面向煤矿井下作业场景的安全帽佩戴识别算法

doi:10.16265/j.cnki.issn1003-3033.2024.03.1985

摘要/Abstract

摘要：

为提高煤矿井下作业场景的工人安全帽佩戴识别准确度,提出一种人体姿态分析的安全帽佩戴识别算法和机器视觉系统优化方法。首先,选择单步多框目标检测(SSD)作为多目标识别的基础模型,利用网络压缩模型(SqueezeNet)削减改进模型参数,以形成高效的识别模型,在提高对工人安全帽的识别准确度的同时,维持识别准确度与计算速度之间的平衡;然后,引入多人姿态估计算法,定位每个工人的人体关节点,判断工人的复杂行为状态;最后,基于多目标识别与多人姿态估计的融合模型,提取目标人员的上肢关节点,结合其与安全帽包围框的空间拓扑结构关系,判定安全帽的佩戴情况,选取3 000条图像数据验证所提方法的有效性。结果表明:机器视觉系统优化方法能优化硬、软件配置,提高系统性能。该方法识别准确度达到91.1%,明显优于最新的安全帽佩戴识别算法,能够满足煤矿井下作业的工人安全帽佩戴识别准确度要求。

关键词: 煤矿井下, 安全帽佩戴, 识别算法, 姿态分析, 模型优化

Abstract:

To improve the accuracy of miners' helmet-wearing recognition in coal mine underground operations, a helmet-wearing recognition algorithm was proposed based on human posture analysis and machine vision system optimization methods. First, Single Shot MultiBox Detector (SSD) was used as the basic model of multi-target identification. The squeezed neural network (SqueezeNet) was used to reduce the model parameters to develop an efficient recognition model, which improved the recognition accuracy of the miners' helmet and maintained the balance between the recognition accuracy and the calculation speed. Then, a multi-person posture estimation algorithm was used to locate the joint points of each miner and determine the miners' complex behavioral status. Finally, the upper limb nodes of the target were extracted based on fusion model of multi-target recognition and multi-person posture estimation, and then the helmet-wearing condition was determined by the spatial topological relationship between the upper limb nodes and the helmet frame. Moreover, 3 000 pieces of image data were selected to validate the proposed method's performance. The results indicated that the machine vision system can optimize the hardware and software configuration to improve the system's performance. Furthermore, the identification accuracy reached up to 91.1%, which was much better than that of the latest helmet-wearing recognition algorithm. Therefore, the proposed system in this study can meet the requirements of helmet-wearing recognition accuracy for underground coal miners.

Key words: coal mine underground, helmet-wearing, recognition algorithm, posture analysis, model optimization

中图分类号:

X936

左明成, 焦文华. 面向煤矿井下作业场景的安全帽佩戴识别算法[J]. 中国安全科学学报, 2024, 34(3): 237-246.

ZUO Mingcheng, JIAO Wenhua. Helmet-wearing recognition algorithm for coal mine underground operation scenarios[J]. China Safety Science Journal, 2024, 34(3): 237-246.

