Model-based detection of progress using D4AR models generated by daily site photologs and building information models

In this paper a new approach that automatically recognizes progress from daily construction photologs and Building Information Models (BIMs) is presented. Currently daily site photologs are being collected at almost no cost on all construction sites; meanwhile BIM models are increasingly turning into binding components of AEC contracts. Using these emerging sources of information, we present a 4 Dimensional Augmented Reality-D 4 AR-modeling approach for integrated visualization of as-built and as-planned models as well as a novel framework for automated recognition of progress. Our approach is based on structure-from-motion, multi-view stereo plus voxel coloring and labeling algorithms to calibrate cameras, reconstruct the building scene, traverse and label the integrated as-built and as-planned scene for occupancy and visibility. Next, a machine learning scheme built upon a Bayesian model is …