|本期目录/Table of Contents|

 TU Siyu,PENG Pingan,JIANG Yuanjian.Research on intelligent recognition technology of abnormal operating conditions in underground environment based on deep learning method[J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2018,14(11):58-63.[doi:10.11731/j.issn.1673-193x.2018.11.009]





Research on intelligent recognition technology of abnormal operating conditions in underground environment based on deep learning method
(中南大学 资源与安全工程学院,湖南 长沙 410083)
TU Siyu PENG Pingan JIANG Yuanjian
(School of Resources and Safety Engineering, Central South University, Changsha Hunan 410083, China)
underground environment unmanned mining abnormal operating conditions deep learning method
为了使装备在井下复杂环境中具有自主精准辨识井下对象和作业环境的能力,实现无轨装备及其他各类型装备无人化作业和开采,提出了基于利用深度学习方法开展井下环境异常工况智能识别分类研究。通过分析装备正常运行和作业的主要影响因素,构建了井下环境异常工况数据集,采用旋转变换、平移变换、缩放变换等数据增强技术,有效防止网络训练过拟合;基于InceptionResnetV2模型采用层冻结方法,重新训练全连接模型,通过不同的迁移策略进行实验对比分析。研究结果表明:添加2层全连接层,且每层包括4 096个神经元的迁移策略模型性能最佳,鲁棒性好,能够精准识别分类井下环境异常工况。
In order to make the equipments own the ability to identify the underground objects and operating environment independently and accurately in the underground complex environment, and realize the unmanned operating and mining of trackless equipments and other types of equipments, the research on intelligent recognition and classification of abnormal operating conditions in the underground environment based on the deep learning method was carried out. The dataset of abnormal operating conditions in the underground environment was established through analyzing the main factors affecting the normal running and operation of equipments, and the data enhancement technologies such as rotation transformation, translation transformation, zoom transformation and others were applied to effectively prevent the network training from overfitting. The fully connected model was retrained by using the layer freeze method based on the InceptionResnetV2 model, and the experimental comparative analysis was conducted with different transfer strategies. The results showed that the transfer strategy model with adding twolayer fully connected layer and containing 4096 neurons in each layer had the optimal performance with a good robustness, and it could identify the abnormal operating conditions in the underground environment accurately.


