JP2000182030A - Image segmenting method and image segmenting device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To increase processing speed by eliminating useless skew correction in an image segmenting device to segment a field image at a specified position from a slip image. SOLUTION: In this image segmenting method, a layout is discriminated for the slip image 1m inputted from an image input part 1 by a bar code retrieving and recognizing part 3 and a kind of the slip is specified by a slip layout retrieving part 6. A position of a field for the purpose of segmenting the object slip is calculated from a slip layout data base 8 and an area including the surroundings of the field is calculated by an image segmenting part 9. Correction corresponding skew quantity Sq of the slip is performed only to the area and a desired field image 25 is segmented from the area at the image segmenting part 9.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is applied to a character reading system or the like for reading characters and bar codes described on a form, and an image cutting method and image for cutting out a field to be read at a predetermined place from a form image. The present invention relates to a cutting device.

[0002]

2. Description of the Related Art A conventional image cutout apparatus cuts out fields to be read from an input form image. However, when a plurality of form images are input, if the form images are not in the correct direction or skew occurs. May have. In this case, the entire form image is rotated and corrected, and a field to be read is cut out from the corrected image.

[0003]

However, in the conventional image clipping apparatus and image clipping method, since the entire form image is rotated and corrected, it takes a lot of processing time.

[0004]

In order to solve the above problems, a first aspect of the present invention is to correct a skew of a field image at a predetermined place in a target form from an input form image of the target form. The following method is employed in an image cutting method for cutting in a state. That is, the form image composed of image data is input, and image input processing for storing the form image in the image data memory, and searching the form image stored in the image data memory to determine the direction and the skew amount of the target form. A layout identification process for detecting and identifying a layout of the target form, and a form layout search process for specifying a form type of the target form from the layout identification result are performed, and the form is identified from the specified form type. Based on the field position information of a predetermined location of the target form, based on the field position information and the direction and the skew amount detected in the layout identification processing, the area including the field image considering the skew amount is stored. Cut out from the scanned form image and correct only the area corresponding to the direction and skew amount Performed, and to perform the image cutting-out processing for cutting out the field image of the predetermined location from the region after the correction.

According to a second aspect of the present invention, in the image extracting method according to the first aspect, in the image extracting process, when the skew amount is equal to or less than a predetermined value, the area is directly extracted as a field image of the predetermined location. Like that. According to the first and second aspects of the invention, since the image clipping method is configured as described above, the result of identifying the direction of the form, the skew amount, and the layout is obtained by the layout identification processing. The form type is specified in the form layout search process based on the layout identification result. Based on the form type, field position information of a predetermined place of the form is obtained by the image cutout processing. In the image clipping process, a region including a field image is further clipped based on the field position information, the direction of the form, and the skew amount, and only the region is corrected according to the direction and the skew amount. That is, correction is not performed on the entire form image. Then, a field image is cut out from the corrected area.

According to a third aspect of the present invention, there is provided an image extracting apparatus for extracting a field image of a predetermined location in a target form from a form image of the input target form in a skew-corrected state as follows. That is, image input means for inputting the form image composed of image data and storing the form image in the image data memory, searching for the form image stored in the image data memory, and determining the direction and the skew amount of the form image. A layout identifying means for detecting and identifying the layout of the form image and outputting key information as an identification result; a first database in which form types and key information are registered in advance in correspondence with each other; A database search unit that searches the first database using the key information output from the server, specifies and outputs the form type of the target form, and the form type and the field position information for each form type. Searching the second database using the second database registered in advance and the specified form type,
The field position information in the form is obtained, and based on the direction and the skew amount detected by the layout identifying means and the field position information, the area including the field image of the predetermined location in consideration of the skew amount is stored. And a cut-out unit that cuts out a field image of the predetermined place from the corrected area by performing a correction corresponding to the direction and the skew amount only on the area.

