Learning Structures in Earth Observation Data with Gaussian Processes
This paper provides a review of GP methods for researchers and practitioners in geoscience and remote sensing, aiming to consolidate recent advancements and their practical applications.
This paper reviews recent theoretical developments and algorithms for Gaussian Processes (GPs) in Earth observation data. It highlights new GP algorithms that respect signal/noise characteristics, provide automatic feature rankings, and enable the applicability of uncertainty intervals for spatio-temporal model transport, illustrating these with examples in geoscience and remote sensing.
Gaussian Processes (GPs) has experienced tremendous success in geoscience in general and for bio-geophysical parameter retrieval in the last years. GPs constitute a solid Bayesian framework to formulate many function approximation problems consistently. This paper reviews the main theoretical GP developments in the field. We review new algorithms that respect the signal and noise characteristics, that provide feature rankings automatically, and that allow applicability of associated uncertainty intervals to transport GP models in space and time. All these developments are illustrated in the field of geoscience and remote sensing at a local and global scales through a set of illustrative examples.