车辆再识别技术综述

doi:10.11959/j.issn.2096-6652.202002

摘要/Abstract

摘要：

车辆再识别是指给定一张车辆图像，找出其他摄像头拍摄的同一车辆，可将车辆再识别问题看作图像检索的子问题。在真实交通监控系统中，车辆再识别可以起到对目标车辆进行定位、监管、刑侦的作用。随着深度神经网络的兴起和大型数据集的提出，提升车辆再识别的准确度成为近年来计算机视觉和多媒体领域的研究热点。从不同角度对车辆再识别方法进行了分类，并从特征提取、方法设计和性能表现等方面对车辆再识别技术进行了概述、比较和分析，对车辆再识别技术面临的挑战及发展趋势进行了预测。

关键词: 车辆再识别, 深度学习, 特征表达, 度量学习

Abstract:

Given a vehicle image,vehicle re-identification aims to find the same vehicle caught by other cameras,it can be regarded as a sub-problem of image retrieval.In the real traffic surveillance system,vehicle re-identification can play a role in locating,supervising and criminal investigation of target vehicles.With the rise of deep neural networks and the release of large-scale dataset,improving the accuracy and efficiency of vehicle re-identification has become a research focus in the field of computer vision and multimedia in recent years.The vehicle re-identification methods from different perspectives were classified,and the overview,comparison and analysis in terms of feature extraction,design and performance were given in detail,and the challenges and future trends of vehicle re-identification were predicted.

Key words: vehicle re-identification, deep learning, feature representation, metric learning

中图分类号:

TP520

刘凯,李浥东,林伟鹏. 车辆再识别技术综述[J]. 智能科学与技术学报, 2020, 2(1): 10-25.

Kai LIU,Yidong LI,Weipeng LIN. A survey on vehicle re-identification[J]. Chinese Journal of Intelligent Science and Technology, 2020, 2(1): 10-25.

