Method for counting coal mine drill pipes based on YOLOv11_OBB
-
摘要:
针对煤矿井下打钻图像识别泛化能力差、钻杆计数不准确等问题，采集并标注了煤矿井下钻杆计数数据集CMDPC_OBB，提出了一种基于YOLOv11_OBB的煤矿钻杆计数方法。该方法包括打钻图像识别模型YOLOv11_OBB和场景自适应的钻杆计数算法2个部分。YOLOv11_OBB采用旋转边界框，精准捕获具有倾斜角度的打钻图像，通过L2正则化处理改进YOLOv11颈部网络，降低权重波动对特征融合的干扰，使模型训练更加稳定；场景自适应的钻杆计数算法通过追踪目标钻杆与钻机尾部关键点之间的运动轨迹及多条件判断峰值实现自动计数，减少了与计数无关的钻杆对计数准确率的影响；将YOLOv11_OBB学习到的图像特征作为潜在知识指导钻杆计数的推理逻辑。在CMDPC_OBB数据集上的实验结果表明，YOLOv11_OBB的图像识别精度为93.5%，与YOLOv5L_OBB，YOLOv8L_OBB相比分别提升了8.2%，3.0%；钻杆计数算法的准确率为97.96%，模型识别速度为34帧/s，计数速度为79帧/s，满足实时计数要求。
-
关键词:
- 钻杆计数 /
- 旋转目标检测 /
- 深度学习 /
- 场景自适应 /
- YOLOv11_OBB
Abstract:To address the issues of poor generalization in downhole drilling image recognition and inaccurate drill pipe counting in coal mines, this study collected and annotated a dedicated dataset CMDPC_OBB, and proposed a method for drill pipe counting based on YOLOv11_OBB. The method consisted of two components: the YOLOv11_OBB-based drilling image recognition model and a scene-adaptive drill pipe counting algorithm. YOLOv11_OBB used rotated bounding boxes to accurately capture drilling images with inclined angles. By applying L2 regularization to improve the neck network of YOLOv11, weight fluctuation interference in feature fusion was reduced, ensuring more stable model training. The scene-adaptive counting algorithm tracked the motion trajectory between target drill pipes and key points at the drill rig's tail while employing multi-condition peak judgment to achieve automatic counting, thereby minimizing the impact of irrelevant pipes on accuracy. In addition, the image features learned by YOLOv11_OBB served as latent knowledge to guide the counting logic. Experimental results on the CMDPC_OBB dataset show that YOLOv11_OBB achieves an image recognition accuracy of 93.5%, outperforming YOLOv5L_OBB and YOLOv8L_OBB by 8.2% and 3.0%, respectively. The counting algorithm achieves an accuracy of 97.96%, with a model recognition speed of 34 FPS and a counting speed of 79 FPS, meeting real-time requirements.
-
Keywords:
- drill pipe counting /
- rotated target detection /
- deep learning /
- scene adaptation /
- YOLOv11_OBB
-
-
![]()
图 5 YOLOv11_OBB打钻图像识别模型训练过程
Figure 5. Training process of YOLOv11_OBB drilling image recognition model
![]()
图 7 YOLOv11_OBB在CMDPC_OBB测试集上的识别结果
Figure 7. Recognition results of YOLOv11_OBB on CMDPC_OBB test set
表 1 CMDPC_OBB数据集划分及标签分布
Table 1 CMDPC_OBB dataset partitioning and label distribution
数据类别 标签 标注数量/张 训练集 验证集 测试集 合计 钻机整体 Drill_body 8 716 1 244 2 496 12 456 钻机头部 Drill_head 8 811 1 158 2 465 12 434 钻机尾部 Drill_tail 8 823 1 236 2 521 12 580 钻机钻杆 Drill_pipe 16 840 2 320 4 814 23 974 合计 43 190 5 958 12 296 61 444 
下载: 导出CSV
表 2 不同目标检测模型在CMDPC_OBB测试集上的识别结果
Table 2 Recognition results of different target detection models on CMDPC_OBB test dataset
模型 精确率 召回率 mAP@
0.5mAP@
[0.5:0.95]参数量/
106个GFLOPs 帧率/
（帧·s−1）YOLOv5L[21] 0.826 0.590 0.663 0.330 2.5 7.810 42 YOLOv6L[22] 0.834 0.588 0.650 0.306 3.9 10.650 46 YOLOv8L[23] 0.833 0.578 0.643 0.308 4.0 10.330 56 YOLOv10L[24] 0.831 0.596 0.649 0.340 2.6 8.488 32 YOLOv5L_OBB 0.853 0.538 0.646 0.336 1.9 5.810 31 YOLOv8L_OBB 0.905 0.710 0.782 0.482 3.5 9.136 38 YOLOv8_OBB_DG[11] 0.929 0.69 0.780 0.493 2.8 7.500 32 YOLOv11_OBB 0.935 0.686 0.796 0.508 2.5 8.390 34
下载: 导出CSV
表 3 钻杆计数方法对比实验结果
Table 3 Comparative experimental results of drill pipe counting methods
视频 时长/s 打钻
类型人工计
数/根基于YOLOv11_OBB的
钻杆计数方法DC_SPC_KEY
计数方法算法计
数/根误计
率/%帧率/
（帧·s−1）算法计
数/根误计
率/%帧率/
（帧·s−1）视频1 300 退钻 105 103 1.904 80 103 1.904 77 视频2 309 退钻 52 52 0 79 51 1.923 78 视频3 200 进钻 12 12 0 81 12 0 79 视频4 300 进钻 16 15 6.250 77 15 6.250 76
下载: 导出CSV
-
[1] 张增辉,张海峰,朱兴林. 煤矿井下打钻视频管理系统研究与设计[J]. 煤矿机械,2024,45(7):184-186. ZHANG Zenghui,ZHANG Haifeng,ZHU Xinglin. Research and design of video management system for underground drilling in coal mine[J]. Coal Mine Machinery,2024,45(7):184-186.
