(Semi-)Autonomous creation of as-built BIM models in bridge construction using drones

BIM-based project management is currently still under development in infrastructure and engineering construction. Little research has been done to date in particular in the realization phase as to how information on the structure can be digitally acquired and matched with building information models (BIMs) in an automated process. The presentation introduces a concept for autonomous drone-based data acquisition, the analysis of point clouds and automated matching with 4D BIMs. The concept includes data acquisition planning based on the 4D BIM, evaluation of the acquired point clouds and matching with the 4D BIM. All sub-steps are to be performed in an automated process to enable real-time information to be provided in short cycles.

Based on the 4D BIM, relevant building elements are identified and used to derive waypoints for the drone-based data acquisition. This is followed by automated planning of the flight path, taking into account the acquisition method and potential obstacles. Data acquisition is performed by means of autonomously operating drones. The acquired point cloud is evaluated geometrically. In addition, objects are recognized directly in the point cloud using artificial intelligence methods and machine learning algorithms. The information acquired is translated into as-built BIMs. With the updated as-built BIMs, disruptions in the construction process and discrepancies in component quality can be identified at an early stage. This allows effective control of digital planning and construction processes.