drone-detector

Drone Detector with YOLO11n

This model was originally made using YOLO8n for a project at my uni, but I have since trained it for longer, and I've upgraded the base model to YOLO11.

• Model Architecture: YOLOv11n
• Task: Object Detection (Drone Detection)
• Image Type: Dataset from Maciej Pawełczyk and Marek Wojtyra
• Classes: 1 (drone)

Training Validation Results

Training and Validation Losses

Confusion Matrix

Precision-Recall Curve

F1 Score Curve

Training Configuration

• Model weight file: yolo11n_drone.pt
• Base model: ultralytics/yolov11n
• Number of Epochs: 150
• Image Size: 640x640

Deployment

How to use this model

from ultralytics import YOLO

# Load the model
model = YOLO("yolo11n_drone.pt")

Limitations

Honestly, this model is good at recognizing drones, but it struggles with limiting the size of the box where it thinks it is. You can train it further if you want, but ¯\_(ツ)_/¯. From experience, I can say that it struggles a bit at a distance, but it works better than I thought it would.

Disclaimer

I would like to direct all possible credit to Pawełczyk and Wojtyra for their dataset. I am however in no way, shape or form related to them in any function. Here is the link if you want to cite their paper. Here is a link to the Github repo where I originally found their dataset.