A real time fuzzy dynamic image understanding system on general roads

The purpose of this study is to realize a real time robotics vision system for, e.g. an ALV (autonomous land vehicle), in a cost effective way using fuzzy technology, that can be used in various circumstances (on general roads in Japan) such as from early morning till late evening and under the sunshine or in the rain. The presented system consists of an image processing part, knowledge base part, and a hardwired implementation part with VHDL. The image processing part is concerned with a dynamic image preprocessing based on mainly a conventional image processing technique. The knowledge base part employs a fuzzy frame knowledge base to express ambiguous information and natural language using fuzzy slot values. The hardwired implementation part designs hardware circuits using VHDL for high speed processing realization