Grundlagen zur automatisierten Ermittlung hoch­genauer Passpunkte für CubeSat-Satellitenbilder mittels Deep-Learning-gestützter Delaunay-Triangulation basierend auf Gebäudedaten des Liegenschaftskatasters

A solution for georeferencing satellite images from line-scan cameras of miniaturized satellites, the so-called CubeSat satellites, with a low pointing accuracy is presented, which allows an automated control point determination in post-processing. Core element of this method is a two-dimensional approach to find similarities between triangulations of aggregated settlement data and consequentially derive building centres as key points. The underlying functional and stochastic model are explained. The proposed solution is demonstrated using a simulator with true digital orthophotos (TrueDOP) and two satellite scenes.     «

A solution for georeferencing satellite images from line-scan cameras of miniaturized satellites, the so-called CubeSat satellites, with a low pointing accuracy is presented, which allows an automated control point determination in post-processing. Core element of this method is a two-dimensional approach to find similarities between triangulations of aggregated settlement data and consequentially derive building centres as key points. The underlying functional and stochastic model are explained...     »