JP2004185129A - Image processor and image processing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processor capable of performing image reading processing to an attached sheet in addition to a main paper in a ledger sheet of specific specification, for example, a receipt with a continued paper. <P>SOLUTION: In this image processor for performing the reading processing of the ledger sheet of specific specification such as the receipt with the continuous paper, an image processing part 100 separately detects the respective image areas of the main paper and the continued paper, and cuts out them on the basis of the edge detection result for an input image. A recognition part 120 executes character recognition processing on each remarks column of the main paper and the continued paper. <P>COPYRIGHT: (C)2004,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates generally to the field of image processing devices, and more particularly, to an image detection technique for detecting an image on a sheet to which an attached sheet is added.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, an image processing apparatus such as an optical character reader (OCR) scans a sheet such as a form with a scanner and inputs an image (a pattern such as a character or a figure) recorded on the sheet.
[0003]
By the way, in OCR, various forms having different sizes and formats are handled as reading targets. In order to read such various forms, it is particularly important to use an image detection technique for detecting and cutting out an image area recorded on the form.
[0004]
Conventionally, a technology for identifying and recognizing various forms has been provided (for example, see Patent Document 1). In addition, a prior art for processing a free format form has also been proposed (for example, see Patent Document 2). In this prior art, a horizontal segment is extracted by connecting a boundary image above or below a horizontal black run on a binary image in the horizontal direction. Similarly, the present invention relates to a technique related to frame recognition in which a boundary image on the left or right side of a vertical black run is vertically connected to extract a vertical segment.
[0005]
[Patent Document 1]
JP-A-2002-109468
[Patent Document 2]
Japanese Patent Publication No. 4-18351)
[0007]
[Problems to be solved by the invention]
Among the various forms described above, there is a form used in medical relations called, for example, a claim (medical fee statement). The receipt is usually A4 size paper in a predetermined format, and is a type of document that a medical institution submits to a municipal government or an insurance union in order to claim medical expenses. A system for reading this receipt by OCR has already been developed and used.
[0008]
By the way, the receipt may be handled in a form of special specification in which an attached sheet called a continuous sheet is added to the main sheet (A4 size sheet). Generally, the format of the receipt includes an item called a summary field. If this summary field cannot be filled in the entry field included in the A4 size book, it can be extended to a continued sheet glued to the book.
[0009]
Conventionally, when this summary field is read by OCR, it is limited to those included in a predetermined A4 size book. Therefore, it was impossible to perform the reading process by the OCR for the summary column on the continuation sheet, which is a sheet attached to the receipt.
[0010]
Therefore, an object of the present invention is to provide an image processing apparatus and an image processing method that can perform image reading processing on attached paper together with the actual paper in a form of special specification such as a receipt with a continuous paper added. It is in.
[0011]
[Means for Solving the Problems]
An aspect of the present invention is to provide an image detection method for distinguishing and detecting an image area on an attached sheet corresponding to a continuous sheet and the like and an image area on the book sheet in a special specification form such as a receipt with a continuous sheet added thereto. The present invention relates to an image processing apparatus provided with means.
[0012]
An image processing apparatus according to an aspect of the present invention includes an image input unit that scans a sheet with an attached sheet added to a book sheet and inputs an image, and includes a certain linear component from the image input by the image input unit. Edge detection means for detecting an edge, and image detection means for distinguishing and detecting an image area on the book paper and an image area on the accessory paper based on a detection result of the edge detection means. is there.
[0013]
With such a configuration, for example, since it is possible to detect an image area on an attached sheet such as a continuous sheet added to the book sheet of the receipt, together with, for example, a summary field recorded on the book sheet, Also, the reading process by the OCR becomes possible. Therefore, since the reading process by the OCR can be performed not only for the receipt of the book sheet but also for the receipt to which the continuous sheet is added, the efficiency of the receipt process can be improved. In short, it is possible to provide a useful image processing apparatus capable of performing an OCR reading process for a special specification form to which an attached sheet is added.
[0014]
BEST MODE FOR CARRYING OUT THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings.
