JPH0377549B2 - - Google Patents

Description

[Detailed Description of the Invention] [Technical Field to which the Invention Pertains] The present invention is directed to a part to be manipulated or processed (target part) by using a two-dimensional imaging means installed in, for example, an assembly robot or an automatic bonding device for semiconductors. The present invention relates to a device that captures an image of a target object, such as an object, and detects handling points, bonding points, etc. as feature points in the contour shape of the object.

[Prior art and its problems]

In the following description of each figure, the same reference numerals indicate the same or corresponding parts.

Conventionally, a technique for capturing an image of an object to be imaged as the target part and detecting contour feature points of the image, which are target points for manipulation or processing in the part, is disclosed in Japanese Patent Application Laid-Open No. 59-201180 by the present applicant. No. (Special Publication No. 1-30183) “Contour Feature Detection Method” (Reference (1)
).

FIG. 3 shows an example of feature points determined by this method. In other words, in the same figure, F is the target part 2.
A binarized image showing the two-dimensional shape of 00, P is this 2
N is a protrusion (hereinafter referred to as a corner point) as a feature point on the contour of the value image F, and N is a depression (hereinafter referred to as a valley point) also as a feature point. Note that G indicates the center of gravity (centroid) of the image F, and X and Y indicate the axes of image scanning.

In this way, the above method detects the types of feature points such as corner points P and valley points N, their positions on the X and Y axes, and their numbers; Even if the detection order of the feature points changes due to movement or rotation around the center of gravity G (hereinafter such displacement of the component 200 or image F will be simply referred to as "placement" or "movement or rotation"). , specific feature points on the contour are not identified. Therefore, in this prior art, the target parts 200 are placed on the workbench in various arrangements.
There is a problem in that the above specific operation or the position of the point to be processed cannot be detected.

[Purpose of the invention]

An object of the present invention is to eliminate the above-mentioned problems, and to provide a contour feature point that can specify the position of a specific operation or processing target point, that is, a predetermined feature point on the contour of the image, regardless of the arrangement of the target part. The purpose of this invention is to provide a numbering (also called labeling) device.

[Key points of the invention]

The gist of this invention is to trace the outline of the target shape in a two-dimensional image of the target part over the entire circumference in a fixed direction, detect shape feature points in the contour part using the contour information obtained at this time, and A polygon is approximated by feature points, and each feature point that appears is sequentially numbered while tracing the circumference of the polygon in a predetermined direction using one of the feature points as a base point.

Further, as the base point, for example, one end of one side having the maximum side length among the sides of the polygon may be used,
Alternatively, among the distances between the center of gravity of the target shape and each feature point, the point at the maximum distance is used. In other words, the gist of the present invention is to
Contour pixels of a target binary image (obtained by capturing an image with a dimensional imaging device, etc., and converting this image signal into a binary image via an A/D converter, etc.) are sequentially converted into the binary values (via a contour tracking circuit, etc.). rotate the converted image in a predetermined direction (e.g. counterclockwise,
Note that this direction is hereinafter referred to as the contour tracking direction), and from the coordinate value of each contour pixel obtained at this time and the tracking direction code indicating the tracking direction to the adjacent contour pixel, (feature The positions of contour pixels (hereinafter referred to as feature points) (such as corner points or valley points) representing a predetermined characteristic portion such as a protrusion or depression on the contour (via a detection circuit, etc.) are determined by moving the contour in a predetermined direction (e.g. In a device that detects feature points in the order of one round (counterclockwise) and stores them in that order, When dividing into a plurality of second groups based on properties (for example, a group consisting only of corner points or only valley points), for each of the first or second groups, the The center of gravity of the binarized image obtained through a side length calculation means that calculates the length of each side of a polygon obtained by sequentially connecting each feature point in the same group, or a center of gravity calculation circuit, etc. a center of gravity distance calculating means for calculating the distance to each feature point in the group (hereinafter referred to as center of gravity distance); and representing the feature points in the group based on the calculated side length or each center of gravity distance. a base point determining means for determining a point (hereinafter referred to as a base point); and a base point determining means for determining a point (hereinafter referred to as base point); a feature point labeling means for sequentially numbering each feature point of (each means of);
(in a feature point numbered circuit, etc.), or furthermore, the base point determining means is located on a predetermined side (for example, clockwise) with respect to the contour tracing direction on both sides of the side having the maximum value among the respective side lengths. (side)), or further, the base point determining means includes means for determining a feature point corresponding to the largest one of the center of gravity distances as the base point. When there are multiple base points belonging to the same group, the base point determination means determines each point of maximum side length with each of the multiple base points on the predetermined side (for example, clockwise side). Means for comparing the side lengths of the adjacent polygons is provided on a predetermined side (for example, counterclockwise side) with respect to the contour tracing direction as a reference, and the polygon has the maximum value among the adjacent side lengths. A point including means for determining a feature point located on a predetermined side (for example, clockwise side) with respect to the contour tracing direction on both sides of the edge as a new base point, or further, the base point determining means may include the same feature point as described above. When the base points belonging to the group are negative numbers, the angle that the side of the polygon that makes each of the plurality of base points on a predetermined side (for example, clockwise side) with respect to the contour tracing direction makes with respect to the center of gravity. The present invention is characterized in that it includes a means for comparing angles (side facing angles, etc.) and a means for determining a base point corresponding to the largest one of the angles as a new base point.

