JPH04188283A - Method and device for processing picture - Google Patents

Abstract

PURPOSE:To prevent the deterioration of picture quality at edges by detecting the directions of edges in a picture and selecting a filter having an optimum characteristic according to the directions of the edges. CONSTITUTION:The binary picture supplied from a scanner, image memory, host computer or facsimile receiver, etc., is almost simultaneously inputted to all edge detection filters 1-4 and smoothing filters 7-10. The edge detection filters 1-4 are primary differentiating filters having specified coefficients, and absolute values are outputted as the outputs. In this case, the edge detection filters 1-4 are corresponding to the smoothing filters 7-10 and when the maximum values of the respective edge detection filters are respectively the filters 1-4, the correspondent one of the smoothing filters 7-10 is selected and outputted by a selector 11. Thus, the picture is processed without the deterioration of picture quality at edge parts.

Description

[Detailed description of the invention] [Industrial application field]
) The present invention relates to an image processing method and apparatus, and more particularly to techniques for restoring a multivalued image from a binary image and for processing other image data.

[Conventional technology]
4 Conventionally, when restoring a binary image to a multivalued image,
” is a smoothing filter using a smoothing filter, for example.
(The filtering results are multi-valued.)
Since it can be obtained using
'It was treated as such. Furthermore, a window is set up around the pixel of interest in the binary image, and the image patterns of "1" and "0" in the window are determined, and the image pattern is determined to be part of the character image. Ka, ah
Determine whether the blur is a halftone image and make that determination.
An image processing method is known that switches the output of different smoothing filters depending on the result, and in particular, when it is determined that the image is part of a character image, a smoothing filter with a relatively weak smoothing effect is used. An image processing method has been proposed to prevent character images from becoming blurred and multivalued.

[Problem to be solved by the invention]

However, in the conventional technology described above, the output of the smoothing filter is determined by simple bang recognition, which may lead to erroneous judgments.Also, since a small filter is used at the edges of the image, it is not possible to Problem 1 to be solved is that the number of gradations cannot be multi-valued, and on the other hand, even if a small smoothing filter is used for character areas and edge areas, it is still insufficient, resulting in deterioration of edges.
was there.

In view of the above-mentioned points, it is an object of the present invention to process image data without deteriorating image quality due to edges. The second objective is to provide an image processing method and apparatus that make it possible.

[Means for Solving the Problems] In order to achieve the above object, the method of the present invention detects edge components present in input image data, and adjusts the characteristics of a spatial filter according to the direction of the detected edge components. The method is characterized in that the image data is processed differently.

Further, in order to achieve the above object, the apparatus of the present invention includes an edge detection means for detecting edge components present in input image data, and a plurality of spatial filters having different shapes and coefficients set depending on the direction of the edge components. and a selection means for selecting one of the spatial filters according to the direction of the edge component detected by the edge detection means.

[For production]

In the present invention, the direction of an edge in an image is detected, and a filter with optimal characteristics is selected depending on the direction of the edge, thereby making it possible to process image data without deteriorating the image quality of edges. be able to.

〔Example〕

Embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 1 shows a circuit configuration of an image processing apparatus according to an embodiment of the present invention. In FIG. 1, a scanner (image reading device), an image memory, and a host computer.

A binary image supplied from a facsimile receiving device or the like is input to all edge detection filters 1-4 and smoothing filters 7-1O almost simultaneously. Edge detection filter 1~
4 is shown in Figure 2 (A), (B), (C), (
It is a first-order differential filter having coefficients as shown in D), and its output is an absolute value. The filters in FIGS. 2A to 2D detect vertical edges, horizontal edges, diagonal edges from upper right to lower left, and diagonal edges from upper left to lower right, respectively.

Outputs A and B of these edge detection filters 1 to 4.

The magnitude relationship between the outputs of C and D is determined by the comprehensive judgment unit 5, and the corresponding 2-bit code is determined depending on which of A, B, C, and D has the maximum output. The signal is output from the selector 11 and given as a signal for controlling the selector 11.

