JPS63257081A - Document input system - Google Patents

Abstract

PURPOSE:To reduce erroneous recognition and to surely find the separation of a document, by identifying an input as the document or a section sheet, and recognizing, if it is the section sheet, the completion of one document. CONSTITUTION:The document desired to be registered is placed on a hopper 1 first. Inputted documents are unified at every separation and a bundle on the forefront of which a black sheet 8 called as the section sheet is placed is generated in advance. The document inputted from the hopper is scanned in order, and after being converted to an electrical signal at a photoelectric transducing part 2, is binarized at a binarization circuit 3, and is transmitted to a threshold circuit 5 via a sunpot counter 4. At the counter, the number of the sunpots is counted from an inputted binary image of one document, and a threshold processing is applied at the threshold circuit. The threshold processing is performed in such a way that the input is decided as the section sheet when a ratio of sunpot to the binary image in one document exceeds 95% and it is transmitted to a control part 6. The documents inputted successively are blocked at every document, and are filed in an optical disk device automatically.

Description

[Detailed Description of the Invention] [Object of the Invention] (Industrial Application Field) The present invention relates to a document system for storing document images continuously input via an image input device or a character recognition device in document units. Regarding input methods.

(Conventional technology) Recently, with the development of information processing technology, it has become possible to perform filing, which used to be done in the world of Gion, in the electronic world. In electronic filing systems, when attempting to register a document, it is necessary to input data from a keyboard. However, since the key human effort required for document registration work is a large amount of time, automation of the input work using a character recognition device is strongly desired in such systems.

Automation of document registration is not only required from the perspective of saving labor in input work, but is also important for improving search sophistication and man-machine performance.

In order to automatically register documents, it is necessary to automatically extract and recognize the type of input document.

This is because - with continuous document data where documents are not filed as one block, the operator must refer to the input documents sequentially from the first input to find the necessary information, which places a great burden on the operator. This is because it takes

Up until now, when attempting to automate document input in a character recognition device, it has been considered to input a special sheet of paper on which necessary information for one document is written at the beginning. However, in this method, the characters written on the first sheet of paper must be recognized, which requires a complicated recognition process, so it takes time to process the first sheet of paper. Further, erroneous recognition such as erroneously recognizing a part of an input document may also occur.

(Problems to be Solved by the Invention) An object of the present invention is to provide a document input method that allows documents each consisting of multiple pages to be input at once and stored in document units. .

[Structure of the invention]

(Means for Solving the Problem) The present invention provides a hopper on which a plurality of input documents can be placed, and a document image signal by scanning input sentences = 3 placed on this hopper sheet by sheet. In an apparatus having a document storage section that sequentially stores a document image signal obtained by the scanning means or a result of character recognition of the document image signal, the apparatus includes a document storage section that sequentially stores a document image signal obtained by the scanning means or a result of character recognition of the document image signal. is the color,
means for detecting the separator sheet by inserting separator notes with different properties such as shape or size and determining the difference in the properties from the document image signal obtained by the scanning means; A control means is provided for sorting and storing one or more document image signals in the document storage section as one unit document, obtained from the scanning means during the time when a sheet is detected. The document input method is

(Operation) In addition to registering a document image, the present invention divides an input document (generally a plurality of pages) into one section for each section and inputs the input documents successively after a specific sheet called a section sheet. A section sheet is a sheet that does not contain any characters that need to be recognized, and is intended to separate consecutively input document data into individual documents. They are different in this respect.

The character recognition device identifies whether the input is a document or a section sheet, and if it is a section sheet, it recognizes that it is the end of one document, and the storage device stores each document. will be blocked and automatically filed. Distinguishing sheets on which no characters to be recognized are written from other documents has the following effects. That is, there is no need to perform complex character recognition processing, and recognition is easy, so processing can be carried out in a short time, and there are fewer erroneous recognitions, making it possible to reliably find document boundaries. This speeds up the document input work and saves labor, and since each document is filed in a block in the storage device, it can be easily searched.

(Example) An example of the present invention will be described below with reference to the drawings.

Figure 1 shows an example of a system for registering documents as images. The document to be registered is first placed on hopper 1. Here, input documents are grouped into categories as shown in Figure 2. , make a bundle with black sheet 8 at the beginning, which is called a section sheet.

Documents input from the hopper are sequentially scanned, converted into electrical signals by a photoelectric conversion section 2, binarized by a binarization circuit 3, and transmitted to a threshold circuit 5 via a black point counter 4. Here, the number of black dots is counted from the input binary image of one document, and threshold processing is performed by a threshold circuit. What is threshold processing?1
If the proportion of black dots in the binarized image of the document is 95 inches or more, it is determined that it is a section sheet, and this is transmitted to the control unit 6. Further, the size of the input document is detected when moving from the hopper to the photoelectric conversion section, and is transmitted to the control section, and the control section determines a threshold value that matches the size of the input document.

