an image output to a display device or the like.

(Third embodiment)
In the third embodiment, a color image of the RGB color system is handled as in the case of the first embodiment described above.

  FIG. 14 is a block diagram illustrating an example of the configuration of the information processing apparatus according to the third embodiment of the present invention. Elements that are the same as those in FIG. 2 are given the same reference numerals, and descriptions of their functions and operations are omitted.

  The information processing apparatus 1 includes the streak noise feature vector calculation unit 11, the streak noise evaluation value addition unit 12, and the streak noise determination unit 13 as various processing functions, and further a streak noise reduction unit. (Noise reduction means) 14 is provided.

  The streak noise reduction unit 14 has a function of correcting the value of each color component in each pixel included in the pixel group determined to have streak noise generated by the streak noise determination unit 13 described above. The streak noise reduction unit 14 uses a vector having each color component value in the correction target pixel as an element, and uses each color component value in a pixel located outside the pixel group in which the streak noise is generated as an element. Based on the vector, the correction is performed up to a certain correction amount.

  The information processing apparatus 1 described above can display or print the image after the correction by the streak noise reduction unit 14 on the output device 3.

  Here, a vector (correction target vector) having the values of the respective color components in the correction target pixel (a + i, j) included in the pixel group related to the streak noise in FIG. 13 as P (a + i, j). . In addition, let P (a-1, j) be a vector whose elements are the values of the respective color components in the pixel (a-1, j) adjacent to the left side of the pixel group related to the streak noise, and the pixel ( Let P (a + w, j) be a vector whose elements are the values of the respective color components in a + w, j). FIG. 15 shows the relationship between these vectors in the RGB color space.

M in FIG. 15 is a target vector (correction target vector) in the correction process of the correction target vector P (a + i, j). The tip of the correction target vector M is located on a line connecting the tip of P (a-1, j) and the tip of P (a + w, j). At this time, the closer the position of the correction target pixel (a + i, j) in FIG. 13 is to the position of the pixel (a-1, j), the closer the position of the tip of the correction target vector M is to P (a− 1, j). On the other hand, the closer the position of the correction target pixel (a + i, j) in FIG. 13 is to the position of the pixel (a + w, j), the more the position of the tip of the correction target vector M is P (a + w, j). Close to the tip position. H is a vector (correction vector) when it is assumed that the correction target vector P (a + i, j) is corrected to the correction target vector M. P ′ is a vector (corrected vector) when the correction target vector P (a + i, j) is corrected with a certain correction amount threshold value h _th as a limit.

  Here, with reference to FIG. 16 and FIG. 17, the operation of the streak noise reduction unit 14 in the information processing apparatus 1 according to the embodiment will be described.

  First, the overall operation will be described with reference to FIG.

  First, 0 is substituted for i (step S81), and 0 is substituted for j (step S82). Next, a color correction process is performed on the correction target pixel (a + i, j) (step S83). Next, j is compared with J (the number of pixels arranged in the j-axis direction in FIG. 13) (step S84). If j <J, 1 is added to j (step S85), and the processing from step S83 is repeated. On the other hand, if not, i and w are compared (step S86). If i <w, 1 is added to i (step S87), and the processing from step 82 is repeated. On the other hand, if not, the process ends.

  Next, details of step S83 in FIG. 16 will be described with reference to FIG.

  First, the correction target vector M is calculated by the following formula (step S91).

M = {(w−i) × P (a−1, j) + (i + 1) × P (a + w, j)} / (w + 1)
Next, the correction vector H is calculated by the following formula (step S92).

H = MP (a + i, j)
Next, the correction amount h is calculated (step S93). The correction amount h is the norm of the correction vector H. As an example, the norm is calculated by the following formula.

h = 0.30h _R + 0.59h _G + 0.11h _B
Here, h _R , h _G and h _B are three components of the correction vector H.

Next, the correction coefficient k is calculated (step S94). When the correction amount h is larger than the correction amount threshold h _th , the correction coefficient k = h _th / h is set. Otherwise, the correction coefficient k = 1.

  Next, the corrected vector P ′ is calculated by the following equation (step S95), and the process is terminated.

P ′ = P (a + i, j) + k × h
As described above, according to the third embodiment, in addition to the effects obtained in the first embodiment, the correction amount is limited in correction so as not to erase the image of the streak pattern erroneously detected at the time of erroneous detection. Thus, it is possible to prevent a person who sees the image output to the display device or the like from feeling uncomfortable.