[0015]
FIG. 1 is a block diagram illustrating a configuration of an image processing apparatus (here, OCR is assumed) according to the present embodiment.
[0016]
(Configuration of OCR)
As shown in FIG. 1, the present OCR is roughly composed of a scanner 20 and a computer 30 such as a personal computer. Normally, the scanner 20 scans each sheet conveyed from many sheets 10 and converts information (patterns such as characters and figures) recorded on the sheet into an image (image data). This is an image input device that converts and inputs the converted data to the computer 30.
[0017]
The computer 30 corresponds to an OCR main body, and includes a microprocessor (CPU) 100 that executes OCR software, and a storage device 200 such as a hard disk drive. In this embodiment, the functions of the OCR software executed by the CPU 100 will be described as an image processing unit 110 and a recognition unit 120 for convenience.
[0018]
The image processing unit 110 processes an image input from the scanner 20 and outputs image data necessary for the character recognition processing in the recognition unit 120. The recognition unit 120 performs a character recognition process on the image data obtained from the image processing unit 110 and outputs a recognition result 210 to the storage device 200. Further, the image processing unit 110 realizes an image detection function of detecting (cutting out) the image areas of the regular paper and the attached paper of the form 10 to be described later separately.
[0019]
Note that the computer 30 can transfer the recognition result output from the recognition unit 120 to an external system via a wired or wireless communication line.
[0020]
(Image processing method)
Hereinafter, the image processing method of the present embodiment will be described with reference to FIGS. 2 to 13 in addition to FIG.
[0021]
In the present embodiment, as shown in FIG. 2, the form 10 is assumed to be a receipt including a main sheet 50 and a continuous sheet 60 that is an attached sheet thereof. The book paper 50 usually has an A4 size format and has various recording items. In this item, there is provided a summary column (the central part on the right side of the book 50 shown in FIG. 2) in which detailed contents regarding treatment such as a drug name, a standard, a unit, and a quantity are entered. On the other hand, the continuation sheet 60 is normally glued to the book sheet 50 on the left end side, and corresponds to an extended portion of the summary field in which a name of a disease or the like that cannot be written in the summary field of the paper sheet is entered.
[0022]
Here, as shown in FIG. 2, it is specified that the continuous paper 60 is smaller in height (vertical size) by, for example, about 50 mm than the A4 size book paper 50 in the vertical direction. However, the length (horizontal size) of the continuous sheet 60 is arbitrary, but has a summary column area in column units as described later. In FIG. 2, the continuation sheet 60 is composed of two columns with a vertical ruled line as a boundary. FIG. 5 shows a case in which a receipt consisting of only the main sheet 50 as shown in FIG. 5A and a continuous sheet 60 of various sizes are added, including the case of only the continued sheet 60 as shown in FIG. A specific example of the receipt will be described.
[0023]
(Image detection processing)
Hereinafter, the procedure of the image detection processing when the reception reading processing is performed will be described mainly with reference to the flowchart of FIG.
[0024]
First, the receipt 20 is scanned by the scanner 20, and an image on the receipt is input (step S1). Here, as shown in FIG. 5, it is assumed that a black area (shaded area) corresponding to the base (bed) of the scanner 20 exists outside the receipt image to be read.
[0025]
Here, as shown in FIG. 5B or 5C, it is assumed that the main sheet 50 and the continuation sheet 60 are scanned at a time and the image input processing is performed. However, as shown in FIGS. 5D to 5H, the image input processing may be performed in a plurality of scans depending on the length of the continuous sheet 60 and the scanning range of the scanner 20. Further, as shown in FIG. 1I, in the case of the continuous sheet 60 that is not glued to the main sheet 50, it is naturally scanned alone.
[0026]
Next, the image processing unit 110 performs an edge detection process on the entire input image (step S2). The edge detection process detects each edge forming a rectangular area (image area) that is a planar shape of the form image. In the edge detection method, usually, scanning is performed from a reference line until a white pixel is detected, and a straight line (a linear function of coordinates x and y) connecting white pixel points detected at regular intervals is obtained. At this time, a line segment having the highest linearity is determined as an edge (upper / lower end and left / right end).
