JP2002057873A - Apparatus and method of removing preprinting - Google Patents

Abstract

PROBLEM TO BE SOLVED: To eliminate such a problem that when an image is expanded or contracted in the transfer direction due to a variation in transfer speed at the time of fetching the image by an image input unit, or when there is a deviation in position of preprinting due to the expansion or contraction of a formed document sheet or the like, a reference image and a preprinting of an inputted image cannot be aligned correctly, thus decreasing the removing accuracy of preprinting. SOLUTION: A ruled line deviation detector 5 detects a quantity of deviation between ruled lines which are judged identical by a ruled line comparison means 4. A positional deviation calculating means 7 calculates a quantity of positional deviation at any point on a formed document, based on the deviation between the ruled lines detected by the ruled line deviation detecting means 5. Based on the quantity of positional deviation calculated by the positional deviation calculating means 7, a coordinate transformation means 8 conducts at least one of operations among parallel movement, rotation, and expansion and contraction of an inputted image, An image comparing means 9 compares the inputted image subjected to coordinate transformation by means of the coordinate transformation means 8 with the reference image, and then remove a preprinting part from the inputted image.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus and method for removing preprints, and more particularly, to an image of an unfilled form including preprinted ruled lines and characters, which is stored as a reference image, and to which a preprinted form has been filled. The present invention relates to a method and an apparatus for generating an image from which a preprint is removed from an input image, using the image of the form as an input image.

[0002]

2. Description of the Related Art When a form printed in non-dropout color is to be read by an OCR, or when it is desired to increase the compression ratio at the time of image compression as much as possible for electronic filing of a large number of forms, the form image can be read from the form image. In some cases, an image in which the preprinted portion is removed and only the entered characters are left is required.

In order to remove preprinting of a form image or to check whether the preprinting of an input image and a reference image match, a process of collating the preprinting of an input image and a reference image is performed. At this time, an operation of performing parallel movement or rotation on the input image is performed in order to absorb the difference between the form position and the inclination of the two. JP-A-11-120288, JP-A-8-77333, JP-A-8-255237, and the like disclose such conventional techniques.

In the image shift detecting method described in Japanese Patent Application Laid-Open No. H11-120288, in order to detect a shift between an input image and a reference image, a ruled line in the reference image is gradually translated by using a black run histogram. In this case, the degree of matching with the ruled line of the input image is checked.

In the electronic filing apparatus described in Japanese Patent Application Laid-Open No. Hei 8-77333, as a method of comparing an input image with a reference image, every time the input image is translated and rotated little by little, the pre-print coincidence with the reference image is determined. The degree is checked, and the method of finding the position with the highest degree of matching is adopted.

[0006] In the data storage device described in Japanese Patent Application Laid-Open No. 8-255237, a form on an input image and a form on a reference image are erected using the inclination of ruled lines, and the erected images are shifted vertically and horizontally. However, a method is employed in which the position of the two is adjusted by finding the position where the degree of overlap is greatest.

[0007]

However, in the above-described conventional method for detecting a displacement of an image or the like, when aligning the pre-printing of the input image and the pre-printing of the reference image, only the parallel movement or the parallel movement and rotation of the entire image is performed. Because, when the image is expanded or contracted in the transport direction due to uneven transport speed during image acquisition by the image input device, the pre-print of the input image and the reference image cannot be accurately aligned, There is a problem that the removal accuracy of the pre-printing is reduced.

In the above-described conventional method for detecting a displacement between images, the means for aligning the two images depends only on the parallel movement and rotation of the entire image. When the pre-print position of the reference image is shifted, there is a problem that the position shift cannot be corrected.

In view of the above, the present invention has been made in view of the problems in the above-described conventional image shift detecting method and the like. It is an object of the present invention to provide a pre-print removal apparatus and method that are robust against a pre-print position shift between an input image and a reference image due to expansion and contraction of the image.

