Detection of Illegal Wearing of Seat Belts Based on Key Points of the Human Body
In order to detect the normative wearing of safety belts by scaffolding workers in the construction envi-ronment,this paper proposes a safety belt violation detection algorithm.This algorithm consists of two parts:object detection and semantic segmentation.In the target detection part,the CenterNet network is used as the main structure,and different prediction branches are used to complete the human key point and scaffold rotation target lo-calization tasks.In the semantic segmentation part,in order to improve the recognition accuracy,the feature fusion structure and the improved CBAM-f attention module are added to the original DeepLabV3+network to realize the pixel-level localization task of the lanyard.Compared with the original model,the improved network is at mIOU,and the F1 score increased by 3.5%and 2.7%respectively.
NETL
NSTL
万方数据
维普
