Aiming at the problem that the traffic infrastructure can’t accurately detect the road users in urban intersection environment

a point cloud feature extraction and classification method based on viewpoint feature histogram and support vector machine（ VFH-SVM） is proposed. According to the perception requirements of the target in the complex environment of the intersection

the point cloud information obtained by roadside 3 D Li DAR is preprocessed by straight through data filtering

the scene point cloud data is segmented by the conditional Euclidean clustering algorithm

the 3 D point cloud features of different targets are extracted based on the viewpoint feature histogram

and viewpoint feature histogram（ VFH） features of different traffic participants and facilities are established. Support vector machine（ SVM） algorithm of linear kernel function is used to classify and recognize the point cloud data in traffic scene. The experimental results show that the accuracy of classification and recognition of typical objects in urban intersection environment is higher than 90 %.