JP3846752B2 - Ruled line removal method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
  The present invention relates to image processing, and more particularly to a technique for erasing ruled lines while preserving characters from an image such as a form in which characters and ruled lines are mixed in a preprocessing unit of a character recognition device.
[0002]
[Prior art]
  When an image such as a form including many ruled lines is processed by the character recognition device, the proximity (contact, intersection, overlap) between the ruled lines and the characters becomes a big problem. In order to realize high-precision character recognition, such close characters must be recognized separately from the ruled lines, and one method is to erase only the ruled lines from the form image in the preprocessing stage. is there.
[0003]
  As a means for realizing such ruled line erasing, FormOut! Software with a trade name of (TiS, Israel) is known. This software reads a form (sample) in which the entry field is not filled, and learns the fixed form information (ruled lines, characters, figures). Thereafter, the same type of form in which the entry field is entered is read, the learned fixed form information is removed from the image, and an image of only the variation information such as characters entered in the entry field is output.
[0004]
[Problems to be solved by the invention]
  The above software is premised on the existence of “samples”, learning them in advance, and processing forms of the same type as learned samples. Cannot be used for processing forms. In addition, it seems that it is possible to cope with the contact between the ruled line and the character to some extent, but in the case of the intersection with the ruled line, especially the intersection with the diagonal line or overlapping, it seems that the performance that can restore the character cannot be expected. .
[0005]
  The purpose of the present invention is to save characters from images of any type of form without using a sample,Crossing and touching the character lineThe object is to provide a method for erasing ruled lines with high accuracy.
[0006]
[Means for Solving the Problems]
  The ruled line erasing method according to the present invention includes:In one embodiment,Recognizing the orthogonal part of the ruled line, erasing the recognized orthogonal part, and eliminating the ruled line while preserving the ruled line close to the ruled line from the image such as a form without using a sample, and The method includes the step of erasing a black run having a length not exceeding the thickness of the ruled line among black runs that vertically cross the vicinity of the center of gravity of the ruled line.In particular, the present inventionA black run that crosses the vicinity of the center of gravity of the ruled line vertically so as not to leave an unerased part of the ruled line at the part where the character line intersects the ruled line diagonally, and a threshold value that is larger than the thickness of the ruled line and sufficiently larger than the thickness of the ruled line A black run with a smaller length and whose center of gravity is shifted from the center of gravity of the ruled line by a predetermined value or more,Obtaining a reference value obtained by subtracting the thickness of the ruled line from the length of the black run, storing a length portion corresponding to the reference value from an end in a direction where the center of gravity is shifted, and erasing the remaining portion. It is characterized by. The present invention also provides:In order not to leave the unerased part of the ruled line in the overlapping or contacting part of the ruled line and the character line, it is a black run that vertically crosses the vicinity of the center of gravity of the ruled line and is almost constant larger by a predetermined amount than the thickness of the ruled line A black run having a length is continuous, and a section where a black run having a length more than a predetermined value sufficiently larger than the thickness of the ruled line is detected before or after the black run in the section is detected.Obtaining a reference value obtained by subtracting the thickness of the ruled line from the length of the black run, storing a length portion corresponding to the reference value from an end in a direction where the center of gravity is shifted, and erasing the remaining portion. It is characterized by.
[0007]
DETAILED DESCRIPTION OF THE INVENTION
  Embodiments of the present invention will be described below with reference to the drawings.
[0008]
  The present invention can be implemented by software in a computer system 10 having a general configuration as shown in FIG. 1, for example. In FIG. 1, 12 is a bus for interconnecting each component, 14 is a central processing unit (CPU), and 16 is a system memory. An external device 20 such as a printer is connected to the bus 12 via a parallel port 22. A display 24 is connected to the bus 12 via a display adapter 26, and a pointing device mouse 38 is connected to the bus 12 through the serial port 28. Reference numeral 30 is a keyboard, 32 is a fixed disk device, and 33 is a floppy disk drive. An image scanner 34 is connected to the bus 12 by an external interface 36.
