JP2823048B2 - Batch processing of multiple formats in OCR using ADF - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

TECHNICAL FIELD The present invention relates to an ADF (Auto
Optical Character Reaper with OCR (Material Document Feeder)
The present invention relates to a der) system, and more particularly, to a processing method for collectively reading a plurality of types of originals using a kanji OCR.

[0002]

2. Description of the Related Art In a conventional kanji OCR with an ADF, only a single format document can be processed on the ADF at a time. That is, the traditional Kanji OCR
Is a system that reads a document in a standard format and outputs recognized character information to a file. Since a processing program is prepared for each standard format of the document, documents that can be processed at a time on the ADF are Limited to a single type.

[0003]

However, in actual operation, it is required that documents of a plurality of types of formats are bundled into a bundle and placed on the ADF, and the processes from reading to recognition and correction are collectively performed. Further, it is required that the number of each document is arbitrary and the order of each document in one bundle is also arbitrary. However, as mentioned above,
It has been difficult to pour a plurality of types of documents bundled from a bundle from the ADF and automatically select and execute different processing for each.

[0004] In view of such a need, a main problem of the present invention is to form a bundle of documents of a plurality of types into one bundle.
An object of the present invention is to provide a multiple format batch processing method which can be loaded on a DF and can collectively perform processes from reading to recognition and correction.

The present invention also provides a multiple format batch processing system capable of processing even if the number of documents in a plurality of formats is arbitrary and the order of each document in one bundle is also arbitrary. It is in.

[0006]

According to the present invention, in an OCR using an ADF, a partition paper is sandwiched between a plurality of types of documents of different formats, and the partition paper has a mark for defining a format of a subsequent document. A processing program corresponding to the different format is prepared in the OCR, and the OCR has a means for determining the partition sheet and a means for reading the mark of the partition sheet. By recognizing the format and starting the corresponding processing program,
A plurality of formats are collectively loaded from the ADF so that processing can be automatically selected and executed in accordance with each format.
Detection of multiple types of marks determined in advance according to
Area, and the format is the plurality of types of matrices.
The black detection area by painting it in black.
In the OCR, the plurality of types of mark detection areas are
Of which the number of black pixels in the area is the largest
This is a multi-format batch processing method in OCR using an ADF, wherein an upper existing area is determined as a designated area .

[0007]

[0008]

[0009]

A separating sheet functioning as a mark sheet for each format is sandwiched in a bundle of a plurality of types of documents. In the OCR, the separating sheet is determined, and the format of the read document is determined from the separating sheet. Means to mark a document type discrimination flag for each partition sheet,
A processing program corresponding to each format of a document to be read from the next page is selectively activated. The end of the document is recognized by the appearance of the next divider or the end of a bundle. Also, ADF
Clears the page number counter for each partition, and OCR searches for the page number for each document type. The determination of the partition paper and the document type is performed by comparing the definition of the partition paper stored in the storage unit of the OCR in advance with the image of the partition paper read from the scanner.

[0010]

BRIEF DESCRIPTION OF THE DRAWINGS FIG.
A preferred embodiment of the present invention for R will be described. In FIG. 1, the description will be made on the assumption that three types of documents A, B, and C have different formats. in this case,
At the top of each of the three types of documents, there is a divider 11, 12, 1
3 is attached. These partition sheets 11 to 13 function as mark sheets for defining three types of document formats. That is, the partition papers 11 to 13
At the same position, a mark detection area that specifies one of the three types of A, B, and C is set, and the format is specified by filling the mark detection area with black.

In the kanji OCR, processing programs corresponding to three types of formats are prepared, and the definition of the partition paper is stored in the storage unit in advance. Then, among the plurality of mark detection areas, the paper in which any of the black areas is painted black is determined to be a partition sheet, and the area having the largest number of black pixels included in the area and the number of black pixels being equal to or more than a predetermined number are present. The area to be processed is regarded as the specified area and is compared with the stored definition. For example, when it is determined that the document is the document A, the data of the read document is processed based on a processing program of a format corresponding to the document A. in this way,
Separators 11 to 1 on three different types of documents
3, the corresponding processing programs can be selected and executed.

In the present invention, documents of three different formats are placed on the ADF as a bundle with the separators 11 to 13 sandwiched at the top for each format, and read in a batch, and processed according to each format. Automatically select and execute a program.

Here, the program to be newly created is a "document discriminating program". The “document discrimination program” includes a “partition sheet discrimination program” that discriminates a partition sheet by detecting a mark detection area painted black. With this document discrimination program, a plurality of types of documents can be collectively read, and recognition and correction can be performed collectively. Therefore, ADF
It is not necessary to limit the types of documents to be loaded to one. When the number of types of documents increases, the format can be flexibly dealt with by additionally registering the format of the partition paper.

Referring to FIG. 1, a plurality of documents in a plurality of formats are bundled out of order and placed on the ADF, and the reading 15 is performed by a batch scanner. The kanji OCR has first determining means for determining a partition sheet, and second determining means for determining a format of a document read from the partition sheet.
The second discrimination means marks the discrimination flag of the document type for each partition sheet. At the end of the document is the appearance of the next divider,
Or recognize at the end of one bundle. In the ADF, the page number counter is cleared for each partition sheet, and in the OCR, the page number is searched for each document type (16 in FIG. 1). The resulting images of the partition sheets 11 to 13 are processed, and the partition sheets 11 to 13 are processed.
A processing program according to the format is started by a document discriminating program 17 which finds a mark detection area having a large number of black pixels for every 13 and automatically discriminates the type of the document.

To explain the operation of the present invention, a mark detection area having a large number of black pixels is found for each partition sheet, and the format of the document is automatically recognized based on the position. After the automatic recognition, a processing program suitable for each is activated, and OCR processing is performed on the document read from the scanner.

In a specific embodiment, for example, in a real estate execution system, an entry certificate (for land), an entry certificate (for building), an entry certificate (for section-owned building), a list of parties, a property OCR processing is performed on a bundle of documents in five types of catalogs. In actual operation, a processing program adapted to the next format is started even when an operator is correcting a character by a processing program adapted to the format, and reading and character recognition can be performed simultaneously in parallel.

[0017]

According to the present invention, the kanji OCR attached AD
Since documents can be processed in a bundle on F, operability in operation is improved. The reason is that the operator does not have to be aware of the format of the document to be placed on the ADF and process it separately.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining batch processing of documents in a plurality of formats according to the present invention.

[Explanation of symbols]

 11,12,13 Partition paper

Continuation of the front page (58) Fields investigated (Int.Cl. ⁶ , DB name) G06K 7/00-9/82 G06K 17/00 B07C 1/00-9/00 B65H 1/00-3/68 B65H 7 / 00-7/20 B65H 33/00-33/18 B65H 43/00-43/08

Claims (1)

(57) [Claims] In an OCR using an ADF, a partition paper is sandwiched between a plurality of types of documents in different formats, and the partition paper is marked with a mark defining a format of a subsequent document. A processing program corresponding to the different format is prepared, and further includes a unit for determining the partition sheet and a unit for reading a mark of the partition sheet to recognize and respond to a format of a document to be subsequently read. by starting the program, and the combined treatment in each of the formats pouring from ADF collectively multiple formats to be automatically selected and executed, the mark
Is predetermined according to the different formats
Are set in a plurality of types of mark detection areas, and
The mat can be any one of the above-described types of mark detection areas.
Designated by painting in black, and in OCR
Black pixels included in the area among multiple types of mark detection areas
Specify the area with the largest number and a certain number or more.
OC using ADF, characterized in that the area is determined to be
Batch processing of multiple formats in R.