Understanding deep learning in land use classification based on Sentinel-2 time series

TLDR

Deep learning is increasingly applied to remote sensing, achieving high accuracy across many tasks, yet its lack of interpretability hampers wider adoption, particularly in policy‑driven contexts that demand accountable decisions. The study seeks to elucidate how a recurrent neural network classifies land use from Sentinel‑2 time series within the European Common Agricultural Policy framework. The authors analyze predictor relevance in the network to better understand its decision‑making behavior. Analysis shows that red and near‑infrared bands, especially summer acquisitions, provide the most informative features, aiding the interpretation of models used in CAP to support the European Green Deal.

Abstract

Abstract The use of deep learning (DL) approaches for the analysis of remote sensing (RS) data is rapidly increasing. DL techniques have provided excellent results in applications ranging from parameter estimation to image classification and anomaly detection. Although the vast majority of studies report precision indicators, there is a lack of studies dealing with the interpretability of the predictions. This shortcoming hampers a wider adoption of DL approaches by a wider users community, as model’s decisions are not accountable. In applications that involve the management of public budgets or policy compliance, a better interpretability of predictions is strictly required. This work aims to deepen the understanding of a recurrent neural network for land use classification based on Sentinel-2 time series in the context of the European Common Agricultural Policy (CAP). This permits to address the relevance of predictors in the classification process leading to an improved understanding of the behaviour of the network. The conducted analysis demonstrates that the red and near infrared Sentinel-2 bands convey the most useful information. With respect to the temporal information, the features derived from summer acquisitions were the most influential. These results contribute to the understanding of models used for decision making in the CAP to accomplish the European Green Deal (EGD) designed in order to counteract climate change, to protect biodiversity and ecosystems, and to ensure a fair economic return for farmers.