Total: 1
Astronomical surveys and the growing involvement of amateur astronomers are producing more sky images than ever before, and this calls for automated processing methods that are accurate and robust. Detecting Deep Sky Objects -- such as galaxies, nebulae, and star clusters -- remains challenging because of their faint signals and complex backgrounds. Advances in Computer Vision and Deep Learning now make it possible to improve and automate this process. In this paper, we present the training and comparison of different detection models (YOLO, RET-DETR) on smart telescope images, using High-Performance Computing (HPC) to parallelise computations, in particular for robustness testing.