A Safety Helmet and Protective Clothing Detection Method based on Improved-Yolo V 3
Xuanyu Wang, Dan Niu, Puxuan Luo, Chao Fu Zhu, Li Ding, Ke‐Wei Huang · 2020
Safety helmet is an important protective tool for workers in the industrial production, and the protective clothing worn by workers can play a role in distinguishing other person. To solve the problem of non-contact on-line monitoring workers to wear safety helmet and protective clothing as required, this paper proposes an improved-YoloV3 method as the safety intelligent detection and identification algorithm. The large-size input layer is added for multi-scale prediction and the size of anchor boxes is adjusted to enhance the detection ability of small-size helmet and protective clothing. The improved-YoloV3 detection algorithm not only meets the real-time supervision requirements with sufficient FPS, but also achieves higher mAP at different resolutions compared with the traditional Yolo V3 algorithm.