JPH04340682A - Optical character reader - Google Patents

Abstract

PURPOSE:To execute the read processing of an object such as a slip or the like at high speed as a whole. CONSTITUTION:In the process of optically scanning the object by a scanner 11, outputting image data B for the unit of a line and storing these image data B for the unit of the line in an image memory 15, the information of edge positions as the right and left edges of the object is detected from the image data B for the unit of the line, the right and left edge position information to be successively detected is temporarily stored for plural lines, position deviation correction information and inclination correction information F of the object such as the slip or the like is calculated based on the relevant right and left edge position information and by using the calculated position deviation correction information and inclination correction information F, address correction is executed to correct position deviation to a read address H. Therefore, a character recognition processing can be immediately started to the obliquely set object such as the slip or the like. Thus, the read processing of the object such as the slip, etc., can be executed at high speed as a whole.

Description

[Detailed description of the invention]

0001

[Industrial Application Field] The present invention relates to an optical character reading device that optically scans an object such as a form and performs character recognition using the read image data, and particularly relates to an optical character reading device that optically scans an object such as a form and performs character recognition using the read image data. The present invention relates to an optical character reading device having a correction function for positional deviation.

0002

2. Description of the Related Art FIG. 3 is a block diagram showing a schematic configuration of an example of a conventional optical character reading device.
Consists of. The scanner 1 outputs line-by-line image data b by optically scanning a document a intended for character reading processing. Image memory 2 is
Image data b in line units from the scanner 1 is sequentially stored for a plurality of lines. The address generation circuit 3 generates a read address c for the image memory 2 based on predetermined format definition information d. The character recognition device 4 identifies a character based on the character image e read out from the image memory 2 and cut out based on the address c, and sets it as a recognition result f.

0003

[Problems to be Solved by the Invention] In the case of such an optical character reading device, if a form is set in the scanner 1 at an angle, the image data is also stored in the image memory 2 at an angle. It is necessary to correct the tilt of the image, but in the conventional method, correction was performed after all image data had been stored in the image memory 2, which caused a delay in the start of processing by the character recognition device 4. As a result, there is a problem in that the reading processing speed decreases.

Another problem is that character recognition processing cannot be started until all image data is stored in the page memory.

The present invention has been made to solve the above-mentioned problems, and an object of the present invention is to immediately start character recognition processing for an object such as a form that is set at an angle. Our goal is to provide technology that makes it possible.

Another object of the present invention is to provide a technique that can perform the entire reading process of an object such as a form at high speed.

The above and other objects and novel features of the present invention will become clear from the description of this specification and the accompanying drawings.

[0008]

[Means for Solving the Problems] In order to achieve the above object, the present invention provides a scanner that optically scans an object such as a form for character reading processing and outputs image data in line units; an image memory for storing image data from the image memory; an address generating means for generating a read address for character extraction for the image memory based on predetermined format definition information; In an optical character reading device equipped with a character recognition means for identifying characters based on character-by-character image data, in the process of storing line-by-line image data output by the scanner into a page memory, line-by-line image data is An edge detection means for detecting edge position information of the left and right edges of an object such as a form from image data; and an edge information register for temporarily storing left and right edge position information sequentially outputted from the edge detection means for a plurality of lines. , a correction information calculation means for calculating positional deviation correction information and tilt correction information of the object such as the form based on the left and right edge position information from the edge information register; The most important feature is that the present invention includes an address converting means for correcting an address for correcting positional deviation with respect to the read address from the address generating means using correction information and tilt correction information.

[0009]

[Operation] According to the above-mentioned means, a process of optically scanning an object such as a form for character reading processing with a scanner, outputting line-by-line image data, and storing this line-by-line image data in a page memory. , detects the edge position information of the left and right edges of an object such as a form from line-by-line image data, temporarily stores the sequentially detected left and right edge position information for multiple lines, and stores the left and right edge position information. Based on the information, the positional deviation correction information and tilt correction information of the object such as the form are calculated, and the calculated positional deviation correction information and tilt correction information are used to correct the positional deviation with respect to the read address. Since the address is corrected, character recognition processing can be immediately started for objects such as forms that are set at an angle. Thereby, the entire reading process of objects such as forms can be performed at high speed.

