ABSTRACT
Atmospheric correction is critical in the processing of hyperspectral data. It is one of the main steps in converting digital numbers to true reflectance values. The conversion to true reflectance is necessary to analyse the spectral signature of the target material, more specifically to determine the diagnostic absorption and reflection of the target material. The effect of the atmosphere on the radiance recorded by satellites is multiplicative and additive. Back-scattering by the atmosphere is additive and absorption and scattering in the direction other than the sensor are multiplicative.
There are two major methods for converting digital numbers or radiance values to reflectance. The first one is a physics-based method and the second is an image-based or empirical method. The physics-based methods model the atmosphere and light interaction using first principles and estimate the absorption and scattering by the atmosphere. The image-based methods are simpler than physics-based methods. They use all the data available in the image to extract the required multiplicative and additive correcting factors. Physics-based methods are complex and require measurements of atmospheric properties which are difficult. Image-based methods do not require any data. Atmospheric correction of the hyperspectral data needs careful attention because of its high spectral resolution.
This chapter will discuss:
o Importance of atmospheric correction
o Atmospheric effects
o Image-based methods
o Dark object,
o Improved dark object,
o Flat field method,
o Internal area relative reflectance,
o Physics-based method – RTC code, 6SV
o Working example for image based on physics-based methods
