COMPONENT IDENTIFICATION AND ORIENTATION ESTIMATION WITH NEURAL NETWORKS AND SINGLE CAMERA VISION SYSTEM
MAHDI VAEZI *
Electronics and Computer Science Faculty, Science and Research Branch of Azad University, Tehran, Iran
*Author to whom correspondence should be addressed.
Abstract
Despite the quick development of artificial intelligence theory and applications yet again a real remarkable deployment like fully automated component assembly of manufacturing lines is still far from reality and the components have to be set at predefined positions on the assembly line prior to be picked and joined together. However if the parts are going to be put on random positions and discovered by the intelligent machine then the essential requirement of such a system is the development of component recognition system and an efficient position estimation method. At this paper a novel approach to resolve this need is discussed which utilizes both pattern recognition and neural networks algorithms to resolve the problem.
Keywords: Assembly components, machine vision, pattern recognition, neural networks