JPS6284390A - Area extraction system - Google Patents

Abstract

PURPOSE:To speedily repeat enlarging and reducing operations of an area at high accuracy and to extract a pattern area by repeating said operations by the prescribed times after a basic area is extracted. CONSTITUTION:When an extractable pattern area is inputted to an area enlargement and reduction part 2 by binarization processing, a control part 3, first, gives instructions that the area enlargement and reduction part 2 reduces the area Q-times, then gives m-times instructions that the part 2 enlarges the area, and finally gives n-times instructions that the part 2 reduces the area. Thus, the area enlargement and reduction part 2 sufficiently removes noise components by the reduction operations Q-times, after which it enlarges the area by the enlargement operations m-times to link the extracted areas, extracts a dropped out part, and finally returns the area to its original size by the reduction operations n-times.

Description

[Detailed Description of the Invention] [Summary] This is a method for extracting a region where a repeating pattern exists in an image in an image processing device. , it has become possible to remove noise and extract regions with repeating patterns, including unextracted regions, at high speed and with high precision.

[Industrial Field of Application] The present invention relates to an image processing apparatus, and particularly to a method for extracting an area where a repeating pattern exists in image processing.

[Prior Art] Conventionally, as a region extraction method for figure recognition, a range of colors or shading levels that are particular to the region is determined in advance, and then a range of colors or shading levels that are unique to the region is found in advance, and then a range of colors or shading levels that are within that range is extracted from the input image. Extract the area, i.e. “■” in binary code
This was to distinguish it from other areas.

[Problems to be Solved by the Invention] The above-mentioned conventional method is effective when the area to be extracted is uniform; When attempting to extract a certain area, there is a drawback that part of the area is regularly omitted depending on the pattern.

The present invention aims to provide a novel area extraction method that eliminates the drawbacks of such conventional methods.

[Means for Solving the Problems] FIG. 1 shows a block diagram of the principle of the region extraction method of the present invention.

In FIG. 1, reference numeral 1 denotes a basic area extraction unit, which sets a darkness value to the color or shading level unique to that area.
Extract pixels within this range.

For example, a dark value is set such that the level of R among the three primary color components R, G, and B is above a certain value, and the level of the G and B components is below a certain value, and pixels satisfying this value are extracted.

Reference numeral 2 denotes a region reduction/enlargement section which performs processing to reduce the region by deleting peripheral pixels of the region or to enlarge the region by adding pixels around the region according to instructions from the control port 3.

3 is a control unit that reduces the area extracted by the basic area extraction unit 1 0 times, enlarges the reduced area m times, and reduces the area again 0 times.
Give instructions to the magnifying section.

FIG. 2 is a diagram illustrating area reduction and enlargement processing.

FIG. 2(a) shows a reduction process for deleting pixels at the periphery of a region, and FIG. 2(a) shows an enlargement process for adding or adding pixels to the periphery of the region.

These processes are well known in the field of image processing.

The number of times Q, m, and n are reduced or enlarged is set depending on the pattern of the area to be extracted.

First, noise components having a small area are removed by repeating the process of deleting pixels in the peripheral area of the extracted basic region 0 times.

Next, the process of adding and adding pixels around the surviving area is repeated m times to enlarge the area and extract regular missing parts that could not be extracted by the binarization process.

Finally, by repeating the process of deleting pixels on the periphery of the enlarged area n times, the enlarged area is reduced and returned to its original size.

[Operation] The above configuration makes it possible to extract a repeating pattern area that is difficult to extract using conventional methods.

That is, extract the pattern part that can be extracted by binarization processing,
After sufficiently removing noise components through 0 reduction processes, the area is enlarged by m enlargement processes, the extracted areas are combined to extract missing parts, and finally the size is returned to the original size by n reduction processes. It is something to return.

[Example] The present invention will be explained in more detail with reference to an example shown in FIGS. 3 and 4.

FIG. 3 is a schematic diagram of an embodiment of the present invention.

In FIG. 3, the same symbols (1 to 3) as in FIG. 1 indicate the same objects.

4 is an input unit which captures an image with a television camera and converts it into a 512×512, 8-bit R, G, B digital image.

