Foreign object recognition of belt conveyor in coal mine based on improved YOLOv7
-
摘要: 带式输送机煤流中会掺杂锚杆、角铁、木条、矸石、大块煤等异物，易导致输送带撕裂、转接处堵塞甚至断带。针对带式输送机巡检机器人难以在井下光照不均及带式输送机高速运行环境中高效、准确识别异物及模型部署不便等问题，以及YOLOv7模型对目标特征提取能力高，但识别速度较慢的特点，提出了一种基于改进YOLOv7的煤矿带式输送机异物识别方法。运用限制对比度自适应直方图均衡化方法对采集的带式输送机监控图像进行增强，提高图像中物体轮廓的清晰度；对YOLOv7模型进行改进，通过在主干提取网络引入轻量化无参注意力机制，提高模型对图像复杂背景的抗干扰能力和对异物特征的提取能力，同时引入深度可分离卷积代替主干特征提取网络中的普通卷积，提高异物识别速度；使用TensorRT引擎将训练后的改进YOLOv7模型进行转换并部署在NVIDIA Jetson Xavier NX上，实现了模型的加速。对煤矿井下分辨率为1 920×1 080的带式输送机监控视频进行识别，实验结果表明：改进YOLOv7模型的识别效果优于YOLOv5L和YOLOv7模型，识别精确率达92.8%，识别速度为25.64帧/s，满足精确、高效识别带式输送机异物的要求。Abstract: The coal flow of the belt conveyor will be mixed with anchor rod, angle iron, wood, gangue, and lump coal. This will easily lead to the tearing of the conveyor belt, the blockage of the transition and even the breakage of the belt. It is difficult for the inspection robot of the belt conveyor to efficiently and accurately recognize foreign objects in the environment of uneven lighting and high-speed running of the belt conveyor. The model deployment is inconvenient. The YOLOv7 model has a high capability to extract target features, but its recognition speed is slow. In order to solve the above problems, a foreign object recognition method of belt conveyor in coal mine based on improved YOLOv7 is proposed. The method of adaptive histogram equalization with limited contrast is used to enhance the collected monitoring image of the belt conveyor to improve the clarity of object contour in the image. The YOLOv7 model is improved by introducing a simple and parameter-free attention module into the backbone extraction network. The improved model can improve the model's anti-interference capability against the complex background of the image and the capability to extract foreign object features. The depthwise separable convolution is introduced to replace the ordinary convolution in the backbone feature extraction network to improve the speed of foreign object recognition. TensorRT engine is used to convert the improved YOLOv7 model after training and deploy it on NVIDIA Jetson Xavier NX, realizing the acceleration of the model. The video of the belt conveyor with the resolution of 1 920 × 1 080 in the underground coal mine is recognized. The experimental results show that the recognition effect of improved YOLOv7 is better than YOLOv5L and YOLOv7. The recognition accuracy rate is 92.8%, and the recognition speed is 25.64 frames/s, meeting the requirements of accurate and efficient recognition of foreign objects in the belt conveyor.
-
表 1 图像清晰度评价结果
Table 1. Evaluation results of image definition
图像 Entropy Brenner 原图 5.113 3 1.909 增强后图像 5.289 9 4.094 
下载: 导出CSV
表 2 不同模型的平均精确率、平均召回率和识别时间
Table 2. Average precision, average recall and recognition time of different models
模型 平均精确率/% 平均召回率/% 识别时间/s YOLOv5L 90.9 85.0 0.047 YOLOv7 89.4 84.5 0.027 改进YOLOv7 93.1 87.4 0.025
下载: 导出CSV
表 3 YOLOv7改进前后异物识别精确率和召回率对比
Table 3. Comparison of foreign object recognition precision and recall before and after YOLOv7 improvement
类别 YOLOv7 改进YOLOv7 精确率/% 召回率/% 精确率/% 召回率/% 锚杆 90.6 85.1 93.4 88.5 角铁 89.2 90.1 90.6 93.1 木条 91.3 84.7 97.7 86.1 矸石 87.4 84.9 88.9 89.3 大块煤 88.5 77.8 94.9 80.0
下载: 导出CSV
表 4 消融实验结果
Table 4. Ablation experimental results
增强 SimAM DWConv 平均精确率/% 识别时间/s 89.4 0.027 √ 90.3 0.027 √ 94.0 0.030 √ 88.1 0.024 √ √ √ 93.1 0.025
下载: 导出CSV
表 5 不同模型平均精确率和识别时间
Table 5. Average precision and recognition time of different models
模型 平均精确率/% 识别时间/s YOLOv5L 90.0 0.049 YOLOv7 88.9 0.041 改进YOLOv7 92.8 0.039
下载: 导出CSV
-
