Automated path planning for UAS based building inspection
Project information
Submitted by
Paul Debus
Co-AuthorsPaul Debus
MentorsProf. Volker Rodehorst, Prof. Guido Morgenthal, Norman Hallermann, Jens Kersten
Winter semester 2017/2018
Faculty:
Civil and Environmental Engineering,
Media
Degree programme:
Civil Engineering (Master of Science (M.Sc.))
Master project
Project description
The chairs of Computer Vision in Engineering and Modeling and Simulation of Structures develop applications of digital methods in condition monitoring of buildings.UAS offer opportunities for the capturing of large structures, as they can reach places inaccessible for humans. The effort for data acquisition is reduced and the results are documented, allowing to not only consider the condition at one point but also the change over time.
The UAS are equipped with cameras and capture the entire building. 3D models are computed from the photos, supporting the analysis of the building condition. To fulfill the requirements for precision and detail, the camera positions are planned carefully. Modern UAS can conduct pre-planned flight missions precisely, better than human pilots. The camera positions are planned such that adjacent images overlap for the reconstruction of the model.
The task for the masters project was to compute suitable flight routes based on available geometric information of the object from existing digital models. Different methods have been developed and tested at the reference object, the school "An der Hart" in Weimar. The results have been used to adapt the methods.
Using the developed methods, the inspection of arbitrary building shapes is possible with high detail and resolution. The resulting models provide new possibilities in condition assessment and documentation over the entire life cycle.