🌌 CosmicQuakes

Seismic Event Detection on the Moon Using Signal Processing and Deep Learning

CosmicQuakes is a scientific project that focuses on detecting seismic events from Apollo 12 lunar mission data using advanced signal processing techniques and deep learning architectures. It aims to improve the efficiency of data transmission in planetary missions by identifying and transmitting only meaningful seismic events from extraterrestrial environments like the Moon.

---

🚀 Overview

Planetary missions often collect vast amounts of seismic data, much of which may be noise or unimportant. Due to limited bandwidth and energy constraints, it’s inefficient to send all this data back to Earth. CosmicQuakes proposes an intelligent system that performs event detection on-device, sending only scientifically valuable segments.

---

🔍 Features

• 📊 Signal Preprocessing with:
  • Bandpass Filtering (0.4 – 1.2 Hz)
  • Empirical Mode Decomposition (EMD)
  • STA/LTA event triggering
• 🔬 Feature Extraction:
  • Time-frequency analysis using Fourier Transform (FFT)
  • Energy-based IMF decomposition
• 🧠 Deep Learning:
  • CNN-based regression model to predict seismic event timing
  • Optional LSTM/Autoencoder extensions
• 📈 Performance Validation:
  • Compared against classic STA/LTA methods
  • Tested on real Apollo 12 seismic data

---

🗂️ Repository Structure

CosmicQuakes/
├── data/             # Raw and processed lunar seismic data (CSV/NPZ)
├── models/           # Trained model weights and outputs
├── notebooks/        # Jupyter notebooks for analysis and visualization
├── src/              # Source code for preprocessing, modeling, evaluation
├── tests/            # Unit tests and performance evaluation scripts
├── requirements.txt  # Python dependencies
└── README.md         # Project documentation (this file)

---

🧪 Technologies & Libraries

• Python (3.10+)
• TensorFlow & Keras
• NumPy, Pandas, SciPy
• Matplotlib & Seaborn
• scikit-learn
• PyEMD

---

⚙️ Installation

Clone the repository:

git clone https://github.com/sedefkjamili/CosmicQuakes.git
cd CosmicQuakes
pip install -r requirements.txt

---

🧠 Model Training & Evaluation

Train the CNN model:

python src/train_cnn.py

Evaluate the model:

python src/evaluate_accuracy.py

Visualize events and performance metrics using Jupyter Notebooks inside /notebooks.

---

📚 Data Source

• NASA Space Apps 2024 – Seismic Detection Challenge Resources
• Apollo 12 passive seismic experiment datasets

---

👩‍🔬 Authors

• İlayda Öcal
• Sedef Kjamili
  Ankara University - Computer Engineering Department

---

📄 License

This project is for academic and research purposes. All Apollo data is courtesy of NASA’s public archives.

---

✨ Acknowledgments

Special thanks to Assoc. Prof. Dr. Yılmaz Ar for supervising this graduation project as part of BLM4061.