This research proposes a fully automated approach for construction progress tracking and as-built model visualization using unordered daily construction images collections as well as Building Information Models (BIM). Such a task currently requires manual data collection and extensive as-planned data extraction, is infrequent and error prone; and if automated can significantly impact the management of a project. Given a set of unordered and uncalibrated site photographs, our approach first reconstruct the building scene, traverses and labels the scene for occupancy. The BIM is subsequently fused into the reconstructed scene by a control based registration-step and is traversed and labelled for expected progress visibility. A machine learning scheme built upon a Bayesian model is proposed that automatically detects physical components in presence of occlusions and demonstrates that component-based tracking at schedule activity level could be fully automated. The resulting D4AR - 4 dimensional augmented reality- model enables the as-planned and as-built models to be jointly explored with an interactive, image-based 3D viewer where deviations are automatically color-coded over the BIM. The D4AR model minimizes challenges of current progress monitoring practice and enables AEC/FM professionals to conduct various decision-enabling tasks in the virtual environment rather than the real world where is time consuming and costly.
For more information, please visit raamac.cee.vt.edu/d4ar
Loading more stuff…
Hmm…it looks like things are taking a while to load. Try again?