JPS63113780A - Peak extracting method - Google Patents

Abstract

PURPOSE:To make the processed result in the next stage obtained by using a peak value accurate by outputting a picture element in a range having a specific density level neighboring the peak value as the picture element having the peak value. CONSTITUTION:When density level data is read out from a picture memory, which stores the density level data by the unit of picture element, by the unit of picture element and the density of a notice element is compared with that of a neighboring picture element so as to extract the picture element having the peak value of the density level, the density level is obtained by subtracting the specific level from the obtained peak value and the range of the picture element having the obtained density level is extracted with the picture element near the picture element having the peak value so as to extract the picture element in the extracted range as the picture element having the peak value. Thus, the processed result in the next stage obtained by using the peak value can be made accurate.

Description

Detailed Description of the Invention] ([Essential] In a peak extraction method for extracting a pixel having a peak density level by comparing the density of a pixel of interest and neighboring pixels, By outputting the pixel as a pixel having a peak value, the result of the next stage processing using the peak value is made more accurate.

[Industrial application field]

The present invention provides a peak extraction method that reads out density level data pixel by pixel from an image memory that stores density level data pixel by pixel, and extracts a pixel having a peak density level value by comparing the density of the pixel of interest and neighboring pixels. Regarding.

In recent years, object recognition technology has been actively developed to enable machines to perform functions similar to those of the human eye. In such object recognition, a method is used in which the object is captured as an image and information such as the object's characteristics is extracted from this image to recognize the object. For this reason, it is generally necessary to image the object with an imaging means such as a television camera, capture the original image, and perform electrical processing based on this image to extract information.

On the other hand, such original images often contain noise, blur, or blur, and it is necessary to remove these noise components and extract structural lines and the like that constitute the original image.

As a method for extracting such an image, an information extraction method using mapping is conventionally known. As an information extraction method using this mapping, a method using spherical mapping (Japanese Patent Application Laid-Open No. 60-21
8183, etc.) and methods using Hough transform are known.

In the case of using spherical mapping, the original image is projected onto a spherical surface (spherical projection), and each point P of the original image projected onto the spherical surface is centered at that point, as shown in Figure 4 (a). large circle C
is drawn, this is expanded on the mapping plane as shown in Figure 4 (bl), and information is extracted from the mutual relationship of the drawn great circle group. As a result, the normal vector of the plane containing each point of the original image and the center of the spherical surface can be obtained, and line segments can be extracted at the same time, and it is also possible to reproduce broken or distorted line segments. By mapping change information to a point on a spherical surface, the trajectory of a point-like object moving in a straight line is extracted.
The projections of a plurality of line segments on the spherical surface are each mapped to a point on the spherical surface, and then by mapping these mapped points, the original intersection of the line segments is reproduced.

Also, the one using Hough transform is shown in Fig. 5 (8).

(As shown in bl, point P on the image, (X, ,Y,
”) to ρ= Xn sin ψ+Yncos ψ=π
τ lower y, t (θ+arctan (Yn /X, l))
Convert it to a curve according to the formula. Then, a curve is drawn for each point of the original image projected onto a plane, and information is extracted from the mutual relationship of the drawn curved parts.

This allows information extraction processing similar to that of spherical mapping to be performed.

By the way, as mentioned above, in this mapping process, the intersection point of each mapping function on the mapping memory in which the mapping function is written during information extraction is determined by the so-called peak value, which is the highest density value (histogram value) of the pixel. Extract as a pixel with .

[Conventional technology]

The peak extraction method for extracting pixels with high density values uses a 3x3 matrix window to calculate the density of the pixel of interest for each pixel written in the mapping memory as a result of mapping by the mapping processor. If the value is higher than or equal to the density values of surrounding pixels, it is extracted as a peak value.

[Problem that the invention seeks to solve]

By the way, for example, if the original image for generating the mapping function is a short line segment or a line segment with low density,
As a result of the mapping, the density values near the intersection of the mapping functions written on the mapping memory are as shown in FIG.

Therefore, when the above-described peak extraction process is performed, the pixel at the position indicated by arrow A in FIG. 2 is extracted as a pixel having a peak value.

However, considering the shape of the concentration histogram,
The pixel at the position indicated by arrow B in the figure seems to be the pixel having the peak value.

In this way, the position of the pixel with the peak value is extremely unstable, and even though the position of arrow B is actually the position of the extracted intersection, if the position of arrow A is extracted as the intersection, the following When performing stage mapping, there was a problem in that the degree of integration of mapping functions such as great circles was poor, making it impossible to obtain correct integration points (intersection points) and making it impossible to extract accurate information.

