• Machine learning
• Object detection
• Orbital debris
• Tracking
• Yolo

Widefield telescopes, that work in staring mode, are widely used for optical astronomical observations. In the observed images, the identification of moving objects, visible as linear features (streaks), is important for several reasons including the cataloging of space debris.
This thesis work consists in the design and development of a system based on machine learning approaches that identify these objects in real-time in the shortest possible time and returns their position within the image through bounding boxes.
This result will be useful for the realization of a tracking system that predicts the direction of movement of the debris and communicates with a second, reduced FOV, telescope that follows it.