Silvio Savarese’s research applying computer vision techniques to construction sites leads to best paper award and a new spinoff company

“We have pioneered an integrated scene understanding framework that enables the automatic tracking of structural changes, allowing data to be collected easily.”

Professor Silvio Savarese Enlarge
Professor Silvio Savarese

Prof. Silvio Savarese received the 2011 Best Paper Award from the Journal of Construction Engineering and Management for research that applies fundamental principles developed for real-world scene understanding to the problem of efficient construction site monitoring. Savarese believes this research will achieve a revolutionary change in the performance analysis of construction sites.

The article, Integrated Sequential As-Built and As-Planned Representation with D4AR Tools in Support of Decision-Making Tasks in the AEC/FM Industry, was co-authored by Prof. Feniosky Peña-Mora (Dean of Engineering and Applied Science at Columbia University), and former student Dr. Mani Golparvar-Fard (Assistant Professor of Civil & Environmental Engineering at Virginia Tech). The researchers formed a startup company built upon the D4AR technology.

The current method of monitoring a construction site involves manual data collection and extensive data extraction from construction drawings and work schedules — a highly time consuming, extremely expensive and erroneous process. Prof. Savarese stated, “We have pioneered an integrated scene understanding framework – called Four-Dimensional Augmented Reality (D4AR) — that enables the automatic tracking of structural changes, allowing data to be collected easily, cheaply and objectively.” The fourth dimension adds the element of time to 3D.

3D site Enlarge

“This new method of performance analysis has the potential to modernize the age-old labor-intensive process in a profound way, freeing up critical human resources, saving costs and improving the quality and efficiency of civil construction projects,” added Savarese.

Describing the research in an earlier stage during his doctoral studies, Dr. Golparvar-Fard stated, “Imagine you are sitting at your office and you would like to conduct a walk-through on your job site, but you are miles away from the site. What you can do now is to make a phone call to your construction site, ask your superintendants and field engineers to take photos of the site and send those back to you. Using our D4AR modeling approach, you would be able to automatically reconstruct actual 3D models of the site using these photos and register the 3D model and the photos with as-designed Building Information Models (BIMs).” This research presented in 2010 earned Dr. Golparvar-Fard a Best Student Paper award at the 2010 International Conference on Innovation in Architecture, Engineering, and Construction. [read more]

The authors have a provisional patent, called D4AR modeling for automation and interactive visualization of construction progress monitoring, and have launched a startup company called Vision Construction Monitoring, LLC, with plans to offer the D4AR modeling technology to the construction industry.