[0027]
In the edge detection method according to the present embodiment, for example, as shown in FIG. 6, in order to detect each portion (rectangular area) of the main sheet 50 and the continuous sheet 60, an edge detection process is performed by a two-step method. That is, the x-coordinates of the points to be plotted at the upper end and the lower end are set to be the same, and a position at which the distance between the upper end and the lower end is shorter than a certain value is detected. Then, the detection point is divided before and after that point, and two straight lines are obtained for the upper and lower lines, respectively, and are determined as the upper and lower edges of the receipt. Similarly, two right and left straight lines are obtained and determined as the left and right edges of the receipt. Here, taking into account the slope of the receptor, the edge of the receptor can be accurately determined by obtaining a straight line except for the first and last white pixel detection points and one before and after the detection point at which the distance between the upper and lower points changes. It is possible to decide.
[0028]
The image processing unit 110 detects a rectangular area based on the edge detection result as described above (Step S3). Further, the image processing unit 110 determines the size of the rectangular area by measuring the height or the like of the rectangular area (step S4). That is, if the size of the detected rectangular area is the A4 size, the image processing unit 110 determines that the sheet is the real paper 50 (YES in step S5). Therefore, in this case, the image processing unit 110 distinguishes an area other than the image area of the book sheet 50 as the image area of the continued sheet 60 (step S6). Specifically, it is in the form of a receipt as shown in FIGS.
[0029]
On the other hand, when the size of the detected rectangular area is not the A4 size, the image processing unit 110 determines that all of the input images are image areas of the continuous sheet 60 (NO in step S5, S7).
[0030]
Here, when the input image is composed of the image areas of the main sheet 50 and the continuous sheet 60, the image processing unit 110 converts the coordinates of the left end of the image area of the continuous sheet 60 into the image of the main sheet 50 as shown in FIG. A process for expanding the area to a predetermined range 70 (for example, about 10 mm) is executed. This is a process in which the left end portion (connection side portion) of the continuation sheet 60 is adhered to the main sheet 50, and thus the portion is considered.
[0031]
As described above, in the present embodiment, the two-step edge detection process using the size difference between the original paper 50 and the continuous paper 60 is performed on the receipt having the continuous paper 60 added to the original paper 50, Each image area of the continuation sheet 60 is detected separately. Accordingly, the image processing unit 110 can cut out the image areas of the main sheet 50 and the continuous sheet 60 constituting the receipt from the input image from the scanner 20.
[0032]
Therefore, the OCR recognition unit 120 performs a character recognition process on the image area (including the summary column) on the book 50 cut out by the image processing unit 110 and outputs the recognition result to the storage unit 200. In addition, the recognition unit 120 performs a character recognition process on a summary column, which is an image area on the continuous sheet 60 cut out by the image processing unit 110, and outputs the recognition result to the storage unit 200.
[0033]
(Image detection processing of continued paper)
As described above, the continued sheet 60 is an extension of the summary field of the book sheet 50, and is usually divided into area units called columns. Therefore, hereinafter, the image detection (cutout) processing in units of columns according to the present embodiment will be described mainly with reference to the flowchart in FIG. 4 and FIGS. 8, 9, 11, 12, and 13.
[0034]
First, as shown in FIG. 11, the range of each column (here, two columns) of the continuous sheet 60 is defined by horizontal ruled lines 600 at the upper and lower ends and vertical ruled lines 601 at the left and right ends. The image processing unit 110 performs a ruled line detection process on the image area of the continued sheet 60 (step S11). Based on the detection result of the horizontal ruled line 600 in the ruled line detection processing, the upper and lower ends of the summary field are set (YES in step S12, S13).
[0035]
Further, the image processing unit 110 sets the left and right ends of the summary column based on the detection result of the vertical ruled line 601 (YES in step S15). Then, the image processing unit 110 divides the summary column into columns based on the set upper and lower ends and the left and right ends (step S16). The image processing unit 110 selects only columns having a certain width or more from the divided columns as columns constituting the summary column (step S18). Thereby, for example, as shown in FIG. 13, a column whose left end portion has not been scanned can be removed.
