Sparse-Sampling Methods For Hyperspectral Infrared Microscopy

A hyperspectral beam-scanning microscope operating in the long wave infrared (LWIR) is demonstrated for future application to stand-off imaging platforms. A 32-channel quantum-cascade laser (QCL) array enables rapid wavelength modulation for fast hyperspectral imaging through sparse sampling in position and wavelength, which when coupled with image reconstruction techniques can enhance frame rate. Initial measurements of dichloromethane and water mixtures are shown, utilizing spectral information for classification across the field of view. Ongoing efforts aim to utilize co-propagating visible and IR beams to enhance spatial resolution for the IR measurements by combining spatial information retrieved from visible images obtained concurrently. Future work will leverage Lissajous trajectories for sparsely-sampled beam-scanning and extend the image interpolation algorithms to arbitrary dimension for sparse sampling in the spectral domain. Simulations of the error associated with various sparse-sampling methods are also presented herein which support the use of Lissajous trajectories as a sparse-sampling method in beam-scanning microscopy.