图/表 14

图1

图2

图3

表1

图4

图5

图6

图7

图8

图9

表2

图10

表3

图11

参考文献 18

[1]	吴天舒, 张志佳, 刘云鹏. 基于改进SSD的轻量化小目标检测算法[J]. 红外与激光工程, 2018, 47(7):37-43.
	WU Tianshu, ZHANG Zhijia, LIU Yunpeng. Lightweight small target detection algorithm based on improved SSD[J]. Infrared and Laser Engineering, 2018, 47(7):37-43.
[2]	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. doi: 10.16265/j.cnki.issn1003-3033.2022.05.0714
	ZHAO Hongcheng, TIAN Xiuxia, YANG Zesen, et al. Improved safety helmet wearing detection algorithm of YOLOv3 in complex construction environment[J]. China Safety Science Journal, 2022, 32(5):194-200. doi: 10.16265/j.cnki.issn1003-3033.2022.05.0714
[3]	颜银慧, 周正锦, 向涛. 基于SqueezeNet的改进SSD模型及其应用[J]. 中国交通信息化, 2018(增1): 189-191.
	YAN Yinhui, ZHOU Zhengjin, XIANG Tao. Improved SSD model and its application based on SqueezeNet[J]. China Transportation Informatization, 2018(S1): 189-191.
[4]	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 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
[5]	刘群, 刘鸿涛, 田小锋. 一种适用于变电站内智能视频安全帽识别算法的研究与应用[J]. 黑龙江科技信息, 2016(31): 26-27.
	LIU Qun, LIU Hongtao, TIAN Xiaofeng. Research and application of an intelligent video safety hat recognition algorithm in substation[J]. Heilongjiang Science and Technology Information, 2016(31): 26-27.
[6]	毕林, 谢伟, 崔君. 基于卷积神经网络的矿工安全帽佩戴识别研究[J]. 黄金科学技术, 2017, 25(4):73-80.
	BI Lin, XIE Wei, CUI Jun. Research on miners' safety helmet wear recognition based on convolutional neural network[J]. Gold Science and Technology, 2017, 25(4):73-80.
[7]	张博, 宋元斌, 熊若鑫, 等. 融合人体关节点的安全帽佩戴检测[J]. 中国安全科学学报, 2020, 30(2):177-182. doi: 10.16265/j.cnki.issn1003-3033.2020.02.028
	ZHANG Bo, SONG Yuanbin, XIONG Ruoxin, et al. Helmet wear detection with human body junction[J]. China Safety Science Journal, 2020, 30(2):177-182. doi: 10.16265/j.cnki.issn1003-3033.2020.02.028
[8]	王雨生, 顾玉宛, 庄丽华, 等. 复杂姿态下的安全帽佩戴检测方法研究[J]. 计算机工程与应用, 2022, 58(1):190-196. doi: 10.3778/j.issn.1002-8331.2007-0218
	WANG Yusheng, GU Yuwan, ZHUANG Lihua, et al. Research on helmet wearing detection method under complex posture[J]. Computer Engineering and Applications, 2022, 58(1):190-196. doi: 10.3778/j.issn.1002-8331.2007-0218
[9]	LIU Wei, DRAGOMIR A, DUMITRU E, et al. SSD: single shot multibox detector[C]. European Conference on Computer Vision, Amsterdam, Springer International Publishing, 2016:21-37.
[10]	FANG Haoshu, XIE Shuqin, TAI Yuwing, et al. RMPE: regional multi-person pose estimation[C]. International Conference on Computer Vision, 2017:2334-2343.
[11]	WOMG A, SHAFIEE M, LI Francis, et al. Tiny SSD: a tiny single-shot detection deep convolutional neural network for real-time embedded object detection[C]. IEEE Conference on Computer and Robot Vision, 2018:95-101.
[12]	KIIRE S. Effect of leg-to-body ratio on body shape attractiveness[J]. Archives of Sexual Behavior, 2016, 45: 901-910. doi: 10.1007/s10508-015-0635-9 pmid: 26474977
[13]	LOU Haitong, DUAN Xuehu, GUO Junmei, et al. DC-YOLOv8: small-size object detection algorithm based on camera sensor[J]. Electronics, 2023, 12(10): DOI: 10.3390/electronics12102323.
[14]	WANG Chienyao, BOCHKOVSKIY A, LIAO Mark. YOLOv7: trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[C]. IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023:7464-7475.
[15]	LI Xingke, WANG Yikai, QU Jiaxin, et al. Intelligent monitoring method of helmet wearing identification in power plants[C]. Asia Conference on Power and Electrical Engineering, 2023:1699-1703.
[16]	GUAN Yaqi, LI Wenqiang, HU Tianyu, et al. Design and implementation of safety helmet detection system based on YOLOv5[C]. Asia Conference on Computers and Communications, 2021:69-73.
[17]	王忠玉. 智能视频监控下的安全帽佩戴检测系统的设计与实现[D]. 北京: 北京邮电大学, 2018.
	WANG Zhongyu. Design and Implementation of helmet wearing detection system under intelligent Video surveillance[D]. Beijing: Beijing University of Posts and Telecommunications, 2018.
[18]	杜思远. 变电站人员安全帽佩戴识别算法研究[D]. 重庆: 重庆大学, 2017.
	DU Siyuan. Research on safety helmet wearing recognition algorithm for substation personnel[D]. Chongqing: Chongqing University, 2017.

层	尺寸与个数	层	尺寸与个数
卷积层2-1	3×3×51	火模型9-Loc	3×3×24
卷积层2-2	3×3×46	火模型9-Con	3×3×12
卷积层3-1	3×3×55	火模型10-Loc	3×3×24
卷积层3-2	3×3×85	火模型10-Con	3×3×12
火模型4-Loc	3×3×16	卷积层2-2-Loc	3×3×24
火模型4-Con	3×3×8	卷积层2-2-Con	3×3×12
火模型8-Loc	3×3×24	卷积层3-2-Loc	3×3×16
火模型8- Con	3×3×12	卷积层3-2-Con	3×3×8

算法	识别准确率/%	行为检测效果
所提算法	91.1	可检测全部行为
YOLOv8	86.8	仅站立、坐立
YOLOv7	85.1	仅站立、坐立
增强的Faster R-CNN	86.6	可检测全部行为
融合姿态分析的YOLOv4	87.9	可检测全部行为
改进YOLOv5	87.2	仅站立、坐立
多策略YOLOv5	85.7	仅站立、坐立
Faster R-CNN	84.3	仅站立、坐立
Vibe-DPM	80	仅站立、坐立

模型	精度/%	尺寸/MB	速度/(帧·s^-1)
原始SSD模型	45	82.1	33
微调SSD模型	72.7	82.1	33
SqueezeNet-SSD	91.1	21.1	100