Fast three dimensional LIDAR target recognition based on projection dimension reduction

Laser radar can obtain the three-dimensional shape information of objects, which is an effective method to identify vehicle on complex ground. A fast laser radar target recognition method based on projection dimension reduction was proposed in this paper. Firstly, the suspected target area was extracted and tested from the scene. Then the digital surface model (DSM) data was obtained. Finally, the combination of contour similarity measure value and the similarity of size measurement acted as the matching recognition criteria for target recognition. The experiments were performed by using the simulated point clouds data of 8 ground armored targets. The results show that the type recognition rate of this method is above 90%, and the real-time performance is far superior to the traditional algorithms. In view of the practical application, the influences of point cloud spatial resolution and laser radar system noise on target recognition were further studied.