The Hyperspectral Analysis Package is designed for processing and analyzing images acquired with airborne and satellite-borne imaging spectrometers. Such images represent many narrow and practically contiguous wavelength intervals, or bands, typically throughout the visible, near-infrared, and mid-infrared wavelength intervals. The Hyperspectral Analysis Package accommodates images with up to 1024 bands.
Hyperspectral sensing, and the processing and analysis of hyperspectral images is a way of discriminating earth surface features that have diagnostic absorption and reflection characteristics over narrow wavelength intervals. These reflection characteristics are lost within the relatively large bandwidths of a conventional multispectral scanner.
The Airborne Visible/Infra-Red Imaging Spectrometer (AVIRIS) and the Landsat 7 Enhanced Thematic Mapper Plus (ETM+) together exemplify the spectral sampling differences between an imaging spectrometer and a multispectral sensor. The AVIRIS has 224 practically continguous bands each of which are less than 10 nm wide, whereas the ETM+ has 7 non-contiguous bands that are no narrower than 150 nm.
Reflectance spectra (reflectance as a continuous function of wavelength) for materials (pure or mixed) in a scene may be estimated from a hyperspectral image of the scene. Reflectance spectra are independent of atmospheric conditions in the scene.
Although reflectance spectra are not necessarily independent of scene illumination and sensor view direction, they may often be treated as an identifying signature of the observed material or material mixture. Much of hyperspectral image processing and analysis is concerned with removing atmospheric effects from image data in order to estimate scene reflectance, and then identifying materials or material mixtures scene by comparing image-derived reflectance spectra with laboratory-measured reflectance spectra (comprising a spectra library) for pure materials.
© PCI Geomatics Enterprises, Inc.®, 2024. All rights reserved.