An improved safety helmet detection algorithm based on YOLOv8

Safety helmets are very important for the life safety of operators in construction sites, mines and other high-risk operating environments. Therefore, based on the YOLOv8 framework, this study proposes the CC-YOLOv8 safety helmet detection algorithm, optimizing the issues of low accuracy and high computational resource consumption faced by traditional detection methods. By introducing the C2fcc module, the backbone network feature extraction capability of the algorithm is significantly enhanced. Meanwhile, the EMA attention mechanism is added to the algorithm, which effectively improves the object localization accuracy. The experimental results show that the algorithm demonstrates superior performance in a variety of scenarios and conditions, and its mAP0.5 reaches 92.6%, which is improved by 0.5% compared with the original algorithm. This research result provides an efficient and accurate new method for safety helmet detection.