图/表 7

图1

图2

表1

图3

图4

表2

表3

参考文献 71

[1]	FERIS R S , SIDDIQUIE B , PETTERSON J ,et al. Large-scale vehicle detection,indexing,and search in urban surveillance videos[J]. IEEE Transactions on Multimedia, 2011,14(1): 28-42.
[2]	YANG L , LUO P , CHANGE LOY C ,et al. A large-scale car dataset for fine-grained categorization and verification[C]// The IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2015: 3973-3981.
[3]	SUN C C , ARR G S , RAMACHANDRAN R P . Vehicle reidentifica tion using multidetector fusion[J]. IEEE Transactions on Intelligent Transportation Systems, 2004,5(3): 155-164.
[4]	LI N , JAIN J J , BUSSO C . Modeling of driver behavior in real world scenarios using multiple noninvasive sensors[J]. IEEE Transactions on Multimedia, 2013,15(5): 1213-1225.
[5]	MEI T , RUI Y , LI S . Multimedia search reranking:a literature survey[J]. ACM Computing Surveys (CSUR), 2014,46(3): 1-38.
[6]	LIU X , LIU W , MA H . Large-scale vehicle re-identification in urban surveillance videos[C]// The 2016 IEEE International Conference on Multimedia and Expo (ICME). Piscataway:IEEE Press, 2016: 1-6.
[7]	KWONG K , KAVALER R , RAJAGOPAL R . Arterial travel time estimation based on vehicle re-identification using wireless magnetic sensors[J]. Transportation Research Part C:Emerging Technologies, 2009,17(6): 586-606.
[8]	LIU H , TIAN Y , YANG Y . Deep relative distance learning:tell the difference between similar vehicles[C]// The IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2016: 2167-2175.
[9]	KHAN S D , ULLAH H . A survey of advances in vision-based vehicle re-identification[J]. Computer Vision and Image Understanding, 2019,182: 50-63.
[10]	WANG Z , TANG L , LIU X . Orientation invariant feature embedding and spatial temporal regularization for vehicle re-identification[C]// The IEEE International Conference on Computer Vision. Piscataway:IEEE Press, 2017: 379-387.
[11]	KLEIN L A , KELLEY M R , MILLS M K . Evaluation of overhead and in-ground vehicle detector technologies for traffic flow measurement[J]. Journal of Testing and Evaluation, 1997,25(2): 205-214.
[12]	DAVIES P . Vehicle detection and classification[J]. Information Technology Applications in Transport, 1986: 11-40.
[13]	SANCHEZ R O , FLORES C , HOROWITZ R ,et al. Vehicle re-identification using wireless magnetic sensors:algorithm revision,modifications and performance analysis[C]// 2011 IEEE International Conference on Vehicular Electronics and Safety. Piscataway:IEEE Press, 2011: 226-231.
[14]	CHARBONNIER S , PITTON A C , VASSILEV A . Vehicle re-identification with a single magnetic sensor[C]// 2012 IEEE International Instrumentation and Measurement Technology Conference. Piscataway:IEEE Press, 2012: 380-385.
[15]	KWONG K , KAVALER R , RAJAGOPAL R ,et al. Arterial travel time estimation based on vehicle re-identification using wireless magnetic sensors[J]. Transportation Research Part C:Emerging Technologies, 2009,17(6): 586-606.
[16]	JENG S T C , CHU L . Vehicle re-identification with the inductive loop signature technology[J]. Journal of the Eastern Asia Society for Transportation Studies, 2013(10): 1896-1915.
[17]	NDOYE M , TOTTEN V , CARTER B ,et al. Vehicle detector signature processing and vehicle reidentification for travel time estimation[C]// Transportation Research Board Meeting.[S.l.:s.n. ], 2008.
[18]	NDOYE M , TOTTEN V , KROGMEIER J V ,et al. A signal processing framework for vehicle re-identification and travel time estimation[C]// The 2009 12th International IEEE Conference on Intelligent Transportation Systems. Piscataway:IEEE Press, 2009: 1-6.
[19]	ALI S S M , GEORGE B , VANAJAKSHI L . Multiple inductive loop detectors for intelligent transportation systems applications:Ramp metering,vehicle re-identification and lane change monitoring systems[C]// 2013 IEEE Symposium on Computers ＆ Informatics (ISCI). Piscataway:IEEE Press, 2013: 176-180.
[20]	PRINSLOO J , MALEKIAN R . Accurate vehicle location system using RFID,an internet of things approach[J]. Sensors, 2016,16(6): 825-848.
[21]	MAZLOUMI E , CURRIE G , ROSE G . Using GPS data to gain insight into public transport travel time variability[J]. Journal of Transportation Engineering, 2009,136(7): 623-631.
[22]	ZHANG Z , TAN T , HUANG K . Three-dimensional deformable-model-based localization and recognition of road vehicles[J]. IEEE Transactions on Image Processing, 2012,21(1): 1-13.
[23]	WOESLER R , . Fast extraction of traffic parameters and re- identification of vehicles from video data[C]// The 2003 IEEE International Conference on Intelligent Transportation Systems. Piscataway:IEEE Press, 2003: 774-778.
[24]	GUO Y , RAO C , SAMARASEKERA S ,et al. Matching vehicles under large pose transformations using approximate 3D models and piecewise MRF model[C]// The IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2008: 1-8.
[25]	ZAPLETAL D , HEROUT A . Vehicle re-identification for automatic video traffic surveillance[C]// The IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2016: 25-31.
[26]	DUBSKá M , HEROUT A , SOCHOR J . Automatic camera calibration for traffic understanding[C]// British Machine Vision Conference (BMVC). Guildford:BMVA Press, 20144(6):8.
[27]	DUBSKá M , HEROUT A , JURáNEK R , . Fully automatic roadside camera calibration for traffic surveillance[J]. IEEE Transactions on Intelligent Transportation Systems, 2015,16(3): 1162-1171.
[28]	FERENCZ A , LEARNED-MILLER E G , MALIK J . Building a classification cascade for visual identification from one example[C]// The IEEE International Conference on Computer Vision. Piscataway:IEEE Press, 2005:1: 286-293.
[29]	LIU X , LIU W , MEI T ,et al. A deep learning-based approach to progressive vehicle re-identification for urban surveillance[C]// The European Conference on Computer Vision. Heidelberg:Springer, 2016: 869-884.
[30]	ZHENG Q , LIANG C , FANG W ,et al. Car re-identification from large scale images using semantic attributes[C]// The IEEE 17th International Workshop on Multimedia Signal Processing. Piscataway:IEEE Press, 2015: 1-5.
[31]	LOWE D G . Distinctive image features from scale-invariant keypoints[J]. International Journal of Computer Vision, 2004,60(2): 91-110.
[32]	VAN DE WEIJER J , SCHMID C , VERBEEK J . Learning color names for real-world applications[J]. IEEE Transactions on Image Processing, 2009,18(7): 1512-1523.
[33]	LIAO S , HU Y , ZHU X . Person re-identification by local maximal occurrence representation and metric learning[C]// The IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2015: 2197-2206.
[34]	FU Y , WEI Y , WANG G ,et al. Self-similarity grouping:a simple unsupervised cross domain adaptation approach for person re- identification[C]// The IEEE International Conference on Computer Vision. Piscataway:IEEE Press, 2019: 6112-6121.
[35]	DENG W , ZHENG L , YE Q ,et al. Image-image domain adaptation with preserved self-similarity and domain-dissimilarity for person re-identification[C]// The IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2018: 994-1003.