[0010]

According to a first aspect of the present invention, there is provided a pre-printing and removing apparatus, wherein the same ruled line among the ruled lines of each form of an input image and a reference image is formed. A ruled line matching means for detecting, a ruled line displacement amount detecting means for detecting a displacement amount between ruled lines determined to be the same ruled line by the ruled line matching means, and a ruled line displacement amount detected by the ruled line displacement amount detecting means. A position shift amount calculating means for calculating a position shift amount at an arbitrary point on a form; and a translation, rotation, expansion and contraction of the input image based on the position shift amount calculated by the position shift amount calculation means. Coordinate conversion means for performing at least one, and image collation means for collating the input image subjected to coordinate transformation by the coordinate transformation means and the reference image and removing a preprinted portion from the input image, I do.

According to a second aspect of the present invention, there is provided a pre-print removal method, comprising the steps of detecting the same ruled line among the ruled lines of each form of the input image and the reference image; Detecting the amount of misalignment, calculating the amount of misalignment at an arbitrary point on the form based on the detected amount of misalignment of the ruled line, and calculating the input image based on the calculated amount of misalignment. And performing a step of performing at least one of translation, rotation, and expansion and contraction, and a step of collating the coordinate-converted input image with the reference image and removing a preprinted portion from the input image. .

According to the first or second aspect of the present invention, the position of the image in the vertical direction of the form is corrected using the horizontal ruled line preprinted on each form of the input image and the reference image, and Since the pre-printing of the input image and the reference image are collated, it is possible to realize a pre-print removal process that is robust against expansion and contraction of the image in the transport direction due to unevenness in the form transport speed of the image input device.

According to the first or second aspect of the present invention, the position of the image in the vertical direction and the horizontal direction of the form is determined by using the horizontal ruled line and the vertical ruled line preprinted on each form of the input image and the reference image. Since the pre-printing of the input image and the reference image is collated after correcting the misalignment, a pre-print removal process that is robust against misalignment of pre-printing in the vertical and horizontal directions due to expansion and contraction of a form can be realized.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, an embodiment of a preprint removal apparatus and method according to the present invention will be described with reference to the drawings.

FIG. 1 shows an embodiment of a pre-print removing apparatus according to the present invention. The pre-print removing apparatus includes a form position detecting means 1, a ruled line coordinate detecting means 2, a ruled line detecting means 3, Ruled line collating means 4, ruled line displacement amount detecting means 5,
A pre-print removing unit 6, a ruled line displacement amount storage memory 10,
And a reference image storage memory 11. The pre-print removal unit 6 includes a displacement amount calculation unit 7, a coordinate conversion unit 8, and an image comparison unit 9.

Next, the function of each means will be described.

The form position detecting means 1 detects the coordinates of the upper left corner of the form on the form image and the inclination of the form. The ruled line coordinate detecting means 2 detects the coordinates of ruled lines preprinted on a form on a form image. The ruled line detection means 3 detects a ruled line preprinted on a form on a form image, and obtains the inclination of the ruled line.

The ruled line matching means 4 associates the same ruled line among the ruled lines detected for each of the input image and the reference image. The ruled line shift amount detecting means 5 detects a positional shift amount of both ruled lines in a direction perpendicular to the ruled line with respect to the ruled line associated by the ruled line matching means 4. The displacement amount calculating means 7 calculates the displacement amount of the input image and the reference image at an arbitrary point on the form based on the displacement amount of the ruled line detected by the ruled line displacement amount detecting means 5.

The coordinate conversion means 8 converts the coordinates of the input image so that the pre-print positions of the input image and the reference image match. The image matching unit 9 compares pixel values at the same position in the input image after coordinate conversion with the reference image.

The pre-print removing means 6 includes the position shift amount calculated by the position shift amount calculating means 7 and the form position detecting means 1.
Based on the coordinates of the upper left corner of each form of the input image and the reference image detected by the above and the inclination of the ruled line detected by the ruled line detecting means 3, the preprinting of the forms on both images is collated, Outputs the image with print removed.

The ruled line shift amount storage memory 10 stores the ruled line shift amount detected by the ruled line shift amount detecting means 5. The reference image storage memory 11 stores an unfilled form image as a reference image.

