The Future of AR Inspections: As-Built Inspections Evolving Further in Collaboration with AI

Table of contents

• Background and challenges that make AR inspections notable

• Innovations in as-built inspections enabled by AR technology

• AR inspections evolving through AI integration

• Future prospects for AR inspections

• Simple surveying using LRTK

• FAQ

In recent years, the construction and civil engineering fields have been seeking digital transformation (DX) to improve work efficiency and quality. One technology that has attracted particular attention is AR (augmented reality). By superimposing digital information such as drawings and 3D models onto real-world images using AR technology, it becomes possible to intuitively share completion images that were difficult to understand from plan views alone and to verify construction accuracy on site. Furthermore, more recently, the collaboration with AI (artificial intelligence) is expected to further advance AR-based onsite inspections. By having AI handle image recognition and data analysis, inspection work can be automated and made more labor-efficient, with a dramatic improvement in accuracy. This article explains in detail how AR technology is bringing innovation to on-site as-built inspections and how it is expected to develop further when combined with AI.

Background and challenges that make AR inspections notable

In the construction industry, the aging of skilled technicians and a shortage of younger personnel have recently led to serious labor shortages on sites. To perform efficient, high-quality construction with limited staff, labor-saving and productivity improvements are indispensable. In on-site as-built inspections (inspections to confirm whether shapes and dimensions after construction match the design) there has traditionally been heavy reliance on the experience of veteran workers and manual labor; measuring dimensions and recording them took time and there was a risk of measurement errors or omissions in records due to human error. Moreover, inspections using conventional surveying equipment such as total stations often require two or more people for certain tasks, and mastering the operation of the equipment also took time.

Also, with paper drawings and two-dimensional plans alone, it is difficult for all stakeholders on site to share a common image of the completed form, and mismatches in understanding can lead to rework or mistakes. For example, items that looked fine on the design drawings sometimes reveal interference with surrounding elements or differences in the finished appearance once constructed; such cases are not uncommon.

AR inspections are attracting attention as a new solution to solve these problems and to improve the accuracy of inspections and construction management. By utilizing AR technology, limited personnel can efficiently confirm and inspect sites, which is expected to suppress variations in construction quality.

Innovations in as-built inspections enabled by AR technology

AR (augmented reality) technology is a technology that uses devices such as smartphones, tablets, and smart glasses to overlay 3D CG models, text information, and other digital content onto real-world images. Adoption in the construction sector has been progressing recently, and it is bringing major changes to as-built inspections.

By using AR, design information that could previously only be checked on drawings can be overlaid and displayed on the actual site, making it easy to intuitively share the completion image and to detect deviations in construction on the spot. For example, if pre-prepared 3D design models such as BIM/CIM are projected onto the local landscape, the position and scale of structures can be visually confirmed and it is possible to immediately judge whether construction matches the design. Slight dimensional differences or tilts that are hard to notice on drawings also become obvious when comparing the virtual model to the real object on AR.

Furthermore, AR-based visualization helps with stakeholder consensus-building and error prevention. Since owners and site staff can share the same AR imagery, it fills the communication gap of “we only find out once it’s finished.” By displaying the expected completed state on AR even during construction and confirming it, problems can be discovered and corrected at an early stage, reducing rework.

AR also contributes directly to onsite work. For example, by combining a tablet AR app with high-precision positioning technology, virtual guide lines and reference points can be displayed on site for layout work without drawing lines on the ground or driving stakes. Even on steep slopes where physical stake driving is difficult, if AR displays “there is a stake here,” workers can perform accurate tasks while looking at the screen. This reduces the labor involved with traditional surveying equipment and marking work, and is expected to reduce human error and improve construction accuracy.

Moreover, by combining AR with smartphone sensor technology, it is possible to scan site topography and automatically calculate required dimensions and volumes. For example, apps that instantly measure excavated soil volume by AR scanning or compute pavement area on the spot have appeared, greatly streamlining manual work using measuring tapes and water-level boards.

There are also initiatives to visualize the positions of buried pipes and cables with AR to enhance excavation safety. Locations of buried utilities that were previously inferred by comparing drawings and the site can now be grasped at a glance via AR displays, reducing the risk of accidental damage.

