KR930005131B1 - Histogram equalization for picture tone - Google Patents

Abstract

The method abstracts half-tone image using improved histogram equalization method to improve the resolution of the half-tone image. It includes a central process unit (10), a manuscript scanning section (20) for converting the analog image signal obtained by scanning to digital image signal, a signal processor (30) for encoding/decoding the image data from the manuscript scanning section (20), a recording section (40) for printing the image data onto thermal sensitive paper, and a transmitting section (50) for transmitting/receiving the image data to/from the other system.

Description

Halftone Image Extraction Using Histogram Equalization

1 is a facsimile system diagram.

2 is a conventional control flowchart.

3 is a histogram smoothing control flowchart of the present invention.

* Explanation of symbols for main parts of the drawings

10: CPU 20: original scanning unit

30: signal processing unit 40: recording unit

50: transmission unit

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a half-tome image extraction method in an image processing system such as a facsimile, and more particularly to a half-tone image extraction method using an improved histogram egualization method.

In general, in an image processing system such as a facsimile, a systematic dither method is widely used to express grayscale images. The above-described systematic dither method refers to a method of determining a threshold value of binarization as coordinate information displayed by the pixel. That is, by introducing a dither matrix, the density of the input image corresponding to the threshold value is compared and the black and white output is determined accordingly.

Recently, image processing systems such as facsimile have been looking for more advanced image processing methods to improve the resolution of picture images.

Accordingly, in the facsimile system schematically shown in FIG. 1, the flowchart of FIG. 2 is performed to represent the half-tone image. In this case, the image of the original document is controlled by the CPU 10. The digital data is converted into digital data by and input to the signal processor 30.

The CPU 10 stores the digital data input to the signal processor 30 in a memory of a RAM area.

At this time, the stored digital data are data having a value of 0-255 when quantized to 8 bits.

Therefore, the CPU 10 obtains a histogram of the stored digital data, which corresponds to step (2b).

After performing step (2b), the histogram is smoothed to obtain an inverse difference value for determining the gradation representation. This corresponds to step (2c) above.

After performing step (2c), the halftone image of step (2d) is output. Here, the gradation of the halftone image output in the step (2d) is often output as 16 gradations.

In other words, the CPU 10 smoothes the histogram to represent the input image stored in the memory as an image of a predetermined output gradation, and then applies the reverse difference values to the dither matrix.

A conventional method of extracting such halftone images is described in application number 16884 (hereinafter referred to as a prior patent application) filed on October 22, 1990 by the applicant of the present application.

The prior patent application has the advantage of extracting an excellent halftone image by smoothing the histogram of the input image and then applying a dither matrix, while the number of pixels obtained before the predetermined output gradation is mostly on the average number of pixels. If it is less than the last pixel number of the output gradation number, there is a problem that greatly exceeds the average pixel number.

In addition, if the number of pixels obtained before the number of output gradations exceeds the average number of pixels, the number of output gradations cannot be normally obtained.

Accordingly, an object of the present invention is to provide a half-tone image extraction method using an improved histogram smoothing method which can improve the resolution of a half-tone image in view of the above problems.

The present invention for achieving the above object is characterized in that to more evenly distribute the number of pixels of the input image.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a system diagram of a facsimile applied to the present invention, and includes a central processing unit (CPU) 10 that performs various operations by processing by a predetermined program and performs various operations of the system; A document scanning unit 20 for scanning an input document under control to generate an analog image signal, converting the generated analog image signal into digital image data, and outputting the analog image signal, and under the control of the CPU 10. A signal processor 30 which encodes and decodes the image data or the received image data output from the document scanning unit 20 to a threshold value of the extracted dither matrix, and an output image stored in the signal processor 30. The recording unit 40 for inputting data under the control of the CPU 10 and recording it on the thermal paper, and the image data stored in the signal processing unit 30 are controlled by the CPU 10. Transmitted to the other system via the network (PSTN), or consists of a transfer unit 50 for receiving from the other systems.

3 is a flow chart of a histogram smoothing control of the present invention, the first process of obtaining a histogram after storing image data scanned from an original or received from another image system and quantized to a predetermined bit in an array form in a memory; A second step of obtaining each pixel sum between neighboring brightnesses having no number of pixels from the histogram obtained in the first step and storing the sum of pixels in a predetermined area of the memory; A third process of setting two brightness values having the pixel sum to the same brightness level and then revising the pixel sum of the brightness values with an adjacent brightness, and having no pixel number of zero among the brightness values revised in the third process; A fourth process of repeating the third process until the number of brightness values becomes a predetermined output gradation number; When the same as the number of output gray levels is made the respective brightness value to obtain the true fifth step for setting the threshold value of the dither matrix.