[0009]
  The image to be processed is read and input from an original by the image scanner 34, for example, or is temporarily recorded on the floppy disk 33A and then input via the floppy disk drive 33. The input image data is temporarily stored in the fixed disk drive 32 and is read into a specific area of the system memory 16 at the time of processing execution or directly read into a specific area of the system memory 16. The image data being processed, the final image data, and other data are temporarily stored in a specific area of the system memory 16. The final image data is stored in the fixed disk drive 30 and output to the external device 20 or the floppy disk 33A when necessary. The input image data and the image data being processed can be displayed on the display 24. The user can specify a ruled line or input other information by operating the mouse 38 or the keyboard 30. A program for ruled line erasing processing, which will be described in detail in connection with an embodiment described later, is loaded into the system memory 16 from the fixed disk drive 32 or the floppy disk 33A, and is executed by the central processing unit 14.
[0010]
  Hereinafter, the ruled line erasing process according to the present invention will be described in detail. However, the relation with the configuration of the computer system 10 shown in FIG.
[0011]
  FIG. 2 is a flowchart showing the flow of processing according to the first embodiment of the present invention. In FIG. 2, first, image data such as a document including characters and ruled lines is input (step 200), and then binarization processing of the image data is performed (step 201). Image compression is performed on the binarized image data (step 202). The image data after the image compression is a target of processing described later. If the input image data has already been binarized, the binarization process in step 201 is not necessary. The image compression in step 202 is intended to reduce the amount of memory necessary for storing image data and shorten the processing time, and may be omitted if not necessary.
[0012]
  Next, a ruled line in the image data is recognized (step 203). Automatic ruled line recognition can be achieved by, for example, integrating a long black run in the horizontal and vertical directions and extracting the circumscribed rectangle as a ruled line, or by projecting black pixels in the horizontal and vertical directions, and using a peak area as a ruled line. This can be done by the extraction method. The information of the ruled line to be obtained is, for example, the position, length, and thickness of the ruled line. The ruled lines recognized by such a method are horizontal and vertical ruled lines. If a ruled line recognition method based on vector analysis is adopted, diagonal ruled lines can also be recognized.
[0013]
  The ruled line position information may be given in advance on the system. Specifically, ruled line position information is given using a ruled line template, or image data is displayed on the display 24, and the position of the ruled line is designated on the screen using the mouse 38 or the keyboard 30. A method of giving information may be adopted.
[0014]
  Next, the orthogonal part of the horizontal and vertical ruled lines is recognized (step 204). When the diagonal ruled line is also recognized by the vector analysis method, the orthogonal part between the diagonal ruled lines is also recognized. As shown in FIG. 3A, the coordinate of the upper left vertex of the rectangle 101 of the vertical ruled line is (Vxs, Vys), the coordinate of the lower right vertex is (Vxe, Vye), and the horizontal ruled line orthogonal to this. If the coordinates of the upper left vertex of the rectangle 100 are (Hxs, Hys) and the coordinates of the lower right vertex are (Hxe, Hye), the orthogonal portion 102 has the coordinates of the upper left vertex (Vxs, Hys) and the coordinates of the lower right vertex. Is obtained as a square region of (Vxe, Hye).
[0015]
  The black pixel in the orthogonal part of the ruled line thus obtained is replaced with a white pixel (step 205). Thus, the ruled line and the orthogonal part of the ruled line are erased from the image. The part to be erased is limited to the part perpendicular to the ruled lines, and the part perpendicular to the ruled lines and the characters is not erased.
[0016]
  For the image data in which the orthogonal part of the ruled line is erased in this way, the unerased part of the ruled line is erased in the processing loop of steps 206 to 212. In step 206, one ruled line is selected, and the ruled line is deleted in the processing loop of steps 207-211. This process is repeatedly executed while changing the processing position in units of one pixel from the start point of the ruled line to the end point, and exits the loop when the process is completed up to the end of the ruled line.
