AR 検査の基本と活用例：初心者でも分かる出来形検査DX

目次

• AR検査とは

• AR検査のメリット

• AR検査を導入するポイント

• AR検査の活用事例

• AR導入時に考慮すべき課題と対策

• LRTKで実現する簡易測量とAR検査

• よくある質問

AR検査とは

"出来形検査" (as-built inspection/acceptance management) is a quality control process in civil engineering and construction that verifies whether completed structures and terrain have been constructed according to the design drawings. Traditionally, as-built inspections used surveying instruments such as total stations (TS) and levels, measuring heights and thicknesses point by point on site, then returning to the office to compare measurements with drawings to determine pass/fail. However, this method often introduced a time lag between measurement and discovery of issues, causing rework. Accurate measurement and judgment also required experienced surveyors, and many tasks were performed by teams of two, making the process inefficient amid labor shortages and an aging workforce.

Recently, AR (Augmented Reality) technology has attracted attention as a trump card to solve these problems. AR overlays three-dimensional digital information (models, drawings, etc.) on real-world imagery. Once an advanced experiment, AR can now be used in routine construction management thanks to improved smartphone and tablet performance. Modern smartphones and tablets often include high-performance cameras and LiDAR sensors, and with dedicated AR apps, as-built checks can be performed intuitively on site. As the industry-wide DX (digital transformation) promoted by initiatives such as the Ministry of Land, Infrastructure, Transport and Tourism’s [i-Construction](https://www.mlit.go.jp/tec/i-construction/) advances, the introduction of AR into as-built inspections is increasingly expected as a powerful solution to simultaneously improve site efficiency and quality.

AR検査のメリット

• リアルタイムでの問題検出: You can detect construction defects or discrepancies with the design immediately on site, enabling instant corrective action. For example, insufficient pavement thickness or slope deficiencies can be color-coded in AR immediately after construction, allowing additional work or trimming the same day. Because PDCA can be run immediately on site, rework is minimized and quality defects are prevented from being left unaddressed.

• 作業時間の短縮・省力化: Tasks that used to require measuring points with paper drawings and surveying instruments are replaced by intuitive checks in AR where you simply hold up a digital model. Because a wide area of as-built conditions can be visualized at once, inspections that used to take days can be greatly accelerated. One person can perform measurement and verification, reducing personnel coordination and saving labor.

• 人材不足への対応: Site personnel can evaluate as-built conditions on the spot without relying on specialized surveyors or veteran technicians. AR app operation is simple—inspection tasks can be completed by following on-screen instructions. Because no special skills are required, AR helps prevent knowledge silos and allows less experienced workers to perform measurements and checks.

• コストの削減: AR using smartphones or tablets eliminates the need to purchase expensive total stations or GNSS surveying instruments. Specialized surveying equipment often requires initial investments in the millions of yen, but recently it has become possible to build a low-cost measurement environment with centimeter accuracy (half-inch accuracy) by combining common mobile devices with relatively inexpensive GNSS receivers. Maintenance and transportation costs for equipment can also be reduced.

• 測定精度と信頼性の向上: AR reduces the risk of human measurement error and recording mistakes. There is no need to transcribe numbers written on-site; digital data can be compared directly, eliminating human error. Combined with high-precision positioning technologies such as RTK-GNSS, measurements can consistently match public coordinate systems with centimeter-level accuracy (half-inch accuracy), enabling more reliable as-built verification than before.

• 記録・報告の効率化: AR allows as-built inspection results to be saved as intuitive visuals, making report creation easier. For example, attaching AR screen screenshots or differential heatmap images to inspection reports produces more understandable documentation than numeric-only reports. In actual field demonstrations by the MLIT, AR use was shown to allow simplification of submission documents such as as-built drawings. Because records are kept digitally, follow-up checks are easy, reducing the burden of reporting tasks.

• 合意形成とコミュニケーション改善: AR visualization also facilitates information sharing with site stakeholders and clients. Showing the completed image overlaid on the real object with a tablet streamlines explanations to clients. Displaying as-built conditions in AR reduces misunderstandings with clients and allows agreement on corrective spots on the spot. MLIT surveys report that AR is beginning to be used not only for construction management but also for pre-construction briefings to residents and information sharing with subcontractors. Such visualization smooths communication inside and outside the site and contributes to improved trust.

AR検査を導入するポイント

• 高精度な位置合わせの確保: Accurate alignment of digital information with the real world is crucial for AR overlays. On large sites or long objects, small positional shifts can cause large errors. Use RTK positioning with GNSS or calibration using known points to maintain centimeter-level alignment between model and reality. RTK-compatible AR systems can project models without needing physical markers on site, enabling stable AR displays where objects do not shift even when moving.

