JPS62222382A - Optical character reading system - Google Patents

Abstract

PURPOSE:To satisfactorily segment a character pattern by controlling information on a read area according to information detecting a read area control mark and newly setting a read start position, a read end position and a stage constitution at every slip so as to read them. CONSTITUTION:An image including a character in the slip 1 is photoelectrically converted by a scan/photoelectric conversion part 2, and stored in an image buffer memory 3. The information detecting the read area control mark controls the information on the read area set by a read area setting part 9, and the character pattern in the read area, which is specified by two pieces of information, is segmented. The segmented character pattern is given to an identifying part 6 through a preprocessing part 5, identified, then stored in a storage part 7.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an optical character reading method that allows the reading area on a form to be changed and set for each form by writing a small number of reading area control marks.

(Prior art) Conventionally, the reading area on a form has been set by specifying the distance from the reference side (top side, right side, etc.) of the form using a keyboard shade, or by writing the reading area. There have been methods such as having the device read the paper and setting based on the reading results, or writing marks in all the reading areas on the reading form and setting based on the positional relationship of the written marks.

For example, as shown in FIG. 2, characters ei'i relatives 101, 102, 103 . In the case of a three-column document, the width is the distance a, C from the right edge of form 1 to the reading start position of each column.

e, the length of each row of rows, d, and f are stored in advance in the device,
Regarding the reading line position, there is a method (Method 1) of specifying the distance Q from the top side of the form 1 to the reading start line position, the line pitch hS, and the number of line reading repetitions i (Method 1), or as shown in Fig. 3 (Method 1).
Regarding the reading line position, read area control marks such as a reading start line mark SLM and a reading end line mark ELM are written.
How to set by detecting this (method 2)
, as shown in FIG. 4, the width of the reading area is specified using reading area control marks as shown in FIG. (Method 3) etc.

On the other hand, existing printed documents have various printing formats, and the reading line position, reading start position and end position, and column structure may vary.

Furthermore, when a bound book is cut out and read continuously, even if the printing style is uniform, the reading start position relative to the right edge of the form may vary depending on the way the book is cut or the front and back of the form. To address these issues,
You can set reading area designation information (format information) individually for each document with a different format, or you can write a reading start position mark and a reading end position mark on every document, or you can use a fixed Once the formatting has finished reading and reading, unnecessary parts had to be deleted.

(Problems that the invention attempts to solve) However, all of the above methods require the work of marking marks as pre-processing (Method 3), the work of creating multiple formats (Method 1 and Method 2), and the work of writing marks as post-processing. This necessitated work to delete unnecessary parts (Method 1 and Method 2), leading to a decrease in work efficiency.

(Objective of the Invention) The object of the present invention is to solve these drawbacks, and for a document with a variable reading area, to determine the reading area from the area set on the device side and the area specified by the mark written on the document. The object of the present invention is to provide an optical character reading method that cuts out a character pattern based on the information set.

(Means for Solving the Problems) In order to achieve the above object, the present invention provides an optical character reading device that reads characters by photoelectrically converting an image including characters on a recording medium. an image buffer memory for storing a read image stored in the image buffer memory; a reading area setting means for specifying a reading area for the image stored in the image buffer memory; and detection of a reading area control mark for controlling the set reading area. means, a cutting means for cutting out a character pattern in the image recorded in the image buffer memory, an identification means for identifying the cut out character pattern, and a storage means for storing the identification result by the identification means, The information on the reading area set by the reading area setting means is controlled by the detection information of the reading area control mark, and the character pattern within the reading area designated by both information is cut out and identified.

(Function) According to the present invention, the information on the reading area set by the reading area setting means is controlled by the detection information of the reading area control mark, and the reading start position, reading end position, and film configuration are updated for each form. settings and perform reading. .

(Embodiment) Fig. 1 is a block diagram of a device that realizes the method of the present invention, in which 2 is a scanning/photoelectric exchange unit, 3 is an image buffer memory, and 4
5 is a reading area control mark detection section, 5 is a preprocessing section, 6 is an identification section, 7 is a storage section, 8 is a correction/editing section, and 9 is a reading area setting section. FIG. 5 shows a case where the form 1 has three columns, and the document has two columns 101, 102. This figure shows an example of writing a reading area control mark on a form for specifying a reading area when the number of columns that are not to be read is one column 104. FIG. 6 (aHb) shows an example of setting fixed format information (2) in the reading area setting section 9 and an example of creating temporary format information (2) by detecting a reading area control mark. Also, the seventh
The figure shows the position occupied by these reading area controls in the character reading operation and the flow of processing. In FIG. 5, SEM indicated by vertical ruled marks, J3, and EM indicated by horizontal ruled marks are respectively the first row reading U11 double position mark and the reading area control that specifies skipping by regarding the reading area as invalid. It's a mark. Also, SLM, E
LM is a reading start line mark and a reading end line mark.

The characters printed on the form 1 and the reading area setting marks entered manually are scanned by the scanning/photoelectric conversion unit 2 based on the fixed format information ■ of the reading area setting unit 9 and converted into electrical signals. It is recorded in the image buffer Ameshiri 3. The reading area control mark detection section 4 scans the reading area designated by the format information (2), detects vertical lines and sets them as reading start marks SEM, and detects horizontal lines and marks them as reading area invalid marks EM. That is, the upper end of the image data is scanned from the left end, and when white information changes to black information and the black information continues vertically longer than a certain threshold value, it is detected as a mark and set as a reading start mark SEM,
Even when black information continues horizontally longer than a certain threshold value, it is detected as a mark and is set as a reading area invalid mark ME.