According to a fourth aspect of the present invention, in the image extracting apparatus according to the third aspect, when the skew amount is equal to or less than a predetermined value, the extracting unit cuts the area as it is as a field image of the predetermined location. I have to. Third
According to the fourth aspect of the present invention, since the image cutout apparatus is configured as described above, the form image is input by the image input means, and the key information of the direction and skew amount of the form image and the layout identification result of the form image are input. Is obtained by the layout identifying means. The form type corresponding to the key information is obtained by the database search means. The image cutout means obtains field position information based on the form type obtained by the database search means, and includes a field image in which the skew amount is considered based on the direction and skew amount detected by the layout identification means and the field position information. An area is cut out from the form image, and correction corresponding to the direction and the skew amount is performed only on the area. That is, correction is not performed for the entire form image. A field image at the predetermined location is cut out from the corrected area by the image cutting means.

[0008]

FIG. 1 is a block diagram of an image extracting apparatus according to an embodiment of the present invention. This image cutout apparatus is, for example, a form image Im of a company's fee payment form in which a plurality of bar codes are described and a receipt is described.
And an image input unit 1 which is an image input device such as an image scanner or a digital camera or an image input means for inputting a form image Im from an image file. And an image data memory 2 for storing the image Im. The image data memory 2 includes a barcode search / recognition unit 3 serving as a layout identification unit via a bus B;
A bar code recognition result memory 4 is connected, and the bar code recognition result memory 4 has a company code database (company code DB) 5 as a first database for specifying the type of a form via a bus B; A form layout search unit 6 as a database search means and a form layout search result memory 7 are connected.

The barcode search / recognition unit 3 searches the form image Im stored in the image data memory 2 for a barcode position, and finds the form direction Dr and the skew amount S of the form.
q and the end point coordinates C of the form image Im are obtained and output, and a company code is obtained and output as key information Ky of the layout. The barcode recognition result memory 4 stores information output by the barcode search / recognition unit 3. In the company code database 5, a company code, a layout name of a business form, and a binarization threshold set for each company are registered in association with each other. The form layout search unit 6 has a function of searching the company code database 5 based on the identification result stored in the barcode recognition result memory 4 and extracting a form layout name. The form layout search result memory 7
The layout name of the form extracted by the form layout search unit 6 is stored.

The bus B further includes a form layout database (form layout D) as a second database.
B) 8, an image extracting unit 9 as an image extracting unit, a recognition result memory 10, and a control unit 11 are connected. In the form layout database 8, a form layout name and field position information F in the form are registered in association with each other. The image cutout unit 9 searches the form layout database 8 using the form layout name to obtain the target field position information F of the form, and cuts out only the image of the target field in a state where the skew correction is performed by processing described later. Has a function. The recognition result memory 10 stores an image of the extracted field. The control unit 11 controls the entire apparatus.

FIG. 2 is a flowchart showing an outline of an image extracting method performed by the apparatus shown in FIG. 1, and FIG. 3 is a flowchart showing details of an image extracting process S40 in FIG. FIG. 4 is a diagram illustrating a configuration example of a form. With reference to these drawings, the operation of each part of the image clipping apparatus and the image clipping method will be described. In this image cutout apparatus, as shown in FIG. 2, an image input process S10, a barcode search / recognition process S20 as a layout identification process, and a form layout search process S30 are performed, and the processes S41 to S48 in FIG. Is performed, and an image of a receipt portion described in a predetermined place of the form is extracted. Here, a configuration example of the target form will be described. The target forms include JAN (Japanese Artic
le Number) code 20 is printed in three or four rows, and the field of the receipt 22 on which the receipt 21 is stamped (or printed) is arranged at a predetermined place.

In the image extracting apparatus, when a form image Im of a target form composed of multi-value image data is input to the image input unit 1 in the first image input processing S10 of FIG. , Form image Im if necessary
Is performed, and the image Im is stored in the image data memory 2. In the barcode search / recognition process S20, the barcode search / recognition unit 3 searches the layout of the image Im stored in the image data memory 2, and determines the position of the barcode 20, the direction Dr of the form in the form image Im, and the like. Form skew amount Sq and form image I
and the end point coordinates C of the m.
To be stored. Further, the barcode search / recognition unit 3 recognizes the barcode 20 to obtain a company code, and stores it in the barcode recognition result memory 4 as layout key information Ky.

