JPH09114990A - Image recognizing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately identify a circle or an ellipse without being affected by noise, etc., by finding the center of the circle or the ellipse from its edge information. SOLUTION: A candidate area extracting part 2 extracts the candidate area of an object from a binary image signal 1, and an edge information extracting part 10 extracts the edge of the circle or the ellipse in the candidate area and calculates its center. A mesh data calculating part 3 divides the inside of the candidate area into meshes and holds the number of black picture elements in the respective meshes in a memory 4. A feature amount calculating part 5 prepares feature amount data for which mesh data on the respective circumferences are arranged along plural centrifugal circles from the center of the circle or the ellipse inside the candidate area. A rotating angle detection part 6 detects a rotating angle by performing matching with the feature amount data of the same radius in a dictionary 7. An object recognizing part 8 recognizes the image based on the degree of certanity in the rotating angle.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image recognition method for recognizing a specific image from color images and monochrome images.

[0002]

2. Description of the Related Art The applicant of the present invention has previously proposed an image recognition method in which the recognition rate of an object is improved by detecting the rotation angle of the object and rotating the object at the detected angle (Japanese Patent Application No. 2000-242242). No. 7-80223). In such an image recognition method, how to accurately match the centers of circles is an important technique in recognition processing.

[0003]

However, in the method proposed above, the center of the circle is set as the center of gravity of the black pixel in the target candidate image area, and therefore, if there is noise in the candidate area, the center of gravity shifts and the accuracy is improved. There is a problem that the center of a circle or an ellipse cannot be extracted often.

An object of the present invention is to provide an image recognition method for accurately identifying a circle or an ellipse without being affected by noise by obtaining the center of the edge information of the circle or the ellipse.

[0005]

In order to achieve the above object, according to the invention of claim 1, a candidate area of an object is extracted from an input binary image signal, and a binary image in the candidate area is extracted. Is divided into a plurality of meshes, the number of black pixels in each of the divided meshes (hereinafter referred to as mesh data) is calculated, and mesh data on a plurality of concentric circles from the center of the candidate area is calculated for each circumference. A feature amount arranged in an order of one round from the base point is calculated, and a feature amount of a circle having a predetermined radius and a circle having the same radius as the predetermined radius in the dictionary are arranged in an order of one circle from the base point. Further, in the image recognition method for detecting the rotation angle of the object by collating with the feature amount arranged in order for one rotation from a position shifted from the base point by a predetermined position, a circle or an ellipse in the candidate area is obtained. Based on the edge information of It is characterized by calculating the center of a circle or an ellipse.

According to a second aspect of the present invention, when the input image signal is a color image signal, the candidate region is extracted after binarizing the color image and the center of the circle or ellipse is calculated. Is characterized by.

According to a third aspect of the present invention, the center of the circle or ellipse is an intermediate coordinate value of two edge pixels on the same scanning line.

According to a fourth aspect of the present invention, the distribution of the intermediate coordinate values is a histogram, and the maximum frequency coordinate of the histogram is the center coordinate of a circle or an ellipse.

According to a fifth aspect of the present invention, the distribution of the intermediate coordinate values is a histogram, and the average value coordinates of the histogram distribution are the center coordinates of a circle or an ellipse.

[0010]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be specifically described below with reference to the drawings. FIG. 1 shows the configuration of an embodiment of the present invention. In the figure, 1 is a binary image signal, 2 is 2
A candidate area extraction unit that extracts an object area (rectangle) from the value image signal 1 is a mesh data calculation unit that divides a binary image in the candidate area into meshes and calculates the number of black pixels in each mesh. 4, a mesh data memory for holding the calculated mesh data, 5 a feature amount calculation unit for arranging the mesh data from the mesh data memory for each predetermined circumference of concentric circles in order for one round to create a feature amount, 7 is a dictionary in which feature amounts are registered in the same manner, 6 is a rotation angle detection unit that detects the rotation angle by performing matching between the feature amount of the object and the dictionary, and 8 is a rotation angle An object recognition unit 9 that recognizes an image based on the certainty factor is a control unit that controls the whole. Each of these parts is similar to that of the above-mentioned application. In the present invention, the edge information extraction unit 10 is further provided.
Is provided.

<First Embodiment> The candidate area extracting section 2 extracts a circumscribed rectangle of a black connected component as a candidate area. The mesh data calculation unit 3 performs mesh division on the image in the candidate area with a predetermined number of pixels. FIG. 2 shows an example in which the image in the candidate area is divided into a mesh (small area) of 2 × 2 pixels. Then, the number of black pixels in each mesh is counted, and the value is stored in the mesh data memory 4 as mesh data. FIG. 3 shows an example mesh data memory for the image of FIG. In each mesh, for example,
A mesh with a value of "4" indicates that the number of black pixels is four.

Next, the characteristic amount calculation section 5 refers to the mesh data memory 4 to create characteristic amount data. That is, a plurality of concentric circles are drawn from the center of the candidate area, and the mesh data on each circumference are arranged in order for one round to create the feature amount data. This is done for a given circumference. FIG. 4 shows feature amount data of meshes on a circle having a radius of 5. In the case of FIG. 4, 28-dimensional data is created as the feature amount data of radius 5.

