Deriving Coefficients of Universal Normalized Vegetation Index and Tasseled Cap Transformation for Sentinel-2 Multispectral Instrument

Traditional vegetation indices (VIs) are widely used with multispectral sensors in contrast to their use with hyperspectral sensors. Despite hyperspectral remote sensing data play an important role in a wide variety of fields such as vegetation monitoring, most of the available VIs are fit for only multispectral sensors. Thus, in contrast to these band-geometric-computing VIs, orthogonal vegetation indices are proved to be representative of both multi- as well as hyperspectral sensors since their ability to utilize all the spectral information. There are two famous indices used in the domain of orthogonal VIs—Tasseled Cap Transformation (TCT) based orthogonal VI consists of three indices and recently developed Universal Normalized Vegetation Index (UNVI) based on Universal Pattern Decomposition Method (UPDM) which works on oblique orthogonal coordinate system. Multispectral Instrument (MSI) on-board Sentinel-2 has a high spatial resolution but its rich spectral information is yet to be discovered in the context of orthogonal vegetation indices. While TCT is derived based on principal component analysis (PCA) and results into three indices—brightness, greenness, and wetness; UNVI is a sensor independent index where each pixel is a linear mixture of water, vegetation, soil, and one supplementary pattern. We calculated the new set of coefficients based on PCA and Procrustes Rotation for TCT and by adopting the methodology of UPDM, entirely new set of coefficient matrices were developed for UNVI. Spectral response function (SRF) of the Sentinel-2 was used in this computation of UNVI. Several research papers have already shown the superiority of UNVI over traditional VIs like NDVI. In this study, we also found the superiority of multi-purpose orthogonal VIs and especially of UNVI over other traditional indices.