Premium
A system for matching moving vehicles across different camera sites in outdoor environments
Author(s) -
Lao Y.,
Leung C. H.,
Hung Y. S.
Publication year - 2005
Publication title -
international journal of imaging systems and technology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.359
H-Index - 47
eISSN - 1098-1098
pISSN - 0899-9457
DOI - 10.1002/ima.20054
Subject(s) - ransac , computer science , computer vision , artificial intelligence , matching (statistics) , process (computing) , shadow (psychology) , tracking (education) , tracking system , image (mathematics) , kalman filter , mathematics , psychology , pedagogy , statistics , psychotherapist , operating system
A method for identifying the same running vehicle from images captured at different camera sites is proposed. This is achieved through an image matching process. With cameras installed at the inlets and exits of highways and at crossroads in a road network, the system can be used for tracking the routes of individual vehicles and computing the times of traveling from one site to another. The system has the advantages of being low cost and not requiring sophisticated imaging hardware. Contributions include (i) a novel background removal method that employs the robust RANSAC algorithm for estimation, (ii) an elaborate shadow removal algorithm, and (iii) an effective license plate matching method applicable to poor quality images.© 2006 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 15, 211–219, 2005