JP2002074264A - Image processing apparatus, image processing method, and recording medium - Google Patents

Abstract

(57) [Summary] [Problem] A ruled line that hinders character recognition is not erased by mistake even if the ruled line intersects with a character. It is an object of the present invention to realize an image processing apparatus capable of securely erasing even in a case. SOLUTION: An image to be processed is fetched by an image input means 1, a position of a ruled line in the fetched image data is estimated by a ruled line position estimating means 3, and a black run in a direction perpendicular to the ruled line is estimated in the vicinity of the estimated position. The regression line passing through the center of the extracted run data extracted by the run extraction means 4 is determined by the regression line calculation means 6, and the distance between the regression line and the run data is determined by the distance calculation means 8. The maximum value of the distance and the threshold value are compared by the maximum distance value examination means 10, and when the maximum value is equal to or larger than the threshold value, the run data giving this maximum value is excluded by the run data editing means 9, and the run erasing means 7 is deleted. Then, the position on the image corresponding to the position of the run data whose maximum value is less than the threshold value is replaced with a white pixel.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus and a method for optically recognizing characters written on a sheet of paper, and a storage medium storing a program for causing a computer to execute the method. TECHNICAL FIELD The present invention relates to an image processing apparatus and an image processing method capable of accurately deleting an image, and a storage medium storing the program.

[0002]

2. Description of the Related Art Hitherto, the main method of erasing a ruled line which hinders recognition is based on a long black run (a continuous black pixel row) constituting a ruled line. JP-A-10-171922 also first extracts a black run longer than a certain length in the ruled line direction (longitudinal direction), and calculates the sum of the lengths of the extracted runs in the direction perpendicular to the direction of the black run. For each block, the sum is divided into a plurality of blocks. If there is one block, it is recognized and deleted as a ruled line. If there are two or more blocks, the largest black run length of the plurality of blocks is used. Blocks having a total value are recognized as a group of black runs constituting a ruled line, and these are erased, so that a scheme is devised to prevent erroneous erasure of characters confused with the ruled line. However, this method can prevent erroneous erasure of a character that does not touch the ruled line, but when the character and the ruled line intersect, even if the intersection is a component of the character, it is erased together with the ruled line. I will. Further, when the image state is poor and the ruled lines are blurred, the black run in the ruled line direction is cut short, and it becomes difficult to extract a long black run. Further, since a long black run is required, it is difficult to erase dotted ruled lines.

[0003]

As described above, in the conventional image processing method for removing ruled lines which hinders character recognition and such an apparatus, ruled lines are recognized based on the line length of a black run. In the case of fading or dotted ruled lines, erasing was difficult. Further, when the character and the ruled line intersect, there is a problem that the constituent elements of the character are erased together with the ruled line. The present invention solves this problem in a relatively simple manner so that even if the ruled line intersects with the character, the character is not erased by mistake and the ruled line is blurred, It is another object of the present invention to realize an image processing apparatus, an image processing method, and a storage medium that stores a program for causing a computer to execute the image processing method, in which the ruled line can be accurately erased even in the case of (1).

[0004]

According to the present invention, there is provided an image processing apparatus for performing optical character recognition, comprising: an image input unit for capturing an image to be processed;
Image data storing means for storing the image data captured by the image input means; ruled line position estimating means for estimating the position of ruled lines in the image data stored in the image data storing means; and ruled line position estimating means. Vertical run extracting means for extracting a black run in a direction perpendicular to the ruled line near the estimated position of the ruled line; run data storing means for storing the run extracted by the vertical run extracting means; Regression line calculation means for obtaining a regression line passing through the center of the run data extracted by the vertical direction run extraction means; and run straight line distance calculation for obtaining a distance between the regression line calculated by the regression line calculation means and the run data Means, distance maximum value examination means for comparing the maximum value of the distance calculated by the run straight line distance calculation means with a predetermined threshold value, and this distance maximum Run data editing means for excluding run data that gives this maximum value when the maximum value is equal to or greater than the predetermined threshold value by comparison in the examination means, and the maximum value is less than a predetermined threshold value. A run erasing unit that replaces a portion corresponding to the position of the run data stored in the run data storage unit with a white pixel on an image in some cases.