[Embodiments of the invention]

Embodiments of the present invention will be described below with reference to FIGS. 1 and 2. FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention, and FIG. 2 is a diagram showing an example of feature points on the outline of a binarized image of a target part to which detection numbers are assigned, based on FIG. 1. be.

In FIG. 1, 200 is the target part, 2
01 is a two-dimensional imaging device that images this two-dimensionally, scans the image in the X and Y axis directions, and converts it into an image signal. 202 is an A/D converter that converts the analog video signal output from the two-dimensional imaging device 201, which changes depending on the brightness and hue on the surface of the target component 200, to a level that obtains an appropriate S/N ratio; Digitize (binarize) at a level corresponding to the target part 200, and convert the background and target part 200.
This is a circuit that separates the area from the target part and obtains a binarized image F of the target part. 203 is a contour tracking circuit, A/D
Based on the data of the target shape in the data of the binarized image F obtained from the converter 202, contour pixels (also referred to as contour points) of the target shape are tracked in a fixed direction (for example, counterclockwise), and each contour is This circuit obtains the coordinate value of a point and the direction code to the next contour point adjacent to each contour point. The details of this circuit are published in Japanese Unexamined Patent Publication No. 59, published by the present applicant.
An example is described in ``Pattern contour tracing method'' (Japanese Patent Publication No. 1-30180). 204 is a feature detection circuit that detects the positions of feature points (corner points P, valley points N, etc.) of the contour in the target shape based on the information obtained from the contour tracking circuit 203, and stores them in the order of detection. It is. An example of the details of this circuit is explained in the above-mentioned document (1). Reference numeral 205 denotes a feature point numbering circuit, which assigns a constant number to each feature point detected by the feature detection circuit 204 so as to be unrelated to the movement or rotation of the image F on the screen. This is the core of the invention. 20
Reference numeral 6 denotes a center of gravity calculation circuit which calculates the center of gravity of the target shape in order to reinforce the function of the feature point numbering circuit 205.

Next, referring to FIG. 2, the operation of numbering each feature point by the component device shown in FIG. 1 will be explained. In Fig. 2, corner points P as detected feature points are P1 to P4.
Similarly, the valley points N are represented by N1 to N8.
Note that in this example, corner points P (P1 to P4) and valley points N (N
1 to N8) are numbered by separate means within the feature point numbering circuit 205.

First, the numbering of valley points N will be explained. Polygonal approximation is performed by connecting the valley points N1 to N8 detected from the target shape (the outline of the external shape of the binarized image F) in the order of contour point tracking direction (counterclockwise direction) A, and at the same time, each side of this polygon is The length of the side is determined, the maximum side length is determined, and one end of the side having the maximum side length in a predetermined direction (this direction is determined based on the tracking direction A, here, for example, clockwise) is used as a base point, and Numbering is performed sequentially in the direction (for example, counterclockwise direction A).

However, in the example in Figure 2, the lengths of side a, which is the longest side, and side c, which is the second longest side, are almost the same and there is no significant difference between them, so it is impossible to distinguish between the two, so in such a case, for example, the following The rules apply. That is, compare the side lengths of sides b and d adjacent to sides a and c in a predetermined direction (for example, counterclockwise direction A), and replace the longer side (d in this case) with the maximum side length, Determination of the base point N1 and each subsequent point N
The sequential numbering from 2 to N8 is performed. By assigning numbers to each feature point in this way, the numbering can be done uniquely regardless of the movement or rotation of the image F on the screen.

Next, the corner points P may be numbered in the same manner as the valley points N, or the valley points N and the corner points P
Connect all the feature points in the order of appearance in the tracking direction A, perform polygonal approximation that includes both valleys and corner points, find the base point of this new polygon by the same method as above, and if this base point is a valley. If it is a point, N
1. If it is a corner point, it is set as P1, and from now on, while tracing each side of the new polygon in a counterclockwise direction (tracking direction A), numbering is done in order according to the type of valley point and corner point. good.

