Updating Field Surveying with LRTK×SfM Processing: Easy, High-Precision 3D Measurement with a Smartphone

Surveying and as-built management on construction sites are indispensable processes for safe, high-quality construction. However, conventional surveying methods have presented many challenges, such as requiring significant manpower and time and producing variable accuracy. Against this backdrop, a new approach combining smartphones and photogrammetry has emerged and is attracting attention. This is “smartphone surveying” that merges LRTK (a high-precision GNSS receiver mounted on a smartphone) and SfM processing (a technique that generates 3D models from photographs). This article explains the issues with traditional surveying, the basics of this technology, specific application examples, and remote management using the cloud, considering the significance of introducing smartphone surveying with LRTK and the possibility of future standardization.

Challenges of Conventional On-Site Surveying

In civil engineering and construction, surveying (as-built management) is carried out to confirm that finished structures have been constructed according to design. Conventional surveying has mainly been done by manual labor, and it came with many problems. In typical methods, surveyors or site technicians use tools such as levels, tape measures, and total stations to measure heights, widths, and thicknesses at key points on the construction site, comparing them with design values on drawings. At the same time, it is also necessary to take and record photos during construction and upon completion. The following issues have been pointed out with such traditional on-site surveying:

• Large manpower and time burden: Measuring many points manually on site requires multiple crews and long hours. Including the clerical work of summarizing measurement results into charts and tables, this was a major burden for field technicians. Arranging experienced surveyors is also necessary, and amid labor shortages, it is not easy to proceed efficiently within the construction schedule.

• Limited measurement points and difficulty in grasping the whole: There is a limit to the number of points that can be measured manually, making it difficult to fully capture wide areas or complex shapes. Measuring only limited points risks missing subtle differences from the design drawings. The larger the structure, the greater the risk of overlooking localized unevenness or small errors, sometimes leading to panic and corrective work during inspections when deviations from the design are pointed out.

• Risk of human error: On busy sites, human errors such as forgotten measurements or recording mistakes can occur. For example, if photos are not taken before burying an embedded object, there will be no evidence after completion, which in the worst case can lead to rework or disputes. Because the work is manual, it can only measure at discrete points and cannot eliminate human errors completely, placing great pressure on site personnel.

• Safety issues: Conventional methods sometimes require workers to enter dangerous areas for measurements, such as high places or slopes, or to survey on roads with heavy traffic. Manual surveying work itself can thus pose safety risks to workers.

Because of these issues, a more efficient and reliable surveying method that can accurately capture existing conditions has long been desired on-site.

Basics of SfM Processing: An Overview of the Technology That Creates 3D Models from Photos

In recent years, photogrammetry, which reconstructs the three-dimensional shape of a site from photographs, has come into focus. ([Photogrammetry](https://ja.wikipedia.org/wiki/%E5%86%99%E7%9C%9F%E6%B8%AC%E9%87%8F%E6%B3%95) is a method for obtaining geometric features of an object from photographic images.) In particular, SfM (Structure from Motion), a type of computer vision technique, is known for reconstructing 3D models of objects or sites from multiple photos. For example, when images continuously shot from a drone or photos taken from various angles on the ground are input into specialized software, the software detects common feature points between photos and calculates their spatial positions by triangulation to generate point cloud data or polygon meshes as 3D models.

Traditionally, 3D measurement commonly relied on expensive laser scanners (LiDAR), but the spread of SfM technology has made it possible to reconstruct three-dimensional data even from photos taken with consumer digital cameras or smartphones. The Ministry of Land, Infrastructure, Transport and Tourism’s *i-Construction* initiative (the MLIT’s DX measures for construction) is also adopting drone aerial photographs for as-built management as a new method. The advantage of SfM is that it enables non-contact, high-density measurement. While manual methods can only measure dozens of points, photogrammetry can acquire tens of thousands to millions of points at once, capturing fine surface details of terrain and structures. Moreover, photography and computation are automated, so anyone who can handle a camera can acquire the data without special skills. In other words, we have entered an era where just taking photos can create full-scale 3D models of a site.

Flow: Shoot with a Smartphone → SfM Processing → Point Cloud Generation (A Labor-Saving Process Anyone Can Do)

Although 3D measurement using photogrammetry requires specialized software, cloud services have recently become robust and the procedure is very simple. Let’s look at the common workflow for smartphone surveying.

• Shoot the site with a smartphone: No special equipment is required; a commercially available smartphone becomes the camera. Photograph the target (construction site or structure) from various angles with a smartphone. The key is to take highly overlapping photos that surround the subject (take multiple shots from slightly different positions so that images overlap by about 60–80% or more). For example, when surveying a road section, take continuous photos alternating from both sides of the road with overlap. Smartphones are easy for anyone to use, and a site worker can easily record necessary locations themselves.