[2] 彭业勋. 煤矿井下钻杆计数方法研究[D]. 西安:西安科技大学,2019. PENG Yexun. Research on counting method of drill pipe in coal mine[D]. Xi'an:Xi'an University of Science and Technology,2019.
[3] 梁运培,郑梦浩,李全贵,等. 我国煤与瓦斯突出预测与预警研究现状[J]. 煤炭学报,2023,48(8):2976-2994. LIANG Yunpei,ZHENG Menghao,LI Quangui,et al. A review on prediction and early warning methods of coal and gas outburst[J]. Journal of China Coal Society,2023,48(8):2976-2994.
[4] REDMON J,FARHADI A. YOLOv3:an incremental improvement[EB/OL]. [2025-01-10]. https://arxiv.org/abs/1804.02767v1.
[5] 姜媛媛,刘宋波. 基于改进YOLOv8n的煤矿井下钻杆计数方法[J]. 工矿自动化,2024,50(8):112-119. JIANG Yuanyuan,LIU Songbo. A coal mine underground drill pipes counting method based on improved YOLOv8n[J]. Journal of Mine Automation,2024,50(8):112-119.
[6] 方杰,李振璧,夏亮. 基于ECO−HC的钻杆计数方法[J]. 煤炭技术,2021,40(11):186-189. FANG Jie,LI Zhenbi,XIA Liang. Drill pipe counting method based on ECO-HC[J]. Coal Technology,2021,40(11):186-189.
[7] 高瑞,郝乐,刘宝,等. 基于改进ResNet网络的井下钻杆计数方法[J]. 工矿自动化,2020,46(10):32-37. GAO Rui,HAO Le,LIU Bao,et al. Research on underground drill pipe counting method based on improved ResNet network[J]. Industry and Mine Automation,2020,46(10):32-37.
[8] 冉庆庆,董立红,温乃宁. 改进YOLOv8的煤矿井下低照度图像钻杆计数方法[J/OL]. 电子测量技术:1-12[2025-05-08]. http://kns.cnki.net/kcms/detail/11.2175.TN.20250425.1132.018.html. RAN Qingqing,DONG Lihong,WEN Naining. Improved YOLOv8 method for counting drill pipe in low illumination image of coal mine underground[J/OL]. Electronic Measurement Technology:1-12[2025-05-08]. http://kns.cnki.net/ kcms/detail/11.2175.TN.20250425.1132.018.html.
[9] 王旭,吴艳霞,张雪,等. 计算机视觉下的旋转目标检测研究综述[J]. 计算机科学,2023,50(8):79-92. DOI: 10.11896/jsjkx.221000148 WANG Xu,WU Yanxia,ZHANG Xue,et al. Survey of rotating object detection research in computer vision[J]. Computer Science,2023,50(8):79-92. DOI: 10.11896/jsjkx.221000148
[10] SHI Pengfei,ZHAO Zhongxin,FAN Xinnan,et al. Remote sensing image object detection based on angle classification[J]. IEEE Access,2021,9:118696-118707. DOI: 10.1109/ACCESS.2021.3107358
[11] 张富凯,赵硕,张强,等. 基于旋转目标检测与显著性峰值推理的煤矿井下钻杆计数方法[J/OL]. 煤炭学报:1-15[2025-05-08]. https://doi.org/10.13225/j.cnki.jccs.2024.1328. ZHANG Fukai,ZHAO Shuo,ZHANG Qiang,et al. Coal mine drill pipe counting method based on oriented object detection and significant peak inference[J/OL]. Journal of China Coal Society:1-15[2025-05-08]. https://doi.org/10.13225/j.cnki.jccs.2024.1328.
