Early Detection of Construction Errors with AR As-Built Inspection: The Key to Reducing Rework

Table of Contents

• What is as-built inspection?

• Challenges of traditional as-built inspection and rework risks

• What is AR as-built inspection?

• Early detection of construction errors with AR as-built inspection

• Reducing rework and improving inspection efficiency

• Visual consensus building enabled by AR

• Points for introducing AR as-built inspection

• AR as-built inspection realized with LRTK simplified surveying

• FAQ

What is as-built inspection?

In civil engineering and construction, as-built inspection (as-built management) is a quality-control process that verifies and records whether completed structures or developed land match the design drawings in shape and dimensions. Especially for public works, the results of as-built management are often critical conditions for passing inspections and for handover, so reliable as-built inspection is required to ensure quality. Traditionally, this inspection involved measuring dimensions and elevations on site using tape measures, staffs (leveling rods), levels, total stations, etc., and then comparing photos taken on site with the drawings back at the office.

However, this method of direct measurement and post-checking is time-consuming and labor-intensive, and because only limited points can be measured it carries the risk of oversights. If there is a time lag between measuring on site and noticing a problem, construction errors may be discovered later and could lead to large-scale rework. In recent years, labor shortages and an aging engineering workforce have further increased the demand for new methods that can confirm as-built conditions efficiently and reliably.

Challenges of traditional as-built inspection and rework risks

Let’s summarize the main challenges of traditional as-built inspection methods. The conventional approach has pointed out the following problems:

• Huge labor and time requirements: Measurements were taken painstakingly point by point by multiple people using surveying instruments and tape measures. For large sites or projects with many measurement points, this could take a whole day. Including the preparation of charts and reports for submission, inspections required a great deal of time.

• Dependence on skilled technicians: Accurate measurement and evaluation require experienced survey technicians, and two-person teams are not uncommon. With chronic labor shortages and an aging technician pool, securing sufficient staff itself was a challenge.

• Expensive equipment required: Measuring differences from design to millimeter precision requires high-performance total stations or GNSS positioning devices. These specialized instruments have high initial costs, creating a barrier for small and medium-sized firms. Maintenance and theft risk are additional burdens.

• Risk of human error: Because work is manual, it is prone to omissions and recording mistakes. For example, forgetting to take photos before backfilling buried utilities or transcribing a memo incorrectly can lead to quality problems.

• Rework due to late problem detection: Because problems are often discovered after returning to the office and comparing to drawings, construction defects may not be noticed on-site and can become irreversible. For instance, if insufficient concrete thickness or inadequate subgrade slope is discovered the next day, materials may already have hardened and large-scale re-pouring may be necessary.

• Burden of document preparation: As-built management requires measurement results to be compiled into drawings and photo logs for submission. Traditionally, preparing these reports took time and effort, placing a heavy burden on site staff.

Thus, traditional methods are inherently inefficient, carrying risks of oversight and rework. To perform precise as-built checks in real time on site, it was indispensable to adopt new technologies different from conventional approaches.

What is AR as-built inspection?

Against this backdrop, a new method called “AR as-built inspection” has attracted attention. By leveraging AR (Augmented Reality) technology, digital information such as design drawings and 3D models is overlaid on smartphone or tablet camera images, enabling direct on-site verification of as-built conditions. Instead of squinting at paper drawings or numeric data, inspectors can compare the real object and digital design information on site, allowing even less experienced technicians to intuitively judge whether the finish is acceptable.

In particular, combining AR with high-precision GNSS (satellite positioning) enables alignment of digital information with the real world within a few centimeters of error. If a small RTK-GNSS receiver attached to a smartphone achieves centimeter-level positioning and preloaded BIM/CIM models or electronic drawings are aligned to site coordinates, design lines and surfaces can be accurately overlaid on the camera view. For example, if the design completion plane or reference surface is displayed in AR over a finished ground or structure, you can immediately judge whether the finish is within tolerance. Because the positioning accuracy is high, the virtual model does not drift away from reality even as the user moves. As a result, as-built management is evolving into real-time on-site verification through digital technology.

Early detection of construction errors with AR as-built inspection

One of the greatest benefits of AR as-built inspection is that construction errors and deviations can be detected on the spot without being missed. By overlaying design data on camera images, height deficiencies or inclination differences of just a few centimeters—difficult to notice with the naked eye—can be detected in real time. Subtle bumps and dips that were previously not captured by limited point measurements can be checked across the entire construction area with AR, ensuring no omissions.

For example, when verifying the fill height in an earthwork project, displaying a virtual horizontal plane corresponding to the design completion elevation in AR allows you to instantly see elevation differences. If the virtual plane appears to float above the ground at a location, that spot indicates insufficient fill; if it appears recessed, it indicates overfilling. Such deviations might have previously been discovered only after surveying post-construction, but with AR they can be noticed immediately after work. If AR shows “this point is 5 cm low,” you can add soil on the spot to correct it. Displayed numeric guides (e.g., “+5 cm fill”) and color-coded heatmaps are visually clear and convey instructions to site workers more effectively than verbal directions. The ability to correct errors on the spot while continuing work is the secret to reducing rework.

