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.
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.
- 📊 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
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)
- Python (3.10+)
- TensorFlow & Keras
- NumPy, Pandas, SciPy
- Matplotlib & Seaborn
- scikit-learn
- PyEMD
Clone the repository:
git clone https://github.com/sedefkjamili/CosmicQuakes.git
cd CosmicQuakes
pip install -r requirements.txtTrain the CNN model:
python src/train_cnn.pyEvaluate the model:
python src/evaluate_accuracy.pyVisualize events and performance metrics using Jupyter Notebooks inside /notebooks.
- NASA Space Apps 2024 – Seismic Detection Challenge Resources
- Apollo 12 passive seismic experiment datasets
- İlayda Öcal
- Sedef Kjamili
Ankara University - Computer Engineering Department
This project is for academic and research purposes. All Apollo data is courtesy of NASA’s public archives.
Special thanks to Assoc. Prof. Dr. Yılmaz Ar for supervising this graduation project as part of BLM4061.