The binarized document images are sequentially written into the data area 11 of the optical disk device 7, as shown in FIG. When a section sheet is detected, the control unit creates a directory for the input document in the directory area 10, and writes the start address of the written document image in the data area.

In this way, one directory corresponds to one input document, and as many directories as there are documents are created. In this way, you can easily search for the desired document by referring to the directory.

The documents that are continuously input through the above processing are divided into blocks one by one and automatically filed by the optical disk device.

FIG. 4 shows an example of a system in which the character information of a document is encoded and then registered. A document to be registered is first placed on the hopper 101. Here, as shown in FIG. 2, the input documents are grouped into sections to create a bundle with the black sheet 13 at the top, which is called a section sheet. Documents input from the hopper are sequentially scanned, converted into electrical signals by a photoelectric conversion unit 102, binarized by a binarization circuit 103, and sent to a recognition unit 108 via a preprocessing unit 105.
8, each character is recognized by comparing it with the dictionary 107. Detects the font, etc. of the document from the obtained recognition results,
A post-processing dictionary suitable for the document is selected from the post-processing dictionary 110, and a post-processing dictionary-1 is selected for each word in the post-processing unit 111.
The final result is determined by comparing it with 1F110. The document image binarized by the binarization circuit 103 is processed by the black point counter 1.
04 to the threshold circuit. Here, the number of black dots is counted from the input binary image of one document, and threshold processing is performed by a threshold circuit 106. Threshold processing means that if the proportion of black dots in the binarized image in one document is 95% or more, it is determined that it is a section sheet, and the result is transmitted to the control unit 109. Further, the size of the input document is detected when the document is moved from the hopper to the photoelectric conversion section and is transmitted to the control section, and the control section determines a threshold value that matches the size of the input document.

Documents that are continuously input through the above processing are divided into blocks one by one and automatically filed as character code information in the disk device j1112. The advantage of a black section sheet with no text written on it is that there is no need for character recognition, so misrecognition does not occur. If a white section sheet is used, there will be some inconvenience. If the document is turned upside down, there may be a white surface, making it difficult to tell whether it is a document or a section sheet. Furthermore, even if the black sheet gets dirty, it does not affect recognition, so it can be used over and over again.

In this way, in the above example, by using the black sheet as the section sheet, it is easy to recognize that the end of one document is the end of the document, whether it is filed as an image or converted to character code using OCR. Documents can be divided into blocks and filed. With this method, OCR input of documents is greatly automated and labor-saving. Furthermore, when searching after filing, the burden on the operator is reduced.

Note that the present invention is not limited to the above embodiments. The section sheet may be made of paper or other materials, and although it was black in the above embodiment, it may be of any color depending on the characteristics of the scanner. Also, the shape and size do not have to be the same as the document. It may have cut corners, holes, or ears. In short, the present invention can be implemented with various modifications without departing from the gist thereof.

〔Effect of the invention〕

Thus, according to the present invention, when inputting a document using OCR or the like to electronically file it, by inputting the section sheet at the beginning of the document, the document input work is automated and labor-saving, and the search after filing is also easy and efficient. The operator's burden is greatly reduced.

[Brief explanation of drawings]

FIG. 1 is a schematic configuration diagram of an apparatus according to an embodiment of the present invention, FIG. 2 is a diagram showing an example of a read document, three stripes are diagrams for explaining the filing process, and FIG. 4 is a diagram showing an example of a document to be read. It is a figure which shows another Example. 1... Hopper, 3... Binarization circuit, 4... Black spot counter, 5... Threshold circuit, 7... Optical disk device. Agent Patent Attorney Nori Chika Yuraku-m--

Claims (2)

[Claims] (1) A hopper on which multiple sheets of input documents can be placed; a scanning means for scanning each input document placed on the hopper to obtain a document image signal; and a document image obtained by the scanning means. In an apparatus having a document storage section that sequentially stores the results of character recognition of a signal or this document image signal, each document is defined as having some property such as color, shape, size, etc. for each arbitrary section of the input documents. means for detecting the separator sheet by inserting a different separator sheet and determining the difference in the properties from the document image signal obtained by the scanning means; and the scanning means between the times when the separator sheet is detected by the means. 1. A document input method comprising: control means for dividing and controlling storage of one or more document image signals obtained from one or more documents in the document storage section as one unit of document. (2) The document storage section includes a document data storage section that stores the document image signal obtained by the scanning means or the result of character recognition of the document image signal, and a document data storage section that classifies the contents of the document data storage section in document units. and a directory storage unit that stores directory information instructing the separator sheet, and the control means is configured to:
2. The document input method according to claim 1, further comprising means for creating new directory information for instructing classification of new document units and writing this into a directory storage section of said document storage section.