[1] 葛世荣,胡而已,裴文良. 煤矿机器人体系及关键技术[J]. 煤炭学报,2020,45(1):455-463. doi: 10.13225/j.cnki.jccs.YG19.1478GE Shirong,HU Eryi,PEI Wenliang. Classification system and key technology of coal mine robot[J]. Journal of China Coal Society,2020,45(1):455-463. doi: 10.13225/j.cnki.jccs.YG19.1478 [2] 方崇全. 煤矿带式输送机巡检机器人关键技术研究[J]. 煤炭科学技术,2022,50(5):263-270. doi: 10.13199/j.cnki.cst.ZN20-056FANG Chongquan. Research on key technology of inspection robot for coal mine belt conveyor[J]. Coal Science and Technology,2022,50(5):263-270. doi: 10.13199/j.cnki.cst.ZN20-056 [3] 吴守鹏,丁恩杰,俞啸. 基于改进FPN的输送带异物识别方法[J]. 煤矿安全,2019,50(12):127-130. doi: 10.13347/j.cnki.mkaq.2019.12.029WU Shoupeng,DING Enjie,YU Xiao. Foreign body identification of belt based on improved FPN[J]. Safety in Coal Mines,2019,50(12):127-130. doi: 10.13347/j.cnki.mkaq.2019.12.029 [4] 吕志强. 复杂环境下煤矿皮带运输异物图像识别研究[D]. 徐州: 中国矿业大学, 2020: 1-60.LYU Zhiqiang. Research on foreign body image recognition of coal mine belt transport under complex environment[D]. Xuzhou: China University of Mining and Technology, 2020: 1-60. [5] 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[2022-09-28]. DOI: 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[2022-09-28]. DOI: 10.14107/j.cnki.kzgc.20200792. [6] 胡璟皓,高妍,张红娟,等. 基于深度学习的带式输送机非煤异物识别方法[J]. 工矿自动化,2021,47(6):57-62,90. doi: 10.13272/j.issn.1671-251x.2021020041HU Jinghao,GAO Yan,ZHANG Hongjuan,et al. Research on the identification method of non-coal foreign object of belt conveyor based on deep learning[J]. Industry and Mine Automation,2021,47(6):57-62,90. doi: 10.13272/j.issn.1671-251x.2021020041 [7] WANG Yuanbin,WANG Yujing,DANG Langfei. Video detection of foreign objects on the surface of belt conveyor underground coal mine based on improved SSD[J]. Journal of Ambient Intelligence and Humanized Computing,2020:1-10. [8] 郝帅,张旭,马旭,等. 基于CBAM−YOLOv5的煤矿输送带异物检测[J]. 煤炭学报,2022,47(11):4147-4156. doi: 10.13225/j.cnki.jccs.2021.1644HAO Shuai,ZHANG Xu,MA Xu,et al. Foreign object detection in coal mine conveyor belt based on CBAM-YOLOv5[J]. Journal of China Coal Society,2022,47(11):4147-4156. doi: 10.13225/j.cnki.jccs.2021.1644 [9] 程德强,徐进洋,寇旗旗,等. 融合残差信息轻量级网络的运煤皮带异物分类[J]. 煤炭学报,2022,47(3):1361-1369. doi: 10.13225/j.cnki.jccs.xr21.1736CHENG Deqiang,XU Jinyang,KOU Qiqi,et al. Lightweight network based on residual information for foreign body classification on coal conveyor belt[J]. Journal of China Coal Society,2022,47(3):1361-1369. doi: 10.13225/j.cnki.jccs.xr21.1736 [10] XIAO Dong,KANG Zhuang,YU Hang,et al. Research on belt foreign body detection method based on deep learning[J]. Transactions of the Institute of Measurement and Control,2022,44(15):2919-2927. doi: 10.1177/01423312221094393 [11] WANG C Y, BOCHKOVSKIY A, LIAO H. YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[J/OL]. [2022-09-28]. https://arxiv.org/abs/2207.02696. [12] 杨骥,杨亚东,梅雪,等. 基于改进的限制对比度自适应直方图的视频快速去雾算法[J]. 计算机工程与设计,2015,36(1):221-226. doi: 10.16208/j.issn1000-7024.2015.01.040YANG Ji,YANG Yadong,MEI Xue,et al. Fast video dehazing based on improved contrast limited adaptive histogram equalization[J]. Computer Engineering and Design,2015,36(1):221-226. doi: 10.16208/j.issn1000-7024.2015.01.040 [13] 舒甜督. 医学CT图像的增强与分类算法研究[D]. 长春: 长春工业大学, 2022.SHU Tiandu. Research on enhancement and classification algorithm of medical CT images[D]. Changchun: Changchun University of Technology, 2022. [14] QIN Xiaoyi, LI Na, WENG Chao, et al. Simple attention module based speaker verification with iterative noisy label detection[C]. IEEE International Conference on Acoustics, Speech and Signal Processing, Singapore, 2021. [15] CHOLLET F. Xception: Deep learning with depthwise separable convolutions[C]. IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, 2017. [16] 顾德英,罗聿伦,李文超. 基于改进YOLOv5算法的复杂场景交通目标检测[J]. 东北大学学报(自然科学版),2022,43(8):1073-1079.GU Deying,LUO Yulun,LI Wenchao. Traffic target detection in complex scenes based on improved YOLOv5 algorithm[J]. Journal of Northeastern University(Natural Science),2022,43(8):1073-1079. [17] MAO Qinghua,WANG Yufei,ZHANG Xuhui,et al. Clarity method of fog and dust image in fully mechanized mining face[J]. Machine Vision and Applications,2022,33(2):1-16. [18] LI Kexin,QIN Liang,LI Qiang,et al. Improved edge lightweight YOLOv4 and its application in on-site power system work[J]. Global Energy Interconnection,2022,5(2):168-180. doi: 10.1016/j.gloei.2022.04.014 -
点击查看大图
图(6) / 表(5)
计量
- 文章访问数: 547
- HTML全文浏览量: 113
- PDF下载量: 169
- 被引次数: 0
