A Fuzzy Approach for Vision-Based Object Perception in Robot Telecontrol

For human being the object perception process is often conducted in a non-crisp manner, in which vagueness plays one of the pivot roles. In this paper, we propose a fuzzy logic based approach for vision-based object perception in enhanced reality, supported by a shape repository. By means of defining the fuzzy matching and reasoning rules, the proposed approach is capable of supporting the fuzzy decision making process in finding the matching shape model in the repository. Experiment results confirm that the proposed method is both valid and effective in the designated application.