Structural health monitoring is an essential step to extend the life of buildings and infrastructure. A central part of structural health monitoring is repetitive visual tasks, such as inspections. Common tasks include taking pictures and analyzing them to detect potential damage and assess their impacts regarding the serviceability and safety of the structure. Such characteristics make structural health monitoring a perfect fit for artificial intelligence applications.
A central research focus at the group is automated detection of damage, especially cracks, for a cost-effective maintenance of concrete structures, such as bridges and buildings. Damage can be detected with the help of Artificial Intelligence, and images of damage can be analyzed live during a maintenance process. Classification of concrete damage can help maintenance workers when rating the condition of the structure. Moreover, image segmentation techniques can determine the size and extend of a crack to track the deterioration of a structure over time. Both would allow for autonomous inspections with higher frequencies, carried out by drones or autonomous robots.
Another related research field is the AI-based monitoring process for real-time localization and tracking workers and construction machinery. It aims to enhance monitoring of work site operations for the workers and safety administrators in order to reduce the risk of accidents. Localization is the process of recognizing and matching one's surroundings to a known model or to previously seen data. Autonomous robotics, indoor navigation, and augmented reality can all benefit from advanced localization procedures, which in turn improves the viability of these technologies for the construction industry. Artificial intelligence opens up new approaches to this problem, including pose regression, semantic recognition, and point cloud processing. Furthermore, a reliable and accurate localization could enable autonomous on-site inspections and improve safety assessment for buildings and infrastructure. We research various AI-supported methods for localization in tandem with new health monitoring and construction site safety processes at the chair.