JPS62287376A - Discriminating method for straight line and curve in image processing - Google Patents

Abstract

PURPOSE:To decrease the quantity of data to be processed, in a computer for controlling an autonomous control type robot, etc., and to execute the processing at a high speed, by measuring angles made by each segment for constituting a linear image, calculating successively a difference of the measured angles, and discriminating a straight line or a curve from the sum total of a difference of the angle. CONSTITUTION:Angles theta1-thetan between each segment which has been connected successively every other piece the points P1-Pn for constituting linear images (A and B) are derived, and differences DELTAtheta1-DELTAthetan of the derived angles, (namely, DELTAthetan=theta2-thetan-1) are measured, respectively. The sum total DELTAtheta1+DELTAtheta2+...DELTAthetan of DELTAthetan is derived, and by comparing a model pattern which has been stored in advance and SIGMADELTAtheta, curvatures of the linear image A and B are decided. In case of SIGMADELTAtheta 0, it can be regarded as roughly a straight line geometrically.

Description

[Detailed Description of the Invention] 3. Detailed Description of the Invention "Industrial illumination field" The present invention relates to a method for determining a straight curve in an image captured by irradiating a target object with slit light. .

``Conventional technology'' For so-called autonomous robots that recognize external conditions and automatically control their movements, technology to recognize three-dimensional objects based on visual information (three-dimensional vision technology) is an important development. This is considered an important point.

FIG. 5 shows the principle of an autonomously controlled robot using slit light.

That is, a robot 2 to be controlled and a target object 3 to be worked by the robot 2 are provided on a workbench 1, while a slit light is projected onto the target object 3 above the workbench 1. A slit light source 4 and a camera 5 that detects the slit light reflected by the target object 3 are provided, and the image information detected by the camera 5 is processed in a computer system 6 to determine the coordinates, shape, etc. of the work target 3. This is B's which supplies information about robot 2 to robot 2.

Furthermore, in the above system, the straight curve of the object image is determined by the following method.

That is, the reflection of the slit light is recognized as a sequence of many points, and the direction of each point is defined as a line segment connecting that point and m points in front. Then, if the direction defined as described above is expressed as a function ψ(s) of the length S of the line segment, this function ψ(s) takes an almost constant value in the line part of the point sequence (image). , and has a characteristic that is sloped in the curved portion. Therefore, a straight line or a curved line is determined by measuring the change in the function ψ(s) along a series of points.

"Problem to be solved by the invention" However, according to the above method, ψ(
Since it is necessary to process all values of S),
There was a problem in that the amount of data to be processed in the computer system 6 was enormous and the processing time would inevitably become long, which would limit the response speed of the robot 2.

Furthermore, it may be possible to obtain the curvature of a point sequence (image) from the slope of the function ψ(s) based on the above method and use it for control data, but in this case, the computer system 6 may have more processing capacity. The problem is that it requires

The present invention was proposed in view of the above circumstances, and an object of the present invention is to provide a determination method that can measure the curvature of an image, which is necessary for precise control of a robot, through a simple process.

[Means for Solving the Problems] In order to achieve the above object, the present invention solves the following problems: are measured, the differences between the measured angles are sequentially calculated, and straight lines and curves are determined from the sum of the calculated angle differences.

"Operation" According to the above method, the data to be processed by the calculation system can be made into simple numerical data.

“Embodiment” An embodiment of an image processing system to which the present invention is applied will be described below. In the figure, parts common to the conventional image processing system are given the same reference numerals to simplify the explanation.

FIG. 1 shows a conventional example of an image processing system to which the present invention is applied.

This device detects slit light reflected by target objects 3A and 3B with a camera 5, performs predetermined processing on the image signal recognized by the camera 5 with an image processing processor 7, and transmits the processed signal to a microcomputer. 8 to recognize the shapes of the target objects 3A and 3B.

Next, the contents of the processing performed in the image processing processor 7 and the microcomputer 8 will be explained with reference to FIG. In the following description, Sn indicates the n-th step in the flowchart.

Sl, start S 7. The camera 5 captures the light rays reflected by the target objects 3A and 3B.
is imported as image data.

S: Binarizes image data expressed by light and shade.

S4: Detect an image drawn by a line from the binarized data.

S6. The image obtained in S4 above, which has a width equal to the number of pixels, is subjected to line thinning processing to reduce the width of the line to one pixel and express it as a series of dots, as shown in Figure 2. In this way, a line image A-B corresponding to the slit light reflected by each of the target objects 3A and 3B is obtained, and the data of this image is supplied to the microcomputer 8.

S8: As shown in FIG. 3, the line images (A and B)
An angle of 0. ~θ. seek.

S7: Difference between the obtained angles Δθ1 to Δθ. , (that is, Δθn=θ.−θn−3), respectively.

The total sum ΣΔθ of S8・Δθn, that is, Δ01+Δθ2
+...Determine Δθn.

S, the pre-stored model pattern and the ΣΔ
The curvatures of the line images A and B are determined by comparing them with θ. Note that, geometrically, the case of ΣΔθ40 can be considered to be approximately a straight line.

Slo: Image identification complete.

Therefore, in determining a straight curve using the above processing process, it is sufficient to have the microcomputer 8 perform the simple processing of calculating the sum of Δθ and comparing the calculated value with the base q value.

Note that after identifying the target object, the identification data is supplied to the robot and used for control, but since the control is not directly related to the present application, a description thereof will be omitted.

Moreover, the above-mentioned discrimination method can be applied not only to robot control but also to various types of control including image processing.

"Effects of the Invention" As is clear from the above explanation, the present invention consists of a process of measuring reflected light of slit light irradiated onto an object, and a sequence of many points corresponding to the measured image data. a process of forming a line image, a process of measuring the angles formed by the straight lines of the number of strokes that connect each point constituting the point sequence continuously or intermittently, and the nth angle and the n-1th angle. The process consists of the process of sequentially calculating the difference between the angles, the process of calculating the sum of the differences between the calculated angles, and the process of comparing the calculation result with the set value. It is possible to judge straight lines and curves in images through the simple processing of numerical calculations and judgments. Therefore, the amount of data that must be processed by computers that control autonomously controlled robots, etc. can be reduced, and the processing speed can be increased. This has the effect of making it possible to improve the responsiveness of the autonomously controlled robot.

[Brief explanation of drawings]

Since FIGS. 1 to 4 show one embodiment of the present invention,
FIG. 1 is a perspective view and a block diagram of an image processing system to which the present invention is applied, FIG. 2 is an explanatory diagram of images to be discriminated,
FIG. 3 is a diagram of the principle of angle all + constant, FIG. 4 is a flowchart of image processing operation, and FIG. 5 is a perspective view of a conventional image processing system. 3A, 3B...Target object, 4...Slit light source, 5...Camera, 7...Image processing processor, 8...Microcomputer.

Claims (1)

[Claims] A process of imaging the reflected light of the slit light irradiated on the object, a process of forming a line image consisting of a sequence of many points corresponding to the imaged image data, and a process of continuously connecting each point forming the sequence of points. The process of measuring the angles formed by multiple straight lines connected regularly or intermittently, the process of sequentially calculating the difference between the nth angle and the n-1st angle, and the sum of the differences between the calculated angles. 1. A method for determining straight and curved lines in image processing, the method comprising the steps of calculating: and determining whether the calculated result is within a set range.