煤矿井下暗光环境人员行为检测研究

董芳凯; 赵美卿; 黄伟龙

doi:10.13272/j.issn.1671-251x.2024090032

煤矿井下暗光环境人员行为检测研究

董芳凯^{1, 2,},
赵美卿^1, ,,
黄伟龙³

1.
山西工程技术学院机械工程系，山西阳泉　045000
2.
中北大学机械工程学院，山西太原　030000
3.
中国科学院海西研究院泉州装备制造研究中心，福建泉州　362000

基金项目: 山西省教育厅2022年度高等学校科技创新项目（2022L704）；阳泉市科技计划项目（2022JH051）。

详细信息

作者简介:
董芳凯（1992—），男，山西寿阳人，助教，硕士，研究方向为机器视觉与图像处理，E-mail：1025915393@qq.com

通讯作者:
赵美卿（1964—），女，山西阳泉人，教授，硕士，研究方向为智能制造、机电控制，E-mail：ruilong1210@126.com。

中图分类号: TD67
计量
- 文章访问数: 67
- HTML全文浏览量: 9
- PDF下载量: 12
出版历程
- 收稿日期: 2024-09-09
- 修回日期: 2025-01-09
- 网络出版日期: 2024-12-05
- 刊出日期: 2025-01-24

Research on mine worker behavior detection in low-light underground coal mine environments

1.
Department of Mechanical Engineering, Shanxi Institute of Technology, Yangquan 045000, China
2.
School of Mechanical Engineering, North University of China, Taiyuan 030000, China
3.
Quanzhou Institute of Equipment Manufacturing Haixi Institutes, Chinese Academy of Sciences, Quanzhou 362000, China

摘要

摘要:
煤矿井下环境复杂，对部分作业现场人员行为进行检测时易出现漏检与误检问题。针对该问题，提出了一种煤矿井下暗光环境人员行为检测方法，包括暗光环境图像增强和行为检测2个部分。暗光环境图像增强基于自校准光照学习（SCI）进行改进，由图像增强网络和校准网络构成。人员行为检测通过引入Dynamic Head检测、跨尺度融合模块和Focal−EIoU损失函数来改进YOLOv8n模型。SCI+网络增强后的图像作为人员行为检测模型检测的对象，完成井下暗光环境人员行为的检测任务。实验结果表明：① 井下暗光环境人员行为检测方法的mAP@0.5为87.6%，较YOLOv8n提升了2.5%，较SSD，Faster RCNN，YOLOv5s，RT−DETR−L分别提升了15.7%，11.5%，0.9%，4.3%。② 井下暗光环境人员行为检测方法的参数量为3.6×10⁶个，计算量为11.6×10⁹，检测速度为95.24 帧/s。 ③ 在公开数据集EXDark上，井下暗光环境人员行为检测方法的mAP@0.5为74.7%，较YOLOv8n提升了1.5%，表明该方法具有较强的泛化能力。
- 暗光环境 /
- 井下人员行为检测 /
- 自校准光照学习 /
- 图像增强 /
- SCI+网络 /
- Dynamic Head /
- 跨尺度融合模块 /
- Focal−EIoU损失函数 /
- YOLOv8n
Abstract:
The underground coal mine environment is complex, leading to missed and false detections when monitoring behaviors of mine workers under certain operational conditions. To address this issue, a method for detecting mine worker behaviors in low-light underground environments is proposed, which includes two parts: a low-light image enhancement and a behavior detection. The low-light image enhancement（SCI+） was improved based on self-calibrated illumination (SCI) learning, which consists ofan image enhancement network and a calibration network. The behavior detection improved the YOLOv8n model by incorporating the Dynamic Head detection, a cross-scale fusion module, and the Focal-EIoU loss function. Enhanced images from the SCI+ network were used as inputs to the behavior detection model to complete the tasks of mine worker behavior detection in low-light underground environments. Experimental results showed that: ① the method for mine worker behavior detection in low-light underground environments achieved an mAP@0.5 of 87.6%, representing an improvement of 2.5% over YOLOv8n, and improvements of 15.7%, 11.5%, 0.9%, and 4.3% compared to SSD, Faster RCNN, YOLOv5s, and RT-DETR-L, respectively. ② The method had a parameter count of 3.6×10⁶, a computational complexity of 11.6×10⁹, and a detection speed of 95.24 frames per second. ③ On the public EXDark dataset, the method achieved an mAP@0.5 of 74.7%, which was 1.5% higher than YOLOv8n, demonstrating strong generalization capability.
- low-light environment /
- underground mine worker behavior detection /
- self-calibrated illumination learning /
- image enhancement /
- SCI + network /
- Dynamic Head /
- cross-scale fusion module /
- Focal-EIoU loss function /
- YOLOv8n

