Research On Target Recognition And Positioning System Of Industrial Robots Based On Binocular Vision

In recent years,the country has put forward the strategic development goal of "Made in China 2025",and the intelligent manufacturing industry has gradually attracted the attention of the country.Industrial robots rely on the advantages of high work efficiency and strong adaptability to the working environment,which have become indispensable for promoting the development of intelligent manufacturing.Machine vision technology has made robots more intelligent,and robotic systems that integrate vision are developing rapidly.Binocular vision,which can obtain three-dimensional information about a target compared to monocular vision,has become a key research topic in this field.It is in this context that this paper designs a binocular vision-based target recognition and positioning system for industrial robots based on practical industrial requirements.First,a complete system solution is designed and built,and the experimental hardware selection is completed according to the system solution.The camera imaging model is analysed and the coordinate system is established,the calibration of the binocular camera is completed,the internal and external parameters and the aberration parameter matrix of the camera are obtained,and finally the hand-eye calibration of the robot vision system is completed by experiment.Secondly,the images of the target workpiece acquired by the camera are pre-processed and the target workpiece is identified and positioned.In the pre-processing stage,the realization principle of different pre-processing algorithms is analyzed,and then realized through programming.The experimental results are compared to select the image pre-processing algorithm suitable for this system.In the identification and positioning of the target,the target recognition is accomplished by template matching,then the principle of the Blob analysis algorithm is introduced and improved by the edge fitting algorithm to locate the shape centre coordinates of the target workpiece,and finally the world coordinates of the target workpiece are obtained by coordinate reconstruction to complete the precise positioning of the target.Finally,the test platform is built according to the system content,the visualisation interface is designed,and the joint programming of the image processing program,robot communication and other modules is completed to complete the system experiment using the visualisation interface.The system designed in this paper is validated by three sets of target recognition and positioning experiments,which control the robot to complete the simulated gripping of the target workpiece.The experimental results show that the robot target recognition and positioning system under binocular vision designed in this paper can identify circular target workpieces well,the robot can complete the target positioning and simulated grasping actions more accurately and the system runs stably.For the environment similar to the industrial scene,this system has certain reference significance in practical application.