黄金科学技术
img

QQ群聊

img

官方微信

高级检索
作者投稿 专家审稿 编辑办公

黄金科学技术 ›› 2024, Vol. 32 ›› Issue (2): 345-355.doi: 10.11872/j.issn.1005-2518.2024.02.150

• 采选技术与矿山管理 • 上一篇    下一篇

基于改进YOLOv8的露天矿区行车障碍物检测

顾清华1,2(),周琼1,2(),王丹1,2   

  1. 1.西安建筑科技大学资源工程学院，陕西 西安 710055
    2.西安建筑科技大学西安市智慧工业感知计算与决策重点实验室，陕西 西安 710055
  • 收稿日期:2023-11-02 修回日期:2024-02-08 出版日期:2024-04-30 发布日期:2024-05-21
  • 通讯作者: 周琼 E-mail:qinghuagu@126.com;zq1121629@163.com
  • 作者简介:顾清华（1981-），男，山东诸城人，教授，从事矿山智能科学与工程研究工作。qinghuagu@126.com
  • 基金资助:
    国家自然科学基金资助项目“金属露天矿无人驾驶多工序多目标协同智能调度方法研究”(52074205);陕西省杰出青年基金资助项目“时空路况下金属露天矿无人驾驶多车协同智能调度集成建模”(2020JC-44)

Traveling Obstacle Detection in Open-Pit Mine Area Based on Improved YOLOv8

Qinghua GU1,2(),Qiong ZHOU1,2(),Dan WANG1,2   

  1. 1.School of Resources Engineering, Xi’an University of Architecture and Technology, Xi’an 710055, Shaanxi, China
    2.Xi’an Key Laboratory of Intelligent Industry Perception, Computing and Decision Making, Xi’an University of Architecture and Technology, Xi’an 710055, Shaanxi, China
  • Received:2023-11-02 Revised:2024-02-08 Online:2024-04-30 Published:2024-05-21
  • Contact: Qiong ZHOU E-mail:qinghuagu@126.com;zq1121629@163.com

RichHTML

1

PDF (PC)

243

摘要：

露天矿区场景复杂，行车障碍物检测受扬尘和颗粒物等粉尘噪声干扰严重，难以准确识别障碍物，尤其是光线较差的夜间，不利于做出正确决策，从而影响无人作业的安全性和整体效率。针对以上问题，提出了一种基于YOLOv8n模型的露天矿区行车障碍物检测算法YOLOv8n-Enhanced。该算法主要从3个方面进行了改进，具体包括：首先，针对受粉尘噪声干扰严重和夜间光线不足的问题，提出了C2fCA模块结构，提高了模型特征提取能力；其次，使用轻量级卷积技术GSConv和VoV-GSCSP模块，减轻模型复杂性，实现检测器更高的计算成本效益；最后，使用WIOU损失函数，提高了模型泛化能力。试验结果表明：改进算法在保持实时性的前提下，可将YOLOv8n的平均精度（mean Average Precision，mAP）分别提高1.8%和2.6%，实现白天与夜间场景下不同尺度的障碍物识别。

关键词: 露天矿区, 无人驾驶, 障碍物检测, YOLOv8检测模型, 矿区复杂场景

Abstract:

The open-pit mining area is a complex scene，and the traveling obstacle detection is seriously interfered by dust noise such as dust and particles，which makes it difficult to accurately identify obstacles，especially at night when the light is poor，which is not conducive to correct decision-making，thus affecting the safety and overall efficiency of unmanned operation.In view of the above problems，a YOLOv8n-based YOLOv8n-Enhanced algorithm for detecting traveling obstacles in open-pit mining areas was proposed.The algorithm is mainly improved in three aspects：Firstly，for the problems of serious interference by dust noise and poor light at night，a C2fCA module structure was proposed instead of the original C2f module，which utilizes the shared weights and context-aware weights to enhance the dependency relationship between different locations of the image，mitigate the noise interference，and improve the feature extraction ability of the model.Secondly，to trade-off the accuracy and real-time performance of the open-pit obstacle detection model，the Neck end of the model was reconstructed，and the lightweight convolutional techniques GSConv and VoV-GSCSP modules were used to reduce the complexity of the computation and network structure，and realize a higher computational cost-effectiveness of the detector.Finally，for the situation that there is a large gap between the quality of data in the open-pit mining area，especially at night when there is insufficient light，and low-quality data will affect the ability of the model to learn features in training，the loss function was optimized to solve the problem of the bounding box regression equilibrium between the samples of different qualities，to improve the ability of the model to generalize and accelerate the convergence.The experimental results show that the improved algorithm in this paper reduces the computational GFLOPs of the model by about 8.5% and the number of parametric params by about 3% while maintaining the real-time performance，and improves the mean Average Precision（mAP） of the YOLOv8n by 1.8% and 2.6% in daytime and nighttime scenarios，respectively，and realizes obstacle recognition at different scales in daytime and nighttime scenes.

