|本期目录/Table of Contents|

[1]韩飞腾,刘永强,房玉东,等.基于注意力机制的安全帽佩戴状态检测模型*[J].中国安全生产科学技术,2024,20(8):196-202.[doi:10.11731/j.issn.1673-193x.2024.08.026]
　HAN Feiteng,LIU Yongqiang,FANG Yudong,et al.Detection model for wearing status of safety helmet based on attention mechanism[J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2024,20(8):196-202.[doi:10.11731/j.issn.1673-193x.2024.08.026]
点击复制

基于注意力机制的安全帽佩戴状态检测模型*
分享到：

《中国安全生产科学技术》[ISSN:1673-193X/CN:11-5335/TB]

卷:
20
期数:
2024年8期
页码:
196-202
栏目:
职业安全卫生管理与技术
出版日期:
2024-08-30

文章信息/Info

Title:
Detection model for wearing status of safety helmet based on attention mechanism
文章编号:
1673-193X(2024)-08-0196-07
作者:
韩飞腾; 刘永强; 房玉东; 冯涛; 郭玮; 薛明; 姬玉成
（1.应急管理部大数据中心，北京 100013；
2.清华大学 自动化系，北京 100084)
Author(s):
HAN Feiteng;  LIU Yongqiang;  FANG Yudong;  FENG Tao;  GUO Wei;  XUE Ming;  JI Yucheng
(1.Ministry of Emergency Management Big Data Center,Beijing 100013,China;
2.Department of Automation,Tsinghua University,Beijing 100084,China)
关键词:
安全生产; 安全帽佩戴状态检测; 目标检测; 注意力机制; 特征金字塔
Keywords:
work safety;  detection on wearing status of safety helmet;  object detection;  attention mechanism;  feature pyramid
分类号:
X947
DOI:
10.11731/j.issn.1673-193x.2024.08.026
文献标志码:
A
摘要:
为缓解安全生产视频监控场景下物体尺寸小、背景复杂、遮挡容易导致的安全帽佩戴状态漏检、误检、定位不准等问题，提出1种基于注意力机制的2阶段高精度安全帽佩戴状态检测模型。提出双向多层连接融合的特征金字塔网络，并设计基于编解码的空间注意力机制去除冗余特征，提升小尺寸目标的召回率；采用多尺度卷积提取候选区域多层上下文特征，并利用注意力机制对不同层级、不同尺度的上下文特征进行显式加权，进而提高模型在复杂背景下的鲁棒性；解耦候选区域分类和定位网络，分别引入通道注意力和空间注意力提升模型分类和定位精度。研究结果表明：基于注意力机制的安全帽佩戴状态检测模型整体上优于当前相对主流的高精度检测框架，如ＹＯＬＯｖ３、ＳＳＤ、ＲｅｔｉｎａＮｅｔ、Ｆａｓｔｅｒ Ｒ-ＣＮＮ、ＴｒｉｄｅｎｔＮｅｔ模型。研究结果可有效缓解安全生产视频监控场景下安全帽佩戴状态的漏检、误检和定位不准等问题。
Abstract:
To alleviate the problems of missing detection,false detection,and inaccurate localization on wearing status of safety helmet in the scenario of work safety video surveillance due to small object size,complex background,and occlusion,a two-stage high-precision detection model for the wearing status of safety helmet based on the attention mechanism was proposed.A feature pyramid network with bidirectional multi-layer connected fusion was proposed,and the spatial attention mechanism based on encoder-decoder was designed to remove the redundant features,thereby enhancing the recall rate of small objects.The multi-scale convolution was used to extract the multi-layer context features of candidate region,and the attention mechanism was employed to explicitly weight the context features with different levels and different scales,thereby improving the robustness of the model in complex background.The classification and localization networks of candidate region were decoupled,and the channel attention and spatial attention were respectively introduced to enhance the classification and localization accuracy of the model.The research results indicate that the helmet wearing status detection model based on attention mechanisms is overall superior to the current related mainstream high-precision detection models such as YOLOv3、SSD、RetinaNet、Faster R-CNN and TridentNet.The research results can effectively mitigate the issues of missing detection,false detection,and inaccurate localization on the wearing status of safety helmet in the work safety video surveillance scenarios.