[36]	PENG P , XIANG T , WANG Y ,et al. Unsupervised cross-dataset transfer learning for person re-identification[C]// The IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2016: 1306-1315.
[37]	WANG J , ZHU X , GONG S ,et al. Transferable joint attribute-identity deep learning for unsupervised person re-identification[C]// The IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2018: 2275-2284.
[38]	WEI L , ZHANG S , GAO W ,et al. Person transfer GAN to bridge domain gap for person re-identification[C]// The IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2018: 79-88.
[39]	LIU X , LIU W , MEI T . PROVID:progressive and multimodal vehicle reidentification for large-scale urban surveillance[J]. IEEE Transactions on Multimedia, 2018,20(3): 645-658.
[40]	GUO Y F , WU L , LU H . Null Foley-Sammon transform[J]. Pattern Recognition, 2006,39(11): 2248-2251.
[41]	ZHANG L , XIANG T , GONG S . Learning a discriminative null space for person re-identification[C]// The IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2016: 1239-1248.
[42]	YAN K , TIAN Y , WANG Y . Exploiting multi-grain ranking constraints for precisely searching visually-similar vehicles[C]// The 2017 IEEE International Conference on Computer Vision. Piscataway:IEEE Press, 2017: 562-570.
[43]	HE B , LI J , ZHAO Y ,et al. Part-regularized near-duplicate vehicle re-identification[C]// The IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2019: 3997-4005.
[44]	REDMON J , DIVVALA S , GIRSHICK R .et al. You only look once:unified,real-time object detection[C]// The IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2016: 779-788.
[45]	JOSEPH REDMON AND ALI FARHADI , . YOLO9000:better,faster,stronger[C]// The IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2017: 6517-6525.
[46]	TANG Z , NAPHADE M , BIRCHFIELD S ,et al. PAMTRI:pose-aware multi-task learning for vehicle re-identification using highly randomized synthetic data[C]// The IEEE International Conference on Computer Vision. Piscataway:IEEE Press, 2019: 211-220.
[47]	SUN K , XIAO B , LIU D ,et al. Deep high-resolution representation learning for human pose estimation[C]// The IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2019: 5693-5703.
[48]	HUANG G , LIU Z , VAN DER MAATEN L ,et al. Densely connected convolutional networks[C]// The IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2017: 4700-4708.
[49]	ZHANG Y , LIU D , ZHA Z J . Improving triplet-wise training of convolutional neural network for vehicle re-identification[C]// The 2017 IEEE International Conference on Multimedia and Expo (ICME). Piscataway:IEEE Press, 2017: 1386-1391.
[50]	GUO H , ZHAO C , LIU Z . Learning coarse-to-fine structured feature embedding for vehicle re-identification[C]// The 3nd AAAI Conference on Artificial Intelligence. Piscataway:IEEE Press, 2018: 6853-6860.
[51]	BAI Y , LOU Y , GAO F . Group-sensitive triplet embedding for vehicle re-identification[J]. IEEE Transactions on Multimedia, 2018,20(9): 2385-2399.
[52]	SUN Y , ZHENG L , YANG Y ,et al. Beyond part models:person retrieval with refined part pooling (and a strong convolutional baseline)[C]// The European Conference on Computer Vision (ECCV) . Heidelberg:Springer, 2018: 480-496.
[53]	WANG G , YUAN Y , CHEN X ,et al. Learning discriminative features with multiple granularities for person re-identification[C]// The 2018 ACM Multimedia Conference on Multimedia Conference. New York:ACM Press, 2018: 274-282.
[54]	CHEN H , LAGADEC B , BREMOND F . Partition and reunion:a two-branch neural network for vehicle re-identification[C]// The IEEE Conference on Computer Vision and Pattern Recognition Workshops. Piscataway:IEEE Press, 2019: 184-192.
[55]	NEWELL A , YANG K , DENG J . Stacked hourglass networks for human pose estimation[C]// The European Conference on Computer Vision. Heidelberg:Springer, 2016: 483-499.
[56]	ZHOU Y , SHAO L , DHABI A . Viewpoint-aware attentive multi-view inference for vehicle re-identification[C]// The IEEE International Conference on Computer Vision. Piscataway:IEEE Press, 2018: 6489-6498.
[57]	CHU R , SUN Y , LI Y ,et al. Vehicle re-identification with viewpoint-aware metric learning[C]// The IEEE International Conference on Computer Vision. Piscataway:IEEE Press, 2019: 8282-8291.
[58]	LIN W , LI Y , YANG X ,et al. Multi-view learning for vehicle re-identification[C]// 2019 IEEE International Conference on Multimedia and Expo (ICME). Piscataway:IEEE Press, 2019: 832-837.
[59]	SHEN Y , XIAO T , LI H . Learning deep neural networks for vehicle re-id with visual-spatio-temporal path proposals[C]// The 2017 IEEE International Conference on Computer Vision. Piscataway:IEEE Press, 2017: 1918-1927.
[60]	DENG J , DONG W , SOCHER R . ImageNet:a large-scale hierarchical image database[C]// The 2009 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2009: 248-255.
[61]	YANG L , LUO P , CHANGE LOY C . A large-scale car dataset for fine-grained categorization and verification[C]// The IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2015: 3973-3981.
[62]	KANACI A , ZHU X , GONG S . Vehicle re-identification in context[C]// The German Conference on Pattern Recognition. Heidelberg:Springer, 2018: 377-390.
[63]	LOU Y , BAI Y , LIU J ,et al. Veri-wild:a large dataset and a new method for vehicle re-identification in the wild[C]// The IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2019: 3235-3243.
[64]	ZHOU Y , SHAO L . Cross-view GAN based vehicle generation for re-identification[C]// British Machine Vision Conference (BMVC). Heidelberg:Springer, 2017.
[65]	LIU X , ZHANG S , HUANG Q ,et al. Ram:a region-aware deep model for vehicle re-identification[C]// The 2018 IEEE International Conference on Multimedia and Expo (ICME). Piscataway:IEEE Press, 2018: 1-6.
[66]	ANTONIO M P , PALAZZI A , BERGAMINI L ,et al. Unsupervised vehicle re-identification using triplet networks[C]// The IEEE Conference on Computer Vision and Pattern Recognition Workshops. Piscataway:IEEE Press, 2018: 166-171.
[67]	BASHIR R M S , SHAHZAD M , FRAZ M M . DUPL-VR:deep unsupervised progressive learning for vehicle re-identification[C]// The International Symposium on Visual Computing. Heidelberg:Springer, 2018: 286-295.
[68]	BASHIR R M S , SHAHZAD M , FRAZ M M . Vr-proud:vehicle re-identification using progressive unsupervised deep architecture[J]. Pattern Recognition, 2019,90: 52-65.
[69]	王茜, 陈一民, 丁友东 . 复杂环境中基于视觉词袋模型的车辆再识别算法[J]. 计算机应用, 2018,38(5): 83-87.
	WANG Q , CHEN Y M , DING Y D . Vehicle re-identification algorithm based on bag of visual words in complicated environments[J]. Com-puter Application, 2018,38(5): 83-87.
[70]	李熙莹, 周智豪, 邱铭凯 . 基于部件融合特征的车辆重识别算法[J]. 计算机工程, 2019(6): 12-20.
	LI X Y , ZHOU Z H , QIU M K . Vehicle re-identification algorithm based on component fusion feature[J]. Computer Engineering, 2019(6): 12-20.
[71]	王盼盼, 李玉惠 . 基于特征融合和 L-M 算法的车辆重识别方法[J]. 电子科技, 2018,31(4): 12-15.
	WANG P P , LI Y H . Vehicle re-identification algorithm based on component fusion feature[J]. Electronic Science and Technology, 2018,31(4): 12-15.

