JP2790525B2 - Linear approximation processor - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Summary] The present invention relates to a linear approximation processing device that extracts a linear component corresponding to a linear state represented by a binary image based on the binary image and generates parameters of the linear line. The purpose of the present invention is to reduce the amount of processing in the straight-line approximation processing unit, and to extract the so-called outline portion of the given binary image from the given binary image. An image processing unit configured to reduce the number of pixels without destroying the image is provided, and a straight line parameter is generated using all of the binary images of the number of pixels reduced by the image processing unit. .

[Industrial applications]

The present invention relates to a straight line approximation processing device that extracts a straight line component corresponding to a straight line state represented by a binary image based on the binary image and generates parameters of the straight line.

In image processing, it is one of important technologies to extract the direction of an object or the boundary between different objects. Since the input image contains a lot of noise, it is necessary to linearly approximate a plurality of points on the object and a plurality of points on the boundary to find the direction and the boundary of the object.

In particular, in the field of unmanned mobile vehicles, it is essential to detect a straight line along which the mobile vehicle should travel.

[Conventional technology]

 FIG. 7 shows the configuration of a conventional linear approximation processing device.

In the figure, a binary image input unit 1 inputs a binary image.
The binary image storage unit 2 stores the binary image input from the binary image input unit 1. The linear approximation processing unit 3 is a binary image storage unit 2
Is approximated to a straight line, and the result is stored in the approximate straight line storage unit 4 as an approximate straight line parameter.

The straight-line approximation processing unit 3 can perform a straight-line approximation by using the least squares method, the Hough transform, the straight-line approximation method, and the like, and calculating the coordinate positions of all the pixels to be processed. The Hough transform uses a space representing the parameters of a straight line for approximation calculation, and is described, for example, in "Introduction to Computer Image Processing" (Soken Publishing), pp. 127-128.

[Problems to be solved by the invention]

In the conventional case, the linear approximation processing unit 3
Since the processing is performed using the value image as it is, the amount of calculation for performing the linear approximation is large, and there is a problem that it takes too much time for the processing. That is, in general, the amount of computation is
This is because the larger the number of pixels to be processed, the larger the number.

An object of the present invention is to reduce the amount of processing in a straight line approximation processing unit.

[Means for solving the problem]

FIG. 1 shows a principle configuration diagram of the present invention. In the figure, reference numeral 1 denotes a binary image input unit, 2 denotes a binary image storage unit, 3 denotes a straight line approximation processing unit, 4 denotes an approximate straight line storage unit, and 5 denotes an image processing unit.

The difference from the conventional configuration (shown in FIG. 7) is that an image processing unit 5 exists. The image processing unit 5 performs a process of reducing the number of pixels for the binary image given and stored in the binary image storage unit 2 without destroying the straight line approximation of the binary image. Do.

(Operation)

The binary image input unit 1 inputs a binary image. The binary image storage unit 2 stores the binary image input from the binary image input unit 1. The input binary image once stored in the binary image storage unit 2 is sent to the linear approximation processing unit 3 after being processed by the image processing unit 5 into an image having fewer pixels. The straight line approximation processing unit 3 extracts a straight line component and generates a parameter corresponding to the straight line. Then, the parameters are stored in the approximate straight line storage unit 4.

〔Example〕

2 (A), 2 (B), 2 (C), and 2 (D) are diagrams for explaining three types of processing modes in the image processing unit.

FIG. 2A shows a given binary image 10, and the black circles in the figure correspond to the black pixels of the original image.

In FIG. 2 (B), with respect to an odd line when horizontal scanning is performed on the binary image 10 shown in FIG. 2 (A), a pixel corresponding to the start end on the scanning line
A process is performed to leave the pixel 11 corresponding to the end on the scan line with respect to the even-numbered line.

Reference numeral 13 denotes a straight line component extracted by the straight line approximation processing unit 3 based on the result of performing the processing as shown in FIG. 2 (B).

In FIG. 2 (C), a logical filter using eight neighboring contours is applied to the binary image 10 shown in FIG. 2 (A). That is, the logical filter using the eight neighboring contours in the case shown in FIG. 2 (C) is used for the upper, lower, left, right, diagonally upper right, diagonally lower right, diagonally lower left, diagonally lower left, and diagonally lower left centered on the black pixel of interest. This is a filter that considers the eight pixels and performs processing to leave the pixel of interest when one or more of the eight pixels are not black pixels. Reference numerals 14 and 15 shown correspond to the remaining pixels, respectively, and 16 represents a straight line component.

In FIG. 2D, a logical filter using a 4-neighborhood contour is applied to the binary image 10 shown in FIG. 2A. That is, the logical filter using the four neighboring contours in the case shown in FIG. 2 (D) considers four pixels (upper, lower, left, right) centered on the black pixel of interest and considers the four pixels. This is a filter that performs a process of leaving the pixel of interest when at least one of the pixels is not a black pixel. Symbol 1 shown
7, 18 correspond to the remaining pixels, respectively, and 19 represents a straight line component.

In the extraction processing using the logical filter, for example, shift registers are arranged in three columns corresponding to three scanning lines.
The problem is solved by a well-known technique such as taking in nine or five pieces of pixel information including the pixel of interest.

FIGS. 3A and 3B are diagrams for explaining the other three processing modes in the image processing unit.

