Abstract:Aiming at the problem of low detection accuracy of vehicle and pedestrian detection models, this paper proposes an im-proved YOLOv4-tiny vehicle and pedestrian target detection algorithm. Attention Module (CBAM) is introduced into CSPDarknet53-tin module to enhance feature extraction capabilities; In addition, the CSP-DBL module is used to replace the original simple convolutional module superposi-tion, which compensates for the high-resolution characteristic information and further improves the detection accuracy of the network. Finally, the test results on the BDD100K traffic dataset show that the mAP value of the final network of the proposed method is 88.74%, and the detection speed reaches 63FPS, which improves the detection accuracy of the network and meets the real-time detection speed.