JPH05225381A - Optical character reader - Google Patents

Abstract

PURPOSE:To improve the performance of an optical character reader by eliminating the necessity of detailed correcting information calculating processing for a business form having no inclination and rapidly executing business form reading processing. CONSTITUTION:The optical character reader having an extracting operation compensating part 7 for calculating the inclination and positional deviation of a business form from business form data inputted by an image scanner 1 and stored in an image memory 2 and compensating the operation of a character data extracting part for extracting character data from the business form data and a character recognizing part 6 for recognizing a character from the extracted character data is also provided with a business form detecting part 8 for detecting the width of the business form based upon the image data inputted from the image scanner 1 and a skew variable measuring part 9 for measuring the inclination of the business form prior to the compensating operation of the compensating part 7 to control the operation of the compensating part 7.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optical character reading device (OCR: Optical C) for performing character recognition based on image data read by optically scanning a form.
The present invention relates to an optical character reading device suitable for efficiently correcting a tilt and a positional shift of a document during scanning.

[0002]

2. Description of the Related Art Conventionally, for correcting the inclination of a paper in a recognition system of a character image printed or handwritten on a paper, for example, "Electronic Information and Communication Handbook" edited by Institute of Electronics, Information and Communication Engineers (1988 Ohmsha) As described on pages 1114 to 1115 in (Issue), the conveyance system mechanically corrects the inclination of the sheet, or detects the inclination of the sheet as shown in FIG. , There is something to correct.

FIG. 3 is a block diagram showing the structure of a conventional optical character reading device. In the figure, reference numeral 1 is an image scanner, which optically scans a form a for character reading processing to output image data b in line units. An image memory 2 stores image data b from the image scanner 1 line by line for a plurality of lines in sequence. Three
Is a correction information calculation unit, which calculates the inclination and displacement of the form from the page image c from the image memory 2 and outputs the correction parameter e. Reference numeral 4 is an address generation unit as a character image data extraction unit, which generates a read address f based on predetermined format definition information d. Reference numeral 5 denotes an address conversion unit, which constitutes the extraction operation correction unit 7 together with the correction information calculation unit 3, and based on the correction parameter e, the inclination and position shift of the form with respect to the read address f from the address generation unit 4 are corrected. The read address g whose address is converted so as to be corrected is output to the image memory 2. A character recognition unit 6 recognizes a character based on the read address f output from the address conversion unit 5 based on the character image h cut out from the image memory 2 and outputs it as a recognition result i. ..

As described above, in the case of the conventional optical character reading device, when the document is set in the image scanner 1 while being tilted, the image data is also stored in the image memory 2 while being tilted. Therefore, it is necessary to detect the inclination of the image memory 2, and in order to change the read address generated by the address generation unit 4, the extraction operation correction unit 7 including the correction information calculation unit 3 and the address conversion unit 5.
Is provided. However, in the related art, the correction information calculation unit 3 calculates detailed correction information that requires a long processing time, regardless of whether or not the document is tilted. Therefore, it causes the start of the processing of the character recognition unit to be delayed, and as a result, there is a drawback that the speed of the entire reading processing decreases.

[0005]

The problem to be solved is that in the prior art, detailed correction information calculation processing is performed regardless of whether or not the inclination of the form has occurred.
The point is that the character recognition process cannot be started immediately. The object of the present invention is to solve these problems of the prior art,
Provided is an optical character reading device capable of speeding up the reading process of a form by immediately starting the character recognition process without performing a detailed correction information calculation process on a form in which an inclination has not occurred. That is.

[0006]

