This project focuses on developing a drone detection model using the Detectron2 framework. The model is trained to accurately identify and locate drones within images and videos.
Data Annotation: Utilizes Labelme for annotating drone images. Model Training: Leverages Google Colab for training on GPU-accelerated servers. Testing: Evaluates model performance on the testing dataset.
Python
Detectron2
Labelme
Google Colab account (for training)
Expand the dataset with more diverse drone images.
Experiment with different model architectures and hyperparameters.
Improve real-time performance for practical applications.
Github
Demo Video
Project Idea
Project Presentation
Arnav Prasad