However, regarding the corner point P in FIG. 2, an example will be shown in which numbering is done in a different way, as described below. In other words, the center of gravity G of the image F and each corner point P (P1~
P4) is the maximum distance among the distances (also referred to as center of gravity distances) and is used as the base point, and the numbers are sequentially numbered in the same manner as above (counterclockwise direction A). However, in the case of FIG. 2, since the center of gravity distances e (corresponding to corner point P4) and f (corresponding to corner point P1) are approximately equal and the maximum distances, the following rules are further applied to determine the base point.

That is, each corner point P4, P1 located at the maximum distance
and sides 4,1 connecting corner points P1, P2 adjacent to each corner point P4, P1 in the tracking direction A,
1. Compare the angles α and β (also referred to as opposite side angles) that P2 makes with respect to the center of gravity G, and set the corner point corresponding to the larger angle β as P1.
-P4 are numbered.

〔Effect of the invention〕

As is clear from the above description, according to the present invention, a base point among the feature points detected on the contour of the target shape is connected to a side of a polygon connecting each feature point along the contour. It is determined based on quantities that are unrelated to the arrangement of the target shape on the screen, such as the length or the distance between each feature point and the center of gravity of the target shape, and
Since the other feature points are sequentially numbered in a predetermined direction along the contour using the base point as a starting point, the numbers of all feature points can be determined regardless of movement or rotation of the target shape. Even when the feature point is used as a target point for operation or processing, it becomes possible to perform necessary operations and processing on the target part regardless of the arrangement of the target part.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention, FIG. 2 is a diagram showing an example of a binarized image of the target part to explain the operation of FIG. 1, and FIG. 3 is a diagram showing the conventional technology. FIG. 2 is a diagram illustrating an example of feature points on the contour of a binarized image detected based on FIG. 200... Target part, 201... Two-dimensional imaging device, 202... A/D converter, 203... Contour tracking circuit, 204... Feature detection circuit, 205... Feature point numbering circuit, 206... Center of gravity Calculation circuit, P
(P1-P4)...corner point, N(N1-N8)...valley point, G...center of gravity, a-d...side, e-g...center-of-gravity distance, α, β...side opposite angle.

Claims (1)

[Claims] 1. Contour pixels of the target binarized image are sequentially
The digitized image is moved in a predetermined direction (hereinafter referred to as the contour tracking direction).
A predetermined characteristic portion, such as a protrusion or depression, on the contour is detected from the coordinate value of each contour pixel obtained at this time and the tracking direction code indicating the tracking direction to the adjacent contour pixel. In a device that detects the positions of representative contour pixels (hereinafter referred to as feature points) in the order of one round of the contour in a predetermined direction and stores them in that order, the feature points are stored in a first group consisting of all the feature points or a first group consisting of all of the feature points. When classifying into a plurality of second groups based on shared properties, for each first or second group, each feature point in the same group is placed around the contour once. Side length calculating means for calculating the length of each side of a polygon obtained by sequentially connecting the polygons, or calculating the distance between the center of gravity of the binarized image and each feature point in the same group (hereinafter referred to as center of gravity distance) a center of gravity distance calculation means for determining a point representing the feature points in the group (hereinafter referred to as a base point) based on the calculated side lengths or center of gravity distances; point, and feature point labeling means for sequentially numbering each feature point in the group so as to go around the contour once in a predetermined direction based on the above order. Feature point labeling device. 2. In the device according to claim 1,
Labeling of contour feature points, wherein the base point determining means includes means for determining, as the base point, a feature point located on a predetermined side with respect to the contour tracing direction on both sides of the side having the maximum value among the respective side lengths. Device. 3. In the device according to claim 1,
The apparatus for labeling contour feature points, wherein the base point determining means includes means for setting the feature point corresponding to the largest one among the respective center-of-gravity distances as the base point. 4. In the device according to claim 2,
When there are a plurality of reference points belonging to the same group, the reference point determining means determines the contour tracing direction as a reference for each side of the maximum side length with each of the plurality of reference points as the predetermined side. A means for comparing the side lengths of the adjacent polygons is provided on a predetermined side of the polygon, and
A labeling device for contour feature points, comprising means for determining, as a new base point, a feature point located on a predetermined side with respect to the contour tracing direction on both sides of the side having the maximum value among the adjacent side lengths. 5. In the device according to claim 3,
When there are a plurality of reference points belonging to the same group, the reference point determining means determines that a side of the polygon having each of the plurality of reference points on a predetermined side with respect to the contour tracing direction is located at the center of gravity. 1. A labeling device for contour feature points, characterized in that the device includes means for comparing the angles extended to the other angles, and means for determining a base point corresponding to the largest one of the angles as a new base point.