分类依据	方法类别	参考文献	优点	缺点
实现方式	基于传感器的方法	参考文献[11-21]等	不需要训练学习	复现难度高；识别率低；需要大量硬件
	基于人工设计特征的方法	参考文献[22-33]等	不依赖特殊硬件；可解释性强	时间复杂度高；受光照变化、视角变化和遮挡的影响较大
	基于深度学习的方法	参考文献[34-68]等	识别率高；实时性强；可以解决光照变化、视角变化和遮挡问题	训练学习时间长；特征抽象，可解释性差
解决车辆再识别问题的着手点	利用多维度信息的方法	参考文献[39-48]等	对车辆特殊外观敏感	容易误判辨识性区域；易受视角、光照变化影响
	基于度量学习的方法	参考文献[49-54]等	识别率高；对难样本的识别效果好	训练学习时间长
	解决视角变化引起的外观偏差	参考文献[55-58]等	对车辆的视角变化有鲁棒性；识别率高	训练学习时间长；系统实时性难以得到保证
	利用时序地理信息的方法	参考文献[59]等	对难样本的识别效果好；方法新颖	需要车辆时序和位置标签；对数据集要求高
是否需要车辆身份标签	有监督学习方法	参考文献[34-65]等	识别率高；符合发展潮流；在不同数据集下的泛化能力强	训练学习时间长；需要大型数据集支持
	无监督学习方法	参考文献[66-68]等	无需人工标注；符合实际需求；拓展性强	识别率较低；时间复杂度高

