JAMES WEBB SPACE TELESCOPE STAR DETECTOR

On Tuesday, July 12, 2022, the first images from NASA's James Webb Space Telescope were released worldwide. Without doubt, this ushers in a new era for data analysis in astronomy, involving studies related to the physics of black holes, celestial bodies and even the origin of the universe.

Taking advantage of the unprecedented images obtained from the world's largest and most powerful space telescope, I was curious on how would it behave on one of the best, deep learning real-time detectors. I applied a YOLOv5, a single-stage deep learning object detector, built by Ultralytics on PyTorch, and with state-of-the-art accuracy.

This project was simply a basic approach, without the intention of optimization, but for pure exploration. It presents a number of opportunities for improvement, which I will list in the end. Still showed a very interesting initial approach, for star detection.