Low-Cost Floating Car Observer Implementation by a Video Camera Based Method

An overall knowledge of the inner-city traffic state is a prerequisite for a functioning traffic management both in private and public transport. The acquisition of traffic data is performed mostly by stationary systems like induction loops, infra-red and video detectors, however these local, stationary measurements do not allow the clear determination of traffic states. This paper offers a new approach to acquire data of real traffic situations using a Floating Car Observer (FCO). The FCO captures data from the traffic situation of oncoming vehicles, such as positions and lengths of traffic hold-ups. The captured data is then used to enhance travel time prognosis for public transport management systems. A simulation framework was used to evaluate different traffic observing devices (laser, ultrasonic, camera) in order to set up a traffic monitoring FCO prototype. The evaluation derives from detection rates, but also from real time capability of the computing process and economical aspects of the devices. The system chosen to set up a FCO prototype comprised of a video processing and an infra-red emitting module. The described technical approach is characterized by low costs compared with conventional video based detection systems. This is achieved by the low effort for number plate recognition and the image processing caused by illuminating the vehicles’ number plate. Hence, the usage of inexpensive microcontrollers is possible. Before the FCO module can be integrated into a public transport vehicle, broad field tests are necessary.