Cellular automata hardware for image processing tasks used in traffic control

The paper presents some practical solutions for diminishing the size of cellular automata hardware dedicated to image processing tasks to be used in a traffic control system. The specificity of visual processing tasks, especially the real-time processing and parallel processing of high dimension and low resolution images lend a big size circuit that can be hardly implemented even with VLSI technologies. Therefore, we present some simple techniques that operate on the size-time product and offer adequate solutions for real systems. These are mainly the common use of the local memories that allow the local functions performing and the function composition that lends to a new structuring of the rule space.