Data analysis of multi-laser standoff spectral identification of chemical and biological compounds

With the availability of tunable broadband coherent sources that emit mid-infrared radiation with well-defined beam characteristics, spectroscopies that were traditionally not practical for standoff detection or for development of miniaturized infrared detectors have renewed interest. While obtaining compositional information for objects from a distance remains a major challenge in chemical and biological sensing, recently we demonstrated that capitalizing on mid-infrared excitation of target molecules by using quantum cascade lasers and invoking a pump probe scheme can provide spectral fingerprints of substances from a variable standoff distance. However, the standoff data is typically associated with random uctuations that can corrupt the fine spectral features and useful data. To process the data from standoff experiments toward better recognition we consider and apply two types of denoising techniques, namely, spectral analysis and Karhunen-Loeve Transform (KLT). Using these techniques, infrared spectral data have been effectively improved. The result of the analysis illustrates that KLT can be adapted as a powerful data denoising tool for the presented pump-probe infrared standoff spectroscopy.