No drawings? No problem! Infrastructure inspection techniques for generating plan views from point cloud data

Have you ever faced a situation where you need to inspect aging infrastructure such as old bridges or tunnels but the crucial drawings are not at hand? As structures built during the high economic growth period reach the time for renewal, cases where drawings have been lost or were never properly maintained are being seen across the country. For deteriorated structures, original as-built drawings may have been misplaced, or repeated repairs may have made the drawings inconsistent with the actual condition — such cases are not uncommon.

Without drawings, deterioration and deformation cannot be accurately recorded, hindering efficient maintenance management and repair planning. This is where the use of point cloud data is attracting attention. If you generate up-to-date plan views from point cloud data acquired by laser scanning or photogrammetry, you can proceed with infrastructure inspections confidently even without original drawings.

Table of contents

• Challenges of inspecting infrastructure without drawings

• What is point cloud data?

• Benefits of creating plan views from point clouds

• Specific methods for generating plan views from point clouds

• Examples of point cloud use in infrastructure inspections

• How the emergence of simple surveying is changing job sites

• Simple surveying with LRTK

• FAQ (frequently asked questions)

Challenges of inspecting infrastructure without drawings

In routine inspections of social infrastructure such as bridges, roads, and tunnels, inspectors usually check inspection points according to pre-prepared drawings and plans, and mark cracks and defects on a plan view for records. Drawings provide the foundation for understanding the overall structure and the dimensions of each part, and are indispensable for efficiently detecting abnormalities.

However, for infrastructure facilities that are several decades old, the essential as-built drawings may be missing. Even when drawings do exist, repeated renovation work often causes discrepancies between the original drawings and the current state. Without drawings, inspectors are forced into inefficient work such as estimating by eye at the site, taking rough measurements with tapes, and sketching rough plans by hand. In practice, it is often pointed out that preparing drawings from scratch requires enormous time and effort in preliminary surveys, delaying repair planning and contracting. As a result, oversights or omissions in records can occur, and preparing repair plans can take a long time.

A promising solution to these problems is the point cloud surveying described later for capturing the current state. Creating plan views from point cloud data enables accurate records and rapid responses even at inspection sites troubled by inadequate drawings.

What is point cloud data?

A point cloud (point cloud data) is three-dimensional data that represents the shape of an object or terrain by a collection of many points in space. Each point includes X, Y, and Z coordinates indicating position, and some points may also include information such as color (RGB values) or laser return intensity. The countless points form a three-dimensional contour, allowing a 3D model of the object to be recreated almost like a photograph.

Point cloud data can be obtained using dedicated measurement devices or from photographs. Major acquisition methods include:

• 3D laser scanners: Ground-based laser surveying devices are set on tripods or similar mounts and irradiate the surroundings with laser light to measure the reflection points. They can acquire millimeter-accurate high-density point clouds in a short time and measure details of structures and terrain. In addition to large fixed units, handheld units and vehicle-mounted mobile mapping systems (MMS) are also in practical use.

• Photogrammetry: This method takes many photographs of the subject from various angles with a camera and converts those images into 3D point clouds through image processing. Point clouds can be generated from drone aerial photography or from DSLR and smartphone photos, enabling wide areas to be recorded at once. Advances in software have made it relatively low cost to create high-density point cloud models.

• Measurement using mobile devices: Recently, initiatives have appeared that utilize smartphones and tablets equipped with GPS and IMU to easily measure point clouds. By combining dedicated devices and apps described later, anyone on site can perform 3D surveying even without specialized equipment.

Thus, point cloud data, which used to be handled only by a subset of specialists, has rapidly spread in the civil engineering and construction industries amid technological innovation and initiatives such as the Ministry of Land, Infrastructure, Transport and Tourism’s *i-Construction*. Point cloud surveying, which digitizes the entire site in 3D, is becoming a new standard in infrastructure maintenance and management.

Benefits of creating plan views from point clouds

Using point cloud data to create plan views of the current state offers many advantages not found in traditional manual methods or reliance on existing drawings. Here are some main benefits.

• Accurate representation of the current state: Plan views derived from point cloud surveys accurately reflect the actual on-site dimensions and shapes down to the millimeter. Because they record renovated parts not updated in old drawings and deformations from aging, they become highly reliable as-built drawings.

• Shorter work time: Point clouds allow unified measurement in a short time even for large structures or complex shapes. This greatly streamlines manual measurement work that would otherwise take hours. As a result, lead time from inspection to reporting is shortened, enabling faster decision-making.

• Comprehensive data coverage: Point cloud data covers the surface of the target object thoroughly, preventing measurement omissions or oversights. Once 3D data is acquired, cross-sections or detail drawings at arbitrary positions can be generated later, avoiding the need to return to the site for additional measurements.