HTML全文

图 1 井下暗光环境人员行为检测网络结构

Figure 1. Structure of mine worker behavior detection network structure in low-light underground environments

下载: 全尺寸图片幻灯片

图 2 SCI+网络结构

Figure 2. SCI+ network structure

下载: 全尺寸图片幻灯片

图 3 校准网络结构

Figure 3. Calibrateion network structure

下载: 全尺寸图片幻灯片

图 4 Dynamic Head 检测过程

Figure 4. Detection process of Dynamic Head

下载: 全尺寸图片幻灯片

图 5 Dynamic Head网络结构

Figure 5. Network structure of Dynamic Head

下载: 全尺寸图片幻灯片

图 6 改进的YOLOv8n Neck网络结构

Figure 6. Structure of improved YOLOv8n Neck network

下载: 全尺寸图片幻灯片

图 7 井下人员检测结果对比

Figure 7. Comparison results of underground mine worker detection

下载: 全尺寸图片幻灯片

图 8 不同模型R−P曲线检测结果

Figure 8. Detection results of R-P curves of different models

下载: 全尺寸图片幻灯片

图 9 热力图对比

Figure 9. Comparison of thermal maps

下载: 全尺寸图片幻灯片

图 10 各图像增强算法图像增强结果对比

Figure 10. Comparison of image enhancement results of different algorithms

下载: 全尺寸图片幻灯片

图 11 各图像增强算法的检测结果对比

Figure 11. Comparison of detection results of different algorithms

下载: 全尺寸图片幻灯片

图 12 各模型可视化检测结果对比

Figure 12. Comparison of visualization detection results of different models

下载: 全尺寸图片幻灯片

图 13 EXDark数据集检测可视化结果

Figure 13. Visualization results of EXDark dataset detection

下载: 全尺寸图片幻灯片

表 1 实验平台

Table 1 Experimental platform

实验平台	版本编号
系统	Ubuntu 20.04
CPU	Intel（R） Xeon（R） E5−2666 v3
GPU	NVIDIA 4060Ti
编程语言	Python3.10
计算设备架构	CUDA 11.6
学习框架	PyTorch−1.12.0

下载: 导出CSV

表 2 消融实验结果比较

Table 2 Comparison of ablation experimental results

模型	P/%	mAP@0.5/%	计算量/10⁹	参数量/10⁶个
YOLOv8n	84.3	85.5	8.7	3.2
M1	85.4	85.7	9.6	3.6
M2	85.7	84.6	8.3	2.7
M3	86.1	86.2	8.7	3.2
M4	86.8	85.1	10.9	3.4
M5	87.2	87.0	14.5	3.7
M6	85.3	84.2	13.4	3.2
M7	88.0	87.6	11.6	3.6

下载: 导出CSV

表 3 各图像增强算法检测性能比较

Table 3 Comparison of detection performance of different algorithms

增强网络	mAP@0.5/%	帧速率/（帧·s⁻¹）
LIME	85.9	2.00
MBLLEN	86.7	0.07
RetinexNet	85.6	7.78
Zero−DCE++	84.9	88.50
SCI	86.1	94.21
SCI+	87.6	95.24

