Drone Detector

High-precision YOLOv8 model for detecting drones, UAVs, birds, and airplanes with 96.1% mAP50 accuracy.

Motivation

Modern conflicts have demonstrated the critical importance of aerial threat detection. Recent events, including the 12-day Iran-Israel conflict in June 2025, highlighted how drone warfare has become a decisive factor in modern military operations. During this period, Iran's infrastructure faced significant challenges from UAV attacks, with multiple critical facilities being targeted by advanced drone systems.

This model was developed to address the urgent need for:

• Early Detection: Identifying aerial threats before they reach critical infrastructure
• Accurate Classification: Distinguishing between friendly aircraft, birds, and hostile drones
• Real-time Response: Enabling rapid counter-drone measures
• Civilian Protection: Safeguarding airports, power plants, and urban areas

The devastating impact of drone strikes on civilian and military targets underscores the necessity for robust, AI-powered detection systems that can operate 24/7 with minimal human intervention.

Performance

• Precision: 94.1%
• Recall: 94.1%
• mAP50: 96.1%
• mAP50-95: 67.4%

Detectable Classes

• Drone
• Bird
• UAV
• Airplane

Quick Start

from ultralytics import YOLO

# Load model
model = YOLO('drone_detector.pt')

# Detect in image
results = model('image.jpg')
results[0].show()

# Detect in video
results = model('video.mp4')

Installation

pip install ultralytics

Sample Detections

Author

Mohammad Rasol Esfandiari

• GitHub: @DeepPythonist
• Email: mrasolesfandiari@gmail.com

License

MIT License - see LICENSE file for details.