Next, the overall operation of the present embodiment will be described in detail mainly with reference to the flowchart of FIG.

The form position detecting means 1 detects a straight line L1 on the upper side and a straight line L2 on the left side of the form on the reference image, and detects an intersection thereof as coordinates (x_ref, y_ref) of an upper left corner of the form. Further, the inclination θsp_ref of the upper side L1 of the form is detected (Step S1 in FIG. 2).

Next, as shown in FIG. 4, the ruled line coordinate detecting means 2 uses the coordinates (x_ref, y_ref) of the upper left corner of the form of the image detected by the form position detecting means 1 as the origin, and the inclination of the upper side of the form. x-axis in θsp_ref direction, y in direction perpendicular to it
In the coordinate system defined by taking the axes, black pixels on the form are projected in each of the x-axis direction and the y-axis direction to generate an x-axis projection histogram and a y-axis projection histogram (step in FIG. 2). S2).

Next, as shown in FIG. 5, among the x-coordinates having the maximum value on the x-axis projection histogram, the coordinates (peak position) having a frequency higher than a predetermined threshold value PEAK_TH.
Is detected as the x coordinate at which the vertical ruled line exists (step S3 in FIG. 2). Similarly, as shown in FIG. 6, among the y-coordinates having the maximum values on the y-axis projection histogram, the coordinates (peak positions) having a frequency higher than a predetermined threshold value PEAK_TH are defined as horizontal ruled lines. Detected as y coordinate.

Next, the ruled line detecting means 3 follows the x-coordinate where the vertical ruled line detected by the ruled line coordinate detecting means 2 exists in the y-axis direction, and successively black pixels extending in the y-axis direction. At the same time as detecting a ruled line, the inclination of the vertical ruled line is obtained. Next, it traces in the x-axis direction on the y coordinate where the horizontal ruled line detected by the ruled line coordinate detecting means exists,
Consecutive black pixels extending in the x-axis direction are detected as horizontal ruled lines, and the inclination of the horizontal ruled lines is obtained. Next, the ruled line detecting means 3 averages the inclination of each of the detected ruled lines and sets the average as the preprint inclination θfm_ref (step S in FIG. 2).
4).

Next, the ruled line coordinate detecting means 2 is applied again,
Using the coordinates (x_ref, y_ref) of the upper left corner of the form of the image detected by the form position detecting means 1 as the origin, x in the direction θfm_ref of the pre-print inclination calculated by the ruled line detecting means 3
In a coordinate system defined by taking the y-axis in a direction perpendicular to the axis (hereinafter, this coordinate system is referred to as a form coordinate system), black pixels on the form are defined for each of the x-axis direction and the y-axis direction. To generate an x-axis projection histogram and a y-axis projection histogram (step S5 in FIG. 2). next,
Among the x-coordinates that take the maximum value on the x-axis projection histogram,
A coordinate having a frequency higher than a predetermined threshold value PEAK_TH is detected as an x coordinate having a vertical ruled line. Similarly,
Of the y-coordinates that take the maximum value on the y-axis projection histogram,
A coordinate having a frequency higher than a predetermined threshold value PEAK_TH is detected as a y-coordinate having a horizontal ruled line (step S6 in FIG. 2).

Next, the form position of the input image, the position of the ruled line, and the like are detected by the same procedure as that performed for the reference image. 2 and the ruled line detection means 3 are applied to the input image (steps S7 to S12 in FIG. 2). As a result, the coordinates (x_in, y_in) of the upper left corner of the form of the input image and the inclination θfm_in
Get. In addition, the coordinates (x_in, y_i
n) and the x coordinate where the vertical ruled line exists and the y coordinate where the horizontal ruled line exist on the input image in the form coordinate system defined by the inclination θfm_in of the pre-printing are detected.