In an image processing apparatus for performing optical character recognition, image input means for capturing an image to be processed, image data storage means for storing image data captured by the image input means, and image data storage means A ruled line position estimating means for estimating the position of a ruled line in the image data stored in the means; and a vertical line for extracting a black run in a direction perpendicular to the ruled line near the estimated position of the ruled line estimated by the ruled line position estimating means. Direction run extraction means, run data storage means for storing the runs extracted by the vertical direction run extraction means, and a histogram indicating the distribution of the length of the run data extracted by the vertical direction run extraction means. A run length histogram calculating means, and a run length mode value calculating means for obtaining a mode value of the length from the histogram obtained by the run length histogram calculating means. A run length data exclusion unit that excludes a run having a length significantly different from the mode value of the length obtained by the run length mode value calculation unit from the data; Regression line calculation means for obtaining a regression line passing through the center of the run data stored in the image data storage means, and a run line for obtaining a distance between the regression line calculated by the regression line calculation means and the run data The distance distance calculating means, the distance maximum value examining means for comparing the maximum value of the distance calculated by the run straight line distance calculating means with a predetermined threshold, and the distance maximum value examining means, Run data editing means for excluding run data that gives the maximum value when the maximum value is equal to or greater than the predetermined threshold value, and the maximum value is predetermined Characterized by comprising a run-erasing means for replacing a portion corresponding to the position of the run data stored in the case is less than the value in the run-data storage means to a white pixel in the image. As a result, a ruled line that hinders character recognition is not accidentally erased even when the ruled line and the character intersect,
It is possible to realize an image processing apparatus capable of accurately deleting even when a ruled line is blurred or a dotted ruled line.

Further, in an image processing method for performing optical character recognition, an image inputting step of capturing an image to be processed, an image data storing step of storing image data captured in the image inputting step, and an image data storing step of storing the image data A ruled line position estimating step of estimating a position of a ruled line in the image data stored in the process, and a vertical run for extracting a black run in a direction perpendicular to the ruled line near the estimated position of the ruled line estimated in the ruled line position estimating process A direction run extraction process, a run data storage process for storing the runs extracted in the vertical run extraction process, and a regression line for finding a regression line passing through the center of the run data extracted in the vertical run extraction process. A calculating step, a run straight line distance calculating step of finding a distance between the regression line calculated in the regression straight line calculating step and the run data,
A distance maximum value examining process for comparing the maximum value of the distance calculated in the run straight line distance calculating process with a predetermined threshold, and the maximum value is determined in advance by comparison in the distance maximum value examining process. The run data editing step of excluding the run data giving the maximum value when the maximum value is equal to or more than the set threshold value is stored in the run data storage step when the maximum value is less than a predetermined threshold value. A run erasing step of replacing a portion corresponding to the position of the run data with a white pixel on the image.

In an image processing method for performing optical character recognition, an image inputting step of capturing an image to be processed, an image data storing step of storing image data captured in the image inputting step, and an image data storing step A ruled line position estimating step of estimating a position of a ruled line in the image data stored in the process, and a vertical run for extracting a black run in a direction perpendicular to the ruled line near the estimated position of the ruled line estimated in the ruled line position estimating process A direction run extraction process, a run data storage process for storing the runs extracted in the vertical run extraction process, and a histogram showing the distribution of the length of the run data extracted in the vertical run extraction process are obtained. A run length histogram calculation process, and a run length mode calculation process for obtaining a mode value of the length from the histogram obtained by the run length histogram calculation process. And a run length data exclusion process in which runs having a length significantly different from the mode of the length obtained in the run length mode calculation process are excluded from the data. A regression line calculation step for obtaining a regression line passing through the center of the run data stored in the image data storage step, and a run line for obtaining a distance between the regression line calculated in the regression line calculation step and the run data. The distance calculation step, a distance maximum value examination step of comparing the maximum value of the distance calculated in the run straight line distance calculation step with a predetermined threshold value, and the comparison in the distance maximum value examination step, A run data editing process for excluding run data that gives the maximum value when the maximum value is equal to or greater than the predetermined threshold value, and the maximum value is predetermined. Characterized by comprising a run erasing process of replacing a portion corresponding to the position of the run data stored in the case it is less than the value in the run-data storage process to a white pixel in the image. As a result, a ruled line that hinders character recognition is not accidentally erased even when the ruled line and the character intersect,
It is possible to realize an image processing method capable of accurately deleting even when a ruled line is blurred or a dotted ruled line.

Further, an image input function for capturing an image to be processed on a recording medium readable by the computer in which a program for realizing a processing method for performing optical character recognition is stored and read by the computer, and an image captured by the image input function. An image data storage function for storing image data, a ruled line position estimating function for estimating a position of a ruled line in the image data stored by the image data storing function, and a ruled line estimated position estimated by the ruled line position estimating function A vertical run extraction function for extracting a black run in a direction perpendicular to the ruled line in the vicinity, a run data storage function for storing runs extracted by the vertical run extraction function, and a vertical run extraction function. A regression line calculation function to find a regression line passing through the center of the extracted run data, and a regression line calculated by this regression line calculation function A line-to-run-line distance calculating function for determining the distance between a line and the run data, and a distance-maximum-value examining function for comparing a maximum value of the distance calculated by the run-line-to-run distance calculating function with a predetermined threshold value, A run data editing function for excluding run data that gives the maximum value when the maximum value is equal to or greater than the predetermined threshold value by comparison with the distance maximum value examination function, and the maximum value is determined in advance. A run erasing function for replacing a portion corresponding to the position of the run data stored in the run data storage function with a white pixel on an image when the value is smaller than the threshold value.

