The Research And Application Of Precision Positioning System Based On Vision

The research based on vision positioning sub-system is a hotspot of machine vision, and it has a comprehensive foreground of applications. Firstly, the basic theory and development status of machine vision is introduced in this paper. Secondly, the design technique for different part of machine vision system is discussed according to the system's logical structure. And then, the vision p ositioning m odule o f t he s ub-project, t he m agnetic field d etecting s ystem w ith v ision positioning, of a 863 project named "Multiply and wide adjustable tube in L continual wave band "(No. 2002AA872020) was turned to reality with the design technique's guidance and put into operation successfully in the magnetic field detecting system. The main research works are as follows:The working principles of vision positioning system are introduced and its normal methods are analyzed in this paper. The particular specifications and feasibility of magnetic field detecting system are analyzed and demonstrated comprehensively according to the practical requirements. The framework o f t he v ision p ositioning sy stem i s p ut forward, and t hen t he d esign o f m ain hardware is introduced.The normal camera calibration methods, which used in computer vision are analyzed and compared. And then the advantages and disadvantages of these methods are summed up. Because of the distance between the camera and the work piece is constant during the working period, a nonlinear calibration method that can reduce the camera distort and improve the precision of inspection is introduced in this testing system. The theory of the nonlinear calibration is demonstrated t horoughly, and experiments are c arried o ut t o t est t he c alibration precision.According to the practical application of the vision positioning system, the auto exposure method, which concern to the hardware is adopted in the first instance. Subsequently, many image-preprocessing methods are introduced and compared. With efficient image-preprocessing methods, the objects can be separated from the background well that prepare for positioning the centers of the circles. To reduce the impact cased by shadow, a method by regrouping edge points with selection is introduced in circle fitting, which can improve positioning precision remarkable.The integration of the magnetic field detecting system with vision positioning is introduced from the hardware and software aspects. To reduce the positioning errors, some corresponding ways are found by many experiments, as a result, the general camera works well. All of these works guarantee that the application of the magnetic field detecting system with vision positioning is successful.