Traffic measurement with a roadside vision system-individual tracking of overlapped vehicles
Takahiro Ikeda, S Ohnaka, Masanori Mizoguchi · 1996
Traffic measurement is an essential technique in moving toward the next stage of vehicular transportation, and one which needs to be more advanced in terms of both safety and efficiency. For this technique, image sensors are more advantageous than conventional sensors because of their portability and the wide variety of information they provide. Until now, however, vehicles overlapping each other within the image plane have caused problems with measurement accuracy. Therefore, a traffic measurement algorithm which features individual vehicle tracking using a rectangular model has been developed. The algorithm is used in real-time on the real-time vision system in order to utilize traffic volume to evaluate its performance.