📝 Citation

If you use the Landsat30-AU dataset, or codebase in your research, please consider citing our AAAI 2026 paper.

Our work introduces a large-scale vision-language dataset bridging the gap in low-resolution, multi-satellite, long-term global monitoring. Built from 30-meter resolution imagery across Landsat 5, 7, 8, and 9 over Australia (spanning 36+ years), it includes:

• Landsat30-AU-Cap: 196,262 image-caption pairs.
• Landsat30-AU-VQA: 17,725 human-verified visual question answering samples across eight remote sensing domains.

BibTeX:

@article{Ma_Li_Taylor_2026, 
  title={Landsat30-AU: A Vision-Language Dataset for Australian Landsat Imagery}, 
  volume={40}, 
  url={[https://ojs.aaai.org/index.php/AAAI/article/view/37724](https://ojs.aaai.org/index.php/AAAI/article/view/37724)}, 
  DOI={10.1609/aaai.v40i10.37724}, 
  number={10}, 
  journal={Proceedings of the AAAI Conference on Artificial Intelligence}, 
  author={Ma, Sai and Li, Zhuang and Taylor, John A.}, 
  year={2026}, 
  month={Mar.}, 
  pages={7809-7817} 
}