0010

Embodiments Hereinafter, embodiments of the present invention will be explained in detail with reference to the drawings.

FIG. 1 is a block diagram showing a schematic configuration of an embodiment of an optical character reading device according to the present invention.

As shown in FIG. 1, the optical character reading device of this embodiment includes a scanner 11 and an edge detection circuit (means) 1.
2, an edge information register 13, a correction information calculation circuit (means) 14, an image memory 15, an address conversion circuit (means) 16, an address generation circuit (means) 17, and a character recognition device (means) 18. The edge detection circuit (means) 12, edge information register 13, correction information calculation circuit (means) 14, image memory 15, address conversion circuit (means) 16, address generation circuit (means) 17, character recognition device (means) ) 18 may be replaced by a computer and software.

In FIG. 1, (a) is a form, (b) is image data in line units, (c) is the X coordinate of the left and right edge points for one line, and (d) is the left and right edges for multiple lines. Point X coordinate, (E) is image data in line units, (F)
are positional deviation correction information and tilt correction information, (g) is a readout address, (ch) is a readout address, (li) is format definition information, (nu) is a character image, and (l) is a recognition result. The image data stored in the image memory 15 described in this embodiment will be described as a multi-tone image.

Next, the functions of each of the above means will be explained. Note that the scanner 11, image memory 15, address generation circuit 17, and character recognition device 18 are the same as the conventional ones, so a description of their functions will be omitted here.

The edge detection circuit 12 is connected to the scanner 11.
Input the line-by-line image data (b) from
A process of detecting the left and right ends of the image data (b) for each input line for a predetermined number of lines from the top of the form (a) as edge point X coordinates (c), and sequentially outputting them to the edge information register 13. I do.

The edge information register 13 has a function of storing a plurality of lines of X coordinates (c) of left and right edge points from the edge detection circuit 12.

The correction information calculation circuit 14 calculates the X coordinates (d) of edge points of a plurality of lines from the edge information register 13.
is input, positional deviation information and tilt information (correction information) (f) are calculated, and these are output to the address conversion circuit 16 as parameters.

The address conversion circuit 16 uses the positional deviation correction information and tilt correction information (f) from the correction information calculation circuit 14 as setting information, and converts the logical coordinates (YL, XL) into physical coordinates (Yp, Xp). do. The relationship is as follows.

Setting information: - Positional deviation information (Ypo, Xpo) - Tilt correction information (Ca, Cb) Relational expression:

[0020]

[Math 1]

That is, the address conversion circuit 16 generates an address (g) for read access of the image memory 15 from the physical coordinates (Yp, Xp) and the read address (h) output from the address generation circuit 17. do.

A method (skew correction method) for calculating positional deviation correction information and skew correction information (f) in the correction information calculation circuit 14 will be explained.

Here, it is assumed that the slope stable region detection threshold THw, the number of detection lines TMG on the top side of a form, and the number TD of detection points on the top side of the form are determined in advance. Further, the positional deviation correction information and the tilt correction information (f) are assumed to be ΔW.

First, based on the X coordinates (D) of the left and right edge points of the plurality of lines, ΔWi=XRi − XLi
(2) ΔXR=XRi − XRi−1
(3) ΔXL=XLi − X
Li-1 is obtained by repeating (4) for i lines.

Next, from the line where ΔW<THw, the coordinate points (XR, YR) and the (X, Y) coordinates of the right edge of the form are continuously observed RD times, and the straight line connecting these points is regarded as the right side of the form. , when the slope of the straight line is positive (as in Figure 3),
From the 0th line + TMG line to the left end of the form (X
, Y) coordinates for the number of detection points on the top side of the form (XL
, YL) to find the equation on the upper side.

The simultaneous equations of the equation on the right side and the equation on the upper side are solved to obtain the coordinates (X0, Y0) of the intersection, which is used as positional deviation correction information.

When the detected inclination of the right side of the form is cotθ, the inclination angle is expressed by θ, and the inclination correction information is expressed as follows.

[0028]

[Math 2]

The operation of the edge detection means 12 will be explained using FIG. 2.