Further, an image input function for capturing an image to be processed on a recording medium storing a program for causing a computer to execute a processing method for performing optical character recognition on a recording medium readable from the computer, and the image input function for capturing the image to be processed. An image data storage function for storing image data, a ruled line position estimating function for estimating a position of a ruled line in the image data stored by the image data storing function, and a ruled line estimated position estimated by the ruled line position estimating function A vertical run extraction function for extracting a black run in a direction perpendicular to the ruled line in the vicinity, a run data storage function for storing runs extracted by the vertical run extraction function, and a vertical run extraction function. A run length histogram calculation function for obtaining a histogram indicating the distribution of the length of the extracted run data,
A run length mode calculation function for obtaining a mode of length from the histogram obtained by the run length histogram calculation function is significantly different from a mode of length obtained by the run length mode calculation function. A different-length run data exclusion function for excluding a run having a length from data, and a regression line passing through the center of the run data stored in the image data storage function that has not been excluded by the different-length run data exclusion function. The regression line calculation function to be obtained, the regression line calculated by the regression line calculation function, the inter-run line distance calculation function to obtain the distance between the run data, and the maximum value of the distance calculated by the inter-run line distance calculation function A case where the maximum value is equal to or greater than the predetermined threshold value by comparing the distance maximum value inspection function with a predetermined threshold value and the distance maximum value inspection function. A run data editing function for excluding the run data giving the maximum value, and a portion corresponding to the position of the run data stored in the run data storage function when the maximum value is less than a predetermined threshold. A run erasing function for replacing white pixels on an image. As a result, the ruled line that hinders character recognition is not erased accidentally even when the ruled line intersects with the character. A computer-readable recording medium that stores a program that causes a computer to execute the erasable image processing method can be realized.

[0010]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an image processing method and an image processing apparatus according to the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a block diagram showing the configuration of the first embodiment of the image processing apparatus of the present invention. In FIG.
Reference numeral 1 denotes an image input unit such as a scanner that takes in an image to be processed, reference numeral 2 denotes an image storage unit that stores an incoming image, and reference numeral 3 denotes a ruled line position estimating unit that roughly determines the position of a ruled line to be erased in the image. Reference numeral 4 denotes a vertical run extracting means for extracting a black run in a direction perpendicular to the ruled line near the ruled line, reference numeral 5 denotes a run data storage means for storing the extracted run data, and reference numeral 6 denotes a regression passing through the center of the run. Regression line calculating means for obtaining a straight line, reference numeral 7 denotes a run erasing means for deleting a run from an image based on the run data, reference numeral 8 denotes a run straight line distance calculating means for obtaining the distance between each run data and the regression line, and reference numeral 9 denotes Reference numeral 10 denotes run data editing means for deleting stored run data, which is a distance maximum value examining means for examining the maximum value of the distance between the run and the regression line and determining the subsequent processing.

FIG. 2 is a flowchart showing the processing operation of this embodiment. The binary image to be processed is captured by the image input means 1 and stored in the image data storage means 2 (step 101). Next, the position of the ruled line to be erased is estimated by the ruled line position estimating means 3 (step 10).
2). Various methods are conceivable for estimating the ruled line position. For example, as described in Japanese Patent Application Laid-Open No. Hei 10-91783, registration is performed with a blank form image of the same format, and position coordinates on this image are used. A method of estimating the position of the character entry column on the image to be processed from the defined character entry column and setting the surrounding area as a range where ruled lines exist is used. When the range in which the ruled line exists is obtained, a black run in the direction perpendicular to the direction of the ruled line is obtained by the vertical direction run extracting means 4 (step 103), and the run data is stored in the run data storing means 5.

In this case, as shown in FIG. 3, run extraction is performed in a slightly wider range than the position where it is estimated that there is a ruled line so that there is no omission of extraction. In such a case, some of the characters are also erased, which adversely affects the subsequent character recognition processing. As shown in FIG. 4, the center of the run tends to be away from the regression line in a portion independent of the ruled line or in a portion where characters intersect. Therefore, it is only necessary to exclude runs that are far from the regression line and erase the ruled line using the remaining run data portion.However, since run data that is not a ruled line is also used for calculating the regression line, multiple runs are excluded at once. Then, there is a possibility that a run that should be a ruled line may be excluded. Therefore,
By excluding only one run farthest from the regression line, calculating the regression line again, and calculating the distance to each run again, the accuracy of the regression line passing through the ruled line can be improved.
By repeating this, only the runs that make up the ruled line remain in the data, and the ruled line can be erased with high accuracy.

