Hyperspectral

Hyperspectral imaging, also known as imaging spectroscopy, has revolutionized the remote sensing industry. The unprecedented detail shown in the spectral curves allows the user to move beyond making gross generalities as to the composition of a material and affix a definite, precise identity.

New Sensor Evaluation

MDA has played a major role in evaluating a state-of-the-art hyperspectral sensor manufactured by SpecTIR, Inc. This sensor utilizes a patented stabilization system resulting in geometrically sound images unaffected by deviations in the aircraft's pitch, yaw, and roll. In addition to validating the geometric stability of the instrument, MDA developed a production pipeline to take the raw imagery and generated distributable radiometrically calibrated and georeferenced datasets.

Pre-Processing

Spectrum Polishing

Hyperspectral sensors not only get a detailed look at the surface, but also at the atmosphere between the surface and the instrument. For surface applications, it is imperative that the atmosphere is accurately characterized and removed from the signal to generate a surface reflectance data set. MDA uses the latest technology to perform this calibration, either with field calibration spectra or without. Residual noise left in the reflectance data is corrected for using a proprietary spectral polishing algorithm.

Algorithm Development

Hyperspectral imaging is an evolving technology being applied to new and ever more challenging applications every day. The software needed to solve such challenges is becoming ever more complex, and is well beyond the reach of existing COTS packages. MDA has developed a library of unique algorithms spanning a wide variety of applications, including:

• Autonomous endmember selection
• Hydrocarbon seep detection
• Vegetation stress mapping
• Spectral polishing
• Linear spectral unmixing enhancements
• Automated optimal spectral ratio generation