SUMMARY OF THE INVENTION In view of the above-mentioned conventional drawbacks, an object of the present invention is to provide a peak extraction method that does not extract erroneous peak value positions and that makes it possible to obtain accurate results in the next stage of processing using peak values. On offer.

[Means for solving problems]

An object of the present invention is to read out density level data pixel by pixel from an image memory that stores density level data pixel by pixel, and to extract a pixel having a peak density level value by comparing the density of the pixel of interest and neighboring pixels. In the extraction method, a density level is obtained by subtracting a predetermined level from the obtained peak value, and in pixels near the pixel having the peak value,
This is achieved by a peak extraction method characterized by extracting a range of pixels having the density level determined by the means, and extracting the pixels in the extracted range as pixels having a peak value.

That is, in the present invention, pixels near arrow A in FIG.
Pixels with a density close to this pixel A are processed in the next stage (e.g.
By performing neighborhood re-extraction processing to output as pixels for mapping (mapping), the shaded area near the peak value in the same figure is re-extracted, thereby preserving the shape of the peak and inheriting it to the next stage of processing. do. By allowing ambiguity in this way, it becomes possible to collect a large amount of ambiguous information and develop it into reliable information in the next stage of processing.

〔Example〕

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

FIG. 1 is an overall configuration diagram of a spherical mapping device to which the present invention is applied.

In the figure, 30 is a microprocessor (hereinafter referred to as MPU), 31 is a spherical camera, 33 is a matrix calculation unit, 34 is a peak value storage table, 35 is an original image memory, 3
6.37 is a mapping processor, and 38.39 is a mapping memory.

The MPU 30 has a spherical camera 31. Matrix calculation unit 33.
The spherical camera 31 photographs the object and outputs the image data after contour extraction processing in polar coordinate (r, θ) format.The original image memory 3
Reference numeral 5 stores image data in polar coordinate format output from the spherical camera 31 in association with each other in a grid pattern, N pieces in the longitudinal direction.

It has M storage areas divided in the latitudinal direction, the area near the equator is stored in one-to-one correspondence, and the area near the north pole is stored in one-to-multiple pixel correspondence. It performs 3 matrix operations, for example,
The pixel of interest E and the four pixels surrounding this pixel E, B, D, F,
A 3×3 matrix is set with H, and the values of the pixel E of interest and each surrounding pixel B-H in this matrix are compared, and the value of the pixel of interest E is larger than the value of each surrounding pixel B-H. If they are equal, the pixel E of interest is determined to be the peak point, and the peak value storage table 34 stores the coordinates and density values of the pixel having the peak value extracted by the matrix calculation unit 33 in the format shown in FIG. The mapping processor 36 generates a spherical mapping function (great circle information) for the image data of each pixel stored in the original image memory 35, and sequentially stores the mapping results in an internal mapping memory. The mapping memory 38 is used to transfer and store the peak points (intersection points) that are the mapping results in the mapping processor 36, and the mapping processor 37 stores the great circle situation for the pixels having the peak points stored in the mapping memory 38. mapping memory 3, which generates the mapping result and sequentially stores the mapping result in the internal mapping memory.
Reference numeral 9 denotes a peak point which is a mapping result in the mapping processor 37 and is transferred and stored therein.

The operation of the configuration described above will be explained.

■ First, the spherical camera 31 photographs the object on the control string of the MPU 30 and outputs image data in polar coordinate (', θ) format, and the MPU 30 receives the image data (coordinate position and density value) from the spherical camera 31. The data is stored pixel by pixel in the corresponding storage area of the original image memory 35.

(2) The MPU 30 issues a command to start the mapping processor 36 when the image data from the spherical camera 31 is stored in the original image memory 35.

- Upon receiving the activation command, the mapping processor 36 sequentially generates a mapping function (great circle information) for the pixels stored in the original image memory 35 (each pixel having the shaded area in FIG. 2) and stores it in the internal mapping memory. Add and store the value.

(2) Once the great circle information corresponding to each pixel in the original image memory 35 is stored in the mapping memory in the mapping processor 36, the MPU 30 reads the histogram value for each pixel from the mapping memory in the mapping processor 36, Peak extraction processing is performed using the matrix calculation section 33.

■ This peak extraction process is performed under the control of the MPU 30. First, the image data at the coordinates (r, 4.θ.) of the mapping memory in the mapping processor 3G is used as the value of the pixel of interest E. Then, image data at coordinates adjacent to the image data in the vertical and horizontal directions are extracted as values of surrounding pixels D to 11.

