A Pedestrian Detection Method for Highways Based on Lightweight FCOS and Hybrid Feature Pyramid Networks

With the rapid development of the transportation industry, pedestrian detection algorithms on highway roads have become an important research direction. To solve the problem of a large number of model parameters and a large amount of calculation, this paper proposes a road pedestrian detection method based on lightweight FCOS. In this method, the hybrid feature pyramid networks(HyFpn) is introduced by improving the neck feature pyramid networks, and the lighter mobilenetV3 is used as the backbone network. Experimental results show that the proposed method significantly reduces the size and computational complexity of the model while maintaining high detection accuracy, which is conducive to real-time pedestrian detection on resource-constrained equipment. Compared with the existing methods, the proposed method achieves a better balance between accuracy and speed.