JPH09305708A - Vertically stretched data correction method for optical character reader - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the recognition rate of an optical character reader by correcting the image data which are read with vertical stretch based on the deformation amount of a reference symbol of the read image data. SOLUTION: When a vertical stretch detection reference symbol 10 is written in a handwritten character entry column of a document, the size of the symbol 10 is equal to the frame size of its next right handwritten character entry column, i.e., equal to an image read range that is set when the document is read. Then the document is read by an optical character reader and the image of the document is acquired with vertical stretch. Therefore, the symbol 10 is previously printed beside the handwritten character entry column of the character read document and then used for correction of the stretch when the data are read. Thus, the stretched data are corrected. As a result, the stretched character data can be accurately read.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to correction of stretched data in an optical character reading device for reading paper or other medium (hereinafter referred to as "form") with a roller or the like while reading it with a video camera, a line sensor or the like. It is about the method.

[0002]

2. Description of the Related Art Conventionally, when a handwritten character or a printed character written on a form is read by an optical character reader, a roller,
We have a conveyor mechanism using a belt inside the device, and we perform reading work while conveying it from the edge of the form. At this time, which position on the form should be read is determined by the sheet ID indicating the form format described in a part of the form.

Further, if the rollers, belts or the like slip during the transportation of the document, the document is not properly transported, and the read image data obtained by the character reading portion may be elongated in the document transport direction. However, in the conventional form, as shown in FIG. 6, a mark is printed on the form to detect whether or not a line gap has occurred between lines in which handwritten characters are written. It was able to detect.

Further, the reading position of the image reading area on the form determined in advance corresponding to the sheet ID is
Although the amount of extension between lines is corrected by the mark as shown in FIG. 6, the extension as shown in FIG. 6 cannot detect the extension of the reading area where characters are actually written. Is not corrected. Further, the optical character reading device is required to have improved functions for high-speed processing of a large amount of data, and it is required to increase the processing speed, improve the reading rate, and downsize the device.

[0005]

However, in the conventional method for correcting the stretched data in the optical character reading apparatus, when the stretch occurs, the stretch between the lines becomes
With the mark for detecting stretch, the coordinates for detecting the amount of stretch and reading on the form can be corrected and read by the amount of stretch, but since conventional stretch detection is performed only between lines, handwritten characters can be written. There is a problem that if the cut lines are extended, they cannot be read correctly. In addition, in order to suppress the occurrence of stretch, it is necessary to take measures such as increasing the precision of the mechanism portion for conveying the form and delaying the conveying speed, which causes a problem of cost increase.

[0006]

According to the present invention, there is provided an optical character reading method for correcting a stretched data, which is an optical method for reading a document in which characters are written as image data and performing character recognition based on the read image data. In a method for correcting lagging data in a character reading device, a reference symbol for detecting lagging is printed on a form in advance, and the lagging and reading image data is corrected based on the deformation amount of the reference symbol of the read image data. It is a thing.

[0007]

BEST MODE FOR CARRYING OUT THE INVENTION

Embodiment 1. According to the present invention, in an optical character reading device, a reference symbol for detection and correction of extension is previously printed beside a handwritten character entry field of a form for reading characters, and the reference symbol is used for correction of extension during data reading. hand,
The extended data is corrected.

FIG. 1 is an explanatory view for explaining a reference symbol for detection and correction of a stretched data correction method according to an embodiment of the present invention. FIG. 1A shows an example of a case in which a reference mark 10 for detecting an extension is printed on the side of a handwritten character entry field in a form. Here, the size of the reference symbol is the same as the frame of the handwritten character entry field arranged on the right side. The size is the same as the range in which the image is read when the form is read.

FIG. 1B shows an example in which the optical character reading device reads a form as shown in FIG. 1A, and when the image of the form is acquired, it is stretched. .

FIG. 2 is a diagram showing an example of the data of the reference symbol when the form as shown in FIG. 1A is normally read. Hereinafter, referred to as black dots) and other portions are represented by white dots. Further, FIG. 3 is a diagram showing an example of data of reference symbols when a form as shown in FIG. 1A is read in an extended state.

Next, the operation of this embodiment will be described. FIG. 4 is a flowchart showing a data correction operation when reading a form. First, read the image of the form with a sensor, camera, etc., and record it in the image memory (S
100), a reference value for the reference symbol is obtained from the sheet ID of the form (S101). Note that this reference value is the maximum value of the number of black dots in the vertical direction when the reading is normal as shown in FIG. 2, that is, the value of XA = 1 shown in FIG. This maximum value (18 in the example of FIG. 2) corresponds to the maximum outer shape in the direction, and is predetermined by the sheet ID.

Here, as the reference symbol read data, if the reading in S100 is normal, the data shown in FIG. 2 is stored in the image memory, and when an extension occurs, it is stored in the image memory. The data shown in FIG. 3 is stored.

Then, the maximum value A (the value of XA = 1) of the number of black dots in the vertical direction of the read reference signal is measured (S10).
2), the difference between A and the reference value is calculated, and the amount of stretch B is calculated (S1
03), it is determined whether B is 0 (S104). S10
If B = 0 in 4 then no data correction is performed and data recognition is performed based on the read data. Here, in the example shown in FIG. 2, the reference value is 18, and the value of A is 18 as well.
Therefore, B = 0, the data is not corrected, and the data recognition is executed based on the read data.

