MykolaL/DelineateAnything

Model Card for Delineate Anything: Resolution-Agnostic Field Boundary Delineation on Satellite Imagery

by Mykola Lavreniuk, Nataliia Kussul, Andrii Shelestov, Yevhenii Salii, Volodymyr Kuzin, Sergii Skakun, Zoltan Szantoi

Delineate Anything Flow (DelAnyFlow) is a resolution-agnostic methodology for fast, country-level agricultural field boundary detection from satellite imagery. It utilizes the DelAny instance segmentation model, which is based on a YOLOv11 backbone and trained on the large-scale Field Boundary Instance Segmentation-22M (FBIS 22M) dataset.

DelAny delivers state-of-the-art accuracy, showing significantly higher mAP and faster inference than alternatives like SAM2. It supports national-scale applications, having been used to generate a complete field boundary layer for Ukraine (603,000 km²) in under six hours.

Paper

Delineate Anything Flow: Fast, Country-Level Field Boundary Detection from Any Source

Citation

@misc{lavreniuk2025delineate,
      title={Delineate Anything Flow: Fast, Country-Level Field Boundary Detection from Any Source}, 
      author={Mykola Lavreniuk and Nataliia Kussul and Andrii Shelestov and Yevhenii Salii and Volodymyr Kuzin and Sergii Skakun and Zoltan Szantoi},
      year={2025},
      eprint={2511.13417},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2511.13417}, 
}