A classification algorithm for vehicle targets applied to complex urban environment is proposed using 3D laser infrared radar（LiDAR） roadside deployment. Firstly

background difference method and density clustering algorithm are used to detect vehicle point clouds. Then

the shape

proportion

height profile and profile of vehicle point clouds are extracted

and Random Forest classifier is trained to achieve accurate classification of vehicle targets. Finally

the detection and classification performances of single feature and comprehensive feature are analyzed and compared. The results show that the accuracy rate of vehicle target classification in urban complex environment is 95. 2 %.