JPS6245092A - Method and apparatus for recognizing position and attitude of part - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Application of the Invention] The present invention relates to a method and apparatus for recognizing the position and orientation of a component, and particularly relates to a method for recognizing the position and orientation of a component, and more particularly, to position and its posture (
The present invention relates to a method and apparatus suitable for recognizing the direction of the lead)'f.

[Background of the invention]

This type of component position and orientation recognition technology is required to perform component position and orientation recognition work associated with automatic electronic component assembly equipment.

A known example of this birch technology is the magazine Automation (1982, Vol. 27, No. 2, published by Nikkan Kogyo Shimbun).
There is a technique described on pages 50 to 55.

According to the above-mentioned known technology, in order to find the direction of an object, take the distribution of each direction code of the object, superimpose it every π/2 to find a distribution with a large peak, and calculate the angle at this peak. It is the direction angle of the object. With this method, only the approximate angle of the object is known, but the exact direction angle is not known, making it difficult to calculate the direction angle when the shape of the object changes in various ways.

Also, known techniques that can be applied to recognize the direction of one part include the method of calculating one center of gravity + pattern matching, and the method of extracting contours + assigning direction codes, but the former method of calculating the center of gravity + pattern matching , the above-mentioned Automation (29% No. 7 in 1984), p. 34, is detailed, but if this is applied to leaded parts, it will not only cause a large error, but also
This requires an arithmetic unit with a huge storage capacity t.

In addition, the latter method of contour extraction + direction code assignment calculates the center of gravity for the center of the part body in the same way as the former, and for the direction of one part, first extracts the contour of the part, and then assigns a direction code (8 This is a method in which the direction mode with the most distribution is assigned as the direction of the part.In this method, the direction is recognized as a rough direction of 22.5°/1 direction, which is less than the 180'/8 direction. , it is practically unsuitable when fine directional accuracy is required.

[Purpose of the invention]

The present invention aims to solve the above-mentioned problems of the prior art.

b) The center position of the component body can be reliably recognized, b) A separate memory function is not required in the calculation means, c) Even if the tip of the lead is bent, it will not be affected significantly, and b) The posture (that is, the direction of the lead) can be precisely calculated. This paper attempts to provide a method and device for recognizing parts position and orientation.

The present invention can be applied to so-called vertical components in which a plurality of leads protrude in parallel from one side of the component body, and also to so-called coaxial components in which parallel leads protrude from two opposing sides. .

[Summary of the invention]

First, the basic principle of the present invention will be briefly described in accordance with the background of its conception.

The present invention provides l1l when binarizing a shadow image of an electronic component.
(Threshold level) We found that the lead image of a component disappears and appears, and first, we binarized it using a threshold such that the lead of the component disappears, and then we converted the image of only the body to the center position of the multi-component body. Detect. 1. Next, the parts lead appears. We discovered that this binarized image now includes the image from which the lead has been deleted, so the left and right coordinates of the body-only image extracted earlier are (about the same distance c5 pixels)? By detecting the root position of one component lead from the taken coordinate position, in the case of coaxial parts, the part is detected from the coordinates of two points on the left and right, and in the case of vertical parts, from the coordinates of the two lead root positions and the coordinates of the body center. Calculate the body direction angle tg. We created a processing program for the above flow and tested it, and confirmed that it was possible to reliably recognize the center of the part body and the direction angle.

In order to achieve the above object based on the above-mentioned principle, the component position and orientation recognition method according to the present invention provides a method for supplying an electronic component having a large difference between the body diameter and the lead diameter to an electronic component automatic assembly device. In the method of automatically recognizing the position and orientation of the electronic component, an overall image of the component is captured by a TV camera and inputted to an image processing device, and the position and orientation of the center point of the component are determined according to the missing procedure. (a) The image information is binarized, and the threshold value is selected so that the image of the lead part is erased, and the resulting binary value is Find the center position of the part body from the information.

(b) The threshold value for binarizing the image IJ2 information is changed to obtain binarized information representing the external shape of the part, and this is subjected to histogram processing to detect the lead root position.

(c) Calculate the position and orientation of the component based on the information obtained in (a) and (b) above.

