For Mobile Robots Research On Key Technology Of Omni-directional Vision System

With the development of science and technology, robot technology plays an increasingly important role in human production, life, military, rescue and counter-terrorism. Robot to complete the scheduled tasks will need to obtain information related to scene, as a result of the robot system scalability, there are many ways to provide decision-making information for the robot, which is the main way of machine vision. The traditional robot vision system has the disadvantages of a single point of view, and cannot obtain global information at once. Robot in many practical applications, however, need to ultra wide Angle even images of the 360° Angle of view, in order to get more info about field. Omni-directional vision technology arises at the historic moment, because it has large characteristic, through the omni-directional vision technology applied in robot vision system design, able to provide more comprehensive and efficient work for robot scene information,has a good application prospect.Fisheye lenses have a larger perspective than ordinary lenses because of their physical structure. In the same area, fisheye lenses capture more information. So in the same field of information request, the number less than ordinary fish eye lens, the corresponding decrease in the number of stitching images, there is less number of joining together, which can significantly reduce computation of video image processing system, improve the real-time performance of system.Based on the analysis of the traditional principle of omnidirectional vision system, on the basis of using fisheye lens for robot work scene information, again through the improvement of distortion correction and image matching algorithm, and then to robot provides a comprehensive and efficient scene information. This article focuses on the following two aspects:Research: (1) image algorithm effective area of the original distortion fisheye images are extracted and correction, and puts forward the improvement of fisheye image fast correction method based on the longitude coordinates; The method of using the SURF algorithm is applied to optimize the image after correction and optimize the process design and implementation of the splicing process.(2) the simulation experimental research: using the professional simulation platform of mobile robot with Webots, simulation modeling and experiment environment for mobile robot, and comparative study on the effects of the proposed algorithm of image. The improved fish-eye image correction and splicing algorithm were implanted in this simulation environment to verify the feasibility of the scheme.Simulation experiment results show that the improved fisheye image correction algorithm and image stitching algorithm after correction, high real-time performance, can get ideal omni-directional look around the image, meet the needs of mobile robot on the omni-directional vision system.