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

TL;DR

This work tackles automated agricultural field boundary delineation from satellite imagery by reframing the task as instance segmentation and introducing FBIS-22M, a multi-resolution dataset with over 672k images and 22.9M field masks. They present Delineate Anything, a resolution-agnostic instance segmentation model trained on FBIS-22M, achieving state-of-the-art performance with $mAP@0.5=0.720$ and $mAP@0.5:0.95=0.477$, while being substantially faster than prior methods and exhibiting strong zero-shot generalization across unseen geographies. The dataset’s diversity in resolutions ($0.25$m–$10$m) and sensor sources, along with the instance-level framing, addresses key limitations of previous semantic segmentation approaches and SAM-based zero-shot methods. Collectively, this approach enables scalable, accurate field boundary mapping suitable for large-scale land administration and precision agriculture applications, with potential for country- or region-level deployment.

Abstract

The accurate delineation of agricultural field boundaries from satellite imagery is vital for land management and crop monitoring. However, current methods face challenges due to limited dataset sizes, resolution discrepancies, and diverse environmental conditions. We address this by reformulating the task as instance segmentation and introducing the Field Boundary Instance Segmentation - 22M dataset (FBIS-22M), a large-scale, multi-resolution dataset comprising 672,909 high-resolution satellite image patches (ranging from 0.25 m to 10 m) and 22,926,427 instance masks of individual fields, significantly narrowing the gap between agricultural datasets and those in other computer vision domains. We further propose Delineate Anything, an instance segmentation model trained on our new FBIS-22M dataset. Our proposed model sets a new state-of-the-art, achieving a substantial improvement of 88.5% in mAP@0.5 and 103% in mAP@0.5:0.95 over existing methods, while also demonstrating significantly faster inference and strong zero-shot generalization across diverse image resolutions and unseen geographic regions. Code, pre-trained models, and the FBIS-22M dataset are available at https://lavreniuk.github.io/Delineate-Anything.

Figures (5)

• Figure 1: Workflow of the Delineate Anything model for field instance segmentation and field boundary extraction from arbitrary resolution satellite imagery, trained on our large-scale Field Boundary Instance Segmentation dataset (FBIS-22M), containing 22M field boundaries.
• Figure 2: Comparison of task formulations and evaluation metrics for field boundary delineation. The top row illustrates field boundary masks (semantic segmentation), while the bottom row shows individual field masks (instance segmentation). Ground truth examples are shown in (a) and (d). Slightly misaligned boundaries result in a boundary IoU of 0.08 (b) and an instance IoU of 0.98 (e). Partially detected boundaries yield a boundary IoU of 0.93 (c) and an instance IoU of 0.54 (f).
• Figure 3: Examples of field boundary instance segmentation from our FBIS-22M dataset. The FBIS-22M dataset contains over 670K+ multi-resolution satellite images (ranging from 0.25m to 10m) and 22M+ field instance masks. Images are grouped by the number of fields to demonstrate the dataset's diversity and scalability, and a challenge of separating fields across varying resolutions and geographies.
• Figure 4: Qualitative results on the FBIS-22M test set. Delineate Anything is compared to MultiTLF kerner2024multi, SAM kirillov2023sam, and SAM2 ravi2024sam2. For a fair comparison, the MultiTLF model was retrained using our FBIS-22M dataset. Different samples are carefully selected and presented, varying in the size and density of the fields, to better illustrate the performance of each model under diverse conditions.
• Figure 5: Qualitative results of zero-shot predictions. Delineate Anything is applied to geographic regions with different climates, terrains, and agricultural practices, highlighting its field boundary delineation capabilities outside the training data.