AR is also beginning to be used for rebar inspection in reinforced concrete. Efforts have appeared that use helmet-mounted cameras or smartphones to capture rebar layouts on site and overlay them with design data models to automatically check for missing rebar or spacing deviations. Inspections that used to require manual verification of each bar can be made more efficient with AR while ensuring quality.

AR inspections evolving through AI integration

What further enhances the effect of AR inspections is integration with AI (artificial intelligence) technologies. By combining AI’s image recognition and analysis capabilities, inspection tasks that previously relied on human eyes can be digitized and automated.

For example, in rebar inspection, AI can automatically count the number and spacing of rebars from camera footage and compare them with design drawings to determine if there are shortages or errors. AI can detect mistakes that humans might overlook and mark them on the AR display in real time, dramatically improving inspection accuracy.

AI can also be applied to visual deterioration of concrete structures, such as cracks and discoloration; AI can identify abnormal areas through image recognition and highlight them on AR. This allows less-experienced technicians to discover defects without omission, standardizing inspections that used to depend on the senses of veterans.

Furthermore, AI can analyze point cloud data obtained by drones or ground LiDAR and photogrammetry results and automatically compare them with design models. By overlaying as-built 3D scan data with design data, AI can instantly identify shape differences or missing parts, and that information can be fed back to AR devices for on-site confirmation. Traditionally, measurement results were taken back to the office for comparison with CAD drawings and deviations were addressed later, but AI-integrated AR makes it possible to know “where it differs from the design” on site immediately.

If such AI-integrated AR inspections are realized, inspection work can be greatly labor- and time-saving. Even with limited personnel, high-frequency, high-accuracy inspections will be possible, making construction quality control more thorough during operations. Inspection results can also be recorded and shared in real time, reducing the administrative burden of report creation. If data collected on site is accumulated directly in the cloud and instantly shared with stakeholders, post-inspection workflows will be greatly streamlined.

Future prospects for AR inspections

Looking ahead, on-site inspections using AR and AI could become the standard in the construction industry. As AR devices become smaller and higher-performance, it may become common to see workers routinely referencing AR information while wearing helmet-type smart glasses. In fact, development of helmet-integrated AR glasses and AR devices equipped with high-precision GNSS is progressing, and in the future AR-based on-site verification could replace paper drawings.

The combination with construction automation and remote management is also expected. If construction machinery that supports high-precision positioning and 3D design data becomes more common, robots could perform automated construction while human supervisors monitor and instruct via AR. It will be conceivable to direct machines to follow virtual guide lines projected on site or to feed back immediate correction commands if as-built results deviate, enabling a real-time quality-control cycle. This approach is attracting attention as a way to maintain stable quality despite labor shortages.

Also, with a shortage of skilled technicians, remote assistance via AR is promising. If a veteran can write instructions in real time on video from AR equipment worn by a worker on site, advanced inspections and judgments can be performed without experts physically present. This allows sites staffed only by younger workers to access veteran knowledge online, improving skills and supporting human resource development.

Such initiatives toward this future have already begun, and with policy support such as the Ministry of Land, Infrastructure, Transport and Tourism’s promotion of i-Construction, numerous demonstrations of AR+AI-based construction management and inspections have been reported domestically and internationally. The fusion of AR inspections and AI is enabling new on-site operations like “continuous inspection during construction” and “taking preemptive action before problems occur.” These approaches contribute not only to quality assurance but also to waste reduction and improved safety, and are likely to lead to a productivity revolution on job sites in the future.

Additionally, it is conceivable to aggregate construction-site data in real time as a digital twin (a twin model in virtual space) in the cloud, allowing appropriate decisions to be made remotely via AR or VR. A future where AI analyzes big data accumulated on site and proposes optimal construction sequences and risk predictions is no longer a fantasy.

Simple surveying using LRTK

As a concrete solution for implementing AR inspections on site, there is a system called LRTK. Since its release in 2022, it has quietly attracted attention among site stakeholders and its adoption has begun. LRTK consists of an ultra-compact RTK-GNSS receiver and a smartphone app; by attaching a receiver with an integrated antenna to a smartphone and receiving correction information from network RTK or satellite augmentation services (for example, the CLAS service of Michibiki), it enables use of a smartphone as a surveying instrument with centimeter-level positioning accuracy (half-inch accuracy).

