Partial least squares regression (PLSR) is a powerful tool that makes use of the information at all wavelengths and provides the means for the quantitative determination of a wide variety of plant canopy properties. Past studies that limited their analysis to the use of spectral indices (e.g. NDVI, SAVI, PVI – see links below for more details) often weren’t able to take advantage of the full spectrum since multispectral systems provide information at a limited set of wavelengths.
With the widespread availability of hyperspectral data, the use of spectral indices alone doesn’t make full use of the available information. Thus, it is not surprising that Lucie Cervena et al. reported at a past EARSeL Symposium that PLSR outperformed a range of vegetation index-based methods in predicting both leaf pigments as well as leaf water content.
You can read the full paper by clicking on the link below:
For more information on PLSR, see the following:
- Partial Least-Squares Methods for Spectral Analyses. 1. Relation to Other Quantitative Calibration Methods and the Extraction of Qualitative Information
- PLS-regression: a basic tool of chemometrics
- Interpreting Vegetation Indices
- Červená, Lucie & Lhotakova, Zuzana & Kupková, Lucie & Kovarova, Monika & Albrechtova, Jana. (2014). Models for estimating leaf pigments and relative water content in three vertical canopy levels of Norway spruce based on laboratory spectroscopy. 10.12760/03-2014-11.