Methods for the Automated Assignment and Comparison of Building Damage Geometries

Abstract

Condition assessment of structures supports effective repair, restoration or reconstruction works on buildings of infrastructural or cultural importance and is therefore an important contribution to building and infrastructure sustainability. Due to increasing ageing and external influences continuous monitoring is obligatory to maintain the structures. Throughout the entire life cycle, digital scans of changing damages and the building geometry, up to the degree of a digital twin, can be acquired using modern inspection methods. This allows for the automation of condition assessment processes to locate, quantify and predict damage progression at a structure related to building information models. However, the usual inspection documentation as performed in the past neither includes an assignment of damages to digital building models nor relates the history of recordings to a damage entity, which consequently makes a direct condition state comparison unfeasible.The paper presents voxel based methods for the automated assignment of geometrical representations of damages to elements of a corresponding building model as well as to past damage scans, which describe the same damage entity. Furthermore, the comparison and computation of discrete geometric changes based on the assignments is shown. Additionally, the clustering of scanned damage geometries into interest areas for the usage in planning processes for further inspections is presented. The resulting set of methods enables an automated condition comparison of damages on each surveyed state of the structure and supports the integration of inspection data to building information models.