A review of information service agents based on remote sensing multimodal foundation models

With the rapid advancement of remote sensing technologies, the acquired data has become increasingly multimodal, multisource, and multiscale, posing significant challenges to traditional processing and interpretation methods. In recent years, the emergence of artificial intelligence—particularly large-scale deep learning models (foundation models)—has opened new avenues for intelligent analysis of remote sensing data. This study presents the current research progress on information service agents powered by multimodal foundation models in the context of remote sensing. Based on existing research, this study proposes an information service agent framework that integrates multimodal large models, lightweight models and knowledge retrieval enhancement. This framework has broad prospects in a wide range of application areas such as disaster response, resource monitoring, and smart city development. This study aims to promote the evolution of remote sensing services from passive tools to active collaborative agents.