Low-cost, vision-based techniques for the quantitative analysis of road traffic data in real-time

• Autorzy: Sgouros N. , Kabolis P.
• Języki publikacji: EN

Abstrakty

EN

Quantitative analysis of traffic data is concerned with the extraction of traffic parameters such as vehicle counts, lane occupancy, speed measurement etc. This article describes a set of low-cost, vision-based techniques for the quantitative analysis of road traffic data in real-time. Analysis consists of two phases; vehicle detection and extraction of traffic parameters. Vehicle detection applies a number of image processing techniques such as background subtraction, binary thresh-olding, morphological processing and blob coloring for the detection of the vehicles passing through a user-specified region in a road. The vehicle detection phase ends with the estimation of a set of characteristic features for each vehicle such as its area, its center of mass and its perimeter. During the next phase these features are fed to a novel tracking algorithm that uniquely identifies each vehicle in time, estimates its speed and determines the traffic flow in the region of interest. All these techniques have been implemented in a traffic monitoring system. The system runs on top of a PC with an AMD 450 K6-2 processor. It has been tested in real traffic conditions and it has achieved a 93rate in real-time speed detection.

Słowa kluczowe

EN

image processing; quantitative analysis; road traffic data in real-time; road traffic system control; traffic control; traffic flow; traffic monitoring system; vehicle detection; vision-based techniques

PL

natężenie ruchu drogowego; przetwarzanie obrazów; system monitorowania ruchu drogowego

• Wydawca: Instytut Telekomunikacji i Informatyki Uniwersytetu Technologiczno-Przyrodniczego w Bydgoszczy
• Czasopismo: Image Processing & Communications
• Rocznik: 2000
• Tom: Vol. 6, no 1-2
• Strony: 13--22
• Opis fizyczny: Bibliogr. 9 poz.,Rys., wykr.,

Twórcy

autor

Sgouros N.

autor

Kabolis P.