Learning Antenna Pointing from Video Data

Abstract

The ground station Neustrelitz is a critical component of DLR's earth observation payload data ground segment. To ensure high reliability, the facility combines modern infrastructure, redundancy measures, comprehensive monitoring, automated processes, and expertise in manual configuration and development.

As part of an advanced monitoring setup, we aim to develop an automated camera-based surveillance system for robust antenna pointing derivation using machine learning. By leveraging and adapting existing approaches, we seek to create a prototype that delivers precise antenna pointing from image data optimized for our specific use case.

This talk provides an overview of the problem, proposed methodology, and current status of this work-in-progress project.