• 3D設計データの整備: AR as-built checks rely on 3D models of design drawings (BIM/CIM data). If 3D data do not yet exist, create simple models from 2D drawings or digitize the existing conditions by point-cloud scanning to prepare comparison data. MLIT-promoted CIM initiatives are pushing 3D data creation from the design stage, and models will become more readily available in most projects in the future. It is important to become familiar with handling 3D data within your company early.

• 運用フローへの組み込み: To make AR checks part of standard site procedures rather than one-off demonstrations, clarify who will use AR, when, and at what stage. For example, incorporate into construction plans and checklists items like "use AR for reinforcement inspection before concrete pouring" or "perform AR finish checks after each embankment completion." Also decide how to record and reflect AR-verified results in reports. For example, automatically adding date/time and location metadata to screenshots and saving them to the cloud for use as inspection evidence is convenient. Integrating AR into existing quality management flows will make it a tool that everyone on site uses routinely.

• 現場スタッフへの教育: To reduce resistance to new technology, ensure site staff understand how to use AR and its benefits. Start with staff who are digitally savvy and trial AR checks in small tasks. Let them experience that "anyone can take measurements by following on-screen instructions." Modern AR apps are intuitive and can be learned in a short training session without specialist knowledge. Share operation procedures through in-house training and on-the-job training (OJT), and present case studies that convey benefits to veterans so acceptance proceeds smoothly.

• 段階的な導入と検証: Rather than deploying AR across all sites and phases at once, start by introducing AR checks on a limited number of sites or phases to verify effects and issues. For example, trial AR measurement on a specific section to demonstrate efficiency gains and error reductions quantitatively compared to traditional methods; this makes it easier to gain internal and external understanding. Start small, accumulate know-how, and after addressing issues (e.g., equipment handling, accuracy verification methods), expand company-wide. Use demonstration results to develop in-house manuals and checklists to smooth future site rollouts.

• クラウドサービスの活用: Using cloud services linked to AR apps enables automatic cloud storage and sharing of measurement data, point-cloud models, and site photos. This allows real-time information sharing between site and office, and remote confirmation of as-built conditions on AR screens. Team members can view and comment on the latest data in the cloud, enabling rapid instructions for corrections or requests for additional investigation. Data are stored as a history in the cloud, useful for future projects or as evidence if problems occur. Activate cloud integration features as much as possible when introducing AR to centralize data management and ensure smooth information sharing.

AR検査の活用事例

• 鉄筋や構造物の位置確認: AR is effective for checking rebar placement before concrete pouring and for detecting positional shifts of structures during construction. For example, when checking whether column rebar positions are correct, displaying rebar placement drawings in AR on site lets you verify counts and spacing at a glance, replacing the traditional work of measuring with a tape. Overlaying a 3D design model on the real object can detect fine discrepancies, allowing construction to proceed with ensured accuracy. There are reported cases where on-site AR verification enabled early correction of defects, reducing rework and material waste.

• 舗装厚や勾配の出来形検査: In road paving, combining AR with point-cloud measurement allows planar evaluation of as-built conditions over wide areas. Scanning the pavement with a smartphone-mounted LiDAR immediately after paving to obtain a high-density point cloud and overlaying the 3D design model can generate an on-the-spot as-built heatmap showing elevation differences in color. You can immediately determine whether pavement thicknesses across the entire section are within design tolerances and detect surface irregularities or insufficient thickness. You can also measure longitudinal grades and widths directly on the point cloud, completing inspections safely and quickly; some sites have achieved zero rework later.

• 埋設物の確認: AR can "see through" to confirm buried items like pipes and cables that are no longer visible after installation. For example, in sewer pipe work, scanning pipes in 3D before burial and saving accurate position and depth point-cloud data to the cloud allows anyone to understand underground pipe alignment and depth after backfilling simply by holding up a smartphone. This can eliminate ground marking immediately after burial and make future maintenance excavations easier by avoiding buried items with AR displays. Visualizing the invisible is a major advantage of AR as-built inspection.

• 法面・地形データの活用: For steep slope works or large-scale earthworks, combining 3D scanning and AR can improve both safety and efficiency in as-built management. For example, scan a slope before construction to obtain baseline data, then rescan after construction or after a disaster and compare to instantly calculate the extent of collapse or changes in embankment volume. Volume calculations that used to take days can now be completed in minutes, supporting restoration planning and as-built evaluation. Overlaying obtained slope point-cloud models onto on-site scenery with AR allows all workers to intuitively share dangerous areas or reinforcement anchor locations. Combining three-dimensional terrain/structure data with AR display enables safe and reliable as-built management and deformation detection over wide areas and at height that was difficult before.