• Safety and labor saving: Laser measurement or photography can be performed remotely in dangerous locations such as high areas or roadways, reducing the risk of workers entering hazardous zones. Because wide areas can be measured by few people, it reduces the burden even at sites with labor shortages.

• Digital utilization: Plan data derived from point clouds is easy to store and share on computers and can be edited directly in CAD software or accumulated as 3D models. Detailed analyses or comparisons of aging behavior that are difficult with paper drawings or photo albums become easy with digital data.

Specific methods for generating plan views from point clouds

So how do you create plan views from point cloud data? Below is a step-by-step outline of the general procedure.

• Acquiring point cloud data: First, perform point cloud surveying on the target structure or site. Set up a laser scanner and scan all around, or take photographs with drones or cameras to generate point clouds; recently, it is also possible to scan easily using a smartphone. Measure from multiple locations as needed to ensure full coverage of the structure.

• Processing and preparing the data: Raw point cloud data often contains noise from measurement errors and extraneous points such as people or vehicles. Use dedicated software to delete unnecessary points and, if multiple point clouds were measured, perform alignment (registration) to produce a single consistent point cloud model. Also adjust the data to suit drawing creation, for example by referencing control points in a surveying coordinate system.

• Plan projection and section extraction: Switch the point cloud viewpoint from side views to an overhead view and extract the information needed as a plan view. For interior floor plans, take a horizontal slice at a fixed height from the floor to read wall layouts. For terrain or structural plan views, use the top-down projection to grasp ground outlines and equipment placement. By setting section lines in analysis software, you can draw contour lines from the points that exist on those sections.

• Drafting and finishing: Based on the extracted contours, trace and tidy the lines in CAD software, add necessary dimension lines and annotations, and finalize the drawings. Where automatically extracted lines from the point cloud are unclear, manually supplement and correct while referencing the point cloud data. This produces plan views faithful to the current shape. Cross-sections and elevations can be created using similar procedures if needed.

Note that automatic analysis technology for point clouds is improving year by year, and software that supports drawing generation by automatically classifying features with AI is emerging. In the future, these advancements will likely reduce manual work further and enable faster, more accurate drawing generation.

Examples of point cloud use in infrastructure inspections

There are many reports of inspection work being significantly streamlined by using point cloud data. In road tunnel inspections, for example, a flattened plan called an “unfolded drawing” is required to record cracks and spalling across the entire wall circumference, but old tunnels often lack such drawings. In such cases, 3D laser scanning inside the tunnel has been used to create unfolded drawings and cross-sections of the inner lining from the resulting point cloud data. High-density point clouds allow detection of minute deformations in the tunnel lining, and measurements such as crack lengths and deflections that were previously taken with a scale and sounding rod can now be measured digitally and accurately. Using point cloud–derived drawings helps prioritize repairs and monitor changes over time.

Point cloud use is also advancing in road maintenance. One local government used a dedicated vehicle to perform laser surveys of roads and created current plan views that included the locations of all roadside assets such as guardrails, signs, and gutters. Whereas road asset information had previously been managed through visual inspections and fragmentary drawings, point cloud data enabled accurate and comprehensive understanding. This has accelerated DX (digital transformation) in maintenance by enabling efficient planning for replacement of aging equipment, among other benefits.

Other examples include recording existing piping and equipment layouts in plants with point clouds during renovations, allowing simulation of potential interference before design changes. In this way, point cloud data is applied across a wide range of infrastructure structures from bridges and tunnels to roads and buildings. Even at sites lacking drawings, necessary plans can be generated from 3D measurements to support appropriate maintenance. National infrastructure DX policies also emphasize 3D data utilization, so this trend is likely to accelerate.

How the emergence of simple surveying is changing job sites

While the usefulness of point cloud surveying is well recognized, traditional methods have required expensive equipment and specialist knowledge. For example, ground-based 3D scanners and RTK-GNSS surveying instruments often require investments in the millions of yen, and measurement typically involves teams including experienced surveyors to transport, set up, and operate the equipment. With labor shortages worsening, these time-consuming and costly traditional methods have limitations, and the field has come to strongly demand “simple surveying tools that anyone can use.” The Ministry of Land, Infrastructure, Transport and Tourism is also promoting ICT use via *i-Construction*, and strong expectations are placed on such simple surveying technologies as a key to on-site DX.

