QuailModel sample: QM Synthetic Airspace: Drone vs Fixed-wing vs Bird

README.md

@@ -0,0 +1,46 @@
+---
+license: cc-by-nc-sa-4.0
+task_categories:
+- object-detection
+tags:
+- drone
+- uav
+- counter-uas
+- bird
+- synthetic
+- yolo
+pretty_name: "QM Synthetic Airspace: Drone vs Fixed-wing vs Bird"
+size_categories:
+- n<1K
+---
+# QM Synthetic Airspace: Drone vs Fixed-wing vs Bird - Free Sample
+
+Synthetic **airspace** images with pixel-exact YOLO boxes for three classes that counter-UAS systems must tell apart:
+**multirotor** drones (4/6/8 arms, 0.12-1.4 m), **fixed-wing** UAVs (0.8-3.5 m span) and **birds** (0.25-1.6 m).
+Rendered by QuailModel with Blender/Cycles from fully procedural, original geometry over physical skies with trees,
+buildings and cloud layers - no photos, no scraped data, no people.
+
+## Contents
+| split | images | boxes |
+|---|---|---|
+| train | 240 | 384 |
+| val | 34 | 44 |
+| test | 26 | 45 |
+
+- Image size: 640x640 JPEG. Labels: YOLO txt, classes: 0 = multirotor, 1 = fixed_wing, 2 = bird. Images with an empty label file are true negatives (33 of 300).
+- Boxes per class: multirotor 241, fixed_wing 119, bird 113
+- `data.yaml` included - train directly with Ultralytics YOLO.
+- Box size distribution (fraction of image width): median 0.060, 26% of boxes are smaller than 16 px.
+
+## How it was made
+- Procedural generator: datasets/air3 (QuailModel original code).
+- Labels from exact projected mesh vertices; occlusion checked by ray casting; objects under 8 px dropped.
+- Train/val/test split 80/10/10 by file-name hash.
+
+
+## Licence
+Sample edition: CC BY-NC-SA 4.0 (non-commercial). The full commercial edition (5000 images, commercial licence) is available from QuailModel: https://huggingface.co/quailrobot
+
+## Disclosure
+Generated with AI assistance: the procedural generator code was written with an AI model and every
+image is computer-rendered. Synthetic data should be validated on real footage before deployment.

data.yaml

@@ -0,0 +1,8 @@
+path: .
+train: images/train
+val: images/val
+test: images/test
+names:
+  0: multirotor
+  1: fixed_wing
+  2: bird