JP2778669B2 - Halftone area detection method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial application field] In a copying machine or a facsimile machine, when reproducing an image in which a halftone image such as a photograph or a picture expressed by a halftone dot and a line drawing such as a character are mixed. In order to improve the quality of a reproduced image, it is desirable to perform a moire removal process on a halftone dot region such as a halftone photograph and a sharpening process on a line drawing region such as a character. Also, when transmitting an image in which a halftone image and a line image are mixed, in order to improve the data compression ratio, etc., the encoding process is performed after performing the process according to the characteristics of each image region. It is desirable to apply such as.

The present invention relates to a dot area detecting method for automatically determining and separating a dot area and a line drawing area in an image when performing the above-described various image processing.

[Conventional technology]

As a method for separating a dot area from an image in which halftone and line drawings are mixed, there is a method proposed by Ueno (Ueno: "Reproduction of a halftone picture using a dot printer" Oki Electric R & D Vol. 132 Vol. .53 No. 4).

In this method, as shown in FIG.
The original image data is taken out as a digital multi-tone signal by raster scanning, and the difference between light and dark density (gradation) between adjacent pixels on the raster is calculated to generate a difference signal. Pixels satisfying any one of the conditions i) to (iii) are detected as poles indicating peaks or valleys of density change as shown in FIGS. 8 (a) and 8 (b).

(I) The pixel when the sign of the difference changes before and after (FIG. 8A) is taken as the extreme point.

(Ii) A pixel when the sign of the difference value changes before and after the difference is zero (FIG. 8 (b)) is defined as an extreme point.

(Iii) A pixel (not shown) when the interval from the previous extreme point has reached a predetermined threshold value L _th0 is defined as an extreme point.

Based on the pole information obtained as described above, when the following conditions (iv) and (v) are simultaneously satisfied, it is determined to be a dot region.

(Iv) The section length L (i) between the poles is equal to a predetermined threshold L _th1 .
L is within the range of _th2 . That is, L _th1 <L (i) <L _th2 (v) The section length L (i) of the current position and the section length L _immediately before it
When the difference from (i-1) is within a predetermined threshold L _th3 .
That is, | L (i) −L (i−1) | ≦ L _th3 Then, based on the above determination result, the output of each pixel becomes a signal corresponding to a line drawing or a signal corresponding to a halftone dot.

[Problem to be Solved by the Invention] The above-described conventional method is to separate halftone dots from an image on the assumption that peaks of peaks and valleys of density levels appear regularly in a halftone dot region. Since there are many extreme points in a character portion other than a halftone dot portion and in a continuous tone photographic portion, there is a problem that a sufficiently high separation rate cannot be expected with the above method.

Further, since the detection is performed by one-dimensional comparison of pixels before and after the raster scan line, a halftone portion having a low halftone ratio, or a halftone portion having a high halftone ratio, or the document rotates. For example, in a halftone portion where the screen angle deviates from the horizontal direction, the section length L (i) between the poles becomes long, and there is a problem that separation from the character portion becomes difficult.

[Means for solving the problem]

In the present invention, a matrix composed of M × M pixels is sequentially applied to a digital multi-gradation input image signal obtained from an input image, and the gradation level of the center pixel of the matrix is the maximum or minimum level in the matrix. The central pixel is a peak or valley extreme pixel based on the difference between the average value of the gray level of the central pixel and the other two pixels symmetrically positioned with respect to the central pixel. The number of extreme pixels indicating peaks and the number of extreme pixels indicating valleys are counted in units of a two-dimensional area consisting of N × N pixels larger than M, and the total number of pixels detected as the extreme pixels is counted. The number of extreme pixels on the side where the numerical value is large is defined as the number of extreme pixels in the two-dimensional area, and the center of the two-dimensional area of interest is determined from the relationship between the number of extreme pixels in the two-dimensional area of interest and the number of extreme pixels in surrounding two-dimensional areas. Halftone dots Door and detection.

[Action]

Using a local two-dimensional pole detection pattern consisting of a matrix of M × M pixels to detect the poles of the pixels, further setting a two-dimensional area consisting of N × N pixels larger than M,
Since the center pixel of the two-dimensional area of interest and the surrounding two-dimensional area are detected as belonging to the halftone part, the one-dimensional comparison of the pixels before and after the raster scan line is performed. It is possible to more accurately separate halftone dots as compared with the conventional method of separating halftone dots only by using the above method.

〔Example〕

FIG. 2 shows an embodiment of a halftone dot area separating apparatus constructed by applying the method of the present invention.

