Extension of Cellular Automata for Dynamic Vehicle Tracking

Abstract

Detecting and tracking moving vehicles in realtime traffic scenes are an emerging research area for intelligent transportation systems. This paper presents the computational model of Fuzzy Cellular Automata (FCA) to address the susceptible to environmental changes' problem associated with the background subtraction methods for dynamic vehicle tracking. The suggested model deploys FCA that is configured with rules supporting least sensitive fuzzy BExclusive orf unction as next state logic to handle degrees of uncertainty in rule similarly operations. Indeed, at each step, the update of background in frame difference approaches is determined according to the number of active cells and fuzzy mapping function; so moving vehicles that their gray level is entirely similar to the background gray level are simply detected. Furthermore, an occlusion handling technique based on visual measurement is employed to detect the classes of the vehicle occlusions and segment the vehicle from each occlusive class. The experimental results show that the suggested method is more robust, accurate and powerful than traditional methods, for real-time vehicle detection and tracking.