In a form layout search process S30 after the barcode search / recognition process S20, the form layout search unit 6 reads the key information Ky from the barcode recognition result memory 4, and uses the key information Ky to read the company code database 5 with the key information Ky. A form layout name N for specifying a form type by searching and a binarization threshold Th are extracted. The form layout name N and the binarization threshold Th are stored in the form layout search result memory 7. In the image cutout processing S40, the image cutout unit 9 reads out the form layout name N and the binarization threshold Th from the form layout search result memory 7, and searches the form layout database 8 using the layout name N.

FIG. 5 is a diagram showing an outline of the form layout database 8. The form layout database 8 stores a form layout name N and field position information F in each form as shown in FIG.
The position information F of the receipt is obtained by searching the form layout database 8. The image cutout unit 9 further reads out the form image Im from the image data memory 2 and also reads out the form direction Dr, the skew amount Sq, and the form end coordinate C from the barcode recognition result memory 4, and outputs the target field as follows. Cut out the image. In steps S41 and S42 of FIG.
The form image Im is binarized by the binarization threshold Th. In this manner, by performing binarization according to the form type, an appropriate image can be obtained even for forms having different print colors, for example.

FIG. 6 is a diagram showing the coordinates of the form image Im, and FIGS. 7A and 7B are diagrams showing the calculation of the position of the receipt 22. FIG. In the processing S43 after the binarization, the image cutout unit 9 calculates the center coordinates O of the form from the form end coordinates C. In the process S44 after the calculation of the center coordinates O, the image cutout unit 9 starts from the center coordinates O based on the field position information F and the skew amount Sq as shown in FIG. The position of the end 23 is calculated. Here, when the direction Dr of the form is opposite to the normal direction, the position of the end 23 of the receipt 22 is inverted from the center O as shown in FIG. .

FIGS. 8A and 8B are views showing the concept of region extraction. The image cutout unit 9 in which the calculation of the end portion 23 of the receipt 22 is completed performs the processing S45, and based on the size of the receipt 22 and the skew amount Sq, the receipt 22 is processed.
An area 24 composed of a portion including a periphery determined according to the skew amount Sq is cut out. The area 24 when the form direction is normal is as shown in FIG. 8A, and when the direction is different, it is as shown in FIG. 8B. When the area 22 is cut out, the image is cut out while being converted into a density suitable for the output target of the image cutting apparatus.

FIG. 9 is a diagram showing correction of an area. The image cutout unit 9 processes the rotation with respect to the region 24 (S46).
And the skew amount Sq is corrected. Further, the image cutout unit 9 rotates the area 24 as necessary to align the length and width of the receipt 22 in a desired direction.

FIG. 10 is a diagram showing extraction of a field image. In the processing S47 after the processing S46,
The image cutout unit 9 determines whether the skew amount Sq is within an allowable range, and if it is within the allowable range, stores the rotated area 24 as a field image 25 of the receipt 22 in the recognition result memory 10 as it is. . If the skew amount Sq exceeds the allowable range, the area of the region 24 is too large, so that in the process S48, the field image 2 of the receipt 22 is set so as not to hinder post-processing.
5 is cut out and stored in the recognition result memory 10. By a series of processing from the image input processing S10 to the image extraction processing S40, the receipt 2 for the extraction purpose is extracted from the form image Im.
2, the skew correction and the rotation correction are stored in the recognition result memory 10. The control unit 11 outputs the field image 25 stored in the recognition result memory 10 to an external device such as a printer.

As described above, in this image extracting apparatus,
The image cutout unit 9 is set to the form layout name N of the form type.
Is used to search the form layout database 8 to obtain field position information F in the form image Im. Based on the direction Dr, the skew amount Sq, and the field position information F, the area 24 including the field of the receipt 22 is converted to the form image. The image is cut out from Im, the correction is performed only on the area 24 in the rotation processing S46, and the field image 25 is cut out. Therefore, as compared with the case where the entire form image Im is corrected, the number of processes is reduced, and the time until the field image 25 is cut can be reduced. In addition, the skew amount Sq is determined in step S47, and the cutout in step S48 is omitted depending on the skew amount Sq, so that the processing speed can be further reduced.

