ASFR-Net: Adversarial Alignment and Spatio-Frequency Refinement Network for Heterogeneous Remote Sensing Image Change Detection

2026-07-08

AI Summary

This work addresses the challenge of pseudo-changes in heterogeneous remote sensing change detection caused by discrepancies in imaging mechanisms. To this end, the authors propose an end-to-end Adversarial Spatial-Frequency Refinement Network (ASFR-Net) that aligns cross-domain disparities through modality-invariant representation learning and introduces a novel spatial-frequency collaborative enhancement module to suppress sensor-specific noise. Multi-level difference features are fused to generate precise change maps. The study also contributes the first benchmark dataset, VisNIR-HCD, specifically designed for building change detection using visible and near-infrared imagery. Extensive experiments demonstrate that ASFR-Net achieves state-of-the-art performance on VisNIR-HCD and several public datasets, significantly outperforming existing methods. The code and dataset are publicly released.

Abstract

The core challenge of heterogeneous change detection in remote sensing imagery lies in effectively decoupling genuine land-cover changes from significant modal disparities caused by distinct imaging mechanisms. These intrinsic inconsistencies are prone to introducing pseudo-changes, thereby constraining detection accuracy. To address this, we propose a novel, end-to-end adversarial spatio-frequency refinement network (ASFR-Net). Initially, a modality-invariant representation learner (MIR-Learner) guides the backbone to extract modality-invariant features, effectively bridging the primary domain gap. Subsequently, to address persistent residual modal differences, we design an innovative spatio-frequency synergistic enhancement module (SFEM), which identifies and suppresses sensor-specific noise and artifacts that are difficult to discern in the spatial domain by leveraging frequency-domain processing. Multi-level difference features are then computed from these refined representations and fed into a decoder equipped with cascaded hierarchical guided fusion module (HGFM) blocks to generate precise change maps. To alleviate the data scarcity in heterogeneous tasks, we construct and release a new high-resolution benchmark specifically focused on building changes: the visible-near-infrared heterogeneous change detection (VisNIR-HCD) dataset. It presents unique scientific challenges arising from deceptive visual similarity and non-linear spectral inversions, providing a robust platform for evaluating model generalization. Extensive experiments on VisNIR-HCD and public datasets demonstrate that ASFR-Net achieves state-of-the-art (SOTA) performance, significantly outperforming existing methods. The source code and the VisNIR-HCD dataset are publicly available at https://github.com/LuoYang2024/ASFR-Net.

Problem

• heterogeneous change detection
• remote sensing imagery
• modality disparity
• pseudo-changes
• domain gap

Innovation

• modality-invariant representation
• spatio-frequency synergistic enhancement
• heterogeneous change detection
• adversarial alignment
• frequency-domain refinement