RSCC: A Large-Scale Remote Sensing Change Caption Dataset for Disaster Events

2025-09-01

AI Summary

Existing remote sensing datasets suffer from a lack of bitemporal image pairs and fine-grained textual annotations, hindering semantic understanding of dynamic disaster impacts. To address this, we introduce the first large-scale pre-disaster–post-disaster remote sensing image–text paired dataset, comprising 62,315 high-quality pairs covering multiple disaster types, each annotated with human-level natural language change descriptions. We propose a “vision-language joint change modeling” paradigm, leveraging human-in-the-loop curation and multi-stage automated pipelines to ensure annotation accuracy and semantic richness. This dataset is the first to systematically support bitemporal vision-language understanding, significantly enhancing model interpretability and fine-grained change description capabilities. Extensive evaluation on multiple vision-language pretraining (VLP) benchmarks confirms its effectiveness for both training and assessment. Our work advances intelligent remote sensing interpretation toward greater readability, granularity, and semantic fidelity.

Technology Category

• Computer Vision: Language and Vision
• Natural Language Processing: Language Grounding & Multi-modal NLP
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• Web Mining and Content Analysis: Large pretrained models with web data
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Remote sensing is critical for disaster monitoring, yet existing datasets lack temporal image pairs and detailed textual annotations. While single-snapshot imagery dominates current resources, it fails to capture dynamic disaster impacts over time. To address this gap, we introduce the Remote Sensing Change Caption (RSCC) dataset, a large-scale benchmark comprising 62,315 pre-/post-disaster image pairs (spanning earthquakes, floods, wildfires, and more) paired with rich, human-like change captions. By bridging the temporal and semantic divide in remote sensing data, RSCC enables robust training and evaluation of vision-language models for disaster-aware bi-temporal understanding. Our results highlight RSCC's ability to facilitate detailed disaster-related analysis, paving the way for more accurate, interpretable, and scalable vision-language applications in remote sensing. Code and dataset are available at https://github.com/Bili-Sakura/RSCC.

Problem

• Lacks temporal image pairs for disaster monitoring
• Absence of detailed textual annotations in datasets
• Need for dynamic disaster impact analysis over time

Innovation

• Large-scale pre-/post-disaster image pairs dataset
• Human-like change captions for temporal analysis
• Vision-language models for bi-temporal disaster understanding