The COVID-19 pandemic is transforming working practices, especially in construction. With these changes comes the opportunity to embrace digital approaches and remote working.
In 2020, researchers from the University of Strathclyde received funding to develop an online platform for remote building inspections from the Construction Scotland Innovation Centre i-Con Challenge. This nine-month project aimed to enable the safe monitoring and inspection of construction sites during the pandemic.
To achieve this, the researchers created immersive, 3D environments by combining mobile technologies and artificial intelligence (AI). This enabled construction sites to operate safely by limiting the need for building standards verifiers to be physically present on site.
When the construction industry should revert to pre-pandemic practice – especially considering the continued need for social distancing where workers are still on site – is currently the subject of industry debate.
If adherence to the government's one-metre-plus social distancing guidelines is not practicable on site, employers should implement the measures set out in the latest COVID-19 guidance.
However, there are clear benefits to using digital technology in remote or difficult-to-access sites, improving time management and productivity. The UK and Scottish governments are encouraging surveyors to conduct inspections and assessments remotely as far as possible.
The Strathclyde project not only minimises operators' need to be on site, it enhances remote inspections by giving greater insight into data collected. It will also help the Scottish government's Building Standards Division work with stakeholders to develop digital verification of construction standards.
The aims of the project were twofold. The first was to analyse how mobile and wearable technology can enable digital cameras to take images to analyse offline. The second was to understand how such methods can be combined with AI to make identification and sizing of any building defects faster and more repeatable using virtual reality (VR).
The focus was on prototyping a 3D immersive environment to inspect wall facade coatings. Such coatings play a significant role in the durability of buildings as they protect the wall against the external environment. This is particularly important in refurbishment and maintenance projects where, for example, building control surveyors need to inspect the adequacy of remedial damp-proofing.
Project partners included the Building Standards Division, Ecosystems Technologies Ltd, Crawford Building Consultants, Robert Gordon University, Balfour Beatty Construction, BDP and the Highland Council, as well as the Construction Scotland Innovation Centre.
Balfour Beatty and the University of Strathclyde's estates office provided the team with access to the learning and teaching building to help develop and test the prototype inspection process.
Andrew Agapiou © University of Strathclyde
Visual inspection has until now been the primary technique for assessing compliance with the Scottish Building Standards. This approach is time-consuming, however.
It requires particular tools to access elements of a building that are unsafe or difficult for inspection in person. In certain cases, it may not even be feasible to use these tools, especially if the building is in a remote location or has a complicated structure. Other drawbacks with visual inspection include how to interpret the results correctly, and the fact that the surveyor remains in one position throughout.
Remote inspection systems can make building facade inspection safer, more efficient and more accurate. Digital platforms can also enable stakeholders to collaborate simultaneously and review specific elements of a building on a remote basis. This gives occupants and the wider community meaningful ways to engage with a building programme.
To exploit this potential, the project established a framework for integrating methodologies and tools, including VR and digital photogrammetry. This allows collection of real-time data and supports automated decision-making.
Central to the new platform is an interactive VR environment. This enables 3D modelling, BIM and machine learning-based image processing and segmentation to work together.
Accordingly, the system architecture includes visual image collection and processing, accurate acquisition of building geometry, and identification and localisation of defects and anomalies in the fabric. It also includes VR and game-like applications for collaboration.
The research established that creating a digital twin of a building can improve the effectiveness of surveying, allow for remote investigation, automate and enhance communications, save on costs and increase productivity.
The research demonstrated how AI could be used to inspect the University of Strathclyde's School of Architecture.
The three-storey building was designed by architect Frank Fielden, a professor of architecture at the university in the mid 1960s, as part of a post-war campus in an otherwise residential area. It was listed as a grade B building by Historic Scotland in 2012.
The school was designed in the modernist style. It has a recessed ground floor with 12 prominent, two-storey projected bay windows above, and a roughly square lecture theatre. An exposed, in-situ concrete frame supports a base course, band courses and 20 parapets. There are also blue and black bricks in stretcher-bond external cavity walls.
The building has copperised felt cladding on lightweight concrete Siporex panels on all elevations. This was installed alongside replacement of external window units, a limited services upgrade, fabric repairs, internal reconfiguration and decoration. Following these works, the building was physically incorporated into the university's learning and teaching hub.
The research project used image processing to detect and localise defects arising from dampness and deterioration of facade cladding elements.
3D model of the University of Strathclyde's architecture building, generated from the laser scan and analysis of aerial photogrammetry data
The AI inspection tool also incorporates 3D laser scanners, drone technology, BIM, VR and gaming software to carry out visual exploration.
We created a database to support this system that includes bibliographic research on the facade materials used, their potential anomalies and common corrective maintenance measures.
The intention was to prototype a virtual environment that combines a 3D model with a machine learning-enabled defect database. This would allow collaboration by parties interested in generating, transforming and analysing data; so, for example, different partners on a project can define and consult inspection reports. The platform links the characteristics of the exterior facade elements to building condition monitoring and maintenance activities.
The first stage of the project is now complete, with the successful development of a desktop prototype. The team has also created a 3D immersive environment for a sample building, minimising the need for quantity surveyors and health and safety inspectors to be physically present on site.
The platform gives building control inspectors and the wider professional community meaningful ways to monitor and interrogate building repair and maintenance activities. The prototype can be accessed online to navigate the 3D environment using mouse and keyboard, as well as view, filter and assign defects on the facade. A mobile app is also in development to complement the desktop version.
One of the initial outputs of the project is a large, public set of images of defects to brick and concrete facades. These can be used by professionals to develop algorithms, carry out tests and compare any defects they encounter with reference images.
The research team is currently working on the second phase of development with the Building Standards Division. This aims to develop new plug-in features to improve the accessibility and functionality of remote building inspection.
a document search and retrieval facility
a plug-in application that will extract compatible 2D and 3D approval drawings from a document management system and transfer them to an inspector's mobile device
a defect extraction format that can be integrated with inspection records to allow for ongoing reporting.
The team will then test the functionality of the upgraded platform with building control inspectors in Scotland. The initial results of these tests will be analysed to develop an accessible, low-cost platform to detect building standards violations from image data. This will combine image-processing technologies, VR interfaces and AI.
These techniques were initially developed to detect and quantify violations of thermal insulation, acoustic insulation and fire-stopping codes from images alone. The same techniques could be readily applied to monitor a range of other conditions and carry out Building Standards inspection activities where image data is routinely collected. This will include among other things the inspection of historic and inaccessible buildings and bridge structures.
The project team hopes to develop a tool that could be rapidly taken to market, providing a timely and efficient means of comparing buildings with their designs. The use of innovative environmental data collection techniques, from low-cost mobile to rapid 3D scanning technologies, will have future applications in asset management.
The team also envisages an overarching, top-level platform that can be used to frame the various information sources and generate an extended digital twin for the building. This will act as a common information model for plug-in modules that focus on specific areas of enquiry.