[0017]
  The processing contents are as follows. At the processing position, the length of the black run that crosses the vicinity of the center of gravity of the ruled line and is perpendicular to the vector direction of the ruled line is detected (step 208). For example, in the case of the horizontal ruled line rectangle 100 shown in FIG. 3 (b), around the straight line 111 connecting the two points (Hxs, (Hys + Hye) / 2) and (Hxe, (Hys + Hye) / 2) The length of the black run that crosses the image in the vertical direction is detected while sequentially shifting the processing position from the left end to the right end. The length of the detected black run and the thickness of the ruled line are compared (step 209). If the length of the black run is equal to or smaller than the thickness of the ruled line, the black pixel of the black run is replaced with a white pixel (step 210). Here, the thickness of the ruled line is the average thickness of the ruled line designated in advance by the system or designated by the user. Instead of such a static threshold process, a dynamic threshold process that determines to delete the noticed black run when the difference in length between the noticed black run and the immediately preceding black run is small. It is also possible to use.
[0018]
  When the erasing process for one ruled line is completed, the process returns to step 206 to select another ruled line and perform the same ruled line erasing process. If it is determined in step 212 that the last ruled line has been processed, the ruled line erasing process is completed, and the image data that has been subjected to the ruled line process is output (step 213).
[0019]
  According to the first embodiment, a ruled line can be erased while storing a character line that is close (contact, intersecting, overlapping) with the ruled line. However, on the other hand, erasure of the part where the character lines of the ruled lines are close is incomplete. For example, when the character line 121 shown in FIG. 4A is a portion 120 that obliquely intersects the ruled line 122, the length 123 of the ruled line 122 shown as a black region in FIG. Since it is larger than the ruled line thickness, it is not erased and remains as noise. If the area of such an erasure omission part is sufficiently small, it is acceptable, but if the area is not ignorable, character recognition for an image after erasure of ruled lines is adversely affected. According to the second embodiment of the present invention to be described next, it is possible to greatly reduce the erasure omission of the ruled line at the portion where the character lines cross obliquely.
[0020]
  FIG. 5 is a flowchart showing the flow of processing of the second embodiment of the present invention. In FIG. 5, an image data input step 300, ruled line recognition step 301, ruled line orthogonal part recognition step 302, and ruled line orthogonal part erased step 303 are the same as steps 200, 203, 204, and 205 in FIG. It is. If the input image data is not binarized, the binarization process is necessary. Further, image data compression processing may be added in order to reduce the amount of memory used and the processing time.
[0021]
  The unerased part of the ruled line is erased by the processing loop of steps 305 to 318 for the image data after the orthogonal part of the ruled line is erased. First, one unprocessed ruled line is selected (step 305), the initial flag is turned off (step 306), and the unerased portions are sequentially erased while shifting the processing position by one pixel unit in the vector direction of the ruled line (step 306). Steps 307-317).
[0022]
  At the processing position, the length of the black run that crosses the vicinity of the center of gravity of the ruled line and is perpendicular to the vector direction of the ruled line is detected (step 308). This is the same as step 208 in FIG. 2 (see FIG. 3B). The size of the detected black run length and ruled line thickness is determined (step 309). If the length of the black run is equal to or less than the thickness of the ruled line, the black run is regarded as a black run constituting the ruled line, the black pixel is replaced with a white pixel (step 310), and the initial flag is turned off (step 311). ). If the processing position has not reached the end of the ruled line, the process returns from step 317 to step 307 to continue the processing by advancing the processing position by one.
