Surface Monitoring at Dams Using Drones and AI-Based Analysis for Advanced Anomaly Assessment

The growing request of renewable energies, in particular hydro power plants, in the last century led to a large number of built dams, which are used to store water in order to generate renewable energy. Due to their age of about 50 to 100 years many dams demand an extensive monitoring of the structures to ensure a safe operation as well as to identify potential problems as early as possible in order to minimize maintenance costs. Therefore, new strategies need to be explored to bring the surface analysis and documentation of large concrete dams to a “next level”. One approach that has emerged is to use drones for visual inspection by capturing high-resolution images of the dam and using artificial intelligence to analyse the huge amount of data. The aim of the master thesis is to examine and develop a suitable method in particular a process that supports dam operators by a drone-based surface documentation and inspection with a suitable evaluating algorithm reducing the manual effort. High quality images of the recorded dams are necessary to identify abnormalities such as cracks, sintering, corrosion, bursting on the dam by automated image analysis using artificial intelligence. To provide a further processable 3D visualization of the dam as well as an overlay with the findings, a 3D model of the dam is generated by a photogrammetric image analysis.