Real-time and image-based bridge inspection using augmented reality and artificial intelligence

Given the high demands of bridge maintenance in the network of federal highways in Germany, in particular bridge inspections are essential for assessing their condition. These hands-on inspections are largely conducted visually and documentation is largely manual using analog documentation methods. The results are a heterogeneous data collection and subjective decisions regarding the choice of measures.Modern technologies like artificial intelligence, augmented reality and BIM promise enormous potential to modernize this process by making it more objective, reliable and economical. This paper presents digital methods for bridge inspection based on the experiences from two research projects of the Federal Highway and Transport Research Institute (BASt) with Ruhr University Bochum and Hochtief Engineering GmbH Vicon including demonstrators developed with practitioners. These investigations focused on detection and localization of visual detectable defects, but also their connection to the digital twin bridge. Furthermore, an ongoing successor study on the chronology of image data sets and their predictive potential for structural damage is presented.