Research On The Rigid Body Automatic Merging System Based On Nonparallel Binocular Vision

Rigid body merging is a common project in the world of machinery, architecture, shipbuilding, chemical, aerospace and so on. And meanwhile it has becoming a trend using the method of advanced image processing and control technology to accomplish the work of high precision automatic merging. A system of rigid body automatic merging based on nonparallel binocular vision is introduced in this paper.The theory and various algorithms about camera calibration were expounded in the first part of this paper. In order to obtain the intrinsic parameters and external parameters of the camera in our system, Zhang algorithm was adopted. Secondly, a systematic discussion of all kinds of image enhancement method was conducted, combined with corresponding experiment, so as to select the suitable one for the system. On the basis of detailed study of the Hough transform algorithm of ellipse detection, a new one based on center extraction was proposed. And then the architecture and models of the merging system were established, by means of which the3D coordinate of the merging surface center was determined. The paper took a examination to the system and analysed its merging accuracy. At last the hardware and software packages that required by the system were listed, and software design process was confirmed, in order to realize the entire system. The experimental results show that the measuring error is0.33mm, which can satisfy the requirement of rigid body high precision merging.