• SfM processing (photogrammetric analysis): Upload the collected photos to a cloud-based SfM processing service. Image analysis is performed automatically on the server, and 3D modeling is executed. No difficult parameter settings are required, and some solutions allow analysis to start with one click of a button. For example, LRTK’s cloud service can process on the order of hundreds of photos into 3D point clouds in about 30–40 minutes with high-speed processing. Because heavy computation is handled by the cloud, a high-performance PC is not required, and you can upload from the field via a tablet and check the results the same day.

• Point cloud data generation and utilization: After processing is complete, high-density point cloud data is obtained. This is a collection of countless 3D coordinate points that reflect the site’s shape and reproduce the surfaces of terrain and structures in detail. Point clouds can be viewed with dedicated viewers or on cloud systems, and various analyses are possible such as measuring distances, angles, areas, volumes, and creating cross-sections. By replacing information that was conventionally measured manually with digital measurement, you can quickly, accurately, and comprehensively grasp the current state.

Through this process, even without experienced surveyors, anyone can perform 3D measurement on-site. This method using smartphones and the cloud delivers significant labor savings and quality improvement for small and medium-sized construction companies and technicians managing sites with few personnel. However, the 3D models obtained by ordinary photogrammetry remain in a relative coordinate system (arbitrary scale and position). This is where the power of RTK positioning for accurate alignment comes into play. Next, let’s look at the benefits of assigning absolute coordinates to photo data using LRTK.

Benefits of Adding High-Precision Coordinates with LRTK (Filling Blind Spots, Automatic Photo Tagging, Design Difference Comparison)

LRTK (Lefixea RTK) is a pocket-sized RTK-GNSS receiver developed by Lefixea, a startup from Tokyo Institute of Technology, that is attached to and used with a smartphone. RTK (Real Time Kinematic) is a satellite positioning correction technique that enables centimeter-level positioning with a smartphone. When photos are taken using this LRTK, high-precision coordinate tags (geotags) can be added to each photo. In other words, the point cloud obtained from SfM processing can be given an accurate public coordinate system from the start. LRTK positioning accuracy is on the order of about 1–2 cm horizontally and vertically, which is a stark contrast to conventional GPS-equipped smartphones (errors of several meters). Here are three concrete benefits this high-precision positioning brings to site surveying.

• Filling blind spots: The biggest advantage of obtaining high-precision absolute coordinates is that datasets can be perfectly overlaid. For example, when integrating a point cloud obtained by drone with one obtained from the ground using a smartphone, if both have absolute coordinates they will automatically align in the correct positional relationship. Areas the drone cannot capture, such as under bridge girders or shaded areas under trees, can be perfectly filled in by ground-based LRTK point clouds. Because data acquired on different days are always placed on the same reference coordinates, there is no worry about shifts between measurements at different construction stages. This unified coordinate-based data integration enables construction of a precise 3D model that comprehensively covers the entire site.

• Automatic photo tagging: Photos taken with the LRTK-linked app automatically record positional information such as latitude, longitude, and elevation at the time of shooting. In the cloud, each photo is pinned on drawings or maps, making it immediately obvious “which location the photo is of.” Traditionally, one had to write numbers on drawings and cross-reference with photo albums, but with LRTK that is unnecessary. Photo management human errors are reduced, and photos with spatial information remain as assets for the future. It is also easy to link photos to point cloud data, so finding “which photo captured this part of the site” is a one-click operation. This kind of automatic tagging greatly improves the efficiency of organizing and sharing site records.

• Comparison with design data: Since point clouds obtained with LRTK are positioned in a public coordinate system, direct comparison with design drawings or BIM/CIM models is possible. By overlaying design data (for example, a completed 3D model or elevation data) with the as-built point cloud in the cloud, you can automatically determine where conditions match the design and where differences exist. Specifically, you can create an as-built heat map that color-codes deviations from the design on the point cloud, making surface irregularities or excesses/deficits immediately identifiable. If the as-built matches the design, colors might show green to blue; if deviations exceed tolerance, they appear red—so construction defects are intuitively visible. The volumes of material lacking or in excess are also calculated instantly, enabling immediate estimation of the rework required. Whereas quality was previously checked only at certain measurement points, point cloud comparisons allow checking across the entire area, dramatically improving the accuracy and reliability of as-built inspections.

By utilizing LRTK in this way, point cloud data obtained with a smartphone can gain accuracy and added value comparable to surveying instruments. In fact, LRTK’s positioning accuracy approaches that of Grade 1 GNSS surveying instruments certified by the Geospatial Information Authority of Japan, and the generated point clouds have the accuracy based on the public surveying reference coordinate system. Next, let’s look at concrete case examples of how this technology is actually used on sites for as-built management.

Concrete Cases of As-Built Management Using LRTK Smartphone Surveying + SfM

Smartphone surveying using LRTK and SfM has begun to be used for as-built management in various civil engineering works such as road construction, slope works, and land development. Here are several representative use cases.