(Fourth embodiment)
In the fourth embodiment of the present invention, a monochrome image (gray image or the like) is handled as in the case of the third embodiment described above.

  FIG. 18 is a block diagram showing an example of the configuration of an information processing apparatus according to the fourth embodiment of the present invention. In addition, the same code | symbol is attached | subjected to the element which is common in FIG. 9, and the description of the function and operation | movement is abbreviate | omitted.

  The information processing apparatus 1 includes the above-described streak noise feature value calculation unit 21, the streak noise evaluation value addition unit 22, and the streak noise determination unit 23 as various processing functions, and further a streak noise reduction unit. (Noise reduction means) 24 is provided.

  The streak noise reduction unit 24 has a function of correcting the value of each pixel included in the pixel group determined to have streak noise generated by the streak noise determination unit 13 described above. The streak noise reduction unit 14 corrects the value of the pixel to be corrected based on the value of a pixel located outside the pixel group in which the streak noise is generated, with a certain correction amount as a limit.

  The information processing apparatus 1 described above can display or print the image after the correction by the streak noise reduction unit 24 on the output device 3.

  Here, the value of the correction target pixel (a + i, j) included in the pixel group related to the streak noise in FIG. 13 is assumed to be p (a + i, j). Further, the value of the pixel (a-1, j) adjacent to the left side of the pixel group related to the streak noise is p (a-1, j), and the value of the pixel (a + w, j) adjacent to the right side is p. Let (a + w, j). The relationship between these values is shown in FIG.

M in FIG. 19 is a target value (correction target value) in the correction process of the correction target value p (a + i, j). This correction target value m is located on a line connecting p (a−1, j) and p (a + w, j). At this time, as the position of the correction target pixel (a + i, j) in FIG. 13 is closer to the position of the pixel (a-1, j), the position of the correction target value m is p (a-1, j) close to the position. On the other hand, the closer the position of the correction target pixel (a + i, j) in FIG. 13 is to the position of the pixel (a + w, j), the closer the position of the correction target value m is to the position of p (a + w, j). Get closer. h is a correction amount when it is assumed that the correction target value p (a + i, j) is corrected to the correction target value. Further, p ′ is a correction amount when the correction target value p (a + i, j) is corrected with a certain correction amount threshold value h _th as a limit.

  Here, with reference to FIG. 20 and FIG. 21, the operation of the streak noise reduction unit 24 in the information processing apparatus 1 according to the embodiment will be described.

  First, the overall operation will be described with reference to FIG.

  First, 0 is substituted for i (step S101), and 0 is substituted for j (step S102). Next, a gray correction process is performed on the correction target pixel (a + i, j) (step S103). Next, j is compared with J (the number of pixels arranged in the j-axis direction in FIG. 13) (step S104). If j <J, 1 is added to j (step S105), and the processing from step S103 is repeated. On the other hand, if not, i and w are compared (step S106). If i <w, 1 is added to i (step S107), and the processing from step 102 is repeated. On the other hand, if not, the process ends.

  Next, details of step S103 in FIG. 20 will be described with reference to FIG.

  First, the correction target value m is calculated by the following formula (step S111).

m = {(w−i) × p (a−1, j) + (i + 1) × p (a + w, j)} / (w + 1)
Next, the correction amount m is calculated by the following equation (step S112).

h = | mp (a + i, j) |
Next, the correction amount h is compared with the correction amount threshold h _th (step S113). If h> h _th , p (a + i, j) is compared with the correction target value m (step S114). Otherwise, the corrected pixel value p ′ is set to m (step S115), and the process ends.

  If p (a + i, j)> m in step S114, the corrected pixel value p ′ is calculated by the following equation (step S116), and the process ends.

p ′ = p (a + i, j) −h _th
Otherwise, the corrected pixel value p ′ is calculated by the following formula (step S117), and the process is terminated.

p ′ = p (a + i, j) + h _th
As described above, according to the fourth embodiment, in addition to the effects obtained in the second embodiment, the correction amount is limited so that the image of the streak pattern erroneously detected at the time of erroneous detection is not erased. Thus, it is possible to prevent a person who sees the image output to the display device or the like from feeling uncomfortable.

