Explainable AI for Earth Observation: Current Methods, Open Challenges, and Opportunities

TL;DR

This paper reviews current explainable AI methods in Earth observation, highlighting challenges and opportunities in making deep learning models more interpretable for remote sensing applications.

Contribution

It provides a comprehensive overview of state-of-the-art explainable AI techniques tailored for remote sensing image analysis across various Earth observation fields.

Findings

Deep learning models dominate remote sensing analysis but lack interpretability.

Explainable AI techniques are increasingly explored to address this gap.

Challenges include balancing model performance with interpretability.

Abstract

Deep learning has taken by storm all fields involved in data analysis, including remote sensing for Earth observation. However, despite significant advances in terms of performance, its lack of explainability and interpretability, inherent to neural networks in general since their inception, remains a major source of criticism. Hence it comes as no surprise that the expansion of deep learning methods in remote sensing is being accompanied by increasingly intensive efforts oriented towards addressing this drawback through the exploration of a wide spectrum of Explainable Artificial Intelligence techniques. This chapter, organized according to prominent Earth observation application fields, presents a panorama of the state-of-the-art in explainable remote sensing image analysis.