Vision-Language Models in Remote Sensing: Current Progress and Future Trends (2305.05726v2)

Abstract: The remarkable achievements of ChatGPT and GPT-4 have sparked a wave of interest and research in the field of LLMs for AGI. These models provide intelligent solutions close to human thinking, enabling us to use general artificial intelligence to solve problems in various applications. However, in remote sensing (RS), the scientific literature on the implementation of AGI remains relatively scant. Existing AI-related research in remote sensing primarily focuses on visual understanding tasks while neglecting the semantic understanding of the objects and their relationships. This is where vision-LLMs excel, as they enable reasoning about images and their associated textual descriptions, allowing for a deeper understanding of the underlying semantics. Vision-LLMs can go beyond visual recognition of RS images, model semantic relationships, and generate natural language descriptions of the image. This makes them better suited for tasks requiring visual and textual understanding, such as image captioning, and visual question answering. This paper provides a comprehensive review of the research on vision-LLMs in remote sensing, summarizing the latest progress, highlighting challenges, and identifying potential research opportunities.