Local 3D Map Building Based on Projection of Virtual Height Line

Abstract

3D map building is an important task of autonomous land vehicle (ALV) for obstacle detection, path planning. Traditional methods using stereo vision usually relies on dense disparity map obtained by stereo matching, and 3D map is built based on the disparity map. Traditional methods are sensitive to mismatch pixels in disparity map, and dense stereo matching increases lots of unnecessary computation. This paper proposes a novel 3D map building method based on projection of virtual height line (VHL). Unlike traditional methods, the proposed needn't obtain the disparity map, so it reduces the computation cost and can be highly real-time. Simulation and experimental result is given to show that the built 3D map reflects the actual scene basically and meets the need of ALV navigation.