Dynamic Obstacle Detection in Autonomous Driving Scenarios Based on Improved YOLOv5

In recent years, the development of autonomous driving technology has made significant progress and has become an important direction for future transportation. Aiming at the problems of complex background and low detection accuracy in dynamic obstacle detection in autonomous driving scenarios, this paper proposes a target detection algorithm based on improved YOLOv5. Firstly, the C3 module in the backbone network is replaced by the lightweight C2F module, which simplifies the reasoning speed of the backbone network lifting model. Secondly, CA attention mechanism is integrated into the Neck layer of the network to enhance the spatial information extraction ability of the model and improve the detection performance. Finally, the original loss function is replaced with the EIoU loss function to improve the positioning accuracy of the predicted boxes, enabling the model to output more accurately. The experimental results show that the improved algorithm improves the precision by 11.7% and the average detection accuracy by 5.3% compared with the original YOLOv5 algorithm, and the overall performance is significantly improved compared with the original algorithm, which proves the effectiveness of the improved YOLOv5 algorithm in dynamic obstacle detection in the automatic driving scene.