Unsafe behavior recognition model of high climbing workers based on vision

doi:10.16265/j.cnki.issn1003-3033.2025.02.0278

Abstract

Abstract:

In order to accurately identify unsafe behaviors during the climbing process of high-altitude workers, this paper proposed an unsafe behavior recognizing method for high climbing workers based on vision, which included the human pose estimation and the one-dimensional convolutional unsafe behavior recognition models. Quantized autoencoder was used to structurally model human key points in human pose estimation, enabling the detection of human key point coordinates. Combining with safety behavior knowledge in high climbing operations, the unsafe behavior recognition model was constructed based on one-dimensional convolutional neural network model, and it was validated by industrial data experiments. Experimental results show that the accuracy of this method is 93.91% and 90.34% on unobstructed and partially obstructed datasets, respectively. Moreover, compared with support vector machines (SVM) and K-nearest neighbor (KNN), this method has stronger generalization capability.

Key words: computer vision, worker high-altitude climbing, unsafe behavior, recognition model, one-dimensional convolutional neural network(1DCNN), human pose estimation

CLC Number:

X910安全人体学

ZHANG Zehui, ZHANG Qianlong, XU Xiaobin, ZHAO Zuguo, WANG Haiquan, LI Hao. Unsafe behavior recognition model of high climbing workers based on vision[J]. China Safety Science Journal, 2025, 35(2): 144-151.

