After successful completion of this course, the student should be able to
- Understand and apply the main steps of the 3D scene analysis related to UAV data;
- Understand and apply the main steps of the semantic scene analysis;
- Design innovative 2D and 3D scene analysis algorithms to extract useful information from images, with specific regards to UAV data;
- Evaluate the geometric and semantic quality of the results achieved using these algorithms;
- Report and critically “discuss” the scientific outcomes by providing relevant referencing.
- The algorithms for 3D reconstruction: Structure from Motion, Simultaneous Localization and Mapping, Visual Odometry, dense image matching, 3D model generation, flight path planning;
- The semantic aspects of scene understanding: semantic segmentation, object detection and tracking;
- The modern deep learning concepts and algorithms;
- The use of UAVs for 3D indoor and outdoor mapping and semantic understanding;
- The development of algorithms and technical solutions to efficiently process real case data for the considered practical applications using small UAVs.