参考文献/References:

［1］ 刘晓慧,叶西宁.肤色检测和Hu矩在安全帽识别中的应用［J］.华东理工大学学报,2014(3):365-370. LIU Xiaohui,HE Xining.Skin color detection and Hu Moments in helmet recognition research［J］.Journal of East China University of Science and Technology,2014(3):365-370.
［2］ REN S Q,HE K M,GIRSHICK R,et al.Faster R-CNN:towards real-time object detection with region proposal networks［J］.IEEE Transactions on Pattern Analysis and Machine Intelligence,2017:1137-1149.
［3］ HE K M,ZHANG X Y,REN S Q,et al.Spatial pyramid pooling in deep convolutional networks for visual recognition［J］.IEEE Transactions on Pattern Analysis and Machine Intelligence,2015:1904-1916.
［4］ CAI Z,VASCONCELOS N.Cascader-cnn:delving into high quality object detection［C］//2018 IEEE Conference on Computer Vision and Pattern Recognition.Salt Lake City：IEEE,2018.
［5］ REDMON J,DIVVALA S,GIRSHICK R,et al.Youonly look once:unified,real-time object detection［C］//2016 IEEE Conference on Computer Vision and Pattern Recognition.Las Vegas,NV：IEEE,2016.
［6］ REDMON J,FARHADI A.YOLO9000:better,faster,stronger［C］//2017 IEEE Conference on Computer Vision and Pattern Recognition.Honolulu:IEEE,2017.
［7］ REDMON J,FARHADI A.YOLOv3:an incremental improvement［J］.Computer Vision and Pattern Recognition,2018（8）：1-6.
［8］ LIU W,ANGUELOV D,ERHAN D,et al.SSD:single shot multibox detector ［M］.Lecture Notes in Computer Science.Berlin:Springer,2016.
［9］ LIN T Y,GOYAL P,GIRSHICK R,et al.Focalloss for dense object detection［C］//2017 IEEE International Conference on Computer Vision.Venice:IEEE,2017.
［10］ 李华，王岩彬，益朋，等.基于深度学习的复杂作业场景下安全帽识别研究［J］.中国安全生产科学技术2021,17(1):175-181. LI Hua,WANG Yanbin,YI Peng,et al.Research on recognition of safety helmets under complex operation scenesbased on deep learning［J］.Journal of Safety Science and Technology,2021,17(1):175-181.
［11］ 赵红成,田秀霞,杨泽森,等.改进YOLOv3的复杂施工环境下安全帽佩戴检测算法［J］.中国安全科学学报,2022,32(5):194-200. ZHAO Hongcheng,TIAN Xiuxia,YANG Zesen,et al.Safety helmet wearing detection algorithm in complex construction environmentbased on improved YOLOv3［J］.China Safety Science Journal,2022,32(5):194-200.
［12］ REZATOFIGHI H,TSOI N,GWAK J,et al.Generalized intersection over union:ametric and aloss for bounding box regression［C］//2019 IEEE Conference on Computer Vision and Pattern Recognition.Long Beach:IEEE,2019.
［13］ 徐先峰，赵万福，邹浩泉，等.基于MobileNet-SSD的安全帽佩戴检测算法［J］.计算机工程,2021,47(10):298-305,313. XU Xianfeng,ZHAO Wanfu,ZOU Haoquan,etal.Helmet wearing detection algorithm based on MobileNet-SSD［J］.Computer Engineering,2021,47(10):298-305,313.
［14］ HOWARD A,ZHU M,CHEN B,et al.MobileNets:efficient convolutional neural networks for mobile vision applications［J］.Computer Vision and Pattern Recognition,2017，7（4）：12670695.
［15］ HE K M,ZHANG X Y,REN S Q,et al.Deep residual learning for image recognition［C］//2016 IEEE Conference on Computer Vision and Pattern Recognition.Las Vegas:IEEE,2016.
［16］ LIU S,QI L,QIN H F,et al.Path aggregation network for instance segmentation［C］//2018 IEEE Conference on Computer Vision and Pattern Recognition.Salt Lake City:IEEE,2018.
［17］ WOO S,PARK J,LEE J Y,et al.CBAM:convolutional block attention module［M］.Lecture Notes in Computer Science.Berlin:Springer,2018.
［18］ WU Y,CHEN Y P,YUAN L,et al.Rethinking classification and localization for object detection［C］//2020 IEEE Conference on Computer Vision and Pattern Recognition.Seattle:IEEE,2020.
［19］ EVERINGHAM M,VAN GOOL L,WILLIAMS C K I,et al.Thepascal visual object classes (VOC) challenge［J］.International Journal of Computer Vision,2010，88:303-338.
［20］ LIN T Y,MAIRE M,BELONGIE S,et al.Microsoft COCO:common objects in context［M］.Lecture Notes in Computer Science.Berlin:Springer,2014.
［21］ DENG J,DONG W,SOCHER R,et al.ImageNet:a large-scale hierarchical image database［C］//2009 IEEE Conference on Computer Vision and Pattern Recognition.Miami:IEEE,2009.
［22］ LI Y H,CHEN Y T,WANG N Y,et al.Scale-aware trident networks for object detection［C］//2019 IEEE International Conference on Computer Vision.Seoul:IEEE,2019.