In response to these needs, a new approach called simple surveying has appeared in recent years. By combining a compact GNSS receiver attachable to a smartphone with a dedicated app, it has become possible to acquire point clouds while measuring positions at the centimeter level accuracy (half-inch accuracy) without specialized surveying instruments. As devices have become smaller and less costly, an era is approaching in which each worker carries a mini surveying device and can measure on the spot whenever needed. The operation is intuitive and simple, so site staff without special qualifications can handle it — a groundbreaking change. If simple surveying that combines speed, ease, and high accuracy becomes widespread, job sites in infrastructure inspection will be transformed. Acquisition and sharing of point cloud data will speed up dramatically, and tasks that used to wait on surveying can proceed in real time.

Simple surveying with LRTK

A representative solution that realizes the simple surveying described above is LRTK, developed by Refixia Co., Ltd. It is a pocket-sized surveying device (weighing approximately 165 g, about 1 cm (0.4 in) thick) that attaches to a smartphone and houses a high-precision GNSS antenna, an RTK receiver, and a battery. Simply attach it to a smartphone via a dedicated case to achieve centimeter-level positioning (half-inch accuracy), turning the phone into a high-precision surveying instrument. With a single button you can record the current coordinates, and transformations to the plane rectangular coordinate system and height corrections are processed automatically, so the resulting data can be used immediately in practice. Measurement data can also be synchronized to the cloud in real time, allowing distant colleagues to instantly share and review on-site point cloud models. It is truly an all-in-one on-site DX tool that enables surveying, drawing creation, and AR-based verification with just a smartphone.

The greatest feature of LRTK is that it makes smartphone-based point cloud measurement easy for anyone. Because LRTK constantly provides high-precision position coordinates to point clouds acquired by a phone’s LiDAR or camera, the entire point cloud model has correct scale and location information. Usually, 3D scans made with a phone alone gradually distort or lose their real-world position, but using LRTK resolves these issues and produces distortion-free point clouds with absolute coordinates. This makes plan views generated from the point cloud highly reliable. In the sense that anyone on site can easily achieve surveying results comparable to specialized equipment, LRTK is truly a powerful ally for infrastructure inspections.

Even at inspection sites troubled by missing drawings, using LRTK lets you digitize the current state in 3D in a short time and immediately produce the required plan views and cross-sections. Tasks that used to rely on veteran technicians can now be handled in-house, reducing costs and improving operational efficiency. As a finishing touch to the techniques for generating drawings from point cloud data, consider introducing such advanced tools. Now is the time for infrastructure management sites to actively adopt point cloud data and simple surveying to build efficient and reliable inspection systems. This will help realize safer and more sustainable infrastructure management.

FAQ (frequently asked questions)

Q: What is point cloud data? A: It is a collection of innumerable three-dimensional measurement points obtained by laser scanners or photogrammetry. It is a digital 3D model that records the shape of objects or terrain in detail, and each point contains information such as coordinates.

Q: Can infrastructure inspections be done without drawings? A: It is possible, but inefficient and with a higher risk of oversights. Having drawings beforehand is preferable, and point cloud surveying can quickly create up-to-date plan views so you can proceed with inspections with confidence.

Q: Can accurate plan views be created from point clouds? A: Yes. Properly measured and processed point cloud data can yield high-precision plan views. As introduced in this article, point clouds reproduce real structures down to the millimeter, so when converted into drawings the dimensional accuracy is very high and reliable plans can be obtained.

Q: What equipment and skills are needed for point cloud surveying? A: Traditionally, large laser scanners and high-performance PCs were required, but today point cloud measurement is possible with just a smartphone and dedicated devices. Technology has advanced so that, with basic equipment operation knowledge, it can be handled without specialized qualifications. However, foundational understanding such as coordinate systems is desirable for accurate surveying.

Q: Can a smartphone really perform precise surveying? A: Yes, with the latest smartphone surveying systems it is possible. For example, technologies that combine a phone’s LiDAR with RTK positioning like LRTK enable 3D measurements with errors on the order of several centimeters (a few inches). While more convenient than traditional equipment, they have been field-proven to provide sufficiently reliable accuracy for typical civil surveying and inspection purposes.

Q: How should acquired point cloud data be handled? A: Use dedicated point cloud processing software or viewers to display and analyze the data. Exporting to common data formats (LAS, PLY, etc.) allows use for drafting in CAD software or integration into BIM models. In the case of LRTK, a cloud-based 3D viewer lets you easily view and share large point cloud datasets from a browser, enabling data use between the field and the office without installing special software.

Q: Will introducing LRTK lead to cost savings? A: Yes. Using LRTK offers significant cost advantages. In many cases it reduces initial investment compared with purchasing dedicated 3D laser scanners or total stations, and efficiency gains in surveying and inspection work reduce labor and days required. Cloud sharing also reduces the need for paper drawings and business travel, so you can save both time and money overall. Depending on site size and use, LRTK is a solution that achieves economical surveying while ensuring sufficient accuracy.