| The problem of robotic visual servoing is one of the most important fields in both domestic and abroad. Its research achievement can not only be used for Industrial production,such as load or unload stuff,transportation,spray,spot welding,precise assemblance,quality verifyication,but can also be used to promote the related subjects.The key technology of the system——image moments pattern extraction and design of visual servoing controller based on image moments,is analysed through the research of the typical joint robot visual servoing system's structure.Considered of these key technology,the whole robotic visual servoing simulation system and related algorisms are completed as follows:(1) The dissertation introduced the classification and general situation of the current robotic visual servoing system,and the robotic kinematics model,visual system imaging model,pose and it's changing velocity between different coordinates,robotic Jacobian matrix were presented;the frame of the typical joint robotic visual servoing system is completed based upon the former models.(2) The image based visual servoing system which was widely opted by researchers in this fileds was chosen in this dissertation to start this research.We chose image moments which was recognized as a useful global image descriptor to describe the image character information ,conclude the formula between zero-order image moments and imaging depth information, propose the image Jacobian matrix based upon image moments;testify the linear mapping between image moments and image dots upon which is based to construct the fundamental matrix in image moments coordinates;a nonlinear robust proportion controller which met the need of wide range lyapunov gradual stability was designed.(3) a simulation modal of the eye in hand robotic visual servoing system which was used to orient an static object and to track a moving object was established to testify the validity of the image Jacobian matrix and visual servoing control law. |
PDF Full Text Request