Visual motion detection with hand-eye manipulator using statistic moments classifiers and fuzzy logic approach-study, application and comparation

In this paper, we address the visual servo control of a robot manipulator under a camera-in-hand configuration. The goal is to recognise several types of different objects in a moving conveyer belt using a camera attached to a robot arm. To accomplish the object recognition three different methods are proposed: statistic based on Hu moments, statistic based on affine moments and a fuzzy based classifier. Validity and effectiveness of the proposed methods are evaluated by real-time experiments using a CCD camera attached to an ADEPT robot arm. Comparisons between the three proposed methods based on experimental results are also presented.