Image based counterfeit coin validation

Abstract

This paper presents the imaging solution for counterfeit and unfit coin detection based on extraction of unique segments on the reference image to be used for cross-correlation matching for validation, calculation of the diameter and thickness of coin and extraction of symbols from the “third” side of a coin or edge. The selection method of the unique segments to be used for cross-correlation is described and analysed. In the validation by cross correlation technique at least two images of one coin are used: face side and edge side are used, providing the high accuracy of proposed solution. This algorithm was implemented and tested in Matlab. The results of experimental tests on real euro coin images are presented and analysed.