下载: 导出CSV

表 4 主流检测模型比较

Table 4 Comparison of mainstream detection models

模型	P/%	R/%	mAP0.5/%	计算量/10⁹	参数量 /10⁶个
SSD	74.2	73.5	75.7	275.8	24.4
Faster RCNN	73.5	77.8	78.6	401.9	136.8
YOLOv5s	84.4	83.9	86.8	16.0	7.0
RT−DETR−L	89.0	82.7	84.0	108.3	32
文献[10]模型	84.7	82.7	87.0	8.5	3.3
文献[31]模型	86.7	79.2	85.3	12.5	4.3
文献[32]模型	85.2	83.4	87.3	68	21.2
人员行为检测模型	88.0	83.2	87.6	11.6	3.6

下载: 导出CSV

表 5 EXDark数据集检测结果比较

Table 5 Comparison of EXDark dataset detection results

模型	mAP@0.5/%	帧速率/（帧·s⁻¹）
YOLOv8n	73.7	75.32
井下暗光环境人员行为检测模型	74.7	81.70

下载: 导出CSV

参考文献(32)

[1]	付恩三,白润才,刘光伟,等. “十三五”期间我国煤矿事故特征及演变趋势分析[J]. 中国安全科学学报,2022,32(12):88-94. FU Ensan,BAI Runcai,LIU Guangwei,et al. Analysis on characteristics and evolution trend of coal mine accidents in our country during "13(th) five-year" plan period[J]. China Safety Science Journal,2022,32(12):88-94.
[2]	王宇,于春华,陈晓青,等. 基于多模态特征融合的井下人员不安全行为识别[J]. 工矿自动化,2023,49(11):138-144. WANG Yu,YU Chunhua,CHEN Xiaoqing,et al. Recognition of unsafe behaviors of underground personnel based on multi modal feature fusion[J]. Journal of Mine Automation,2023,49(11):138-144.
[3]	王国法,徐亚军,张金虎,等. 煤矿智能化开采新进展[J]. 煤炭科学技术,2021,49(1):1-10. WANG Guofa,XU Yajun,ZHANG Jinhu,et al. New development of intelligent mining in coal mines[J]. Coal Science and Technology,2021,49(1):1-10.
[4]	黄瀚,程小舟,云霄,等. 基于DA−GCN的煤矿人员行为识别方法[J]. 工矿自动化,2021,47(4):62-66. HUANG Han,CHENG Xiaozhou,YUN Xiao,et al. DA-GCN-based coal mine personnel action recognition method[J]. Industry and Mine Automation,2021,47(4):62-66.
[5]	刘浩,刘海滨,孙宇,等. 煤矿井下员工不安全行为智能识别系统[J]. 煤炭学报,2021,46(增刊2):1159-1169. LIU Hao,LIU Haibin,SUN Yu,et al. Intelligent recognition system of unsafe behavior of underground coal miners[J]. Journal of China Coal Society,2021,46(S2):1159-1169.
[6]	温廷新,王贵通,孔祥博,等. 基于迁移学习与残差网络的矿工不安全行为识别[J]. 中国安全科学学报,2020,30(3):41-46. WEN Tingxin,WANG Guitong,KONG Xiangbo,et al. Identification of miners' unsafe behaviors based on transfer learning and residual network[J]. China Safety Science Journal,2020,30(3):41-46.
[7]	李伟山,卫晨,王琳. 改进的Faster RCNN煤矿井下行人检测算法[J]. 计算机工程与应用,2019,55(4):200-207. DOI: 10.3778/j.issn.1002-8331.1711-0282 LI Weishan,WEI Chen,WANG Lin. Improved Faster RCNN approach for pedestrian detection in underground coal mine[J]. Computer Engineering and Applications,2019,55(4):200-207. DOI: 10.3778/j.issn.1002-8331.1711-0282
[8]	延晓宇,董立红,厍向阳,等. 改进的FCOS煤矿井下行人检测算法[J]. 矿业研究与开发,2022,42(4):160-165. YAN Xiaoyu,DONG Lihong,SHE Xiangyang,et al. Improved FCOS pedestrian detection algorithm in underground coal mine[J]. Mining Research and Development,2022,42(4):160-165.
[9]	SHAO Xiaoqiang,LIU Shibo,LI Xin,et al. Rep-yolo:an efficient detection method for mine personnel[J]. Journal of Real-Time Image Processing,2024,21(2). DOI: 10.1007/S11554-023-01407-3.
