img

QQ群聊

img

官方微信

高级检索

黄金科学技术 ›› 2021, Vol. 29 ›› Issue (1): 136-146.doi: 10.11872/j.issn.1005-2518.2021.01.089

• 采选技术与矿山管理 • 上一篇    

基于计算机视觉技术的矿井电机车无人驾驶障碍物检测技术

王京华(),王李管,毕林()   

  1. 中南大学资源与安全工程学院,湖南 长沙 410083
  • 收稿日期:2020-05-18 修回日期:2020-08-03 出版日期:2021-02-28 发布日期:2021-03-22
  • 通讯作者: 毕林 E-mail:185511062@csu.edu.cn;mr.bilin@163.com
  • 作者简介:王京华(1996-),男,河南洛阳人,硕士研究生,从事数字矿山和轨道障碍物检测等方面的研究工作。185511062@csu.edu.cn
  • 基金资助:
    国家重点研发计划项目“深部集约化开采生产过程智能管控技术”(2017YFC0602905)

Obstacle Detection Technology of Mine Electric Locomotive Driverless Based on Computer Vision Technology

Jinghua WANG(),Liguan WANG,Lin Bi()   

  1. School of Resources and Safety Engineering,Central South University,Changsha 410083,Hunan,China
  • Received:2020-05-18 Revised:2020-08-03 Online:2021-02-28 Published:2021-03-22
  • Contact: Lin Bi E-mail:185511062@csu.edu.cn;mr.bilin@163.com

摘要:

针对传统计算机视觉方法难以实现障碍物实时检测和定位的问题,提出传统计算机视觉技术与深度学习目标检测算法YOLOv3相结合的障碍物智能检测方法。首先,采集电机车行驶区域(即有效检测区域)存在的障碍物数据并制作VOC格式数据集,使用YOLOv3训练数据集,得到障碍物检测模型;然后,采用传统计算机视觉技术定位到轨道,使用“3邻域”搜索法获得轨道线坐标值,根据距离信息向轨道外侧扩展一定距离,提取有效检测区域,同时网格化图片,将障碍物的坐标换算为实际距离;最后,使用障碍物检测模型对有效检测区域进行检测。试验结果表明:该方法可以识别行驶区域内多种特征差异很大的目标物体,如电机车、人和大块落石等;该方法每秒可以处理6帧图片,现场采集的实际数据测试平均精确率达到93.2%。

关键词: 地下矿, 无人驾驶电机车, 障碍物智能检测, 计算机视觉, YOLOv3, 有效检测区域

Abstract:

The mining of mineral resources is getting deeper and deeper,the working environment is bad,the employees are aging seriously,and the human cost is rising,which brings great challenges to the development of mining industry.Intelligent mine operation has become an inevitable trend.As a part of mine intellec-tualization,unmanned transportation system is very important for mine,which means the improvement of safety production efficiency can achieve zero injury and zero time loss.Considering the development demand of driverless electric vehicle and the traditional computer vision method,it is difficult to realize the real-time detection and location of obstacles.An intelligent obstacle detection method based on the combination of traditional computer vision technology and deep learning target detection algorithm YOLOv3 was proposed.First of all,the video data of obstacles in the driving area of electric locomotive(this paper calls it effective detection area) were collected,using an image annotation tool namely labelimg to make VOC data set,using YOLOv3 to train data set,according to the feedback results,adjust the parameters continuously to obtain the relative optimal parameters,and finally get the obstacle detection model.Then use the traditional computer vision technology to locate the track by edge,texture and other information,using the “3 neighborhood” search method to get the track line coordinate value of left and right track lines,and expand a certain distance to the outside of the track according to the distance information,extracting the effective detection area,thus reducing the computational complexity of the later obstacle detection,at the same time,gridding images,converting the coordinates of obstacles to the actual distance.Finally,using the obstacle detection model to detect the effective detection area and respond to the detection results.Experimental results show that the method can identify many objects with different characteristics in the driving area,such as electric locomotive,people,large falling rocks,etc.It can process 6 pictures per second,and the average accuracy of the actual data collected in the field reaches 93.2%,it has good performance in real-time and accuracy,and has a good effect in the underground mine scene.

Key words: underground mine, driverless electric locomotive, intelligent obstacle detection, computer vision, YOLOv3, effective detection area

中图分类号: 

  • TD524

图1

障碍物智能化检测技术构架"

表1

YOLOv3和Mask R-CNN试验结果对比"

模型名称平均单张测试时间/s平均精确率MAP/%
YOLOv30.0293.2
Mask R-CNN0.1194.5

图2

基于YOLOv3的轨道障碍物检测模型结构"

图3

有效检测区域提取总路线"

图4

轨道定位主要步骤"

图5

轨道定位过程图"

图6

“3邻域”搜索法示意图"

图7

有效检测区域提取"

图8

网格化计算距离示意图"

表2

障碍物类别距离和指令关系"

障碍物类别不同距离下的指令
>10 m5~10 m<5 m
鸣笛鸣笛减速刹车
石块减速减速刹车
电机车鸣笛鸣笛减速刹车

图9

训练集和验证集的损失率折线图"

图10

训练过程平均损失率和平均交叠率曲线"

图11

测试精确率和召回率折线图"

表3

障碍物检测结果"

障碍物类型检测精确率/%
92.3
石块90.0
电机车97.3
平均精度均值93.2
平均单帧检测时间0.02 s

图12

障碍物智能化检测示例1"

图13

障碍物智能化检测示例2"

