REAL TIME VEHICLE COUNTRY OF ORIGIN CLASSIFICATION BASED ON COMPUTER VISION

izvorni znanstveni rad

izvorni znanstveni rad

REAL TIME VEHICLE COUNTRY OF ORIGIN CLASSIFICATION BASED ON COMPUTER VISION

Vrsta prilog sa skupa (u zborniku)
Tip izvorni znanstveni rad
Godina 2014
Nadređena publikacija Proceedings of 22st International Symposium on Electronics in Transport – ISEP 2014, ITS for Seamless and Energy Smart Transport
Status objavljeno

Sažetak

In order to efficiently monitor and control road traffic, relevant traffic parameters values are needed in real time. Traffic parameters include traffic flow speed profile, vehicle type classification, vehicle gap, vehicle length, origin-destination matrix, etc. Using these parameters advanced traffic control approaches from the domain of intelligent transportation systems (ITS) can be applied. Additionally, traffic planners can incorporate more accurate traffic forecasts when planning new roads. Also ITS related services can be provided to users, especially precise real time traffic information. Surveillance video cameras combined with advanced computer vision (CV) algorithms are nowadays becoming a tool used more frequently to obtain mentioned traffic parameters. They are used as standalone sensors or in combination with other sensors like radars, inductive loops, etc. In this article architecture of a computer vision based road traffic surveillance system is proposed. Application of such a system for vehicle counting and vehicle country of origin classification using license plate recognition is described. Implemented system is tested using real world road traffic video footage from a Croatian highway near the city of Zagreb. Proposed system quality is analysed using vehicle detection processing time and accuracy.

Ključne riječi

Road traffic monitoring; Vision sensor; Vehicle detection; License plate recognition