JPS61100873A - Optical character/graphic reading device - Google Patents

Abstract

PURPOSE:To prepare data for OCR in business form design image by writing a character reading frame and a graphic inputting frame matched to an actual business form in a spacing business form of drop-out color printing, and making the OCR read the form. CONSTITUTION:Light signals read by an optical section are converted to binary data by a photoelectric converting section 61 and stored in a picture memory GM4 through a preprocessing section 5. A hand-written character inputting frame, a printing type character inputting frame and a picture inputting frame are discriminated, and characteristic parameter of respective frame is stored in the same work memory WM3. Characters written in respective inputting frame are recognized, and the characteristic parameter is stored in the work memory 3 correlating them with respective character frame information. A relevant business form format table is prepared basing on various data stored in the work memory 3, and stored in a file memory FM8.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to an optical character/figure reading device, and in particular, uses a spacing form to read form data written on the side of a form design image to read optical characters. /Relates to a method for creating a form format for a graphic reader (OCR).

In conventional optical character/figure reading devices, the input method for manually creating form formats has been thoroughly generalized, so it can easily handle forms in a variety of formats. Although it had advantages, there were many entries and optical character/graphic reader (hereinafter referred to as OCR)
There is a problem that the amount of processing increases, the memory capacity increases, and the cost increases due to the increase in program development costs. A method for inputting takes to create a format has been awaited.

[Conventional technology]

The configuration diagram of a general OCR is shown in Figure 3, where 1 is a microprocessor (MPtl), 2 is a program memory (PM), 3 is a work memory (WM),
4 is image memo U (GM), 5 is preprocessing IL,
61 is a photoelectric conversion section, 6 is an optical section, 7 is a mechanism section, and 71 is a mechanism control section.

8 is a file memory (FM), 9 is an external interface section, 10 is a display device (DISP), 11
For example, it is a keyboard.

In the conventional method, ■When creating a form format for OCR, while looking at the screen displayed on the display device (DISP) 10, the required data is entered on the keyboard 1 using so-called conversation processing.
Starting from scratch, a form format for the ocR was created based on the input data.

Specifically, for example, by operating the key fob 11 provided in this OCR, the microprocessor (MPU)
l is stored in program memory (PM) 2,
When the format analysis program 21 is executed, first O
The screen format definition data is read from the file memory (FM) 8 in the CR, and the display device (DISP) 10 displays a dialog processing screen for creating a form format for OCR, so the operator can: While looking at the screen, the user sequentially enters the required data from the keyboard 11 in a conversational manner.

The input data is stored in the format creation program 2.
3, and created a 7-ormat document in the form of a table for OCR.

■Another manual method when creating a form formatter for OCR is to change the size of the above form by specifying the left end position of the 1 character field, 1 character interval, character type, etc. in a specific format input form (format system: lU). OC sheet) is filled in manually, and the form is read by OCR.
I was creating a form format for R.

Specifically, when the mechanism control program 24 is started by the format analysis program 21 and the format control sheet is read by the OCR, each data on the sheet is recognized by the character recognition and recognition program 22, and the format is created. By program 23, OC
I was creating a form format for R in the form of a table.

[Problem that the invention seeks to solve]

Therefore, in the conventional method, conversation processing, Otj? Regardless of the input processing method, there is a problem that it is not possible to obtain an image of the form that will actually be used, and because numerical parameters are entered manually on the processing form, precision of dimensions etc. is required. For Baranota, there is a problem in that if an entry error occurs, it is necessary to redo a series of input processes.

In view of the above-mentioned conventional drawbacks, the present invention aims to provide a method for creating form format data for OCR by directly reading and recognizing data written in a form design image. .

[Means for solving problems]

This purpose is to provide a spacing form printed in dropout color with a form that is filled in according to the filling rules with an image of the form that will actually be used, and a means for directly recognizing characters and figures on the form. This is achieved by the optical character/graphic reader (OCR) of the present invention, which is equipped with means for creating a form format for an optical character/graphic reader (OCR) for a large number of forms to be processed from the recognized spacing form. Ru.

[Effect]

That is, according to the present invention, a spacing form printed in dropout color is used in the format input form, and the spacing form is printed in accordance with the actual form.
Fill in the character reading frame 1 figure, write the processing method in a part of each text, have the OCR read it, and the OCR detects each frame data to calculate the reading position and Since the reading method is determined from the characters written in the frame and the form is formatted for OCR, the form design image filled in the spacing form is used to create the data for the form format for OCR. The present invention has the advantage of providing a data input method for creating a form format for OCR that is simple, easy to use, and has a good man-machine interface.

〔Example] Embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a diagram explaining the concept of creating a form format for OCR by implementing the present invention, and FIG.
OCR program memory (P) 1) explained in Figure 3
2 is a diagram showing an example of each processing program necessary to implement the present invention, which is stored in the computer.

In Figure 1, (a) shows a spacing form printed in trombe-out color, where the vertical grid spacing is, for example, 1.02 mm, and the horizontal grid spacing is, for example, 0.635 mm. Even when a form is read using OCR, the grid is not read.