Next, the ruled line collating means 4 determines the absolute coordinate difference between the x-coordinates of the reference image and the input image detected at the second application of the ruled line coordinate detecting means 2 and the vertical ruled lines. Threshold PEAK_LAG_LIM with pre-set value
Those smaller than IT are associated with each other, and they are paired with the x coordinate [pkx_in (i), pkx_ref (i)], i = 1,2,
, M (step S13 in FIG. 2). Here, m is the number of pairs successfully associated. Also, pkx_in (i)
<pkx_in (i + 1), pkx_ref (i) <pkx_ref (i + 1).

Next, the ruled line collating means 4 determines the absolute value of the coordinate difference among the y-coordinates where the horizontal ruled lines exist for each of the reference image and the input image detected by the ruled line coordinate detecting means 2 at the time of the second application. Threshold PEAK_LAG_LIM with pre-set value
Those smaller than IT are associated with each other, and they are paired with the y coordinate representing the same horizontal ruled line [pky_in (j), pky_ref (j)], j = 1,2,
…, N. Here, n is the number of pairs successfully associated.

Next, as shown in FIG. 7, the ruled line shift amount detecting means 5 calculates an equation (Px_in (i), pkx_ref (i)) for each of the x coordinate pairs [pkx_in (i), pkx_ref (i)] generated by the ruled line matching means 4. According to 1), a relative ruled line shift amount lag_x (i) on the x coordinate where the vertical ruled line corresponding to the i-th pair exists is obtained (step S14 in FIG. 2).

Lag_x (i) = pkx_in (i) -pkx_ref (i) (Equation 1) Further, pkx_ref (i) and lag_x (i) are stored in the ruled line deviation amount storage memory in association with each other. 8 is stored in the ruled line deviation amount storage memory.
It is assumed that pkx_ref (i) <pkx_ref (i + 1) is stored while satisfying.

Next, the ruled line shift amount detecting means 5 calculates the relative ruled line shift on the y coordinate where the horizontal ruled line corresponding to the j-th pair exists, by using equation (2) for each pair of the y coordinates. Find the quantity lag_y (j).

Lag_y (j) = pky_in (j) -pky_ref (j) (Equation 2) Also, pky_ref (j) and lag_y (j) are stored in the ruled line deviation amount storage memory in association with each other. In addition, the ruled line deviation amount storage memory contains
It shall be stored while satisfying pky_ref (j) <pky_ref (j + 1).

Next, the pre-print removing means 6 applies the displacement calculating means 7, the coordinate converting means 8 and the image comparing means 9 to each pixel position (x, y) on the form of the reference image. As a result, the input image is compared with the reference image, an image in which the pre-print is removed from the input image is generated, and the image is output as a result image.

First, the pre-print removing means 6 selects one pixel on the form in the reference image as a point of interest (x, y) (step S20 in FIG. 3).

The position shift amount calculating means 7 calculates the ruled line shift amount lag_x (i) in the x-axis direction detected by the ruled line shift amount detecting means 5.
And the coordinate (Xs, xs) of the point of interest (x, y) on the reference image in the form coordinate system based on the x coordinate pkx_ref (i) of the vertical ruled line position on the reference image stored in association with the , Ys), the displacement amount dx (Xs) of the form coordinate system in the x-axis direction is obtained (FIG. 3).
Step S21). Similarly, the ruled line displacement amount lag_y (j) in the y direction detected by the ruled line displacement amount detection means 5 and the horizontal ruled line position on the reference image stored in association with it.
Based on the y-coordinate pky_ref (j), a displacement amount dy (Ys) in the y-axis direction of the form coordinate system at the coordinates (Xs, Ys) on the form in the form coordinate system is obtained (step S22 in FIG. 3). ).

Next, the coordinates (X in the form coordinate system on the reference image)
A method for obtaining the displacement amounts dy (Xs) and dy (Ys) in (s, Ys) will be described.

The displacement dx (Xs) in the x-axis direction of the form coordinate system at the x-coordinate Xs can be obtained by one of the following three methods depending on the position of the x-coordinate Xs.

(A) When x = Xs is located in an area sandwiched between two adjacent vertical ruled lines, the amount of displacement dx (Xs) in the x-axis direction is determined by the x-axis offset of x = Xs. Interval [pkx_ref (i), pkx_ref
(i + 1)], linear interpolation is performed between the ruled line shift amounts [lag_x (i), lag_x (i + 1)] in the x-axis direction at the vertical ruled line position using equation (3).

