JPS59178872A - Binary-coding system of contrast picture - Google Patents

Abstract

PURPOSE:To attain binary-coding without missing of contrast picture information at each part even if there exists density gradient in a contrast picture by dividing the contrast picture into plural small regions, calculating histogram at each divided region and obtaining a threshold value of binary-coding. CONSTITUTION:The contrast picture is stored in an input memory 1 and the entire content of the input memory 1 is transmitted to a histogram operator 2 through a data line (a) and the entire histogram is calculated at the device 2. A threshold value deciding circuit 3 inputs the histogram obtained at the histogram operator 2 via a data line (b) to decide the threshold value for binary- coding. On the other hand, the histogram obtained by the histogram operator 2 is fed to a threshold width deciding circuit 4 via a data line (c) and the threshold width is decided by referencing the threshold value data line (d) binary-coded from the entire picture being an output of the threshold value deciding circuit 3. Threshold value deciding circuits 6a, 6b,-,6n decide the threshold value for binary-coding in each small region based on the content obtained via data lines f1, f2,-,fn.

Description

DETAILED DESCRIPTION OF THE INVENTION Technical Field of the Invention The present invention relates to image processing and to a method for converting an image having a given gray scale (hereinafter referred to as a gray scale image) into a binary image.

PRIOR ART AND PROBLEMS Conventional methods for binarizing grayscale images include determining a threshold value for binarization from the histogram of the image and binarizing the entire image using that value, as shown in Figure 1. The relationship between the histogram and the binarization threshold is shown. In the figure, the vertical axis shows the concentration, and the horizontal axis shows the frequency (
1iii prime number). In FIG. 1, α is the number of pixels with low density.β is the difference between the number of pixels with high density and the number of pixels with low density.

Next, we will discuss the problems when binarizing a grayscale image using the X histogram. First, let us consider the case where the grayscale image has a full density gradient as shown in FIG. In Figure 2, the area near the coordinates (0,0) has low density (light);
The closer to the coordinates (vx, wy) the higher the concentration (darker)
Suppose there is a concentration gradient like this. In this case, the density histogram shows that the difference β between the numbers of gray pixels is smaller than α, as shown in FIG. When a grayscale image as shown in FIG. 2 is binarized using one threshold value, the image shown in FIG. 4 becomes an image in which the light areas are faint and the edges are blurred.

In order to eliminate all of these drawbacks, the conventional method is to divide a grayscale image into several small areas, obtain a histogram like the one shown in Figure 1 for each divided area, and then perform binarization for each divided area. was adopted.

However, the conventional image division method has the disadvantage that, for example, when noise is partially mixed in at a low level, this noise level is taken advantage of.

OBJECTS OF THE INVENTION In view of the shortcomings of the prior art, it is an object of the present invention to provide a system that can binarize the grayscale image without losing the entire grayscale image in each part even if the grayscale image has a gradient in density. It is.

According to the present invention, the object is to provide a memory for storing grayscale images, a histogram calculator for extracting a histogram of the memory and an arbitrary small area on the memory, and a histogram extracted by the histogram calculator. A threshold value determining circuit is provided to calculate all threshold values for binarizing the memory and the small area on the memory, and a histogram is obtained from the entire memory, the binarization threshold value for the entire memory is determined, and the threshold value is determined from the shape of the histogram. Set the upper and lower limits of the binarization threshold according to the density distribution of the image,
Divide the grayscale image into a plurality of small regions, find a threshold value for binarizing all histograms for each divided region, and if the value of the threshold is within the range of the upper and lower limits, use the value to binarize all the small regions. By providing a binarization method for a grayscale image, which is characterized in that the small area is binarized using a value that is closer to the upper limit or the lower limit of the range if it is not within the range. achieved.

Examples of the invention Embodiments according to the present invention will be described in detail below.

The gist of the present invention is to first take a histogram of the entire grayscale image and set it as the all center α, including the allowable range of the binarization threshold.
If β is thicker, it is narrower, and if β is smaller, it is set wider. Next, the entire image is divided into several small regions, and the threshold values for binarization are determined for each divided small region, and the values are within the allowable width determined from the entire one-plane image. If it is within that range, then use that value as the threshold, and if it is not within that range, then use the one of the upper and lower limits of that range that is closer to the binarization threshold found for each divided area as the threshold for full binarization. This is to binarize each area.

FIG. 5 shows a diagram of one embodiment of the present invention.

In FIG. 5, an input memory 1 stores grayscale images.

The entire contents of the input image 1 are sent to a histogram calculator 2 via a data line a, where a histogram of the entire image is obtained. The threshold value determining circuit 3 inputs the histogram obtained by the histogram calculator 2 via the data line b"4 and determines all thresholds for binarization. On the other hand, the histogram obtained by the histogram calculator 2 is input via the data line cf. The threshold width is sent to the threshold width determining circuit 4, and the threshold width of the divided small area is determined by referring to the threshold value for binarization of the entire image, which is the output of the threshold determining circuit 3, using the data line d.

On the other hand, the image in the input memory 1 is divided into n parts, and the contents of the divided small areas are data lines eI+e2・'=
Histogram calculators 5a, 5b via e1.

. . , 5n, which takes a histogram in each small area and sends it to threshold value determination circuits 5a and 5b.

..., 6n (the data line f1 + f2 +...
. . . Determine the threshold value for binarization in the small area based on the content obtained through +fn.

The threshold comparison circuits 7a, 7b,..., 7n refer to the allowable width of the threshold for binarizing the entire image from the data line g, and perform binarization in a small area from the data lines h1+h2+...+hn. The threshold for binarization of the small area is finally determined by comparing the threshold with the threshold. The binarization circuit 8 reads the target small area from the data line e1 + ``21...en, and calculates the threshold value for binarization in the small area from the data line '+ + '2 r...+'n. Reference data line jt+jz
A binary image is stored via +...+jn.

Claims (1)

[Claims] a memory for storing a grayscale image; a histogram calculator for extracting a histogram of the memory and any small area on the memory; and a histogram for extracting the memory and the small area on the memory from the histogram extracted by the histogram calculator. A dark value determination circuit is provided to calculate the threshold for binarization,
The binarization threshold for the entire histogram memory is determined from the entire memory, and the upper and lower limits of the binarization threshold are set according to the density distribution of the image from the shape of the histogram. is divided into a plurality of small regions, a histogram is taken for each divided region, a threshold value for binarization is determined, and if the value of the threshold value is within the range of the upper and lower limits, the small region "f: binary value , and if it is not within the range, the small area is converted to 2(tl!i) with a value that is closer to the value of the upper or lower limit of the range.