FIG. 3A is similar to the case of FIG. 2A,
1 shows a given binary image 10.

FIG. 3B shows an example in which a logical filter using four neighboring contours is used as in the case of FIG. 2D. However, the logical filter used in FIG. 3 (B) considers four pixels, that is, the upper, lower, left, and right pixels of the pixel of interest, and if the pixel to the left of the pixel of interest is not a black pixel, This is a filter that leaves the pixel of interest. The black pixels 101, 102, 103, 104 shown in FIG. 3B correspond to the black pixels 101, 102, 103, 104 shown in FIG. 3A.

When the straight-line approximation processing unit 3 performs straight-line approximation based on the black pixels shown in FIG. 3B, a straight-line component as indicated by a dotted line 20 is extracted. Therefore, when the processing shown in FIG. 3 (B) is performed, correction as described later (FIG. 5) is performed to extract a desired linear component 21.

4 (A) and 4 (B) are explanatory diagrams of processing related to the processing shown in FIG. 3 (B).

In the case shown in FIG. 4A, 2 in FIG.
This shows a case where noises 23 and 24 are present in the value image 10.

The binary image 10 'shown in FIG.
When the same processing as shown in FIG. 4B is performed, a part of the noise may remain as shown in FIG. 4B, and the linear component 20 'is extracted under the influence of this. Then, the corrected linear component 21 'is obtained. The fourth
The black pixels 101, 102, and 109 shown in FIG. 4B correspond to the black pixels 101, 102, and 109 shown in FIG. 4A.

FIG. 5 shows another embodiment of the present invention. Reference numerals 1, 2, 3, 4, and 5 in the figure correspond to FIG. 1, 6 indicates an approximate straight line correction unit, and 7 indicates an approximate straight line relation unit.

The configuration shown in FIG. 5 corrects the straight line components 20, 20 ', 20 "extracted by the straight line approximation processing unit 3 as shown in FIG. 3 (B) and FIG. , 21 ′, 21 ″.

The approximate linear relation unit 7 calculates the correction amount in the case shown in FIG. 3B and the correction amount in the case shown in FIG. 4B from the line width of the original binary image 10 and the pixel interval. , Are obtained in advance and held. Then, the approximate straight line correction unit 6 performs desired position correction on the straight line components 20, 20 ', 20 "extracted by the straight line approximation processing unit 3 to obtain straight line components 21, 21', 21".

FIG. 6 shows a processing mode for obtaining a plurality of linear components.

When a binary image 10 "as shown in FIG. 6 (A) is present, a binary filter as shown in FIG. 6 (B) is used by using a logical filter which performs processing as shown in FIG. 3 (B). An image is obtained.

In addition, regarding a logical filter having a 4-neighborhood contour as shown in FIG. 3 (B), when the pixel above the black pixel of interest is a black pixel and the other lower, left and right pixels are black pixels, A binary image shown in FIG. 6 (C) is obtained by using a logical filter that leaves pixels to be processed.

〔The invention's effect〕

As described above, according to the present invention, the parameters are generated using the coordinate positions of all the pixels in the binary image with the reduced number of pixels. For this reason, during processing in the image processing unit, undesired noise components are reduced without destroying the linear approximation of the original binary image, and in the processing in the linear approximation processing unit, The adverse effects introduced forcefully in the treatment can be eliminated as far as possible.

[Brief description of the drawings]

FIG. 1 is a diagram showing the principle of the present invention, and FIGS. 2 (A) and 2 (B).
FIGS. 3 (C) and 3 (D) are diagrams for explaining three examples of processing modes in the image processing unit. FIGS. 3 (A) and 3 (B) are diagrams for explaining other example processing modes in the image processing unit. Figure (A)
(B) is an explanatory diagram of a process related to the process shown in FIG. 3 (B), FIG. 5 is an embodiment diagram, FIG. 6 is a diagram showing a processing mode for obtaining a plurality of line segments, and FIG. 1 shows a configuration of a straight line approximation processing device. In the figure, 1 is a binary image input unit, 2 is a binary image storage unit, 3 is a straight line approximation processing unit, 4 is an approximate line storage unit, 5 is an image processing unit, 6 is an approximate line correction unit, and 7 is an approximate line. The related part 10 represents the original binary image.

Claims (2)

(57) [Claims] A linear approximation processing unit for extracting a linear component included in a binary image (10) based on the binary image (10) and generating a parameter of a linear line corresponding to the linear component. In the linear approximation processing device having (3), when the binary image (10) is viewed in a predetermined coordinate direction with respect to the binary image (10), the leading edge of the binary image (10) The corresponding pixels (11,14,17) and / or the pixels (12,15,15) corresponding to the trailing edge of the binary image (10)
By extracting only 18), an image processing unit (5) for performing a pixel number reduction process for reducing the number of pixels is provided. The linear approximation processing unit (3) is reduced by the image processing unit (5). A linear approximation processing device characterized in that the parameters are generated using the coordinate positions of all pixels in an image having the number of image data. 2. An approximation straight line correction unit (6) for correcting the approximation line obtained from the approximation line obtained by the line approximation processing unit (3). The straight line approximation processing device according to claim 1, wherein the position correction accompanied by the parallel movement is performed on the approximate straight line obtained by the straight line approximation processing unit (3).