JPH0620093A - Character recognition device - Google Patents

Abstract

PURPOSE:To easily and quickly correct the inclination of image data obtained from a document. CONSTITUTION:Image data of the document inputted from an input part 1 is stored in a memory 2 as a two-dimensional matrix, and this image data stored in the memory 2 is stepwise read out in accordance with an angle thetaof inclination dependent upon values (a) and (b) read out from an angle table 3 by an inclination extracting part 4, and the sum of black points of each image data is obtained as an inclination value, and the inclination value which maximizes the sum of black points is given to an inclination correcting part 5, and thereby, the matrix of image data in the memory 2 is shifted to correct the inclination of image data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character recognition device used for recognizing printed characters and the like.

[0002]

2. Description of the Related Art Recent advances in character recognition technology are remarkable and have been widely used in various fields.

When performing character recognition on a printed document, for example, each character is read as image data in the character string direction using an optical character reading means, and each character is recognized using this image data. I am trying to do it.

In this case, it is necessary for the character document to have a proper character image data because the character string direction of the printed document is the same as the reading direction of the character reading means, so that accurate character recognition can be performed. Is considered important in.

However, in actuality, when setting a print document or the like for performing character recognition to the character reading means, image data is often read while the set state of the print document is tilted, and the image data is read as it is. In some cases, data was skewed and accurate character recognition was not possible.

Therefore, conventionally, a method of correcting the inclination of the image data has been considered, but the conventional inclination correction method is to correct the inclination of the image data by using a mathematical method such as a trigonometric function. Therefore, the amount of calculation for such inclination correction becomes enormous, and the processing takes a lot of time.

[0007]

As described above, in the correction of the inclination of the image data by using the conventional mathematical method such as the trigonometric function, the amount of calculation for the correction becomes enormous, and the processing takes a lot of time. There was a problem. The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a character recognition device that can easily and quickly correct inclination of image data.

[0008]

SUMMARY OF THE INVENTION According to the present invention, storage means for storing image data of a document in a two-dimensional array, and stairs according to inclination angles set respectively for the two-dimensional array of image data stored in the storage means. And a gradient extracting means for reading out the sum of black points of each image data as a gradient value and outputting a gradient value at which the sum of the black points is maximum,
It is composed of a tilt correcting means for correcting the tilt of the image data by shifting the array of the image data in the memory according to the tilt value extracted by the tilt extracting means.

[0009]

As a result, according to the present invention, the image data stored in the two-dimensional array is read out stepwise according to the inclination angles set, and the inclination value that maximizes the sum of the black points of each image data is obtained. The arrangement of the image data is shifted by the obtained inclination value and the inclination is corrected. As a result, the inclination can be corrected without using a mathematical method such as a trigonometric function, so that the amount of calculation for correcting the inclination of the image data can be reduced and the processing time can be shortened. .

[0010]

First, the concept of the present invention will be briefly described.

Considering the projection of the black dots of the image data of the character string for each line in a horizontally written document, the black dots cannot be distinguished for each line when the document is tilted as shown in FIG. On the other hand, in the case where the document is not tilted as shown in FIG. 2, the black dot projection is obtained for each line, while the image is moderately obtained. This is
The projection of black dots changes depending on the degree of inclination of the document, and the smaller the inclination, the sharper the projection becomes so that each line can be distinguished. Therefore, by utilizing this property, the inclination of the image data due to the inclination of the document can be corrected. An embodiment of the present invention will be described below with reference to the drawings based on this concept.

FIG. 3 shows a schematic configuration of the same embodiment.
In the figure, reference numeral 1 denotes an input unit, and the input unit 1 supplies to a memory 2 image data of a print document read by an optical character reading means (not shown). The memory 2 stores the image data supplied from the input unit 1 in a two-dimensional array on a page-by-page basis. The memory 2 is connected to a tilt extraction unit 4 and a tilt correction unit 5 to which an angle table 3 is connected.

Here, the angle table 3 includes a tilt extraction unit 4
Is a plurality of values a and b for determining the tilt angle θ (FIG. 4) when the image data stored in the memory 2 is read out.

The inclination extraction unit 4 extracts the inclination value D of the image data of the memory 2 according to the inclination of the document, and the inclination angle for reading the image data of the memory 2 from the values a and b read from the angle table 3. θ is determined, each image data is read stepwise according to this inclination angle θ, and the size of the sum of the black points of each image data is obtained as the inclination value D. Such processing is performed for all the values a and b stored in the angle table 3, and the inclination value D of the inclination angle θ having the largest sum of black points is output.

The inclination correction unit 5 shifts the arrangement of the image data in the memory 2 based on the maximum inclination value D extracted by the inclination extraction unit 4 to correct the inclination of the image data.