In order to achieve the above object, the optical character reading apparatus of the present invention comprises (1) optically scanning a form to be subjected to character reading processing to obtain image data in line units. An image scanner that outputs the image data, an image memory that stores the image data from the image scanner, a character image data extraction unit that extracts character-by-character image data from the image data stored in the image memory, and an image memory that stores the image data. Based on the image data, the inclination and position shift of the form are calculated, and the extraction operation correction unit that corrects the extraction operation of the image data in character units of the character image data extraction unit, and the correction operation of the extraction operation correction unit, Based on the character-based image data extracted from the image memory by the character image data extraction unit, In an optical character reading device equipped with a character recognition unit for identifying characters, a form width detection unit that detects the width of a form using image data output by an image scanner in line units, and this form width detection unit Based on the change in the width of the form to be detected, a skew amount measuring unit that measures the inclination of the form is used before the operation of calculating the inclination and position shift of the form using the image data stored in the image memory by the extraction operation correction unit. It is characterized in that the calculation operation of the extraction operation correction unit is controlled based on the measurement result of the skew amount measurement unit. (2) In the optical character reading device according to (1) above, the form width detection unit selects either a white level or a black level for each line from the image data output from the image scanner in line units. One of the features is that the number of dots is sequentially detected, and the skew amount measuring unit measures the inclination of the form based on the change in the number of dots for one line sequentially detected by the form width detecting unit.

[0007]

In the present invention, the inclination of the form is detected before the correction information is calculated in the process of storing the image data for each line output from the image scanner in the image memory. Then, only when the inclination is large and exceeds the predetermined threshold value, the inclination correction processing based on the correction information calculation is performed, and when the inclination of the form is small, the correction information is not calculated and the character is immediately output. Perform recognition processing. As a result, the character reading operation can be speeded up. Further, the correction information can be calculated with higher accuracy by performing the correction information calculation process according to the tilt situation using the tilt information detected here.

[0008]

Embodiments of the present invention will now be described in detail with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of a configuration according to the present invention of an optical character reading device according to the present invention. The optical character reader of this embodiment is the image scanner 1 described in the conventional optical character reader shown in FIG. 3, the image memory 2, the address generator 4 as a character image data extractor, and the correction information calculator 3. An extraction operation correction unit 7 including an address conversion unit 5 and a character recognition unit 6 are provided with a form width detection unit 8 and a skew amount measurement unit 9 for detecting the inclination of a form according to the present invention. ing. With such a configuration, the optical character reading device according to the present embodiment stores the image data from the image scanner 1 in the image memory 2,
Before the correction information calculation unit 3 calculates the correction information, the inclination of the form is detected. If there is no inclination of the form, the correction information is not calculated and the character recognition process is immediately performed to speed up the entire reading process. Turn into Hereinafter, such operation will be described. The image data stored in the image memory 2 of this embodiment is a multi-gradation image.
It is not limited to this. Further, the image scanner 1, the image memory 2, the correction information calculation unit 3, the address generation unit 4, the address conversion unit 5, and the character recognition unit 6 are the same as those of the conventional one shown in FIG. Is omitted.

The form width detection unit 8 receives the image data b in line units from the image scanner 1 as an input, and receives the form a.
After detecting the tip of the
For each input line, the number j of white level dots for one line, which is information indicating the width of the form, is detected, and the skew amount measuring unit 9
, And sequentially output. Then, the skew amount measuring unit 9 measures the tilt information k based on the number j of white level dots for one line from the form width detecting unit 6 and sends it to the extracting operation correcting unit 7. The operation of measuring skew information of the form by the skew amount measuring unit 9 will be described.

First, the difference value for each line of the white level dot number j for one line input from the form width detection unit 8 is calculated. Next, this difference value is accumulated for a predetermined number of lines, and this is set as the increase amount. Then, the average value av of the increase amount is obtained, and this average value av is compared with a predetermined threshold value to obtain inclination information k. For example, the tilt information k
Is no slope = av ≧ threshold value α (1) Slope (middle) = threshold value α> av ≧ threshold value β (2) Slope (large) = av <threshold value β ... (3).

In this way, the extraction operation correction unit 7 including the correction information calculation unit 3 and the address conversion unit 5 operates the correction information calculation unit 3 based on the tilt information k obtained by the skew amount measurement unit 9. To control. For example, when the tilt information k is “no tilt” in (1) above, the correction information calculation unit 3
Correction information is not calculated. As a result, when the form is set normally and there is no inclination, the operation of the correction information calculation unit 3 becomes unnecessary, the character recognition unit 6 immediately performs the character recognition process, and the reading process is performed at high speed. be able to. It is also possible that the correction information calculation unit 3 performs the correction information calculation processing according to the tilt state based on the tilt information k from the skew amount measurement unit 9, and the correction information calculation can be performed more accurately. Can be done.

