Use of Computer Vision to Improve Construction Safety

Implementation Time:
24 months
Solution Provider: AI Singapore

The Housing & Development Board (HDB) is Singapore’s public housing authority. We plan and develop Singapore’s housing estates; building homes and transforming towns to create a quality living environment for al. We provide various commercial, recreational, and social amenities in our
towns for our residents’ convenience.

  • Falling from height is one of Singapore’s most common causes for fatal injuries at workplace and crane-related fatalities
  • Two unsafe scenarios were identified to be detected from CCTV cameras:
    • Workers working near unbarricaded open edge of construction floor
    • Workers under lifted load

How can AI help provide automated and reliable round-the-clock supervision on construction sites and a faster alerting mechanism for identification of unsafe situations?

Computer vision was used to help monitor the sites. Cutting edge deep-learning technology was developed to identify (i) construction workers (ii) lifted pre-casted loads (iii) the 3-dimensional position of area under a lifted load from a 2-dimensional image, and (iv) unbarricaded open edges on construction sites. Automated alerts would then be sent to site supervisors when the two unsafe scenarios were detected. In particular, unique algorithms were designed for the latter 2 use cases.

Outcomes

  • An MVP to detect two construction unsafe scenarios was deployed and trialled at 4 construction sites
  • Computer vision was applied onto existing CCTV system, reducing the reliance on manual safety supervision thus allowing more consistency and accuracy
  • Safety culture improved as workers are more mindful that there is a smart camera surveillance system

PDF Document

This is the last content of this tab. If you do not see any resources above, it means the solution provider have not provided any resources. Feel free to contact Solution Provider for more information.

Contact Us
The enquirer name
Company Name

Implementation Time

24 months

Use Case Brochure