Evaluation through field trials of a computer vision based automatic incident detection system

Summary form only given. Reliable Automatic Incident Detection (AID) Systems are elements in integrated urban and inter-urban traffic control management. A new approach using image processing especially oriented to AID was developed in the frame of the CEC DRIVE Programme. The AID system is based on the analysis of video images from closed circuit television (CCT) cameras situated in strategic sites along side the road, carried out in processing units (LSM) which produce meaningful real time spatial referenced data. These data are further processed at central level by the central process system (CPS) to produce their spatial and temporal tracking, to detect and follow up incidents and congestions throughout the road network. Implementation of both, LSM and CPS was accomplished by evaluating the whole system at many different operational conditions. The set up and first results of the field trial currently on progress in Spain, within the project ARTIS (Advanced RTI in Spain) are discussed.