相似文献/References:

[1]何川,刘功智,任智刚,等.企业安全生产分级监管模型研究*[J].中国安全生产科学技术,2011,7(2):84.
　HE Chuan,LIU Gong zhi,REN Zhi gang,et al.Study on grading supervision model of work safety in enterprises[J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2011,7(8):84.
[2]吴凯.安全生产标准化激励约束机制的建立探讨[J].中国安全生产科学技术,2011,7(2):164.
　WU Kai.Discussion on the establishment of incentive and restraint mechanisms for work safety standardization[J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2011,7(8):164.
[3]刘铁民,任智刚.安全生产“十二五”规划若干统计指标考量[J].中国安全生产科学技术,2011,7(5):5.
　LIU Tie-min,REN Zhi-gang.Research on several indicators in the twelfth five-year program for work safety[J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2011,7(8):5.
[4]李小杰,叶成明,刘迎娟,等.地质勘查劳动防护和野外救生、特殊生活用品（用具）配备标准编制要点[J].中国安全生产科学技术,2011,7(6):63.
　LI Xiao-jie,YE Cheng-ming,LIU Ying-juan,et al.Introduction on standard for labor protective equipment, field lifesaving supplies and special living supplies in China geological exploration[J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2011,7(8):63.
[5]李绍亮,王用杰,徐冠武.企业安全生产行为路径管理探讨[J].中国安全生产科学技术,2011,7(6):176.
　LI Shao-liang,WANG Yong-jie,XU Guan-wu.Approach a subject of behaviour path management[J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2011,7(8):176.
[6]张楠,邹昭晞.创建本质安全型煤矿探讨[J].中国安全生产科学技术,2011,7(6):180.
　ZHANG Nan,ZOU Zhao-xi.Probe into the establishment of essence safety coal mine[J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2011,7(8):180.
[7]刘斌,杨辉,许文,等.航空工业安全生产标准体系研究[J].中国安全生产科学技术,2011,7(7):106.
　LIU BinYANG HuiXU WenZHANG Lin.Study on [J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2011,7(8):106.
[8]杨乃莲.对改进我国生产安全事故统计工作的思考[J].中国安全生产科学技术,2011,7(7):159.
　YANG Nai-lian.Study on improving statistics of work safety accidents accidents in China[J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2011,7(8):159.
[9]梅国栋.基于现代数学思想对安全生产控制指标分解方法的研究[J].中国安全生产科学技术,2011,7(8):60.
　MEI Guo-dong.Research on decomposition method of control index in safety production based on modern mathematics[J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2011,7(8):60.
[10]曾明荣,李钢,吴宗之,等.安全生产“十一五”规划实施评估技术研究[J].中国安全生产科学技术,2011,7(8):104.
　ZENG Ming-rong,LI Gang,WU Zong-zhi,et al.Evaluation on progress of carrying out national eleventh five-year program for work safety[J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2011,7(8):104.

备注/Memo

备注/Memo:
收稿日期： 2024-03-04
* 基金项目： 国家重点研发计划项目（2021YFC3001304）
作者简介： 韩飞腾，博士,工程师，主要研究方向为基于计算机视觉的安全生产风险检测、自然灾害风险检测等。
通信作者： 郭玮，博士，高级工程师，主要研究方向为安全生产风险检测、人群疏散、公共安全等。
更新日期/Last Update: 2024-08-26