Research On Map Construction Method For The Intelligent Vehicle Based On Stereo Vision

For the smart vehicle, in order to achieve the autopilot, you must have the ability to perceive and to construct a map of the road environment, and map accuracy for intelligent vehicle navigation and path planning has a very important impact. In recent years, as a result of stereo vision system can perform three-dimensional reconstruction of the environmental information directly, are increasingly being used for scientific and technical personnel in the field of intelligent vehicle map building. Therefore, based on stereo vision smart car road map building environment has a very important role and significance.This paper mainly deals in unknown road environment, research in motion the process of intelligent vehicle road environment map building problem. First, in the analysis summarizes the status of the domestic and foreign in this field of research, pointing out the important and difficult issue when constructing a map; then, zhengyou Zhang calibration method using stereo camera calibration; then, a series of coordinate conversion, and establish feature point coordinates and pixel coordinates relations between the world; the same time, the development of stereo vision based on mileage calculation method, using sparse achieve matching feature points matching and tracking, smart car traveling estimated mileage information.Finally, global matching algorithm, respectively, about a moment ago, and after a moment’s view parallax calculation and three-dimensional reconstruction, and the front and rear views reconstructed matching stitching computing will eventually discrete point cloud elevation data regularization, Build a Smart vehicle can be expressed through sexual raster. Experimental results show that the proposed method can be completed based on the smart vehicle stereo vision map Construction Methods.