The Research And Application Of Atmospheric Correction Methods Based On Aerosol Optical Thickness Parametric Model

Hyperspectral imaging is the new research field which developed in the remote sensing in the early of 20th century 80s. It is also the leading-edge technology of remote sensing at turn of the century. Hyperspectral remote sensing is up to nanometer scale in the visible spectral resolution, and it often has more bands that can be seen as many as dozens or even hundreds of spectral channels between visible and the near infrared spectral regions. The radiation from the ground received by Hyperspectral imaging sensors includes not only the ground reflectance spectral information, but also the change information of surface irradiance reflection due to the effects of atmospheric radiation transfer. In particular, under the absorption and scattering of atmospheric molecules, aerosols and other substances conditions, electromagnetic waves in the target-atmosphere-sensor occurs distortion during transmission, electromagnetic signal waves transmitted or reflected is attenuated, so that it will bring more uncertainty to the key parameters of the remote sensing of surface. In order to get the real surface features information, we must do some atmospheric corrections stripping "atmosphere's contributions" from the signal sensors received. Atmospheric correction will have a direct impact on the subsequent quantitative analysis, information extraction and other remote sensing applications.This paper takes a deep explanation and analysis to the basic theory and method of the hyperspectral image atmospheric correction, and elaborates in detail for a wide range of applications of typical of relatively correction algorithm, and two absolute calibration algorithm the 6S and MODTRAN radiation model. Based on the current variety of atmospheric correction for hyperspectral image processing method, this paper introduce a atmospheric correction algorithm which is based on the aerosol optical thickness parameterized model, deeply analysis of the theoretical basis of the model, detailed description of the software input and output, multi-angle analysis of experimental results and fully proving the correctness and feasibility of the method. The images after atmospheric correction obtain good correction results not only in terms of subjective visual but also in terms of pixel spectral characteristic curves.