Building-Guided Pseudo-Label Learning for Cross-Modal Building Damage Mapping

2025-05-08

AI Summary

Accurate post-earthquake building damage assessment remains challenging due to the cross-modal discrepancy between pre-event optical and post-event SAR imagery. Method: This paper proposes a building-prior-guided pseudo-label learning framework. It introduces, for the first time, a building-mask-constrained low-uncertainty pseudo-label refinement strategy that jointly leverages multi-model building segmentation, test-time augmentation (TTA), and a cross-modal change detection network—explicitly embedding structural building priors into the damage classification process. Contribution/Results: Evaluated on the 2025 IEEE GRSS Data Fusion Contest dataset, the method achieves a new state-of-the-art mean Intersection-over-Union (mIoU) of 54.28%, securing first place in the competition. It significantly improves both accuracy and robustness of post-disaster damage mapping, particularly under severe cross-modal domain shifts.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Accurate building damage assessment using bi-temporal multi-modal remote sensing images is essential for effective disaster response and recovery planning. This study proposes a novel Building-Guided Pseudo-Label Learning Framework to address the challenges of mapping building damage from pre-disaster optical and post-disaster SAR images. First, we train a series of building extraction models using pre-disaster optical images and building labels. To enhance building segmentation, we employ multi-model fusion and test-time augmentation strategies to generate pseudo-probabilities, followed by a low-uncertainty pseudo-label training method for further refinement. Next, a change detection model is trained on bi-temporal cross-modal images and damaged building labels. To improve damage classification accuracy, we introduce a building-guided low-uncertainty pseudo-label refinement strategy, which leverages building priors from the previous step to guide pseudo-label generation for damaged buildings, reducing uncertainty and enhancing reliability. Experimental results on the 2025 IEEE GRSS Data Fusion Contest dataset demonstrate the effectiveness of our approach, which achieved the highest mIoU score (54.28%) and secured first place in the competition.

Problem

• Mapping building damage from pre- and post-disaster images
• Improving building segmentation using multi-model fusion
• Enhancing damage classification with building-guided pseudo-labels

Innovation

• Building-Guided Pseudo-Label Learning Framework
• Multi-model fusion and test-time augmentation
• Building-guided low-uncertainty pseudo-label refinement