Infrared-based system for vehicle classification

This paper presents a low-cost infrared-based system developed for vehicle classification in urban areas. The system contains an equipment that is placed above the roadway and emits down-looking pulsed infrared beams through infrared LEDs (Light-Emitting Diodes). Vehicle profile is obtained detecting and analyzing the infrared light reflected by the vehicle passing directly below the equipment. Profiles of different vehicles were obtained to build a template databank. Vehicle classification is achieved by comparing a vehicle profile with all of the templates. Dynamic Time Warping (DTW) technique is used for carrying out the comparisons. Results obtained are presented and discussed.