  Further, according to each of the above-described embodiments, it is possible to prevent the streak noise having a relatively short length from being removed by mistake. Further, according to each of the above-described embodiments, streak noise can be detected or reduced even in an image with few flat portions. Further, according to each of the embodiments described above, it is possible to detect or reduce the streak noise that changes in a direction parallel to the streak direction.

  The various processing procedures according to the present invention described in the above-described embodiments are stored in a storage medium (for example, a magnetic disk, an optical disk, or a semiconductor memory) that can be read by a computer (information processing apparatus) as a computer program. If necessary, it may be read and executed by the processor. Such a computer program can also be distributed by transmitting from one computer to another computer via a communication medium.

  The present invention is not limited to the above-described embodiments as they are, and can be embodied by modifying the constituent elements without departing from the scope of the invention in the implementation stage. In addition, various inventions can be formed by appropriately combining a plurality of components disclosed in the embodiment. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, constituent elements over different embodiments may be appropriately combined.

The figure for demonstrating the stripe noise of the detection target common to each embodiment of this invention. 1 is a block diagram showing an example of the configuration of an information processing apparatus according to a first embodiment of the present invention. The figure which shows the 1st example for demonstrating the method of calculating an evaluation value. The figure which shows the 2nd example for demonstrating the method of calculating an evaluation value. 6 is a flowchart showing an overall operation of the information processing apparatus according to the embodiment. The flowchart which shows the detail of step S12 in FIG. The flowchart which shows the detail of the process common to step S21, step S22, and step S23 in FIG. The flowchart which shows the detail of step S13 in FIG. The block diagram which shows an example of a structure of the information processing apparatus which concerns on the 2nd Embodiment of this invention. 6 is a flowchart showing an overall operation of the information processing apparatus according to the embodiment. 11 is a flowchart showing details of step S52 in FIG. The flowchart which shows the detail of step S53 in FIG. The figure for demonstrating the pixel group used as the object of reduction of the stripe noise in the 3rd Embodiment and 4th Embodiment of this invention. The block diagram which shows an example of a structure of the information processing apparatus which concerns on the 3rd Embodiment of this invention. The figure for demonstrating the relationship of the various vectors used for correction | amendment. 6 is a flowchart showing an overall operation of a streak noise reduction unit in the information processing apparatus according to the embodiment. The flowchart which shows the detail of step S83 in FIG. The block diagram which shows an example of a structure of the information processing apparatus which concerns on the 4th Embodiment of this invention. The figure for demonstrating the relationship of the various values used for correction | amendment. 6 is a flowchart showing an overall operation of a streak noise reduction unit in the information processing apparatus according to the embodiment. The flowchart which shows the detail of step S103 in FIG.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 11 ... Streaky noise feature vector calculation part, 12, 22 ... Streaky noise evaluation value addition part, 13, 23 ... Streaky noise determination part, 14, 24 ... Streaky noise reduction part, 21 ... Streaky noise characteristic value calculation , 111, 211 ... streak noise evaluation value calculation unit, 112 ... norm calculation unit, 113, 212 ... streak noise evaluation value adjustment unit.

Claims (8)

