The Learning Camera

Automating routine checks on sites.

Last updated: 9th May 2021

Date uploaded:

Approved for use

Innovation Lead: Sherrie Rad
Project number: 104794
UKRI funding: £190,239

Website:
thelearningcamera.bamnuttall.co.uk/about/


Edit this story

Summary

To ensure the safety and compliance of construction sites, resource has to be dedicated to routine monitoring of equipment. This is repetitive, time consuming and low-skilled work. The Learning Camera can be fixed in place to monitor such situations using artificial intelligence, alerting the site team only if the condition changes in a way that requires attention. This is particularly useful in hazardous areas.

Innovation type: Digital
Organisation type: Construction tier 1 contractors

Story building blocks:

Story building blocks:

Learn more about this innovation below and use these building blocks to craft your own story to share.

Open allClose all

Project pioneers

BAM Nuttall is one of Europe's largest contractors and operates a large number of complex construction projects. It is driven to innovate and improve safety and productivity across all their sites.

The problem

Managing construction sites is resource intensive. Checking everything is where it should be, and safe, requires regular human monitoring - whether that's noting the presence of safety equipment or taking readings on a critical piece of equipment such as a compressor. This kind of monitoring is time consuming and risky when it involves hazardous parts of site or poor weather conditions. Automating these checks using a static system would improve safety and productivity for the construction industry. 

Vision

A camera monitoring system that learns to spot dangerous or sub-optimal situations across a site would free up workers to concentrate on the task in hand. It would also enable continuous remote monitoring. Removing humans from the process would mean more regular and consistent data on site situations, and potentially a safer environment. 

Key Insight

BAM Nuttall wanted to understand how well equipment is working onsite, to ensure safety for the workforce and improve efficiency. It partnered with academic and digital twin partners, Cranfield University and Iotic, to test different approaches to automated checking, using cameras to monitor sites and machine learning to train digital systems to alert teams when things are wrong. 

First step

A simple, but critical site monitoring challenge was selected to demonstrate the principle: checking that the safety board was properly stocked and functioning. These boards provide instructions and equipment for use in first aid and emergencies and can often be accessed in a hurry and then need to be restocked.

Barrier

Manual ways of checking safety equipment onsite is critical but takes time and attention away from progress and can increase risk to the workforce. Automating such processes would increase safety and productivity.

Digital Innovation

The Learning Camera captures images of critical equipment continuously and, through machine learning, can be trained to recognise working and defective images. The data feeds back to a database, where the team can analyse problems and establish fixes. The camera also feeds in information such as local air quality and weather data. The physical components include a standard web camera, Raspberry Pi, a GPS sensor and a proximity detector and can be positioned in difficult to reach places where monitoring would be particularly challenging or risky. This is of course specifically for equipment onsite that has no digital and internet connection. The Learning Camera has been testing on equipment such as compressors on uninterruptible power supplies and site health and safety incident response boards. It demonstrated the viability of using a fixed camera that could use artificial intelligence to monitor an image and interpret whether the monitored equipment required attention.

Collaborators

Iotic and Cranfield University established the application of Digital Twin technology within the Learning Camera. The Environmental Agency has supported the project and is working with BAM Nuttall on One Source of Truth, finding a means of limiting physical interaction on site due to Covid19 restrictions.

  • BAM Nuttall
  • Cranfield University
  • Environment Agency
  • Iotics

Lead support

The Transforming Construction Challenge funded the project and, as this first innovation developed, it led to another TCC investment – One Source of Truth, a project between BAM Nuttall and Cranfield University.  

Long Term Vision

Time, resource and risk can all be reduced through greater automation. By combining camera systems and machine learning, BAM Nuttall has created a way that teams can continually and remotely monitor sites to ensure safety and compliance. The Learning Camera improves the accuracy and frequency of the data, frees up teams to work on the project, and creates a safer environment. 

Human Stories

It can be time-consuming for teams to carry out regular checks onsite to ensure everything is where it should be. Often this kind of monitoring can also be risky when it involves hazardous parts of site or poor weather conditions. The Learning Camera is a static system that does the monitoring so teams don't have to, improving safety and productivity for the project and the industry.

Powerful Processes

BAM Nuttall wanted to improve the health and safety onsite and also reduce the time and human intervention currently needed to carry out checks on site equipment. The Learning Camera captures images of critical equipment continuously and, through machine learning, can be trained to recognise working and defective images. The data feeds back to a database, where the team can analyse problems and establish fixes. The camera also feeds in information such as local air quality and weather data.

Fascinating Facts

At least 80% of BAM Nuttall’s current sites have at least one application of the Learning Camera.

Benefits

Productivity
Removing the need for routine, repetitive, low-skilled tasks frees up productive time for the site workforce.

Safety
At least 80% of BAM Nuttall’s current sites have at least one application of the Learning Camera. A typical use is monitoring the accident response board. The site Health and Safety Office is alerted when it is used and the camera detects any items missing or displaced items to be replaced. Therefore, should a serious incident occur on site, the relevant equipment is always available.