JPH0283785A - Method for recognizing object - Google Patents

Abstract

PURPOSE:To recognize a three-dimensional shape with simple processing and composition by switching the depth of focus of an image pickup means and storing an image at a part in which the difference in the lightness of a picture is closed to zero as contour line information. CONSTITUTION:A focal length L of an ITV camera 2 is set to a scanning start distance by a control signal from a picture processor 5, and next, variables to indicate a scanning space, the number of the times of scanning, a threshold value, and the current number of the times of scanning are set. Next, the focal length is changed by a switching mechanism 3, and the depth of focus is changed by a switching mechanism 4, respectively. Thus, in the camera 2, the depth of focus is either deepened or shallowed, and respective received pictures are stored in the memory of the device 5. At this time, the device 5 calculates the differences in the lightness of picture elements corresponding to the respective pictures, determines some picture elements at parts in which the differences are smaller than the threshold value determined from a zero point as the picture elements at the position of a contour line, links the picture elements at the position of the contour line with each other, and stores the linked picture elements as contour line information 6. Further, after the whole of the scanning of an object 1 completes, the three-dimensional shape of the object 1 is recognized from the bent state and the interval of the information 6.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention relates to an object recognition method for three-dimensionally recognizing the shape of an object.

[Prior Art] The following methods are known as conventional methods for three-dimensionally recognizing the shape of an object.

(a) The three-dimensional shape of the object is determined based on a lattice-like image obtained by irradiating light such as a laser beam onto the object to be recognized in a band-like manner and shifting the irradiation position in steps at equal intervals. Laser cutting method that facilitates recognition, (b) An object is irradiated with light that has passed through a grating, and the object is imaged through an equally spaced grating, and a three-dimensional image of the object is recognized based on the image obtained. (C) A modified lattice method in which the lattice spacing on the image receiving side is different from the lattice spacing on the light irradiation side in the above Moire method.

(d) A method of recognizing the three-dimensional shape of an object using the principle of trigonometry using two ITV (Industrial Television) cameras.

However, methods (a) to C) require a special lighting device in addition to the ITV camera that receives images of the object, and method (d) requires two ITV cameras. In either case, equipment costs would increase, and there would be problems such as shadows being cast on objects depending on the position of the lighting and the field of view of the ITV camera, making it impossible to recognize the shape of the shadowed parts.

Therefore, in order to solve and overcome such problems, (e) the object to be recognized is imaged by a single ITV camera while changing the focal length in stages, and the image of the in-focus part is determined from the obtained image. A method has been proposed in which a contour line report is extracted and the three-dimensional shape of an object is recognized and read from this contour line report.

[Problems to be Solved by the Invention] However, the method of extracting a contour column using only one image for a certain focal length has a problem in that the processing becomes complicated.

An object of the present invention is to provide an object recognition method that can recognize a three-dimensional shape with simple processing and a simple configuration.

(Means for Solving the Problems) The present invention processes an object to be recognized by changing the focus processing step by step using a single imaging means, changing the depth of focus into two stages: deep depth of focus and shallow depth of focus for each focal length. The system is constructed so that the image of the area where the difference in brightness is close to zero is taken out as a contour line report, and the three-dimensional shape of the object is recognized and covered from this contour line report. .

(Function) When the depth of focus is switched to two levels, deep and shallow, at each of a plurality of focal lengths, the brightness of parts at the same distance remains almost the same even when the focal length is switched.

Therefore, when the focal length is changed, the brightness of the images is compared, and the parts where the brightness is almost the same are located at the same distance from the imaging means. Therefore, if parts of the same brightness are converted into contour lines, the three-dimensional shape of the object can be easily recognized from these contour lines.

[Example] Figure 1 is a schematic diagram showing an example of the north gate.
1 is an object whose three-dimensional shape is to be recognized; 2 is an ITV camera as an imaging means for receiving an image of the object 1; .ITV) J Tara 2 has a depth of focus switching mechanism 3 that allows the focal length 1i1111L- to be adjusted in stages at equal intervals, and also a depth of focus switching mechanism!
4, the depth of focus can be adjusted in two stages, deep and shallow.The image of the object 1 taken by the ITV camera 2 is processed by the image processing device 5 as follows
Dimensional shapes are recognized.

