The Research About The Recognition And Location Of Workpiece Based On Machine Vision

In order to adapt the application of machine vision in industrial, the articlerespectively studies the sorting and positioning when the workpiece is in thestationary state and the identification and location of the workpiece in themovement state, finally it is completed that the static sorting system anddynamic classification system software development which not only canprovide help for automated production detection and identification pipeline,but also can further improve the machine vision applications. In the process toachieve this purpose, mainly researching the things as the followings:According to the working principle of machine vision, the articleresearches the general layout design of the recognition and location system,and then completed the table structures of the recognition system.Characteristics of the system, it introduce mainly the calibration of the insideand outside parameters of the camera calibration and the calibration of robot“eye-hand” system based on the operator provided by HALCON software, thecalibration algorithm is simple and feasible, and high precision. In order toachieve the requirements of static sorting and recognition of workpiece, the classification algorithm is researching based neural network. And ultimatelydetermine the classification algorithm of the MLP and complete theexperimental interface design and the debugging of the recognition system. Inaddition, this article makes a study of location and classification of dynamicworkpiece based the sorting and identification of static workpiece. Thismethod is based on the spatial position of triggering time to determine the graba space time phasor translation methods, then use the image processingsoftware to control the electric signals of a robotic arm to complete theclassification of the workpiece.It is a dynamic recognition system in C++development platform that isbased on the image processing operator of the HALCON software. Afterdebugging, this system can complete real-time online detection, and canachieve production requirements precision.