Also, modern smartphones are capable of 3D scanning with built-in LiDAR scanners and high-performance cameras. Scanning a finished structure or terrain with a smartphone to obtain point cloud data (a current-state model made up of numerous 3D points) and comparing it with the design 3D data allows automatic generation of an as-built heatmap showing deviations in color. If this heatmap is overlaid in AR on the actual site view, you can immediately see which parts are how many centimeters higher or lower than the design. For example, inadequate pavement thickness or slope shortages on a slope can be identified instantly, enabling immediate corrective actions such as adding pavement or trimming, thereby avoiding large-scale repairs later. Traditionally, point clouds were analyzed back at the office to create color-coded maps on plan views, and then field teams had to locate problem points again on site; with AR you can visualize defects directly on site.

Reducing rework and improving inspection efficiency

Because construction errors can be detected early and corrected immediately, substantial reductions in rework are expected. Rework means redoing nonconforming areas later, which wastes materials and labor and can cause project delays and lowered morale. By adopting AR as-built inspection, problems can be addressed while they are still small on site, minimizing cases that require large-scale rework after completion. This leads to prevention of recurring quality issues and contributes to increased trust from clients.

At the same time, the time and effort spent on as-built inspection are dramatically reduced. Because a wide area can be scanned at once and as-built conditions confirmed, there is no need to measure point by point as before. Software automatically analyzes measurement data and even makes pass/fail judgments, leaving inspectors only to review results and make decisions. For example, slope as-built inspections that previously took several days could be completed in a few hours using AR and point cloud measurements. Shorter inspection times directly reduce labor and surveying costs, making AR a major productivity driver amid chronic staff shortages.

Furthermore, AR as-built inspection automatically records measurement results as digital data, simplifying report preparation. If you cloud-share coordinate-attached AR composite photos and differential heatmap data captured on site, you can avoid the effort of reassembling drawings back at the office. With the inspection process digitized and automated, the burden on staff is reduced while data accuracy improves—providing two benefits at once.

Visual consensus building enabled by AR

AR as-built inspection is also powerful for building consensus with clients and site supervisors. Discrepancies in perception sometimes occur—“the drawing shows it’s within tolerance, but seeing it on site doesn’t convince me.” With AR, inspection results can be visualized on the spot, enabling all stakeholders to share the same information. For example, projecting the design model in AR onto a completed structure lets clients intuitively understand the workmanship, greatly reducing explanation time.

If you can look together at a tablet on site and say, “This area is a few centimeters higher than the design but within tolerance,” or “This area is a bit low, so we will repair it later,” consensus can be reached far more smoothly than via paperwork. The visualization effect of AR enables reliable consensus building with clients and dramatically streamlines the approval process. Also, as-built data and AR images captured on site can be shared to the cloud in real time, allowing remote supervisors or managers to grasp the situation instantly. Remote inspections can be conducted as needed, enabling a new site management system without communication loss.

Points for introducing AR as-built inspection

AR as-built inspection is innovative, but what preparations are needed to introduce it on site? Basically, if you have digital design data and an AR surveying system, you can start relatively easily. The following points summarize concrete considerations.

• Digital design models and drawing data: First, digitize the design data for the structure to be inspected. BIM/CIM 3D models or CAD drawings are ideal, but if unavailable, it is desirable to convert key dimensions and alignment information into coordinate-attached data.

• AR-capable smart devices: Prepare smartphones or tablets that support AR. Recent iPhones, iPads, and Android devices support AR display, and some models have LiDAR sensors for high-precision point cloud measurement.

• High-precision GNSS receiver: Prepare an RTK-GNSS compatible small antenna that can be attached to the device. This can correct typical GPS errors of several meters down to several centimeters, enabling accurate alignment between real-world coordinates and design data. In Japan, augmentation signals from the Quasi-Zenith Satellite System “Michibiki” or network RTK services can be used.

• Dedicated application software: Install a surveying app that supports AR as-built inspection. It should handle reading design data, integrating GNSS positioning information, point cloud scanning, and AR display. Choosing user-friendly software allows personnel without surveying expertise to operate it intuitively.

• Reference information for site coordinate systems: Confirm reference point coordinates and compatibility with the surveying coordinate system. To correctly tie design data to the site, perform coordinate alignment (calibration) in the app using known points if necessary. The use of 3D design data is increasing in public works, and in many cases coordinate-attached models are available.

Once these are prepared, inspectors can start checking as-built conditions by simply displaying design data in AR on a smartphone on site. Although there may be some initial confusion, operation is similar to taking a photo. With brief training and practice, site personnel will become accustomed and able to use it routinely. The Ministry of Land, Infrastructure, Transport and Tourism has promoted ICT construction and 3D as-built management under the *i-Construction* initiative, and in the 2020s it formally recognized the use of smartphones as as-built measurement devices, among other institutional developments. As implementation progresses, AR as-built inspection is expected to become one of the standard methods in the near future.

