Datasets:

Spatiallysaying
/

rwy_obb-300-65-65_v2

license: mit
task_categories:
  - image-segmentation
tags:
  - computer-vision
  - segmentation
  - runway
  - airport
  - aerial
  - satellite
size_categories:
  - 1K<n<10K

Runway Object-Oriented Bounding Box (OBB) Segmentation Dataset v2

Dataset Description

This dataset contains aerial/satellite imagery for runway segmentation using oriented bounding boxes (OBB). The dataset is specifically designed for semantic segmentation tasks focusing on runway detection and delineation in aerial imagery.

Dataset Statistics

Split	Images	Labels
Train	303	303
Val	65	65
Test	65	65

Classes

{
  "0": "_background_",
  "1": "rwy_obb"
}

Usage

from datasets import load_dataset

dataset = load_dataset("Spatiallysaying/rwy_obb-300-65-65_v2")

# Load class mappings
import json
with open('id2label.json', 'r') as f:
    id2label = json.load(f)
with open('label2id.json', 'r') as f:
    label2id = json.load(f)

Data Format

Images: RGB images in JPEG format
Labels: Grayscale segmentation masks in PNG format
- Pixel values correspond to class IDs (0=background, 1=runway)
Resolution: Variable (original image dimensions preserved)

Citation

@dataset{rwy_obb_300_65_65_v2,
  title={Runway Object-Oriented Bounding Box Segmentation Dataset v2},
  author={Spatiallysaying},
  year={2026},
  url={https://huggingface.co/datasets/Spatiallysaying/rwy_obb-300-65-65_v2}
}

License

MIT License