%A %T Real-time Detection and Tracking of Road Information for Outdoor Vision-based Robot Based on Improved Otsu %0 Journal Article %D 2011 %J Science & Technology Review %R 10.3981/j.issn.1000-7857.2011.08.009 %P 58-62 %V 29 %N 11-08 %U {http://www.kjdb.org/CN/abstract/article_3282.shtml} %8 2011-03-18 %X The detection and tracking of road information are the basis and prerequisite of environmental exploration and autonomous navigation by mobile robot. In view of the complexity of the outdoor working environment of the robot and the variety of the outdoor illumination conditions, a suitable and novel way of real-time detection and tracking for outdoor robot is presented. First, the video images obtained from CCD camera are preprocessed. Then the Region of Interest (ROI) is extracted and transformed into the HSV color space. The edge information of the road image is obtained from the ROI by using the improved Otsu algorithm. With the anti-jamming ability of the Hough transform, the region where robot can pass through is shown. Finally the weighted Mahalanobis distance discrimination method is used to track the passing area. Experiment results show that this algorithm enjoys a good performance in the detection and tracking of road information in outdoor environment in real-time.