Research On The Application Of Binocular Vision System In Disc Parts Loading Robot

In recent years,with the development of intelligent manufacturing,industrial production has gradually developed from automation to intelligent,and the application of vision system in industrial robots is an important aspect of the transition to intelligent.As far as industrial robots loading are concerned,the traditional method is to use manual teaching to load materials,When the position and shape of the loading objects are fixed,it can achieve a certain automation and can achieve high efficiency;However,it takes a great deal of time to prepare for the early stage of teaching,and the accuracy of the teaching is closely related to the technical level of the workers;In addition,when the position and shape of the loading object change,it is necessary to teach again,which will affect the production efficiency and increase the labor cost.Combining the binocular vision system with the industrial robot is an effective method to solve this problem,The binocular vision system can recognize the position and shape of the loading object in real time,and it is processed through related image processing and positioning algorithms,the three-dimensional coordinates of the loading object in space can be obtained,and the coordinate value can be sent to the robot control system to guide the robot to perform intelligent loading.The method has the advantages of fast loading speed,high precision and low labor cost.According to the request of disc parts loading system and the overall function,determine the overall scheme,main including feeding robot,binocular vision control system,the camera fixed way,such as scraping scheme design.The hardware selection of the loading robot system based on binocular vision was completed,which mainly includes the selection of the loading robot,the camera and the lens;The key technologies of camera calibration,image processing,hand-eye relationship calibration,and target three-dimensional positioning algorithm were derived.The relationship between the coordinate systems in the vision system is analyzed,and the relationship between them is determined by the camera calibration;And the method of image processing is studied,according to the actual situation of disc parts carries on the comparison to different image processing algorithms,select the appropriate algorithm,completed the disc parts center extraction;Based on existing algorithms,a feature extraction and matching algorithm is optimized and improved toimprove the accuracy of matching;The hand-eye relationship matrix was determined by calibrating the eye-hand relationship.Through the research of the target 3d localization algorithm,the spatial 3d localization of the disk parts is completed.VS2012 and OpenCV2.4.9 were used to complete the design of the testing software for the binocular vision positioning system of the loading robot,and the three-dimensional positioning and loading experiment of the disc parts were completed,the feasibility of the algorithm was verified,the experimental measurement data is analyzed and the causes of errors is obtained.