Deep Learning in Astronomy
Deep Learning solutions for critical astronomical challenges: morphology and edge detection of the most distant galaxies observed by the Hubble telescope
PyTorch
Python
About the Project
This project represents a collaboration with the University of Sciences to address two critical challenges in modern astronomy through Deep Learning techniques.
The main objective is to develop artificial intelligence models capable of analyzing images of the most distant galaxies captured by the Hubble Space Telescope, providing accurate information about their morphology and detecting their edges with unprecedented precision.
Challenges Addressed
- Edge Detection: Precise identification of galaxy boundaries in high-resolution Hubble images
- Morphological Classification: Automatic analysis of the structure and shape of distant galaxies
- Large Volume Processing: Efficient analysis of extensive catalogs of astronomical images
Technologies Used
The project uses PyTorch to implement deep neural networks specialized in astronomical image processing, combining computer vision techniques with astronomical knowledge to achieve state-of-the-art results.
The Hubble Space Telescope
The Hubble Space Telescope (HST) is one of the most important scientific instruments ever built, launched into space in 1990 by NASA and ESA. Orbiting at approximately 547 kilometers above Earth, the Hubble has revolutionized our understanding of the universe by capturing images of extraordinary clarity, free from the atmospheric distortion that affects ground-based telescopes.
Relevance to This Project
The Hubble has been fundamental to this project by providing high-resolution images of the most distant and ancient galaxies in the universe. Its observations in the Deep Field and Ultra Deep Field have revealed thousands of galaxies that formed more than 13 billion years ago. These extraordinarily high-quality images are the perfect dataset for training and validating our Deep Learning models.
Interactive Demo
Work in Progress
We're working on an amazing interactive demo. Soon you'll be able to explore our galaxy catalog and observe how the AI model analyzes each image.
Thesis Document
View the complete thesis document that documents the entire research process, methodology, and results obtained in this Deep Learning project applied to astronomy.