/Military-Vehicles-Image-Recognition

Fine-Tuning YOLOv5 to detect Military Vehicles in Aerial ARMA 3 Imagery

Primary LanguagePythonGNU General Public License v3.0GPL-3.0

Military Vehicles Image Recognition

Fine-Tuning YOLOv5 to detect Military Vehicles in Aerial ARMA 3 Imagery

Compound

Surveillance gif

A Youtube devlog of the project is available here:

Youtube Devlog

Currently, the model is able to detect the following classes:

  • CSAT Varsuk
  • CSAT Marid
  • CSAT Zamak (Transport)

The model is a YOLOv5 fine tuned using 100 images of each class using various environments and angles at noon clear sky using a UAV at around 100 meters altitude.

The dataset used is available on Kaggle.

How to use

  1. Clone the repository
git clone https://github.com/AlexandreSajus/Military-Vehicles-Image-Recognition.git
  1. Install the requirements
pip install -r requirements.txt

  1. Add your images to the input folder

  2. Run the model

python detect.py --source ./input/ --weights runs/train/yolo_arma4/weights/best.pt --conf 0.5 --name yolo_arma
  1. The results will be available in a runs/detect/yolo_armaX folder