If B ≧ 1 in S104, it means that the extension has occurred, and therefore the correction operation of the data is performed. For the following correction operation, see FIG.
Will be described as an example. In FIG. 3, the maximum outer shape value in the vertical direction is 22, this value is A, and the amount of extension B is a value obtained by subtracting the reference value (18) from A = 22, and B = 4. Has become. This indicates that the read image data extends for 4 dots with respect to the reference value 18.

Then, in order to know the stretched portion, all the addresses of XA showing values other than the maximum outer shape value: A and the normal value: 1 out of the vertical black dot values are searched, and the address DXAN (N = 1, 2 ...) is set (S105),
The amount of extension BN (N =
1, 2 ...) and the number N of extension points are set (S10).
6).

In the example shown in FIG. 3, the number of extended portions is N = 1, and XA = 9: DXA1. And DX
The number of black points of A1 is 5, and a value obtained by subtracting a normal value from the number of black points is the amount of extension B1, and B1 is B1 = 5-1 = 4.

The extension start point DYAN (N = 1,
2 ...) are values obtained by adding 1 to the value of DXAN, so that the extension start point DYAN = DXAN + 1 is obtained (S107), and the extension start point DY obtained in S107 is obtained.
The image corresponding to the extension amount BN address is deleted from the AN from the image memory (S108), and the extension start point DYAN is deleted.
After that, the image data after DYAN + BN + 1 is transferred (S109), and S107 to S109 are repeated by the number N of the extending points, and all the numbers N of the extending points are
When the data correction is completed, data recognition is executed based on the read data (S110).

In the example shown in FIG. 3, the starting point D of the extension is D.
YA1 is a value obtained by adding 1 to the value of DXA1, and DYA1:
9 + 1 = 10. Then, in order to delete the extension area, B1 = 4 addresses (Y
The images of A = 10 to 13) are deleted from the image memory, and the images of YA = 14 to YA = 23 are moved to YA = 11 to YA = 19.

In this embodiment, the deformation amount of the reference symbol is measured, the stretch is detected, and the read image data is corrected based on the deformation amount. Therefore, the stretched character data is accurately read. Therefore, the recognition rate of character recognition can be improved.
Further, since the recognition failure due to the extension is eliminated, it is possible to improve the form conveyance speed, it is possible to improve the recognition processing speed as the device, and the conveyance system speed control accuracy is lowered to cause the extension. Also, since it is readable, the cost of the transport system can be reduced.

In this embodiment, the reference symbol is used for only a part of the form, but the reference symbol may be used for the entire form as shown in FIG. It is also possible to correct the extension between lines.
Further, in this embodiment, the example applied to the optical character reading device has been described, but the processing method of detecting the stretch and changing the read image area is described by detecting the stretch when the image is read. If the processing method is changed to correct the image data while deleting the image data, it can be applied to a facsimile apparatus having no image data storage function of transmitting the image data as it is to the communication line.

[0021]

As described above, according to the present invention, a reference symbol for detecting stretch is previously printed on a form, and based on the deformation amount of the reference symbol of the read image data,
Since the image data read while being stretched is corrected, not only the line spacing but also the stretched character data can be corrected and accurately read, and the recognition rate of the optical character reading device can be improved. Has the effect.

Further, since the recognition failure due to the extension is eliminated, the conveyance speed of the form can be improved, the recognition processing speed as an apparatus can be improved, and further, the speed control accuracy of the conveyance system can be lowered to extend the extension. Even if it occurs, since it can be accurately read, there is an effect that the cost of the transport system can be reduced.

[Brief description of drawings]

FIG. 1 is an explanatory diagram for explaining a reference symbol for detection and correction of the present invention.

FIG. 2 is a diagram showing an example of reference symbol data when normally read.

FIG. 3 is a diagram showing an example of data of reference symbols when read in an extended state.

FIG. 4 is a flowchart showing a data correction operation.

FIG. 5 is a diagram showing another example of reference symbols.

FIG. 6 is an explanatory diagram for explaining a conventional mark for detecting an inter-row extension.

[Explanation of symbols]

 10 Stretch detection reference symbol

Claims (2)

[Claims] 1. A stretched data correction method in an optical character reading device for reading a form in which characters are written as image data, and performing character recognition based on the read image data, wherein a reference symbol for detecting the stretch in advance is added to the form. Is printed, and the image data read by being extended and read is corrected based on the deformation amount of the reference symbol of the read image data. 2. A stretched data correction method in an optical character reading device for reading a form in which characters are written as image data and performing character recognition based on the read image data, wherein the form is preliminarily horizontal to the reading direction. And diagonal 2
Based on the number of data horizontal to the reading direction of the reference symbol of the read image data and the amount of deformation of the data oblique to the reading direction, a reference symbol for detecting the stretch of the book line is printed. And correcting the image data that has been read while being stretched, and a method of correcting stretched data in an optical character reading device.