Perpustakaan Digital ITB

Recently, money detection using ultraviolet device has commonly known and had good ccuration by mean combining with human ability as observer. The difficulty raised when we attempted to detect on several sheet of money that spent more time to accomplish. In this paper We proposed a new alternative method to detect the counterfeit of new money using regular image scanner. Assumed that all new true money had special features that must be granted, they were the quality and correctly of printing. This special features will detect further by mean of pattern searching techniques i.e. image processing and artificial neural network. The special features we chosen were money value and carved pattern. This research were done on new 20,000 rupiahs and gave result 90% success for 10 data; 6 true money and 4counterfeit.