Research On Image Processing And Calibration Of Robot Visual Servo System

A robot visual servo system is broadly used and rapidly developed. Image processing and calibration are the key part of a robot visual servo system. In this thesis, image processing and calibration of robot visual servo system which are the part of the project"Research on Visual Servo Manipulator Catching Moving Object"(BK2002405) of Jiang-su Natural Science Foundation are studied and are experimentized in the testing jig. The process from image capture to the coordinate output is finished.The system structure and testing jig are designed on the basis of analyzing the requirements and research contents of the project. After many image processing methods are tested and are compared in the system, a series of image processing methods are chosen, which can divide up the object from background and output the coordinates of the object.Considering the conditions of the project, a simple applied calibration method is put forward after traditional calibration methods and characters are studied. Image aberration is mended with the method of digital image processing。The depth to the workspace is assumed constantly, so nonlinear model is transferred to linear model, and then calibration is finished by using linear relationship method.Experimental results from applying the methods of image processing and calibration show that the procedure is simple and useful, and the demands on real-time and precision of robot visual servo system are also met.