Non-Destructive Inspection Revolution: Structural Inspection DX Realized with AR and 3D Point Clouds

The Importance of Non-Destructive Inspection and the Limits of Traditional Methods

As social infrastructure ages, non-destructive inspection (inspections that examine condition without causing damage) has become increasingly important to ensure the safety of structures such as bridges and tunnels. Especially since 2014, laws have mandated regular close-up visual inspections for road bridges and tunnels every 5 years, greatly increasing the frequency and scope of inspections nationwide. Many structures built intensively during the high-growth period are now entering a high-aging phase of over 50 years, and the importance and burden of inspections are expected to increase further. Traditionally, structural inspections have been carried out by experienced technicians using visual inspection and simple tools to check for abnormalities. However, the following limitations and challenges of these conventional methods have been pointed out.

• Subjectivity: Inspection results can be influenced by the individual inspector’s experience and judgment, leading to variability between veterans and newcomers.

• Ambiguous records: Inspection result recording methods and formats are not standardized, and the precision and granularity of photos and notes vary by person, making it difficult to accurately reconstruct conditions later.

• Low reproducibility: It is difficult to compare and verify against past inspections, making it hard to objectively grasp the degree of deterioration over time.

For example, even for a single concrete crack, the conventional approach was commonly to mark it with chalk, measure its length, and record it by hand. With this method, it is hard for others to imagine exactly where and how large the crack was from the record, and it is not easy to identify the same spot in the next inspection. Relying on subjective methods limits the improvement of infrastructure maintenance precision. Therefore, there is a growing need to realize inspection DX (digital transformation of inspections) using digital technologies to build systems that allow anyone to perform and record high-precision inspections in the same way.

The Importance of “Visualization” and “Recordability” in Non-Destructive Inspection

The key to overcoming the above challenges is improving inspection information “visualization” and “recordability.” “Visualization” means making previously invisible things visible. In traditional inspections, the presence or severity of abnormalities remained in the inspector’s head or on paper reports and was not visualized in a way that all stakeholders could intuitively share. For example, even if a report states “a 0.3 mm (0.01 in) crack in the bearing of XX Bridge,” it is difficult to grasp at a glance exactly where and what shape the crack is. If information is not sufficiently shared and visualized, necessary repair decisions may be delayed or overlooked. In fact, investigations into infrastructure accidents have pointed out cases where warning signs existed but insufficient records prevented countermeasures.

On the other hand, improving “recordability” means leaving inspection results in sufficient detail so that they can be reproduced by anyone later. With clear records, comparing the same location 5 or 10 years later becomes easy, allowing quantitative evaluation of deterioration progress. Low recordability risks losing valuable inspection knowledge when personnel change.

Incorporating digital technology into non-destructive inspection can dramatically enhance both visualization and recordability. Specifically, two pillars are key to inspection DX: intuitive on-site visualization using AR (augmented reality) and precise digital records using 3D point cloud data. The next sections examine the benefits each technology brings to the field in detail.

Intuitive Inspection Support and Construction History Visualization with AR

AR (augmented reality) is a technology that can overlay digital information onto real-world images via smartphones, tablets, or AR glasses. Introducing AR to non-destructive inspection sites makes inspection tasks markedly more intuitive and easier to understand.

For example, imagine being able to overlay past inspection records on-site via AR during a bridge inspection. If the locations of previously detected cracks and repair history information are displayed like holograms on the actual structure, the need to flip through paper records and search while thinking “was it around here…?” is eliminated. If previous damage locations are highlighted in AR, inspectors can immediately identify the points they need to check by looking at the camera image. At the same time, they can make real-time comparisons by overlaying the physical object with digital records to see whether “the crack has widened compared to last time,” preventing missed deterioration.

AR technology can also be used for task navigation, not just comparing past data. For example, items to be checked according to an inspection procedure manual can be displayed sequentially in AR space, and arrows or markings can indicate the next target the inspector should view. By following a digital guide rather than relying on the intuition or experience of veterans, even newcomers can perform inspections thoroughly and reliably. Additionally, for large structures such as bridges, AR can transparently display as-built drawings or design information, allowing checks that overlay the “original intended position or shape” with the current condition. For structures with internal cables or piping, displaying layout on drawings via AR lets inspectors visually understand what lies behind a wall and carry out appropriate non-destructive inspections (e.g., flaw detection), helping avoid unnecessary openings or excavations.

AR supports fieldwork in ways that align with human senses. Its biggest advantage for inspectors is that information can be obtained immediately on site in an easily understood form. Necessary data appears in the space in front of the inspector without spreading paper forms or carrying drawings, greatly improving inspection efficiency and accuracy. As wearable devices such as AR glasses become more widespread in addition to tablets, inspectors will be able to have necessary information constantly displayed in their field of view while keeping both hands free, further improving safety and efficiency.