For this purpose, as a next step, a regression line passing through the center of the run is determined by the regression line calculation means 6 (step 104). The regression line is obtained as follows. Assuming that the main scanning direction and the sub scanning direction are the X direction and the Y direction, respectively, in the case of the ruled line in the main scanning direction, the equation of the regression line is

Let y = a + bx, and let the center coordinates of the run be

(Xi, Yi) (i = 1, 2, 3,... N) (N is the total number of runs)

[0017]

(Equation 1)

The coefficients a and b are calculated using

[0019]

(Equation 2)

Can be written as Similarly, in the sub-scanning direction,

The coefficients c and d of x = c + dy are

[0022]

(Equation 3)

[0023]

Next, the maximum value of the distance between each run and the regression line is obtained by the run straight line distance calculating means 8 (step 10).
5), to reach the maximum distance examination means 10; Here, the predetermined value is compared with the maximum value (step 106). If the maximum value is larger, the run giving the maximum value is excluded by the run data editing means 9 (step 107). The process returns to step 104 for obtaining a regression line again with the remaining runs. If the maximum value is smaller, the run erasing means 7 replaces the run data portion on the screen with white pixels based on the stored run data information (step 108). This operation is performed for all the ruled lines to be processed (step 109). When the operation is completed, a result image is output (step 110), and the processing is terminated.

Although the ruled line to be erased is a solid line in the figure, the ruled line can be erased in exactly the same manner as a dotted line. Further, since run information in the ruled line direction is not used in this method, even if the ruled line is blurred and short, the accuracy of erasure is less reduced than in the case of using run information in the ruled line direction.

As shown in FIG. 5, even when characters intersect, if a run exceeds the width of the ruled line search range, the center of the run accidentally comes near the regression line, and Despite being a copy, it may be erroneously targeted for erasure. To avoid this, you can increase the width of the search range so that runs do not exceed this width,
If the width is widened, it will take time for run extraction and distance calculation. Therefore, as shown below, the width of the ruled line is first estimated from the mode value of the run histogram, and runs with a certain distance from them are excluded before calculating the regression curve. Erroneous erasure can be prevented without significantly increasing the time.

FIG. 6 is a block diagram showing the configuration of the second embodiment of the present invention. Basically, it is the same as the first embodiment, but the portion on the left side of the figure is a portion added in the present embodiment. In FIG. 6, the portions denoted by reference numerals 1 to 10 are the same as those in FIG. Reference numeral 21 denotes a run length histogram calculating unit that obtains a histogram of the run length based on the run data, reference numeral 22 denotes a run length mode calculating unit that obtains a mode of the histogram, and reference numeral 23 denotes a most frequent run. This is a different-length run data exclusion means for selecting run data greatly different from the length and deleting the run data from the run data.

FIG. 7 is a flowchart showing the processing operation of this embodiment. Also in this case, the major flow is the same as the flowchart of FIG.
Since a new step is added during the step 04, only the added part is shown. First, in step 103, a run in the direction perpendicular to the ruled line is extracted, and a run length histogram is obtained by the run length histogram calculating means 21 (step 201). Next, the mode value of the run length is obtained by the run length mode value calculating means 22 based on the obtained histogram (step 202). Since this is considered to be a value representative of the width of the ruled line, it is considered that the length of the run is far from the width is not a ruled line.
In step 3, a run having a length difference equal to or larger than a predetermined value is excluded from the run data so as not to be used in the subsequent processing (step 203). By adding the processing shown in FIG. 7, the runs having different lengths are not included in the calculation of the regression line from the beginning as shown in FIG.

FIG. 9 is a block diagram showing the configuration of the third embodiment of the present invention. This configuration is a configuration when the present invention is realized by software. 9, reference numeral 31 denotes a CPU, reference numeral 32 denotes a memory, reference numeral 33 denotes a hard disk, reference numeral 34 denotes an input device, and reference numeral 35 denotes a CD-RO.
M drive, 36 is a display, 37 is a CD
-A recording medium such as a ROM. In this configuration, a general-purpose processing device is used, and the recording medium 37 stores a program for realizing the processing function and processing procedure of ruled line erasing of the present invention. The document image to be processed is stored in, for example, the hard disk 33 or the like. The CPU 31
A program for realizing the processing function and processing procedure of ruled line erasure is read out from the recording medium 37 and executed, and the result of ruled line erasure is output to the display 36 or the like.

