Skip to Main Content
A great number of roadside-cameras, which are disposed in large scale road net, are the critical component among Intelligent Transport Systems. These roadside-cameras also can be used to track suspicious vehicles by license plate recognition systems. Suspicious vehicle may perform a wide movement on road net. However, most video recognition systems focus on one or several cameras to perform the recognition operation. It results in that a wide vehicle tracking cannot be performed. Therefore, we propose a mechanism to make an overall plan to perform vehicle tracking on large scale roadside camera network according the recognition information acquired from each camera, which includes (i) a logical roadside-camera network topology (LRNT), and (ii) Suspicious vehicle tracking algorithm using LRNT.