基于MES−YOLOv5s的综采工作面大块煤检测算法

徐慈强; 贾运红; 田原

doi:10.13272/j.issn.1671-251x.2024030009

基于MES−YOLOv5s的综采工作面大块煤检测算法

doi: 10.13272/j.issn.1671-251x.2024030009

徐慈强^1,,
贾运红^{2, 3},
田原^{2, 3}

1.
煤炭科学研究总院，北京　100013
2.
中国煤炭科工集团太原研究院有限公司，山西太原　030032
3.
煤矿采掘机械装备国家工程实验室，山西太原　030032

基金项目: 国家重点研发计划资助项目（2020YFB1314003）；山西省自然科学基金资助项目（201801D121189）。

详细信息

作者简介:
徐慈强（2000—），男，江西上饶人，硕士研究生，主要研究方向为煤矿电气自动化，E-mail：905079262@qq.com

中图分类号: TD67
计量
- 文章访问数: 64
- HTML全文浏览量: 17
- PDF下载量: 23
- 被引次数: 0
出版历程
- 收稿日期: 2024-03-04
- 修回日期: 2024-03-22
- 网络出版日期: 2024-04-11

Large block coal detection algorithm for fully mechanized working face based on MES-YOLOv5s

XU Ciqiang^1
,,
JIA Yunhong^{2, 3},
TIAN Yuan^{2, 3}

1.
China Coal Research Institute, Beijing 100013, China
2.
CCTEG Taiyuan Research Institute, Taiyuan 030032, China
3.
National Engineering Laboratory for Coal Mining Machinery, Taiyuan 030032, China

摘要

摘要: 综采工作面的目标具有高速运动、多尺度、遮挡等特点，现有的目标检测算法存在精度低、模型占用的内存大、硬件依赖强等问题。针对上述问题，提出了一种基于MES−YOLOv5s的综采工作面大块煤检测算法。采用轻量化设计，将MobileNetV3作为主干网络，以减小模型占用的内存，提高CPU端的检测速度；在颈部网络添加高效多尺度注意力（EMA）模块，融合不同尺度的上下文信息，并进一步减少计算开销；采用SIoU损失函数代替CIoU损失函数，以提高训练速度和推理准确性。消融实验结果表明：MobileNetV3大幅减少了模型占用的内存和检测时间，但mAP损失严重；EMA模块和SIoU损失函数可在一定程度上恢复损失的精度，同时保证模型在CPU上具有较高的检测速度，满足煤矿井下目标实时检测需求。对比实验结果表明，与DETR，YOLOv5n，YOLOv5s，YOLOv7模型相比，MES−YOLOv5s模型综合性能最好，mAP为84.6%，模型占用的内存为11.2 MiB，在CPU端的检测时间为31.8 ms，在高速运动、多尺度、遮挡和多目标的工况环境下能够保持较高的召回率和精度。
- 综采工作面 /
- 目标检测 /
- 大块煤检测 /
- YOLOv5s /
- MobileNetV3 /
- 高效多尺度注意力模块 /
- SIoU损失函数
Abstract: The objects in the fully mechanized working face have the features of high-speed motion, multi-scale, occlusion, etc. The existing object detection algorithms have problems such as low precision, large memory of models, and strong hardware dependence. In order to solve the above problems, a large block coal detection algorithm based on MES-YOLOv5s is proposed in fully mechanized working face. The method adopts a lightweight design, uses MobileNetV3 as the backbone network to reduce the memory occupied by the model and improve the detection speed on the CPU side. The method adds an efficient multi-scale attention (EMA) module to the neck network, fuses contextual information of different scales, and further reduces computational overhead. The method uses SIoU loss function instead of CIoU loss function to improve training speed and inference accuracy. The ablation experiment results show that MobileNetV3 significantly reduces the memory and detection time occupied by the model, but the mAP loss is severe. The EMA module and SIoU loss function can restore the precision of the loss to a certain extent, while ensuring that the model has a high detection speed on the CPU, meeting the real-time detection needs of coal mine underground objects. The comparative experimental results show that compared with DETR, YOLOv5n, YOLOv5s, and YOLOv7 models, the MES-YOLOv5s model has the best overall performance, with an mAP of 84.6%. The model occupies 11.2 MiB of memory and has a detection time of 31.8 ms on the CPU side. It can maintain high recall and precision in high-speed motion, multi-scale, occlusion, and multi-object working environments.
- fully mechanized working face /
- object detection /
- large block coal detection /
- YOLOv5s /
- MobileNetV3 /
- efficient multi-scale attention module /
- SIoU loss function