A new reading area is set based on the detected reading start mark SEM and the format information ■ of the reading area setting section 9,
Control the reading area as temporary format information ■. Further, in the temporary format information (2), information to be controlled by a reading fE4 area invalid mark ME that instructs whether or not to treat the newly set reading area as a character area and perform reading is also added.

That is, the range of horizontal ruled lines from the start point MES to the end point MEE where the reading area invalidation mark ME is written, or that column, is skipped from reading the next line and processed, or only that part is read.

The vertical direction of form 1, that is, the line reading range, is the same as before, with a reading start line mark SLM and a reading end line mark EL.
This is done by detecting M. In the preprocessing unit 5, the area necessary for reading and the character pattern are cut out based on the newly set temporary format information ■, and the area and character pattern necessary for reading are cut out, and
sent to. The identification section 6 identifies the cut out character pattern and records the result in the storage section 7.

In the three-column example shown in FIG. 5, line detection nine characters are extracted in the order from the second line of the first column to the end line, and from the start line to the end line of the second column, and identified by the identification unit 6. The results are recorded in the storage section 7. However, the third row is edited into a continuous sentence by processing it according to the instructions of the format information without performing character extraction processing.

The reading results obtained in this manner are finally sent to the correction/editing section 8.

In the above example, an apparatus that imports data into the image buffer memory 3 page by page is shown, but similar processing can be performed with an apparatus that imports data into the image buffer memory 3 line by row. In that case, the starting rows of the first to third stages are first cut out by the scanning/photoelectric conversion section 2 (or the preprocessing section 4). The cut out character pattern is identified by the identification section 6 and output to the storage section 7. At this time, column separation information is added between the identification results of the first and second columns and between the second and third columns based on the format information received from the reading area control mark detection section 4, and the storage section 7 recorded in Next, column break information is detected from the recorded identification results, and information about the number of columns is used to automatically re-edit into a continuous sentence.

Note that in the present invention, vertical ruled marks SEM, horizontal ruled marks EM,
Although we have shown an example in which the start line mark SLM and the end line mark ELM are used independently as reading area control marks,
If there are charts or tables in the middle of columns or rows in a form and you want to skip them, set a part of the form, for example the top area of the form in Figure 5, as the format information specification area, and add more detailed lines to that area. You can control skipping by entering marks to control reading of columns or columns. It can also be applied to cases in which special information on paper, such as ruled lines, is specified and read automatically, and reading control is performed based on that information. Furthermore, as shown in Figure 5, the marking method is to use the reading area control mark 1:, M to specify the reading area, detect the start and end points of the mark, and read only that area. You can also have them do this. Even in this type of reading area control, when a document with no mark attached becomes the reading target during continuous reading, the area specified by the fixed format information is set as the reading target, or the area specified by the fixed format information is read, or the document read immediately before is read. They can also be read as being in the same format. By performing such control, it is also possible to further reduce the effort required to add marks.

(Effects of the Invention) As explained above, according to the present invention, format information is set for specifying a reading area on a form, and reading area control is performed for controlling a preset reading area. The reading area on a form can be changed and set for each form by writing a few reading area control marks, so it is possible to change or set the reading area for each form by writing a few reading area control marks. There is an advantage that reading operations can be performed continuously without changing the format even when there are mixed formats. Furthermore, by detecting the column structure, discontinuous sentences can be edited into continuous sentences.

[Brief explanation of drawings]

FIG. 1 is a block diagram of a device showing an embodiment of the present invention, FIGS. 2, 3, and 4 are explanatory diagrams of a reading area setting method in a conventional device. FIG. 5 is a block diagram of a reading area setting method in a conventional device. FIG. 6 is an example of writing the area control mark, FIG. 6 is an example of setting format information of the reading area setting section, and FIG. 7 is a diagram for explaining the positioning of reading area control in character reading operation. In the figure, 1 is a recording medium (form>, 101°102
．． 103 is a three-column text area, 104 is an advertising area, 2 is a scanning/photoelectric exchange unit, 3 is an image buffer memory, 4
is a reading area control mark detection unit, 5 is a preprocessing unit, 6 is an identification unit, 7 is a storage unit, 8 is a correction/editing unit, 9 is a reading area m setting unit, SEM is a reading start position mark, and SLM is a reading start The line mark ELM is a reading end line mark, and EM is a reading necessity control mark. Also, ■ and ■ are fixed format information and temporary format information, respectively.

Claims (1)

[Claims] An optical character reading device that reads characters by photoelectrically converting an image containing characters on a recording medium, an image buffer memory that stores an image photoelectrically converted by an optical scanning means, and an image stored in the image buffer memory. a reading area setting means for specifying a reading area for a reading area, a reading area control mark detecting means for controlling the set reading area, and a cutting means for cutting out a character pattern in the image recorded in the image buffer memory. and an identification means for identifying the cut-out character pattern, and a storage means for storing the identification result by the identification means, and the information on the reading area set by the reading area setting means is used as the detection information of the reading area control mark. An optical character reading method that is characterized by controlling the information of both parties and cutting out and identifying character patterns within a reading area designated by information from both.