Physics-Driven Probabilistic Deep Learning for the Inversion of Physical Models With Application to Phenological Parameter Retrieval From Satellite Times Series

• Published in: IEEE transactions on geoscience and remote sensing Vol. 61; pp. 1 - 23
• Main Authors: Zerah, Yoel, Valero, Silvia, Inglada, Jordi
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.01.2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Autoencoders (AEs); Bayesian analysis; Bayesian physics-guided learning; Biological system modeling; Constraints; Data models; Decoding; Deep learning; generative models; Inverse problems; large scale; Learning algorithms; Machine learning; Mathematical models; Methodology; Methods; Normalized difference vegetative index; Parameters; phenology monitoring; Physical properties; Physics; Probability theory; Remote sensing; Representation learning; Retrieval; Satellite constellations; satellite image time series (SITS); Satellites; self-supervised representation learning; Statistical analysis; Statistical inference
• ISSN: 0196-2892; 1558-0644; 1558-0644
• DOI: 10.1109/TGRS.2023.3284992

Recent Sentinel satellite constellations and deep learning methods offer great possibilities for estimating the states and dynamics of physical parameters on a global scale. Such parameters and their corresponding uncertainties can be retrieved by machine learning methods solving probabilistic inverse problems. Nevertheless, the scarcity of reference data to train supervised methodologies is a well-known constraint for remote sensing applications. To address such limitations, this work presents a new generic physics-guided probabilistic deep learning methodology to invert physical models. The presented methodology proposes a new strategy to combine probabilistic deep learning methods and physical models avoiding simulation-driven machine learning. The inverse problem is addressed through a Bayesian inference framework by proposing a new physically constrained self-supervised representation learning methodology. To show interest in the proposed strategy, the methodology is applied to the retrieval of phenological parameters from normalized difference vegetation index (NDVI) time series. As a result, the probability distributions of the intrinsic phenological model parameters are inferred. The feasibility of the method is evaluated on both simulated and real Sentinel-2 data and compared with different standard algorithms. Promising results show satisfactory accuracy predictions and low inference times for real applications.