Photogrammetric Computer Vision in Remote Sensing

Many applications in the field of Remote Sensing can only be solved by incorporating methodologies from the fields of Photogrammetry and Computer Vision. Thus, this habilitation thesis presents significantly advances in those three fields, namely Photogrammetry, Computer Vision, and Remote Sensing. The main aim is to develop enhanced methods from Photogrammetry and from Computer Vision that are then implemented in various Remote Sensing applications. Therefore, this work is assigned to the scientific subject called Remote Sensing. Major topics under investigation are sensor modeling, epipolar geometry, and radargrammetry (from Photogrammetry), and image matching, 3D object detection, and object density estimation (from Computer Vision). Those technologies enable numerous applications from the domain of Remote Sensing, where this thesis focuses on digital surface model generation, ortho-rectification, digital surface model fusion, digital terrain model generation, forest assessment, landslide mapping, ground control point retrieval, and disaster control. In summary, a novel field of research called Photogrammetric Computer Vision in Remote Sensing is presented that bridges the three interdisciplinary topics of interest. Various improvements are proposed in all those genres which is additionally illustrated by more than 100 publications of the author.