基于改进YOLOv4的综采工作面目标检测

王科平; 连凯海; 杨艺; 费树岷

doi:10.13272/j.issn.1671-251x.2022070080

基于改进YOLOv4的综采工作面目标检测

doi: 10.13272/j.issn.1671-251x.2022070080

王科平^{1, 2,},
连凯海^{1, 2, ,},
杨艺^{1, 2},
费树岷³

1.
河南理工大学电气工程与自动化学院, 河南焦作　454003
2.
河南理工大学河南煤矿装备智能检测与控制重点实验室, 河南焦作　454003
3.
东南大学自动化学院, 江苏南京　210096

基金项目: 河南省科技攻关计划项目（212102210390）。

详细信息

作者简介:
王科平（1976—），女，河北张家口人，副教授，博士，主要研究方向为图像清晰化处理、目标检测、深度学习，E-mail：wangkp@hpu.edu.cn

通讯作者:
连凯海（1997—），男，河南南阳人，硕士研究生，主要研究方向为目标检测、深度学习，E-mail：15139027117@163.com

中图分类号: TD67
计量
- 文章访问数: 369
- HTML全文浏览量: 103
- PDF下载量: 43
- 被引次数: 0
出版历程
- 收稿日期: 2022-07-29
- 修回日期: 2023-02-16
- 网络出版日期: 2022-09-23

Target detection of the fully mechanized working face based on improved YOLOv4

WANG Keping^{1, 2
,},
LIAN Kaihai^{1, 2
, ,},
YANG Yi^{1, 2},
FEI Shumin³

1.
School of Electrical Engineering and Automation, Henan Polytechnic University, Jiaozuo 454003, China
2.
Henan Key Laboratory of Intelligent Detection and Control of Coal Mine Equipment, Henan Polytechnic University, Jiaozuo 454003, China
3.
School of Automation, Southeast University, Nanjing 210096, China

摘要

摘要: 综采工作面关键设备及人员的准确检测是实现煤炭智能化开采信息感知的重要环节。传统目标检测算法通过人工提取特征实现目标检测，易受环境影响，不具有普适性。基于卷积神经网络的目标检测算法可以自适应地提取深层信息，但复杂环境下检测精度不高、网络参数多、计算量大。针对上述问题，提出了一种改进YOLOv4模型，并将其应用于综采工作面目标检测。为准确从综采工作面复杂环境中检测到目标，在CSPDarkNet53网络中融入残差自注意力模块，保证参数共享及高效局部信息聚合的同时增强全局信息获取能力，提升图像关键目标特征表达能力，进而提高目标检测精度；为适应综采工作面目标检测高效性需求，引入深度可分离卷积替代传统卷积，以减少模型参数量和计算量，有利于模型的工业部署，提高目标检测速度。实验结果表明，与YOLOv3、CenterNet及YOLOv4模型相比，改进YOLOv4模型平均精度均值最高，达92.59%，且在参数量、计算量、检测精度上具有更优的平衡，可在煤尘干扰、光照不均、目标运动等复杂环境下对目标准确检测。
- 综采工作面 /
- 目标检测 /
- YOLOv4 /
- 残差自注意力 /
- 深度可分离卷积
Abstract: The accurate detection of key equipment and personnel in the fully mechanized working face is an important link to realize the information perception of intelligent coal mining. The traditional target detection algorithm realizes the target detection by extracting the features manually. But it is easily affected by the environment and it is not universal. The target detection algorithm based on the convolutional neural network can extract deep information adaptively. But the detection precision is not high, the network parameters are too many, and the calculation is too large in complex environment. In order to the above problems, an improved YOLOv4 model is proposed and applied to the target detection of the fully mechanized working face. In order to accurately detect targets in the complex environment of a fully mechanized working face, a residual self-attention module is integrated into the CSPDarkNet53 network. The capability of acquiring global information is enhanced while parameter sharing and efficient local information aggregation are ensured. The capability of expressing the features of key targets in an image is improved, and the target detection precision is further improved. In order to meet the requirement of high efficiency of target detection in the fully mechanized working face, depthwise-separable convolution is introduced to replace traditional convolution. The model parameter quantity and calculation quantity are reduced. It is beneficial to the industrial deployment of the model. And it improves target detection speed. The experimental results show that compared with YOLOv3, CenterNet and YOLOv4 models, the average precision of the improved YOLOv4 model is the highest, up to 92.59%. It has better balance in parameter quantity, calculation quantity and detection precision. It can accurately detect the target in the complex environment such as coal dust interference, uneven lighting and motion blur.
- fully mechanized working face /
- target detection /
- YOLOv4 /
- residual self-attention /
- depthwise-separable convolution

