JPS59186467A - Image discriminating method - Google Patents

Abstract

PURPOSE:To discriminate automatically on whether an image is a dot image or not with high precision by discriminating the fact that the image is the dot image when the rate of black level picture elements when picture elements are sampled at reference intervals is larger than a specific value and the rate of white level picture elements is larger than a specific value. CONSTITUTION:The lateral axis in a figure represents picture elements on one scanning line and the longitudinal axis represents the level of the tentative binary-coding of each picture element by circuits S and D using a single fixed threshold value. The coordinate P=(a+b)/2 of the center picture element of a black-level picture element block B1 is calculated from the coordinate (a) of the starting picture element and coordinate (d) of the ending picture element of said block B1, and the coordinate Q= (c+d)/2 of the center picture element of a block-level picture element of a black-level picture element block B2 is calculated from the coordinate (c) of the starting picture element and coordinate (d) of the ending picture element of said block B2 to calculate the interval D=Q-P between the center picture elements from those P and Q. This D is regarded as a reference interval to sample picture elements at reference intervals from the center picture element (e.g. picture element at coordinate P), calculating the rate of block-level picture elements and the rate of white-level picture elements.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an image discrimination method for determining whether or not an image is a halftone dot image. This invention relates to a method for determining expressed halftone images.

(Prior art) A dither processing method using a dot concentration threshold matrix is
This is an effective method for reproducing pseudo-halftones using binary dots. However, in this method, the spatial frequency tends to be biased towards a specific component, so moiré patterns are likely to occur in the reproduced image. That is, when a halftone image is binarized using a dot concentration type threshold matrix, a moiré pattern is generated and the image quality often deteriorates significantly. This impression is especially hidden when we try to copy each generation. Therefore, when a halftone image is to be binarized, it is necessary to use a processing method that is less likely to produce moiré patterns.

By the way, it is difficult to judge just by looking at the document whether it has continuous or discrete gradation, or whether it is formed with halftone dots. Particularly for ordinary operators, a great deal of attentiveness is required when making this judgment.

However, in reality, an operator performs this image discrimination visually and selects a processing method that matches the discrimination result. For this reason, wasteful work due to discrimination errors is often performed. There is a demand for speeding up and improving the efficiency of this type of work, and copying machines equipped with automatic document feeders (ADF devices) are now widely used. In processing operations, it is practically impossible to respond to the above-mentioned increases in speed and efficiency.

(Object of the Invention) The present invention has been made in view of this point, and its purpose is to provide an image discrimination method that can automatically discriminate whether or not it is a halftone image with high reliability. be.

(Structure of the Invention) The image discrimination method of the present invention that achieves this objective binarizes pixels on one scanning line using a single fixed threshold, and from among the black level pixel blocks obtained by this, adjacent do 2
Select one black level pixel block, set a reference interval based on the interval between the two black level pixel blocks, and sample pixels at each reference interval from the center pixel of any black level pixel block. When the ratio of pixels is larger than a predetermined value, and the ratio of white level pixels obtained by reassigning pixels at each reference interval from the middle pixel of two arbitrary adjacent overlapping black level pixel blocks is larger than the predetermined value, The image is characterized in that it is determined to be a halftone image.

(Example) Hereinafter, a detailed description of the present invention will be given with reference to the drawings.

FIG. 1 is a diagram for explaining the discrimination principle of the present invention. The horizontal axis shows pixels on one scanning line, and the vertical axis shows the level when each pixel is temporarily binarized using a single fixed threshold. It shows. That is, the level of II B II indicates the black level, and “°W
'' level indicates the white level. Through this binarization, black level pixel blocks B+ to B4 and white level pixel blocks W and to W5 are constructed. In the method of the present invention,
Any two adjacent black level pixel blocks are selected from among these black level pixel blocks to perform image discrimination. For example, if 8+ and B2 are selected as two adjacent black level pixel blocks and tc, the coordinates of the center pixel of the black level pixel block B1 are given by the coordinates a of the first pixel and the coordinates of the last pixel of the black level pixel block B1. P=
(a + b), /2 is calculated, and similarly, the coordinate C of the first pixel and the coordinate d of the last pixel of the black level pixel block B2 are calculated.
From, the coordinates Q of the center pixel of the black level pixel block B2 =
(c + d)/2 is calculated, and from these P and Q, the distance between the center pixels D=Q-P is calculated. Each center pixel approximately coincides with the center of the halftone dot, and the distance is equal to the halftone dot distance. . Therefore, when pixels are sampled from the center pixel (for example, the pixel at coordinate P) at each reference interval using this D as a reference interval, the sampling interval is equal to the halftone dot interval, so the sampling pixel becomes the black level pixel. In many cases, this is the case. On the other hand, since the intermediate pixel between black level pixel blocks B+ and B2 (the coordinates R of this pixel is indicated by R= (P+Q)/2) is a pixel located at the intermediate position of the halftone dot, If pixels are sampled every time, the sampling pixels will be white level pixels in many cases. Therefore, in the above two samplings, the proportion of black level pixels in the former sampling is larger than the predetermined value, and the proportion of white level pixels in the latter sampling is thicker than the predetermined value (if It can be determined that the percentage of black level pixels is greater than or equal to a predetermined value, and the comparison of whether the percentage of white level pixels is greater than or equal to a predetermined value has the same meaning. For a specific comparison, pixels of either black or white level may be selected and their ratios taken.