[12] 张睿,李允臣,王家宝,等. 多尺度特征融合的双模态目标检测方法[J]. 计算机工程与应用,2024,60(17):233-242. ZHANG Rui,LI Yunchen,WANG Jiabao,et al. Multiscale feature fusion approach for dual-modal object detection[J]. Computer Engineering and Applications,2024,60(17):233-242.
[13] LUO Yi,SHAO Feng,MU Baoyang,et al. Dynamic weighted fusion and progressive refinement network for visible-depth-thermal salient object detection[J]. IEEE Transactions on Circuits and Systems for Video Technology,2024,34(11):10662-10677. DOI: 10.1109/TCSVT.2024.3414170
[14] 沈记全,陈相均,翟海霞. 基于改进边界框回归损失的YOLOv3检测算法[J]. 计算机工程,2022,48(3):236-243. SHEN Jiquan,CHEN Xiangjun,ZHAI Haixia. YOLOv3 detection algorithm based on the improved bounding box regression loss[J]. Computer Engineering,2022,48(3):236-243.
[15] 张震,彭景昊,田鸿朋,等. 考虑数据分布损失的图像分割[J]. 计算机工程与应用,2024,60(19):242-249. DOI: 10.3778/j.issn.1002-8331.2307-0076 ZHANG Zhen,PENG Jinghao,TIAN Hongpeng,et al. Design of loss function considering data distribution[J]. Computer Engineering and Applications,2024,60(19):242-249. DOI: 10.3778/j.issn.1002-8331.2307-0076
[16] 朱旭东,熊赟. 基于样本分布损失的图像多标签分类研究[J]. 计算机科学,2022,49(6):210-216. DOI: 10.11896/jsjkx.210300267 ZHU Xudong,XIONG Yun. Study on multi-label image classification based on sample distribution loss[J]. Computer Science,2022,49(6):210-216. DOI: 10.11896/jsjkx.210300267
[17] 张富凯,孙一冉,武旭峰,等. 基于深度学习的煤矿钻杆实时计数方法[J/OL]. 煤炭科学技术:1-12[2025-02-21]. https://kns.cnki.net/kcms/detail/11.2402.TD.20240617.1542.006.html. ZHANG Fukai,SUN Yiran,WU Xufeng,et al. Real time counting method for coal mine drill pipes based on deep learning[J/OL]. Coal Science and Technology:1-12[2025-02-21]. https://kns.cnki.net/kcms/detail/11.2402.TD.20240617.1542.006.html.
[18] BUSLAEV A,IGLOVIKOV V I,KHVEDCHENYA E,et al. Albumentations:fast and flexible image augmentations[J]. Information,2020,11(2). DOI: 10.3390/info11020125.
[19] 唐晓彬,沈童. 深度学习框架发展综述[J]. 调研世界,2023(4):83-88. TANG Xiaobin,SHEN Tong. Review of development of deep learning framework[J]. The World of Survey and Research,2023(4):83-88.
[20] 张少康,王超,孙芹东. 基于多类别特征融合的水声目标噪声识别分类技术[J]. 西北工业大学学报,2020,38(2):366-376. DOI: 10.3969/j.issn.1000-2758.2020.02.018 ZHANG Shaokang,WANG Chao,SUN Qindong. Underwater target noise recognition and classification technology based on multi-classes feature fusion[J]. Journal of Northwestern Polytechnical University,2020,38(2):366-376. DOI: 10.3969/j.issn.1000-2758.2020.02.018
[21] 戴得恩,朱瑞飞,陈长征,等. 基于改进Yolov5l的航空小目标检测算法[J]. 计算机工程与设计,2023,44(9):2610-2618. DAI De'en,ZHU Ruifei,CHEN Changzheng,et al. Aerial small target detection method based on improved Yolov5l[J]. Computer Engineering and Design,2023,44(9):2610-2618.
[22] AHMED A,IMRAN A S,MANAF A,et al. Enhancing wrist abnormality detection with YOLO:analysis of state-of-the-art single-stage detection models[J]. Biomedical Signal Processing and Control,2024,93. DOI: 10.1016/j.bspc.2024.106144.
[23] 韩慧妍,张秀权,况立群,等. 基于YOLOv8L遥感图像旋转目标检测[J]. 激光与红外,2024,54(9):1462-1468. DOI: 10.3969/j.issn.1001-5078.2024.09.018 HAN Huiyan,ZHANG Xiuquan,KUANG Liqun,et al. Rotating object detection of remote sensing image based on YOLOv8L[J]. Laser & Infrared,2024,54(9):1462-1468. DOI: 10.3969/j.issn.1001-5078.2024.09.018
[24] 安徽工程大学. 基于改进YOLOv10网络的道路缺陷检测方法:CN202411303739. X[P]. 2025-01-07. Anhui Polytechnic University. Road defect detection method based on improved YOLOv10 network:CN202411303739. X[P]. 2025-01-07.
计量
- 文章访问数: 79
- HTML全文浏览量: 23
- PDF下载量: 19
