Rasterize the Lidar Point Cloud on The Ground Out Method Optimization Analysis

Abstract This paper is mainly aimed at the over-segmentation or under-segmentation phenomenon of ground point cloud in moving scene unmanned platform to identify target. proposes an optimized rasterization method remove clouds appropriately. First, partitioned data by quadrant processed based on asymmetric distribution front-to-back and left-to-right data. Then, estimation lowest carried out drilling into laser layer data, connection between estimated points detection segmented using road reflection intensity correction. Next, filtering raster elevation difference. Finally, obstacle continuum hypothesis model used improve that occurs inside raster. The overall effect, algorithm has a certain degree universality ideal effect within road, achieve desired goal.