Fast vehicle detection using orientation histogram and segmented line projection

Fast detection of moving vehicle is essential for autonomous driving. In this paper we propose fast vehicle classification algorithm for autonomous vehicle using orientation histogram and segmented line projection, and also works well with in urban environment. The rapidness of our approach stems from the sequential laser data processing and orientation histogram techniques. Whereas accurate classification of point cloud into vehicle object is done in a 2D grid with segmented line projection. Experimental results show verification on data from outdoor scenes acquired from our electrical vehicle equipped with 2D laser scanners.