Early detection of actual or potential performance deviations in field construction activities is critical to project management as it provides an opportunity to initiate proactive actions to avoid them or minimize their impacts. Despite the importance, current monitoring practice includes manual data collection and extensive data extraction, non-systematic and generic reporting, and visually/spatially complex representations. This talk addresses these challenges by introducing the underlying hypotheses and algorithms for automated generation of D4AR – 4D augmented reality – models for automating and visualizing monitoring of sustainable built environments. These models assembled through superimposition of 4D point clouds generated from unordered daily construction photo collections and 4D Building Information Models, visualize performance deviations and allow Architecture/Engineering/Construction professionals to intuitively observe problems, conduct various decision-making tasks, and minimize detrimental impacts of performance deviations in an augmented-reality environment rather than the real world which is time-consuming and costly. Moreover, application of D4AR models, developed with several challenging building construction photo collections captured under different lighting conditions and server occlusions, demonstrates that component-based tracking of progress at schedule-activity level could be automated. These models generate a new research paradigm by allowing researchers further develop visual and spatial sensing techniques to automatically track productivity, safety, quality, and carbon footprint of operations.
Dr. Golparvar-Fard has a Bachelor of Science and a Master of Science degree in Civil Engineering from Iran University of Science and Technology (2002, 2005), a Master of Applied Science degree in Project and Construction Management (Civil Engineering) from University of British Columbia (2006), a Master of Science in Computer Science and a Ph.D. in Civil Engineering from the University of Illinois at Urbana-Champaign (May and Aug 2010). Dr. Golparvar-Fard has also worked for five years in private construction industry and his most recent appointment was assistant project manager at Turner Construction Company. His research focuses on visual sensing, semantic analysis, and augmented reality visualization of construction and building performance metrics. Particular focus is on 1) creating computer vision, image processing and machine learning techniques to automatically reconstruct as-built 3D and 4D models as well as track construction progress, productivity, quality, and safety from static images as well as video streams; and 2) building information modeling to reason about building elements, systems and contents to support model-based assessment of performance metrics. He has also been the recipient of the 2010 best poster from Construction Industry Institute (CII), the best conference paper from 2010 International Conference on Innovation in AEC, 2010 FIATECH CETI award in outstanding researcher category and finally the best poster award from the 2009 ASCE Construction Research Congress.
