A study on the 3D dimension measurement using robot hand-eye calibration

Abstract

With the continuous progress of science and technology, industrial robots are entering factories to replace human beings. Industrial robots are highly efficient, but they mainly detect and locate fixed targets on a single production line in a specific environment. For this reason, the vision sensor and industrial robot are combined to complete the fixed target work by visual guidance. This paper mainly studies the techniques of camera calibration, feature point extraction and three-dimensional positioning in the binocular stereo vision system. A visual positioning and ranging system suitable for industrial production is constructed, and the distance measurement between multiple target points is realized. The results show that the average error of three-dimensional positioning ranging is less than 0.5 mm, which proves the feasibility of the method. At the same time, the hand-eye matrix is calculated by matrix direct product, and the target point pose of the world coordinate system is converted into the pose of the robot coordinate system, which provides a basis for industrial production. The main work is as follows: (1) Build a simple hardware platform with a robotic arm, binocular camera and computer. (2) The theoretical basis and principle of camera calibration are studied.Based on the linear calibration method of perspective transformation model in MATLAB software environment obtain camera parameters and distortion parameters. (3) The two images are acquired by using a binocular vision camera. The pre-processing of the acquired image is completed, and the specified target point is searched in the image by template matching to obtain the two-dimensional image parallax. (4) Based on the internal and external parameters of the camera and matching points, the depth of the feature points is calculated according to the stereo vision measurement principle, and corresponding three-dimensional space information is obtained. (5) The hand-eye transformation matrix of the robot is calculated by using the direct product of the matrix and the eigenvector. The spatial relationship between the camera coordinate system and the robot arm coordinate system is determined by hand-eye calibration, and the relative position of the target point relative to the robot end tool is calculated, which provides a data basis for the spatial operation.