Why are hyperspectral images converted into log10(1/R)?

Would the data variation be different between the spectra extracted from log10(1/R) transformed or non-transformed hyperspectral images?

[Rik]

This doesn't really sound like a Matlab question to me. This not being my field of expertise I can't tell you where this question would be better suited.

Answer 1

[Abhishek Tripathi]

Hi Luqman

The spectra obtained from hyperspectral images that have undergone a log10(1/R) transformation are expected to exhibit different data variation compared to those obtained from non-transformed hyperspectral images.

The log10(1/R) transformation is commonly used in hyperspectral data analysis to account for the non-linear relationship between reflectance and radiance. This transformation can help to normalize the data and reduce the impact of atmospheric effects, such as scattering and absorption.

The log10(1/R) transformation can compress the data by amplifying values near zero. This compression can reduce data variation, which can be advantageous in certain scenarios, such as when working with noisy data. However, in certain cases, data compression may lead to information loss, particularly in regions of the spectrum where reflectance values are low.

Therefore, whether the data variation would be different between the spectra extracted from log10(1/R) transformed or non-transformed hyperspectral images depends on the specific characteristics of the data and the research question being addressed. It is important to carefully consider the potential impact of the log10(1/R) transformation on the data and the specific analysis being performed before deciding whether or not to apply this transformation.

download our Image Processing Toolbox™ Hyperspectral Imaging Library MATLAB®

For more info. see the documentation here