Fast and efficient FPGA-based Euro coin identification

Abstract

This paper describes an efficient hardware implementation of a cross-correlation algorithm on a Field-Programmable Gate Array (FPGA) platform with the purpose of visually identifying coins. This method has the ability to compare images obtained via a video camera with those stored in memory, thereby, accepting/rejecting coins-under-test at the very high-speed demanded by today's coin validation mechanisms. In addition, the user is offered the option to refresh the FPGA memory, i.e. add/remove coin images, on-site as well as store the test results on a host-machine if desired so. The implementation described here constitutes part of a wider coin-validation system that utilises multiple techniques, such as this, operating inparallel with the sole purpose of Euro-coin identification at an acceptable performance to today's standards and criteria.