Note that the present invention is not limited to the above embodiment, and various modifications are possible. For example, in the above-described embodiment, the form and the company code of the form are determined from the position and the like of the barcode 20 on the form on which the barcode 20 is printed. This is applicable as long as a mark or the like that can detect the direction is attached to the mark or a ruled line or sheet identification symbol that can identify the layout is attached to the mark.

[0021]

As described above in detail, according to the first and second aspects, the image input processing, the layout identification processing, and the form layout search processing are performed to specify the form type of the target form. In the image cutout process, the entire form image is not corrected, but an area including a field at a predetermined location of the target form is cut out, and only the area is corrected according to the direction and skew amount of the form. The field image of the place is cut out.
Therefore, the processing until the target field image is output can be reduced, and the processing time can be shortened accordingly.

According to the third and fourth aspects, the image input means, the layout identification means, the database search means, the first and second databases, and the image cutout means are provided. A configuration in which a field including a field at a predetermined location of a target form is cut out without performing correction on the entire image, a correction corresponding to the direction and skew amount of the form is performed only on the area, and a field image at a predetermined location is cut out. I have to. Therefore, the processing until the target field image is output can be reduced, and the processing time can be shortened accordingly.

[Brief description of the drawings]

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

FIG. 2 is a flowchart showing an outline of an image clipping method performed by the apparatus of FIG.

FIG. 3 is a flowchart showing details of an image cutting process S40 in FIG. 2;

FIG. 4 is a diagram illustrating a configuration example of a form.

FIG. 5 is a diagram showing an outline of a form layout database 8;

FIG. 6 is a diagram showing coordinates of a form image.

FIG. 7 is a diagram showing position calculation of a receipt.

FIG. 8 is a diagram showing a concept of cutting out an area.

FIG. 9 is a diagram illustrating correction of an area.

FIG. 10 is a diagram illustrating extraction of a field image.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Image input part 2 Image data memory 3 Barcode search / recognition part 4 Barcode recognition result memory 5 Company code database 6 Form layout search part 7 Form layout search result memory 8 Form layout database 9 Image extraction part 10 Recognition result memory 20 Bar Code 22 Receipt 24 Area 25 Field image S10 Image input processing S20 Barcode search / recognition processing S30 Form layout search processing S40 Image extraction processing

Claims (4)

[Claims] 1. From a form image of an input target form,
An image extracting method for extracting a field image at a predetermined location in a target form in a skew-corrected state, comprising: inputting the form image composed of image data and storing the form image in an image data memory; A layout identification process for detecting the direction and skew amount of the target form by searching the form image stored in the target form and identifying a layout of the target form; and determining a form type of the target form from the layout identification result. A form layout search process to be specified is performed, and field position information of a predetermined place of the target form is obtained from the specified form type, and the field position information and the direction and skew detected in the layout identification process are obtained. Based on the skew amount. A region corresponding to the direction and skew amount is cut out only from the stored form image, and a field image at the predetermined location is cut out from the corrected region. An image cutting method characterized by the above-mentioned. 2. The image cutting method according to claim 1, wherein in the image cutting processing, when the skew amount is equal to or smaller than a predetermined value, the area is cut as it is as a field image of the predetermined location. 3. From the input form image of the target form,
An image cutout device for cutting out a field image at a predetermined location in the target form in a state where the skew is corrected; an image input means for inputting the form image composed of image data and storing the form image in an image data memory; Layout identifying means for searching the form image stored in the form image, detecting the direction and skew amount of the form image, identifying the layout of the form image, and outputting key information of the identification result, A first database in which key information is registered in advance in correspondence with the first database, using the key information output from the layout identifying means, searching for the first database, specifying a form type of the target form; The database search means to be output, and the form type and field position information for each form type are stored in advance. A second database registered accordingly, and a search of the second database using the specified form type, obtaining the field position information in the form, the direction detected by the layout identifying means, Based on the skew amount and the field position information, an area including the field image at the predetermined location in consideration of the skew amount is cut out from the stored form image, and only the area corresponds to the direction and the skew amount. An image extracting device for performing a correction and extracting a field image at the predetermined location from the area after the correction. 4. The image clipping device according to claim 3, wherein the image clipping means is configured to clip the area as it is as a field image of the predetermined location when the skew amount is equal to or less than a predetermined value. apparatus.