• Quality control in road construction: Smartphone surveying is effective for measuring subgrade and pavement thicknesses. Traditionally, after construction the road width, thickness, and height were measured every tens of meters to check whether they were within prescribed tolerances. However, unevenness between measured points could not be fully captured, risking missed variations in flatness. Scanning the road surface with an LRTK-equipped smartphone allows continuous acquisition of the height distribution across the entire road surface. Comparing the obtained point cloud with the design elevation immediately visualizes on a color map where thickness is insufficient or where there are unwanted bulges. As a result, corrective work can be precisely targeted before inspection, reducing rework.

• Shape confirmation in slope works: Smartphone surveying is a safe and effective way to confirm slope gradients and shaping. Measuring as-built slopes manually required dangerous tasks like entering the slope to measure length with a tape or using a total station from a distance to measure point heights. With an LRTK smartphone, you can photograph the entire slope from a safe distance and obtain the 3D shape in a short time. By cutting arbitrary longitudinal and cross-sections from the generated point cloud, you can instantly determine whether the slope gradient matches the design. Moreover, comparing the point cloud with the design model immediately reveals shortages or excesses in fill or cut, allowing prompt correction of as-built variations.

• Earthwork volume management in land development: In housing development or large excavation sites, knowing the volumes of material moved and confirming as-built shape is important. Traditionally, pre- and post-construction surveys were used to calculate volumes, but manual surveying led to coarse mesh estimates with limited accuracy. Using smartphone surveying, you can scan the terrain in detail before and after or during work and compare to calculate massive fill and cut volumes in a few clicks. For example, at one site, LRTK point clouds were used to automatically calculate daily backfill volumes in the cloud, supporting real-time progress management. This enabled revising hauling plans and streamlining the preparation of submission documents for as-built inspections.

As shown above, as-built management with a smartphone + LRTK allows one person to safely and quickly grasp site conditions, visualizing construction quality and defects with data. For municipalities and small contractors that manage sites with limited staff, this becomes a powerful ally for quality control.

Remote Management via Cloud Integration: Division of On-Site Shooting and Office Analysis

To maximize the power of smartphone surveying, integration with cloud services is indispensable. By processing and sharing data on the cloud, role division between the field and the office and remote progress checks become easy.

First, photos taken with an LRTK smartphone and the generated point cloud data are uploaded to the cloud for centralized management. On the cloud, not only the on-site staff but also supervisors or clients located remotely can view site data in real time. For example, a site technician can shoot photos in the morning and convert them into a point cloud, and in the afternoon staff at headquarters or government offices can review the model on an office PC. This makes it possible to understand and instruct on construction status remotely without traveling to the site, reducing transit time and enabling faster decision-making.

Next, the division of on-site shooting and office analysis is another cloud-specific benefit. On-site personnel can perform the quick photo capture, while heavy processing and detailed drawing checks are handled by office staff. If on-site workers collect data and headquarters technicians analyze the data and prepare reports, few people can efficiently manage multiple sites. Even small companies and municipalities that cannot station surveyors at each site can realistically operate with local staff performing photography and central staff validating the data.

Furthermore, accumulating data on the cloud promotes information sharing and reuse. As-built point clouds and photos can be instantly shared among project stakeholders and used directly for electronic submission to clients. For long-term projects, intermediate data can be stored chronologically to help with plan changes and as-built evaluations. In future projects where the same location is excavated again, past 3D data saved in the cloud can be referenced to predict the positions of buried objects. With cloud integration that combines remote management and data accumulation, smart construction management that transcends time and place constraints becomes possible.

Conclusion: The Significance of Introducing Smartphone Surveying with LRTK and the Future Standardization

The new surveying method that combines smartphones, SfM, and LRTK is fundamentally transforming on-site surveying, which previously required significant effort and experience. As shown in this article, smartphone surveying addresses the traditional issues of labor shortages, workload, and variable accuracy and points the way to solutions. The ability to perform centimeter-accurate 3D surveying with a tool as universally accessible as a smartphone is revolutionary for improving field productivity.

There are two major points to the significance of introducing smartphone surveying with LRTK. First is the immediate on-site improvement effect. By saving labor in surveying, personnel and time can be conserved, and the comprehensiveness and precision of the obtained data raise the level of quality control. Early detection and correction of construction errors become easier, contributing to safety assurance and cost reduction. Second is the establishment of a technology foundation for the future. The construction industry is expected to increasingly standardize the use of three-dimensional data and digital construction. Initiatives like the MLIT’s *i-Construction* promotion are supporting ICT adoption in the field. In this context, adopting LRTK smartphone surveying now can be seen as a forward investment toward future operational standards.

Field DX offers great benefits not only to large companies but especially to small and medium-sized construction firms and municipalities. Affordable, smart surveying using smartphones and the cloud is an ideal solution for these stakeholders. If smartphones become ubiquitous as a “universal surveying device” for each worker, surveying and as-built management will be fundamentally updated. Consider introducing smartphone surveying with LRTK and applying it to next-generation site operations. It will likely become the new normal style of future construction sites.