(b> indicates the entry rules when designing a form for OCR on the above spacing form.
For each type frame 2 handwriting frame 2 image frame indicated by ``frame attribute'', ``vertical/horizontal grid/number of gutter'', etc. are shown.

An example of filling in according to the above entry rules is (
c), ■ indicates an example of a handwritten character input frame,
■ indicates an example of an image input frame, and ■ indicates an example of a manually input frame for printed characters.

(d) shows an example of processing the characters (1 to 1) written in the handwriting frame ■ shown in (c), for example, when processing a large number of forms input to OCR at the operational stage. Since the processing method is shown, each character is stored in the file memory (FM) 8 as a form former 71-table data for OCR in correspondence with each character frame.

Next, a specific method of processing the spacing form will be explained with reference to FIG.

First, the temple structure control program 25 is activated by the keyboard 'll of this OCR, and the mechanism control section 71 controls the mechanism section 7 to input the spacing form into the optical section 6.

At this time, the optical signal read by the optical section 6 is converted into binary data by the photoelectric conversion section 61, and stored in the image memory (GM) 4 via the preprocessing section 5.

When the storage of the spacing form data in the image memory (GM) 4 is completed, the enlarged frame detection program 26 is started, and various frame data written on the spacing form is read and stored in the work memory (GM). )3.

In this case, "line cutting" is first performed. Specifically, the binary information of the above-mentioned spacing form is read out from the image memory (GM) 4 on a line-by-line basis, and it is determined that there is a continuous black 1, that is, 1' in the pixel data constituting the line. Detect to find the first horizontal frame.

By repeating the same operation and detecting the next horizontal frame, "line cutting" for a certain character input frame is completed.

Thereafter, "line cutting" of a plurality of character input frames and one image input frame is performed in the same manner.

Next, for the frame that has been subjected to the above-mentioned "row cutting", "column cutting" is performed within the range where the row frame exists, and if a continuous black pixel string is detected for each column, multiple character frames can be extracted. The "line cutting" information and "column cutting" information for each character frame are stored in the work memory (Wl'l) 3.

Next, the character/figure recognition program 27 is started, and the "row cutting" and "column cutting" information stored in the work memory (WM) 3 and the table shown in FIG. From the frame rule information explained in (1)), the characteristics of each frame are recognized, the aforementioned handwritten character input frame, printed character input frame, and image input frame are identified, and the characteristic parameters of each frame are determined by Stored in the same work memory (WM) 3.

In the same way, for example, characters written in a handwritten character input box are recognized, and their feature parameters are stored in the work memo IJ (WM) 3 in association with each character box information. The recognized character indicates the method of processing the data within the frame, as explained in FIG.

At this stage, the form format creation program 28 for OCR is activated, creates the form format table based on the various data stored in the work memory 1) 3, and creates the form format table in the file memory (F) 1). Store in 8.

As described above, the present invention is characterized in that a form format file necessary for reading the form during operation can be created directly based on the form image written on the spacing form.

During operation, in the same manner as above, the form format table stored in the file memory (FM) 8 is referred to from the form data read into the image memory (GM) 4 using a known method. Then, the data entered in each frame of the form is recognized and sent to the host computer through the external interface section 9.

〔Effect of the invention〕

As described above in detail, the optical character/figure reading device of the present invention uses a spacing form printed in drop-out color on a formatted form,
In the spacing form, fill in the character reading frame 1 figure input frame according to the actual form, write the processing method in a part of each frame, have the OCR read it, and OCR will read each frame data. By detecting it, the reading position is calculated, and the reading method is determined from the characters written in each frame.
Since it is designed to be a form format for OCR, data in the form format for OCR can be created based on the form design image entered in the spacing form, which is easy and has a good man-machine interface. ,0(,H) has the effect of providing a data input method for creating a form format for H.

[Brief explanation of the drawing]

FIG. 1 is a diagram illustrating the concept of creating a form format for OCR by implementing the present invention. FIG. 2 is a diagram showing each processing program necessary to implement the present invention. Figure 3 shows a block diagram of an example of the configuration of an optical character/graphic reader (OCR). In the drawing, l represents a microprocessor (MPU). 2 is a program memory (PM). 3 is work memory (WM), 4 is image memory (GM)
. 5 is a preprocessing section, and 6 is an optical section. 61 is a photoelectric conversion section, and 7 is a mechanism section. 71 is a mechanism control unit, 8 file memory (FM>
. 9 External interface section. 10 is a display device (DISP). 11 is the keyboard. 21 to 28 are various processing programs. are shown respectively. Ear 1 (2) fα)

Claims (1)

[Claims] A spacing form printed in dropout color is provided with a form filled in with the image of the form actually used according to the entry rules, and a means for directly recognizing the characters and figures on the form, and the recognized An optical character/graphic reading device comprising means for creating a form format for an optical character/graphic reading device (OCR) for a large number of forms to be processed from the spacing form.