A control unit 6 stores a predetermined program in advance. The control unit 6 outputs a predetermined control command to the memory 2, the tilt extraction unit 4, and the tilt correction unit 5 according to the program. Has become. Next, the operation of the embodiment configured as described above will be described.

When the control section 6 gives a read command to the input section 1, the input section 1 reads the print characters of a document (not shown). Then, the output from the input unit 1 is stored in the memory 2. In this case, the memory 2
Image data for one page is stored in a two-dimensional array. In this state, the first values a and b for determining the tilt angle θ of the image data read from the angle table 3 according to the instruction of the control unit 6
Is read out and sent to the inclination extracting unit 4.

Then, the inclination extracting unit 4 uses these values a,
As shown in FIG. 4B, the tilt angle θ for reading the image data of the memory 2 is determined, and each image data is read stepwise according to the tilt angle θ, and the sum of the black dots is obtained from each of these image data. The size is the slope value D
Is required as.

Such a processing operation is performed for all the values a and b stored in the angle table 3. In this case, the inclination angle θ of the values a and b when the inclination value D becomes maximum is stored. I'll do it. Finally, the inclination extraction unit 4 outputs the inclination value D of the inclination angle θ that gives the largest sum of the black dots.

Thus, for example, when the document is tilted as shown in FIG. 1, each value a from the angle table 3,
When the tilt angle θ for reading the image data of the memory 2 determined from b matches the tilt of the document, this tilt angle θ
The sum of the black points of each image data of the memory 2 read in a stepwise manner along the maximum becomes maximum, and this is the slope value D.
Will be output as.

As a result, if the image data in the memory 2 is shifted by b / a by the inclination correction unit 5 based on the inclination value D, the inclination of the image data can be corrected. Will be able to.

Further, when the document is not tilted as shown in FIG. 2, when the tilt angle θ for reading out the image data of the memory 2 determined from the values a and b from the angle table 3 is 0, this tilt angle is 0. The sum of the black points of each image data of the memory 2 read along θ = 0 becomes the maximum, and this is output as the slope value D. In this case, the tilt correction unit 5 does not correct the image data in the memory 2.

Therefore, in this way, with respect to the two-dimensionally arrayed image data obtained from the print document, the inclination extraction unit 4 reads out stepwise according to the inclination angles θ of the values a and b read from the angle table 3. At the same time as reading, the sum of the black points of each image data is obtained as a tilt value, the tilt value that maximizes the sum of the black points is extracted, and this is given to the tilt correction unit 5 to shift the image data array and correct the tilt of the image data Therefore, the processing amount for correcting the inclination of the image data can be dramatically reduced as compared with the conventional method that requires a mathematical method such as a trigonometric function and the amount of calculation becomes enormous. Since it can be made simple and the processing time can be shortened, high-speed processing can be realized.

The present invention is not limited to the above-mentioned embodiments, and can be carried out by appropriately modifying it without departing from the scope of the invention.
For example, in the above description, a horizontal document is described, but the same applies to a vertical document. Further, in the above description, the image data of the document for one page is directly stored in the memory 2
However, the present invention can be applied to a memory which is compressed to about 1/4 and stored in the memory 2 at the time of writing.

[0025]

According to the present invention, the image data stored in the two-dimensional array is read stepwise according to the inclination angles set, and the inclination value that maximizes the sum of the black points of each image data is obtained. The inclination of the image data is shifted by the inclination value to correct the inclination. This makes it possible to correct the tilt without using a mathematical method such as a trigonometric function, so that the amount of calculation for tilt correction of image data can be dramatically reduced and the processing can be simplified. Since the processing time can be shortened, high-speed processing can be realized.

[Brief description of drawings]

FIG. 1 is a diagram for explaining the concept of the present invention.

FIG. 2 is a diagram for explaining the concept of the present invention.

FIG. 3 is a diagram showing a schematic configuration of an embodiment of the present invention.

FIG. 4 is a diagram for explaining the operation of the embodiment.

FIG. 5 is a diagram for explaining the operation of the embodiment.

[Explanation of symbols]

1 ... Input unit, 2 ... Memory, 3 ... Angle table, 4 ... Tilt extraction unit, 5 ... Tilt correction unit, 6 ... Control unit.

Claims (1)

[Claims] 1. A storage unit for storing image data of a document in a two-dimensional array, and the image data of the two-dimensional array stored in the storage unit are read out stepwise according to a tilt angle set for each image data. A slope extracting unit that obtains the sum of black points as a slope value and outputs a slope value that maximizes the sum of the black points, and shifts the array of image data in the memory according to the slope angle extracted by the slope extracting unit. A character recognition device, comprising: a tilt correction means for correcting the tilt of image data.