Complex terrain classification algorithm based on multi-sensors fusion

Terrain detection under complex environment is important to environment perception for autonomous vehicle. This paper presents a terrain classification method based on multi-sensors data fusion. Raw data received from 3D laser ranger and camera is applied to get the feature of terrain firstly. Then the driving space is divided into stereo unit and each unit includes some inherent feature characteristics. Hidden Markov model describe the structure of the driving space and model parameters are trained by Baum-Welch algorithm. Experiment results show that the method can classify the complex terrain effectively.