[0036]
Here, in the above-described vertical ruled line detection processing, the image processing unit 110 converts the plurality of divided ruled lines 80a to 80c into one ruled line due to the gluing operation of the continuous sheet 60, as shown in FIG. Execute the integration process. In the horizontal ruled line detection process, similarly, a process of integrating a plurality of divided ruled lines as one ruled line is executed.
[0037]
When the horizontal ruled line 600 cannot be detected in the ruled line detection process, the image processing unit 110 normally determines all of the image area of the continuous sheet 60 from the top to the bottom as the summary column (NO in step S12). ). However, as shown in FIG. 12, some of the continuous sheets 60 include a description different from the description in the summary column such as a medical institution code in the upper part and a blank 610 having a certain width. In this case, the image processing unit 110 sets the lower part of the blank 610 as the upper end of the summary column instead of the horizontal ruled line (step S14). In this case, the lower end of the summary field is set to be below the image area.
[0038]
Further, as shown in FIG. 12, when the vertical ruled line 601 does not exist in the continuous sheet 60, the image processing unit 110 detects a blank instead of the vertical ruled line 601 and executes a division process for each column. (No in step S15, S17). Specifically, as shown in FIG. 9, the image processing unit 110 detects a blank 90 included in a range larger than a predetermined width of one column, and sets the blank 90 as a column boundary. When the leftmost or rightmost vertical ruled line exists, the image processing unit 110 detects a blank 90 included in the range of one column based on the vertical ruled line. If the vertical ruled line does not exist, a blank 90 included in the range of one column is detected based on the left end or the right end of the image area.
[0039]
Here, the width of one column that divides the summary column of the continuous sheet 60 is usually about half the width of the A4 size area. After determining the upper and lower ends of the summary field as described above, the image processing unit 110 cuts out an image included in the range of the width of one column as an image in column units based on the left end or right end of the image area. Processing may be performed. According to such a method, even when the blank 90 cannot be detected, it is possible to detect and cut out an image in column units.
[0040]
As described above, according to the method of the present embodiment, it is possible to divide and cut out the image area of each column constituting the summary column from the image area of the continuous sheet 60 detected from the input image. Thus, the OCR recognizing unit 120 can execute the column-based character recognition process on the summary column on the continuous sheet 60 together with the summary column on the main sheet 50. Then, the OCR can output the recognition result of the summary field included in both the main sheet 50 and the continued sheet 60 to the storage unit 200.
[0041]
(Modified example of column division processing)
FIG. 10 is a diagram showing a modified example of the column division processing for dividing the image area of the column constituting the summary field from the image area of the continuous sheet 60. The present modified example is a processing method in which a dynamic programming (DP) matching method is applied to divide a summary column on the continuous sheet 60 into columns.
[0042]
In this method, as shown in FIG. 10, a cost is set between the ruled lines, and the cost is associated with the distance between the ruled lines. Then, the smallest cost is set between the ruled lines for one column, and the cost is set in consideration of the cost between the ruled lines at intervals of two columns and three columns. According to this method, when a ruled line other than a ruled line such as noise is detected as a ruled line, a column dividing process not affected by the ruled line can be performed.
[0043]
【The invention's effect】
As described in detail above, according to the present invention, for example, in a special specification form such as a receipt to which an attached sheet such as a continuous sheet is added, the image areas of the original sheet and the attached sheet are separately detected and cut out. be able to. Therefore, the image reading process can be performed on the attached paper together with the real paper. In particular, a form with a special specification such as a receipt to which a continuation sheet is added can be collectively read by OCR or the like, which is useful from a practical viewpoint.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a configuration of an OCR according to an embodiment of the present invention.