FIG. 2 shows that the edge detection circuit 12 detects a form (
FIG. 3 is an explanatory diagram schematically showing how the edge point X coordinate (c) of (a) is detected and stored in the edge information register 13;

As shown in FIG. 2, when the X-axis direction is the main scanning direction and the Y-axis direction is the sub-scanning direction, when tracing individual pixels in one line in the i-th line, Ne consecutive pixels are traced. When the multi-tone pixel values exceed the slope stable area detection threshold THw, that is, when it is considered that the white area inside the form has entered from outside the form, the first pixel is set as the left edge point XLi. , In addition, the right edge point XRi is the area in which Ne consecutive pixels do not exceed the value of the stable slope area detection threshold THw, that is, the pixel immediately before the last point when it is considered to have gone outside the form. (Form edge determination threshold THe
It is assumed that the number of detected form edges Ne is predetermined).

As described above, according to the optical character reading device of this embodiment, the document (a) for character reading processing is sent to the scanner 11.
to optically scan and generate line-by-line image data (b)
In the process of outputting and storing this line-by-line image data (b) in the image memory (page memory) 15,
Edge position information, which is the left and right edges of the form (a), is detected from the line-by-line image data (b), and the sequentially detected left and right edge position information is temporarily stored for multiple lines, and the left and right edge position information is stored. Based on this, the positional deviation correction information and tilt correction information (f) of the object such as the form are calculated, and the read address (chip) is calculated using the calculated positional deviation correction information and tilt correction information (f). ), character recognition processing can be immediately started for the form (a) that is set at an angle. In other words, the form (a)
By detecting the positional deviation and inclination of the image data (e) in the process of writing it into the image memory 15, the positional deviation correction information and the inclination correction information (f) are outputted to the address conversion circuit 16, and then they are immediately detected. It becomes possible to perform character recognition processing. Thereby, the entire reading process of objects such as forms can be performed at high speed.

Although the present invention has been specifically explained above based on examples, it goes without saying that the present invention is not limited to the above-mentioned examples and can be modified in various ways without departing from the gist thereof. .

[0034]

[Effects of the Invention] As explained above, according to the present invention, the positional deviation and inclination of an object such as a form are detected in the process of writing image data to the image memory, so that the positional deviation information and inclination can be detected. After the correction information is output to the address conversion means, character recognition processing can be performed immediately. Thereby, the entire reading process of objects such as forms can be performed at high speed.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing a schematic configuration of an embodiment of an optical character reading device of the present invention;

FIG. 2 is a schematic diagram for explaining how edge point coordinates are detected from image data and stored in an edge information register in this embodiment,

FIG. 3 is a diagram for explaining problems with conventional optical character reading devices.

[Explanation of symbols]

11...Scanner, 12...Edge detection circuit (means), 13
...Edge information register, 14...Correction information calculation circuit (means), 15...Image memory, 16...Address conversion circuit (
(Means), 17... Address generation circuit (Means), 18... Character recognition device (Means), (A)... Form, (B)... Image data in line units, (C)... Left and right edge points X for one line. Coordinates, (d)...X coordinates of left and right edge points for multiple lines, (e)...Image data in line units, (f)...positional deviation correction information and tilt correction information, (g)...reading address, (chi) )...reading address, (li)...format definition information, (nu)...character image, (ru)...recognition result.

Claims (1)

[Claims] Claim 1: A scanner that optically scans an object for character reading processing and outputs image data line by line, an image memory that stores image data from the scanner, and predetermined format definition information. Address generation means for generating a read address for character extraction for the image memory, and character recognition means for identifying a character based on image data in units of characters extracted from the image memory using the read address. In the optical character reading device equipped with the above-mentioned optical character reading device, in the process of storing line-by-line image data output by the scanner in the page memory, edge detection detects edge position information, which is the left and right edges of the object, from the line-by-line image data. means, an edge information register that temporarily stores left and right edge position information for a plurality of lines sequentially output from the edge detection means, and a position of the object based on the left and right edge position information from the edge information register. Correction information calculation means for calculating deviation correction information and tilt correction information, and positional deviation correction for the read address from the address generation means using the positional deviation correction information and tilt correction information from the correction information calculation means. What is claimed is: 1. An optical character reading device comprising an address conversion means for correcting an address.