JPH02179401A - Measuring method of object - Google Patents

Abstract

PURPOSE:To prevent false detection by a method wherein an image of a numbered contour line is prepared beforehand for each object of measurement and only the coordinates of the intersection of the contour line having the same number with that of the object in detection are determined when intersections of the contour line and segments drawn on the image of the contour line are determined. CONSTITUTION:An input image information from an ITV camera 100 is converted into a binary-code information in a binary-coding circuit 101 and stored in a data memory 103. By using the binary-coded image stored in this memory 103, the input image information is processed in a processing device 102. Consequently, the binary- coded image is raster-scanned to detect a point of change from a black pixel to a white pixel, or vice versa, and thereby an image of a contour line is obtained. Next, on the image of the contour line, a segment of a prescribed length is drawn at every prescribed angle around the center of gravity of an object to be measured. When this segment intersects the contour line, only the coordinates of the intersection of the contour line having the number of the object to be measured are determined and a distance between the intersections thus determined is calculated. In this way, measured data on each object are obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method for measuring an object, and more particularly, to a method for measuring an object suitable for measuring the length of an arbitrary part of a target object.

[Conventional technology]

Image processing in robot vision systems, etc.
There is a demand for finding a specific part of a target object and measuring its length even if the position or orientation of the target object imaged by an ITV camera changes. This type of conventional technology involves drawing line segments based on predetermined geometric relationships on a binary image of an input image obtained by photographing with an I'I'V camera, and identifying specific parts of the target object. Methods for determining length are generally known.

[Problem to be solved by the invention]

Even with the above conventional method, there is no problem when there is only one object in the screen taken by the ITV camera or there is no noise at all.However, if there are multiple objects in the screen or there is noise. In this case, when you draw a line segment and find the intersection with the contour, you need to know whether the intersection is with the object you are currently trying to detect or with another object, and whether it is an intersection with noise. It was impossible to distinguish whether it was an intersection or not, and there was a risk of false detection.

An object of the present invention is to provide a method for faithfully determining the length of a designated specific portion of a plurality of objects within a screen, even when the positions and postures of the objects change.

[Means to solve the problem]

In order to achieve the above object, the present invention processes an input image from an imaging device such as an ITV camera to extract a contour line of a target, and sets the constituent pixels of the contour line differently for each contour line of each target. Create a contour image with a number given to it, draw a line segment on the contour image based on the center of gravity of the object to be measured, etc., and draw a' among the intersections of the line segment and the contour line. The method is characterized in that only the coordinates of the intersection of contour lines having the same number as the object to be measured are determined, and the length of a specific portion of the object to be measured is determined from the coordinates.

[For production]

Create a numbered contour image for each object, draw a line segment that serves as a so-called ruler on the contour image, and draw a line segment with the same number as the object currently being detected among the intersections between this and the contour line. Only the coordinates of the intersection of contour lines are valid. With this, when a line segment is drawn and intersects with the contour line, it is possible to distinguish whether the intersection point is the intersection of the current object to be measured or the intersection of another object or noise. This eliminates the risk of false detection even when multiple objects or noise are present in the image captured by the camera.

〔Example〕

An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 shows a flowchart 1- of an embodiment of the present invention.

Figure 1 shows the case where line segments are drawn at constant angles around the center of gravity of the object, but by using the center of gravity of the object in this way, it is possible to perform the total Lil1l regardless of the position or posture of the object. Can be done. Note that in addition to the center of gravity, the direction of the principal axis of inertia of the object may be used.

FIG. 2 shows a schematic configuration of hardware for realizing the method of calculating d1 of the present invention. The ITV camera 100 photographs a target object and inputs image information. Usually, in the screen obtained by shooting with this type of ITV camera 100,
There are multiple objects and noises. The binarization circuit 101 is
This is a circuit that binarizes the black pixels of the manually inputted image information into II I 11 and the white pixels into '0'. Data memory 103
The program memory 104 is composed of a RAM, etc., and is used to store binary input image information and contour image information, as well as intermediate result data of processing in the processing device 102. etc., and the processing device 10
It holds the programs required for the process in step 2.