Key words: open-pit mining area, unmanned driving, obstacle detection, YOLOv8 detection model, complex scene of mining area

中图分类号: 

  • TP391

图 1

模型结构示意图"

图2

AttnConv模块和C2fCA模块"

图3

GSConv结构和VoV-GSCSP结构"

图4

不同质量数据样本"

图5

不同类别检测目标检测实例"

图6

数据集不同类别的实例分布"

表1

主干网改进后性能对比"

模型

计算量

/GFLOPs

参数量/MmAP50/%
白天夜间
YOLOv8n（Baseline）8.23.0195.792.5
+2 C2fCA8.43.1397.095.0
+1 C2fCA对照8.23.0995.194.3
+3 C2fCA对照8.53.1596.095.4
+4 C2fCA对照8.63.1594.595.0

表2

Neck端改进后性能对比"

模型

计算量

/GFLOPs

参数量/MmAP50/%

计算速度

/FPS

白天夜间
YOLOv8n（Baseline）8.23.0195.792.580
+new-Neck7.42.8097.193.787

表3

损失函数改进后性能对比"

模型

计算量

/GFLOPs

参数量/MmAP50/%
白天夜间
YOLOv8n（Baseline）8.23.0195.792.5
+WIoU8.23.0196.894.2
+2 C2fCA+WIoU8.43.1397.295.0
+new-Neck+WIoU7.42.8097.094.6

表 4

消融试验结果"

模型C2fCAnew-NeckWIoU

计算量

/GFLOPs

参数量

/M

mAP50/%检测速度/FPS
白天夜间
YOLOv8n（Baseline）---8.23.0195.792.580
N1√--8.43.1397.095.061
N2-√-7.42.8097.193.787
N3--√8.23.0196.894.280
N4√√-7.52.9296.394.769
YOLOv8n-Enhanced√√√7.52.9297.595.169

图7

模型改进前后mAP50曲线对比"

图8

改进前后算法效果对比"

表5

不同模型检测性能对比"

