Represent and Acquire Knowledge for the Development of Autonomous Vision System

T h e visual perceptional ability of computer systems t o understand the environment i s desirable in engineering design and m a n u facturing where automat ion i s anticipated. I n order t o empower a n engineering syst e m wi th the heuristic capabilities, adequate knowledge representation techniques m u s t be employed t o resolve ambiguity in representa t ion and uncertainty in decision making. Discussed in this paper is how t o represent environmental knowledge as object models so that they can be utilized in spatial reasoning, and how t o check knowledge completeness so that autonomous vision systems can be developed wi th self-regulated knowledge acquiring mechanism.