AR導入時に考慮すべき課題と対策

• 精度への不安: Concerns such as "Can AR really measure accurately?" are common. Indeed, judgments cannot be correct if alignment is off, so precision management is important. Countermeasures include RTK corrections using GNSS and rigorous calibration with known points to eliminate discrepancies between the digital model and real space. Combining a GNSS rover with an AR system allows you to align design data and the as-built exactly in space, and with proper operation AR checks have been demonstrated to achieve accuracy comparable to traditional surveying (plane and elevation within a few centimeters (a few inches)). In the early stages of introduction, use traditional measurements at key points to verify errors and pay attention to ensuring accuracy.

• デジタルデータ準備の手間: AR requires preparation of digital materials such as 3D models and point clouds, and preparing these can be labor-intensive. Although BIM/CIM design data are gradually becoming more common, many small to mid-size projects still lack 3D data. In such cases, LiDAR scanning of the site to obtain as-built point clouds can be used as an ad hoc 3D model. There are also apps that can create simple AR models from 2D drawings by displaying baseline lines and planes. MLIT guidance is promoting transition to as-built management using 3D measurement technologies, so digital data will become easier to obtain. Although it may feel burdensome initially, once data are created they are valuable for subsequent process management and future maintenance, so treat digitization as an investment.

• 端末・現場環境への対応: Pay attention to physical challenges when using smart devices on site. For example, screens can be hard to see under direct summer sunlight and batteries drain quickly. Use tablet sunshades or carry mobile battery packs to mitigate these issues. For rainy weather, prepare waterproof cases or splash-proof covers. In dusty environments, clean cameras and sensor areas frequently. If it is difficult to hold a tablet for a long time, using a neck holder can help. Adopt accessories and operating methods suited to the site environment to ensure devices can perform adequately.

• 現場スタッフの抵抗感: Psychological barriers to new technology cannot be ignored. Veteran workers, in particular, may say "manual methods are more reliable." The best solution is to let them experience the benefits of AR firsthand. Share concrete results like "an inspection that used to take half a day finished in 30 minutes with AR" or "a rebar error that had been overlooked was found on the spot." Tools like LRTK that allow "anyone to easily survey alone" can turn two-person tasks into single-person operations and are often welcomed on site. Start with younger employees using the technology and let its convenience spread through the site to gradually reduce resistance.

• 導入コストとROI: Introducing new hardware and software involves costs, but AR can leverage existing smartphones and tablets, significantly lowering initial investment barriers. As noted above, you may avoid buying expensive dedicated surveying equipment and can start with costs for a GNSS receiver and software subscription. Considering reductions in rework and labor costs from labor savings, ROI can be expected relatively early. For those concerned about cost-effectiveness, start with a limited introduction and demonstrate visible results (e.g., X% reduction in man-hours, Y fewer defect corrections) internally. Calculating ROI from actual results provides persuasive material for management and clients and facilitates further investment decisions.

• 公式な検査への適用: Currently, as-built management guidelines sometimes require traditional measurement and drawing preparation to be performed in parallel. Some inspectors may be cautious about accepting tablet-only checks as sufficient. However, MLIT has confirmed the effectiveness of AR-based as-built inspections in field demonstrations, and AR-based labor-saving methods are expected to be incorporated into guidelines in the future. Even now, software that automatically creates as-built documentation from AR-obtained point clouds and photos can produce deliverables equivalent to manual work. The key is to explain AR results to clients and inspectors in an understandable way and obtain their acceptance. Showing as-built heatmaps on a tablet during inspection meetings provides more convincing evidence of quality than paper drawings. With growing understanding in both public and private sectors, early adopters who accumulate know-how will gain future advantages.

LRTKで実現する簡易測量とAR検査

One solution gaining attention for making AR as-built inspections easier and more accurate is "LRTK." LRTK is a cutting-edge tool that enables centimeter-level high-precision positioning (half-inch accuracy) by simply attaching a compact GNSS receiver to a smartphone, allowing RTK positioning and enabling survey tasks that previously required specialized equipment and skilled operators to be completed by a single person. It supports Japan’s quasi-zenith satellite system "Michibiki" CLAS corrections and network RTK, maintaining high precision even in mountainous areas outside of communication coverage. The major strength is that even without veteran surveyors, a single smartphone can handle tasks from reference point surveying to as-built verification.

