Intelligent identification and positioning of steel belt anchor hole in coal mine roadway support
-
摘要: 煤矿井下掘进巷道使用钢带辅助锚杆支护时，如果支护钢带锚孔的定位不准确，钻头打在钢带或锚网上易造成设备损坏，存在较大安全隐患。针对上述问题，提出了一种基于改进YOLOv5s模型的煤矿巷道支护钢带锚孔智能识别与定位方法。① 通过超分辨率重构技术（SR）增加锚孔图像的清晰度，防止因图像模糊而丢失图像中锚孔边缘高频信息。② 由于锚孔较小且摄像头距锚孔有一定距离，在卷积神经网络中易丢失较小锚孔的特征信息，影响锚孔检测效果。在YOLOv5s模型的Backbone网络中添加坐标注意力机制（CA）模块，增加YOLOv5s网络中特征提取网络的网络层数，将目标物体的坐标信息融入到卷积网络中，可有效提取锚孔小目标特征信息，从而提高锚孔检测成功率。③ 利用嵌入CA模块的YOLOv5s网络训练经SR重构后的锚孔数据集，得到改进后的YOLOv5s模型，即SR−CA−YOLOv5s模型。④ 采用SR−CA−YOLOv5s模型结合双目摄像头对锚孔进行实时识别与定位。实验结果表明：相较于YOLOv5s模型，SR−CA−YOLOv5s模型的平均精度均值为96.8%，较YOLOv5s模型提高了3.1%；SR−CA−YOLOv5s模型有更好的检测能力，在一定程度上避免了漏检；虽然SR−CA−YOLOv5s模型的每秒填充图像的帧数（FPS）降低了18.5帧/s，但其FPS仍保持在166.7帧/s，并不影响模型的实时检测功能。实际测试结果表明：SR−CA−YOLOv5s模型能够在不同光照条件下准确检测出锚孔并获得锚孔相对于摄像头的三维坐标，坐标误差在6 mm以内，且FPS满足实时性要求。Abstract: When the steel belt auxiliary bolt is used in the coal mine underground heading roadway, if the positioning of the steel belt anchor hole is not accurate, the drill bit is easy to cause equipment damage when hitting the steel belt or anchor net. There are large potential safety hazards. In order to solve the above problems, an intelligent identification and positioning method of steel belt anchor hole in coal mine roadway support based on improved YOLOv5s model is proposed. ① The definition of the anchor hole image is increased by the super-resolution(SR). The high-frequency information of the anchor hole edge in the image is prevented from being lost due to image blurring. ② Because the anchor hole is small and the camera has a certain distance from the anchor hole, it is easy to lose the characteristic information of the small anchor hole in the convolutional neural network. This affects the detection effect of the anchor hole. The coordinate attention mechanism (CA) module is added to the Backbone network of YOLOv5s model. The network layers of the characteristic extraction network in the YOLOv5s network are increased. The coordinate information of the target object is integrated into the convolutional network. The characteristic information of the anchor hole small target can be effectively extracted, and the success rate of anchor hole detection is improved. ③ The YOLOv5s network embedded in the CA module is trained to the anchor hole dataset reconstructed by SR, and the improved YOLOv5s model, namely SR-CA-YOLOv5s model, is obtained. ④ The SR-CA-YOLOv5s model combined with the binocular camera is used to identify and locate the anchor hole in real-time. The experimental results show that compared with the YOLOv5s model, the mean average precision of the SR-CA-YOLOv5s model is 96.8%, which is 3.1% higher than the YOLOv5s model. The SR-CA-YOLOv5s model has better detection capability and avoids missing detection to a certain extent. Although the frames per second (FPS) of the SR-CA-YOLOv5s model is reduced by 18.5 frames/s, its FPS remains at 166.7 frames/s, which does not affect the real-time detection function of the model. The actual test results show that the SR-CA-YOLOv5s model can accurately detect the anchor hole and obtain the three-dimensional coordinate of the anchor hole relative to the camera under different lighting conditions. The coordinate error is within 6 mm, and the FPS meets the real-time requirements.