Figures/Tables 14

Fig.1

Fig.2

Fig.3

Fig.4

Fig.5

Table 1

Fig.6

Fig.7

Fig.8

Fig.9

Table 2

Fig.10

Fig.11

Fig.12

References 24

[1]	国家统计局. 中华人民共和国2023年国民经济和社会发展统计公报[EB/OL].(2024-02-29). https://www.stats.gov.cn/xxgk/sjfb/tjgb2020/202402/t20240229_1947923.html.
[2]	SABRINA S O, WILLAMES D A S, BIANCA M V. Fatal fall-from-height accidents: statistical treatment using the human factors analysis and classification system-HFACS[J]. Journal of Safety Research, 2023: DOI: 10.1016/j.jsr.2023.05.004.
[3]	YANG Jingjing, YE Gui, XIANG Qingting, et al. Insights into the mechanism of construction workers' unsafe behaviors from an individual perspective[J]. Safety Science, 2021: DOI: 10.1016/j.ssci.2020.105004.
[4]	郑霞忠, 王晓宇, 陈述, 等. 高处坠落事故的人因失误与干预策略研究[J]. 中国安全生产科学技术, 2017, 13(6):139-144.
	ZHENG Xiazhong, WANG Xiaoyu, CHEN Shu, et al. Study on human error and intervention strategies of high falling accidents[J]. Journal of Safety Science and Technology, 2017, 13(6):139-144.
[5]	ZHOU Longfei, ZHANG Lin, KONZ N. Computer vision techniques in manufacturing[J]. IEEE Transactions on Systems, Man, and Cybernetics: Systems, 2022, 53(1):105-117.
[6]	伍麟, 郝鸿宇, 宋友. 基于计算机视觉的工业金属表面缺陷检测综述[J]. 自动化学报, 2024, 50(7):1261-1283.
	WU Lin, HAO Hongyu, SONG You. A review of metal surface defect detection based on computer vision[J]. Acta Automatica Sinica, 2024, 50(7):1261-1283.
[7]	FANG Weili, DING Lieyun, LOVE P E D, et al. Computer vision applications in construction safety assurance[J]. Automation in Construction, 2020: DOI: 10.1016/j.autcon.2019.103013.
[8]	张萌, 韩豫, 刘泽锋. 深度学习下建筑工人高空安全防护装备检测方法[J]. 中国安全科学学报, 2022, 32(5):140-146. doi: 10.16265/j.cnki.issn1003-3033.2022.05.1141
	ZHANG Meng, HAN Yu, LIU Zefeng. Detection method of high-altitude safety protective equipment for construction workers based on deep learning[J]. China Safety Science Journal, 2022, 32(5):140-146. doi: 10.16265/j.cnki.issn1003-3033.2022.05.1141
[9]	程淑红, 马晓菲, 张仕军, 等. 基于多任务分类的吸烟行为检测[J]. 计量学报, 2020, 41(5):538-543.
	CHENG Shuhong, MA Xiaofei, ZHANG Shijun, et al. Smoking detection algorithm based on multitask classification[J]. Acta Metrologica Sinica, 2020, 41(5):538-543.
[10]	LI Li, ZHONG Boxuan, HUTMACHER J C, et al. Detection of driver manual distraction via image-based hand and ear recognition[J]. Accident Analysis & Prevention, 2020:DOI: 10.1016/j.aap.2020.105432.
[11]	ZHANG Xiaojun. Application of human motion recognition utilizing deep learning and smart wearable device in sports[J]. International Journal of System Assurance Engineering and Management, 2021, 12(4):835-843. doi: 10.1007/s13198-021-01118-7
[12]	XIE Yadong, LI Fan, WU Yue, et al. HearSmoking: smoking detection in driving environment via acoustic sensing on smartphones[J]. IEEE Transactions on Mobile Computing, 2021, 21(8):2847-2860.
[13]	SHEN Jie, XIONG Xin, LI Ying, et al. Detecting safety helmet wearing on construction sites with bounding-box regression and deep transfer learning[J]. Computer-Aided Civil and Infrastructure Engineering, 2021, 36(2):180-196.
[14]	SENER A S, INCE I F, BAYDARGIL H B, et al. Deep learning based automatic vertical height adjustment of incorrectly fastened seat belts for driver and passenger safety in fleet vehicles[J]. Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering, 2022, 236(4):639-654.
[15]	GB 30871—2022,危险化学品企业特殊作业安全规范[S].
	GB 30871-2022, Safety specifications of special work in hazardous chemicals enterprises[S].
[16]	刘勃, 孔韦韦, 肖家钦, 等. 基于RBF神经网络的跌倒检测算法[J]. 计算机技术与发展, 2022, 32(6):167-178.
	LIU Bo, KONG Weiwei, XIAO Jiaqin, et al. Fall detection algorithm based on rbf neural network[J]. Computer Technology and Development, 2022, 32(6):167-178.
[17]	GENG Zigang, WANG Chunyu, WEI Yixuan, et al. Human pose as compositional tokens[C]. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023: DOI: 10.1109/CVPR52729.2023.00071.
[18]	HU Qiyu, ZHANG Guangyi, QIN Zhijin, et al. Robust semantic communications with masked VQ-VAE enabled codebook[J]. IEEE Transactions on Wireless Communications, 2023, 22(12):8707-8722.
[19]	LIU Ze, LIN Yutong, CAO Yue, et al. Swin transformer: hierarchical vision transformer using shifted windows[C]. Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021:DOI: 10.1109/ICCV48922.2021.00986.
[20]	YU Jianbo, ZHOU Xingkang. One-dimensional residual convolutional autoencoder based feature learning for gearbox fault diagnosis[J]. IEEE Transactions on Industrial Informatics, 2020, 16(10):6347-6358.
[21]	池毅, 陈光武. 基于一维卷积神经网络的实时道岔故障诊断[J]. 计算机工程与应用, 2022, 58(20):293-299. doi: 10.3778/j.issn.1002-8331.2103-0348
	CHI Yi, CHEN Guangwu. Real-time turnout fault diagnosis based on one-dimensional convolutional neural network[J]. Computer Engineering and Applications, 2022, 58(20):293-299. doi: 10.3778/j.issn.1002-8331.2103-0348
[22]	SONG Yifan, ZHANG Zhang, SHAN Caifeng, et al. Constructing stronger and faster baselines for skeleton-based action recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022, 45(2):1474-1488.
[23]	SONG Yifan, ZHANG Zhang, SHAN Caifeng, et al. Richly activated graph convolutional network for robust skeleton-based action recognition[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2020, 31(5):1915-1925.
[24]	ZHAO Baining, LAN Haijuan, NIU Zhenwen, et al. Detection and location of safety protective wear in power substation operation using wear-enhanced YOLOv3 algorithm[J]. IEEE Access, 2021: DOI: 10.1109/ACCESS.2021.3104731.

网络模型参数	参数配置值
检测数	2
输入尺寸	(34,1)
批大小	50
优化器	Adam
学习率	0.001
卷积层数	2
池化层数	2
损失函数	交叉熵损失
训练轮次	80

检测模型	A	P	R
PCT-SVM	88.44	81.63	73.97
PCT-KNN	89.12	81.85	74.31
PCT-1DCNN	92.18	96.16	95.32