DETEKSI KEASLIAN UANG KERTAS BERDASARKAN WATERMARK DENGAN METODE SUPORT VEKTOR MACHINE (SVM)

Abstract

Money is a legal means of exchange or standard for measuring value (a unit of account), issued by the government of a country in the form of paper, gold, silver or other metals printed with certain shapes and images. Detection is a process of checking or examining something using certain methods and techniques. Along with advances in information technology, crimes that utilize technology are also developing. Currently, digital image processing has developed its usefulness to carry out a recognition system for possible criminal disturbances, especially for recognizing unique objects, watermarks on rupiah banknotes. In image segmentation there are also several methods, for example canny edge detection. Canny edge detection is a method that produces a different image display by displaying a relief effect in it. The aim of this research is to detect the authenticity of banknotes with watermarks using the canny edge detection method. The process of using the method above involves image acquisition,grayscale operations, morphology operations, then canny edge detection. There are 21 images used in this research consisting of nominal banknotes of 1,000, 2,000, 5,000, 10,000, 20,000, 50,000 and 100,000. The final result of the canny edge detection process is a collection of pixels that are used to determine whether the image has a watermark or not. From this research, the results of the accuracy of the watermark detection program on banknotes using the canny edge detection method to detect the authenticity of money were 85.71%.

Keywords: canny, detection, image processing, money, watermark,