联合DD-GAN和全局特征的井下人员重识别方法

doi:10.19665/j.issn1001-2400.2021.05.023

西安电子科技大学学报 ›› 2021, Vol. 48 ›› Issue (5): 201-211.doi: 10.19665/j.issn1001-2400.2021.05.023

联合DD-GAN和全局特征的井下人员重识别方法

孙彦景^1,²(),魏力¹(),张年龙³(),云霄¹(),董锴文¹(),葛敏¹(),程小舟^1,⁴(),侯晓峰⁵()

1.中国矿业大学信息与控制工程学院,江苏徐州 221116
2.徐州市智能安全与应急协同工程研究中心,江苏徐州 221116
3.安徽马钢罗河矿业有限责任公司生产技术部,安徽合肥 231562
4.中钢集团马鞍山矿山研究院股份有限公司选矿及自动化研究所,安徽马鞍山 243000
5.无锡沃爱思科技有限公司,江苏无锡 214125

收稿日期:2020-09-20 出版日期:2021-10-20 发布日期:2021-11-09
通讯作者: 云霄
作者简介:孙彦景(1977—),男,教授,博士,E-mail: yjsun@cumt.edu.cn|魏力(1996—),女,助理教师,硕士,E-mail: lwei@cumt.edu.cn|张年龙(1985—),男,高级工程师,学士,E-mail: anhuilj@qq.com|董锴文(1996—),男,硕士,E-mail: dongkaiwen@cumt.edu.cn|葛敏(1994—),女,助理工程师,硕士,E-mail: 1462220177@qq.com|程小舟(1981—),男,正高级工程师,硕士,E-mail: cxz3005@163.com|侯晓峰(1980—),男,硕士,E-mail: houxiaofeng@voicon.cn
基金资助:
江苏省自然科学基金青年项目(BK20180640);国家自然科学基金(61902404);国家自然科学基金(51734009);国家自然科学基金(51504255);国家自然科学基金(61771417);国家自然科学基金(62001475);国家重点研发计划(2016YFC0801403);江苏省重点研发计划(BE2015040)

Person re-identification method combining the DD-GAN and Global feature in a coal mine

SUN Yanjing^1,²(),WEI Li¹(),ZHANG Nianlong³(),YUN Xiao¹(),DONG Kaiwen¹(),GE Min¹(),CHENG Xiaozhou^1,⁴(),HOU Xiaofeng⁵()

1. School of Information and Control Engineering,China University of Mining Technology,Xuzhou 221116,China
2. Xuzhou Engineering Research Center of Intelligent Industry Safety and Emergency Collaboration,Xuzhou 221116,China
3. Anhui Magang Luohe Mining CO.,LTD,Hefei 231562,China
4. Institute of Mineral Processing and Automation,Sinosteel Maanshan Institute of Mining Research CO.,LTD,Ma’anshan 243000,China
5. Wuxi Voicon Technology CO.,LTD,Wuxi 214125,China

Received:2020-09-20 Online:2021-10-20 Published:2021-11-09
Contact: Xiao YUN

摘要/Abstract

摘要：

对煤矿井下各重要区域多个监控摄像头获取到的视频数据进行管控分析,定位和身份识别视频中的工作人员,对煤矿安全智能化生产具有重要意义。为解决矿井下光线暗淡、光照不均匀,现有常规行人重识别方法不能满足井下人员重识别的问题,提出一种联合双鉴别式生成对抗网络和全局特征的井下人员重识别方法。首先,采用双鉴别式生成对抗网络DD-GAN对井下暗光或光照不均的图像进行增强与复原,为后续重识别提供更具辨别力的图像基础;其次,在图像增强的基础上,设计一种基于全局特征描述的重识别网络以解决井下作业人员的身份识别问题,并使用随机擦除与k互近邻重排序方法来进一步提高重识别网络的鲁棒性和识别精度;最后,构建适用于井下特殊场景的Miner-CUMT数据集,解决了现有样本集场景单一的问题,同时也提高了该方法的泛化性。该方法在构建数据集Miner-CUMT进行了有效性验证,实现了煤矿巷道低照度场景下精准的人员重识别任务,为推进煤矿智能化安全生产的发展打下重要基础。

关键词: 智能化煤矿, 行人重识别, 卷积神经网络, 图像增强, 生成对抗网络

Abstract:

It is of great significance to the smarter safe production of coal to control and analyze the video data obtained by multiple surveillance cameras in each important area of the coal mine,and to locate and identify the workers in the video.However,due to the dim light and uneven illumination in the mine,the existing conventional method of person re-identification (Re-ID) cannot meet the requirements in the coal mine.In order to solve the above problems,this paper proposes a Re-ID method combining the Dual-Discriminator Generative Adversarial Network and global feature in coal mine.First,the Dual-Discriminator Generative Adversarial Network(DD-GAN) is designed to enhance and restore images with dim light or uneven illumination,providing a more discriminating image foundation for person re-identification.Second,the Global Feature Network for Re-ID is proposed in the coal mine to solve the miner’s identification problem,and the methods of Random erasing and Re-ranking of k-reciprocal nearest neighbors are used to improve further the robustness and accuracy of the Re-ID network.Finally,the Miner-CUMT dataset suitable for special downhole scenes is constructed,which solves the problems of the single scene of existing sample sets and improves the generalization of the method presented in this paper.The proposed method has achieved good results in the Miner-CUMT dataset and actual scene in the coal mine,which lays an important foundation for the development of intelligent and safe production in coal mines.

