Research on 3D Target Detection Algorithm For Vehicle-road Collaborative Perception

To eliminate the blind area of vehicle environment perception and achieve real over-the-horizon global perception, this paper formulates a three-dimensional object detection algorithm based on vehicle-road collaborative perception. Firstly, the vehicle-road simultaneous sequential sensing 3D detection dataset V2X-Seq and the evaluation index are presented. Then, the neural network structure based on point cloud and image is designed, and the post-fusion approach is adopted. Finally, the training test and experimental results are analyzed. The road sensing equipment can precisely transmit the road information in the blocked area that is challenging for intelligent vehicles to detect to the vehicle.