Comparison of partial least squares and support vector regressions for predicting leaf area index on a tropical grassland using hyperspectral data

Abstract

.Leaf area index (LAI) is a key biophysical parameter commonly used to determine vegetation status, productivity, and health in tropical grasslands. Accurate LAI estimates are useful in supporting sustainable rangeland management by providing information related to grassland condition and associated goods and services. The performance of support vector regression (SVR) was compared to partial least square regression (PLSR) on selected optimal hyperspectral bands to detect LAI in heterogeneous grassland. Results show that PLSR performed better than SVR at the beginning and end of summer. At the peak of the growing season (mid-summer), during reflectance saturation, SVR models yielded higher accuracies (R2=0.902 and RMSE=0.371 m2 m−2) than PLSR models (R2=0.886 and RMSE=0.379 m2 m−2). For the combined dataset (all of summer), SVR models were slightly more accurate (R2=0.74 and RMSE=0.578 m2 m−2) than PLSR models (R2=0.732 and RMSE=0.58 m2 m−2). Variable importance on the projection scores show that most of the bands were located in the near-infrared and shortwave regions of the electromagnetic spectrum, thus providing a basis to investigate the potential of sensors on aerial and satellite platforms for large-scale grassland LAI prediction.

Quantifying Vegetation Biophysical Variables from Imaging Spectroscopy Data: A Review on Retrieval Methods

“…Beyond individual vegetation attributes, PLSR was recently used to predict landscape-scale fluxes of net ecosystem exchange (NEE) and gross primary productivity (GPP) across multiple timescales (Matthes et al 2015), and also for the estimation of floristic composition of grassland ecosystems (Harris et al 2015;Neumann et al 2016;Roth et al 2015). At the same time, thanks to its PLS-vectors, PLSR is also increasingly applied for band sensitivity analysis of spectroscopic datasets in view of the targeted application (e.g., Feilhauer et al 2015;Kiala et al 2016;Kira et al 2016;Li et al 2014a;Neumann et al 2016). Various experimental studies demonstrated the superior predictive power of PLSR as opposed to VIs for the prediction of multiple vegetation properties, including above-ground biomass, LAI, leaf pigments (chlorophyll, carotenoids), GPP and NEE fluxes, leaf rust disease detection and nutrients concentration (nitrogen and phosphorus concentrations) (Capolupo et al 2015;Dreccer et al 2014;Foster et al 2017;Hansen and Schjoerring 2003;Matthes et al 2015;Wang et al 2017a;Yue et al 2017).…”

Effects of Growth Stage Development on Paddy Rice Leaf Area Index Prediction Models

“…There are two main approaches to build LAI estimation models from remote sensing data; the empirical statistical approach and the radiative transform model (RTM) approach [8]. The former approach includes univariate regression models built on a vegetation index (VI) and multivariate-calibration-based models using the full reflectance spectrum [14][15][16]. These multivariate calibration techniques include the partial least squares regression (PLSR) methods and modern machine learning methods such as support vector regression (SVR), random forests (RF), and artificial neural networks (ANN).…”

“…PLSR has been considered to be a powerful alternative to univariate methods and provides better performance in most cases [27][28][29], although there is a study that reported the opposite results [22]. Moreover, the potential performance of the state-of-the-art machine learning methods, such as SVR, RF and ANN, has been explored in several studies [14,15,30]. These studies showed that the state-of-the-art machine learning techniques appear to be more efficient than the VI and PLSR methods in most LAI estimation cases.…”

“…These studies showed that the state-of-the-art machine learning techniques appear to be more efficient than the VI and PLSR methods in most LAI estimation cases. Furthermore, several studies [14,15] also point out that performance of PLSR and machine learning method is affect by the development of growth stage. Kiala et al [14] built SVR and PLSR models at different growth stages (early, mid, late and combined) of tropical grassland for LAI estimation.…”

“…Furthermore, several studies [14,15] also point out that performance of PLSR and machine learning method is affect by the development of growth stage. Kiala et al [14] built SVR and PLSR models at different growth stages (early, mid, late and combined) of tropical grassland for LAI estimation. Their results showed that the SVR model yielded better accuracy at the mid-growth stage and whole growing season, whereas the PLSR models performed better at the early and late growth stages.…”