Design And Implementation Of Workpiece Location Recognition And Sorting System Based On Machine Vision

With the continuous development of machine vision technology,industrial robots are gradually applied to industrial processing and assembly environment.They can automatically identify the types and positions of workpieces by using visual technology,and complete the sorting of spare parts by combining mechanical grabbing equipment.However the existing technology usually relies on high-precision visual imaging equipment or laser scanner which can detect depth information.The expensive hardware,complex algorithm and long-time production make this can't be widely used in industrial production.Aiming at the problems of industrial robots in the technology of workpiece recognition and sorting,this paper designs and implements a workpiece location recognition and sorting system based on machine vision.The system uses monocular camera and industrial manipulator as hardware basis,combines visual calibration and artificial neural network technology to complete the work of workpiece location.recognition and sorting.Firstly,the calibration of the internal parameters of the camera is completed by the traditional Zhang Zhengyou calibration method,and the external reference calibration is transformed into the PnP problem.A nonlinear optimization calibration method is proposed.Aiming at the difficulty of solving the calibration equation in the traditional linear hand-eye calibration method,a method for solving the calibration equation based on Lie algebra is proposed,and the positioning accuracy of the vision system is verified in the experiment.Secondly,the feature extraction method of the workpiece to be sorted is introduced,and the MLP-based workpiece recognition neural network is constructed.It can determine the type and number of each workpiece in the captured image according to the actual characteristics of the workpiece image.Finally,the problem of image segmentation of bonded workpieces is studied.The algorithm of segmentation based on watershed and region calculation is proposed.Then the sorting work of static workpieces is realized in the actual hardware environment.The rapidity and accuracy of the system are verified by experiments.The main contribution of this paper is to build a workpiece recognition and sorting system based on HALCON vision software platform.After debugging the hardware platform,it can achieve higher accuracy and shorter completion time.The method used in the system is low cost,simple equipment debugging,short software development cycle,and can meet the requirements of industrial accuracy and real-time.It provides a new solution for automatic assembly of industrial parts.