Dx (Xs) = ((pkx_ref (i + 1) -Xs) * lag_x (i) + (Xs-pkx_ref (i)) * lag_x (i + 1)) / (pkx_ref (i + 1) -pkx_ref (i)) (Equation 3) (b) If x = Xs is smaller than the x coordinate pkx_ref (1) of the leftmost vertical ruled line, the displacement dx (Xs) at x = Xs is The ruled line shift amount lag_x (1) of the leftmost vertical ruled line is set.

(C) x = Xs is the x coordinate pkx_ of the rightmost vertical ruled line
If it is larger than lag (m), the displacement amount at x = Xs
dx (Xs) is the ruled line shift amount lag_x (m) of the rightmost vertical ruled line.

Similarly, y in the form coordinate system at y coordinate Ys
The axial displacement dy (Ys) is determined by the position of the y coordinate Ys.
It is obtained by one of the following three methods.

(A) When y = Ys is located in an area sandwiched between two adjacent horizontal ruled lines, the displacement amount dy (Ys) in the y-axis direction is calculated based on the y-axis sandwiching y = Ys. Interval [pky_ref (j), pky_ref
(j + 1)], the amount of displacement of the ruled line in the y-axis direction [lag_y (j), lag_y (j + 1)] at the horizontal ruled line position is obtained by linear interpolation using equation (4).

Dy (Ys) = ((pky_ref (j + 1) -Ys) * lag_y (j) + (Ys-pky_ref (j)) * lag_y (j + 1)) / (pky_ref (j + 1) -pky_ref (j)) (Equation 4) (b) If y = Ys is smaller than the y coordinate pky_ref (1) of the top horizontal ruled line, the displacement dy (Ys) at y = Ys is The ruled line shift amount lag_y (1) on the uppermost horizontal ruled line is set.

(C) y = Ys is the y coordinate pky_ of the lowermost horizontal ruled line
If it is larger than lag (n), the amount of displacement at y = Ys
dy (Ys) is a ruled line shift amount lag_y (n) on the lowest horizontal ruled line.

Next, the coordinate conversion means 8 uses the coordinate conversion formula shown in FIG. 9 to perform coordinate conversion on the input image so that the pre-print position of the input image matches the pre-print position of the reference image. Find pixel values on the image (step S2 in FIG. 3)
3). The coordinate conversion formula shown in FIG. 9 means not only moving and rotating the input image but also expanding and contracting the input image so as to correct the displacement of the ruled line.

Next, the image collating means 9 illuminates the pixels at the same pixel position with the converted image generated by the image converting means 8 and the reference image, and removes the preprinted portion from the converted image. An image leaving only characters is output as a pre-print removal image. That is, the pixel of interest (x, y) on the reference image is a white pixel, and the interest point (x, y) on the converted image is
If the pixel of (y) is a black pixel, the pixel of the point of interest (x, y) on the converted image is determined to be an input character pixel, and the pixel of the point of interest (x, y) on the result image is determined. Let the value be a black pixel. In other cases, the pixel value of the target point (x, y) on the result image is set as a white pixel (step S24 in FIG. 3).

Next, it is determined whether or not all pixels on the form in the reference image have been selected (step S25 in FIG. 3). If there is an unselected pixel, the process returns to step S20, and the unselected pixel is replaced with the next pixel. One is selected as a point of interest. If all pixels have been selected, the result image is output as a pre-print removal image.

Next, the effect of this embodiment will be described.

In this embodiment, the displacement amount calculating means 7
Is calculated based on the ruled line, and the coordinate conversion means 8 performs not only parallel movement and rotation but also expansion and contraction of the input image such that the ruled line matches. ing. Therefore, it is possible to correct not only the image displacement in the transport direction due to the unevenness of the form transport speed at the time of image acquisition, but also the displacement of the ruled lines in the vertical and horizontal directions of the form due to the expansion and contraction of the paper.

