Domain-Incremental Remote Sensing Change Detection via Difference-Guided Adaptation and Frequency-Decoupled Distillation

2026-07-14

AI Summary

This work addresses the challenge of catastrophic forgetting in cross-domain incremental learning for remote sensing change detection, a problem exacerbated by the neglect of bitemporal discrepancy cues in existing methods. To mitigate this, the authors propose the DG-FDD framework, which introduces a Difference-Guided Dynamic Adapter (DGDA) to explicitly model feature discrepancies between bitemporal inputs and guide adaptive learning. Additionally, they design a Frequency-Domain Decoupled Knowledge Distillation strategy with Channel Separation (FDKD-CS) that disentangles structural and stylistic information in the frequency domain, enabling stable knowledge transfer without access to historical data. Experimental results demonstrate that under two- and three-domain incremental settings, the proposed method achieves average relative performance drops of only −0.23%/−0.45% and −0.69%/−1.31% in F1 score and IoU, respectively, significantly alleviating catastrophic forgetting.

Abstract

Remote sensing change detection (RSCD) models are prone to catastrophic forgetting when incrementally adapted to new domains. Existing domain-incremental learning (DIL) methods mainly preserve image-level representations but often overlook bitemporal discrepancy cues, which are critical for robust change detection under domain shifts. To address this limitation, we propose DG-FDD, a domain-incremental change detection framework that integrates Difference-Guided Adaptation and Frequency-Decoupled Distillation. Specifically, the Difference-Guided Dynamic Adapter (DGDA) models bitemporal feature discrepancies to promote change-aware feature adaptation and reduce domain-specific interference. Meanwhile, the Frequency-Decoupled Knowledge Distillation strategy with Cross-domain Synthesis (FDKD-CS) separates structural information from domain style in the frequency domain, enabling stable knowledge transfer without historical data. Extensive experiments on three public high-resolution RSCD datasets under two- and three-domain incremental protocols demonstrate that DG-FDD effectively mitigates catastrophic forgetting. Compared with independently trained single-task models, DG-FDD records mean relative changes in F1 and IoU of only -0.23% and -0.45%, respectively, across six two-domain sequences, and -0.69% and -1.31%, respectively, across the three evaluated three-domain sequences. These results indicate a favorable stability-plasticity balance between historical knowledge retention and new-domain adaptation in continual cross-domain change detection.

Problem

• catastrophic forgetting
• domain-incremental learning
• remote sensing change detection
• bitemporal discrepancy
• domain shift

Innovation

• Difference-Guided Adaptation
• Frequency-Decoupled Distillation
• Domain-Incremental Learning
• Remote Sensing Change Detection
• Catastrophic Forgetting