HTML全文

图 1 YOLO系列网络模型性能对比

Figure 1. Performance comparison of YOLO series network models

下载: 全尺寸图片幻灯片

图 2 MES−YOLOV5s网络模型结构

Figure 2. MES-YOLOV5s network model structure

下载: 全尺寸图片幻灯片

图 3 bneck模块结构

Figure 3. Structure of bneck module

下载: 全尺寸图片幻灯片

图 4 EMA模块结构

Figure 4. Structure of efficient multi-scale attention（EMA） module

下载: 全尺寸图片幻灯片

图 5 长壁综采工作面数据集处理流程

Figure 5. The processing process of the dataset of the longwall fully mechanized working face

下载: 全尺寸图片幻灯片

图 6 不同模型部分检测结果

Figure 6. Partial detection results of different models

下载: 全尺寸图片幻灯片

表 1 消融实验结果

Table 1. Ablation test results

改进策略			mAP/%	占用内存/MiB	CPU检测时间/ms
MobileNetV3	EMA	SIoU	mAP/%	占用内存/MiB	CPU检测时间/ms
×	×	×	85.1	54.1	68.5
√	×	×	80.9	11.1	28.8
×	√	×	87.3	54.7	75.7
×	×	√	85.7	54.1	64.9
√	√	√	84.6	11.2	31.8

下载: 导出CSV

表 2 对比实验结果

Table 2. Comparative experimental results

模型	mAP/%	占用内存/MiB	CPU检测时间/ms
DETR	81.3	149	496.2
YOLOv5n	83.9	14	31.9
YOLOv5s	85.1	54.1	68.5
YOLOv7	86.5	284.6	320.8
MES−YOLOv5s	84.6	11.2	31.8

下载: 导出CSV

参考文献(22)

