6533b86ffe1ef96bd12cd147

RESEARCH PRODUCT

Machine Learning Methods for Spatial and Temporal Parameter Estimation

subject

description

Monitoring vegetation with satellite remote sensing is of paramount relevance to understand the status and health of our planet. Accurate and constant monitoring of the biosphere has large societal, economical, and environmental implications, given the increasing demand of biofuels and food by the world population. The current democratization of machine learning, big data, and high processing capabilities allow us to take such endeavor in a decisive manner. This chapter proposes three novel machine learning approaches to exploit spatial, temporal, multi-sensor, and large-scale data characteristics. We show (1) the application of multi-output Gaussian processes for gap-filling time series of soil moisture retrievals from three spaceborne sensors; (2) a new kernel distribution regression model that exploits multiple observations and higher order relations to estimate county-level crop yield from time series of vegetation optical depth; and finally (3) we show the combination of radiative transfer models with random forests to estimate leaf area index, fraction of absorbed photosynthetically active radiation, fraction vegetation cover, and canopy water content at global scale from long-term time series of multispectral data exploiting the Google Earth Engine cloud processing capabilities. The approaches demonstrate that machine learning algorithms can ingest and process multi-sensor data and provide accurate estimates of key parameters for vegetation monitoring.