In addition, the device of the present invention created to be suitable for carrying out the above-mentioned method of the invention supplies the electronic component automatic assembly device with a component having a large difference between the body diameter and the lead diameter. A device that automatically recognizes a posture is provided with a TV camera that images the shape of the part, and an arithmetic device that processes image information captured by the camera, and the image processing arithmetic device has the following features: (a) (b) A function of binarizing image information 1, selecting a threshold value such that the image of the lead portion is erased, and determining the center position of the main part of the component from the 2@ conversion information obtained thereby; (b) Image information 1r, a function of changing the binarization threshold to obtain binarized information representing the external shape of the part, and performing histogram processing on this to detect the lead root position; (c)
It is assumed that the system is equipped with a function of calculating the position and orientation of the parts based on the information obtained in (a) and (b) above.

[Embodiments of the invention]

FIG. 4 is an explanatory diagram of the overall structure and operation of an electronic part assembly device, which is a device to which the present invention is applied.

The parts discharged from the parts supply device 1 are conveyed by a belt conveyor and placed into a transparent parts supply pallet 6.

Then, the position of the component is recognized and the vacuum pickup head 2 picks up the component.

The suctioned parts are sent to the conveyor arm position according to the lead root position information to the main body of the automatic parts assembling machine.

There, the parts are subjected to various forms, transported by the transport conveyor 3, and subjected to constant polarity checks (partly.

After the direction is reversed), the lead-cut prints supplied from the printed board conveying device 4 are inserted into the insertion head 5.

FIG. 2 is an enlarged detailed view of the component position recognition portion shown in FIG. 4.

The electronic components 7 contained in the transparent acrylic pallet 6 are illuminated from the bottom by the lighting 8, and the shaded images of the components are illuminated by the ITV from the top.
An image is taken by the camera 9. The captured image is processed by an image processing device 110, and data on the body center position and direction angle of the part is sent to the vacuum pickup head 11. The vacuum pickup head 11 described above is supported by a robot (not shown), and conveys and supplies the vacuum-adsorbed part 7 to the insertion machine. It is necessary to correctly recognize the center position of the part and the direction of the lead, adjust it to a predetermined posture, and then hand the receiver to the insertion machine. Therefore, in the image processing device 10, a technique for analyzing the imaging information of the ITV camera 29 is necessary and important. The method and apparatus of the present invention meet this need.

FIG. 3 is a schematic diagram for explaining an embodiment of detecting the position and orientation of a component by the method of the present invention.

When a shadow image of a component is binarized, the leads of the component may disappear or appear depending on the threshold value. At that time, it is known that the lead erased image is obtained by uniformly deleting the periphery of the entire component image, as shown in FIG. 3, and is completely included in the entire component image. Therefore, in order to detect the center position of the component body, it is sufficient to obtain the image with the leads removed. To calculate the center of the lead erasure image, take a histogram (column distribution) on the Y axis and the Y axis, and from the coordinates of the intersection with each axis, XM = (X1 + X2) /
2. The center position is calculated as YM=(Y1+Y2)/2.

Next, if the position of the base of the lead is known, the body direction angle of the component can be determined from this and the center position of the component, so the position of the base of the lead is detected as follows. In other words, from the intersection points X1 and X2 with the Y axis found earlier, a point taken a certain distance (about 5 pixels) to the left and right is X3.
．． X4, and detect the Y coordinates of these two points.

This example shows the case of a vertical component, in which it is recognized that there is a lead on the X4 side by comparing the image data at X3°X4, and the image data existing on that position is By searching for two y coordinates Y*r
Find it in Y4'.

These two coordinates y4. If y4′ is known, take the average basket and find y
, we find a point with coordinates (X4.y, ,) and a point with coordinates (X
The arrow (A) connecting the points M, YM) is the body direction angle (
lead protrusion direction).

FIG. 1 is an explanatory diagram of the processing algorithm in the above-mentioned calculation, and shows a method for recognizing the orientation and orientation of a vertical component.

DL shows the flow of processing, and d and t are attached with charts corresponding to the above.

First, an image d of the part is input (flow D).

Perform binarization (Flow E), L histogram processing (Flow F), and t- with such low distress that the component lead disappears, and detect all body center positions of the component from the coordinates of the intersection with the X and Y axes (Flow 〇).

