ABSTRACT
To ensure the structural safety and serviceability is a top priority for large infrastructures. To determine the load bearing capacity for the current state and to predict the structural behaviour for the future life span, regular inspections and evaluations of the structural condition are required. Recently, many different approaches using modern digital technologies were developed to support the engineer with the acquisition and interpretation of structural information. This article focuses on the evaluation of visual data acquired by using small unmanned aircraft systems (UAS) equipped with high-quality cameras. The images obtained are used amongst others to reconstruct a dense 3D point cloud, which contains local (e.g. cross sectional shape, spalling areas) as well as global (e.g. bending line with respect to a reference condition or to earlier observation) geometric information. Furthermore, surface anomalies (e.g. cracks), automatically detected by image analysis algorithms, are mapped to the 3D structure. This paper presents a methodology, which processes and combines the acquired data from consecutive UAS-based inspections to quantify identifiable parameters usable for a validation or calibration of a numerical model. This parameters cover for example cross sectional values, displacements and curvature at each position of the element axis as well as magnitudes of load and effective bending stiffness. The introduced procedure systematically shows on different structural levels and based on the number of data acquisitions and available a priori information about the structure which parameters are identifiable and how they are obtained. The quantification is performed by a combination of mathematical optimization strategies with linked cross sectional and system analysis. Case studies on a laboratory experiment and a semi-integral bridge are used to show the functionality and potential of the proposed method.
