Discrimination of vapors based on direct comparison of passive Fourier transform infrared interferograms

摘要

A real-time compound identification system for Fourier Transform Infrared Spectroscopy data has been based on the direct comparison of interferograms. Statistical pattern- recognition methods are applied to the feature extraction of infrared interferograms. Using large training sets, a real- time classification filter has been developed that is able to discriminate the specific-compound of infrared interferograms, and it has a very high probability. Our system can be used to identify several specific-compounds (several chemical vapors) of infrared interferograms of remote detection.