Vehicle Recognition Method Based on Multi-sensor Information Fusion

A multi-sensor fusion algorithm for vehicle recognition and detection is proposed to solve the problem that a single sensor is prone to false detection or missing detection in a complex driving environment. The Ghost Conv module is used to improve the traditional Conv and C3 module to achieve lightweight. An improved PointNet++ algorithm based on ResMLP is proposed to increase network depth to solve the problem of low obstacle detection accuracy caused by too small obstacle detection model. Millimeter wave radar and camera features are associated based on the matching method between image detection frame and radar projection point. The radar point cloud is projected onto the pixel coordinate system to realize the fusion of multi-sensor information. The simulation results show that the detection accuracy of the multi-sensor fusion vehicle identification algorithm is 9.8% higher than that of PointPillars and 11.7% higher than that of CenterFusion, which can effectively obtain vehicle information.