A counterfeit detection method based on deep learning infrared detector

Abstract

The invention discloses a depth learning based and infrared geminate transistor employed authenticity identification method of a currency detector and belongs to a field of infrared authenticity identification. The method includes performing signal correctness verification according to multiple channels of infrared geminate transistor waveforms; performing pretreatment on the multiple channels ofthe infrared geminate transistor waveforms; generating a two-dimensional array, converting to an image and processing; performing currency authenticity identification by utilizing the infrared geminate transistor waveforms and adopting CNN training and a detection process in depth learning. The method provided by the invention can realize a good authenticity identification function at a comparatively high processing speed with high accuracy.

Description

technical field

[0001] The invention belongs to the technical field of infrared counterfeit identification, and in particular relates to an infrared counterfeit identification method for banknote detectors based on deep learning. Background technique

[0002] At present, the technique of making counterfeit banknotes with false ones makes people hard to guard against, especially the artificial anti-counterfeit marks in banknotes have been mostly imitated by counterfeiters. Therefore, only by identifying some conventional anti-counterfeiting features, such as: ultraviolet fluorescent reaction, magnetic ink and security thread to identify counterfeit, its accuracy rate is greatly reduced. Because the optically variable inks used in genuine banknote printing are highly confidential, it is difficult for counterfeiters to imitate the optically variable features of the ink areas of genuine banknotes. By comprehensively analyzing and judging the difference in infrared transmission ...

Images