For the sake of simplicity, a case where a monochrome image is used as a processed image will be described as an example. If you want to apply to a color image, you can decompose the color original image into three primary colors such as RGB or YMC according to the display format of the reproduced image, such as CRT display, print display, etc., and perform the dot separation processing according to the present invention for each color Good.

In FIG. 2, an input image signal 1 is a digital multi-level image composed of luminance signals corresponding to density (gradation) levels obtained by raster-scanning an original image in which a halftone image such as a halftone photograph and a line image such as a character are mixed. Is converted into an input image signal, and at least scan lines required for subsequent separation processing, for example, N
This is a circuit for storing in a line memory or the like the data of × 3 scan lines (N is the pixel size of a unit block composed of N × N pixels for halftone dot detection described later).

The pole detection unit 2 sequentially applies a matrix of predetermined M × M pixels to each pixel of the digital multi-tone input image sent from the input image signal unit 1 so that the center pixel of the matrix is This is a circuit for detecting whether or not it is an extreme point indicating a peak or a valley of a density change from a density relationship with surrounding pixels.

The following conditions (Ia) and (Ib) are employed as AND conditions as the extreme point detection conditions in the extreme point detection unit 2, and detection is performed using a local two-dimensional extreme point detection pattern.

As shown in pole detection condition (I a) Figure 3 (a) (b), M × M concentration level the M × M of the central pixel m ₀ matrix comprising pixels
When greater or less than the other concentration level of all the pixels m ₁ ~m _i in the pixel region, to detect the center pixel m ₀ as pole. That, m _0> m ₁ ~m _i or m ₀ <m ₁ ~m _i pole detection conditions (I b) as shown in FIG. 3 (a) (b), the center pixel of the matrix of M × M pixels When the absolute value of the difference between m ₀ , the density level, and the average value of the density levels of the other two pixels located symmetrically with respect to the center pixel is larger than the fixed threshold Δm _TH , the center pixel m ₀ is detected as an extreme point giving a peak or a valley of the density change. That is, when in the in center pixel m ₀ the other two pixels at the positions symmetrical m _a, and m _b, It is.

The pole detection unit 3 divides an image into blocks B each having N × N pixels (where N> M), for example, a block B having a 9 × 9 pixel size (N = 9) as shown in FIG. Then, the number of extreme pixels indicating peaks and the number of extreme pixels indicating valleys are counted for each block, and the number of extreme pixels on the side with the larger count value is determined as the number of extreme pixels of the block. the pole number of pixels P ₀ of the target block B _0, the center pixel n ₀ of the block of interest B ₀ from the relationship between the pole number of pixels P of the upper, lower, left and right diagonal of each surrounding blocks B ₁ .about.B ₈ surrounding the
(See FIG. 4), or all of the pixels n ₀ ~n ₈₀ of the said block B ₀ is a circuit determines whether belonging to the halftone dot region.

The halftone dot detection condition in the halftone dot area detection unit 4 is determined by the following conditions (IIa) to (IIa) depending on whether the processing is performed one pixel at a time or one block at a time.
Either of d) is adopted, and detection is performed using a local two-dimensional halftone dot detection pattern.

N × N block B as shown in Figure 5 when dot detecting a condition for executing a halftone separation process (II a) while moving pixel by pixel of the pixel, the block of interest B ₀ and surrounding blocks
In B _{1 to} B ₈ , when a block B in which the number of extreme pixels P is equal to or more than a predetermined threshold P _TH is equal to or more than a predetermined threshold B _TH ,
The center pixel n ₀ (see FIG. 4) of the block of interest B ₀ is detected as a halftone dot. That is, as shown in> B _TH dot detecting condition (II b) Figure 5 Number ΣB blocks of P> P _TH], the block of interest B ₀ and surrounding blocks
In B _{1 to} B ₈ , the block of interest B ₀ and each surrounding block B ₁ to B ₈
Sum of absolute values of the difference ΔP in the number of pole pixels from B ₈ Σ | ΔP |
Is less than or equal to the predetermined threshold value ΔP _TH , the center pixel n ₀ in the block of interest B _{0 is} defined as a halftone dot portion. That is, P | ΔP | <ΔP _{TH When the} dot separation processing is executed while moving the block B of the above L × L pixels one by one. Halftone dot detection condition (IIc) As shown in FIG. B ₀ and surrounding blocks
In B _{1 to} B ₈ , when a block B in which the number of extreme pixels P is equal to or more than a predetermined threshold P _TH is equal to or more than a predetermined threshold B _TH ,
All pixels n _{0 to} n ₈₀ in the block of interest B ₀ are detected as halftone dots.