FIG. 2 shows an example of a specific discrimination procedure that should be taken by a discrimination device that actually executes the above-described discrimination method. The determination operation when binary data as shown in FIG. 1 is input to this device will be described next.

First, as shown in Fig. 1, when an image take of a scanning line segment that has been provisionally binarized is received from an image memory or the like, the first and second
For the black level pixel blocks Bl and B2, the coordinates a and C of the pixels rising from the white level to the black level and the coordinates S and d of the pixels falling from the black level to the white level are detected, and p= (a +11 > /2.Q7 (c +
d) Calculate /2, and further, R= (P+Q), /2
．． Calculate D=(Q-P).

Next, pixels whose coordinates are p + n - D (n = O to n l) are sampled one after another, and when the pixel is at the desired level, the built-in counter is counted up to find the cumulative number of black levels. . Once this L is determined, a predetermined number L
○ and L are compared, and when L<LO, it is determined that the image is a non-halftone image.

On the other hand, when L≧L○, the coordinates are R+n −D (n=
The pixels indicated by o to n2) are sampled one after another, and when they are at the white level, they are counted up, the cumulative number of times M of the white level is determined, and it is compared to see if this is greater than a predetermined value M○. Then, if M < Mo, then
If it is a non-halftone image, Y is determined, and if M≧Mo, it is determined that it is a halftone image.

In this determination procedure, the black level pixel block B1 appearing at the beginning of the scanning line and the next black level pixel block B
2 is used as the reference interval (sampling interval), but in order to prevent the influence of noise etc. that appears at the beginning of the scanning line, even if it is determined that the image is a non-halftone image, the discrimination start pixel is shifted and the determination is repeated again. If the configuration is such that the determination is made and the image is finally determined based on the results of these determinations, the determination accuracy will be further improved. Also, if the center pixel ISI interval is determined for various adjacent black level pixel blocks and the value with the highest frequency is adopted, it is possible to obtain a white halftone dot.
The degree of discrimination in the case of

FIG. 3 is a block diagram of an image processing apparatus that binarizes image data by employing the above-described discrimination method.

In this figure, an image input section I generates a gradation signal corresponding to one scan of the original, and a binary image forming section S generates a tone signal corresponding to one scan of the original.

Send to D. The binary image forming section S binarizes the input signal using a single fixed threshold value, and the binary image forming section binarizes the input signal using a dot concentrated threshold matrix, and stores the input signal in the image memory Ml. , stored in M2. Next, using the image data in the image memory M1 that has been converted into a binary signal using a single fixed threshold value, the arithmetic processing unit C determines whether or not it is a halftone image. This discrimination algorithm is as described above. As a result of this discrimination, for images determined to be halftone images, 2
The value image forming unit S binarizes all input data using a single fixed threshold, and for images determined to be non-halftone images,
The value image forming section binarizes all input data using a dot-concentrated threshold matrix.

The image output unit O includes image memories Ml, M-2,
Image data is sent from an image memory in which optimally binarized data is stored, and the image is then reproduced as a soft copy or hard copy or transmitted to an external device. In this case, it goes without saying that a moire pattern will occur in the reproduced image.In addition, when converting all the data input to the binary image forming section S into binarization, a random dither threshold value or a dot-distributed threshold matrix is used. May be used.

(Effects of the Invention) As described above, according to the present invention, it is possible to automatically determine whether an image is a halftone image or not with high reliability.

[Brief explanation of the drawing]

Fig. 1 is an explanatory diagram of the principle of the method of the present invention, Fig. 2 is an explanatory diagram showing an example of a specific discrimination procedure of the method of the present invention, and Fig. 3 is an image obtained by binarizing an image signal using the method of the present invention. FIG. 2 is a block diagram of a processing device. ■...Image input section S, D...Binary image forming section Ml, M2...Image memory C...Arithmetic processing section O...Image output section Patent applicant Konishiroku Photo Industry Co., Ltd.

Claims (2)

[Claims] (1) In an image discrimination method for determining whether an image is a halftone image, pixels on one scanning line are binarized using a single fixed threshold, and the black level of the resulting black level pixel block is , two adjacent black level pixel blocks were selected, a reference interval was set based on the interval between the two black level pixel blocks, and pixels were sampled at each of the reference intervals from the center pixel of an arbitrary black level pixel block. when the proportion of black level pixels is larger than a predetermined value, and when pixels are sampled at each reference interval from the middle pixel of any two adjacent black level pixel blocks, the proportion of white level pixels is larger than a predetermined value. An image discrimination method characterized in that the image is determined to be a halftone image. (2) Selecting the interval between the center pixel of the first black level pixel block and the center pixel of the second black level pixel block adjacent to the first black level pixel block as the value of the reference interval. An image discrimination method according to claim 1, characterized in that: