A Research Of Industrial Robot Application Based On Multi-vision

Industrial robots have driven human industrial production into a whole new field,and can liberate humans and increase production efficiency.However,the environment is more complex and the precision requirements are higher in the production field.Traditional industrial robots and even ordinary single-binocular vision-guided robots can no longer meet the production needs.Therefore,researchers at home and abroad are committed to the research of multi-vision and robot combination.In this case,this paper proposes the application research of industrial robot based on multi-vision,which combines the global monocular camera with the local binocular camera to form multi-vision,realizes the visual guidance of the robot arm,and makes the robot arm movement more intelligent.This paper studies the multi-vision guide industrial robots for related operations,which can be divided into the following aspects:1)Calibration of the multi-vision system is a prerequisite for visual guidance.First we explain the relationship between each coordinate system of the camera calibration,use the Zhang Zhengyou calibration method to calibrate the camera,and obtain the internal reference coefficient of the camera.On this basis,then we complete the calibration of the global monocular camera and the local binocular camera,and calculate their hand eyes respectively relationship and the relative pose relationship of the two cameras in the binocular camera.2)Monocular vision target monitoring and localization.The image taken by the camera is processed to highlight the region of interest,and then the position of the target on the image is obtained by template matching based on the SURF feature point,and the calculated hand-eye relationship is used to obtain the spatial position of the target in the base coordinate system of the robot;A marker-based localization algorithm is designed for large and complex objects.3)Binocular stereoscopic three-dimensional positioning.The binocular stereo vision is used to stereo-match the target,and the precise position of the target is obtained by three-dimensional reconstruction,and then the position in the base coordinate system is derived by the local camera hand-eye relationship.4)The kinematics model of 6-DOF industrial robot is established by D-H method,and the forward kinematics and inverse kinematics are studied.According to thecalculated motion node,the interpolation method is used for trajectory planning.5)Firstly,we designe the application system experiment and the corresponding software system,then apply the multi-vision robot system to the automatic demonstration of the particle detection machine.Lastly,we complete the automatic loading and unloading and start-stop operation,and carry out the result analysis.