A map of terrain surface reflectance as a function of wavelength band is a valuable remote sensing product (even if practical considerations restrict the map to reflectance as a function of a single view direction and illumination direction). Such a map representing the many contiguous wavelength bands that an imaging spectrometer senses in is particularly valuable. However, image sensors do not measure scene reflectance, or even the surface radiance. At best, a calibrated sensor can measure the radiation that has reached its external optical element along the view direction that intersects each detector element at the instant that the signal from the detector element is captured (the at-sensor or apparent radiance). The layer of atmosphere between the sensor and the scene absorbs and scatters radiation (and even significantly emits radiation at some wavelengths) and causes the at-sensor radiance to differ from scene radiance. So, an image (i.e., an at-sensor radiance map) is a measurement of a combination of ephemeral atmospheric effects and relatively stable reflectance properties of the sensed surface. Atmospheric correction is the process of reducing or removing atmospheric effects from the image.
Techniques for atmospheric correction may be divided into empirical techniques that are relatively simple to apply but are relatively limited in the cases in which they may produce useful results, and more rigourous model-based techniques that are more complicated to apply but are relatively robust. The specific technques that are provided by the Hyperspectral Analysis Package are described in the following sections:
© PCI Geomatics Enterprises, Inc.®, 2026. All rights reserved.