Key words: smart coal mines, person re-identification, convolutional neural network, image enhancement, generative adversarial networks

中图分类号:

TP391

孙彦景,魏力,张年龙,云霄,董锴文,葛敏,程小舟,侯晓峰. 联合DD-GAN和全局特征的井下人员重识别方法[J]. 西安电子科技大学学报, 2021, 48(5): 201-211.

SUN Yanjing,WEI Li,ZHANG Nianlong,YUN Xiao,DONG Kaiwen,GE Min,CHENG Xiaozhou,HOU Xiaofeng. Person re-identification method combining the DD-GAN and Global feature in a coal mine[J]. Journal of Xidian University, 2021, 48(5): 201-211.

图/表 10

图1

图2

图3

图4

图5

图6

表1

表2

图7

图8

参考文献 24

[1]	张立亚. 基于动目标特征提取的矿井目标监测[J]. 煤炭学报, 2017, 42(S2):603-610.
	ZHANG Liya. Mine Target Monitoring Based on Feature Extraction of Moving Target[J]. Journal of China Coal Society, 2017, 42(S2):603-610.
[2]	朱礼义. 矿井图像增强和井下人员检测算法的研究[D]. 江苏:中国矿业大学, 2019.
[3]	DALAL N, TRIGGS B. Histograms of Oriented Gradients for Human Detection[C]// Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE, 2005:886-893.
[4]	LOWE D G. Object Recognition From Local Scale-Invariant Features[C]// Proceedings of the Seventh IEEE International Conference on Computer Vision.Piscataway:IEEE, 1999:1150-1157.
[5]	LIAO S, HU Y, ZHU X, et al. Person Re-identification by Local Maximal Occurrence Representation and Metric Learning[C]// Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE, 2015:2197-2206.
[6]	MARTIN K, HIRZER M, WOHLHART P, et al. Large Scale Metric Learning from Equivalence Constraints[C]// Proceedings of the 2012 IEEE Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE, 2012:2288-2295.
[7]	闫林, 刘凯, 段玫妤. 一种用于点云分类的轻量级深度神经网络[J]. 西安电子科技大学学报, 2020, 47(2):46-53.
	YAN Lin, LIU Kai, DUAN Meiyu. Lightweight Deep Neural Network for Point Cloud Classification[J]. Journal of Xidian University, 2020, 47(2):46-53.
[8]	ZHONG Z, ZHENG L, ZHENG Z, et al. CamStyle:A Novel Data Augmentation Method for Person Re-Identification[J]. IEEE Transactions on Image Processing, 2019, 28(3):1176-1190. doi: 10.1109/TIP.2018.2874313
[9]	CHENX S, FUC M, ZHAO Y, et al. Salience-Guided Cascaded Suppression Network for Person Re-Identification[C]// Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE, 2020:3300-3310.
[10]	LU Y, WU Y, LIU B, et al. Cross-Modality Person Re-Identification with Shared-Specific Feature Transfer[C]// Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE, 2020:13379-13389.
[11]	王艳芬, 朱绪冉, 云霄, 等. 面向公共安全监控的多摄像机车辆重识别[J]. 西安电子科技大学学报, 2019, 46(4):190-196.
	WANG Yanfen, ZHU Xuran, YUN Xiao, et al. Vehicle Re-Identification by Multi-Cameras for Public Security Surveillance[J]. Journal of Xidian University, 2019, 46(4):190-196.
[12]	JIANG Y, GONG X, LIU D, et al. Enlightengan:Deep Light Enhancement without Paired Supervision[J/OL]. [2019-06-17]. https://arxiv.org/pdf/1906.06972v1.pdf.
[13]	WEI C, WANG W, YANG W, et al. Deep Retinex Decomposition for Low-Light Enhancement[C]// British Machine Vision Conference.Newcastle:BMVA, 2018: 155.
[14]	CHEN Y S, WANG Y, KAO M, et al. Deep Photo Enhancer:Unpaired Learning for Image Enhancement from Photographs with Gans[C]// Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE, 2018:6306-6314.
[15]	RONNEBERGER O, FISCHER P, BROX T. U-net:Convolutional Networks for Biomedical Image Segmentation[C]// International Conference on Medical Image Computing and Computer-Assisted Intervention.Berlin:Springer, 2015:234-241.
[16]	DENG J, DONG W, SOCHER R, et al. Imagenet:A Large-scale Hierarchical Image Database[C]// Proceedings of the 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE, 2009:248-255.
[17]	SIMONYAN K, ZISSERMAN A. Very Deep Convolutional Networks for Large-Scale Image Recognition[C]// Proceedings of the 2015 International Conference on Learning Representations.San Diego:Elsevier, 2015:1-13.
[18]	JOLICOEUR-MARTINEAUA. The Relativistic Discriminator:A Key Element Missing from Standard GAN[J/OL]. [2018-09-10]. https://arxiv.org/pdf/1807.00734v3.pdf.
[19]	MAOX D, LI Q, XIE H, et al. Least Squares Generative Adversarial Networks[C]// Proceedings of the IEEE International Conference on Computer Vision.Piscataway:IEEE, 2017:2794-2802.
[20]	HE K M, ZHANG X Y, REN S Q, et al. Deep Residual Learning for Image Recognition[C]// Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE, 2016:770-778.
[21]	ZHONG Z, ZHENG L, KANG G L, et al. Random Erasing Data Augmentation[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2017, 34(7):1-10.
[22]	ZHONG Z, ZHENG L, CAO D L, et al. Re-Ranking Person Re-Identification with k-Reciprocal Encoding[C]/ Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE, 2017:1318-1327.
[23]	ZHENG L, SHEN L Y, TIAN L, et al. Scalable Person Re-Identification:A Benchmark[C]// Proceedings of the 2015 IEEE International Conference on Computer Vision.Piscataway:IEEE, 2015:1116-1124.
[24]	WEI L H, ZHANG S L, GAO W, et al. Person Transfer GAN to Bridge Domain Gap for Person Re-Identification[C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Piscataway:IEEE, 2018:79-88.