Subsequently, the rotation angle detection unit 6 matches the feature amount data in the dictionary 7 similarly created and the feature amount data of the candidate area with respect to the circumference having the same radius. In this matching, while shifting the matching start position of the feature amount data in the dictionary 7, the sum of the distances in each dimension of the feature amount data is obtained, and the rotation angle is detected based on the shift position having the smallest distance.

In the present invention, in order to obtain the center of a circle or an ellipse included in the candidate area, the edge information extracting section 10 first detects the black pixel of each scanning line (horizontal direction = X direction) of the candidate area. Among them, the black pixel of the maximum pixel and the minimum pixel of the X coordinate is extracted as the edge coordinate of the circle or the ellipse. By performing the same processing in the vertical direction (= Y direction),
As shown in FIG. 5, edge information within the candidate area rectangle can be obtained.

<Embodiment 2> FIG. 7 shows another structure of an embodiment of the present invention. The difference from the first embodiment is that a binary image generation unit 12 that generates a binary image from a color image signal (R, G, B) 11 is provided, and a memory 13 that stores the binary image is provided. It is a point. In the present embodiment, the input image is a color image signal (R, G, B), and a binary image in which an object is extracted is created (for example, the brightness of the color image signal is obtained, and the brightness and a predetermined threshold value are set. After a binary image is created by comparing (1) and (2), the same processing as that in the first embodiment is performed.

<Embodiment 3> The edge information extraction unit 10 of the present embodiment obtains the intermediate coordinate value of the pixel of each scanning line (horizontal direction = X direction) from the edge data of a circle or an ellipse. For example,
Edge data of two points (x1,
In the case of y) and (x2, y), the intermediate coordinates are ((x1 + x
2) / 2, y). In the ideal case where there is no noise, the X coordinate of the value of the intermediate coordinate for each scanning line (horizontal direction = X direction) becomes the same value, and this becomes the X coordinate of the center of the circle or ellipse to be obtained. Become.

Similar processing is performed in the vertical direction (= Y direction). The Y coordinate of the intermediate coordinate value obtained here has the same value, and this becomes the Y coordinate of the center of the circle or ellipse to be obtained.

<Embodiment 4> In the edge information extraction unit 10 of this embodiment, a histogram is created for each of the intermediate coordinates obtained in Embodiment 3 in the vertical and horizontal directions, and the maximum frequency coordinate is the center of a circle or an ellipse. And The maximum value is taken to eliminate the influence of the noise in the candidate area rectangle, because the coordinate value of the intermediate coordinate obtained in the third embodiment is displaced.

<Embodiment 5> In the edge information extraction unit 10 of the present embodiment, a histogram is created for each of the intermediate coordinates obtained in Embodiment 3 in the vertical and horizontal directions, and the average value coordinates of the histogram distribution are the centers of circles or ellipses. Set as coordinates. This is because, ideally, the histogram distribution should have a normal distribution shape, and the center of the histogram distribution can be obtained by taking the average thereof.

In the above embodiment, a dedicated processing unit for executing each function is provided, but the present invention is not limited to this, and calculation and processing are performed on a general-purpose processor. The configuration can be changed as described above.

[0021]

As described above, according to the present invention, it is possible to accurately identify a circle or an ellipse without being affected by noise or the like.

[Brief description of the drawings]

FIG. 1 shows a configuration of an embodiment of the present invention.

FIG. 2 shows an example in which an image in a candidate area is divided into a mesh (small area) of 2 × 2 pixels.

FIG. 3 shows an example mesh data memory for the image of FIG.

FIG. 4 shows feature amount data of meshes on a circle having a radius of 5.

FIG. 5 is a diagram showing an example of extracting edge information.

FIG. 6 shows another configuration of the embodiment of the present invention.

[Explanation of symbols]

 1 Binary image signal 2 Candidate area extraction unit 3 Mesh data calculation unit 4 Mesh data memory 5 Feature amount calculation unit 6 Rotation angle detection unit 7 Dictionary 8 Object recognition unit 9 Control unit 10 Edge information extraction unit

Claims (5)

[Claims] 1. A candidate region of an object is extracted from an input binary image signal, a binary image in the candidate region is divided into a plurality of meshes, and the number of black pixels in each divided mesh ( (Hereinafter, referred to as mesh data), and calculate the mesh data on a plurality of concentric circles from the center of the candidate area,
For each circumference, the feature amount arranged in order from the base point for one round is calculated, and for the feature amount of the circle having a predetermined radius and the circle having the same radius as the predetermined radius in the dictionary, the feature amount is sequentially turned for one round from the base point. In the image recognition method for detecting the rotation angle of the object by collating the arrayed feature value and the feature value arrayed in order for one rotation from a position displaced from the base point by a predetermined position, in the candidate area, An image recognition method characterized by calculating a center of a circle or an ellipse based on edge information of the circle or the ellipse. 2. When the input image signal is a color image signal, the candidate region is extracted after binarizing the color image, and the center of the circle or ellipse is calculated. 1. The image recognition method described in 1. 3. The image recognition method according to claim 1, wherein a center of the circle or ellipse is an intermediate coordinate value of two edge pixels on the same scanning line. 4. The image recognition method according to claim 3, wherein the distribution of the intermediate coordinate values is a histogram, and the maximum frequency coordinate of the histogram is a center coordinate of a circle or an ellipse. 5. The image recognition method according to claim 3, wherein the distribution of the intermediate coordinate values is a histogram, and the average value coordinates of the histogram distribution are the center coordinates of a circle or an ellipse.