The paper describes a system for detecting vehicles in urban traffic scenes in daytime and at night by means of image analysis and rule-based reasoning. The strength of the proposed approach is its formal separation between the low-level image processing modules (detecting moving vehicles under day and night light) and the high-level module, which provides a single framework for tracking vehicles in the scene. The image processing modules perform spatio-temporal analysis on moving templates in daytime images, and morphological analysis of headlight pairs in night images. The high-level module is designed as a forward chained production rule system, working on symbolic data, i.e. vehicles and their attributes, and exploiting a set of heuristic roles tuned to urban traffic conditions. The synergy between the artificial intelligence techniques of the high level and low-level image analysis techniques provides the system with flexibility and robustness.