A fuzzy clustering method for efficient 2-D object recognition

Advances in image processing architecture have provided speed and inspection capabilities previously not attainable for machine vision applications. Methods for improving the quality of visual data have stored great interest. Image acquisition has become increasingly important for the analysis of complex scenes where grey scale, colour, depth, texture and/or motion information is present. In this paper, we illustrate our design of a prototype system for the diagnosis of high tolerances in machined or cast components that copes with uncertainty and performs approximate reasoning since information used in decision-making or reasoning processes in advanced manufacturing metrology could be uncertain, imprecise, or incomplete. In the design, we employ fuzzy logic based on fuzzy sets theory. Inference procedures that incorporate uncertainty are becoming more important in rule-based expert-like systems. The design is extensible to handle a large number of rules, and the speed of inference is almost independent of the number of rules.