Enhancing Defect Detection, Record Precision, and Reproducibility with 3D Point Cloud Data

The other important pillar is the method of recording and analyzing a structure’s shape in detail as 3D point cloud data. Point cloud data represent the surface of a structure or terrain as a collection of innumerable points (a set of coordinates). In recent years, with the spread of laser scanners, photogrammetry, and LiDAR-equipped smart devices, it has become easy to perform high-precision three-dimensional scans on-site and acquire point cloud data.

Using 3D point clouds dramatically improves the precision of defect detection and recordkeeping in non-destructive inspection. For example, analyzing point clouds obtained from scanning a tunnel inner wall can reveal minute deflections and surface irregularities that might be missed by visual inspection. Crack widths and delaminated areas can be measured accurately on the point cloud. Dimensions that were traditionally measured with scales or crack gauges can be quantified digitally down to the millimeter, reducing human measurement errors.

Point cloud data also greatly enhance the reproducibility of inspection records. Once a point cloud is saved, comparing newly acquired point clouds during inspections 5 or 10 years later allows changes to be captured as objective numerical values. For instance, you can visualize the progression of deterioration by showing “a displacement of ◯◯ mm compared to several years ago” or “an increase in the area of the defect.” While photo records are difficult to compare precisely because shooting angles and distances vary each time, point clouds record the entire space, making it easy to analyze differences at the same cross-section later.

There are many other advantages to point cloud data. Representative ones include:

• Intuitive 3D visualization: Point clouds can be displayed in a three-dimensional view as if the structure were copied whole, allowing instant understanding of depth and shape. Information that is hard to convey in drawings or photos becomes easy to share with clients and stakeholders in 3D.

• High-precision measurement: Distances between any two points, areas of defects, and volumes can be measured freely and more accurately than manual work, reducing human error. Advanced evaluations, such as calculating the volume of fallen concrete from the point cloud, are possible.

• Comprehensive information coverage: Point cloud surveys record all on-site locations without omission, eliminating concerns like “we forgot to measure that part.” They function as digital archives for later additional analysis or secondary use as needed.

Against this background, 3D point cloud technology is expected to be a core technology driving innovation in infrastructure maintenance management in construction DX initiatives promoted by the Ministry of Land, Infrastructure, Transport and Tourism, such as *i-Construction*. In the field of non-destructive inspection, research and demonstrations of crack detection using 3D point clouds and AI, automatic deformation analysis, and more are progressing, and in the near future part of the inspection work that relied on human eyes and hands may be automated on digital twins.

Streamlining the Inspection Process with Cloud Integration and Remote Support

Digital data acquired with AR and 3D point clouds can be further leveraged through cloud integration. By centrally managing inspection data in the cloud, it becomes possible to understand field conditions from the office, have multiple personnel view and analyze data simultaneously, and provide remote support.

For example, if high-resolution point clouds or photos captured on-site are uploaded to the cloud on the spot, supervisors or specialist technicians in remote locations can check them immediately. Field personnel can proceed with inspections without omissions while receiving real-time advice, and in difficult cases they can consult veterans on the spot. Processes that used to take days—such as “take it back to the office and then consult”—are increasingly being completed the same day through cloud-based information sharing.

Using cloud-based inspection data management systems, stakeholders can view 3D point cloud models and photos via a web browser without installing dedicated software, add annotations, and organize information. Since everyone can access the latest information at all times, discrepancies such as “the field was using a new drawing while the office was working from an old version” are avoided. In infrastructure inspection, multiple parties are typically involved—road managers, bridge-owning municipalities, inspection contractors, and construction consultants—so the value of aggregating data and expertise in the cloud is considerable.

Remote on-site participation, which has seen growing demand, enables remote support without physically visiting the site. If inspections or guidance can be conducted from the office using field images and data, it helps address labor shortages and reduce costs. For example, AR remote support, where a field inspector shares live camera footage and a remote specialist marks up or annotates that footage, is becoming practical. This enables less experienced technicians to work with constant veteran support, making knowledge sharing across regions and generations a reality. The Ministry of Land, Infrastructure, Transport and Tourism has also actively promoted the introduction of remote technologies to infrastructure inspection and began trials of remote “close-up visual inspections” for bridges from 2021, supporting DX from a regulatory perspective. With wider social implementation of remote participation, uniform inspection quality nationwide irrespective of geographic conditions could become a reality.

