Spectral Classification of Chandrayaan-2 IIRS Data

Project Overview

This project addresses the spectral classification of Chandrayaan-2 Imaging Infrared Spectrometer (IIRS) data using AI/ML techniques to enhance our understanding of the Moon's geological diversity. Developed for the ISRO Bharatiya Antariksh Hackathon 2024, our solution combines advanced data processing, machine learning, and visualization to analyze lunar spectral data.

Key Features

• Holistic integration of spectral analysis with spatial visualization
• Accurate geolocation for precise mapping of lunar features
• User-friendly GUI for data exploration and analysis
• Comprehensive toolset utilizing QGIS, MATLAB, Python, and Orfeo Toolbox

Technical Approach

1. Data Processing: Acquisition and preparation of IIRS data from ISSDC
2. Spatial Analysis: Overlay of IIRS data on lunar basemap using MATLAB
3. Spectral Analysis: Extraction and plotting of spectral profiles
4. Machine Learning: Application of CNNs for spectral classification
5. Visualization: Integration of results using MATLAB, Python, and QGIS

Technologies Used

• Python (NumPy, Pandas, SciPy, scikit-learn, PyQt)
• MATLAB (Image Processing Toolbox, Statistics and Machine Learning Toolbox)
• QGIS for geospatial visualization
• TensorFlow/Keras for deep learning models

Impact and Applications

• Provides insights into lunar geological features and mineral composition
• Demonstrates the effectiveness of ML in processing hyperspectral data

For more details on the Bharatiya Antariksh Hackathon 2024, visit (https://isro.hack2skill.com/2024/).