FUSAR-GPT : A Spatiotemporal Feature-Embedded and Two-Stage Decoupled Visual Language Model for SAR Imagery

2026-02-22

AI Summary

Existing vision-language models exhibit limited performance on synthetic aperture radar (SAR) imagery, primarily due to the complex SAR imaging mechanism, sensitivity to scattering characteristics, and a scarcity of high-quality image–text paired data. To address these challenges, this work constructs the first SAR image–text–AlphaEarth feature triplet dataset and introduces FUSAR-GPT, a specialized vision-language model. FUSAR-GPT integrates geospatial baseline priors, multi-source remote sensing temporal features, and spatiotemporal anchor embeddings to enable dynamic feature compensation. Furthermore, it employs a two-stage supervised fine-tuning strategy that effectively decouples knowledge injection from task execution. Experimental results demonstrate that FUSAR-GPT significantly outperforms state-of-the-art baselines across multiple remote sensing vision-language benchmarks, achieving performance gains exceeding 12%.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Research on the intelligent interpretation of all-weather, all-time Synthetic Aperture Radar (SAR) is crucial for advancing remote sensing applications. In recent years, although Visual Language Models (VLMs) have demonstrated strong open-world understanding capabilities on RGB images, their performance is severely limited when directly applied to the SAR field due to the complexity of the imaging mechanism, sensitivity to scattering features, and the scarcity of high-quality text corpora. To systematically address this issue, we constructed the inaugural SAR Image-Text-AlphaEarth feature triplet dataset and developed FUSAR-GPT, a VLM specifically for SAR. FUSAR-GPT innovatively introduces a geospatial baseline model as a 'world knowledge' prior and embeds multi-source remote-sensing temporal features into the model's visual backbone via 'spatiotemporal anchors', enabling dynamic compensation for the sparse representation of targets in SAR images. Furthermore, we designed a two-stage SFT strategy to decouple the knowledge injection and task execution of large models. The spatiotemporal feature embedding and the two-stage decoupling paradigm enable FUSAR-GPT to achieve state-of-the-art performance across several typical remote sensing visual-language benchmark tests, significantly outperforming mainstream baseline models by over 12%.

Problem

• Synthetic Aperture Radar
• Visual Language Model
• SAR imagery
• intelligent interpretation
• text corpora

Innovation

• Spatiotemporal Feature Embedding
• Two-Stage Decoupled SFT
• Visual Language Model for SAR
• Geospatial Prior
• SAR Image-Text Dataset