In an information processing apparatus that obtains an image by scanning the surface of a paper sheet with a scanner,
Computation means for calculating an evaluation value of noise for each pixel in an image for each color component, wherein the value of the pixel to be evaluated is calculated between two pixels located on both sides of the pixel to be evaluated in a predetermined direction When the value is larger than any of the values, the difference between the value of the pixel to be evaluated and the larger value of the two pixels is set as the evaluation value of the pixel to be evaluated, while the two pixels Is smaller than any of the two values, the difference between the smaller value of the two pixels and the value of the pixel to be evaluated is used as the evaluation value of the pixel to be evaluated. Calculating means for performing, for each color component, processing for setting 0 as the evaluation value of the pixel to be evaluated;
Addition means for performing, for each color component, processing for adding the evaluation value of each pixel obtained by the calculation means in a direction perpendicular to the predetermined direction, and calculating a vector norm having the addition result of each color component as an element. When,
An information processing apparatus comprising: a determination unit that determines that a streak noise is generated in a pixel group that has generated a norm equal to or greater than a certain threshold among the individual norms obtained by the addition unit. . The computing means is
A norm calculating means for calculating a norm of a vector whose elements are evaluation values of each color component in each pixel in the image;
Evaluation value adjusting means for performing, for each color component, a process for reducing the evaluation value of the corresponding pixel when there is a norm exceeding a certain threshold among the individual norms obtained by the norm calculation means. The information processing apparatus according to claim 1.   Noise reduction means for correcting the value of each color component in each pixel included in the pixel group in which the streak-like noise is generated, wherein a vector having the value of each color component in the pixel to be corrected as an element And further comprising noise reduction means for correcting a fixed correction amount as a limit, based on a vector having each color component value in a pixel located outside the pixel group in which the noise is generated as an element. Item 3. The information processing apparatus according to item 1 or 2. In an information processing apparatus that obtains an image by scanning the surface of a paper sheet with a scanner,
An arithmetic means for calculating an evaluation value of noise for each pixel in an image, wherein a value of a pixel to be evaluated is greater than any value of two pixels located on both sides of the pixel to be evaluated in a predetermined direction Is larger, the difference between the value of the pixel to be evaluated and the larger one of the values of the two pixels is set as the evaluation value of the pixel to be evaluated, while any value of the two pixels Is smaller than the value of the smaller value of the two pixels and the value of the pixel to be evaluated as the evaluation value of the pixel to be evaluated; A calculation means for setting an evaluation value of a pixel to be evaluated;
Adding means for adding evaluation values of individual pixels obtained by the calculating means in a direction perpendicular to the predetermined direction;
Determining means for determining that streak noise has occurred in a pixel group that has generated an addition result equal to or greater than a certain threshold value among the individual addition results obtained by the addition means. Processing equipment. The computing means is
5. The information according to claim 4, further comprising: an evaluation value adjusting unit that reduces an evaluation value of a corresponding pixel when there is an evaluation value exceeding a certain threshold among evaluation values of individual pixels in the image. Processing equipment.   Noise reduction means for correcting the value of each pixel included in the pixel group in which the streak noise is generated, wherein the value of the pixel to be corrected is set outside the pixel group in which the streak noise is generated. 6. The information processing apparatus according to claim 4, further comprising noise reduction means for correcting a fixed correction amount as a limit based on a value of a pixel located at. In an information processing method for detecting streak noise included in an image obtained by scanning the surface of a paper sheet with a scanner by an information processing device,
A calculation step of calculating an evaluation value of noise for each pixel in an image for each color component, wherein the value of the pixel to be evaluated is calculated between two pixels located on both sides of the pixel to be evaluated in a predetermined direction When the value is larger than any of the values, the difference between the value of the pixel to be evaluated and the larger value of the two pixels is set as the evaluation value of the pixel to be evaluated, while the two pixels Is smaller than any of the two values, the difference between the smaller value of the two pixels and the value of the pixel to be evaluated is used as the evaluation value of the pixel to be evaluated. Is a calculation step for performing, for each color component, a process of setting 0 as the evaluation value of the pixel to be evaluated,
An addition step of performing, for each color component, a process of adding the evaluation value of each pixel obtained in the calculation step in a direction perpendicular to the predetermined direction, and calculating a norm of a vector having the addition result of each color component as an element. When,
A determination step of determining that streak noise is generated in a pixel group that has generated a norm equal to or greater than a certain threshold among the individual norms obtained by the addition step. . In an information processing method for detecting streak noise included in an image obtained by scanning the surface of a paper sheet with a scanner by an information processing device,
A calculation step of calculating an evaluation value of noise for each pixel in the image, wherein the value of the pixel to be evaluated is greater than any of two pixels located on both sides of the pixel to be evaluated in a predetermined direction Is larger, the difference between the value of the pixel to be evaluated and the larger one of the values of the two pixels is set as the evaluation value of the pixel to be evaluated, while any value of the two pixels Is smaller than the value of the smaller value of the two pixels and the value of the pixel to be evaluated as the evaluation value of the pixel to be evaluated; A calculation step as an evaluation value of a pixel to be evaluated;
An addition step of adding evaluation values of individual pixels obtained by the calculation step in a direction perpendicular to the predetermined direction;
A determination step of determining that streak noise has occurred in a pixel group that has generated an addition result equal to or greater than a certain threshold value among the individual addition results obtained by the addition step. Processing method.

Images