Aiming at the problem of poor detection effect and real-time performance of existing laser point cloud target,a real-time vehicle point cloud detection algorithm based on attention mechanism is proposed.The detection algorithm proposed in this paper combines the attention mechanism algorithm with YOLOv3,uses the attention mechanism to distribute the weight of the features of the bird′s-eye view of the point cloud,so as to learn the correlation of the features in different channels and spaces,and improves the loss function of the detector through CIOU loss and focal loss.The experimental results show that the detection speed of the vehicle point cloud detection algorithm based on the attention mechanism can reach 30 f/s,and the average detection accuracy of the vehicle target can reach 92.5 %.In the real vehicle data test,the algorithm can quickly and accurately identify vehicles in a certain range,and achieve real-time detection effect.