Study On The Road Detection Technology Based On Graph Theory In The Scope Of Monocular Vision

The Intelligent Transportation System(ITS)is the development direction of the modern transportation. The world developed countries invest lots of resources in the area of ITS one after another. The automatic vehicle control technology is one of the most important technologies of the ITS.And the vision-based road navigation technology is one of the most important parts of the automatic vehicle control technology. This study is undertaken to find a rapid and highly-efficient algorithm to satisfy the demand of the intelligent car traveled at a high speed.The road images are gathered by a monocular vision based information collect system, and then processed by graph theory based image segmentation algorithm. Establish the rule of mergence of the subblocks of the segmented road image and merge the subbloks into the road area. Based on the result of the road area detection, the road edge detection algorithm uses the least square method to fit the road edge. The edge detection algorithm is rapid, reliable and flexible.Be different from the traditional road detection algorithm, the algorithm in this paper is independent of the road traffic marking. It may get a satisfying result when was applied to the unstructured road. Experiment results prove that the algorithm can get a good result in the condition of various road types, diverse weather conditions and complicated traffic conditions. In addition, the algorithm is rapid. The detecting time of each frame is 26 millisecond, can meet the requirements of real-time operation of the intelligent vehicle.