Abstract
Image processing entered the era of artificial intelligence, and machine learning algorithms emerged as attractive alternatives for time series data processing. Satellite image time series processing enables crop phenology monitoring, such as the calculation of start and end of season. Among the promising algorithms, Gaussian process regression (GPR) proved to be a competitive time series gap-filling algorithm with the advantage of, as developed within a Bayesian framework, providing associated uncertainty estimates. Nevertheless, the processing of time series images becomes computationally inefficient in its standard per-pixel usage, mainly for GPR training rather than the fitting step. To mitigate this computational burden, we propose to substitute the per-pixel optimization step with the creation of a cropland-based precalculations for the GPR hyperparameters θ. To demonstrate our approach hardly affects the accuracy in fitting, we used Sentinel-2 LAI time series over an agricultural region in Castile and Leon, North-West Spain. The performance of image reconstructions were compared against the standard per-pixel GPR time series processing. Results showed that accuracies were on the same order (RMSE 0.1767 vs. 0.1564 [m2/m2], 12% RMSE degradation) whereas processing time accelerated about 90 times. We further evaluated the alternative option of using the same hyperparameters for all the pixels within the complete scene. It led to similar overall accuracies over crop areas and computational performance. Crop phenology indicators were also calculated for the three different approaches and compared. Results showed analogous crop temporal patterns, with differences in start and end of growing season of no more than five days. To the benefit of crop monitoring applications, all the gap-filling and phenology indicators retrieval techniques have been implemented into the freely downloadable GUI toolbox DATimeS.
Highlights
Earth observation (EO) is used to monitor and assess continuously the status of, and changes in, natural and managed vegetated lands [1,2]
A diversity of interpolation and fitting methods can perform this task, but the difficulty lies in the choice of the one that successfully reconstruct vegetation indices and retrieve reliable phenology indicators such as dates of start and end of growing season (SOS and EOS, respectively), maximum peak, day of maximum value (DOM), amplitude, length of the season (LOS), etc. [13], which are narrowly related to essential sources of information including start of senescing, harvest day, productivity estimates, irrigation management, nutrient management, health management, yield prediction, and crop type mapping [14,15,16]
The core part of this study is ascertaining how using precalculated hyperparameters optimized over crop areas affects the leaf area index (LAI) estimates and performance of Gaussian processes regression (GPR) models
Summary
Earth observation (EO) is used to monitor and assess continuously the status of, and changes in, natural and managed vegetated lands [1,2]. The usage of optical EO time series has opened the door to global-scale crop monitoring through their spectral properties using different kinds of Agronomy 2020, 10, 618; doi:10.3390/agronomy10050618 www.mdpi.com/journal/agronomy. The generation of continuous fields in time and space (i.e., gap-filling) starting from irregularly distributed data is of critical importance. These time series are associated with significant uncertainties and incomplete because of inadequate climatic conditions (e.g., clouds, snow and aerosols), and the long interval needed for the satellites to revisit and acquire data for the exact same location [8]. A diversity of interpolation and fitting methods can perform this task (e.g., see review of [12]), but the difficulty lies in the choice of the one that successfully reconstruct vegetation indices and retrieve reliable phenology indicators such as dates of start and end of growing season (SOS and EOS, respectively), maximum peak, day of maximum value (DOM) (when the largest value per cycle occurs), amplitude (difference between the maximum and the average of the left and right minimum values per season), length of the season (LOS), etc. [13], which are narrowly related to essential sources of information including start of senescing, harvest day, productivity estimates, irrigation management, nutrient management, health management, yield prediction, and crop type mapping [14,15,16]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