FIG. 3 is a flowchart showing the recognition processing procedure. First, the focal length of the ITV camera 2 is set to the scanning start distance by a control signal from the image processing device 5 (step 1).
00). Next, the scan interval DL is set, the number of scans N and the threshold Th are roughly set, and the variable n representing the current number of scans is set to n-1 (steps 101 to 104).
. Next, the switch I[M3 is commanded to set the focal length to L (n) = L (1), and the switch 114M4 is commanded to further deepen the depth of focus (step 105
゜106). As a result, the focus of the ITV camera 2 is set to the first focal length 1 as shown in FIG. 2 (as shown in the longitudinal cross-sectional view of the object 1), and the depth of focus is set to be deep. Here, the image IM1 as shown in FIG. 3 received by the ITV camera 2 is sent to the image processing device 5 and stored in a memory (not shown) (steps 107 and 108). Next, a command to reduce the depth of focus is sent from the image processing device 5, the depth of focus of the ITV camera 2 is made shallower, and the image I as shown in FIG.
M2 is also sent to the image processing device 5 and stored in the memory (
Steps 109-111).

Therefore, the image processing device 5 calculates the difference in brightness (absolute value) between corresponding pixels in the images IM1 and 1M2, and determines the pixels in the portion where the difference is smaller than the threshold Th determined from the zero point as pixels at the contour line position. Then, the pixels at these contour line positions are connected and stored in the memory as contour column report 6 (steps 112 and 113).

The same operation is repeated for the second focal length L(2) to the nth focal length L(2) to n), and the image obtained from the object 1 is stored as a contour report.

When N=n and all scanning of the object 1 is completed, the obtained contour line report 6 is analyzed and the shape of the object 1 is recognized from the degree of curvature and spacing of the contour line report 6.

The recognized contour line report or shape is output to a monitor or the like and reproduced as a wire frame image of the object 1.

In addition, although the focal length is adjusted in this example, I
It is equivalent to changing the position of TV nonomera 2 or changing the position of object 1, and no matter which method is used, object 1
Recognition of three-dimensional shapes on all sides of a single r - r v
Camera 2 'U - Possible roar.

Further, the depth of focus can also be changed by changing the F number of the lens 2a of the ITV camera 2. (Effects of the Invention) As explained above, in the present invention, a single imaging means is used, and the recognition The object to be detected is imaged by changing the focal length step by step, and the depth of focus is changed to deep and shallow at each focal length, and the difference in brightness of the images obtained by changing the depth of focus is close to zero. The image of the part is extracted as a contour line report, and the three-dimensional shape of the object is recognized from this contour line report, so it is possible to recognize the object with a single imaging means, without requiring special lighting, etc., and with simple processing. The three-dimensional shape of can be recognized. As a result, the shape of the object can be recognized with high accuracy without being affected by shadows, etc., and since it uses simple equipment, it becomes possible to perform actual M- at low cost.

[Brief explanation of drawings]

Fig. 1 is a schematic diagram showing an embodiment of the present invention; Fig. 2 is a vertical cross-sectional view of a Lee object; Fig. 2 is a longitudinal cross-sectional view of a Lee object; FIG. 5 is a flowchart showing a three-dimensional image recognition procedure. 1...object, 2...ITV camera. 3... Focal length switching mechanism, 4... Focal depth switching 1a mechanism, 6...
・Contour column report.

Claims (1)

[Claims] The object to be recognized is imaged by a single imaging means while changing the focal length in stages, changing the depth of focus into two stages: deep depth of focus and shallow depth of focus for each focal length, and the image obtained by changing the depth of focus. A method for recognizing an object, characterized in that an image of a portion where the difference in brightness is close to zero is extracted as contour line information, and the three-dimensional shape of the object is recognized from this contour line information.