Hereinafter, the present invention will be described in detail with reference to the flowcharts of FIG. 2 and FIG. 3 based on the above-described configuration. First, the present invention uses the improved histogram smoothing method for an input image to obtain respective threshold brightnesses of the dither matrix and then inputs them. It compares with the brightness value which each pixel of an image has, and determines whether it is "black" or "white".

In FIG. 1, the CPU 10 stores image data scanned from an original or quantized from another image system into predetermined rows in the form of M rows and N columns (MXN, where M and N are natural numbers) in a predetermined region of the memory.

Here, the document is converted into a digital image by the document scanning unit 20, and the analog image data received from another image system is converted into digital image data by the transmission unit 50.

The predetermined area of the memory refers to the RAM area of the signal processor 30. Subsequently, the CPU 10 obtains a histogram according to the density of the image data stored in the predetermined memory area in the signal processor 30.

Obtaining the histogram is an important task for the preprocessing of the image, and is to investigate how the density of each pixel value of the image to be processed is distributed.

The task for obtaining the histogram is performed by the processing of the CPU 10, and the concentration distribution of the telephone station in the M row and N columns is represented by the concentration frequency.

The process of obtaining the histogram corresponds to the first process.

The histogram obtained in the first step is used to obtain a predetermined number of threshold brightnesses to be used as the dither matrix threshold using the improved histogram smoothing method of the present invention.

The predetermined number of threshold brightnesses are typically output at 16 threshold brightnesses in 8-bit quantization.

Here, the threshold brightnesses are obtained by the second, third and fourth processes of the present invention.

When the threshold brightnesses are all obtained in the series of processes, the dither matrices correspond one-to-one and use the threshold values.

This corresponds to the fifth process.

Hereinafter, an embodiment of the operation will be described in detail with reference to the control flowchart of FIG. 3. For convenience of explanation, it is assumed that the histogram obtained in the first process is shown in Table 1 below, and the total pixel number is 160.

TABLE 1

When the eight gradations of the input image of Table 1 are output as four gradations by the histogram smoothing method, the average pixel number is 160 ÷ 4 = 40.

In step 3a of FIG. 3, data of Table 1 is already stored in a predetermined area of the memory.

The CPU 10 performs a second process of obtaining pixels between neighboring brightnesses having no number of pixels of zero from the stored data of Table 1 and storing them in a predetermined area of the memory.

This is the step (3b) → (3k) of FIG.

That is, in the second process, the number of pixels of Table 1 6 + 13 = 19, 3 + 41 = 54, 41 + 20 = 61, 20 + 35 = 55, 35 + 32 = 67, 31 + 7 = 38, 7 This is the process of finding + 6 = 13.

For example, if the number of pixels of the first brightness is not 13 but is 0, 6 + 41 = 47 is obtained.

In the third process, two brightnesses having the smallest pixel sum among the stored pixel sums are converted to the same brightness level, and the pixel sums of adjacent brightness values of the brightness values are revised. That is, in the third process, the smallest sum of pixels among the 19, 54, 61, 55, 67, 38, and 13 obtained in the second process is 13, and thus the brightness level having the sixth and seventh brightnesses of Table 1 is set. The sixth brightness is set to 13 and the seventh brightness is set to 0. This is the step (3l) → (3m) of FIG. Thereafter, if the number of non-zero brightness values in the step (3n) exceeds the predetermined number of four gray levels, looping is continued in step (3l) until the number is four, which is the fourth process. Therefore, when the fourth process is completed, threshold values of 19, 41, 55, and 45 are obtained. If four are output in step (3n), step (3o) is performed, which is the fifth process.

Here, four thresholds are generated when the above-described fourth process is performed. Therefore, if the effect is compared with the conventional method, since the average pixel number is 40 in the conventional method, only three threshold values having a value of 60, 55, and 45 are output, whereas four threshold values are generated in the present invention. It appears that the number was distributed more evenly. Therefore, it will be appreciated by those skilled in the art that the improved histogram smoothing method can be used in an efficient manner in an image processing system such as a copier or a facsimile.

Therefore, as described above, the present invention has an advantage of improving the resolution of halftone images using an improved histogram smoothing method.

Claims (1)

A method for extracting halftone images in an image processing system, the method comprising: obtaining an histogram after storing image data scanned from an original or received from another image system and quantized with a predetermined bit in an array form in a memory; A second process of obtaining pixel sums between neighboring brightnesses having no number of pixels from the histogram obtained in the first process and storing the sums of pixels in a predetermined corresponding area of the memory; and each stored in the second process. After the two brightness values having the smallest pixel sum among the pixel sums are set to the same brightness level, the pixel sum of the brightness values with the adjacent brightness value is revised in the third step and zero of the revised brightness values in the third step. A fourth step of repeating the third step until the number of brightness values having no number of pixels becomes a predetermined output gradation number; and And a fifth process for setting each of the obtained brightness values as threshold values of the dither matrix when the output gray level is equal to the output gray level in the fourth process.

Images