Dissertation > Industrial Technology > Automation technology,computer technology > Automation technology and equipment > Robotics > Robot

Research on Binocular Stereo Vision System for Indoor Mobile Robot on Complex Environment

Author MaZuo
Tutor ZhuShiQiang
School Zhejiang University
Course Mechanical and Electronic Engineering
Keywords Binocular stereo vision Camera Calibration OpenCV Stereo matching Disparity map Obstacle map
Type Master's thesis
Year 2012
Downloads 70
Quotes 0
Download Dissertation

Chapter on binocular stereo vision technology Research , summarizes the main map created vision - based obstacle detection methods , and research at home and abroad binocular stereo vision for mobile robot navigation system as well as existing binocular stereo vision system products . Clarify the research significance and main content . Chapter II , the first linear pinhole model of camera imaging lens distortion : radial distortion and tangential distortion . Checkerboard flat image experiment , using OpenCV calibration parameters and structural parameters of the space inside and outside of the left and right cameras . And use the the Bouquet corrected image on a binocular epipolar rectification . Chapter III , the introduction of a stereo matching algorithm popular international criteria and standards of experimental images provide an objective basis for the judgment of the pros and cons of various stereo matching algorithm . The stereo matching algorithm is divided into two categories , local matching algorithm and global matching algorithm , winning Select local matching algorithm as well as the more popular dynamic programming method and graph cut method global matching algorithm . Correlation algorithm qualitative and quantitative assessment . Chapter IV , according to the application requirements of the mobile robot , key research area - based local matching algorithm to analyze the various factors that affect the matching results , the qualitative similarity measure the accuracy of the function and the size of matching window and quantitative analysis , in-depth study of image preprocessing and post-processing methods to improve the disparity map of stereo matching results to the box filtering multiresolution computing acceleration techniques applied to the implementation of a stereo matching algorithm , and to use OpenCV further improve the algorithm real-time. The fifth chapter analyzes the the ideal binocular stereo vision system , depth measurement error model , local map to remove surface noise to a reasonable height model . Integrated application and validation of the stereo matching algorithm used in the previous chapters and calibration results . The depth calculation results show that the measurement error can meet the actual needs of the use of the actual scene depth . The results of the three-dimensional reconstruction and raster map well to provide environmental information for mobile robots to lay the foundation for the future path planning .

Related Dissertations
More Dissertations