FIG. 2 is an exemplary view showing a specific example of a receipt according to the embodiment;
FIG. 3 is an exemplary flowchart for explaining the procedure of an image detection process according to the embodiment;
FIG. 4 is an exemplary flowchart for explaining the procedure of a continuous sheet image cutting process according to the embodiment;
FIG. 5 is an exemplary view showing specific examples of various receipts according to the embodiment;
FIG. 6 is an exemplary view for explaining edge detection processing according to the embodiment;
FIG. 7 is an exemplary view for explaining extension processing of an image area of a continuous sheet according to the embodiment;
FIG. 8 is an exemplary view for explaining a continuous sheet column dividing process according to the embodiment;
FIG. 9 is a view for explaining the column division processing.
FIG. 10 is an exemplary view for explaining a modification of the column dividing process of the embodiment.
FIG. 11 is an exemplary view for explaining a specific example of a column on a continuation sheet according to the embodiment;
FIG. 12 is an exemplary view for explaining a continuous sheet column dividing process according to the embodiment;
FIG. 13 is an exemplary view for explaining a continuous sheet column dividing process according to the embodiment;
[Explanation of symbols]
10… Form (paper)
20 scanner 30 computer 50 book (recept)
60 ... continued paper (attached paper)
100 ... Microprocessor (CPU)
110 image processing unit 120 recognition unit 200 storage device 600 horizontal ruled line 601 vertical ruled line

Claims (14)

Image input means for inputting an image by scanning a sheet on which an attached sheet is added to the main sheet;
Edge detection means for detecting an edge including a certain linear component from the image input by the image input means,
An image processing apparatus, comprising: an image detecting unit that distinguishes and detects an image region on the book and an image region on the attached paper based on a detection result of the edge detecting unit. The image detecting means includes means for detecting an image area on the main paper and an image area on the auxiliary paper based on a size difference between the main paper and the attached paper, and cutting out each of the image areas. The image processing apparatus according to claim 1, wherein: 2. The image processing apparatus according to claim 1, further comprising an image dividing unit that divides the image on the attached paper detected by the image detecting unit into a predetermined recognition target area. 4. The image processing apparatus according to claim 3, wherein the image dividing unit divides the image into a predetermined recognition target area based on an image corresponding to a horizontal ruled line or a vertical ruled line recorded on the attached paper. 4. The image according to claim 3, wherein the image dividing unit divides the image into a predetermined recognition target area based on an image corresponding to a horizontal ruled line or a vertical ruled line recorded on the attached paper and a blank area. Processing equipment. 2. The image detecting device according to claim 1, further comprising a unit that, when detecting an image area on the attached paper, expands and cuts the image area to a predetermined range corresponding to a connection portion of the book. An image processing apparatus according to claim 1. The book is a predetermined size receipt,
4. The attached sheet is a continuation sheet for expanding a summary field recorded on the receipt, and the summary field can be divided in a predetermined column unit. The image processing apparatus according to claim 1. An image processing method applied to an image processing apparatus that processes an input image by scanning a sheet to which an attached sheet is added to a main sheet,
An edge detection step of detecting an edge including a certain linear component from the input image,
An image processing method, comprising: an image detecting step of distinguishing and detecting an image area on the main paper and an image area on the attached paper based on a detection result by the edge detecting step. The image detecting step detects an image area on the main paper and an image area on the auxiliary paper based on a size difference between the main paper and the attached paper, and cuts out each image area. The image processing method according to claim 8, wherein: The image processing method according to claim 8, further comprising an image dividing step of dividing an image on the attached paper detected by the image unit into a predetermined recognition target area. The image processing method according to claim 10, wherein the image dividing step divides the image into a predetermined recognition target area based on an image corresponding to a horizontal ruled line or a vertical ruled line recorded on the attached paper. The image according to claim 10, wherein the image dividing step divides the image into a predetermined recognition target area based on an image corresponding to a horizontal ruled line or a vertical ruled line recorded on the attached paper and a blank area. Processing method. 9. The image detecting method according to claim 8, wherein, when detecting the image area on the attached paper, the image detecting step cuts out the image area by expanding the image area to a predetermined range corresponding to a connection portion of the book. Image processing method. The book is a predetermined size receipt,
11. The attached sheet is a continuation sheet for expanding a summary field recorded on the receipt, and the summary field can be divided in a predetermined column unit. 2. The image processing method according to claim 1.

Images