Next, set a high threshold so that the leads of the component appear.
Value conversion (flow H,)L, histogram processing (
70-d), Recognizing the lead direction as described above (3rd
Calculate arrow (a) in Figure B) (flow J),
! Detection of J-dot root position (flow K) and component angle θ
is calculated (flow L).

FIG. 5 shows an example in which the same processing as described above (FIG. 1) is performed on a coaxial component, and the operations from flow D' to flow E' are similar to those from 70-D to 70-E in the previous example. It will be processed.

When the histogram is processed in flow 1' and the root position is detected in flow ', two points (b) and (c) are found as the roots of the two leads in the case of a coaxial component (see k'), so... It is possible to calculate the component angle (flow L) without performing calculations equivalent to lead direction recognition (flow J) for vertical components
), the lead protrusion direction (component angle θ') can be determined.

〔Effect of the invention〕

As described above in detail, image information is binarized by applying method a of the present invention, and the threshold value is selected so that the image in the lead part is erased, and the resulting binarized 21i, which calculates the center position of the main part of the part from the information, and then changes the threshold for binarizing the image information to represent the external shape of the part.
By obtaining i conversion information, processing it into a histogram, and performing arithmetic processing based on the above information, a) the center position of the component body can be reliably recognized; without the need for memory functions.

3) Even if the tip of the lead is bent, it will not be affected much, and 2) The posture (i.e., the direction of the lead) can be accurately calculated, which are excellent practical effects.

Further, by applying the device according to the present invention, a TV camera that images the shape of a part and a calculation device that processes image information taken by the camera are provided, and the image processing calculation device described above is (a ) Binarize the image information, select the threshold value so that the image of the lead part is erased, and calculate the center position of the main part of the part from the binarized information obtained thereby; (b) A function that changes the image information to a threshold value to binarize it to obtain binary information representing the external shape of the part, and processes this into a histogram to detect the lead root position, and (c) the above (a), (b) ) If a recognition device is configured that has a function of calculating the position and orientation of parts based on the information obtained by .

[Brief explanation of the drawing]

FIG. 1 shows the position of a vertical part by applying the method of the present invention. FIG. 6 is an explanatory diagram showing an example of an operation for recognizing a posture. FIG. 2 is a general explanatory diagram regarding component position recognition. FIG. 3 is a diagram for explaining the principle of the method of the present invention, and FIG. 4 is an explanatory diagram of an electronic component assembly apparatus as an example of an object to which the method of the present invention is applied. FIG. 5 shows an embodiment different from FIG. 1, and is an explanatory diagram when the method of the present invention is applied to a coaxial component. DESCRIPTION OF SYMBOLS 1... Parts supply device, 2... Vacuum pickup head, 3... Conveyor, 4... Printed board transfer device, 5... Insertion head, 6... Transparent acrylic pallet, 7...・Electronic parts, 8...Lighting, 9...ITV
Camera, 10... Image processing device, 11... Vacuum pickup head.

Claims (1)

[Claims] 1. In order to supply an electronic component with a large difference between a body diameter and a lead diameter to an electronic component automatic assembly device, a method for automatically recognizing the position and orientation of the electronic component, An overall image of the image is captured by a TV camera and inputted to an image processing device,
A method for recognizing the position and orientation of a part, characterized by calculating the position and orientation of the center point of the part according to the following steps. (a) Binarize the image information, select a threshold value so that the image of the lead portion is erased, and determine the center position of the component body portion from the resulting binary information. (b) Change the threshold value for binarizing the image information to obtain binarized information representing the external shape of the part, process this with a histogram to detect the lead root position, (c) Above (a), ( Calculate the position and orientation of the part based on the information obtained in b). 2. In order to supply an electronic component with a large difference between the body diameter and the lead diameter to an electronic component automatic assembly device, a TV that images the shape of the component in a device that automatically recognizes the position and orientation of the electronic component. A camera and a calculation device that processes image information captured by the camera are provided, and the image processing calculation device (a) binarizes the image information, and the threshold value is set so that the image of the lead portion is erased. (b) expressing the external shape of the part by changing the threshold for binarizing the image information; A function that obtains binarized information and processes it into a histogram to detect the lead root position, and (c) a function that detects the position and orientation of the part based on the information obtained in (a) and (b) above. 1. A component position and orientation recognition device, characterized in that it is equipped with a function of calculating.