Optimal design of IR sensor's spectral bands for material classification

The spectral matched filtering is based upon the principle of forming a superposition of basis functions such as a set of sensor's spectral bands to perform target classification. We report a band-engineering routine, which is critical to find the optimal design setting for each sensor band to improve the estimation of target's signature such as reflectance of material. The quality of engineered sensor's spectral bands is validated as they apply to the various material classification problems.