Vision-Language Models for Satellite Image Understanding and Natural-Language Reporting: A Survey and a Lightweight LoRA-Based Captioning Pipeline

Author

A. Ratna Rishika and Dr. S. Akila Agnes


Abstract

 The rapid growth of high-resolution satellite data has made manual interpretation slow and dependent on expert analysts. Vision-Language Models (VLMs) address this problem by combining computer vision with natural language processing, so that an image can be described, questioned and reported on in plain language. This paper surveys twenty recent publications on vision-language learning for remote sensing, grouped into foundation models, large instruction-tuned multimodal models, image–text datasets, cross-modal alignment and task-specific studies. The surveyed work is compared by objective, limitation, advantage, evaluation metric and remaining research gap, and five representative methodologies (AddressCLIP, LHRS-Bot, VHM, a generative change-detection framework and SARCLIP) are analysed in detail. On this basis, a lightweight pipeline is proposed in which a pre-trained BLIP captioning model is adapted to remote-sensing imagery using Low-Rank Adaptation (LoRA) on the RSICD dataset. On 200 test images the BLEU score improved from 3.92 for the un-tuned baseline to 10.23 for the fine-tuned model, and the captions changed from a generic template to descriptions that name objects and layout. Applications, open challenges (scarce captions, domain gap, model size and hallucination) and future directions are also discussed.



Keywords

Vision-Language Models, Satellite Image Understanding, Remote Sensing, Image Captioning, Multimodal Artificial Intelligence, Earth Observation, LoRA



Full Text:

Download Paper PDF


References


  • Xu, C. Zhang, L. Fan, G. Meng, S. Xiang, and J. Ye, “AddressCLIP: Empowering vision-language models for citywide image address localization,” in Proc. European Conf. Computer Vision (ECCV), 2024, pp. 76–92.
  • Weng, C. Pang, and G.-S. Xia, “Vision-language modeling meets remote sensing: Models, datasets and perspectives,” IEEE Geoscience and Remote Sensing Magazine, 2025.
  • Li, C. Wen, Y. Hu, Z. Yuan, and X. X. Zhu, “Vision-language models in remote sensing: Current progress and future trends,” IEEE Geoscience and Remote Sensing Magazine, vol. 12, no. 2, pp. 32–66, 2024.
  • Radford et al., “Learning transferable visual models from natural language supervision,” in Proc. 38th Int. Conf. Machine Learning (ICML), PMLR vol. 139, 2021, pp. 8748–8763.
  • Muhtar, Z. Li, F. Gu, X. Zhang, and P. Xiao, “LHRS-Bot: Empowering remote sensing with VGI-enhanced large multimodal language model,” in Computer Vision – ECCV 2024, Springer, pp. 440–457.
  • Pang et al., “VHM: Versatile and honest vision language model for remote sensing image analysis,” Proc. AAAI Conf. Artificial Intelligence, vol. 39, no. 6, pp. 6381–6388, 2025.
  • Zan, S. Ji, S. Chao, and M. Luo, “Open-vocabulary generative vision-language models for creating a large-scale remote sensing change detection dataset,” ISPRS J. Photogrammetry and Remote Sensing, vol. 225, 2025.
  • “SARCLIP: The first vision-language foundation model for SAR image,” IEEE Trans. Geoscience and Remote Sensing, 2025.
  • Cha, D. Yu, J. Seo, H. Choi, and T. Jeon, “Pushing the limits of vision-language models in remote sensing without human annotations,” Int. Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, vol. XLVIII-G2025, pp. 249–254, 2025.
  • Hu, J. Yuan, C. Wen, X. Lu, et al., “RSGPT: A remote sensing vision language model and benchmark,” ISPRS J. Photogrammetry and Remote Sensing, vol. 224, pp. 272–286, 2025.
  • Zhan, Z. Xiong, and Y. Yuan, “SkyEyeGPT: Unifying remote sensing vision-language tasks via instruction tuning with large language model,” ISPRS J. Photogrammetry and Remote Sensing, vol. 221, 2025.
  • Zavras, D. Michail, B. Demir, and I. Papoutsis, “Mind the modality gap: Towards a remote sensing vision-language model via cross-modal alignment,” ISPRS J. Photogrammetry and Remote Sensing, vol. 228, pp. 270–287, 2025.
  • Choudhury, P. Kurkure, and B. Banerjee, “Improving visual grounding in remote sensing images with adaptive modality guidance,” ISPRS J. Photogrammetry and Remote Sensing, vol. 224, pp. 42–58, 2025.
  • Chen et al., “A vision-language foundation model-based multimodal retrieval-augmented generation framework for remote sensing lithological recognition,” ISPRS J. Photogrammetry and Remote Sensing, 2025.
  • Ge, X. Zhang, Y. Zheng, K. Guo, and J. Liang, “RSTeller: Scaling up visual language modeling in remote sensing with rich linguistic semantics,” ISPRS J. Photogrammetry and Remote Sensing, 2025.
  • Zhao, M. Zhang, B. Yang, et al., “LuoJiaHOG: A hierarchyoriented geo-aware image caption dataset for remote sensing image-text retrieval,” ISPRS J. Photogrammetry and Remote Sensing, 2025.
  • Zhang et al., “Frequency domain transfer learning for remote sensing visual question answering,” Expert Systems with Applications, vol. 291, p. 128395, 2025.
  • Kuckreja, M. S. Danish, M. Naseer, A. Das, S. Khan, and F. S. Khan, “GeoChat: Grounded large vision-language model for remote sensing,” in Proc. IEEE/CVF Conf. Computer Vision and Pattern Recognition (CVPR), 2024, pp. 27831–27840.
  • Zhang, T. Zhao, Y. Guo, and J. Yin, “RS5M and GeoRSCLIP: A large-scale vision-language dataset and a large vision-language model for remote sensing,” IEEE Trans. Geoscience and Remote Sensing, vol. 62, pp. 1–23, 2024.
  • Wang, R. Prabha, T. Huang, J. Wu, and R. Rajagopal, “SkyScript: A large and semantically diverse vision-language dataset for remote sensing,” Proc. AAAI Conf. Artificial Intelligence, vol. 38, no. 6, pp. 5805–5813, 2024.
  • Liu et al., “RS-LLaVA: A large vision-language model for joint captioning and question answering in remote sensing imagery,” Remote Sensing, vol. 16, no. 9, p. 1477, 2024.

Share your valuable work from Social Media Buttons