LRTK also integrates seamlessly with AR functionality. By using high-precision GNSS position information, 2D/3D design data can be overlaid precisely on site, eliminating the cumbersome task of alignment and preventing object drift. For example, simply walking around the site with a tablet can accurately indicate the virtual stake positions from the design model on the actual ground, allowing target coordinates to be visually identified even at distant points. It is also possible to automatically overlay acquired as-built point-cloud data and the design model in the LRTK cloud for differential comparison to instantly check "whether construction is proceeding as planned."

LRTK provides a cloud platform where measured and scanned data are synchronized to the cloud on site. Team members can view site 3D point clouds and measurement point information from office PCs in real time, enabling collaborative verification among stakeholders. You can also measure distances, areas, and volumes in the cloud, and link photos with location information for one-click listing. This breaks down the barriers between site and office and dramatically improves the efficiency of as-built inspections.

In addition, LRTK offers diverse features beyond as-built management, such as a "coordinate navigation" function that guides a single user to stake positions, features to calculate embankment volumes from LiDAR-scanned point clouds, and cloud sharing of high-precision geotagged photos. In other words, LRTK is designed to let you complete workflows that used to require multiple devices—from surveying to verification, recording, and as-built inspection—using a single iPhone. Data obtained on site can be used and delivered in formats compliant with the MLIT’s as-built management guidelines, and many construction companies are adopting LRTK to achieve both labor savings and quality improvement.

By using LRTK, a smartphone surveying + AR system, anyone can easily perform high-precision as-built checks, overcoming many constraints associated with surveying and inspection. Even sites struggling with labor shortages can realize shorter work times, reduced human error, and improved stakeholder communication through one-device-per-person smart surveying tools and AR. These innovations strongly support the DX of construction sites and are fundamentally changing as-built management. The key to successful AR as-built inspection is to incorporate advanced tools like these and link them to overall site productivity improvements. Take advantage of the latest technology and let "as-built AR checks" demonstrate their full value on your sites.

よくある質問

Q: 出来形ARチェックを始めるには何が必要ですか？

A: Basically, you need a smartphone or tablet capable of AR display, a GNSS receiver to improve measurement accuracy, and a corresponding AR surveying app. Modern iOS/Android devices have capable cameras and sensors suitable for AR use. If centimeter accuracy (half-inch accuracy) is required, combine the device with a Bluetooth-connected compact GNSS rover for RTK positioning (for example, LRTK devices that can be attached to a smartphone). Also prepare digital comparison data such as 3D design models and point clouds. With these set up, you can immediately try as-built AR checks on site.

Q: ARによる出来形チェックの精度は信頼できますか？

A: Yes—if operated appropriately, high reliability can be ensured. Systems using GNSS RTK corrections can achieve positioning accuracy on the order of a few centimeters (a few inches) in both plane and elevation. This falls within the accuracy typically required for as-built inspections. Even when verifying differences in AR, heatmap displays and similar tools allow you to obtain quantitative information such as "which point is X centimeters higher/lower." The important things are to align with site control points beforehand and, when necessary, verify selected points using traditional methods. Doing so will provide sufficient basis to trust AR check results.

Q: 3Dの設計モデルが無い現場でもARを活用できますか？

A: Even without a 3D model, AR can be utilized with some creativity. Some apps can overlay 2D drawing data (e.g., DXF) in AR space to visualize key lines and positions on site. If the finished shape is not very complex, a simple approach is to mark major dimensions on site before construction and overlay those marks on AR-captured images. However, AR’s true value is greatest with 3D models. As the creation of CIM models in public works increases, request 3D data from the client or consider making simple models in-house. Alternatively, for comparing measured as-built data (point clouds) to design drawings, you may use point-cloud processing software to detect differences without relying on AR. The goal is intuitive on-site verification, so choose the method best suited to whether a model is available.

Q: ARチェックの結果は公式な検査に使えますか？

A: Currently, relying solely on AR as the only basis for official inspection is still early, but acceptance is gradually increasing. MLIT conducted field demonstrations in FY2023 confirming that AR technology can allow omission of certain as-built documentation. At present, traditional records (measurement drawings and photo logs) may still be required in parallel, but using specialized software to automatically create as-built documentation from AR-obtained point clouds and photos can produce deliverables equivalent to manual methods. The key is to explain AR results appropriately to the client and inspectors and obtain their understanding. For example, showing an as-built heatmap on a tablet during an inspection meeting can demonstrate quality more convincingly than paper drawings. With growing understanding in both public and private sectors, early adoption and accumulation of know-how will provide future advantages.