[1]FLORES O, ACUA E. Improving monitoring and control hardware cost at Totten Mine[C]//Canadian Institute of Mining (CIM): proceedings of the Maintenance, Engineering and Reliability/Mine Operators (MEMO) 2016 Conference, Sudbury, Canada, 2016.
[5]来自建龙重工集团思山岭铁矿的纵深报道之三[EB/OL].(2015-08-17).国土部网站. http://money.163.com/15/0817/13/B17NHN0F00253B0H.html.
[6]HON V, DOSTL M, SLANINA Z. The use of RFID technology in mines for identification and localization of persons [J]. IFAC Proceedings Volumes, 2010, 43(24): 236-239.
[7]叶晨成,校景中,肖丽.基于RFID的井下人员定位系统[J].武汉理工大学学报,2010,32(15):146-149. YE Chencheng,XIAO Jingzhong,XIAO Li. Personnel Positioning system of underground mines based on RFID[J]. Wuhan Ligong Daxue Xuebao(Journal of Wuhan University of Technology), 2010, 32(15): 146-149.
[8]BANDYOPADHYAY L K, CHAULYA S K, MISHRA P K. Wireless communication in underground mines: RFID-based sensor networking[M]. Springer Science & Business Media, 2009.
[9]YALCIN O, SAYAR A, ARAR O F, et al. Detection of road boundaries and obstacles using LIDAR[C]//Computer Science and Electronic Engineering Conference (CEEC), 2014 6th. IEEE, 2014: 6-10.
[10]金焱飞,张会林,杨迪瑞,等.有轨电车激光雷达障碍物探测的决策方法[J].电子测量技术,2017,40(3):197-200. JIN Yanfei,ZHANG Huilin,YANG Dirui, et al. Lidar-based tram's decision-making method for obstacle detection[J]. Electronic Measurement Technology, 2017, 40(3): 197-200.
[11]孟宇, 刘立, 李文辉. 用于井下移动设备定位的路标设计及其视频识别方法[C]//2011年中国智能自动化学术会议论文集 (第一分册),2011:770-772.
[12]KRIZHEVSKY A, SUTSKEVER I, HINTON G E. Imagenet classification with deep convolutional neural networks[C]//Advances in neural information processing systems, 2012: 1097-1105.
[13]RUSSAKOVSKY O, DENG J, SU H, et al. Imagenet large scale visual recognition challenge[J]. International Journal of Computer Vision, 2015, 115(3): 211-252.
[14]HE K, ZHANG X, REN S, et al. Deep residual learning for image recognition[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2016: 770-778.
[15]YAN S, XU D, ZHANG B, et al. Graph embedding and extensions: A general framework for dimensionality reduction[J]. IEEE transactions on pattern analysis and machine intelligence, 2007, 29(1): 40-51.
[16]KIM S, JI M, KIM H. Noise-robust speaker recognition using subband likelihoods and reliable-feature selection[J]. ETRI journal, 2008, 30(1): 89-100.
[17]IOFFE S, SZEGEDY C. Batch Normalization: Accelerating deep network training by reducing internal covariate shift[C]//International Conference on Machine Learning. 2015: 448-456.
[18]MAIER K L, FILDANI A, PAULL C K, et al. Deep-sea channel evolution and stratigraphic architecture from inception to abandonment from high-resolution autonomous underwater vehicle surveys offshore central California[J]. Sedimentology, 2013, 60(4): 935-960.
[19]FERREIRA C A, MELO T, SOUSA P, et al. Classification of breast cancer histology images through transfer learning using a pre-trained inception resnet V2[C]//International Conference Image Analysis and Recognition. Springer, Cham, 2018: 763-770.
[20]CHEN J, WANG Y, WANG D L. Noise perturbation improves supervised speech separation[C]//International Conference on Latent Variable Analysis and Signal Separation, Springer, Cham, 2015: 83-90.
[21]SCHERER D, MüLLER A, BEHNKE S. Evaluation of pooling operations in convolutional architectures for object recognition[C]//Artificial Neural Networks-ICANN 2010, Springer, Berlin, Heidelberg, 2010: 92-101.
[22]YU D, WANG H, CHEN P, et al. Mixed pooling for convolutional neural networks[C]//International Conference on Rough Sets and Knowledge Technology, Springer, Cham, 2014: 364-375.
[23]STEWART R T, QUINTANA I O. Probabilistic opinion pooling with imprecise probabilities[J]. Journal of Philosophical Logic, 2018, 47(1): 17-45.
[24]GULCEHRE C, CHO K, PASCANU R, et al. Learned-norm pooling for deep feedforward and recurrent neural networks[C]//Joint European Conference on Machine Learning and Knowledge Discovery in Databases, Springer, Berlin, Heidelberg, 2014: 530-546.
[25]PAN S J, YANG Q. A survey on transfer learning[J]. IEEE Transactions on knowledge and data engineering, 2010, 22(10): 1345-1359.
[26]王帅,周国民,王健.主题爬虫相关度算法研究综述[J].计算机与现代化,2013(4):27-30,39. WANG Shuai, ZHOU Guomin, WANG Jian. Reviews of relevance algorithm in focused crawler[J]. Computer and Modernization, 2013(4):27-30,39.



收稿日期: 2018-11-01
基金项目: 国家重点研发计划(2017YFC0602905)
作者简介: 涂思羽,硕士研究生,主要研究方向为安全隐患智能识别。
通信作者: 彭平安,博士研究生,主要研究方向为数字矿山和矿山安全。
更新日期/Last Update: 2018-12-03