遥感大模型:综述与未来设想  

Remote Sensing Large Models:Review and Future Prospects

作　　者：张帅豪 潘志刚 ZHANG Shuaihao;PAN Zhigang(Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100190,China;University of Chinese Academy of Sciences,Beijing 100049,China)

机构地区：[1]中国科学院空天信息创新研究院,北京100190 [2]中国科学院大学,北京100049

出　　处：《遥感技术与应用》2025年第1期1-13,共13页Remote Sensing Technology and Application

基　　金：国家重点研发计划项目(2017YFB0503001)。

摘　　要：深度学习极大地推动了遥感图像处理技术的发展,在精度和速度方面展现了显著优势。然而,深度学习模型在实际应用中通常需要大量人工标注的训练样本,且其泛化性能相对较弱。近年来,视觉基础模型和大语言模型的发展为遥感图像处理的大模型研究引入了新的范式。遥感大模型也称为遥感基础模型,基础模型因其在下游任务中的卓越迁移性能而备受瞩目,这些模型首先在大型数据集上进行与具体任务无关的预训练,然后通过微调适应各种下游应用。基础模型在语言和视觉及其他领域已经得到了广泛应用,其在遥感领域的潜力也正逐渐引起学术界的重视。然而,目前针对这些模型在遥感任务中的全面调查和性能比较仍然缺乏。由于自然图像与遥感图像之间存在固有差异,这些差异限制了基础模型的直接应用。在此背景下,本文从多个角度对常见的基础模型以及专门针对遥感领域的大模型进行了全面回顾,概述了最新进展,突出了面临的挑战,并探讨了未来发展的潜在方向。Deep learning has significantly advanced remote sensing image processing technology,demonstrating notable improvements in both accuracy and speed.However,deep learning models typically require large amounts of manually labeled training samples in practical applications,and their generalization performance is relatively weak.In recent years,the development of visual foundation models and large language models has in⁃troduced a new paradigm for research on large models in remote sensing image processing.Remote sensing large models,also known as remote sensing foundation models,have garnered attention for their outstanding transfer performance in downstream tasks.These models are first pretrained on large datasets unrelated to spe⁃cific tasks and are then fine-tuned to adapt to various downstream applications.Foundation models have already been widely applied in language,vision,and other fields,and their potential in the field of remote sensing is in⁃creasingly gaining attention from the academic community.However,there is still a lack of comprehensive sur⁃veys and performance comparisons of these models in remote sensing tasks.Due to the inherent differences be⁃tween natural images and remote sensing images,these differences limit the direct application of foundation models.Against this backdrop,this paper provides a comprehensive review of common foundation models and large models specifically designed for the field of remote sensing from multiple perspectives.It outlines the latest advancements,highlights the challenges faced,and explores potential future directions for development.

关 键 词：遥感基础模型 微调 下游任务 预训练 

分 类 号：TP75[自动化与计算机技术—检测技术与自动化装置]