[1]	葛世荣,胡而已,李允旺. 煤矿机器人技术新进展及新方向[J]. 煤炭学报,2023,48(1):54-73. GE Shirong,HU Eryi,LI Yunwang. New progress and direction of robot technology in coal mine[J]. Journal of China Coal Society,2023,48(1):54-73.
[2]	谢和平,王金华,王国法,等. 煤炭革命新理念与煤炭科技发展构想[J]. 煤炭学报,2018,43(5):1187-1197. XIE Heping,WANG Jinhua,WANG Guofa,et al. New ideas of coal revolution and layout of coal science and technology development[J]. Journal of China Coal Society,2018,43(5):1187-1197.
[3]	王国法,庞义辉,任怀伟,等. 智慧矿山系统工程及关键技术研究与实践[J/OL]. 煤炭学报:1-23[2024-02-22]. https://doi.org/10.13225/j.cnki.jccs.2023.1355. WANG Guofa,PANG Yihui,REN Huaiwei,et al. System engineering and key technologies research and practice of smart mine[J/OL]. Journal of China Coal Society:1-23[2024-02-22]. https://doi.org/10.13225/j.cnki.jccs.2023.1355.
[4]	詹召伟. 煤矿综采工作面智能化开采关键技术和发展方向[J]. 能源与节能,2023(1):82-86. ZHAN Zhaowei. Key technologies and development direction of intelligent mining in fully mechanized mining face[J]. Energy and Energy Conservation,2023(1):82-86.
[5]	葛世荣,张晞,薛光辉,等. 我国煤矿煤机智能技术与装备发展研究[J]. 中国工程科学,2023,25(5):146-156. doi: 10.15302/J-SSCAE-2023.05.013 GE Shirong,ZHANG Xi,XUE Guanghui,et al. Development of intelligent technologies and machinery for coal mining in China's underground coal mines[J]. Strategic Study of CAE,2023,25(5):146-156. doi: 10.15302/J-SSCAE-2023.05.013
[6]	崔卫秀,穆润青,解鸿章,等. 500 m超长工作面刮板智能输送技术研究[J/OL]. 煤炭科学技术:1-10[2024-02-22]. http://kns.cnki.net/kcms/detail/11.2402.TD.20230921.1538.002.html. CUI Weixiu,MU Runqing,XIE Hongzhang,et al. Research on intelligent conveying technology of 500 m ultra-long face scraper[J/OL]. Coal Science and Technology:1-10[2024-02-22]. http://kns.cnki.net/kcms/detail/11.2402.TD.20230921.1538.002.html.
[7]	苗长云,李佳. 基于机器视觉的带式输送机落料口堆煤检测[J]. 辽宁工程技术大学学报(自然科学版),2023,42(5):617-624. MIAO Changyun,LI Jia. Machine vision-based coal pile detection at drop port of belt conveyor[J]. Journal of Liaoning Technical University(Natural Science),2023,42(5):617-624.
[8]	程德强,寇旗旗,江鹤,等. 全矿井智能视频分析关键技术综述[J]. 工矿自动化,2023,49(11):1-21. CHENG Deqiang,KOU Qiqi,JIANG He,et al. Overview of key technologies for mine-wide intelligent video analysis[J]. Journal of Mine Automation,2023,49(11):1-21.
[9]	宋军强. OpenCV耦合机器视觉的背光板表面异物检测算法研究[J]. 组合机床与自动化加工技术,2015(11):83-87. SONG Junqiang. The study on image objects location and compensation technology based on OpenCV and machine vision[J]. Modular Machine Tool & Automatic Manufacturing Technique,2015(11):83-87.
[10]	刘孝军,王飞. 基于AI的煤矿视频智能分析技术[J]. 煤炭科学技术,2022,50(增刊2):260-264. LIU Xiaojun,WANG Fei. Application of video intelligent analysis technology in coal mine based on computer vision[J]. Coal Science and Technology,2022,50(S2):260-264.
[11]	曹现刚,李虎,王鹏,等. 基于跨模态注意力融合的煤炭异物检测方法[J]. 工矿自动化,2024,50(1):57-65. CAO Xiangang,LI Hu,WANG Peng,et al. A coal foreign object detection method based on cross modal attention fusion[J]. Journal of Mine Automation,2024,50(1):57-65.
[12]	高涵,赵培培,于正,等. 基于特征增强与Transformer的煤矿输送带异物检测[J/OL]. 煤炭科学技术:1-11[2024-02-22]. http://kns.cnki.net/kcms/detail/11.2402.td.20240119.1515.012.html. GAO Han,ZHAO Peipei,YU Zheng,et al. Coal mine conveyor belt foreign object detection based on feature enhancement and transformer[J/OL]. Coal Science and Technology:1-11[2024-02-22]. http://kns.cnki.net/kcms/detail/11.2402.td.20240119.1515.012.html.
[13]	张立亚. 基于生成对抗网络的带式输送机异物检测方法[J]. 工矿自动化,2023,49(11):53-59. ZHANG Liya. Foreign object detection method for belt conveyor based on generative adversarial nets[J]. Journal of Mine Automation,2023,49(11):53-59.
[14]	王科平,连凯海,杨艺,等. 基于改进YOLOv4的综采工作面目标检测[J]. 工矿自动化,2023,49(2):70-76. WANG Keping,LIAN Kaihai,YANG Yi,et al. Target detection of the fully mechanized working face based on improved YOLOv4[J]. Journal of Mine Automation,2023,49(2):70-76.
[15]	李江涛,张康辉,沙特. 煤中异物识别的深度学习模型轻量化策略[J]. 煤炭工程,2023,55(增刊1):220-224. LI Jiangtao,ZHANG Kanghui,SHA Te. Lightweight deep learning model compression strategy for coal foreign object recognition[J]. Coal Engineering,2023,55(S1):220-224.
[16]	桂方俊,李尧. 基于CBA−YOLO模型的煤矸石检测[J]. 工矿自动化,2022,48(6):128-133. GUI Fangjun,LI Yao. Coal gangue detection based on CBA-YOLO model[J]. Journal of Mine Automation,2022,48(6):128-133.
[17]	LIU Shu,QI Lu,QIN Haifang,et al. Path aggregation network for instance segmentation[C]. IEEE/CVF Conference on Computer Vision and Pattern Recognition,Salt Lake City,2018:8759-8768.
[18]	HOWARD A,SANDLER M,CHEN Bo,et al. Searching for MobileNetV3[C]. IEEE/CVF International Conference on Computer Vision,Seoul,2019:1314-1324.
[19]	OUYANG Daliang,HE Su,ZHANG Guozhong,et al. Efficient multi-scale attention module with cross-spatial learning[C]. IEEE International Conference on Acoustics,Speech and Signal Processing,Rhodes Island,2023:1-5.
[20]	YU Jiahui,JIANG Yuning,WANG Zhangyang,et al. UnitBox:an advanced object detection network[C]. The 24th ACM International Conference on Multimedia,2016:516-520. DOI: 10.1145/2964284.2967274.
[21]	GEVORGYAN Z. SIoU loss:more powerful learning for bounding box regression[EB/OL]. [2024-02-12]. https://arxiv.org/abs/2205.12740.
[22]	YANG Wenjuan,ZHANG Xuhui,MA Bing,et al. An open dataset for intelligent recognition and classification of abnormal condition in longwall mining[J]. Scientific Data,2023,10. DOI: 10.1038/s41597-023-02322-9.