A recent story reported by MIT’s Technology Review (among others) has sparked some spirited conversation across the construction industry. It revolves around a joint venture between Boston-based construction company, Suffolk, and computer vision company, SmartVid, from the same New England city. They’ve been working together for the last year to develop an exciting new technology designed to help resolve a problem that’s plagued construction for its entire storied history.
The safety problem in construction
It’s no secret that working on a construction site is among the most dangerous jobs out there. Even with advanced protective gear, significant regulation in place, and the best of intentions by all involved, the construction industry routinely records five times as many fatal accidents as any other industry in the U.S.
In many cases, post-accident investigation reveals that the situation could have been avoided. But, the combination of over-confidence (which is often supported by experience) and a hectic, speed-favoring workflow, even the most safety-conscious individuals can cut corners, overlook warning signs, or otherwise put themselves in a dangerous position. When inexperienced workers intermingle with their more experienced peers, it’s also common for them to get in over their heads, perhaps without others nearby — or even they themselves — realizing the dangerous situation that creates.
In all these cases and more, safety concerns and the accidents they promote happen fast. Regrettably, it’s only after the fact that everything can be sorted out and, hopefully, lessons learned.
But, maybe that won’t be the case for long.
Suffolk’s unique solution
The system being developed by Suffolk and SmartVid could potentially predict — and therefore, prevent — accidents before they happen. It incorporates a deep-learning algorithm (a form of artificial intelligence, or AI) and image-recognition software, along with cameras that are routinely taking shots of ongoing work throughout an active construction site. The system analyzes these images continually, and compares what it “sees” to well over a decade of data, correlating pictured conditions and actions to documented jobsite accidents.
As a result, “it can then be put to work monitoring a new construction site and flagging situations that seem likely to lead to an accident, such as workers not wearing gloves or working too close to a dangerous piece of machinery.”
While the technology has thus far been developed using Suffolk’s own massive database of images and accident reports, the company recently reached out to a number of their competitors in a collaborative effort to reduce safety incidents across the industry. Joining together in what’s been dubbed the Predictive Analytics Strategic Council, a number of companies have agreed to share their own accident data and test the solution on their jobsites, while sharing feedback to help enhance accuracy and reduce any workflow issues that may arise.
With more data to work with and more resources at its disposal, the AI behind Suffolk and SmartVid’s solution will no doubt improve exponentially.
A safer future for construction?
Construction safety has long been a subject we’re passionate about. Access control and safety monitoring, such as that provided through a current solution like Trimble CrewSight, goes a long way toward ensuring only those workers who are trained, qualified, and properly equipped for a given task can access a jobsite or location where that work needs to be done.
As this story illustrates, AI and video-based monitoring offer an exciting, viable solution to that problem that future releases of CrewSight and related solutions may be able to leverage. We look forward to a brighter future where access control and accident prediction solutions can make real progress in reducing the number of accidents that occur every day on construction sites around the world.
For more exciting tech advancements impacting construction, read our free ebook, “7 Tech Trends Driving the Construction Industry.”