GeoLLaVA: Efficient Fine-Tuned Vision-Language Models for Temporal Change Detection in Remote Sensing

TL;DR

GeoLLaVA introduces an efficient fine-tuning approach for vision-language models to accurately detect and describe temporal changes in remote sensing imagery, aiding environmental and urban planning.

Contribution

The paper presents a novel dataset and applies advanced fine-tuning techniques to improve vision-language models' ability to analyze temporal geographical changes.

Findings

Achieved a BERT score of 0.864 in change detection.

Improved ROUGE-1 score to 0.576 for land-use descriptions.

Enhanced model performance on remote sensing temporal change tasks.

Abstract

Detecting temporal changes in geographical landscapes is critical for applications like environmental monitoring and urban planning. While remote sensing data is abundant, existing vision-language models (VLMs) often fail to capture temporal dynamics effectively. This paper addresses these limitations by introducing an annotated dataset of video frame pairs to track evolving geographical patterns over time. Using fine-tuning techniques like Low-Rank Adaptation (LoRA), quantized LoRA (QLoRA), and model pruning on models such as Video-LLaVA and LLaVA-NeXT-Video, we significantly enhance VLM performance in processing remote sensing temporal changes. Results show significant improvements, with the best performance achieving a BERT score of 0.864 and ROUGE-1 score of 0.576, demonstrating superior accuracy in describing land-use transformations.