HTML全文

图 1 改进YOLOv4模型结构

Figure 1. Improved YOLOv4 model structure

下载: 全尺寸图片幻灯片

图 2 RSA模块结构

Figure 2. Residual self-attention module architecture

下载: 全尺寸图片幻灯片

图 3 深度可分离卷积操作

Figure 3. Depthwise separable convolution operation

下载: 全尺寸图片幻灯片

图 4 不同模型检测结果

Figure 4. Detection results of different models

下载: 全尺寸图片幻灯片

表 1 不同模型在井下综采工作面数据集上的检测结果

Table 1. Test results of different models on data set of underground fully-mechanized mining face

类别	AP/%
类别	YOLOv3	CenterNet	YOLOv4	改进YOLOv4
护帮板	97.28	89.92	97.97	98.50
采煤机	94.86	94.84	97.51	97.90
滚筒	94.46	95.80	96.23	96.87
大煤块	90.57	90.90	93.44	94.65
行人	84.14	87.41	89.67	91.17
线槽	81.41	82.91	87.21	89.93
刮板输送机	65.64	72.75	73.14	79.14
mAP/%	86.91	87.79	90.74	92.59

下载: 导出CSV

表 2 不同模型检测性能对比

Table 2. Comparison of detection performance of different models

模型	输入大小	参数量/10⁶	模型大小/MB	FLOPs/10⁹	mAP/%
YOLOv3	416×416	61.56	246.5	32.78	86.91
CenterNet	416×416	32.67	130.9	23.14	87.79
YOLOv4	416×416	63.97	256.3	29.90	90.74
改进YOLOv4	416×416	33.11	133.2	19.48	92.59

下载: 导出CSV

表 3 消融实验结果

Table 3. Ablation experiment results

模型	mAP/%	参数量/10⁶
YOLOv4	90.74	63.97
YOLOv4+深度可分离卷积	90.25	35.71
YOLOv4+深度可分离卷积+RSA	92.59	33.11

下载: 导出CSV

参考文献(21)