Next, the operation of this embodiment will be described using a specific example.

A completed form 20 as shown in FIG.
1 is a binary image captured by an image input device such as a scanner 2
A case where pre-print removal is performed for 00 will be described. At this time, as shown in FIG. 11, a binary image 300 obtained by the image input device with the blank form 301 preprinted with the same format information is stored in the reference image storage memory 1.
1 is stored.

The form position detecting means 1 detects the coordinates (x_ref, y_ref) of the upper left corner of the blank form 301 and the inclination θsp_ref of the top side of the form in the reference image 300 stored in the reference image storage memory 11 ( Step S1 in FIG. 2).

The ruled line coordinate detecting means 2 generates an x-axis projection histogram (FIG. 5) and a y-axis projection histogram (FIG. 6) of the reference image 300 (step S2 in FIG. 2). Next, the ruled line coordinate detecting means 2 determines two peak positions px1 and px2 on the x-axis projection histogram shown in FIG.
Detect as coordinates. Similarly, six peak positions py1, py2, py3, py4, and py on the y-axis projection histogram shown in FIG.
5. py6 is detected as the y coordinate where the horizontal ruled line exists (step S3 in FIG. 2).

Next, the ruled line detecting means 3 detects a vertical ruled line based on the x-coordinate where the vertical ruled line detected by the ruled line coordinate detecting means 2 exists, and at the same time, obtains the inclination of the vertical ruled line (see FIG. 2 step S4). Next, a horizontal ruled line is detected based on the y-coordinate where the horizontal ruled line detected by the ruled line coordinate detecting means 2 exists, and at the same time, the inclination of the horizontal ruled line is obtained. Next, the ruled line detection means 3 calculates the average value of the slopes of the detected ruled lines, and calculates the average value of the preprinted slope θfm_r.
ef.

Next, the ruled line coordinate detecting means 2 calculates the coordinates of the upper left corner of the form of the image detected by the form position detecting means 1.
In the form coordinate system defined by (x_ref, y_ref) and the preprint inclination θfm_ref obtained by the ruled line detection means 3, black pixels on the form are projected in each of the x-axis direction and the y-axis direction. Then, an x-axis projection histogram and a y-axis projection histogram (not shown) are generated (step S5 in FIG. 2). Next, the peak position on the x-axis projection histogram is detected as the x coordinate where the vertical ruled line exists. Similarly, the peak position on the y-axis projection histogram is detected as the y coordinate where the horizontal ruled line exists (step S6 in FIG. 2).

Next, the position of the form and the position of the ruled line of the input image 200 are detected (steps S7 to S12 in FIG. 2). As a result, the coordinates (x_in, y_in) of the upper left corner of the form of the input image and
Obtain the preprint inclination θfm_in. Also, the coordinates (x_in, y_in) of the upper left corner of the form of the input image and the inclination θfm_in
In the form coordinate system defined by the above, the x coordinate where the vertical ruled line exists and the y coordinate where the horizontal ruled line exists on the input image are detected.

Next, the ruled line collating means 4 compares the reference image 300 detected by the ruled line coordinate detecting means 2 with the input image 2.
Among the x-coordinates where the vertical ruled lines exist for each of 00, a pair [pkx_in (i), pkx_ref (i)] of the x-coordinate representing the same vertical ruled line is found.

Next, the ruled line collating means 4 compares the reference image 300 detected by the ruled line coordinate detecting means 2 with the input image 2.
Among the y-coordinates where horizontal ruled lines exist for each of 00, a pair of y-coordinates [pky_in (j), pky_ref (j)] representing the same horizontal ruled line is found (step S13 in FIG. 2).

Next, the ruled line shift amount detecting means 5 calculates the relative ruled line shift amount lag_x (i) on the x coordinate where the vertical ruled line exists and the relative ruled line shift amount on the y coordinate where the horizontal ruled line exists. lag_y
(j) is obtained (step S14 in FIG. 2).

