On the Symbolic Interpretation of Traac Scenes

We present a high-level traac scene interpretation system based on an Object-Oriented concurrent logic programming approach. The system is written in Parlog++ and is based on the idea of a network of concurrent objects , each dealing with a class of structures in a picture , and exchanging messages in both a top-down and bottom-up direction. The system is tested on real images obtained from aerial video of a city intersection. These are preprocessed to nd the positions and ori-entations of the cars. The output from the high level system is in the form of traac law interpretations of the car conngurations.