Research on Fast Road Detection Problem for Visual Robots

SHI Peng-bo,GUO Dong-yan,ZHAO Chun-xia
DOI: https://doi.org/10.3969/j.issn.1006-9348.2012.11.061
2012-01-01
Abstract:Common road detection algorithm for visual robots in complex environment has the shortcomings of bad anti-interference and poor speed.In order to suppress noise disturbance and improve the execution speed of the algorithm,a road detection algorithm based on the combination of edge and region was proposed.The algorithm firstly uses the Canny operator to find the image edge,and performs the adaptive region segmentation according to the color information of the road,then combines edge information with region segmentation information to identifies the road boundary region.Finally,it performs Hough transform based on edge image of the region to detect the road.The experimental results show that compared with the traditional road detection algorithm,this algorithm can effectively improve the ability of anti-noise and running speed with a better road detection results.
What problem does this paper attempt to address?