In order to detect the surrounding vehicles for intelligent vehicle high reliably increase environmental adaptive ability of vehicle detection algorithm

a vehicle detection algorithm based on multi-feature fusion is studied. Segmentation on Li DAR sensor data is carried out by extracting road surface point

grid

and spatial clustering method

several candidate targets are obtained. candidate target is classified. detecting vehicles targets are obtained. While constructing input of classifier

on the basis of basic geometric and reflectivity features

Lshape feature and a reflective distribute features are employed for the training classifier in order to improve the accuracy of the vehicle detection. NWFE-transformation is performed on the input feature vectors before the training classifier

which further improves the accuracy of the vehicle detection. The experimental results verify that the proposed algorithm has strong robustness and environmental adaptivity.