Simuletic/Military-Drone-Swarm-Saturation-Attack-Dataset

🦟 Simuletic Drone Swarm & Saturation Attack Dataset

Synthetic Benchmark for High-Density Counter-UAS

Overview

This is an open-source synthetic dataset designed to solve the hardest problem in Counter-UAS (C-UAS): Saturation Attacks.

Most drone datasets feature single, large drones against clear skies. This dataset provides high-density drone swarms (5–15 units per frame) in complex, photorealistic environments. It is specifically engineered to benchmark Small Object Detection and Swarm Tracking algorithms where standard models fail due to object size, occlusion, and environmental noise.

---

🚀 Need more data?

This is a sample dataset by Simuletic. We provide hyper-realistic synthetic data to solve "edge cases" in Defense and Security AI—from FPV threats to adverse weather.
👉 Get full-scale datasets & video sequences: simuletic.com/datasets

---

✨ Key Features

• 🦟 High Density: Frames contain multiple distinct targets (Swarms), not just single objects.
• 🎯 Tiny Object Detection: Drones appear as small, distant threats (10–50 pixels), simulating real-world long-range radar/optical acquisition.
• ❄️ Adverse Weather: Includes rare conditions like heavy snow, rain, fog, and low-light "Golden Hour" scenarios to stress-test vision models.
• 🔒 Privacy-First: Fully synthetic. No flight restrictions were violated, and no sensitive real-world locations were filmed.

---

📂 Dataset Structure

The dataset follows the standard YOLOv8 / YOLO11 format.

• images/: High-fidelity synthetic .jpg files.
• labels/: .txt files containing class ID and normalized bounding boxes.

Class Map:

ID	Class Name	Description
0	drone	Includes quadcopters, FPVs, and loitering munitions

---

⚙️ YAML Configuration

To train immediately with YOLO, copy this into your data.yaml:

# Simuletic Drone Swarm Configuration
path: /path/to/dataset
train: images
val: images

nc: 1
names: ['drone']

Use Cases

    Counter-UAS (C-UAS): Train detection systems to handle "Saturation Attacks" where multiple bogeys appear simultaneously.

    Small Object Detection: Benchmark model performance on tiny, fast-moving objects in noisy backgrounds (rubble, forests, urban ruins).

    FPV Defense: Detect kamikaze/FPV drones that don't look like standard consumer quadcopters.

⚖️ Ethics & License

    Synthetic Nature: This data is 100% computer-generated by Simuletic. It is free from GDPR concerns and export control restrictions (ITAR/EAR).

    License: CC BY 4.0. You are free to use and adapt this data for research or commercial proofs-of-concept, provided you give appropriate credit to Simuletic.

📝 Citation

If you use this dataset in your research, please cite:
Kodavsnitt

@dataset{simuletic_drone_swarm_2025,
  author = {Simuletic Team},
  title = {Simuletic Synthetic Drone Swarm & Saturation Attack Dataset},
  year = {2025},
  url = {[https://simuletic.com](https://simuletic.com)}
}

Feedback? Reach out via simuletic.com or the "Issues" tab here on Kaggle.