Dataset · yaraalaa0
Synthetic aerial imagery with depth and segmentation maps
yaraalaa0/TopAir
Bird's-eye view frames from the AirSim simulator for depth estimation and segmentation.
What's inside
Contains between 10,000 and 100,000 top-down nadir aerial frames captured in rural and urban environments using the AirSim simulator. Each image is paired with a depth map, semantic segmentation map, and external camera coordinates. Collected across various flight trajectories and published by MaLga Lab.
What you can do with it
- Training monocular depth estimation models for drones
- Fine-tuning semantic segmentation algorithms for aerial imagery and terrain maps
- Testing UAV visual odometry and navigation systems
- Tasks
- Depth estimation, Image segmentation
- Size
- 10K–100K rows
- Downloads this month
- 46K
- Likes
- 4
How to load it
from datasets import load_dataset
ds = load_dataset("yaraalaa0/TopAir")