[0030]

As described above, according to the first aspect of the present invention, in an image processing apparatus for performing optical character recognition, image input means for capturing an image to be processed and the image input means for capturing the image to be processed. Image data storage means for storing image data; ruled line position estimating means for estimating ruled line positions in the image data stored in the image data storing means; and ruled line estimated position estimated by the ruled line position estimating means A vertical run extracting means for extracting a black run in a direction perpendicular to the ruled line, a run data storing means for storing runs extracted by the vertical run extracting means, and a vertical run extracting means. Regression line calculation means for obtaining a regression line passing through the center of the run data, and run line distance calculation means for obtaining the distance between the regression line calculated by the regression line calculation means and the run data A distance maximum value examining means for comparing the maximum value of the distance calculated by the run straight line distance calculating means with a predetermined threshold, and a maximum value determined in advance by comparison with the distance maximum value examining means. A run data editing unit that excludes run data that gives this maximum value when the maximum value is equal to or greater than a set threshold value; and a run data editing unit that stores run data stored in the run data storage unit when the maximum value is less than a predetermined threshold value. A run erasing means for replacing a portion corresponding to a position with a white pixel on an image is provided. As a result, a ruled line that hinders character recognition is removed even if a character is in contact with the ruled line by leaving a large distance between the regression line and the run and deleting runs near the other regression line. Image processing that can identify the pixels that make up the ruled line with high accuracy, so that characters are not erased accidentally, and even if the ruled lines are blurred or dotted ruled lines can be erased with high accuracy The device can be realized.

In an image processing apparatus for performing optical character recognition, image input means for capturing an image to be processed, image data storage means for storing image data captured by the image input means, and image data storage means Ruled line position estimating means for estimating the position of a ruled line in the image data stored in the means; and a vertical run for extracting a black run in a direction perpendicular to the ruled line near the estimated position of the ruled line estimated by the ruled line position estimating means. Extraction means, run data storage means for storing the runs extracted by the vertical run extraction means, and run length histogram for obtaining a histogram indicating the distribution of the length of the run data extracted by the vertical run extraction means Calculation means; run length mode value calculation means for obtaining a mode value of the length from the histogram obtained by the run length histogram calculation means; Different-length run data exclusion means for excluding runs having a length significantly different from the mode of the length determined by the long mode calculation means from data, and image data not excluded by the different-length run data exclusion means Regression line calculation means for obtaining a regression line passing through the center of the run data stored in the storage means; and run line distance calculation means for obtaining a distance between the regression line calculated by the regression line calculation means and the run data; Distance maximum value examination means for comparing the maximum value of the distance calculated by the run straight line distance calculation means with a predetermined threshold, and the maximum value was previously determined by comparison with this distance maximum value examination means. A run data editing unit for excluding run data that gives the maximum value when the maximum value is equal to or more than the threshold value; and a run data storage unit when the maximum value is less than a predetermined threshold value. And characterized by providing a run-erasing means for replacing the white pixels to a portion corresponding to the position of the run data stored in the image to. As a result, ruled lines that hinder character recognition
If the run length is significantly different from the mode value, objects with a large distance between the regression line and the run are excluded from the deletion target, and runs near the other regression lines are deleted, so that even if characters touch the ruled line, Since the pixels that make up the ruled line can be identified with high precision, characters can be erased with accuracy, even if the ruled line is blurred or a dotted ruled line is not erased by mistake. An image processing device can be realized.

Further, in an image processing method for performing optical character recognition, an image inputting step of capturing an image to be processed, an image data storing step of storing image data captured in the image inputting step, and an image data storing step of storing the image data A ruled line position estimating process for estimating the position of the ruled line in the image data stored in the process, and a vertical run for extracting a black run in a direction perpendicular to the ruled line near the estimated position of the ruled line estimated in the ruled line position estimating process An extraction process, a run data storage process for storing the runs extracted in the vertical run extraction process,
A regression line calculation process for obtaining a regression line passing through the center of the run data extracted in the vertical direction run extraction process, a run line distance calculation process for obtaining a distance between the regression line calculated in the regression line calculation process and the run data, The maximum value of the distance calculated in the run straight line distance calculation process is compared with a predetermined threshold value, and the maximum value is determined in advance by comparison in the distance maximum value inspection process. A run data editing process that excludes run data that gives the maximum value when the maximum value is equal to or greater than the set threshold value, and a run data storage process that stores the run data stored in the run data storage process when the maximum value is less than a predetermined threshold value. A run erasing step of replacing a portion corresponding to the position with a white pixel on the image. As a result, a ruled line that hinders character recognition is removed even if a character is in contact with the ruled line by leaving a large distance between the regression line and the run and deleting runs near the other regression line. Since the pixels that make up the ruled line can be identified with high accuracy, characters are not accidentally erased.
It is possible to realize an image processing method capable of accurately deleting even a dotted ruled line.

