Hyperspectral Data Characterization for Object Detection Applications 1

Hyperspectral imaging (HSI) electro-optical sensors, with their hundreds of contiguous spectral bands, are a rich data source useful in a variety of applications. Numerous examples have demonstrated their potential. Understanding the characteristics of HSI data is important not only for algorithm design and assessment, but also for accurate modeling in a performance prediction context. Initial results are presented from an investigation exploring the applicability of characterizing the data with the first and second spectral statistical moments.