The processing in the flowchart of FIG. 1 will be described in detail below.

Input image information from the ITV camera 100 is input to the binarization circuit 1
01, it is converted into binary image information (hereinafter simply referred to as a binary image) and stored in the data memory 103 (step 1
1.12). Using the binary image stored in the data memory 103, the processing device 102 executes the following processing.

First, a binary image is rask-scanned to detect a change point (outline part) from a black pixel to a white pixel or from a white pixel to a black pixel to obtain one outline image (step 13). In the transverse ζ line image, only black pixels or white pixels in one contour are set to ``1''.
All other values are represented by 1101+.

Normally, an input image (screen) obtained by photographing with the IT camera 100 includes several objects and noise of 1Mi. Therefore, a plurality of objects and noises in the obtained contour image are separated (step 14). This is the pixel that constitutes the contour part of the axis contour image (i.e., II 111
This is achieved by tracing the contours that form a closed loop sequentially, and assigning numbers to the pixel groups of the contours forming each closed loop, for example, in the order of the found closed loops. Note that the contour search is performed while calculating its length, and among contour lines forming a closed loop, those whose length is shorter than a predetermined threshold are removed as noise. FIG. 33 shows an example of a contour image obtained by separating each object when there are three objects in the input image.

Next, as shown in FIG. 3, a total of i11! contour images with different numbers added for each 11-contour line of each object!
Geometric features representing the position and orientation of the object to be measured are extracted (step 15).In this step, the center of gravity of the object to be measured is detected, but the direction of the principal axis of inertia of the object may also be detected. Next, after initially setting the angle θ to 0 (step 16), line segments of a predetermined length are drawn at constant angles Δ0 (for example, 10°) around the center of gravity of the object to be measured, and the outline of the object is Measure the distance (steps 17-20)
.

FIG. 4 is a diagram for explaining this, in which G indicates the center of gravity and a indicates the length of the line segment. 2, line segment 2 plays the role of a so-called ruler, and this line segment is +16 as shown in Figure 3.
When this is drawn on the contour image and the contour line intersects with the contour line, only the coordinates of the intersection point of the contour line of the target number currently to be 819 are determined, and the distance between these intersection points is calculated (step 18).
, This allows measurement data for each object to be obtained. The distance measurement ends when the angle 0 reaches 360° (step 19). Note that the obtained data is used for identifying the shape of the object, etc., but since this is not the purpose of the present invention, the explanation will be omitted.

〔Effect of the invention〕

As explained above, according to the present invention, a numbered contour image is created for each object, and when a line segment is drawn on one contour image to find the intersection with the contour, the current detection Since only the intersection coordinates of the contour lines with the same number as the object are calculated, IT
Even if there are multiple objects in a screen photographed with a V-camera or the like, there is no possibility of false detection.

[Brief explanation of the drawing]

FIG. 1 is a flowchart for explaining an embodiment of the object measuring method of the present invention, and FIG. 2 is a diagram showing a schematic configuration of hardware for realizing the method of the present invention. FIG. 3 is a diagram showing an example of a contour image in which each object is numbered, and FIG. 4 is a diagram showing an example of how to draw line segments on the contour image. 100... ITV camera, 101... Binarization circuit, 102... Processing device, 103... Data memory, 104... Program memory. Figure 1 σji

Claims (1)

[Claims] (1) Processing the input image from the imaging device to extract the contour of the object, and creating a contour image in which the constituent pixels of the contour are given different numbers for each contour of the object, A line segment is drawn on the contour image, and among the intersections between the line segment and the contour line, only the coordinates of the intersection point of the contour line with the same number as the object to be measured are determined, and the object to be measured is identified from the coordinates. A method of measuring an object characterized by finding the length of a part.