Next, the processing in the form width detecting section 8 will be described with reference to FIG. FIG. 2 is an explanatory diagram showing an embodiment of the processing operation according to the present invention of the form width detection unit in FIG. This figure shows that the form width detection unit 8 of FIG. 1 uses the image data of the line unit from the image scanner 1 of FIG.
It shows how to detect the white level dot number j for one line. When the X-axis direction is the main scanning direction and the Y-axis direction is the sub-scanning direction, in one line of the n-th line, a pixel exceeding a predetermined threshold value is a white level pixel,
The sum is calculated and used as the number of white level dots in the nth line. In this embodiment, the number of white level dots in the first line is 4, and hereinafter, 12 in the second line, 21 in the third line, 28 in the fourth line, 34 in the fifth line and 36 in the sixth line. .., and 37 on the 7th line and the 8th line, which is measured by the skew amount measuring unit 9 in FIG. 1 as if the document is inclined.

As described above with reference to FIGS. 1 and 2, the optical character reading apparatus of this embodiment is provided with the form width detecting section and the skew amount measuring section.
Since the inclination of the form is detected in the process of writing the form image data in the image memory, it is possible to know in advance whether the form is inclined before the calculation of the detailed correction information is started. As a result, when there is no inclination of the form, the correction information is not detected and the character recognition process can be started immediately, and the entire reading process can be speeded up. Further, if the correction information calculation processing according to the tilt situation is performed based on the tilt information from the skew amount measuring unit, the correction information calculation can be performed accurately. The present invention is based on FIG.
The present invention is not limited to the embodiment described with reference to FIG. For example, in the form width detection unit of the present embodiment, the number of dots of the white level according to the form is used to detect the width of the form, but the skew amount is measured even if the number of dots of the black level outside the form is used. The department can measure the inclination of the form.

[0014]

According to the present invention, detailed correction information calculation processing is not performed on a form in which no inclination has occurred,
By starting the character recognition processing immediately, it is possible to speed up the processing of reading the form and improve the performance of the optical character reading device.

[0015]

[Brief description of drawings]

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

2 is an explanatory diagram showing an example of a processing operation according to the present invention of a form width detection unit in FIG. 1. FIG.

FIG. 3 is a block diagram showing a configuration of a conventional optical character reading device.

[Explanation of symbols]

 1 image scanner 2 image memory 3 correction information calculation unit 4 address generation unit 5 address conversion unit 6 character recognition unit 7 extraction operation correction unit 8 form width detection unit 9 skew amount measurement unit a form b image data in line units c page image d Format definition information e Correction parameter f Read address g Read address converted address h Character image i Recognition result j White level dot number for one line k Tilt information

Claims (2)

[Claims] 1. An image scanner for optically scanning a form to be subjected to a character reading process to output image data in units of lines, an image memory for storing image data from the image scanner, and the image memory. Character image data extracting means for extracting character-by-character image data from the image data stored in the image data, and the character image data extracting means for calculating the inclination and displacement of the form based on the image data stored in the image memory. Based on the image data in character units extracted from the image memory by the character image data extraction means based on the correction operation of the extraction operation correction means for correcting the extraction operation of the image data in character units , An optical character reader equipped with character recognition means for identifying characters In the above arrangement, using the line-by-line image data output by the image scanner, a form width detection means for detecting the width of the form in line units and a change in the width of the form in line units detected by the form width detection means are used. Based on the above, the skew amount measuring means for measuring the inclination of the form is provided before the operation of calculating the inclination and the positional deviation of the form using the image data stored in the image memory by the extraction operation correcting means, and the skew amount is measured. An optical character reading device, characterized in that, based on a measurement result of the means, the operation of calculating the inclination and position shift of the form of the extraction operation correcting means is controlled. 2. The optical character reading device according to claim 1, wherein the form width detecting means selects either a white level or a black level for each line from the image data output from the image scanner in line units. The number of dots on one side is sequentially detected, and the skew amount measuring means is
An optical character reading device characterized in that the inclination of a form is measured based on a change in the number of dots for one line, which the form width detecting means sequentially detects.