Abstract :
The dynamic nature of the cosmos is vividly illustrated by transient astronomical events, which include phenomena such as supernovae, gamma-ray bursts, or microlensing events. Detecting and identifying such events requires wide-field imaging, often conducted in suboptimal observing conditions from the ground.This presentation will explore the synergistic potential of video imaging and deep learning techniques in wide-field imaging of the deep sky, particularly for time-domain astronomy. I will demonstrate how such a combination can provide key benefits in many areas, from the ability to filter out artifacts to detecting transient events and monitoring the atmosphere. Emphasis will be placed on the mitigation of atmospheric turbulence, potentially bridging the gap between speckle holography and adaptive shift-and-add methods. I will present model prototypes trained with realistic image simulations that already provide some of these capabilities, and outline future directions for development.