By using LRTK, as-built measurement and verification tasks that previously required total stations, laser scanners, and PC-based analysis can be completed on site with just a smartphone. For example, at one construction site, workers surveyed control points with a smartphone equipped with LRTK, then scanned the completed structure with the built-in LiDAR to acquire point cloud data, and overlaid the design 3D model in AR to immediately confirm the as-built condition. The process from measurement to drawing verification that used to take several days was completed on the same day.

At another site, CAD data for the planned excavation area was imported into an LRTK-compatible app, and by displaying virtual excavation guide lines on the smartphone screen to guide heavy machinery, they achieved accurate excavation shapes without setting batter boards. For small-scale projects or sites with few workers, LRTK allows efficient one-person surveying and as-built checks, making it an attractive tool to promote DX while compensating for the shortage of skilled surveyors.

Thus, the emergence of simple surveying using smartphones and AR technology is making AR inspections more accessible and practical. Consider adopting AR-based as-built inspections at your sites using the latest solutions. New inspection methods enabled by AR and AI will become the standard that balances quality and efficiency.

FAQ

Q: What is an AR inspection? A: An AR inspection uses AR (augmented reality) technology on site. Through the screen of a smartphone or AR-capable device, design drawings and 3D models are overlaid onto real-world imagery so you can intuitively confirm whether as-built works match the design. It is characterized by the ability to visually grasp dimensional or positional deviations that are hard to understand from drawings alone.

Q: What are the benefits of introducing AR inspections? A: The main benefits are increased efficiency of on-site confirmation work and reduced errors. Because AR allows direct overlay of design and on-site conditions, small differences are easier to notice. Early prevention of rework reduces variability in quality. Since all stakeholders can share the same AR imagery, communication becomes smoother and explanations to owners are easier. Replacing measurements in dangerous areas with AR can also improve worker safety.

Q: What equipment and preparations are needed for AR inspections? A: Basically, you need an AR-capable smartphone or tablet and a dedicated app that can display AR. In situations requiring outdoor accuracy, use of high-precision GNSS (RTK) is desirable. For example, using an RTK receiver that attaches to a smartphone (such as LRTK-type products) enables stable AR display with minimal positional deviation. Preparation of design data (digital 3D models or drawing data) is also necessary. Nowadays, AR can be used easily on smartphones and tablets, so you can consider introduction without large capital investment.

Q: What changes when AR is integrated with AI? A: Combining AI advances automation and sophistication of inspections. For example, AI can read component dimensions and quantities from camera footage and compare them with design values to automatically detect deviations. AI can point out subtle defects that humans might miss, dramatically improving inspection accuracy. As a result, a small team can perform efficient, high-quality inspections.

Q: I’m worried whether my site can handle AR and AI. How should we deal with this? A: You may feel uneasy with new technologies at first, but recent AR apps and AI-equipped systems are designed for intuitive operation. It is good to conduct hands-on training beforehand or to start with smartphone AR to gradually introduce it on site. Younger staff generally have less resistance to digital devices, and once they try it they will likely find it convenient. Experiencing the effect on site even once will make AR and AI reliable daily tools. Actively engaging with digital technology also boosts motivation among young engineers and is expected to be beneficial for securing human resources.

Q: What are the challenges or points to note when introducing AR inspections? A: To maximize the benefits of AR inspections, there are some caveats. First, preparation of design data for AR display is necessary. If you only have drawings, you will need to create 3D models in advance or overlay drawing data aligned to reference points. AR accuracy depends on the device’s positioning accuracy. Indoors, markers or LiDAR can provide reasonably high positional alignment, but outdoors, normal GPS has large errors, so improving positioning accuracy with RTK-GNSS is important. Using a high-precision positioning system like LRTK enables reliable AR inspections with minimal model misplacement. Also, when introducing new technologies, it is important to provide training to site staff and carry out phased trial operations to familiarize the site. Once the benefits are experienced, these challenges can be overcome.

Q: What technologies or trends should we pay attention to going forward? A: Going forward, the miniaturization and weight reduction of AR devices (such as the spread of smart glasses) and faster communication environments (5G, etc.) will further smooth AR use on site. AI will also become more advanced, expanding applications beyond inspections to construction planning optimization and predictive detection (early detection of defects). Combined with IoT sensors, vast amounts of real-time data from sites can be collected and analyzed to enable “smart construction” that always issues optimal decisions and instructions. The development of these cutting-edge technologies will greatly transform future construction sites. The industry is paying attention to these trends, and DX in construction sites is expected to accelerate further.