[0023]
  If it is determined in step 309 that the length of the detected black run is larger than the ruled line thickness, the length of the black run is sufficiently larger than the ruled line thickness or character thickness (specified in advance by the system or specified by the user). Whether the threshold value is large or not is determined (step 312). If it is determined that the length of the black run is equal to or greater than the threshold, it is determined that the character line is orthogonal to the ruled line, and the process proceeds to the next processing position without deleting the black run. For example, at the intersection 130 between the character line 132 and the ruled line 131 as shown in FIG. 6A, the length of the black run is much larger than the ruled line thickness and the character thickness, and is not deleted. Therefore, as shown in an enlarged view in FIG. 6B, only the portion excluding the intersection of the ruled line 131 and the character line 132 is erased, and the character line 132 is stored.
[0024]
  If it is determined in step 312 that the length of the black run is smaller than the threshold, a deviation between the center of gravity of the black run (middle point) and the center of gravity of the ruled line (middle point in the thickness direction) is detected (step 313). It is determined whether or not the deviation is larger than a predetermined threshold (step 314). If the deviation is larger than the threshold, the process proceeds to step 315 for black run erasure. If not, the process position is advanced by one and the process is continued. .
[0025]
  Steps 313, 314, and 315 will be specifically described with reference to FIG. In FIG. 7A, 140 is a ruled line, and 141 is a character line intersecting with the ruled line 140. Reference numeral 142 denotes a line connecting the center of gravity of the ruled line in the thickness direction, and reference numeral 143 denotes a line connecting the center of gravity of the black run in a direction perpendicular to the vector direction of the ruled line 140. The black run is erased in step 310 from the beginning of the ruled line 140 to just before the position P1 that intersects the character line. However, since a long black run exceeding the threshold is detected at the P1 position, the process proceeds to step 313, and a shift in the center of gravity (shift between the line 142 and the line 143) is detected. Then, as in the example illustrated in FIG. 7A, when the deviation of the center of gravity is large, the process proceeds to step 315.
[0026]
  In step 315, since the initial flag is initially off, a reference value X obtained by subtracting the thickness b of the ruled line from the length h of the black run at the processing position is obtained (step 315A). FIG. 7B illustrates how to obtain the reference value X. Then, the initial flag is turned on (step 315B), the length corresponding to the reference value X is stored from the end of the black run at the processing position in the direction where the center of gravity is shifted, and the remaining portion is deleted (step 315C). ). In the example of FIG. 7A, the length corresponding to the reference value X is stored from the lower end of the black run at the P1 position, and the remaining upper black pixels are replaced with white pixels.
[0027]
  Similar black run partial erasing is performed while the processing position is sequentially advanced to the right. The black run is partially erased up to the P2 position where the deviation of the center of gravity is smaller than the threshold. From the P2 position, the black center partial erasure (step 315) is not performed because the deviation of the center of gravity is smaller than the threshold value. As a result, the triangular portion 145 above the character line 141 of the ruled line 140 is deleted as shown in FIG.
[0028]
  The process position is advanced without proceeding to step 315 until just before the P3 position where the deviation of the center of gravity exceeds the threshold value, but when the P3 position is reached, the process proceeds to step 315. At this time, since the initial flag is in the on state, the reference value X is not recalculated, the reference value x obtained at the P1 position is used, and the reference value X of the black run is shifted from the end on the shifted side. The part exceeding the length is erased (step 315C). In the example of FIG. 7A, the lower end portion of the black run is erased. This partial erasure is sequentially performed up to the P4 position. As a result, the triangular portion 146 below the character line 141 of the ruled line 140 is erased as shown in FIG. In this way, the ruled line portion that obliquely intersects the character line and the like that could not be erased in the first embodiment is erased. Since the length of the black run is equal to or smaller than the ruled line thickness from the processing position next to the P4 position, it is erased in step 310.
[0029]
  FIG. 8 shows an example in which the deviation of the center of gravity is reduced. 150 is a ruled line, 151 is a line connecting the centers of gravity, 152 and 153 are lines such as character lines intersecting the ruled lines 150, and 154 is a line connecting the centers of black runs in the ruled line thickness direction. Since the intersecting portion of the line as in this example is excluded from the target of partial erasure in step 315, the ruled line portions (shaded areas 155 and 156 in FIG. 8) that are in contact with the intersecting line segment remain as noise. .