[10]	XIN Fangfang,HE Xinyu,YAO Chaoxiu,et al. A real-time detection for miner behavior via DYS-YOLOv8n model[J]. Journal of Real-Time Image Processing,2024,21(3). DOI: 10.1007/S11554-024-01466-0.
[11]	WANG Zheng,LIU Yan,DUAN Siyuan,et al. An efficient detection of non-standard miner behavior using improved YOLOv8[J]. Computers and Electrical Engineering,2023,112. DOI: 10.1016/J.COMPELECENG.2023.109021.
[12]	孙亚琳,孙鹏翔,薛晔,等. 基于SCI−XDNet−CFF轻量化网络的井下运煤皮带异物识别[J]. 煤矿现代化,2025,34(1):40-46,51. SUN Yalin,SUN Pengxiang,XUE Ye,et al. Identification of foreign objects in underground coal transportation belt based on SCI-XDNet-CFF lightweight networks[J]. Coal Mine Modernization,2025,34(1):40-46,51.
[13]	REDMON J,DIVVALA S,GIRSHICK R,et al. You only look once:unified,real-time object detection[C]. IEEE Conference on Computer Vision and Pattern Recognition,Las Vegas,2016. DOI: 10.1109/CVPR.2016.91.
[14]	GIRSHICK R. Fast R−CNN[C]. IEEE International Conference on Computer Vision,Santiago,2015. DOI: 10.1109/ICCV.2015.169.
[15]	HE K M,GKIOXARI G,DOLLAR P,et al. Mask R−CNN[C]. Proceedings of the IEEE International Conference on Computer Vision,Piscataway,2017:2980-2988.
[16]	WANG Ao,CHEN Hui,LIU Lihao,et al. Yolov10:Real-time end-to-end object detection[EB/OL]. arxiv preprint arxiv:2405.14458,2024. https://arxiv.org/abs/2405.14458v2.
[17]	VARGHESE R,SAMBATH M. YOLOv8:a novel object detection algorithm with enhanced performance and robustness[C]. International Conference on Advances in Data Engineering and Intelligent Computing Systems (ADICS),Chennai,2024. DOI: 10.1109/ADICS58448.2024.10533619.
[18]	DAI Xiyang,CHEN Yinpeng,XIAO Bin,et al. Dynamic head:unifying object detection heads with attentions[C]. IEEE/CVF Conference on Computer Vision and Pattern Recognition,Nashville,2021:7373-7382.
[19]	赵小虎,张狄,谢礼逊,等. 基于改进YOLOv8的煤矿皮带异物检测方法[J/OL]. 工程科学与技术,2025:1-16,[2024-08-14]. https://kns.cnki.net/kcms/detail/51.1773.tb.20250114.1313.001.html. ZHAO Xiaohu,ZHANG Di,XIE Lixun,et al. Detection method of foreign body in coal mine belt based on improved YOLOv8[J/OL]. Advanced Engineering Sciences,2025:1-16,[2024-08-14]. https://kns.cnki.net/kcms/detail/51.1773.tb.20250114.1313.001.html.
[20]	NING Shiyong,HAN Xu. An improved YOLOv8-based safety helmet wearing detection algorithm[C]. 7th International Conference on Computer Information Science and Application Technology ,Hangzhou,2024:75-79.
[21]	ZHENG Zhaohui,WANG Ping,REN Dongwei,et al. Enhancing geometric factors in model learning and inference for object detection and instance segmentation[J]. IEEE Transactions on Cybernetics,2021,52(8):8574-8586.
[22]	ZHANG Yifan,REN Weiqiang,ZHANG Zhang,et al. Focal and efficient IoU loss for accurate bounding box regression[J]. Neurocomputing,2022,506:146-157. DOI: 10.1016/j.neucom.2022.07.042
