Framework for automated UAS-based structural condition assessment of bridges

Abstract

Abstract This paper presents a coherent framework for automated unmanned aircraft system based inspections of large bridges to facilitate an automated condition assessment. Modern camera equipped unmanned aircraft systems are used to generate high-resolution digital image data of the structural surface. The flight path is automatically computed from a basic 3D model and ensures that the image set will satisfy defined quality parameters according to the desired information extraction. State-of-the-art photogrammetry and machine learning based feature detection methods are employed to automatically compute high-resolution geo-referenced 3D structural geometries and to identify typical damage patterns such as cracks. Further framework components dedicated to condition assessment allow the mapping of damages to structural parts and the calibration of mechanical numerical simulation models used to compute the internal structural demand under design loads. Data models are proposed that allow a consistent data storage and management to serve as a basis for all algorithmic components. The application of the framework to a large bridge structure showcases how the integration of digital systems and algorithms forms the basis for an intelligent and potentially autonomous safety assessment of very large infrastructures.