Halftone detection condition (II d), as shown in FIG. 5, block of interest B ₀ and surrounding blocks
In B _{1 to} B ₈ , the block of interest B ₀ and each surrounding block B ₁ to B ₈
Sum of absolute values of the difference ΔP in the number of pole pixels from B ₈ Σ | ΔP |
There when it is less than a predetermined threshold value [Delta] P _TH, detects all pixels n ₀ ~n ₈₀ of the block of interest B within ₀ as halftone dot portion.

The area determination signal output unit 4 is a circuit that outputs a determination signal as to whether each pixel is a halftone part or a line drawing part based on the detection result of the halftone area detection unit 3.

The operation of the embodiment having the above configuration will now be described with reference to the flowchart of FIG. Note that the above (IIa) is used as a condition for detecting a halftone dot region in the halftone dot region detection unit 3. The input image is composed of a halftone portion and a line drawing portion, and does not include a continuous tone portion such as a continuous tone photograph.

The input image signal unit 1 raster-scans the original image and converts it into a digital multi-tone input image signal composed of luminance signals.
This image signal is stored for N × 3 scan lines required for the subsequent halftone dot separation processing (step [1]). For example, the fourth
As shown in the figure, when N = 9 is adopted as the pixel size of the block B of N × N pixels, N × 3 = 9 × 3 = 27
The image data for the scan line is stored.

The pole detection unit 2 sequentially applies, for example, a matrix of 3 × 3 pixels shown in FIG. 3A to each pixel of the pixel data stored in the input image signal unit 1, and performs the contact inspection condition (Ia). Center pixel of the matrix based on (I b)
It is determined whether or not m ₀ is the extreme point of the density change for all pixels on the 27 scan lines (step [2]).

When the contact detection is completed for all the pixels in step [2], in steps [3] to [11], the area determination as to whether or not each pixel is a halftone part is performed according to the halftone detection condition (IIa). Perform based on.

That is, first, in step [3], the image in which the extreme points are detected is divided in units of a block B having a size of 9 × 9 pixels as shown in FIG. The number of pixels is counted, and the number of extreme pixels on the side with the larger count value is set as the number of extreme pixels P of the block.

Next, in step [4], the number of the pole number of pixels P for the target block B ₀ and block B ₁ .about.B ₈ surrounding shown in FIG. 5 is equal to or greater than a predetermined threshold P _TH block Σ
B is obtained, and in step [5], the number of blocksΣB
Is greater than or _equal to a predetermined threshold BTH.

In step [5] above, the number of blocks ΣB is equal to the threshold
Central pixel n of the block of interest B ₀ when greater than B _TH
0 (see FIG. 4) is determined to be a halftone dot portion, and the process proceeds to step [6]. On the other hand, the center pixel n ₀ of the block of interest B ₀ when the number of blocks ΣB is smaller than the position B _TH is determined not to be a halftone dot portion, the process proceeds to step [7].

Area identification signal output unit 4 which has received the detection results from the dot region detecting unit 3, the center pixel n ₀ outputs a determination signal whether the meshed portion (Step [6] [7]) .

As described above, the halftone dot separation process is repeatedly performed for every pixel of the input image for every N × 3 scan lines (steps [8] and [9]), and the halftone portion and the line drawing are obtained for all the pixels of the input image. Separate into regions.

The above embodiment has been described by way of example when the original image is composed of halftone portions such as halftone photographs and line drawing portions such as characters. However, in actuality, continuous tone photographs and pictures are included in the original image. There are cases where continuous tone portions are mixed. In such a case, the method of the present invention may be applied after first differentiating the input image signal portion and removing the continuous tone portion from the edge density (see, for example, JP-A-58-115975). By doing so, it is also possible to separate the three parts, the halftone part, the character part, and the continuous tone part.

The value of the pixel size M of the matrix composed of M × M pixels and the value of the pixel size N of the block composed of N × N pixels can of course be arbitrarily adopted.

FIG. 6 shows an example of a copying machine constructed by using the halftone dot area separating apparatus of the above embodiment. In the drawing, the block indicated by reference numeral 5 is the halftone dot area separating apparatus shown in FIG. is there. In FIG. 6, reference numeral 6 denotes a sharpening processing circuit for character area processing;
Is a Bayer-type dither processing circuit emphasizing resolution for clearing characters, etc., 8 is a smoothing processing circuit for halftone dot area processing, and 9 is emphasis on gradation for producing halftones such as halftone pictures. The vortex-type dither processing circuit 10 is an image signal selection circuit that receives the halftone dot determination signal from the area separation device 5 and selects one of the dither processing circuits 7 and 9.

