Compressive Spectral Video in the LWIR

Physical Sciences Inc. and Colorado State University will develop an innovative sensor that enables low-cost infrared hyperspectral imaging though the use of novel sampling algorithms which provide real-world chemical plume detection capability with compressed data and a hardware configuration which enables high frame rate capture of full 2D spatial and 1D spectral data. Compressive sensing techniques will be used to capture complete hyperspectral data cubes at video frame rates without the need for a costly Long Wave Infrared (LWIR) Focal Plane Array. PSI"s experience in chemical plume detection and Colorado State University"s expertise in reduced order modeling of large data sets will be used to move beyond the standard compressive sensing techniques employed in visible imaging to detect the low thermal contrast signatures of chemical clouds. The proposed system will provide standoff (less than or equal to 5 km) wide-area detection of these clouds for dramatically decreased size, weight, power consumption, and cost compared to current LWIR imaging systems. In Phase I, a compressive LWIR sensor will be designed and modeled, a trade study and down-select of key components performed, relevant algorithms prototyped, and risk-reduction experiments conducted. In Phase II, a TRL 5 prototype system will be developed, integrated, and tested in relevant environments.