By combining cloud integration and remote support, overall inspection process efficiency and sophistication can be expected to improve. The sequence of data collection → analysis → reporting can be performed seamlessly without spatial or temporal constraints, allowing more infrastructure to be properly maintained with limited human resources.

LRTK Features and Use Cases in Non-Destructive Inspection (undersides of structures, heights, complex shapes, etc.)

A solution gaining attention for easily practicing the above DX technologies in the field is our LRTK. LRTK is an integrated platform that combines high-precision RTK-GNSS positioning technology with smartphone-based 3D scanning and AR functions, enabling anyone to easily perform point cloud measurements with absolute coordinates and AR display. By linking a dedicated positioning device with a smartphone and simply walking the site as if shooting a video, wide-area structures can be three-dimensionally scanned and digitally recorded. Since the acquired point clouds have absolute coordinates based on the Geospatial Information Authority of Japan’s standards, the data are reliable with high positional accuracy. It is easy to accurately align and compare data acquired from multiple inspections or overlay them with design drawings or CIM models for review.

LRTK demonstrates its power in various non-destructive inspection scenarios. For example, for parts that are normally difficult to approach—such as the undersides of structures or narrow spaces—LRTK can obtain 3D point clouds of target areas by simply pointing a camera from a safe location. For inaccessible heights, photogrammetric analysis of zoomed images from the ground can generate 3D models that allow detailed measurement later. Also, by aligning a crosshair on the smartphone screen and pressing the shutter, it is possible to remotely measure the coordinates of distant objects. With such target positioning functions, recording the coordinate position of a crack located 6 m (19.7 ft) away at a high location from the ground becomes easy. Inspections that previously required aerial work platforms or scaffolding can be performed with LRTK while greatly reducing personnel, time, and cost and improving safety. In addition, LRTK can integrate with drones and 360-degree cameras, making it applicable to inspecting areas such as bridge undersides or dam crowns that are difficult for people to enter.

LRTK is also effective for inspecting structures with complex shapes. Even curved or intricate steel structures can be scanned from all angles to acquire accurate point cloud data, enabling the entire structure, including blind spots easily missed by visual inspection, to be digitally recorded without omission. Point clouds acquired on site can be immediately reviewed in a cloud-based 3D viewer, and distance or area measurements and cross-section creation are one-touch operations. The ability to record large areas in a short time is another notable feature. For example, in mountain slope inspections, LRTK can scan slopes extending over 100 m (328.1 ft) in roughly 1 minute and record surface irregularities and displacements in detail. Areas that would take a full day manually can be efficiently covered digitally. Rivet patterns or weld shapes on steel bridges can also be observed in detail on the point cloud, aiding early detection of deterioration signs.

LRTK is already being trialed in various field settings, including regular inspections of bridges and tunnels and plant equipment maintenance, contributing to labor-saving and sophistication of inspection processes. The operation is simple, and even technicians unfamiliar with the equipment can master it after a short lecture. This helps address the problem of technical succession caused by veteran retirements and labor shortages, enabling organizations to maintain stable inspection quality.

Conclusion

The world of non-destructive inspection, which supports infrastructure safety, is now undergoing a revolutionary transformation. With the introduction of digital technologies such as AR, 3D point clouds, and cloud services, inspection work that used to depend on people is shifting to objective, data-driven processes. This structural inspection DX is indispensable for efficiently and reliably maintaining aging social infrastructure. Digitalization not only improves inspection efficiency but also enables the use of collected big data to develop preventive maintenance and predictive maintenance, reducing long-term maintenance costs and risks. Promoting DX is not merely about labor saving; it represents a shift to smarter management that considers the entire infrastructure lifecycle. If repair and renewal priorities can be objectively determined based on data, unnecessary works can be reduced and limited budgets used effectively. DX is expected to be a key to achieving both cost reduction and improved safety.

In this context, LRTK has emerged as a reliable partner to root non-destructive inspection DX in the field. Its ease and accuracy—enabling simple surveying and precise recording with a smartphone in hand—have the potential to overturn conventional wisdom at inspection sites. By eliminating subjectivity, standardizing records, and ensuring reproducibility, it contributes to solving all the challenges discussed in this article and is expected to raise the quality of infrastructure inspection. Such advanced technologies will also greatly assist skill transfer and mitigate workforce shortages in the field.

The digital revolution in non-destructive inspection has only just begun. As technology progresses further, new developments such as AI-based automatic analysis and robotic inspections may emerge. However, to maximize the benefits of DX, the starting point is accurately acquiring and sharing high-precision field data. Take this opportunity to step into smart inspections using LRTK and get ahead in the future of infrastructure maintenance management.