-
图 1 锚孔智能识别与定位流程
Figure 1. Process of intelligent identification and positioning of anchor hole
表 2 YOLOv5s模型与SR−CA−YOLOv5s模型对比结果
Table 2. Comparison results of the YOLOv5s model and the SR-CA-YOLOv5s model
% 模型 准确率 召回率 mAP YOLOv5s 93.9 92.8 93.7 SR−CA−YOLOv5s 96.6 97.0 96.8 
下载: 导出CSV
表 3 消融实验结果
Table 3. Results of the ablation test
模型 SR 加入CA机制 mAP/% FPS/(帧·s−1) YOLOv5s × × 93.7 185.2 SR−YOLOv5s √ × 94.5 119.0 CA−YOLOv5s × √ 95.5 169.5 SR−CA−YOLOv5s √ √ 96.8 166.7
下载: 导出CSV
表 4 锚孔识别与定位结果
Table 4. Anchor hole identification and positioning results
序号 实际坐标/mm 检测坐标/mm 误差/mm FPS/(帧·s−1) 1 (225,36,542) (221,34,538) (4,2,4) 30.1 2 (518,−10,1012) (519,−13,1015) (1,3,3) 28.5 3 (−355,237,680) (−350,236,675) (5,1,5) 28.7 4 (−247,−106,735) (−251,−108,741) (4,2,6) 29.8 5 (−112,209,1021) (−109,206,1020) (3,3,1) 28.1
下载: 导出CSV
-
[1] 王步康. 煤矿巷道掘进技术与装备的现状及趋势分析[J]. 煤炭科学技术,2020,48(11):1-11. doi: 10.13199/j.cnki.cst.2020.11.001WANG Bukang. Current status and trend analysis of readway driving technology and equipment in coal mine[J]. Coal Science and Technology,2020,48(11):1-11. doi: 10.13199/j.cnki.cst.2020.11.001 [2] 薛光辉,程继杰,管健,等. 深部综掘巷道机器人化超前支护方案与最佳支护时机研究[J]. 矿业科学学报,2019,4(4):349-356.XUE Guanghui,CHENG Jijie,GUAN Jian,et al. Robotized advance support scheme and optimum support opportunity for deep fully mechanized roadway[J]. Journal of Minging Science and Technology,2019,4(4):349-356. [3] 康红普,王金华,林健. 煤矿巷道支护技术的研究与应用[J]. 煤炭学报,2010,35(11):1809-1814. doi: 10.13225/j.cnki.jccs.2010.11.008KANG Hongpu,WANG Jinhua,LIN Jian. Study and applications of roadway support techniques for coal mines[J]. Journal of China Coal Society,2010,35(11):1809-1814. doi: 10.13225/j.cnki.jccs.2010.11.008 [4] 康红普,王金华,林健. 煤矿锚杆支护应用实例分析[J]. 岩石力学与工程学报,2010,29(4):649-664.KANG Hongpu,WANG Jinhua,LIN Jian. Case studies of rock bolting in coal mine roadways[J]. Chinese Journal of Rock Mechanics and Engineering,2010,29(4):649-664. [5] 刘送永,张德义. 巷道掘进机智能化技术研究现状及展望[J]. 工矿自动化,2019,45(10):23-28. doi: 10.13272/j.issn.1671-251x.2019030031LIU Songyong,ZHANG Deyi. Research status and prospect of intelligentization technology of roadheader[J]. Industry and Mine Automation,2019,45(10):23-28. doi: 10.13272/j.issn.1671-251x.2019030031 [6] 王传武. 超重型悬臂式掘进机在岩巷中的应用[J]. 煤矿机械,2019,40(7):154-155. doi: 10.13436/j.mkjx.201907052WANG Chuanwu. Application of super heavy boom-type roadheader in rock roadway[J]. Coal Mine Machinery,2019,40(7):154-155. doi: 10.13436/j.mkjx.201907052 [7] 贺安民,武利民,杜善周,等. 煤矿全自动两臂顶锚杆钻车研制[J]. 煤炭科学技术,2019,47(增刊2):165-169.HE Anmin,WU Limin,DU Shanzhou,et al. Development of fully automatic two-arm top roof bolter in coal mine[J]. Coal Science and Technology,2019,47(S2):165-169. [8] 吴浩,刘锋,王士伟,等. 四臂锚杆钻车的设计[J]. 煤矿机械,2011,32(7):23-25. doi: 10.3969/j.issn.1003-0794.2011.07.012WU Hao,LIU Feng,WANG Shiwei,et al. Design of four-arms bolt driller[J]. Coal Mine Machinery,2011,32(7):23-25. doi: 10.3969/j.issn.1003-0794.2011.07.012 [9] 范要辉. 六臂掘锚机的设计研究[J]. 煤炭技术,2021,40(8):205-207.FAN Yaohui. Design and research of six-arm bolter-miner[J]. Coal Technology,2021,40(8):205-207. [10] 王威. 十臂锚杆钻车的研制与应用[J]. 煤矿机电,2021,42(2):21-23.WANG Wei. Development and application of bolt drill truck with ten arm[J]. Colliery Mechanical & Electrical Technology,2021,42(2):21-23. [11] 张改莲. 基于PLC的锚杆钻机智能可视化控制系统应用研究[J]. 机床与液压,2018,46(16):130-132,129. doi: 10.3969/j.issn.1001-3881.2018.16.036ZHANG Gailian. Intelligent visualization of anchor drill based on PLC[J]. Machine Tool & Hydraulics,2018,46(16):130-132,129. doi: 10.3969/j.issn.1001-3881.2018.16.036 [12] 唐会成. EPEC控制器在锚杆钻臂电液控制中的应用[J]. 煤矿机械,2020,41(7):179-181. doi: 10.13436/j.mkjx.202007060TANG Huicheng. Application of EPEC controller in electro-hydraulic control of anchor drill arm[J]. Coal Mine Machinery,2020,41(7):179-181. doi: 10.13436/j.mkjx.202007060 [13] GOODFELLOW I,POUGET-ABADIE J,MIRZA M,et al. Generative adversarial nets[J]. Neural Information Processing Systems,2014,3(1):2672-2680. [14] WANG Zhihao,CHEN Jian,HOI S C H. Deep learning for image super-resolution:a survey[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2021,43(10):3365-3387. doi: 10.1109/TPAMI.2020.2982166 -
点击查看大图
图(3) / 表(4)
计量
- 文章访问数: 440
- HTML全文浏览量: 123
- PDF下载量: 51
- 被引次数: 0