[23]	GUO Xiaojie,LI Yu,LING Haibin. LIME:low-light image enhancement via illumination map estimation[J]. IEEE Transactions on Image Processing,2016,26(2):982-993.
[24]	田子建,阳康,吴佳奇,等. 基于LMIENet图像增强的矿井下低光环境目标检测方法[J]. 煤炭科学技术,2024,52(5):222-235. DOI: 10.12438/cst.2023-0675 TIAN Zijian,YANG Kang,WU Jiaqi,et al. LMIENet enhanced object detection method for low light environment in underground mines[J]. Coal Science and Technology,2024,52(5):222-235. DOI: 10.12438/cst.2023-0675
[25]	WEI Chen,WANG Wenjing,YANG Wenhan,et al. Deep retinex decomposition for low-light enhancement[EB/OL]. [2024-08-14]. https://arxiv.org/abs/1808.04560v1.
[26]	LI Chongyi,GUO Chunle,CHEN C L. Learning to enhance low-light image via zero-reference deep curve estimation[J]. IEEE Transactions on Software Engineering,2021. DOI: 10.1109/TPAMI.2021.3063604.
[27]	LIU Wei,ANGUELOV D,ERHAN D,et al. SSD:single shot multibox detector[J]. CoRR,2015. DOI: 10.1007/978-3-319-46448-0_2.
[28]	CHEN Xinlei,GUPTA A. An implementation of faster rcnn with study for region sampling[EB/OL]. [2024-08-06]. https://arxiv.org/abs/1702.02138v1.
[29]	YANG Guanhao,FENG Wei,JIN Jintao,et al.Face mask recognition system with YOLOV5 based on image recognition[C].IEEE 6th International Conference on Computer and Communications ,Chengdu,2020:1398-1404.
[30]	ZHAO Yian,LYU Wenyu,XU Shangliang,et al. DETRs beat YOLOs on real-time object detection[C]. IEEE/CVF Conference on Computer Vision and Pattern Recognition,Seattle,2024:16965-16974.
[31]	寇发荣,肖伟,何海洋,等. 基于改进YOLOv5的煤矿井下目标检测研究[J]. 电子与信息学报,2023,45(7):2642-2649. DOI: 10.11999/JEIT220725 KOU Farong,XIAO Wei,HE Haiyang,et al. Research on target detection in underground coal mines based on improved YOLOv5[J]. Journal of Electronics & Information Technology,2023,45(7):2642-2649. DOI: 10.11999/JEIT220725
[32]	陈伟,江志成,田子建,等. 基于YOLOv8的煤矿井下人员不安全动作检测算法[J/OL]. 煤炭科学技术:1-19[2024-08-02]. http://kns.cnki.net/kcms/detail/11.2402.td.20240322.1343.003.html. CHEN Wei,JIANG Zhicheng,TIAN Zijian,et al. Unsafe action detection algorithm of underground personnel in coal mine based on YOLOv8[J/OL]. Coal Science and Technology:1-19[2024-08-02]. http://kns.cnki.net/kcms/detail/11.2402.td.20240322.1343.003.html.

施引文献(10)

期刊类型引用(6)

1.	陈书航，王世博，葛世荣，王赟，马广军. 综采工作面刮板输送机煤流时空分布研究. 工矿自动化. 2024(09): 98-107 . 本站查看
2.	王桂忠，叶隆浩. 基于煤流量识别的带式输送机节能控制系统设计与研究. 煤矿机械. 2023(01): 14-17 . 百度学术
3.	周爱民，叶飞，施旭东，赵培成. 基于超声波传感器的带式输送机烟丝瞬时流量监测系统的设计. 现代信息科技. 2022(03): 149-152 . 百度学术
4.	唐文杰. 带式输送机运行状态智能监控体系的研究. 机械管理开发. 2022(09): 300-301+304 . 百度学术
5.	毛清华，毛金根，马宏伟，张旭辉，李铮. 矿用带式输送机智能监测系统研究. 工矿自动化. 2020(06): 48-52+58 . 本站查看
6.	李瑶，王义涵. 带式输送机煤流量自适应检测方法. 工矿自动化. 2020(06): 98-102 . 本站查看