现代电子技术2025，Vol.48Issue(18)：177-186,10.DOI:10.16652/j.issn.1004-373x.2025.18.027
基于遮挡感知的安全帽细粒度穿戴检测算法
Safety helmet fine-grained wearing detection algorithm based on occlusion perception
摘要
Abstract
In order to cope with the diversity and complexity of power operation scenarios and solve the difficult problem of safety helmet supervision,an safety helmet fine-grained detection algorithm based on occlusion perception(EHD-Net)is proposed.In allusion to the problem of insufficient multi-scale feature extraction capability in occlusion scenarios,a large separable kernel convolution module(EDKA)based on occlusion perception is proposed,solving the issue of inaccurate detection caused by the occlusion of safety helmet.In allusion to the problems of the insufficient feature extraction and fusion capability of the model,a separation and enhancement attention module(DAAM)is proposed,and a new feature amplification detection head(FA-Head)is constructed,resolving the problem of poor small-object detection performance caused by the limited model receptive fields.To address the model's insufficient convergence capability,a loss function based on distance and scale factors(DLS-IoU)is proposed,solving the problem of slow convergence speed during training.In allusion to the insufficient generalization capability,a scheme of fine-grained dataset partitioning is proposed,which can divide the dataset into five different categories based on the norms of safety helmet wearing and the status of the safety helmet chinstrap,thereby enhancing the practical application ability of the model.The experimental results show that,in comparison with the baseline model(YOLOv8n),the average accuracy of the proposed algorithm can reach 94.5%,an improvement of 6.3%.关键词
细粒度/遮挡感知/安全帽佩戴检测/目标遮挡检测/小目标检测/卷积神经网络Key words
fine granularity/occlusion perception/safety helmet wearing perception/object occlusion detection/small object detection/convolutional neural network分类
信息技术与安全科学引用本文复制引用
杨学周,张军,陈习文,金淼,刘莉,余锋,姜明华..基于遮挡感知的安全帽细粒度穿戴检测算法[J].现代电子技术,2025,48(18):177-186,10.基金项目
国家自然科学基金项目(62202346) （62202346）
国家留学基金资助项目(202208420109) （202208420109）
武汉市应用基础前沿研究项目(2022013988065212) （2022013988065212）
湖北省安全生产专项资金科技项目(SJZX20220908) （SJZX20220908）
