University of Twente Student Theses

Login

Airborne hyperspectral data for estimation and mapping of forest leaf area index

Xie, Rui (2020) Airborne hyperspectral data for estimation and mapping of forest leaf area index.

[img] PDF
1MB
Abstract:Forest ecosystems cover about 30% of the earth’s land surface and provide a significant contribution to the terrestrial biodiversity, biomass and carbon storage, as well as timber production. Quantitative timely information about the forest canopy cover and characteristics is important for ecologists and decision-makers to assess the influence of climate change and expanding human activities on forest ecosystems. However, traditional field sampling of plant traits is often laborious and limited to small areas. Remote sensing, because of its repetitiveness, cost-effectiveness, and non-destructive characterization of land surfaces, has been recognized as a prevalent technology and a practical mean for monitoring forest canopy characteristics over a large scale. Among many characteristics, leaf area index (LAI) is a widely used biophysical parameter to quantify forest health and growth. Thus, accurate estimating LAI and mapping its spatial distribution is crucial for forest management and many ecological studies. Among existing remote sensing-based methods, machine learning algorithms, in particular, kernel-based machine learning methods, such as Gaussian processes regression (GPR), have shown to be promising alternatives to conventional empirical methods for retrieving vegetation parameters. However, the performance of GPR in predicting forest biophysical parameters has hardly been examined in the literature. The main objective of this study was to evaluate the potential of GPR to estimate forest LAI using airborne hyperspectral data. To achieve this, field measurements of LAI were collected in the Bavarian Forest National Park (BFNP), Germany, concurrent with the acquisition of the Fenix airborne hyperspectral images (400-2500 nm) in July 2017. The performance of GPR was further compared with three commonly used empirical methods (i.e. narrowband vegetation indices (VIs), partial least square regression (PLSR), and artificial neural network (ANN)). The cross-validated coefficient of determination (R2 CV) and root mean square error (RMSEcv) between the retrieved and field-measured LAI were used to examine the accuracy of the respective methods. The results showed that using the entire spectral data (400-2500 nm), GPR yielded the most accurate LAI estimation (R2 CV = 0.67, RMSEcv = 0.53 m2 m-2) compared to the best performing narrowband vegetation indices SAVI2 (R2 CV = 0.54, RMSEcv = 0.63 m2 m-2), PLSR (R2 CV = 0.74, RMSEcv = 0.73 m2 m-2) and ANN (R2 CV = 0.68, RMSE = 0.54 m2 m-2). Consequently, when a spectral subset obtained from the analysis of VIs was used as input, the predictive accuracies were generally improved (GPR RMSEcv = 0.52 m2 m-2; ANN RMSEcv = 0.55 m2 m-2; PLSR RMSEcv = 0.69 m2 m-2), indicating that extracting the most useful information from vast hyperspectral bands is crucial for improving model performance. In general, there was an agreement between measured and estimated LAI using different approaches (p > 0.05). The generated LAI map for BFNP using GPR and the spectral subset endorsed the LAI spatial distribution across the dominant forest classes (e.g. deciduous stands were generally associated with higher LAI values). The accompanying LAI uncertainty map generated by GPR shows that higher uncertainties were observed mainly in the regions with low LAI values (low vegetation cover) and forest areas which were not well represented in the collected sample plots. The results of this study demonstrated the potential utility of GPR for estimating LAI in forest stands using airborne hyperspectral data. Owing to its capability to generate accurate predictions and associated uncertainty estimates, GPR is evaluated as a promising candidate for operational retrieval applications of vegetation traits. The generated trait maps can offer spatially explicit and continuous information of vegetation to effectively support sustainable forest management and resource decision-making.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Programme:Spatial Engineering MSc (60962)
Link to this item:https://purl.utwente.nl/essays/84939
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page