[0030]
  In this way, when the erasing of one ruled line is completed by the processing loop of steps 307 to 317, the process returns to step 305 to select one next ruled line and execute the erasing process. When the erasing process is completed up to the last ruled line, the process loop is exited at step 318, and the image data after the ruled line process is output (step 319). When this is completed, the process is completed.
[0031]
  9 and 10 are flowcharts showing the flow of processing according to the third embodiment of the present invention. In this embodiment, it is possible to reduce the overlap of the ruled lines with the character lines or the erasure omission of the contact portions.
[0032]
  9, an image data input step 400, ruled line recognition step 401, ruled line orthogonal part recognition step 402, and ruled line orthogonal part erased step 403 are the same as steps 200, 203, 204, and 205 in FIG. is there. If the input image data is not binarized, the binarization process is necessary. Further, image data compression processing may be added in order to reduce the amount of memory used and the processing time.
[0033]
  A black run erase region that is considered to be an unerased part of the ruled line is created by the processing loop of steps 405 to 413 for the image data after the orthogonal part of the ruled line is erased. Actual erasure is performed in the processing loop of steps 415 to 424 shown in FIG.
[0034]
  First, one unprocessed ruled line is selected (step 405), and a black run erase region that is considered to be an unerased part is created while shifting the processing position by one pixel unit in the vector direction of the ruled line (steps 406 to 412). . At the processing position, the length of the black run that crosses the vicinity of the center of gravity of the ruled line and is perpendicular to the vector direction of the ruled line is detected (step 407). This is the same as step 208 in FIG. 2 (see FIG. 3B). The size of the detected black run length and ruled line thickness is determined (step 408). If the length of the black run is equal to or smaller than the thickness of the ruled line, the black run is likely to be a black run constituting the ruled line, and the process proceeds to step 409. In this step 409, a black run erase region is created and updated by integrating the continuous black runs determined to be highly likely to be the ruled line constituent black run in step 408. If the processing position has not reached the end of the ruled line, the process returns from step 412 to step 406 to continue the process by advancing the processing position by one.
[0035]
  If it is determined in step 408 that the length of the detected black run is larger than the ruled line thickness, the size of the black run length and a threshold selected sufficiently larger than the ruled line thickness or character thickness is determined (step). 410). If it is determined that the length of the black run is equal to or greater than the threshold value, it is regarded as the black run of the intersected character, the black run is excluded from the erasure target, and the process returns from step 412 to step 406 to the next processing position. Proceed to the process. If it is determined in step 410 that the length of the black run is smaller than the threshold, the process proceeds to step 411. In this step 411, a black run erase area including a continuous black run whose length is determined to be smaller than the threshold value in step 410 is created and updated (step 411). Then, the process returns from step 412 to step 406 to continue the process at the next process position.
[0036]
  When the processing is completed up to the end of the noticed ruled line, the processing loop of steps 406 to 412 is exited, and the process returns to step 405 to start processing for the next unprocessed ruled line. When the processing of the last ruled line is completed, the processing loop of steps 405 to 413 is exited, and the processing loop of steps 415 to 424 is entered.
[0037]
  One black run erase area is selected (step 415), and it is checked whether its size (size in the length direction of the ruled line) is smaller than a predetermined threshold (step 416). If the size is smaller than the threshold value, the black run erase region is regarded as not a ruled line, and is excluded from the erase target. The process returns from step 424 to step 415 to proceed to the next black run erase region.
[0038]
  When the black run erase area is larger than a predetermined threshold, the black run length distribution (thickness distribution) in the black run erase area, and the black run center of gravity and ruled lines described with reference to FIG. A deviation from the center of gravity is obtained (step 417).