模型

参数量

/M

mAP50/%

检测速度

/FPS

白天夜间
YOLOv8n-Enhanced2.9297.595.169
Faster-Rcnn4083.053.228
SSD30026.264.352.434
YOLOv5s7.0793.591.565
YOLOv7-tiny6.0395.092.354
Bochkovskiy A， Wang C， Liao H M，2020.YOLOv4：Optimal speed and accuracy of object detection［J］.ArXiv：.
Dong Lijuan，2023.Research on Mine Roadway Obstacle Detection Based on Infrafed Binocular Vision［D］.Xi’an：Xi’an Unversity of Architecture and Technology.
Fan Q， Huang H， Guan J，et al，2023.Rethinking local perception in lightweight vision transformer［J］.ArXiv： .
Girshick R，2015.Fast R-CNN［C］//Proceeding of the IEEE International Conference on Computer Vision.Piscataway：IEEE Press： 1440-1448.
Girshick R， Donahue J， Darrell T，et al，2014.Rich feature hierarchies for accurate object detection and semantic segmentation［C］//Proceeding of IEEE Conference on Computer Vision and Pattern Recognition.Piscataway：IEEE Press： 580-587.
Li H， Li J， Wei H，et al，2022.Slim-neck by GSConv：A better design paradigm of detector architectures for autonomous vehicles［J］.ArXiv： .
Liu W， Anguelov D，Erhabd，et al，2016.SSD：Single Shot MultiBox Detector［M］//Computer Vision-ECCV 2016.Cham：Springer International Publishing： 21-37．
Liu Yongfeng，2022.Research on Nighttime Road Obstacle Image Semantic Segmentation Based on Attention Mechanism［D］.Changsha：Central South University.
Pathak A R， Pandey M， Rautaray S，2018.Application of deep learning for object detection［J］.Procedia Computer Science，132：1706-1717.
Qin Xiaohui， Huang Qidong， Chang Dengxiang，et al，2023a.Object detection method for open-pit mine based on improved YOLOv5［J］.Journal of Hunan University（Natural Sciences Edition），50（2）：23-30.
Qin Xuebin， Xue Yuqiang， Jing Ningbo，et al，2023b.Research on front obstacle detection algorithm for autonomous mining trucks in open-pit coal mines ［J/OL］.Metal Mine：1-12 ［2023-08-14］..
Redmon J， Divvala S K， Girshick R B，2015.You only look once：Unified，real-time object detection［C］// 2016 IEEE Conference on Computer Vision and Pattern Recognition （CVPR），June 27-30，2016，Las Vegas，NV，USA：779-788.
Redmon J， Farhadi A，2018.YOLOv3：An incremental improvement［J］.ArXiv：.
Ren S， He K， Girshick R，et al，2016.Faster R-CNN：Towards real time object detection with region proposal networks ［J］.IEEE Transactions on Pattern Analysis and Machine Intelligence，39（6）：1137-1149.
Ruan Shunling， Li Shaobo， Gu Qinghua，et al，2023.Road obstacle detection in open-pit mining area based on two-way feature fusion［J］.Journal of China Coal Society，48（3）：1425-1438.
Ruan Shunling， Li Shaobo， Lu Caiwu，et al，2021.Road negative obstacle detection in open-pit mining areas with multi-scale feature fusion［J］.Journal of China Coal Society，46（Supp.2）：1170-1179.
Tong Z， Chen Y， Xu Z，et al，2023.Wise-IoU：Bounding box regression loss with dynamic focusing mechanism［J］.Ar Xiv：2301.10051.
Wang H， Yu Y， Cai Y，et al，2019.A comparative study of state-of-the-art deep learning algorithms for vehicle detection ［J］.IEEE Intelligent Transportation Systems Magazine，11（2）：82-95.
Wang Jinghua， Wang Liguan， Bi Lin，2021.Obstacle detection technology of mine electric locomotive driverless based on computer vision technology［J］.Gold Science and Technology，29（1）：136-146.
Zhang Rui， Gao Shibo， Zhao Xia，et al，2023.Nighttime vehicle object detection algorithm for unmanned driving based on improved YOLOv5s［J］.Electronic Measurement Technology，46（17）：87-93.
Zhang Xi， Liang Bin， Yu Miao，et al，2022.Research on the current situation and development trend of unmanned driving transportation technology in open-pit mines［J］.Coal Engineering，54（6）：132-138.
Zheng Z， Wang P， Liu W，et al，2020.Distance-IoU loss：Faster and better learning for bounding box regression［C］//Proceedings of the AAAI Conference on Artificial Intelligence.Palo Alto，California：AAAI Press： 12993-13000.
Zhu X， Lyu S， Wang X，et al，2021.TPH-YOLOv5：Improved YOLOv5 based on transformer prediction head for Object detection on drone-captured scenarios［C］//2021 IEEE/CVF International Conference on Computer Vision Workshops（ICCVW），Montreal，B C，Canada，2021，pp.2778-2788.doi：10.1109/ICCVW 54120.2021.00312 .
doi: 10.1109/ICCVW 54120.2021.00312
董莉娟，2023.基于红外双目视觉的矿井巷道障碍检测研究［D］.西安：西安建筑科技大学.
刘永锋，2022.融合注意力机制的夜间道路障碍物图像语义分割方法研究［D］.长沙：中南大学.
秦晓辉，黄启东，常灯祥，等，2023.基于改进YOLOv5的露天矿山目标检测方法［J］.湖南大学学报（自然科学版），50（2）：23-30.
秦学斌，薛宇强，景宁波，等，2023.露天煤矿自动驾驶矿卡车前障碍物检测算法研究［J/OL］.金属矿山：1-12［2023-08-14］..
阮顺领，李少博，顾清华，等，2023.基于双向特征融合的露天矿区道路障碍检测［J］.煤炭学报，48（3）：1425-1438.
阮顺领，李少博，卢才武，等，2021.多尺度特征融合的露天矿区道路负障碍检测［J］.煤炭学报报，46（增2）：1170-1179.
王京华，王李管，毕林，2021.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［J］.黄金科学技术，29（1）：136-146.
张蕊，高诗博，赵霞，等，2023.基于改进YOLOv5s的无人驾驶夜间车辆目标检测算法［J］.电子测量技术，46（17）：87-93.
张晞，梁斌，于淼，等，2022.露天矿山无人驾驶运输技术现状及发展趋势研究［J］.煤炭工程，54（6）：132-138.
[1] 顾清华, 杜艺凡, 李萍丰, 王丹. 基于加权双向特征融合的矿区道路落石检测[J]. 黄金科学技术, 2023, 31(6): 953-963.
[2] 阮顺领,董莉娟,卢才武,顾清华. 基于RCR _YOLOv4的矿井巷道红外障碍检测研究[J]. 黄金科学技术, 2022, 30(4): 603-611.
[3] 王鹏飞,毕林,王李管. 露天矿无人驾驶矿卡速度规划研究[J]. 黄金科学技术, 2022, 30(3): 460-469.
[4] 王京华,王李管,毕林. 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[J]. 黄金科学技术, 2021, 29(1): 136-146.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!

©2018 黄金科学技术编辑部

电话：0931-8277791　

E-mail: hjkx@lzb.ac.cn 邮编：730000　

陇ICP备17001318号-1    甘公网安备 62010202000672号
访问总数:    当日访问:    当前在线: