Real-time Hyperspectral Data Compression Using Principal Components Transformation

The advantages of Hyperspectral Imaging (HSI) over conventional remote sensing imaging has been well recognized (Schowengerdt, 1983). HSI provides the user with spatial and spectral information about scene objects and facilitates improved detection and identification over panchromatic and multispectral imagery. The predecessors of present day HSI systems were the broad band spaceborne multi-spectral imaging systems like MSS and the more recent TM, that provided data in 4 and 7 spectral bands respectively. Present airborne HSI systems like AVIRIS and HYDICE operate in the VIS-SWIR spectral region (0.4-2.5 μm). They acquire data at a higher spectral resolution (~ 10 nm) and produce data cubes in 224 and 210 bands respectively. The SEBASS sensor acquires data in the MWIR and LWIR regions in 100 and 128 spectral bands each.