Owing to accuracy and capability of noise-free, the laser scanner has been used in order to recognize objects in real industrial field. In this thesis, new recognition algorithm for truck automation system, which use range image obtained by laser scanner, is presented. Pre-stage analysis gives a specific information of trucks, and by conferring that information, range image segmentation based on plane approximation is performed. Proposed segmentation method is appropriate to appling to real industrial field because it is faster than other segmentation method. Also, for reliable recognition of trucks at noisy environment, the algorithm which uses Winner-Takes-All Filter at Hough Domain, is proposed. Computer simulation and experimental results show that proposed algorithms are ones of efficient, accurate, and fast method of recognition system.
