Solar Radio Burst Classification

Classification of solar radio bursts from spectrogram images using transfer learning models.

Overview

This project classifies solar radio burst spectrograms into three categories:

• Type II: Slow-drifting bursts associated with CMEs
• Type III: Fast-drifting bursts from electron beams
• Empty: Background spectrograms without bursts

Setup

# Clone repository
git clone https://github.com/Hermanlrx/SRBClassificationCodeAndData.git
cd SRBClassificationCodeAndData

# Create conda environment
conda create -n solar-bursts python=3.9
conda activate solar-bursts

# Install dependencies
conda install tensorflow keras opencv matplotlib pandas numpy scikit-learn jupyter
pip install ultralytics  # For YOLO

Usage

Navigate to any model folder and run the corresponding Jupyter notebook:

jupyter notebook "Code and results of each model/1.Comparison Transferlearning Densenet201/1. Model+code/TransferLearningTest.ipynb"

Raw data files

Zenodo