The input image signal is judged by the dot area separating device 5 as to whether each pixel belongs to a dot portion or not, and the judgment result is sent to the image signal selection circuit 10. The image signal selection circuit 10 selects and outputs a pixel signal from the dither processing circuit 7 when the determination signal is not a halftone area signal, and outputs a dither processing circuit when the determination signal is a halftone area signal. The image signal on the ninth side is selected and output.

As a result, an image signal such as a character explained by the dither processing circuit 7 is selectively output in the line drawing area, and an image signal such as a halftone picture processed by the halftone processing by the dither processing circuit 9 is selected in the halftone area. Is output. Therefore, if the image signal output from the image signal selection circuit 10 is subjected to necessary processing and then copied and reproduced, the line drawing portion such as a character is sharpened, and the dot portion such as a halftone photograph looks natural. A pseudo-halftone high quality binary image is obtained.

〔The invention's effect〕

According to the present invention, the extreme detection of the pixels of the digital multi-tone input image is performed using the local two-dimensional extreme detection pattern, and the obtained extreme detection result is further converted to a local two-dimensional network. Since a dot portion is detected using a dot detection pattern, a higher separation rate can be achieved compared with the conventional method, and a dot portion having a lower dot ratio or a dot portion having a higher dot ratio can be achieved. Further, even in the case where the screen angle is deviated from the horizontal direction due to the rotation of the original or the like, the halftone dots can be accurately separated.

Furthermore, when detecting the extreme point, not only the density level of the target central pixel and the density level of each peripheral pixel are individually compared, but also the density level of the central pixel and the central pixel are symmetrically positioned with respect to the central pixel. Since the extreme point is detected on condition that the absolute value of the difference between the average value of the density levels of the two pixels and the average value is larger than a certain threshold value, it is only necessary to compare the levels individually with the surrounding pixels. As a result, it is possible to provide a halftone dot area detection method that is more resistant to digital noise.

[Brief description of the drawings]

FIG. 1 is a flowchart of the operation of the embodiment of the present invention, FIG. 2 is a diagram showing an embodiment of the present invention, FIG. 3 is a diagram showing an example of a matrix composed of M × M pixels, and FIG. FIG. 5 is a diagram showing an example of a block composed of N pixels, FIG. 5 is a diagram showing the relationship between a target block and surrounding blocks, FIG. 6 is a diagram showing an example of a copying machine constituted by applying the present invention, FIG. FIG. 8 is an explanatory diagram of a conventional system, and FIG. 8 is a diagram illustrating an example of pole detection by the conventional system.

Claims (2)

(57) [Claims] 1. A matrix composed of M × M pixels is sequentially applied to a digital multi-gradation input image signal obtained from an input image, wherein a gradation level of a center pixel of the matrix is a maximum or minimum level in the matrix. And the other two pixels having a gray level and being symmetrical with respect to the center pixel.
The center pixel is detected as a peak pixel of a peak or a valley from the difference between the average value of the gradation levels of the two pixels and the pixel detected as the pole pixel is composed of N × N pixels larger than M. The number of extreme pixels indicating peaks and the number of extreme pixels indicating valleys are counted in units of two-dimensional areas, and the number of extreme pixels on the side with the larger count value is defined as the number of extreme pixels in two-dimensional areas. A halftone dot detection method, comprising: detecting a center pixel of the two-dimensional region of interest as a halftone dot portion from the relationship between the number of pixels and the number of extreme pixels in a two-dimensional region around the halftone dot. 2. A matrix consisting of M × M pixels is sequentially applied to a digital multi-tone input pixel signal obtained from an input image, wherein the tone level of the center pixel of the matrix is the highest or lowest level in the matrix. And the other two pixels having a gray level and being symmetrical with respect to the center pixel.
The center pixel is detected as a peak pixel of a peak or a valley from the difference between the average value of the gradation levels of the two pixels and the pixel detected as the pole pixel is composed of N × N pixels larger than M. The number of extreme pixels indicating peaks and the number of extreme pixels indicating valleys are counted in units of two-dimensional areas, and the number of extreme pixels on the side with the larger count value is defined as the number of extreme pixels in two-dimensional areas. A halftone dot area detecting method, wherein all pixels of the two-dimensional area of interest are detected as halftone dots from the relationship between the number of pixels and the number of extreme pixels in a two-dimensional area around the halftone dot area.