基于yolov5的高感染风险行为检测算法

A real-time detection of mask wearing and safety distance based on YOLOv5 algorithm

随着计算机视觉领域的创新和发展,针对传染病爆发类公共卫生事件对人类身体健康的威胁,文章提出一种基于YOLOv5算法实现对口罩佩戴情况和安全距离的实时检测。通过添加CA注意力机制对目标检测进行优化,实验结果表明,所提出算法检测的精确性高,其中对小目标口罩map值提升较大,高达99.1%,实时性强,能够满足实际使用需求。

With the innovation and development in the field of computer vision,aiming at the threat of public health events such as infectious disease outbreaks to human health,this paper proposes a real-time detection of mask wearing and safety distance based on YOLOv5 algorithm.The target detection is optimized by adding CA attention mechanism.The experimental results show that the proposed algorithm has high accuracy and real-time performance,and meets the actual use requirements.