The Research Of Vision Navigation Based On Path Following

With the development of robot technology and the market's push , the intelligent robot becomes one research focus , and mobile robot is widely used in many fields . As the development , the way of navigation on the based of computer vision is widely concerned in robot navigation . Compared with other methods , vision navigation has the advantages of informative , high-level intelligent , and the good applicability .This paper takes the wheel mobile robot as research object , and focuses on the navigation and control problems of robot system , the main work in this paper is :First of all , this paper summarize the status of at home and abroad . Systemically analyzes the monocular vision system characteristics , the types , working principle and the vision navigation system technical composition , and carries on the detailed explanation to structure and dynamic model .Secondly , path recognition is the key of vision navigation technology , this paper attempts to use a variety of image processing algorithms . The image have noise , must to carry through the image pre-processing ; Then compared several edge detection operators , choose the Prewitt operator to detect the road edge ; At last road edge extraction used Hough transform .Then , taking the human driving experiences , the path following based on fuzzy control is established . The deviation of distance and angle are selected as two input values , difference of wheel velocity as the output value . The experiment results show that the fuzzy control approach is a ideal control method for mobile robot system .Finally , the algorithms of the path recognition and following developed by VC++ , the result show that the algorithm and the following system can efficiently perform .