Euclid space telescope computer vision
AI & Data
Semester programme:Open Learning/Innovation
Client company:European Space Agency (ESA)
Project group members:Saamie Vincken
Amal Khairunnisa
Mardhiyyah Indarto,
incent van Gent
Cristian Măgureanu
Stan Rareș
Malwina Raczyńska
Project description
The Euclid Space Telescope is a mission by the European Space Agency (ESA) that maps one third of the observable universe to understand the properties of dark matter and dark energy. One of the main ways to study these is through strong gravitational lensing, which happens when the light from a distant galaxy is bent by the gravity of a massive object in front of it. By studying these effects, scientists can learn more about the distribution of dark matter and improve measurements of the Hubble Constant, which describes how fast the universe is expanding.
The purpose of this project is to develop a research-oriented software pipeline that supports ongoing scientific research into strong gravitational lensing using data from the Euclid Quick Data Release 1 (Q1). The developed pipeline will consist of several (beta-level) software features that address specific research questions formulated together with the SWG. These questions focus on improving data processing, machine learning analysis, and understanding systematic differences between simulated and real Euclid data.
Context
The project takes place in collaboration with the Euclid Strong Lensing Working Group (SWG), which forms part of the wider Euclid Consortium coordinated by the European Space Agency (ESA).
The project combines technical implementation with scientific exploration. With a main focus on data science and deep learning (ML) techniques. Its outcomes are intended to provide both practical tools and research insights for members of the Euclid Consortium, supporting future data releases and model development within the Strong Lensing Discovery Engine (SLDE).
Results
The project is currently between TRLevel 3 and 4. A lot of time was spent on research and understanding of the complex topics. With the most significant outcomes this semester being proof of concepts for:
• Image compression comparison: A pipeline feature that processes raw Euclid data and converts photometric color bands into an optimized compressed format.
• Color processing method: A pipeline feature that implements an improved color mapping method to create more accurate representations of the photometric bands captured by Euclid.
• Cross-mission image transformation:A pipeline feature that converts gravitational lens images from other missions into Euclid-like versions for use in machine learning training datasets.
4. SLDE catalogue classification: A pipeline feature that performs classification and visualization of subclasses within the SLDE catalogue to gain insight into their distribution and characteristics.
5. Unsupervised learning application: A proof of concept on applying unsupervised learning methods for grouping or clustering lens systems in the SLDE catalogue. Depending on the results, this may later be integrated into the pipeline as a feature.
6. Domain shift analysis: A research task to study the differences between simulated and real Euclid data and how this affect model performance. With a goal of gaining a better understanding on the impact of domain shifts and defining main issues in the current simulation pipelines of the SWG.
About the project group
A mixed group with students from different semesters and backgrounds within ICT. This project is very complex in data science and ML related topics, but even though this constraint, every person in the group (even the semester 2 students!) managed to excel. This has mostly been because of the support within the group and communication between members. We worked around 2 to 3 days a week on this project for one whole semester.