Based on analyzing existing objection recognition and location methods,The main contents of this thesis for the characteristics of clean-up robot are as follows:1.Analysis of status in Obiective Recognition and Location for disposal robots in the dangerours environment,And given component of the system.A double CPU parallel structure was designed to increase the accuracy and efficiency of transmission system.Introduce the framework of software and hardware design.2.Using histogram equalization and gauss filtering method pre-processing dud CCD image,After comparing canny operator algorithm to other edge detection algorithms in edge feature,Canny operator algorithm is used in this paper.Towards Shape for the characteristics of dud,Hough transform is adopted to gain the lines in edge space.Definite mean region by means of cut up image according to the detected lines.3.A splitting algorithm can effectively extract the turning point of dud and link them.The method to solve the existing problem of dealing with sharp turning point in the image contour extraction.4.An self-adapting genetic algorithm and improved "Part Hausdorff distance" recognition is offered in this paper.The algorithm uses component edge feature as match feature,Adopt improved Part Hausdorff distance as the norm of target similarity measurement,Adopt self-adapting genetic algorithm to search the best object to match.5.We Studied the visual localization algorithm of dud,KLT will be movement of visual feature tracking in three-dimensional visual.Which carry out the objectives of the three-dimensional reconstruction. |