■ Then, the value of each pixel is calculated by the matrix calculation unit 33.
Set the 3×3 matrix in the matrix calculation section 3.
3, the peak extraction process is performed by determining whether the value of the pixel of interest E is greater than or equal to the value of the surrounding pixels DH based on the set value, and the pixel having the extracted peak value is The coordinates and density values (eg, 256 gradations) are stored in the peak value storage table 34 in the format shown in FIG.

(2) Thereafter, the processes of (1) and (2) are sequentially performed to extract the peak value of the image data in the mapping memory in the mapping processor 36.

■ Then, after the peak extraction process is completed, the MPU 30
The peak values stored in the peak value storage table 34 are sequentially read out, and a value obtained by multiplying the stored peak value by 0.8 is obtained and stored in the internal memory of the MPtJ 30 as a subtraction level value.

- Based on the read peak value coordinates, the MPU 30 sets a 9×9 matrix range (Fig. 2) centered on these coordinates (coordinates of the pixel in the part indicated by arrow A in Fig. 2), and sends the map to the mapping processor 36. Image data of pixels within this range is read from the mapping memory within.

[Phase] Next, the image data of the pixels in the read range is sliced using the subtraction level value obtained in the above 0 term, pixels having a value equal to or higher than this subtraction level value are extracted, and the matrix calculation unit 3
This extracted pixel is added to the pixel having the peak value extracted in step 3, and these are added to the pixel having the peak value (
The mapping memory 3B
Write to.

■ From now on, the above ■, [
[phase] term is processed.

0 Next, the mapping processor 37 generates great circle information for each pixel having a peak value stored in the mapping memory 38 (each pixel having a shaded area in FIG. 2), and stores it in the internal mapping memory.

@ The MPU 30 performs the same processing as 0 to 0 described above on the great circle information stored in the mapping memory in the mapping processor 37 and stores it in the mapping memory 39 .

As explained above, in this embodiment, for a pixel having an obtained peak value, neighboring pixels having a density value obtained by subtracting a predetermined level from the peak value are also extracted as pixels having the peak value. Since the mapping processor 37 performs mapping based on such ambiguous information, the resulting mapping result becomes information on these intersection points, and as a result, it becomes possible to obtain reliable information.

In this embodiment, an example of spherical mapping has been described, but the present invention is not limited to this, and can be applied to any process in which the next stage of processing is performed based on the result of peak extraction. Needless to say, it is possible.

Furthermore, although the spherical mapping process is explained using two mapping processors, it is not limited to this, and the number can be increased as appropriate depending on the type of information to be extracted. The mapping result of the mapping processor is
It goes without saying that the present invention can be applied when storing data in a mapping memory.

〔Effect of the invention〕

As described above, according to the present invention, in a peak extraction method for extracting a pixel having a peak value of a density level by comparing the density of a pixel of interest and neighboring pixels, a range having a predetermined level of 1 degree near the peak value is provided. Since the pixels are output as pixels having a peak value, it is possible to obtain accurate results in the next stage of processing using the peak value.

[Brief explanation of the drawing]

Fig. 1 is a block diagram of the configuration of the spherical mapping device, Fig. 2 is an explanatory diagram of the peak center, Fig. 3 is an explanatory diagram of the peak value storage table, Fig. 4 is an explanatory diagram of the principle of spherical mapping, and Fig. 5 is an explanatory diagram of the peak center. FIG. 3 is a diagram explaining the principle of conversion. In the figure, 30 is an MPU, and 31 is a spherical camera. 33 is a matrix calculation unit, 34 is a peak value storage table, 35 is an original image memory, 36.37 is a mapping processor, 3
8.39 is a mapping memory. Patent applicant: Director of the Agency of Industrial Science and Technology Kozo Iizuka 1st [1n photo'JJtJ/) The configuration is 77" between the locks 1st (!
] Hi0-ri center nt Umei Asu 2 Figure Bifui Gokaku Funai Chi 7゛1 Shi no Sho V ni r Hippo 14yo Figure 3 (a) (b) Explanation of the principle of spherical images Hough exchange Thread diagram'$ 5 (2)

Claims (1)

[Scope of Claims] A peak system that reads out density level data pixel by pixel from an image memory that stores density level data pixel by pixel, and extracts a pixel having a peak density level value by comparing the density of the pixel of interest and neighboring pixels. In the extraction method, a density level is obtained by subtracting a predetermined level from the obtained peak value, and a range of pixels in the vicinity of the pixel having the peak value and having the density level obtained by the means is extracted, and the extraction 1. A peak extraction method characterized by extracting pixels in a range of pixels as pixels having a peak value.