JPS59117862A - Binary-coding method of picture data - Google Patents

Abstract

PURPOSE:To obtain a binary-coding picture data without shift of position to a picture data by splitting each horizontal line of the pictue data into (n), and processing an average value of the picture data on the split lines as threshold values of the middle position of the small lines. CONSTITUTION:Each horizontal line of the picture data is split into n-set of same length of blocks to form (n) small lines. The middle point and the average value of the picture data are obtained from each small line and the average value is taken as the threshold value at the middle position of the samll line. The threshold value of the middle point of each small line is interpolated linearly as to other points on the lines and taken as the threshold values 5 on each point of the lines. The binary-coding picture data without position shift from the original picture is obtained by comparing the picture data 4 on each point of lines with the threshold value 5.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a method for binarizing image data, in which a floating threshold value is determined and binarized for each horizontal line of image data. be.

The structure of the conventional example and its problems The image binarization method in the conventional example will be described below. Arrange image data (d+j)1U2. , □≦, ≦8.

A and B are the image sizes in the X coordinate direction and the X coordinate direction, respectively. Initial threshold value t. By giving the relation & to O<k<1, and using the following recurrence formulas (1) and (2), the threshold value (tij) at each point is 9. □Calculate 6 and 9.

tl, j= to(izj ot B) −(1)
ti,j=Ct-k)t,-□1. 10 to 1 ton, j (2i i4A, 1 otsu j4B)...(2) Image data (d
ti) and the threshold value (tij), by equation (3), 2
Value image data (b+j)□≦6. □Yu, ask for Yu8.

Incidentally, the horizontal image size Blr is assumed to be 40, and the initial value of the threshold value is t. 'ilo, set the coefficient k to 0, 2,
An example of determining the threshold value for each point is shown in FIG. (1) in Figure 1 is image data on one horizontal line, (2) in Figure 1 is image data on one horizontal line.
) is the threshold value on the same line obtained by the above method. FIG. 2 shows the image data of FIG. 1 and binary image data obtained from the threshold value.

The disadvantage of this method is that it uses a recurrence formula, so the threshold value for each point is calculated using past data, so the threshold value lags behind changes in image data, and binary image data The main problem is that image data tends to shift to the left.

Purpose of the Invention In view of the above-mentioned drawbacks, the present invention provides an image binarization method in which the latter is not misaligned with respect to the former when the image data is fully binarized to create binary image data. The purpose is to provide all the necessary information.

Structure of the Invention In order to achieve all of the above objects, the present invention divides each line of image data in the horizontal direction into t-n small lines, and for all the divided small lines, the image data on the small line is , and use this as the threshold value at the midpoint of the small line.For ground points on the line, the threshold value at the midpoint of each small line is linearly interpolated, and each point on the line is By comparing the image data value and the above threshold value at each point on the line, a binary image or image data is created. It is possible to obtain binary image data with no positional deviation with respect to the original image data.

DESCRIPTION OF THE EMBODIMENTS An image binarization method according to an embodiment of the present invention will be described below with reference to the drawings. The image data is represented by an array (dt3) □YoYU, □U, I. A and B are image sizes in the X and Y coordinate directions, respectively. Figure 3 is a rough image. Focusing on the horizontal l line as shown in (3) in FIG. 3, the following processing is performed.

As shown in FIG. 4, the l line is divided into n blocks of the same length to form n small lines.

For each small line, let the midpoint position and the average e[me of the image data be C and mean, respectively, and the threshold value at C be mean. Threshold values are determined for the n points on the line as described above, and by performing linear interpolation 1 on these, the threshold values at the remaining points are determined.

The method of linear interpolation is as follows. Let the xl marks of the midpoint positions of adjacent blocks be xl and x2, respectively, and the threshold values t1 and t2 at the midpoint positions x1 and x2, then any arbitrary point between these two points (this point ximarkff1
x) is given by equation (4).

2(t2-tx)(x-Xi)/(X2-xx)+t
t・-(4) Linear interpolation is not performed for the points that are to the left of the midpoint of the leftmost minor line l, and the points that are to the right of the midpoint of the rightmost minor line n, respectively. Let the threshold value at the midpoint position of the small line to which the point belongs be the threshold value for that point. Let the threshold value at point (i, D be rjij, and according to equation (5), find binary image data (bt4) 1 ui<A, □23.J (formula (5) is (3) ).

By the way, the number of small lines is &n't4, and for each small line, the first line is 1 to 10, the second line is 11 to 20, and the third line is 21.
Figure 5 shows an example of determining the threshold value assuming that the 4th test consists of 31 to 400 points. (4) in FIG. 5 is image data on one horizontal line, and numeral 5 in FIG. 5 is a threshold value on the same line obtained by the above method. FIG. 6 shows binary image data obtained from the image data of FIG. 5 and the threshold value.

FIG. 7 is a flowchart of an algorithm for realizing the method of the embodiment described above. Here, dg is the point (i
, image data at D, tIj is point (i, threshold value at D, 1) ij is point (i, binary image data at D, m is the number of data on one small line, l is the kakuto of the small line) The number mean is the average of image data of small lines,
C is the midpoint of the small line, and A and B are the image sizes in the horizontal and vertical directions.

Effects of the Invention As described above, according to the present invention, it is possible to obtain a two-Buddhist image data without any positional deviation with respect to the original image data.

[Brief explanation of the drawing]

Fig. 1 is a diagram showing image data and threshold values determined by the conventional method, Fig. 2 is a diagram showing binary image data based on the data in Fig. 1, and Fig. 3 is a diagram showing the size of the image. figure,
Figure 4 is a diagram showing one line extracted from image data.
Fig. 5 is a diagram showing image data and threshold values determined by the method of the embodiment, Fig. 6 is a diagram showing binary image data based on the data of Fig. 5, and Fig. 7 is a flowchart of the algorithm of the embodiment. It is. (4)...Image data on the horizontal l line, (5)...
... Threshold agent on the same line Yoshihiro Moriki Figure 1 Figure 3 Figure 4 Figure 5 Figure 7 (and 2 Figure 7 (b)

Claims (1)

[Claims] 1. Divide each horizontal line of image data into n small lines, calculate the average of the image data on the small lines for all the divided small lines, and calculate this as the midpoint position of the small line. For other points on the line, the threshold value at the midpoint of each small line is linearly interpolated, and the threshold value at each point on the line is set as the threshold value at each point on the line. By comparing with the above threshold value,
A method for binarizing image data to create binary image data.