In an image processing method for performing optical character recognition, an image inputting step of capturing an image to be processed, an image data storing step of storing image data captured in the image inputting step, and an image data storing step A ruled line position estimating process for estimating the position of the ruled line in the image data stored in the process, and a vertical run for extracting a black run in a direction perpendicular to the ruled line near the estimated position of the ruled line estimated in the ruled line position estimating process An extraction process, a run data storage process for storing the runs extracted in the vertical run extraction process, and a run length histogram for obtaining a histogram indicating the distribution of the length of the run data extracted in the vertical run extraction process Calculating a run length mode from the histogram obtained by the run length histogram calculating process. A different-length run data exclusion process for excluding runs having a length significantly different from the mode of the length determined by the long mode calculation process, and image data not excluded in the different-length run data exclusion process A regression line calculation process for obtaining a regression line passing through the center of the run data stored in the storage process, a run line distance calculation process for obtaining a distance between the regression line calculated in the regression line calculation process and the run data, A maximum distance value examination process in which the maximum value of the distance calculated in the run straight line distance calculation process is compared with a predetermined threshold value, and a maximum value is determined in advance by comparison in the distance maximum value examination process. A run data editing process for excluding the run data that gives the maximum value when the maximum value is equal to or more than the threshold value, and a run data storage error when the maximum value is less than the predetermined threshold value. Characterized by comprising a run erasing process to replace the white pixel portion corresponding to the position of the run data stored in the image to.
As a result, ruled lines that hinder character recognition are excluded from runs that have run lengths that are significantly different from the mode, and that have a large distance between the regression line and the run. By erasing the run, even if a character is in contact with the ruled line, it is possible to identify the pixels constituting the ruled line with high accuracy, so that the character is not erased by mistake and the ruled line is blurred. It is possible to realize an image processing method capable of accurately deleting even a case or a dotted ruled line.

Further, an image input function for taking in an image to be processed in a recording medium which stores a program for realizing a processing method for performing optical character recognition in a computer and which is readable by the computer, An image data storage function for storing image data; a ruled line position estimating function for estimating a ruled line position in the image data stored by the image data storing function; and a ruled line estimated position estimated by the ruled line position estimating function The vertical run extraction function that extracts black runs in the direction perpendicular to the ruled line, the run data storage function that stores the runs extracted by this vertical run extraction function, and the vertical run extraction function A regression line calculation function for finding a regression line that passes through the center of the run data, and the regression line Distance-to-run-line distance calculation function to find the distance of the data, a distance-maximum value examination function to compare the maximum value of the distance calculated by the run-line-to-line distance calculation function with a predetermined threshold value, and this distance maximum value A run data editing function that excludes run data that gives this maximum value when the maximum value is equal to or greater than a predetermined threshold value by comparison with the examination function, and a case where the maximum value is less than the predetermined threshold value A run erasing function for replacing a portion corresponding to the position of the run data stored in the run data storage function with a white pixel on an image. As a result, a ruled line that hinders character recognition is removed even if a character is in contact with the ruled line by leaving a large distance between the regression line and the run and deleting runs near the other regression line. Since the pixels that make up the ruled line can be identified with high accuracy, characters are not accidentally erased,
It is possible to realize a computer-readable recording medium that stores a program for causing a computer to implement an image processing method capable of accurately deleting even a ruled line that is blurred or a dotted ruled line.

Further, an image input function for capturing an image to be processed on a recording medium which stores a program for causing a computer to execute a processing method for performing optical character recognition and which is readable from the computer, and an image input function for capturing the image to be processed. An image data storage function for storing image data; a ruled line position estimating function for estimating a ruled line position in the image data stored by the image data storing function; and a ruled line estimated position estimated by the ruled line position estimating function The vertical run extraction function that extracts black runs in the direction perpendicular to the ruled line, the run data storage function that stores the runs extracted by this vertical run extraction function, and the vertical run extraction function Run length histogram calculation function for obtaining a histogram indicating the distribution of run data lengths, and run length histogram calculation function A run length mode value calculation function for finding the mode value of the length from the obtained histogram, and a run having a length significantly different from the mode value of the length obtained by this run length mode value calculation function A regression line calculation function for finding a regression line that passes through the center of the run data stored in the image data storage function that has not been excluded by the different-length run data removal function. A run straight line distance calculating function for calculating a distance between the regression line calculated by the straight line calculating function and the run data is compared with a predetermined threshold value and a maximum value of the distance calculated by the run straight line distance calculating function. Distance maximum value examination function,
A run data editing function that excludes run data that gives this maximum value when the maximum value is greater than or equal to a predetermined threshold by comparison with this distance maximum value examination function, and a maximum value that is less than a predetermined threshold And a run erasing function for replacing a portion corresponding to the position of the run data stored in the run data storage function with a white pixel on the image. As a result, ruled lines that hinder character recognition are excluded from runs that have run lengths that differ significantly from the mode, and that have a large distance between the regression line and the run. By erasing the run, even if a character is in contact with the ruled line, the pixels constituting the ruled line can be identified with high accuracy, so that the character was not erased accidentally and the ruled line was blurred. It is possible to realize a computer-readable recording medium that stores a program for causing a computer to implement an image processing method capable of accurately deleting even a dotted line or a ruled line.

