摘 要:人为因素是煤矿事故的最主要原因,预防和制止煤矿井下人员的不安全行为是降低事故率的根本途径。针对违规进入危险区域、吸烟行为、未佩戴安全帽三种主要不安全行为,提出了一种结合 AlphaPose 和 YOLOv5 的不安全行为监测方法,实现公共模块复用,并建立不安全行为判定算法。经过实验对比分析,对上述三种不安全行为监测的准确率分别达到 92.33%、92.76%、95.12%,监测速率为 11.91 f/s,均优于对比算法,算法具有较高的准确率和实时性。
关键词:实时监测;目标检测;姿态估计;姿态关节点;深度学习
DOI:10.19850/j.cnki.2096-4706.2022.21.018
中图分类号:TP391.4;TD76 文献标识码:A 文章编号:2096-4706(2022)21-0072-06
A Monitoring Method of Unsafe Behavior under the Shaft Based on AlphaPose
ZHENG Wen, DONG Haizhi
(School of Mechanical Electronic and Information Engineering, China University of Mining and Technology-Beijing, Beijing 100083, China)
Abstract: Human factors are the most main causes of coal mine accidents. Preventing and stopping the unsafe behavior of personnel under the coal mine shaft is the fundamental way to reduce the accident rate. Aiming at the three main unsafe behaviors of entering dangerous areas illegally, smoking behavior and not wearing a helmet, an unsafe behavior monitoring method combined with AlphaPose and YOLOv5 is proposed. It realizes common module reuse and establishes unsafe behavior judgment algorithms. After comparative experimental analysis, the accuracy rates of the above three unsafe behavior monitoring are 92.33%, 92.76% and 95.12%, respectively, and the monitoring rate is 11.91 f/s, which are all better than the comparison algorithms. The algorithm has higher accuracy and real-time performance.
Keywords: real-time monitoring; object detection; pose estimation; pose joint point; deep learning
参考文献:
[1] 任玉辉 . 煤矿员工不安全行为影响因素分析及预控研究[D]. 北京:中国矿业大学(北京),2014.
[2] 佟瑞鹏,张艳伟 . 人工智能技术在矿工不安全行为识别中的融合应用 [J]. 中国安全科学学报,2019,29(1):7-12.
[3] 杨蒙超,宗伟林,孙凯,等 . 基于 Hi3516 的矿井运动目标检测系统的设计与实现 [J]. 中国煤炭,2016,42(4):81-85.
[4] DALAL N,TRIGGS B.Histograms of Oriented Gradients for Human Detection [C]//2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.San Diego:IEEE,2005,1:886-893.
[5] 周飞燕,金林鹏,董军 . 卷积神经网络研究综述 [J]. 计算机学报,2017,40(6):1229-1251.
[6] 孔丽丽,易春求 . 矿用智能安全头盔的设计 [J]. 中国矿业,2020,29(12):95-98+115.
[7] 万雪芬,杨义,周红伟,等 . 用于煤矿井下安全告警的智能腰带 [J]. 电子测量技术,2014,37(9):98-103.
[8] 张黎 . 矿用本安救援及安全生产巡检物联手表 [J]. 煤矿安全,2017,48(8):80-82.
[9] 叶晨成,校景中,肖丽.基于RFID的井下人员定位系统 [J].武汉理工大学学报,2010,32(15):146-149.
[10] 陈新科,喻川,文智力 .UWB 定位技术在煤矿井下的应用 [J]. 煤炭科学技术,2018,46(S1):187-189.
[11] 唐世泽 . 面向运动评估的多深度摄像头人体姿态跟踪算法研究 [D]. 成都:电子科技大学,2021.
[12] FANG H S,XIE S Q,TAI Y W,et al.RMPE:Regional Multi-Person Pose Estimation [C]//2017 IEEE International Conference on Computer Vision.Venice:IEEE,2017:553.
[13] GKIOXARI G,HARIHARAN B,GIRSHICK R B, et al.Using K-Poselets for Detecting People and Localizing their Keypoints [C]//2014 IEEE Conference on Computer Vision and Pattern Recognition.Columbus:IEEE,2014:3582-3589.
[14] ANDRILUKA M,PISHCHULIN L,GEHLER P,et al.2D Human Pose Estimation:New Benchmark and State of the Art Analysis [C]//2014 IEEE Conference on Computer Vision and Pattern Recognition.Columbus:IEEE,2014:3686-3693.
[15] 马敬奇,雷欢,陈敏翼 . 基于 AlphaPose 优化模型的老人跌倒行为检测算法 [J]. 计算机应用,2022,42(1):294-301.
[16] HE K M,SUN J,TANG X O.Single Image Haze Removal Using Dark Channel Prior [J].IEEE Trans Pattern Anal Mach Intell, 2011,33(12):2341-53.
[17] 涂翠红 . 煤矿井下作业人员不安全行为的分析与评价[D]. 淮南:安徽理工大学,2011.
[18] 金小汉 . 煤矿瓦斯爆炸的火花诱因分析与应对措施 [J].矿业安全与环保,2008(5):66-68+80.
[19] REDMON J,FARHADI A.YOLOv3:An Incremental Improvement [J/OL].arXiv:1804.02767 [cs.CV].[2022-06-08].https:// arxiv.org/abs/1804.02767.
[20] 董文轩,梁宏涛,刘国柱,等 . 深度卷积应用于目标检测算法综述 [J]. 计算机科学与探索,2022,16(5):1025-1042.
作者简介:郑雯(2001—),女,汉族,江西赣州人,本科在读,研究方向:机器学习、信息处理;董海志(2001—),男,汉族,吉林吉林人,本科在读,研究方向:深度学习、图像识别。