[0039]
  With reference to the obtained black run length distribution, it is examined whether the black run length is substantially constant over the entire region and whether the black run length exceeds the ruled line thickness (step 418). If this condition is satisfied, the black run erase region can be determined as a ruled line. Therefore, all the black run pixels in the region are replaced with white pixels, and the black run is entirely erased (step 419). End processing for the region.
[0040]
  If the condition of step 418 is not satisfied, it is checked whether there is a section in which black run lengths slightly larger than the ruled line thickness and small changes are continuous, and a black run length considerably larger than the ruled line thickness exists before and / or after (step) 420). If there is such a section, it is considered that there is an overlap or contact between the character and the ruled line, so the black run in the section is partially erased so as to leave a part corresponding to the character line ( Step 421) and then proceed to Step 422.
[0041]
  This partial erasure will be described with reference to FIG. FIG. 11A shows an example in which the character 161 overlaps or contacts the ruled line 160. The overlapping or contacting portion 162 is shown enlarged in FIG. A section 163 shown in FIG. In this example, the center of gravity of the black run in the section 163 is shifted upward with respect to the center of gravity of the ruled line. Therefore, only the length corresponding to the reference value is left from the upper end of the black run in the section 163, thereby erasing the lower end portion. As the reference value, the thickness of a character given in advance by the system or designated by the user can be used. The result of such partial erasure is as shown in FIG. In FIG. 11C, the shaded portion is the erased portion. It should be noted that the sections 164 and 165 of the ruled line 160 are extracted as separate black run erasure areas and are erased by step 419, respectively.
[0042]
  Next, with reference to the information on the center of gravity deviation acquired in step 417, it is checked whether there is a portion where the difference between the center of gravity of the black run and the center of gravity of the ruled line is larger than a predetermined threshold (step 422). This threshold value may be the same value as the threshold value used in step 314 of FIG. If there is a part where the deviation of the center of gravity is large, it is considered that the character line intersects the ruled line diagonally. Therefore, in step 423, as in step 315 of FIG. A process is performed in which the black pixel in the triangular area is replaced with a white pixel so that it does not remain as noise.
[0043]
  In this way, the black run erase areas are processed one by one, and when the process of the last black run erase area is finished, the process loop is exited at step 424. Then, the image data after the erasure process is output (step 425), and the process is completed.
[0044]
【The invention's effect】
  As is clear from the above explanation,According to the present invention,Without using a sample, it is possible to erase a ruled line from an image such as an arbitrary form while saving a character line close to the ruled line.In particular,Precise ruled line erasure is possible without leaving an unerased part of the ruled line where the character line intersects the ruled line diagonally.Also,Precise ruled line erasing is possible without leaving the unerased part of the ruled line at the overlapping or contacting part of the ruled line and the character line.
[Brief description of the drawings]
FIG. 1 is a block diagram showing an example of a computer system that can be used when the present invention is implemented by software.
FIG. 2 is a flowchart of ruled line erasing processing according to the first embodiment of the present invention.
FIG. 3A is a diagram for explaining detection of an orthogonal part of a ruled line;
  (B) It is a figure for demonstrating the detection of the length of the black run of a ruled line.
FIG. 4A is a diagram illustrating an example in which a character line obliquely intersects a ruled line.
  (B) It is an enlarged view for demonstrating the noise by ruled line deletion omission in the part which a ruled line and a character line cross | intersect diagonally.
FIG. 5 is a flowchart of ruled line erasing processing according to the second embodiment of the present invention.
FIG. 6A is a diagram illustrating an example in which a character line is orthogonal to a ruled line.
  (B) It is an enlarged view for demonstrating the process with respect to the orthogonal part of a ruled line and a character line.
FIG. 7 is a diagram for explaining a shift between the center of gravity of a ruled line and the center of gravity of a black run at a portion where a character line obliquely intersects with the ruled line, and ruled line erasing processing;
FIG. 8 is a diagram illustrating an example of an oblique intersection of a ruled line and a character line, and a shift between the center of gravity of the ruled line and the center of gravity of a black run at that portion;
FIG. 9 is a first half of a flowchart of ruled line erasing processing according to the third embodiment of the present invention.