[1]	边文越,陈挺,陈晓怡,等. 世界主要发达国家能源政策研究与启示[J]. 中国科学院院刊,2019,34(4):488-496. doi: 10.16418/j.issn.1000-3045.2019.04.014 BIAN Wenyue,CHEN Ting,CHEN Xiaoyi,et al. Study and enlightenment of energy policies of major developed countries[J]. Bulletin of Chinese Academy of Sciences,2019,34(4):488-496. doi: 10.16418/j.issn.1000-3045.2019.04.014
[2]	葛世荣,郝尚清,张世洪,等. 我国智能化采煤技术现状及待突破关键技术[J]. 煤炭科学技术,2020,48(7):28-46. GE Shirong,HAO Shangqing,ZHANG Shihong,et al. Status of intelligent coal mining technology and potential key technologies in China[J]. Coal Science and Technology,2020,48(7):28-46.
[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. doi: 10.13199/j.cnki.cst.2021.01.001 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. doi: 10.13199/j.cnki.cst.2021.01.001
[4]	张强,张润鑫,刘峻铭,等. 煤矿智能化开采煤岩识别技术综述[J]. 煤炭科学技术,2022,50(2):1-26. doi: 10.13199/j.cnki.cst.2021-1333 ZHANG Qiang,ZHANG Runxin,LIU Junming,et al. Review on coal and rock identification technology for intelligent mining in coal mines[J]. Coal Science and Technology,2022,50(2):1-26. doi: 10.13199/j.cnki.cst.2021-1333
[5]	李首滨. 智能化开采研究进展与发展趋势[J]. 煤炭科学技术,2019,47(10):102-110. doi: 10.13199/j.cnki.cst.2019.10.012 LI Shoubin. Progress and development trend of intelligent mining technology[J]. Coal Science and Technology,2019,47(10):102-110. doi: 10.13199/j.cnki.cst.2019.10.012
[6]	高有进,杨艺,常亚军,等. 综采工作面智能化关键技术现状与展望[J]. 煤炭科学技术,2021,49(8):1-22. GAO Youjin,YANG Yi,CHANG Yajun,et al. Status and prospect of key technologies of intelligentization of fully-mechanized coal mining face[J]. Coal Science and Technology,2021,49(8):1-22.
[7]	李爽,薛广哲,方新秋,等. 煤矿智能化安全保障体系及关键技术[J]. 煤炭学报,2020,45(6):2320-2330. doi: 10.13225/j.cnki.jccs.zn20.0356 LI Shuang,XUE Guangzhe,FANG Xinqiu,et al. Coal mine intelligent safety system and key technologies[J]. Journal of China Coal Society,2020,45(6):2320-2330. doi: 10.13225/j.cnki.jccs.zn20.0356
[8]	李伟. 深部煤炭资源智能化开采技术现状与发展方向[J]. 煤炭科学技术,2021,49(1):139-145. LI Wei. Current status and development direction of intelligent mining technology for deep coal resources[J]. Coal Science and Technology,2021,49(1):139-145.
[9]	殷华,虎晓龙. 煤矿无人综采模式及关键技术分析[J]. 工矿自动化,2021,47(增刊2):23-25. YIN Hua,HU Xiaolong. Analysis of unmanned fully mechanized mining mode of coal mine and its key technologies[J]. Industry and Mine Automation,2021,47(S2):23-25.
[10]	任怀伟,孟祥军,李政,等. 8 m大采高综采工作面智能控制系统关键技术研究[J]. 煤炭科学技术,2017,45(11):37-44. REN Huaiwei,MENG Xiangjun,LI Zheng,et al. Study on key technology of intelligent control system applied in 8 m large mining height fully-mechanized face[J]. Coal Science and Technology,2017,45(11):37-44.
[11]	刘贝. 矿井人员目标检测与跟踪研究[D]. 西安: 西安科技大学, 2020. LIU Bei. Research on target detection and tracking of mine pedestrian[D]. Xi'an: Xi'an University of Science and Technology, 2020.
[12]	赵谦. 煤矿井下动态目标视频监测图像处理研究[D]. 西安: 西安科技大学, 2014. ZHAO Qian. Study on video monitoring and image processing of coal mine dynamic targets[D]. Xi'an: Xi'an University of Science and Technology, 2014.
[13]	程健,王东伟,杨凌凯,等. 一种改进的高斯混合模型煤矸石视频检测方法[J]. 中南大学学报(自然科学版),2018,49(1):118-123. doi: 10.11817/j.issn.1672-7207.2018.01.016 CHENG Jian,WANG Dongwei,YANG Lingkai,et al. An improved Gaussian mixture model for coal gangue video detection[J]. Journal of Central South University(Science and Technology),2018,49(1):118-123. doi: 10.11817/j.issn.1672-7207.2018.01.016
[14]	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
[15]	刘备战,赵洪辉,周李兵. 面向无人驾驶的井下行人检测方法[J]. 工矿自动化,2021,47(9):113-117. doi: 10.13272/j.issn.1671-251x.17830 LIU Beizhan,ZHAO Honghui,ZHOU Libing. Unmanned driving-oriented underground mine pedestrian detection method[J]. Industry and Mine Automation,2021,47(9):113-117. doi: 10.13272/j.issn.1671-251x.17830
[16]	付燕,窦晓熠,叶鸥. 基于YOLO的井下人员速度测量方法研究[J]. 煤炭工程,2022,54(2):160-165. FU Yan,DOU Xiaoyi,YE Ou. A measurement method of underground workers speed based on YOLO[J]. Coal Engineering,2022,54(2):160-165.
[17]	任志玲,朱彦存.改进CenterNet算法的煤矿皮带运输异物识别研究[J/OL].控制工程:1-8[2022-09-08]. https://doi.org/10.14107/j.cnki.kzgc.20200792. REN Zhiling,ZHU Yancun.Research on foreign objects recognition of coal mine belt transportation with improved CenterNet algorithm[J/OL].Control Engineering of China:1-8[2022-09-08]. https://doi.org/10.14107/j.cnki.kzgc.20200792.
[18]	王燕平. 基于改进YOLOv3的煤岩界线识别及采高仿人智能控制研究[D]. 太原: 中北大学, 2021. WANG Yanping. Coal-rock boundary identification based on improved YOLOv3 and research on human-simulating intelligent control of mining height[D]. Taiyuan: North University of China, 2021.
[19]	BOCHKOVSKIY A, WANG C Y, LIAO H Y M. YOLOv4: Optimal speed and accuracy of object detection[EB/OL]. [2022-01-20]. https://doi.org/10.48550/arXiv.2004.10934.
[20]	SRINIVAS A, LIN T Y, PARMAR N, et al. Bottleneck transformers for visual recognition[C]. IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021: 16519-16529.
[21]	SANDLER M, HOWARD A, ZHU Menglong, et al. MobileNetV2: Inverted residuals and linear bottlenecks[C]. IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, 2018: 4510-4520.