[Brief description of the drawings]

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

FIG. 2 is a flowchart showing a processing operation of the embodiment of FIG. 1;

FIG. 3 is an explanatory diagram showing a ruled line erasing operation according to the embodiment of FIG. 1;

FIG. 4 is an explanatory diagram showing a ruled line erasing operation according to the embodiment of FIG. 1;

FIG. 5 is an explanatory diagram showing a ruled line erasing operation according to the embodiment of FIG. 1;

FIG. 6 is a block diagram showing a configuration of another embodiment of the image processing apparatus of the present invention.

FIG. 7 is a flowchart showing an addition of the processing operation of the embodiment in FIG. 6 to FIG. 2;

FIG. 8 is an explanatory diagram showing a ruled line erasing operation according to the embodiment of FIG. 6;

FIG. 9 is a block diagram showing a configuration of still another embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Image input means 2 Image storage means 3 Ruled line position estimation means 4 Vertical direction run extraction means 5 Run data storage means 6 Regression line calculation means 7 Run elimination means 8 Run straight line distance calculation means 9 Run data editing means 10 Distance maximum value examination Means 21 Run length histogram calculating means 22 Run length mode value calculating means 23 Different length run data excluding means 31 CPU 32 Memory 33 Hard disk 34 Input device 35 CD-ROM drive 36 Display 37 Recording medium

Claims (6)