FIG. 10 is a second half of a flowchart of ruled line erasing processing according to the third embodiment of the present invention.
FIG. 11A is a diagram showing an example of overlapping or contact between ruled lines and character lines.
  (B) It is an enlarged view for demonstrating the ruled line erasure | elimination process with respect to the overlap or contact part of a ruled line and a character line.
  (C) It is an enlarged view which shows the result of the ruled line erasure | elimination process with respect to the overlap or contact part of a ruled line and a character line.
[Explanation of symbols]
  14 Central processing unit
  16 System memory
  24 display
  30 keyboard
  32 Fixed disk drive
  33 Floppy disk drive
  33A floppy disk
  34 Image scanner
  100, 101 Ruled rectangle
  102 Orthogonal part of ruled line
  121 character line
  122 Ruled Line
  131 Ruled Line
  132 character line
  140 Ruled Line
  141 Character line
  142 Line connecting the center of gravity of the ruled line
  143 Line connecting the center of gravity of black run
  150 Ruled Line
  151 Line connecting the center of gravity of the ruled line
  152,153 Character line
  154 Line connecting the center of gravity of the black run
  155, 156 Unerased part of ruled line
  160 Ruled Line
  161 Text line
  163 Overlapping or contacting part of character line and ruled line

Claims (3)

A ruled line erasing method that erases the ruled line of the part that intersects and touches the character line from an image in which characters and ruled lines are mixed,
Detecting a black run perpendicular to the ruled line;
Detecting a deviation of the center of gravity of the ruled line and the black run;
Among black runs, a black run having a length that is sufficiently smaller than a threshold value that is larger than the thickness of the ruled line and the deviation of the center of gravity is greater than a predetermined threshold, and a standard obtained by subtracting the thickness of the ruled line from the length of the black run Obtaining a value, storing a length portion corresponding to the reference value from an end in a direction where the center of gravity is shifted, and erasing the remaining portion;
A ruled line erasing method comprising: A ruled line erasing method that erases the ruled line of the part that intersects and touches the character line from an image in which characters and ruled lines are mixed,
Detecting a black run perpendicular to the ruled line;
Detecting a deviation of the center of gravity of the ruled line and the black run;
For the black run, a black run with a substantially constant length that is a predetermined amount larger than the thickness of the ruled line continues, and there is a black run with a length that is sufficiently larger than the thickness of the ruled line before or after the black run. A reference value obtained by subtracting the thickness of the ruled line from the length of the black run, and a length portion corresponding to the reference value from the end in the direction in which the center of gravity is shifted. Saving and deleting the rest,
A ruled line erasing method comprising: A ruled line erasing method that erases the ruled line of the part that intersects and touches the character line from an image in which characters and ruled lines are mixed,
Detecting a black run perpendicular to the ruled line;
Detecting a deviation of the center of gravity of the ruled line and the black run;
The black run has a length that is smaller than a threshold sufficiently larger than the thickness of the ruled line, and for a black run in which the deviation of the center of gravity is larger than a predetermined threshold, a reference value obtained by subtracting the thickness of the ruled line from the length of the black run Storing the length portion corresponding to the reference value from the end of the direction of the center of gravity shift, and erasing the remaining portion;
For the black run, a black run with a substantially constant length that is a predetermined amount larger than the thickness of the ruled line continues, and there is a black run with a length that is sufficiently larger than the thickness of the ruled line before or after the black run. For the black run in the section, a reference value obtained by subtracting the thickness of the ruled line from the length of the black run is obtained, and the length corresponding to the reference value from the end in the direction where the center of gravity is shifted is obtained. Saving the part and erasing the remaining part,
A ruled line erasing method comprising:

Images