Dual Ccd Humanoid Eye Neck Motion Analysis And Control

Computer, network, image processing, communication and multimedia technologies have developed rapidly in the decades. These technologies that are gathered and applied in mechanical and electronic products have improved industry greatly. After successfully researching and manufacturing an intelligent dome camera that can follow and track object in all orientations, supported by aerospace technology support foundation (No.HT2001-zjdx), the author researchs on Dual-CCD Vision System Simulating Human Eyes and Neck. The system uses two CCD cameras and four freedoms of rotate framework to simulate human eyes and neck. The research includes many theories and technologies such as mechanical designing, movement controlling, stereo vision measure and so on. So it supports experiment equipment and theoretic basis for our further research, 3D vision recognizing and specise object tracking.In chapter 1, the research's significance of a dual-CCD simulating human eyes and neck vision system is introduced, as well as related research background at home and abroad. And related key technologies and main research content are summed up.In chapter 2, the whole scheme and theoretic frame of dual-CCD simulating human eyes and neck vision system are introduced. The dual-CCD simulating human eyes and neck framework is designed and its mathematic model is built, then the robot's movement equations are deduced.In chapter 3, CCD imaging technology is introduced and the optical model for distance measuring using two CCD cameras is built. Using binocular intersection measuring method, its solid vision measuring equations are worked out and these error factors are analyzed.In chapter 4, the mathematic model of stepping motor are researched in general, subdivision control arithmetic of stepping motor are introduced. For clearance of gear side brought by decelerating framework of stepping motor being applied, its compensatingarithmetic while system resetting is put forward, and it can satisfy the system after successful experiments. The communicating ans controling software between PC and bottom CPU are also developed.In chapter 5, a new camera calibration technique for 3D machine vision with the radialalignment constrain (RAC) is described. The method separates camera parameters by using RAC, in order to provide the solution through a linear algorithm according to the reasonable order instead of traditional nonlinear optimization. After experiments, the camera calibration technique is proved to be a good method.In chapter 6, a 3D measuring model based on vision sensors is build. After existing methods of characteristic point matching being improved, the related 3D measuring theory and software flow designing are put forward.In chapter 7, the project's main research content and solutions are summed up, and some prospect and explore for further discuss and research are described.