基于多信息融合的可驾驶道路检测算法研究.pdfVIP

M ay COM PU T ER ENGINEERING A ND DESIGN Vol No RA NSAC KIT T I T P A doi issn j Drivable road detection algo rithm based on multi inf ormation f usion Z H U Zhenw en Z H O U Li LIU Jian C HEN Jie R D Center for Green Energy Automotive Electronics Institute of M icroelectronics of Chinese Academy of Sciences Beijing China School of Electronic Electrical and Communication Engineering U niversity of Chinese Academy of Sciences Beijing China Abstract T o enhance the performance of drivable road detection in advanced driver assistance sy stems ADAS in complex ur ban road environment a robust road detection algorithm w as proposed For color image its illuminantinvariant feature w as ex tracted af ter removing the influence of illuminant variance and shadow s in the w ay of the feature space transformation w hich w as used for building the modelbased classifier Likelihood distribution w as computed by filtering the binary images w hich w as ob tained using the modelbased classifier Road region of interest RoadROI w as estimated by Bayesian probability analysis com bining likelihood distribution and empirical spatial priors distribution Drivable road region w as generated using depth information in RoadROI and RANSAC to filter out the nonroad pixels Ex perimental results on KIT T IRoad datasets verify the effective ness and robust