Next, the pre-print removing unit 6 applies the displacement calculating unit 7, the coordinate converting unit 8, and the image comparing unit 9 to each pixel position (x, y) on the form of the reference image. Thereby, the input image is compared with the reference image, a result image 400 obtained by removing the pre-printing from the input image is generated, and the result image is output as the result image (step S1 in FIG. 2).
5).

[0063]

As described above, according to the present invention,
In addition to the parallel movement and rotation of the input image, expansion and contraction can be performed, so even if there is expansion or contraction of the image in the conveyance direction due to uneven conveyance speed at the time of image acquisition, expansion or contraction of the form itself, etc. It is possible to provide a pre-print removal apparatus and method capable of detecting a pre-print displacement between an image and a reference image and performing pre-print removal.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of an embodiment of a preprint removal apparatus according to the present invention.

FIG. 2 is a flowchart showing the operation of an embodiment of the preprint removal apparatus according to the present invention.

FIG. 3 is a flowchart showing an operation of a pre-print removing unit of the pre-print removing apparatus according to the present invention.

FIG. 4 is a diagram showing a method for generating a pixel projection histogram in the preprint removal apparatus according to the present invention.

FIG. 5 is a diagram illustrating a method of detecting a peak of an x-axis pixel projection histogram in the pre-print removal apparatus according to the present invention.

FIG. 6 is a diagram showing a peak detection method of a y-axis pixel projection histogram in the pre-print removal apparatus according to the present invention.

FIG. 7 is a diagram illustrating a method for detecting a ruled line shift amount in the pre-print removal apparatus according to the present invention.

FIG. 8 is a diagram showing a storage method of a ruled line shift amount storage memory in the pre-print removal apparatus according to the present invention.

FIG. 9 is an equation showing coordinate transformation in the preprint removal apparatus according to the present invention.

FIG. 10 is a diagram showing an example of an input image in one embodiment of the preprint removal apparatus according to the present invention.

FIG. 11 is a diagram showing an example of a reference image in an embodiment of the preprint removing apparatus according to the present invention.

FIG. 12 is a view showing an example of a result image in one embodiment of the pre-print removal apparatus according to the present invention.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 form position detecting means 2 ruled line coordinate detecting means 3 ruled line detecting means 4 ruled line collating means 5 ruled line displacement amount detecting means 6 preprint removing means 7 positional displacement amount calculating means 8 coordinate converting means 9 image collating means 10 ruled line displacement amount storage memory 11 Reference image storage memory 100 Reference image 101 Blank form 200 Input image 201 Filled form 300 Reference image 301 Blank form 400 Result image

 ────────────────────────────────────────────────── ─── Continued on the front page F term (reference) 2C087 AA12 BD01 BD24 CB03 CB04 5B057 AA11 BA02 CA12 CA16 CB12 CB16 CC01 CD02 CD03 CD05 CH11 DA07 DA08 DB02 DC03 DC19 DC32 5C076 AA03 AA21 AA22 AA24 BA01 BA06 CB05 FA03FA03 GA03

Claims (2)

[Claims] A rule matching means for detecting the same ruled line among the ruled lines of each form of an input image and a reference image; and a ruled line for detecting an amount of displacement between ruled lines determined to be the same ruled line by the ruled line matching means. Displacement amount detection means; displacement amount calculation means for calculating a displacement amount at an arbitrary point on a form based on the displacement amount of the ruled line detected by the displacement amount detection means; Based on the displacement calculated by (1), a coordinate conversion unit that performs at least one of translation, rotation, and expansion / contraction of the input image; and the input image and the reference image that have been coordinate-converted by the coordinate conversion unit. An image collating means for collating and removing a preprint portion from an input image. 2. A step of detecting the same ruled line among the ruled lines of each form of the input image and the reference image; a step of detecting a shift amount between the ruled lines determined to be the same ruled line; Calculating a displacement amount at an arbitrary point on the form based on the displacement amount; and performing at least one of translation, rotation, and expansion / contraction of the input image based on the calculated displacement amount. And a step of collating the coordinate-converted input image with the reference image to remove a preprint portion from the input image.