img

QQ群聊

img

官方微信

高级检索

黄金科学技术

• 采选技术与矿山管理 • 上一篇    下一篇

基于卷积神经网络的矿工安全帽佩戴识别研究

毕林1,2,谢伟1,2,崔君1,2   

  1. 1.中南大学资源与安全工程学院,湖南  长沙    410083;
    2.中南大学数字矿山研究中心,湖南  长沙    410083
  • 收稿日期:2016-06-27 修回日期:2017-02-15 出版日期:2017-08-30 发布日期:2017-10-30
  • 作者简介:毕林(1975-),男,四川通江人,讲师,从事数字矿山研究工作。mr.bilin@163.com
  • 基金资助:

    国家自然科学基金项目“基于深度学习和距离场的复杂金属矿体三维建模技术研究”(编号:41572317)资助

Identification Research on the Miner’s Safety Helmet Wear Based on Convolutional Neural Network

BI Lin 1,2,XIE Wei 1,2,CUI Jun 1,2   

  1. 1.School of Resources and Safety Engineering,Central South University,Changsha    410083,Hunan,China;
    2.Center of Digital Mine Research,Central South University,Changsha    410083,Hunan,China
  • Received:2016-06-27 Revised:2017-02-15 Online:2017-08-30 Published:2017-10-30

摘要:

为了解决矿山安全监控监测主要靠人工对视频数据进行识别而存在的诸多人为因素限制的问题,通过构建卷积神经网络实现计算机智能识别矿工安全帽的佩戴,在不增加任何辅助装置的条件下实现矿工安全着装智能识别。从视频数据中提取图像,通过对图像进行旋转、偏移、剪切等预处理,将图像分为矿山背景、戴安全帽的矿工和不戴安全帽的矿工3类。通过构建3种不同深度层次的卷积神经网络进行实验对比,“4个卷积层+3个池化层+3个全连接层”组成的深层网络识别准确率较高,达到91.2%。实验表明利用卷积神经网络可以较好地实现对矿工是否正确佩戴安全帽的智能识别。研究方法为人工智能应用于矿山的安全监控、安全行为及安全状态的智能识别研究提供借鉴。

关键词: 矿山安全, 卷积神经网络, 智能识别, 安全帽, 安全生产, Caffe深度学习框架

Abstract:

In order to solve human factors restrictions of mine safety monitoring,which relied on manual monitoring of video data to identify risk factors,a method of construction of deep convolution network to identity whether miners wear helmets without adding any auxiliary device.Images was extracted from video data,which was divided into three categories:background,miners wearing helmets and miners without helmets,through rotated offset and sheared for image.The experimental companison is made by constructing three different levels of convolutional neural networks.Experiment shows that deep convolution network which was developed by“4 convolution layers+3 pooling layers+3 fully connected layers”has a highest recognition rate,reached 91.2%.Convolution neural network can achieve intelligent identification of miners dress safety.Research show that intelligent recognition of mine has an important reference for safety monitoring,safe behavior and security status .

Key words: mine safety, convolution neural network, intelligent identification, helmet, safety in production, Caffe deep learning framework

中图分类号: 