[Claims] 1. An image processing apparatus for performing optical character recognition, comprising: image input means for capturing an image to be processed; image data storage means for storing image data captured by the image input means; Ruled line position estimating means for estimating the position of the ruled line in the image data stored in the means; and a vertical line for extracting a black run in a direction perpendicular to the ruled line near the estimated position of the ruled line estimated by the ruled line position estimating means. Direction run extraction means, run data storage means for storing runs extracted by the vertical direction run extraction means, regression line for finding a regression line passing through the center of the run data extracted by the vertical direction run extraction means Calculating means; run straight line distance calculating means for calculating the distance between the regression line calculated by the regression straight line calculating means and the run data; Means for examining the maximum value of the distance calculated by the means and a predetermined threshold, and a means for examining the maximum value, wherein the maximum value is greater than or equal to the predetermined threshold by comparison with the means for examining the maximum distance. A run data editing unit that excludes run data that gives the maximum value in a certain case; and a position corresponding to the position of the run data stored in the run data storage unit when the maximum value is less than a predetermined threshold. An image processing apparatus comprising: run erasing means for replacing a portion to be replaced with a white pixel on an image. 2. An image processing apparatus for performing optical character recognition, comprising: image input means for capturing an image to be processed; image data storage means for storing image data captured by the image input means; Ruled line position estimating means for estimating the position of the ruled line in the image data stored in the means; and a vertical line for extracting a black run in a direction perpendicular to the ruled line near the estimated position of the ruled line estimated by the ruled line position estimating means. Direction run extraction means, run data storage means for storing runs extracted by the vertical direction run extraction means, and a histogram showing the distribution of the length of the run data extracted by the vertical direction run extraction means A run length histogram calculating means, and a run length mode value calculating means for obtaining a mode value of the length from the histogram obtained by the run length histogram calculating means. A run length data exclusion unit for excluding a run having a length significantly different from the mode value of the length obtained by the run length mode calculation unit from the data; Regression line calculation means for obtaining a regression line passing through the center of the run data stored in the image data storage means, and a run line for obtaining a distance between the regression line calculated by the regression line calculation means and the run data Distance distance calculating means, distance maximum value examining means for comparing the maximum value of the distance calculated by the run straight line distance calculating means with a predetermined threshold, and said distance maximum value examining means Run data editing means for excluding run data that gives the maximum value when the maximum value is equal to or greater than the predetermined threshold value; and the maximum value is predetermined. And a run erasing unit for replacing a portion corresponding to the position of the run data stored in the run data storage unit with a white pixel on the image when the value is smaller than the threshold value. 3. An image processing method for performing optical character recognition, comprising: an image inputting step of capturing an image to be processed; an image data storing step of storing image data captured in the image inputting step; A ruled line position estimating step of estimating a position of a ruled line in the image data stored in the process, and a vertical run for extracting a black run in a direction perpendicular to the ruled line near the estimated position of the ruled line estimated in the ruled line position estimating process A direction run extraction process, a run data storage process for storing runs extracted in the vertical run extraction process, and a regression line for finding a regression line passing through the center of the run data extracted in the vertical run extraction process. A calculation step; a run-to-run-line distance calculating step for obtaining a distance between the regression line calculated in the regression-line calculation step and the run data; A distance maximum value examination process for comparing the maximum value of the distance calculated in the process with a predetermined threshold value, and the maximum value is equal to or greater than the predetermined threshold value by comparison in the distance maximum value examination process. A run data editing process for excluding run data that gives the maximum value in some cases; and a run data editing process corresponding to the position of the run data stored in the run data storage process when the maximum value is less than a predetermined threshold. A run erasing step of replacing a portion to be replaced with a white pixel on an image. 4. An image processing method for performing optical character recognition, comprising: an image inputting step of capturing an image to be processed; an image data storing step of storing image data captured in the image inputting step; A ruled line position estimating step of estimating a position of a ruled line in the image data stored in the process, and a vertical run for extracting a black run in a direction perpendicular to the ruled line near the estimated position of the ruled line estimated in the ruled line position estimating process A direction run extraction process, a run data storage process for storing the runs extracted in the vertical run extraction process, and a histogram showing the distribution of the length of the run data extracted in the vertical run extraction process. A run length histogram calculation process, and a run length mode value calculation process for obtaining a mode value of the length from the histogram obtained by the run length histogram calculation process. The run length mode calculation step excludes runs having a length significantly different from the mode of the length obtained by the run length mode calculation process from the data. A regression line calculation step for obtaining a regression line passing through the center of the run data stored in the image data storage step, and a run line for obtaining a distance between the regression line calculated in the regression line calculation step and the run data. A distance maximum value examining process for comparing the maximum value of the distance calculated in the run straight line distance calculating process with a predetermined threshold value; and a comparison in the distance maximum value examining process. A run data editing process for excluding run data that gives the maximum value when the maximum value is equal to or greater than the predetermined threshold value; and wherein the maximum value is predetermined. A run erasing step of replacing a portion corresponding to the position of the run data stored in the run data storing step with a white pixel on the image when the difference is less than the threshold value. 5. A recording medium storing a program for causing a computer to implement a processing method for performing optical character recognition, comprising: an image input function for capturing an image to be processed; and image data captured by the image input function. An image data storage function; a ruled line position estimating function for estimating a position of a ruled line in the image data stored by the image data storing function; and a ruled line near the estimated position of the ruled line estimated by the ruled line position estimating function. A vertical run extraction function for extracting a vertical black run; a run data storage function for storing runs extracted by the vertical run extraction function; and run data extracted by the vertical run extraction function. A regression line calculation function for obtaining a regression line passing through the center of the line, and a distance between the regression line calculated by the regression line calculation function and the run data. A function for calculating the distance between run straight lines, a function for examining the maximum value of the distance calculated by the function for calculating distance between run straight lines and a predetermined threshold value, and a function for examining the maximum distance value. A run data editing function for excluding run data that gives the maximum value when the maximum value is equal to or greater than the predetermined threshold value by comparison; and when the maximum value is less than a predetermined threshold value, A run erasing function for replacing a portion corresponding to the position of the run data stored in the run data storage function with a white pixel on an image. recoding media. 6. A recording medium storing a program for causing a computer to execute a processing method for performing optical character recognition, an image input function for capturing an image to be processed, and image data captured by the image input function. An image data storage function; a ruled line position estimating function for estimating a position of a ruled line in the image data stored by the image data storing function; and a ruled line near the estimated position of the ruled line estimated by the ruled line position estimating function. A vertical run extraction function for extracting a vertical black run; a run data storage function for storing runs extracted by the vertical run extraction function; and run data extracted by the vertical run extraction function. Run length histogram calculation function to obtain a histogram indicating the distribution of the length of the histogram, and the histogram obtained by the run length histogram calculation function Run length mode value calculation function to find the mode value of the length from the system, and runs with lengths significantly different from the mode value of the length obtained by this run length mode value calculation function are excluded from the data. A regression line calculation function for obtaining a regression line passing through the center of the run data stored in the image data storage function that has not been excluded by the different-length run data removal function; A run straight line distance calculating function for finding the distance between the regression straight line calculated by the straight line calculating function and the run data, and comparing the maximum value of the distance calculated by the run straight line distance calculating function with a predetermined threshold value And a run for excluding run data that gives the maximum value when the maximum value is equal to or greater than the predetermined threshold value by comparison between the distance maximum value examination function and the distance maximum value examination function. A data editing function, and a run erasing function for replacing a portion corresponding to the position of the run data stored in the run data storage function with a white pixel on an image when the maximum value is less than a predetermined threshold value. And a computer-readable recording medium storing a program for an image processing method.