RSICCLLM: A Multimodal Large Language Model for Remote Sensing Image Change Captioning
2026-06-26 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors focus on describing changes between two remote sensing images, which is useful for many applications. They note that existing methods are limited because they use smaller models and struggle with detailed change detection. To improve this, the authors created a new system called RSICCLLM that uses large vision-language models tailored for this task by generating special training data and techniques to better recognize changes. Their experiments show that their smaller 7-billion-parameter model outperforms much larger ones. They also plan to share their code and datasets publicly.
Remote SensingImage Change CaptioningVision-Language ModelsPost-TrainingFine-TuningDifference RepresentationNegative SamplingBenchmark DatasetLarge ModelsTemporal Change Detection
Authors
Yelin Wang, Zijia Song, Shuo Ye, Chuanguang Yang, Miaoyu Wang, Yong Xu, Zhulin An, Yongjun Xu, Zitong Yu
Abstract
Remote Sensing Image Change Captioning (RSICC) aims to describe changes between bi-temporal remote sensing images and holds significant research and application value. However, most existing methods rely on conventional deep learning architectures, and the limited model capacity constrains performance. Although large-model post-training techniques have achieved great success in general domains, their direct transfer to RSICC remains challenging due to data scarcity and the need for fine-grained change understanding. To address this, we propose RSICCLLM, the first post-training framework for large vision-language models in RSICC. Specifically, we design a data generation paradigm, release the instruction dataset RSICI, and establish a task-specific RSICC benchmark. We further introduce Difference-aware Supervised Fine-tuning to explicitly extract change representations and guide the model in perceiving and understanding temporal differences. In addition, we propose Dual-Negative Preference Optimization (DNPO), which employs two complementary negative-sample construction strategies to construct the preference dataset RSICP and further refine model performance. Extensive experiments validate the superior capability of RSICCLLM, which achieves outstanding results with only 7B parameters, surpassing models of substantially larger scales. The code and dataset will be made publicly available at https://github.com/keaill/RSICCLLM.