Atmospheric Correction

Basic principles of remote sensing: This includes the electromagnetic spectrum, radiation, wavelengths, and sensors used in remote sensing.

Atmospheric scattering and absorption: Understanding how the atmosphere scatters and absorbs radiation is crucial to atmospheric correction, as it affects the way the signal is received by sensors.

Atmospheric models: There are several atmospheric models used in atmospheric correction, such as the standard atmosphere model, MODTRAN, and 6S. Understanding how these models work is essential.

Radiative transfer: Radiative transfer refers to the transfer of energy through the atmosphere. Knowledge of this process is required to understand atmospheric correction algorithms.

Atmospheric correction algorithms: There are many different algorithms used in atmospheric correction, including the Dark Target algorithm, the Aerosol Robotic Network algorithm, and the Radiative Transfer for Retrieval of Aerosol and Cloud algorithm.

Aerosols: Aerosols play a significant role in atmospheric correction, and understanding their properties and how they affect remote sensing data is critical.

Surface reflectance: Once atmospheric correction has been applied, the data needs to be converted to surface reflectance. Understanding how this is done, and the different methods available, is essential.

Quality assessment and validation: Finally, it is important to assess the quality of the atmospheric correction, and to validate the results. This includes understanding the different metrics that can be used to assess the quality of the data.

Dark Object Subtraction (DOS): This technique involves subtracting the radiance due to the dark object within the image, which is a known value, from the measured radiance.

Empirical Line Method (ELM): It uses a linear regression between the spectral reflectance of a ground reference site and the corresponding digital numbers in the image.

FLAASH: Fast Line-of-sight Atmospheric Analysis of Spectral Hypercubes, a method to correct atmospheric effects in images using MODTRAN radiation transfer codes in order to compute the atmospheric path of radiance.

Atmospheric Radiative Transfer Simulator (ARTS): It is a fast radiative transfer model for simulating the atmospheric effects on satellite-borne imaging systems.

Second Simulation of a Satellite Signal in the Solar Spectrum (6S): A radiative transfer model for calculating the atmospheric path of radiance.

Spectral Angle Mapping (SAM): A method of comparing spectra of pixels in an image with a known spectrum in order to classify them.

Principal Component Analysis (PCA): This technique creates a linear combination of bands to reduce the dimensionality of the data, while at the same time maintaining the maximum amount of information, hence remove atmospheric effects.

Modified Tight-Frame (MTF) algorithm: A variant of the PCA that usually performs better. It involves creating a linear combination of basis spectra via a tight frame.

Minimum Noise Fraction (MNF): Another variant of PCA technique that uses decorrelation to identify the minimum noise fraction from the high-dimensional data.

Hyperspherical Convex Hull (HCH) technique: A linear optimization approach to the problem of atmospheric correction. It assumes that convex spectra lie in a hypersphere in a high-dimensional spectral space.