posted on 2020-03-25, 16:48authored byRuodan Lu, Ioannis Brilakis
This paper describes the design, implementation and benchmarking of a framework to automate the process of geometric digital twinning for existing slab and beam-and-slab bridges. Called Lukis, the framework followed a top-down strategy to detect and twin bridge concrete elements in point clouds into an established data format. Existing software packages require modellers to spend many hours generating shapes to fit point-cloud sub-parts. Previous methods can generate surface primitives combined with rule-based classification to produce cuboid and cylinder models. While these methods work well in synthetic datasets or simplified cases, they encounter challenges when dealing with real-world point clouds. This challenge was tackled by investigating the entire workflow of geometric digital twinning for bridges and proposing a new framework to auto-generate bridge objects without needing to generate low-level surface primitives. The framework was implemented on a single software platform. Experiments demonstrated its ability rapidly to twin geometric bridge concrete elements. Compared to manual operation, the framework reduced the overall twinning time by at least 95% while the twinning quality (spatial accuracy) was improved. Lukis is the first framework of its kind to have achieved geometric digital twinning for primary concrete elements of bridges on one platform. It has laid foundations for researchers to generate semantically enriched digital twins.
Funding
This research work is supported by EPSRC, Infravation Seebridge project under grant no. 31109806.0007, and Cambridge Trimble Fund.
History
School
Architecture, Building and Civil Engineering
Published in
Proceedings of the Institution of Civil Engineers - Smart Infrastructure and Construction
This paper was accepted for publication in the journal Proceedings of the Institution of Civil Engineers - Smart Infrastructure and Construction and the definitive published version is available at https://doi.org/10.1680/jsmic.19.00012.