MoBaP - Model-based digital Inspection of Structures
MoBaP is the acronym for model-based digital Inspection of Structures, which is a research project funded by the Bavarian Ministry of Economic Affairs, Regional Development and Energy's program for information and communication technology. The team of MoBaP consists two industrial partners: Ilp2 and K3, and the Bundeswehr University Munich as academic partner.
The project’s goal is to develop a product that serves as a reliable toolkit during the inspection of built structures and the design of building surveys by automatizing important elements. There are four main components to name: modelling a digital twin, recognizing damages, localizing the defects and generating their evaluation.
The Bundeswehr University’s main aim is the classification and measurement of the damages by utilizing deep learning models, mainly Convolutional Neural Networks. Currently, the focus is on defects occurring on reinforced concrete bridges exclusively (e.g. cracks, spalling, scaling and efflorescence), whereby further damage classes will follow.
The most recent multi-target damage classifier -called DACL- performs superb. At present, the DACL shows an accuracy of approximately 90%.
Confident classification of a crack by the DACL
Stay tuned for more an the latest information about the MoBaP project on LinkedIn.