On the other hand, the edge detection filters 1, 2, and 3.4 above correspond to the smoothing filters 7, Jl, and 9.10, and the maximum value of each edge detection filter is filter 1, 2, and 3, respectively.
Smoothing filters 7, 8, 9, and 10 corresponding to the angle .degree. 4 are selected by the selector 11 and output. Smoothing filters 7.8 and 9.10 are shown in Figure 3 (A) and (B).
, (C) and (D).

In other words, when an image has many vertical edge components, the image shown in Figure 2 (
The filter output of A) is thick, and the filter shown in Figure 3 (A) is used to smooth the edges (multilevel conversion is performed, and a smoothing filter that is long in the vertical direction is used to obtain sufficient levels). Similarly, the image is as shown in Figure 2 (B).
When the image has many components on horizontal edges such as , the horizontally elongated smoothing filter shown in FIG. The output of the smoothing filter shown in FIG. 3(C) or (D) is selected by the selector 11 according to .

As another configuration example, if all the values of the edge detection filters 1 to 4 do not exceed a certain value, the output of the smoothing filter 13 as shown in FIG. 3(E) is selected by the selector 14. , it may be used as a final multilevel output.

FIG. 4 shows a specific internal configuration of the comprehensive judgment section 5 shown in FIG. In the comprehensive judgment section 5, the input A, B,
The data of C and D are combined with each other (six sets in total), and the comparators 21 to 26 judge the size of this combined data, and the judgment result is output in a judgment section 27. This judgment section 27 is a ROM ( read-only memory), R
Needless to say, this can be realized using a lookup table such as AM (Random Access Memory).

FIG. 5 shows an example of the internal configuration of each of the smoothing filters 7 to 1O shown in FIG.
4 and is input to the line buffer 41.

The image data is delayed by one line in each of line buffers 41 to 43, and then sent to each latch circuit 4.
5.46. . . 47 and is delayed by one pixel. The outputs of the latch circuits 44 to 47 are input to the next stage latch circuits 48 to 51, and are further delayed by one pixel. Similarly, the delay is repeated for each pixel, and the final delay is performed by the latch circuits 52 to 55.

Since the image data is supplied in rask order, data of a continuous rectangular area around the pixel of interest is obtained from the latch circuits 44 to 55 at the final stage, and a filter coefficient is calculated by the sum-of-products calculation unit 60 for the data of this rectangular area. A sum-of-products operation is performed.

FIG. 6 shows another embodiment of the present invention which realizes a circuit having the same function as the filter operation section 12 shown in FIG. The output of 61 is given to four parallel-connected product-sum calculation units 60-1 to 60-4, respectively. For this reason, all of the smoothing filters 7 to 10 shown in FIG. 1 can be used in common without having a line buffer section.

FIG. 7 shows another embodiment of the invention for obtaining the same effect as the embodiment of FIG. Edge detection filters 28 to 31
The outputs A, B, C, and D are processed by a preprocessing section 33 and then input to a smoothing filter 32 to obtain a final output. This smoothing filter 32 has a structure similar to the sum-of-products calculation section 60 in FIG.
B, C, and D are converted into filter coefficients in the pre-output section 33 and then provided to the sum-of-products calculation section 60 . The filter coefficients of the smoothing filter 32 are as shown in FIG. That is, depending on the intensity of each direction component of the detected edge, the coefficient in the direction in which that edge component remains becomes higher.

In addition, although omitted for the sake of simplifying the explanation, Figs. 3 (A), (B), (C), (D), (
E) The actual coefficients in j15 and FIG. 8 are each divided by the sum of each coefficient of each filter so that the sum of each coefficient after this division is 1.

In addition, in the above embodiment of the present invention, among the edges in the image,
Although a total of four edge directions (vertical, horizontal, and two diagonal directions) and four edge directions were determined, it goes without saying that this number is not limited and can be freely set according to the purpose.

As explained above, according to the above embodiment of the present invention, the edge direction in the image is detected, and a filter with the optimal shape is selected depending on the edge direction. This has the effect of enabling multivalued gradation with a large number of gradations and improving image quality.