The basic area extraction section 1 is composed of a threshold value setting section 11 and a binarization circuit 12, and the darkness value setting section 11 selects one of several colors that make up the repetitive pattern to be extracted, The level range of the corresponding R, G, and H components is set as the dark value for binarization.

In the binarization circuit 12, the pixel value of the pixel whose level of each R, G, and B component of the input image is within the range of the darkness value set by the darkness value setting unit 11 is set to "1", and the pixel value of the other pixels is set to "1". The pixel value of “
Convert to 0”.

An image memory 5 stores images at each stage of processing.

The control unit 3 first issues an instruction to reduce the area to the area reduction/enlargement unit 2 Q times, then issues an instruction to enlarge the area m times, and finally issues an instruction to reduce the area n times.

When the area reduction/enlargement unit 2 receives an instruction to reduce the area from the control unit 3, it scans the image with a 3×3 window,
When at least one pixel within the window is "0", the pixel value of the pixel at the center of the window is set to "0", and at other times it is set to "1".

When an instruction to enlarge the area is received from the control unit 3, the image is scanned with a 3×3 window, and when at least one pixel in the window is “1”, the pixel at the center of the window is Set the value to “1′, and set the value to “1” at other times.
Processing to set the value to "O" is performed.

The number of reductions Q is set depending on the pattern of the area to be extracted so that noise is removed and the extracted basic area remains.

The number of enlargements m is set so that basic regions are combined with each other by region enlargement.

The number of reductions n is set to return to the original size, therefore, n=m-Q becomes.

FIG. 4 is a diagram illustrating processing according to an embodiment of the present invention.

FIG. 4(a) is an image captured by a television camera, and in this example, surveying rods ■, ■, ■ are to be extracted from the image.

The surveying rods ■, ■, ■ are colored alternately red and white, and the darkness value setting section sets the darkness value of the R component level to 10, for example.
The red part of the surveying rod is extracted by setting the darkness value of the levels of 0 or more and G and B components to 10 or less.

Figure (b) shows the result of extracting the red part of the image, which includes noise in addition to the surveying rod.

FIG. 4C shows that noise has been removed as a result of the reduction process Q times.

In the same figure, dl indicates a situation in which the regions have been expanded and combined as a result of the enlargement process m times.

FIG. 6(e) shows that as a result of the three reduction processes, an image of the surveying rod of the original size has been extracted.

[Effects of the Invention] As explained above, according to the present invention, noise can be removed and unextracted areas can be re-extracted using relatively simple processing logic, and repetitive pattern areas can be extracted at high speed and with high precision. Therefore, the practical effect is great.

[Brief explanation of drawings]

Fig. 1 is a block diagram of the principle of the present invention, Fig. 2 is a diagram illustrating area reduction and enlargement processing, Fig. 3 is a block diagram of an embodiment of the present invention, and Fig. 4 is a processing according to an embodiment of the present invention. FIG. In the drawings, 1 is a basic region extraction section, 2 is an area reduction/enlargement section, 3 is a control section, 4 is a human power section, 5 is an image memory,
Reference numeral 11 indicates a dark value setting section, and reference numeral 12 indicates a binarization circuit. (a) Chew ・1・ (b) 1 mu force 1 s or 0 syllable, l...
2 Illustrated book S - Sun and Moon no Mi Yamura Itayu 07" 177 illustration 3
Same (1)) CC)
(, t mail)
(0'Z Kokushi Shibei Ita-j by Ira-j of the Honkai Sun-Moon) Figure 4

Claims (1)

[Scope of Claim] A method for extracting a region of a regular repeating pattern in an image, comprising: a basic region extraction unit (1) for extracting a region whose color or shade level is within a predetermined range; an area reduction/enlargement unit (2) that reduces the area by deleting peripheral pixels of the area or enlarges the area by adding pixels around the area; and a preset number of basic areas extracted by the basic area extraction unit. After the reduction, the control unit (3) instructs the area reduction/enlargement unit (2) to enlarge the reduced area a preset number of times and to reduce the enlarged area again a preset number of times. , A region extraction method characterized in that it is configured to include.