Vehicle Ego-Localization Based on Streetscape Image Database Under Blind Area of Global Positioning System

doi:10.1007/s12204-018-2008-8

摘要/Abstract

摘要： Vehicle positioning is critical for inter-vehicle communication, navigation, vehicle monitoring and tracking. They are regarded as the core technology ensuring safety in everyday-driving. This paper proposes an enhanced vehicle ego-localization method based on streetscape image database. It is most useful in the global positioning system (GPS) blind area. Firstly, a database is built by collecting streetscape images, extracting dominant color feature and detecting speeded up robust feature (SURF) points. Secondly, an image that the vehicle shoots at one point is analyzed to find a matching image in the database by dynamic programming (DP) matching. According to the image similarity, several images with higher probabilities are selected to realize coarse positioning. Finally, different weights are set to the coordinates of the shooting location with the maximum similarity and its 8 neighborhoods according to the number of matching points, and then interpolating calculation is applied to complete accurate positioning. Experimental results show that the accuracy of this study is less than 1.5m and its running time is about 3.6 s. These are basically in line with the practical need. The described system has an advantage of low cost, high reliability and strong resistance to signal interference, so it has a better practical value as compared with visual odometry (VO) and radio frequency identification (RFID) based approach for vehicle positioning in the case of GPS not working.

关键词: vehicle ego-localization, global positioning system (GPS) blind area, dominant color, dynamic programming (DP) matching, weighted interpolation

Abstract: Vehicle positioning is critical for inter-vehicle communication, navigation, vehicle monitoring and tracking. They are regarded as the core technology ensuring safety in everyday-driving. This paper proposes an enhanced vehicle ego-localization method based on streetscape image database. It is most useful in the global positioning system (GPS) blind area. Firstly, a database is built by collecting streetscape images, extracting dominant color feature and detecting speeded up robust feature (SURF) points. Secondly, an image that the vehicle shoots at one point is analyzed to find a matching image in the database by dynamic programming (DP) matching. According to the image similarity, several images with higher probabilities are selected to realize coarse positioning. Finally, different weights are set to the coordinates of the shooting location with the maximum similarity and its 8 neighborhoods according to the number of matching points, and then interpolating calculation is applied to complete accurate positioning. Experimental results show that the accuracy of this study is less than 1.5m and its running time is about 3.6 s. These are basically in line with the practical need. The described system has an advantage of low cost, high reliability and strong resistance to signal interference, so it has a better practical value as compared with visual odometry (VO) and radio frequency identification (RFID) based approach for vehicle positioning in the case of GPS not working.

Key words: vehicle ego-localization, global positioning system (GPS) blind area, dominant color, dynamic programming (DP) matching, weighted interpolation

中图分类号:

TP 391.4

ZHOU Jingmei *(周经美), ZHAO Xiangmo (赵祥模), CHENG Xin (程鑫), XU Zhigang (徐志刚), ZHAO. Vehicle Ego-Localization Based on Streetscape Image Database Under Blind Area of Global Positioning System[J]. Journal of Shanghai Jiao Tong University (Science), 2019, 24(1): 122-129.

ZHOU Jingmei *(周经美), ZHAO Xiangmo (赵祥模), CHENG Xin (程鑫), XU Zhigang (徐志刚), ZHAO Huaixin (赵怀鑫). Vehicle Ego-Localization Based on Streetscape Image Database Under Blind Area of Global Positioning System[J]. Journal of Shanghai Jiao Tong University (Science), 2019, 24(1): 122-129.

参考文献 16

[1]	TARD′OS J D, NEIRA J, NEWMAN P M, et al. Robustmapping and localization in indoor environmentsusing sonar data [J]. International Journal of RoboticsResearch, 2002, 21(4): 311-330.
[2]	SHIBUHISA N, SATO J, TAKAHASHI T, et al.Accurate vehicle localization using DTW betweenrange data map and laser scanner data sequences[C]//Proceedings of the 2007 IEEE Intelligent VehiclesSymposium. Istanbul, Turkey: IEEE, 2007: 975-980.
[3]	AGRAWAL M, KONOLIGE K. Real-time localizationin outdoor environments using stereo vision and inexpensiveGPS [C]//Proceedings of the 18th InternationalConference on Pattern Recognition. Hong Kong,China: IEEE, 2006: 1063-1068.
[4]	MATSUMOTO Y, SAKAI K, INABA M, et al. Viewbasedapproach to robot navigation [J]. Journal of theRobotics Society of Japan, 2000, 20(5): 1702-1708.
[5]	ZHANG E Z, KUANG Y J, JIANG W L, et al.Active RFID positioning of vehicles in road traffic[C]//Proceedings of the 11th International Symposiumon Communications and Information Technologies.Hangzhou, China: IEEE, 2011: 222-227.
[6]	HUANG J F, LI X, SUN Y H, et al. A highly-reliablecombined positioning method for vehicle in urban complexenvironments [C]//Proceedings of the InternationalConference on Vehicular Electronics and Safety.Dongguan, China: IEEE, 2013: 153-158.
[7]	BOUCHER C, NOYER J C. A hybrid particle approachfor GNSS applications with partial GPS outages[J]. IEEE Transactions on Instrumentation andMeasurement, 2010, 59(3): 498-505.
[8]	NIST′ER D, NARODITSKY O, BERGEN J. Visualodometry for ground vehicle applications [J]. Journalof Field Robotics, 2010, 23(1): 3-20.
[9]	FARENZENA M, BARTOLI A, MEZOUAR Y.Efficient camera smoothing in sequential structurefrom-motion using approximate cross-validation[C]//European Conference on Computer Vision.Marseille, France: Springer-Verlag, 2008: 196-209.
[10]	ZHOU J M, ZHAO X M, XU Z G, et al. Algorithmof vehicle trajectory extraction using fusion of Hu momentsand grayscale feature [J]. Information TechnologyJournal, 2014, 13(11): 1770-1778.
[11]	SATO J, TAKAHASHI T, IDE I, et al. Change detectionin streetscapes from GPS coordinated omnidirectionalimage sequences [C]//Proceedings of the18th International Conference on Pattern Recognition.Hong Kong, China: IEEE, 2006: 935-938.
[12]	UCHIYAMA H, TAKAHASHI T, IDE I, et al.Frame registration of in-vehicle normal camera withomni-directional camera for self-position estimation[C]//International Conference on Innovative ComputingInformation and Control. Dalian, China: IEEE,2008: 7.
[13]	UCHIYAMA H, DEGUCHI D, TAKAHASHI T, etal. Ego-localization using streetscape image sequencesfrom in-vehicle cameras [C]//Intelligent Vehicles Symposium.Xi’an, China: IEEE, 2009: 185-190.
[14]	ZHOU J M, ZHAO X M, WANG Z, et al. An automaticcalibration system of vehicle motion error underGPS blind area [C]//International Conference on IntelligentTransportation Systems. Las Palmas, Spain:IEEE, 2015: 238-243.
[15]	SUN J. Research on content-based image retrieval [D].Xi’an, China: School of Computer Science, XidianUniversity, 2005 (in Chinese).
[16]	LIU Y, EMOTO H, FUJII T, et al. A methodfor content-based similarity retrieval of imagesusing two dimensional DP matching algorithm[C]//International Conference on Image Analysis andProcessing. Palermo, Italy: IEEE, 2001: 236-241.