AR as-built inspection realized with LRTK simplified surveying

To maximize the benefits of AR as-built inspection, the underlying surveying and data-processing system is crucial. LRTK simplified surveying is an all-in-one solution designed to easily implement AR as-built inspection. LRTK is a smartphone-based high-precision positioning and measurement system that can complete site surveying and as-built verification using a single standard iPhone and a small GNSS receiver. With RTK GNSS, positioning can be determined with horizontal accuracy of about ±1–2 cm (±0.4–0.8 in) and vertical accuracy of about ±3 cm (±1.2 in), delivering positioning performance on par with dedicated surveying equipment on a smartphone. Based on this high-precision positioning, AR projection of design models and comparison with point cloud data can be performed on site. Stable AR display without positional drift enables intuitive as-built checks that paper drawings cannot provide.

LRTK also has a point cloud scanning function that utilizes the iPhone’s built-in LiDAR sensor and camera. Even for complex-shaped structures, you can generate a high-precision 3D point cloud model simply by walking around while holding the smartphone; the acquired point cloud is automatically tagged with absolute coordinates derived from RTK-GNSS. This allows immediate comparison with design data and volume calculations on site, enabling instant as-built difference checks. Additionally, recorded coordinates of measured points can later be used to guide piling or equipment installation within centimeter-level accuracy using a coordinate navigation function. By integrating AR display, point cloud acquisition, coordinate guidance, and as-built verification into a single platform, LRTK simplified surveying enables seamless workflows that previously required separate devices and software. Uploading point clouds scanned on site to the cloud and immediately checking differences in AR within a single smartphone app accelerates on-site digital transformation.

In practice, LRTK is being deployed at construction sites nationwide, contributing to faster disaster recovery work and more efficient construction management. Even those who think “I’d like to try AR as-built inspection but preparation seems difficult…” should be able to start operations in a short time using LRTK. Combining the advantages of the latest technology with ease of use, LRTK simplified surveying will be a reliable ally for future construction sites.

FAQ

Q: What do I need to start AR as-built inspection on site? A: Basically, you need a tablet or smartphone device, a high-precision GNSS receiver, and a dedicated app that supports AR as-built inspection. For example, with solutions like LRTK you can attach a small GNSS antenna to a commercial iPhone or iPad to achieve centimeter-level positioning, and handle 3D design data and point cloud data within a dedicated app. Once you load prepared design drawings (BIM/CIM models or electronic drawings) and align them with site reference coordinates, you can begin AR-based as-built checks immediately.

Q: Is the accuracy of AR-based as-built inspection reliable? A: Yes. When combined with high-precision positioning such as RTK-GNSS, AR can achieve sufficiently reliable accuracy. Standard smartphone GPS can have errors of several meters, but augmentation or network RTK corrections can reduce errors to a few centimeters. In fact, LRTK simplified surveying has confirmed horizontal positioning accuracy of about 1–2 cm (0.4–0.8 in), comparable to conventional first-class surveying instruments. Stable AR display based on high-precision coordinates enables detection of centimeter-level offsets and gaps. For critical areas, combining AR with acquired point cloud data allows millimeter-level accuracy verification.

Q: Can AR technology be used for as-built inspection in public works? A: In recent years, the Ministry of Land, Infrastructure, Transport and Tourism has actively promoted ICT construction and 3D as-built management, and trials of AR technology are progressing. In practical trials for direct-managed projects, 3D models created during construction have been projected on-site via tablet AR screens and used for as-built inspections. Although AR inspection is not yet explicitly written into official as-built management guidelines, examples integrating point cloud surveying, as-built management methods, and remote on-site inspection are increasing. In the 2020s, the use of simple mobile devices such as smartphones as as-built measurement equipment has been formally permitted in the guidelines. As further guidelines are developed, AR as-built inspection is likely to become one of the accepted inspection methods for public works.

Q: Is operating AR as-built inspection difficult? Can young or inexperienced workers use it? A: Operation itself is simple, and even those unfamiliar with digital devices can master it with short training. The smartphone app simply overlays design data on the camera view, so no special surveying skills are required. Because the data handled are preprepared design models and drawings, you just select the file in the app and follow the prompts to complete checks. Tools with thoughtfully designed UIs, such as LRTK simplified surveying, let anyone perform accurate as-built checks by following on-screen guides. Results are visually presented, making them easy to understand and easy to share within the company.

Q: For what types of projects and sites is AR as-built inspection effective? A: AR is useful in any situation across civil and building construction where you want to verify discrepancies between design and reality on site. For large-scale earthworks such as roads and land development, AR heatmaps are powerful for managing extensive elevations. In tunneling or dam construction, 3D comparisons with design models prevent overlooking defects. In building construction, you can check positions of columns and walls against BIM models or pre-check equipment piping clashes in AR. In short, AR is effective wherever you want to confirm construction results on the spot. The benefits of introducing AR as-built inspection are particularly large in processes where remeasurement or rework is costly.