Ireland’s Evercam monitors construction sites via IP cameras and is integrating live imagery with BIM and using AI to track site activities.
Dublin, Ireland-based company Evercam provides cameras for construction sites to provide regular feeds of images that can also be used to create time-lapse and project management video sequences.
Established in 2010 by Marco Herbst and Vinnie Quinn, the company looked at various potential applications of high resolution IP (internet protocol) cameras (which are more resilient as long-term solutions than digital SLRs or GoPro-type devices). In 2016, a commission to support a Dublin contract for a student accommodation project highlighted opportunities in the construction sector, and since 2017, the business has been solely focused on supporting construction projects. “Main contractors, in particular, are interested in using cameras to support communications and collaboration, and to avoid disputes,” says Herbst. “These are a perfect fit for what we have built.”
Evercam’s clients now include several of Ireland’s leading main contractors (John Paul, Bam Ireland and Sisk, for example). Its cameras are being used on Ireland’s biggest healthcare project, the National Children’s Hospital in Dublin. And the firm has also been commissioned by several of Ireland’s high-tech industry clients including Google, Facebook, AirBnB, LinkedIn and Oracle.
Customers pay a fixed monthly fee per camera per calendar month (currently £350/mth), plus an installation charge. Cameras may be mounted on adjacent buildings, on other site structures (sometimes alongside installations installed to monitor noise, dust and vibration), and on tower cranes (“Imagery covering groundworks often delivers the most value for the least investment,” says Herbst, right). They may also be installed inside buildings to show, for example, the fit-out of large spaces inside pharmaceutical facilities. On the Royal London Hospital project, Evercam has seven camera feeds covering both external and internal views.
Cameras securely transmit imagery via the internet to a cloud-computing storage facility in Frankfurt, Germany. “Our hosting is predominantly used for storage, and to host the software application that sits over the top,” says Herbst. “Authorised users can then access their projects online through our web interface, or we can also embed Evercam into third-party solutions.” Another Dublin-based construction technology firm, Zutec (posts), was among the early adopters incorporating an Evercam feed into their interface, but Herbst says integration involves little more than cutting-and-pasting some code into the HTML of the destination web page, mentioning Procore and Viewpoint as platforms where Evercam imagery was already being shared on an “ad hoc” basis.
Imagery can be used for marketing videos – showing “before and afters”: the progress of construction from empty site to finished building, to do weekly updates or comparisons, or to produce videos of particular operations (the Evercam website case studies include the controlled demolition by explosion of power station cooling towers, for instance). It can also be used by professionals for project management, to plan work or highlight issues on site, and to provide digital evidence of the sequence of events leading to problems that might cause disputes.
AI and BIM applications of Evercam
Evercam’s development team is applying artificial intelligence (AI) and machine learning (ML) to extend the uses of the camera imagery and software. Using the AI image recognition tools, managers can, for example, monitor vehicle movements in and out of a site. On one project, Herbst showed me how the software captured time- and date- stamped images showing the arrival and departure of all large vehicles at a UK hotel construction site, and then narrowed the search to show just the movements of concrete mixers. “We are also thinking about applying the AI to people movements. In the current coronavirus pandemic situation, we could potentially use tower crane imagery to check that site workers are maintaining social distancing, perhaps to identify hot-spots where additional measures might be needed.”
Herbst also demonstrated how Evercam imagery could produce live and historic visual evidence to support building information modelling (BIM), enabling detailed comparison of what was designed and scheduled with what was delivered and when – effectively validating so-called 4D BIM programs. “We are typically taking outputs from Autodesk’s Navisworks or Bentley’s Synchro, adjusting the models’ viewpoint to ensure alignment with the camera view, and providing a real-time view showing how construction is progressing against what was scheduled.”
Evercam has grown to a 34-strong company and has started to expand internationally. A London-based team was established in 2018 to grow the firm’s UK footprint in 2018; new operations are being established in Singapore and in Australia; and Herbst is also looking at opportunities in the United States (“We already have a couple of US customers who want us to expand to support their projects in the States, but we will need to think about how we fund this expansion.”). Channel partners are a key aspect of the company’s delivery model – the company currently has a dozen firms across the UK and Ireland helping supply and install equipment.
The Extranet Evolution view
Project collaboration platforms have long been used to share site progress photography in the same way as they have shared documents, drawings, spreadsheets and form-based process data. Incorporating live webcam views is also well-established (my one-time employer BIW Technologies worked with Camvista in the early 2000s to deliver a webcam module), but the cameras, connectivity and software have advanced rapidly, and – as Evercam demonstrates – is now pushing into new possibilities enabled by AI and ML.
The technology also complements other reality capture tools, including:
- conventional digital still and video photography, perhaps extended using photogrammetry techniques (Bentley Systems has talked extensively about its context capture technologies in recent years)
- laser-scanning and point-clouds
- 360-degree photography (see 2017 Holobuilder post, and Panono: December 2018 post)
- body-or helmet-mounted cameras (Buildots and OpenSpace: December 2019 post)
- drone, aerial and satellite imagery (OnePlace—formerly UnEarth—for example: May 2017 )
- machine controls (Infrakit: April 2016) and
- a host of mobile smartphone and tablet-based image and data-capture applications