方法	DD-GAN	mAP	Rank-1	Rank-5
GFN	×	58.7	87.2	94.4
GFN	√	64.2	89.7	95.4

方法	mAP	Rank-1	Rank-5	Rank-10
GFN+None	58.7	87.2	94.4	96.3
GFN+RE	62.9	89.8	96.3	97.5
GFN+Re-ranking	69.1	90.0	95.4	96.5
GFN+RE+Re-ranking(文中方法)	78.3	92.5	95.7	96.7

联合DD-GAN和全局特征的井下人员重识别方法

Person re-identification method combining the DD-GAN and Global feature in a coal mine

RichHTML

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献 24

相关文章 15

Metrics

本文评价

推荐阅读 10

[1]	李源,崔玉爽,王伟. 一种基于字词双通道网络的文本情感分析方法[J]. 西安电子科技大学学报, 2021, 48(6): 179-186.
[2]	于浩洋,尹良,李书芳,吕顺. 生成对抗网络小样本雷达调制信号识别算法[J]. 西安电子科技大学学报, 2021, 48(6): 96-104.
[3]	董如婵,焦李成,赵进,沈维燕. 一种深度融合机制的遥感图像目标检测技术[J]. 西安电子科技大学学报, 2021, 48(5): 128-138.
[4]	程德,郝毅,周靖宇,王楠楠,高新波. 利用混合双通路神经网络的跨模态行人重识别[J]. 西安电子科技大学学报, 2021, 48(5): 190-200.
[5]	王军军,孙岳,李颖. 一种生成对抗网络的遥感图像去云方法[J]. 西安电子科技大学学报, 2021, 48(5): 23-29.
[6]	杨静波,赵启军,吕泽均. 维度情感模型下的表情图像生成及应用[J]. 西安电子科技大学学报, 2021, 48(5): 30-37.
[7]	周鹏,杨军. 采用神经网络架构搜索的遥感影像分割方法[J]. 西安电子科技大学学报, 2021, 48(5): 47-57.
[8]	杨云航,闵连权. 采用空洞卷积的多尺度融合草图识别模型[J]. 西安电子科技大学学报, 2021, 48(5): 92-99.
[9]	陈昌川,王海宁,黄炼,黄涛,李连杰,黄向康,代少升. 一种基于局部表征的面部表情识别算法[J]. 西安电子科技大学学报, 2021, 48(5): 100-109.
[10]	宋建锋,苗启广,王崇晓,徐浩,杨瑾. 注意力机制的多尺度单目标跟踪算法[J]. 西安电子科技大学学报, 2021, 48(5): 110-116.
[11]	张宇浩,程培涛,张书豪,王秀美. 一种自适应权重学习的轻量超分辨率重建网络[J]. 西安电子科技大学学报, 2021, 48(5): 15-22.
[12]	回海生,张雪英,吴泽林,李凤莲. 一种主辅路径注意力补偿的脑卒中病灶分割方法[J]. 西安电子科技大学学报, 2021, 48(4): 200-208.
[13]	王平,江雨泽,赵光辉. 目标检测的多尺度定位提升算法[J]. 西安电子科技大学学报, 2021, 48(3): 85-90.
[14]	魏梓钰,杨曦,王楠楠,杨东,高新波. 互惠双向生成对抗网络用于跨模态行人重识别[J]. 西安电子科技大学学报, 2021, 48(2): 205-212.
[15]	郭泽坤,田隆,韩宁,王鹏辉,刘宏伟,陈渤. 采用CNN-SSD的雷达HRRP小样本目标识别方法[J]. 西安电子科技大学学报, 2021, 48(2): 7-14.