Three-dimensional reconstruction and classification of high-resolution optical satellite data.

Operation, planning and protection of infrastructure requires 3D geoinformation of high semantic, temporal and geometric quality. This cannot be recorded over a large area, in remote or dangerous areas, and with high temporal resolution using terrestrial or airborne methods. Likewise, there is no automatic and efficient processing and interpretation for road planning, pipelines and evacuation applications. ReKlaSat 3D makes high-resolution satellite images available through research and development in 3D processing, semantic classification/machine learning and change detection. The participating project partners WU and TU Vienna, Vermessung Schmid, Siemens and BMLVS bring in complementary competencies to cover the entire chain from satellite image acquisition to the interpretation of status and change.

Artificial Intelligence (Machine Learning and Deep Learning) is used to efficiently process satellite images for various use cases like road planning, pipelines, as well as evacuation applications.