Enhancing Road Pothole Detection Algorithm of YOLOv5

To address the issues of low accuracy and difficult deployment in road pothole detection, an improved algorithm for road pothole detection based on YOLOv5s is proposed in this paper. The algorithm incorporates a GhostNet backbone feature extraction network to reduce model complexity. Additionally, GSConv (Ghost Shuffle Convolution) is introduced in the Neck module to ensure model accuracy while reducing model weight. Moreover, a weighted bidirectional feature pyramid network (BiFPN) is employed to capture more feature information and enhance the information representation capability of feature maps, thereby increasing the detection accuracy of the network.Experimental results demonstrate that the improved model achieves a 30% reduction in algorithm complexity and a 33.3% decrease in computational workload. The mAP0.5:0.95 is improved by 4%, and the precision is increased by 5.3%. The improved algorithm achieves a balance between lightweight design and accuracy. Overall, the proposed algorithm achieves a significant reduction in complexity while maintaining high accuracy. This improvement contributes to the effective detection of road potholes, addressing the challenges of accuracy and deployment in practical applications.