That is, according to the above-described embodiment of the present invention, when restoring a multivalued image from a binary image, it is possible to preserve edge data and expand the filter size in a direction that does not cause edge deterioration. This has the effect of providing gradation reproducibility with a sufficient number of gradations during image processing.

Note that the input image is not limited to binary data, but may be multivalued data. Further, color data, particularly RlG, B, or color component data such as Y, M, C, etc., may be used.

Furthermore, the edge detection direction is not limited to the above example, and any appropriate matrix size, coefficient, etc. may be used.

Further, the spatial filter is not limited to the above-mentioned one, and the shape, coefficients, etc. can be arbitrarily set. Further, the filter is not limited to a smoothing filter, but may be an edge enhancement filter.

Further, it may be configured with a gate circuit or may be implemented as computer software.

INDUSTRIAL APPLICABILITY The present invention is suitable for use in image processing devices, particularly printers serving as output means for facsimile receiving devices, computers, copying machines, and the like.

〔Effect of the invention〕

As described above, according to the present invention, it is possible to process an image without deteriorating the image quality of the edge portion.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the circuit configuration of an embodiment of the present invention, FIGS. 2(A) to (D) are explanatory diagrams showing an example of the contents of the edge detection filter in FIG. ) to (E) are explanatory diagrams showing an example of the coefficients of the smoothing filter in FIG. 1, FIG. 4 is a block diagram showing an example of the circuit configuration of the comprehensive judgment section in FIG. 1, and FIG. FIG. 6 is a block diagram showing an example of the circuit configuration of a smoothing filter; FIG. 6 is a block diagram showing another example of the configuration of the filter operation section in FIG. 1; FIG. 7 is a block diagram showing a circuit configuration of another embodiment of the present invention. FIG. 8 is an explanatory diagram showing an example of coefficients of the smoothing filter of FIG. 7. 1-4... Edge detection filter, 5... Comprehensive judgment section, 6... Edge detection section, 7-IO... Smoothing filter, 11... Selector, 12... Filter calculation section, 13... Smoothing filter, 14... Selector, 21-26... Comparator, 27... Judgment unit, 28-31... Edge detection filter, 32... Smoothing filter, 33. ... Preprocessing section, 41-43... Line buffer, 44-55... Latch circuit, 60... Product-sum operation section, 61... Line buffer section, 60-1-60-4... - Product-sum operation section. Figure 4 Figure 5 Figure 7

Claims (1)

[Claims] 1) An edge component present in input image data is detected, and the image data is processed by changing the characteristics of a spatial filter according to the direction of the detected edge component. Image processing method. 2) The spatial filter is a smoothing filter, and when the direction of the detected edge component is vertical, the shape is vertically long, when it is horizontal, the shape is horizontally long, and when the detected edge component is in the diagonal direction, the shape is diagonal. 2. The image processing method according to claim 1, wherein the image data is smoothed using the smoothing filter having a long length. 3) If no edge component is detected in the input image data, the image data is smoothed using a large-sized smoothing filter in all vertical, horizontal and diagonal directions. The image processing method according to claim 2. 4) The image processing method according to claim 2, wherein the input image data is binary image data, and the smoothing filter performs multi-value conversion to obtain a predetermined halftone level. 5) edge detection means for detecting edge components present in input image data; a plurality of spatial filters having different shapes and coefficients set depending on the direction of the edge components; An image processing apparatus comprising: a selection means for selecting one of the spatial filters according to a direction. 6) The spatial filter is a smoothing filter, and the selection means has a vertically elongated shape when the direction of the edge component detected by the edge detection means is vertical.
5. The smoothing process of the image data is performed by selecting a smoothing filter having a horizontally long shape in the case of the horizontal direction, and selecting a smoothing filter having a shape long in the diagonal direction in the case of the diagonal direction.
The image processing device described in . 7) If the edge detection means does not detect the edge component, the selection means selects smoothing filters of large size in all vertical, horizontal and diagonal directions;
7. The image processing apparatus according to claim 6, wherein the image processing apparatus performs smoothing processing on the image data. 8) The image processing apparatus according to claim 6, wherein the input image data is binary image data, and the smoothing filter performs multi-value conversion to obtain a predetermined halftone level.