JPS6228402B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a distance measuring method, and particularly to a distance measuring method for non-contactly measuring the distance to a target object using image data of the target object, which is required for industrial robots and the like.

Conventionally, methods for measuring the distance to a target object using image data have been based on the principle of triangulation using stereo vision. Where,
The problem is finding corresponding points that correspond to the same object on both the left and right images, and various methods have been devised for this purpose, including the correlation method, but all of them utilize image patterns near the corresponding points. The method was to do so. In this method, in order to improve measurement accuracy, it is necessary to increase the distance between the measurement points of the left and right images, but doing so may increase the deformation of the image pattern and make the correspondence difficult. Defects that become difficult to measure appear, and if the image pattern is made smaller, detailed measurements can be made, but there is a risk of finding false corresponding points.

In view of the above, the present invention has been made for the purpose of providing a method that uses image data of a target object and can easily and precisely measure distance even if the image pattern of the target object changes depending on the position of the viewpoint. It is.

The present invention uses a plurality of images obtained by taking more detailed imaging points between the imaging points of both the left and right images used in conventional stereo vision.
Distance is measured using the principle of motion stereopsis.

The configuration and function of this device will be explained based on the drawings. As shown in the first diagram, move the camera 1 to the camera moving table 2.
The coordinates of the imaging system are x'-y'-z', and the coordinates of the image are x-z. Now, place an object plane O perpendicular to the optical axis at a distance y′ _p from the camera. Assume that the object plane is divided into two planes, black and white, O _W and _OB by a straight line, as shown in the figure. In reality, for example, one surface O _B may be the object surface, and the other surface O _W may be the surrounding space surface. When the camera 1 is moved in the x'-axis direction, a plurality of images I _p are captured for each constant minute movement amount δw.
~I _i ~I _o are arranged in order as shown in the second diagram to form a kind of stereoscopic image. A point p' on the target object is mapped onto each image I _i , but its height (z value) on the screen is constant. This is because, if the camera position is c (w, 0, 0) and the focal length is c (w, 0, 0), then the point p'(x',y',z') on the object is the point p (x , w, z) using the following equation.

w=−y′/・x+x′ −(1) z=/y′・z′ −(2) According to equation (2), the value of z is constant. Therefore, if we take a horizontal cross-section of a stereoscopic image at height z and create a composite image, point p is a point on the straight line in equation (1), and its slope -y'/ is proportional to distance y' It is quantity. If point p′ is a boundary point of a black and white region, then the set of corresponding points p U _p
corresponds to a straight edge on the composite image as shown in the third figure. In other words, the straight edge U on the composite image
Distance can be measured by detecting the slope of _p . FIG. 4 shows a schematic configuration of a processing device based on the above measurement principle.

The camera 1 is driven by a camera moving device 2 and moves in a straight line. The distance image shown in FIG. 3 is synthesized from a plurality of images taken every certain minute movement distance using the image synthesizing device 3. From the obtained distance image, the straight edge direction detection device 4 obtains the direction of the straight edge using the above equation, and calculates the distance.

By using the present invention, the problem of distance measurement can be reduced to the problem of finding the inclination of an edge on a single composite image, and it is possible to perform detailed measurements on points on an object that correspond to edges.

[Brief explanation of the drawing]

Fig. 1 is an explanatory diagram of the imaging system, Fig. 2 is an explanatory diagram of a three-dimensional image constructed by sequentially arranging multiple images taken as the imaging point moves, and Fig. 3 is an explanatory diagram of the synthesized distance image. The explanatory diagram, FIG. 4, is a schematic configuration diagram of the entire processing device. In the figure, 1 is a camera and 2 is a camera moving device.

Claims (1)

[Claims] 1. Obtaining a straight line edge formed by a set of predetermined points on the target object by composing multiple images taken every fixed minute movement distance with a camera aimed at the target object, and detecting the slope of this straight line edge. A distance measurement method characterized by calculating the distance between the camera and a target object.