Deep Learning Based Domain Adaptation Methods in Remote Sensing: A Comprehensive Survey

TL;DR

This work addresses the pervasive problem of domain shift in remote sensing by offering a unified, taxonomy-driven survey of deep learning–based domain adaptation methods. It comprehensively categorizes methods by supervision paradigm, input mode, algorithm granularity, and task type (classification, segmentation, detection, and change detection), and it connects these with widely used benchmark datasets and performance analyses. Key contributions include formal definitions and notations, a four-part methodological taxonomy, a synthesis of benchmark results across RS modalities, and a forward-looking view toward transformer architectures and foundation-model–driven DA-RS. By consolidating methods, datasets, and results, the paper provides a practical roadmap for researchers and practitioners to design robust, cross-domain RS systems with improved generalization and transferability.

Abstract

Domain adaptation is a crucial and increasingly important task in remote sensing, aiming to transfer knowledge from a source domain a differently distributed target domain. It has broad applications across various real-world applications, including remote sensing element interpretation, ecological environment monitoring, and urban/rural planning. However, domain adaptation in remote sensing poses significant challenges due to differences in data, such as variations in ground sampling distance, imaging modes from various sensors, geographical landscapes, and environmental conditions. In recent years, deep learning has emerged as a powerful tool for feature representation and cross-domain knowledge transfer, leading to widespread adoption in remote sensing tasks. In this paper, we present a comprehensive survey of significant advancements in deep learning based domain adaptation for remote sensing. We first introduce the preliminary knowledge to clarify key concepts, mathematical notations, and the taxonomy of methodologies. We then organize existing algorithms from multiple perspectives, including task categorization, input mode, supervision paradigm, and algorithmic granularity, providing readers with a structured understanding of the field. Next, we review widely used datasets and summarize the performance of state-of-the-art methods to provide an overview of current progress. We also identify open challenges and potential directions to guide future research in domain adaptation for remote sensing. Compared to previous surveys, this work addresses a broader range of domain adaptation tasks in remote sensing, rather than concentrating on a few subfields. It also presents a systematic taxonomy, providing a more comprehensive and organized understanding of the field. As a whole, this survey can inspire the research community, foster understanding, and guide future work in the field.

Figures (7)

• Figure 1: Illustration of the cases to explain the main challenges of domain adaptation methods in remote sensing. Different colors represent specific sections. Best viewed in color.
• Figure 2: Illustration of the pipeline of this survey. Different colors represent specific sections. Best viewed in color.
• Figure 3: Overview on the method taxonomy on domain adaptation remote sensing tasks of this survey.
• Figure 4: Overview on the four domain adaptation remote sensing tasks of this survey.
• Figure 5: Illustration of the common-utilized distribution measurement paradigm. $E$ and $\{f_{S}, f_{T}\}$ respectively denote encoders and feature mapping modules.