  • X936

[1] Ji Xuewen.Study on the intelligent construction of new mines[J].China Mine Engineering,2015,44(2):60-64.[吉学文.新矿山智能化建设问题探讨[J].中国矿山工程,2015,44(2):60-64.]
[2] Hu Tian,Wang Xingang.Analysis and design of safety helmet recognition system based on wavelet transform and neural network[J].Software Guide,2006(23):37-39.[胡恬,王新刚.基于小波变换和神经网络的安全帽识别系统分析与设计[J].软件导刊,2006(23):37-39.]
[3] Liu Yunbo,Huang Hua.Research on monitoring of workers’helmet wearing at the construction site[J].Electronic Science and Techndogy,2015,28(4):69-72.[刘云波,黄华.施工现场安全帽佩戴情况监控技术研究[J].电子科技,2015,28(4):69-72.]
[4]  Yu Kai,Jia Lei,Chen Yuqiang,et al.Deep learning:Yester- day,today and tomorrow[J].Journal of Computer Research and Development,2013,50(9):1799-1804.余凯,贾磊,陈雨强,等.深度学习的昨天、今天和明天[J].计算机研究与发展,2013,50(9):1799-1804.
[5] Hinton G E,Osindero S,Teh Y W.A fast learning algo- rithm for deep belief nets[J].Neural Computer,2006,18(7): 1527-1554.
[6] Li Zhenhua,Xu Yanchun,Li Longfei,et al. Forecast of the height of water flowing fractured zone based on BP neural networks[J].Journal of Mining & Safety Engineering,2015, 32(6):905-910.[李振华,徐延春,李龙飞,等.基于BP神经网络的导水裂隙带高度预测[J].采矿与安全工程学报,2015,32(6):905-910.]
[7] Li Huimin,Li Zhenlei,He Rongjun,et al.Rock burst risk eva- luation based on particle swarm optimization and BP neural network[J].Journal of Mining & Safety Engineering,2014,31(2):203-207,231.[李慧民,李振雷,何荣军,等.基于粒子群算法和BP 神经网络的冲击危险性评估[J].采矿与安全工程学报,2014,31(2):203-207,231.]
[8] Wang Qian,Zhang Haixian.The depth of vehicle recognition based on neural network[J].Modern Computer,2015(35):61-64.[王茜,张海仙.基于深度神经网络的汽车车型识别[J].现代计算机(专业版),2015(35):61-64.]
[9] Fan Heng,Xu Jun,Deng Yong,et al.Behavior recognition of human based on deep learning[J].Geomatics and Information Science of Wuhan University,2016,4(4):492-497.[樊恒,徐俊,邓勇,等.基于深度学习的人体行为识别[J].武汉大学学报(信息科学版),2016,4(4):492-497.]
[10] Jia Y Q,Shelhamer E,Donahue F,et al.Caffe:Convolutional architecture for fast feature embedding[C]//Proceedings of the 22nd ACM International Conference on Multimedia,2014:675-678.
[11] Tompson J,Jain A,Lecun Y,et al.Joint training of a convolutional network and a graphical model for human pose estimation[J].Eprint Arxiv,2014:1799-1807.
[12]  Fathi A,Mori G.Action recognition by learning mid-level motion features[C]//IEEE International Conference on Com- puter Vision,2009:1-8.
[13] Krizhevsky A,Sutskever I,Hinton G E.ImageNet classi- fication with deep convolutional neural networks[C]// Advances in Neural Information Processing Systems,2012, 25:1-8.
[14] Chen Xianchang.Research and Application of Deep Learning Algorithm Based on Convolutional Neural Network[D].Hangzhou:Zhejiang Gongshang University,2013.[陈先昌.基于卷积神经网络的深度学习算法与应用研究[D].杭州:浙江工商大学,2013.]
[15] He K,Zhang X,Ren S,et al.Scalable,spatial pyramid pooling in deep convolutional network for visual recognition[J].IEEE Transactions on Pattern Analysis & Machine Intelligence,2014,37(9):1904.
[16] Chen C C,Aggarwal J K.An adaptive background model initialization algorithm with object moving at different depths[C]//IEEE International Conference on Image Proces- sing,2006:1-7.
[17] Assael Y M,Wahlstrom N,Schon T B,et al.Data-efficient learning of feedback policies from image pixels using deep dynamical models[J].Computer Science,2015,48(28):1059-1064.
[18] LeCun Y,Bengio Y,Hinton G.Deep learning[J].Nature,2015,521:436-444.
[19] Bengio Y.Deep generative stochastic networks trainable by backprop[J].Department of Computer Science,Cornell Uni- versity,2013:5-7.
[20] LeCun Y,Bottou L,Bengio Y,et al.Gradient-based learning applied to document recognization[J].Proceedings of the IEEE,1998,86(11):2278-2324.
[21] Lian Xiaoyan,Deng Fang.CAPTCHA recognition based on image recognition and neural networks[J].Journal of Cen- tral South University(Science and Technology),2011,42(S):49-52.[连晓岩,邓方.基于图像识别和神经网络的验证码识别[J].中南大学学报(自然科学版),2011,42(增):49-52.]
[22] LeCun Y.Generalization and network design strategy[R].North-Holland:Connections in Perspective,1989:1-6.
[23]  LeCun  Y,Bengio Y,Szegedy G H,et al.Going deeper with convolution[J].Preprint arxiv,1409.4842,2014:5.
[24] Nair V,Hinton G E.Rectified linear units improve restricted boltzmann machines[J].Toronto:University of Toronto,2010: 3-6.
[25] Robinson A E,Hammon P S,De Sa V R.Explaining brightness illusions using spatial filtering and local response normalization[J].Vision Research,2007,47(12):1631-1644.
[26] Srivastava N,Hinton G,Krizhevsky A,et al.Dropout:A simple way to prevent neural networks from overfitting[J].The Journal of Mac

[1] 杨超, 史秀志. 基于Fisher判别法模型的矿山安全标准化等级评价[J]. 黄金科学技术, 2018, 26(2): 187-194.
[2] 林伟,胡汉华* . 人机学视阈下的矿井通风系统设计与思考[J]. 黄金科学技术, 2017, 25(3): 121-127.
[3] 刘连生,兰德沛,张声辉,潘荣森,蔡建德 . 频分多址盲炮检测技术与系统在深孔爆破中的应用研究[J]. 黄金科学技术, 2017, 25(2): 54-61.
[4] 李金玲,王李管,陈鑫. 露天台阶爆破矿岩交界处损失贫化控制系统[J]. 黄金科学技术, 2016, 24(3): 14-20.
[5] 冯雪飞,罗周全,汪伟,路凡. 采空区失稳辨析的RS-CPM方法研究及应用[J]. 黄金科学技术, 2015, 23(3): 67-72.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!