方法	参考文献序号	评价指标
方法	参考文献序号	mAP	Rank-1	Rank-5
FACT	[6]	19.9	59.6	75.2
XVGAN	[64]	24.7	60.2	77.0
FACT+SNN+STR	[6]	27.7	61.4	78.7
OIFE	[10]	51.4	68.3	89.7
PROVID	[39]	53.4	81.5	95.1
FDA-Net	[63]	55.4	84.2	92.4
VGG+C+T+S	[49]	57.4	86.5	92.8
S-CNN+PathLSTM	[59]	58.2	83.5	90.0
GS-TRE	[51]	59.4	96.2	98.9
VAMI	[56]	61.3	85.9	91.8
AAVER	—	61.1	90.1	94.7
RAM	[65]	61.5	88.6	94.0
VANet	[57]	66.3	89.7	95.9
PNVR	[43]	74.3	94.3	98.7
MRL	[58]	78.5	94.3	99.0
PGAN	—	79.3	96.5	98.3
PRN	[54]	85.8	97.1	99.4

方法	参考文献序号			评价指标
		小尺寸		中尺寸		大尺寸
		Rank-1	Rank-5	Rank-1	Rank-5	Rank-1	Rank-5
DRDL	[8]	49.0%	73.5%	42.8%	66.8%	38.2%	61.6%
C2F-Rank	[50]	61.1%	81.7%	56.2%	76.2%	51.4%	72.2%
VAMI	[56]	63.1%	83.2%	52.8%	75.1%	47.3%	70.2%
FDA-Net	[63]	64.0%	82.8%	57.8%	78.3%	49.4%	70.4%
VGG+C+T+S	[49]	69.9%	87.3%	66.2%	82.3%	63.2%	79.4%
OIFE	[10]	—	—	—	—	67.0%	82.9%
GS-TRE	[51]	75.9%	84.2%	74.8%	83.6%	74.0%	82.7%
AAVER	—	74.6%	93.8%	68.6%	89.9%	63.5%	85.6%
PNVR	[43]	78.4%	92.3%	75.0%	88.3%	74.2%	86.4%
PGAN	—	—	—	—	—	77.8%	92.1%
PRN	[54]	78.9%	94.8%	74.9%	92.0%	71.5%	88.4%
VANet	[57]	83.2%	95.9%	81.1%	94.7%	77.2%	86.7%
MRL	[58]	84.8%	96.9%	80.9%	94.1%	78.4%	92.1%