Contractors and site managers need to conduct quality inspections
regularly during construction. Such work is time-consuming and
labour-intensive, and the resultant quality records are prone to
be incomplete or faulty. Besides, contractors nowadays tend to
adopt multi-layer subcontracting for construction projects. The
presence of multiple subcontractors intensifies the complexity of
quality management, and quality deviations may occur under such
circumstances. In light of this, researchers from PolyU developed
a smart construction quality management system called “PI”,
an acronym for both “Project Eyes” and “Project Intelligence”.
Integrating machine vision and deep learning technologies, PI
detects construction-related objects and identifies the activities of
workers and construction machinery from the videos captured by
surveillance cameras on the construction site. It can also determine
if there are quality deviations or dangers on the construction site.
The PI system can greatly improve the quality management process,
construction productivity as well as safety performance.