Two-wheeler Helment Wearing Detection Alogrithm Based on Improved YOLOv8

The two-wheeled vehicle helmet-wearing detection algorithm is an important research content of target recognition algorithms in road traffic applications; this study for the existing helmet detection algorithms real-time detection speed is slow, leakage of wrong detection and other issues, proposed based on YOLOv8n improved helmet detection algorithm YOLOv8-Slimneck-PIoU.Based on the experimental comparison selected YOLOv8n model as the basis, the Slim-neck module is used to optimise the Neck module in the base network, and the original Conv module is replaced by the lighter convolution module of GSConv, which improves the detection speed of the whole algorithm under the premise of maintaining the accuracy; further optimising the loss function for PIoU enhances the detection accuracy of the algorithm. After a comparative test, it is found that in the case of the same data set, the average accuracy mAP@0.5 of the improved algorithm can reach 87.2%, and the overall performance is higher than other mainstream algorithms, which meets the real-time detection of helmet-wearing under road traffic.