Deng J,Dong W,Socher R,al et,2009.ImageNet:A large-scale hierarchical image database[C]//2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2009).Miami:IEEE:248-255.
Dewan A,Caselitz T,Tipaldi G D,al et,2016. Motion-based detection and tracking in 3D LiDAR scans[C]// 2016 IEEE International Conference on Robotics and Automation (ICRA). Washington DC:IEEE:4508-4513.
Guo Chunming,2019.Obstacle detection method in front of track based on video image[J].Electronic Measurement Technology,42(12):55-59.
Güzel M S,Zakaria W,2013.A hybrid architecture for vision-based obstacle avoidance[J].Advances in Mechanical Engineering,5:545-550.
He K M,Zhang X Y,Ren S Q,al et,2016.Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition(CVPR). Las Vegas:IEEE:770-778.
He K,Gkioxari G,Dollár P,al et,2020.Mask R-CNN[J].IEEE Transactions on Pattern Analysis & Machine Intelligence,42(2):386-397.
Lin T Y,Dollár P,Girshick R,al et,2017.Feature pyramid networks for object detection[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).Honolulu:IEEE:936-944.
Lin T,Maire M,Belongie S,al et,2014.Microsoft COCO:Common objects in context[C]//Computer Vision-ECCV 2014.Cham:Springer:740-755.
Liu L,Ouyang W,Wang X,al et,2018.Deep learning for generic object detection:A survey[J].International Journal of Computer Vision,128(2):261-318.
Liu Wenqi,2016.Detection Algorithm of Railway Foreign Body Based on Depth Neural Network [D].Beijing:Beijing Jiaotong University.
Lou Xinyu,Wang Hai,Cai Yingfeng,al et,2019. Research on real-time road obstacle detection and classification algorithm using 64 line lidar [J]. Automotive Engineering,41(7):779-784.
Redmon J,Divvala S,Girshick R,al et,2016.You only look once:Unified,real-time object detection[C]// 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas:IEEE:779-788.
Redmon J,Farhadi A,2017. YOLOv3:An incremental improvement[C]// IEEE Conference on Computer Vision and Pattern Recognition(CVPR).Honolulu:IEEE:779-788.
Ruder M,Mohler N,Ahmed F,2003.An obstacle detection system for automated trains[C]//Intelligent Vehicles Symposium. Washington DC:IEEE:180-185.
Song Huaibo,He Dongjian,Xin Xiangjun,2011.Unstructured road detection and obstacle recognition based on machine vision [J].Journal of Agricultural Engineering,27(6):225-230.
Tong Lei,Zhu Liqiang,Yu Zujun,al et,2012.Detection of track foreign matters based on vehicle mounted forward-looking camera[J].Transportation System Engineering and Information,12 (4):79-83.
Wang Xinzhu,Li Jun,Li Hongjian,al et,2016.Automatic obstacle detection method based on 3D lidar and depth image [J]. Journal of Jilin University(Engineering Edition),46(2):360-365.
Xie Desheng,Xu Youchun,Wang Rendong,al et,2018.Obstacle detection and tracking of unmanned vehicle based on 3D lidar[J]. Automotive Engineering,(8):952-959.
Zhuang Fuzhen,Luo Ping,He Qing,al et,2015.Research progress of transfer learning[J].Journal of Software,26(1):26-39.
郭春明,2019.基于视频图像的轨道前方障碍物检测方法[J].电子测量技术,42(12):55-59.
刘文祺,2016.基于深度神经网络的铁路异物检测算法[D].北京:北京交通大学.
娄新雨,王海,蔡英凤,等,2019.采用64线激光雷达的实时道路障碍物检测与分类算法的研究[J].汽车工程,41(7):779-784.
宋怀波,何东健,辛湘俊,2011.基于机器视觉的非结构化道路检测与障碍物识别方法[J].农业工程学报,27(6):225-230.
同磊,朱力强,余祖俊,等,2012.基于车载前视摄像机的轨道异物检测[J].交通运输系统工程与信息,12(4):79-83.
王新竹,李骏,李红建,等,2016.基于三维激光雷达和深度图像的自动驾驶汽车障碍物检测方法[J].吉林大学学报(工学版),46(2):360-365.
谢德胜,徐友春,王任栋,等,2018.基于三维激光雷达的无人车障碍物检测与跟踪[J].汽车工程,(8):952-959.
庄福振,罗平,何清,等,2015.迁移学习研究进展[J].软件学报,26(1):26-39.
[1] 戚伟,李威,李振阳,赵国彦. 基于CRITIC-CW法的地下矿岩体质量评价[J]. 黄金科学技术, 2020, 28(2): 264-270.
[2] 毕林,段长铭,任助理. 基于RANSAC的地下矿山巷道边线检测算法[J]. 黄金科学技术, 2020, 28(1): 105-111.
[3] 胡建华,徐朔寒,徐泽林,韩磊. 城市地下矿山采矿方法的数值与熵权耦合优选[J]. 黄金科学技术, 2019, 27(4): 513-521.
[4] 刘定一, 王李管, 陈鑫, 钟德云, 徐志强. 地下矿中长期计划多目标优化及应用研究[J]. 黄金科学技术, 2018, 26(2): 228-233.
[5] 聂兴信,张国丹. 基于熵值法—突变理论的地下矿山紧急避险系统可靠性研究[J]. 黄金科学技术, 2016, 24(6): 72-77.
[6] 陈建宏,曾闵,李涛,江时雨. 基于物元分析—未确知测度理论的地下矿山安全避险“六大系统”可靠性评估方法[J]. 黄金科学技术, 2015, 23(1): 80-84.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!