Updating Site Surveying with LRTK×SfM Processing: Easy High-Accuracy 3D Measurement with a Smartphone

Surveying and as-built management on construction sites are essential processes for safe, high-quality construction. However, traditional surveying methods have faced many challenges, such as requiring manpower and time and producing variable accuracy. Amid this, a new approach combining smartphones and photogrammetry has emerged and attracted attention. This is “smartphone surveying” that merges LRTK (a smartphone-mounted high-precision GNSS receiver) and SfM processing (a technique that generates 3D models from photos). This article explains the issues with conventional surveying, the basics of this technology, concrete use cases, and remote management using the cloud, and examines the significance of introducing smartphone surveying with LRTK and the potential for future standardization.

Challenges of Conventional On-Site Surveying

In civil engineering and construction sites, surveying (as-built management) is performed to confirm that completed structures have been constructed according to design. Traditional surveying was mainly done by manual methods and had many issues. Typical approaches had surveyors and site technicians using tools such as levels, tape measures, and total stations to measure height, width, thickness, and so on at key points of the work area one point at a time and compare them with design values on drawings. It was also necessary to take photos during and after construction for records. The following challenges have been pointed out for this kind of conventional on-site surveying.

• High manpower and time burden: Manually measuring many points on site requires multiple crews and long hours. Office work to compile measurement results into tables and charts also placed a heavy burden on site technicians. Arranging skilled surveyors was necessary, and with staff shortages, progressing efficiently within the construction schedule was not easy.

• Limited measurement points make it hard to grasp the whole picture: The number of points that can be measured manually is limited, so it is difficult to completely capture wide areas or complex shapes. Measuring only limited points risks missing subtle differences from the design drawings. The larger the structure, the higher the risk of overlooking localized unevenness or slight errors, leading to situations where differences are only discovered at inspection and hurried corrective work is required.

• Risk of human error: On busy sites, human errors such as forgotten measurements or recording mistakes can occur. For example, if a photo is not taken before burying a buried object, there will be no evidence after completion, which in the worst case can lead to rework or disputes. Because measurements are manual, there are weaknesses such as “only being able to measure points” and “human recording cannot be made error-free,” placing great pressure on site personnel.

• Safety issues: Conventional methods sometimes force workers to enter dangerous areas, such as high places, slopes, or busy roads for measurements. Manual surveying work itself can lead to safety risks for workers in many cases.

Because of these issues, there has been a strong demand on sites for a more efficient and reliable surveying method that can grasp the actual situation.

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

Recently attracting attention is photogrammetry, a method for reconstructing three-dimensional shapes of sites from photos ([photogrammetry](https://ja.wikipedia.org/wiki/%E5%86%99%E7%9C%9F%E6%B8%AC%E9%87%8F%E6%B3%95) is a technique for obtaining geometric features of objects from photographic images). In particular, SfM (Structure from Motion), a type of computer vision technology, is known as a method for reconstructing 3D models of objects and sites from multiple photos. For example, when images continuously captured by a drone from the air or a set of photos taken on the ground from various angles are input into specialized software, the software detects common feature points between images, calculates their spatial positions by triangulation, and generates 3D models such as point cloud data or polygon meshes, which are aggregates of points.

Traditionally, 3D measurement generally meant using expensive laser scanners (LiDAR), but with the spread of SfM technology, three-dimensional reconstruction is now possible from commercially available digital cameras and smartphone photos. The Ministry of Land, Infrastructure, Transport and Tourism’s *i-Construction* (the construction DX initiative proposed by the Ministry of Land, Infrastructure, Transport and Tourism) is also adopting drone aerial photos for as-built management as a new method ([Ministry of Land, Infrastructure, Transport and Tourism](https://www.mlit.go.jp/tec/i-construction/index.html)). The advantage of SfM is non-contact, high-density measurement. Where manual methods can measure only a few dozen points, photogrammetry can acquire tens of thousands to millions of points at once, capturing minute surface irregularities of terrain and structures. Moreover, shooting and computation are automated, and anyone who can operate a camera can acquire data without special skills. In other words, we have entered an era where you can create a full-scale 3D model of a site simply by taking photos.

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

3D measurement by photogrammetry requires specialized software, but recently cloud services have become more robust and the procedure is very simple. Let’s look at the typical flow of smartphone surveying.

• Shoot the site with a smartphone: No special equipment is required; a commercial smartphone serves as the camera. Photograph the target (work area or structure) from various angles with the smartphone. The key is to take highly overlapping photos that surround the subject (take multiple photos from slightly different positions so that the images overlap by 60–80% or more). For example, to measure a road section, take continuous photos with overlapping shots alternately from both sides of the road. Smartphones are easy for anyone to use, and site workers themselves can easily record necessary areas.

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

• Point cloud data generation and utilization: After analysis is complete, you obtain a high-density point cloud as the result. This is a collection of countless 3D coordinate points reflecting the site’s shape, reproducing the surfaces of terrain and structures in detail. Point clouds can be viewed in dedicated viewers or cloud systems, and various analyses are possible, such as measuring distance, angle, area, and volume, or creating cross sections. By replacing information that was previously measured manually with digital measurement, you can quickly, accurately, and comprehensively grasp the current situation.

With the above process, even without seasoned surveyors, anyone can perform on-site 3D measurement. This smartphone-and-cloud method will bring significant labor savings and quality improvements to small and medium-sized construction firms and technicians who manage sites with few people. However, 3D models obtained by ordinary photogrammetry remain in a relative coordinate system (arbitrary scale and position). This is where high-precision positioning by RTK shines. Next, let’s look at the benefits of giving photo data absolute coordinates using LRTK.

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

LRTK (Lefixea RTK) is a pocket-sized RTK-GNSS receiver developed by Lefixea, a venture company originating from the Tokyo Institute of Technology, designed to be attached to a smartphone. It leverages the satellite positioning correction technique RTK (Real Time Kinematic) to enable centimeter-level positioning with a smartphone. By shooting photos using this LRTK, you can attach high-precision coordinate tags (geotags) to each photo. In other words, point clouds obtained from SfM processing can be given precise public coordinate system positions from the start. LRTK positioning accuracy is on the order of about 1-2 cm (1-2 cm (0.4-0.8 in)) horizontally and vertically, which is far beyond that of GPS-equipped smartphones (errors of several meters). Here are three concrete benefits that this high-precision positioning brings to on-site surveying.

• Filling blind spots: The greatest advantage of obtaining high-precision absolute coordinates is that data can be perfectly overlaid. For example, when integrating point clouds obtained by drone and point clouds obtained from the ground with a smartphone, both datasets will automatically align correctly if they have absolute coordinates. Areas that the drone point cloud cannot capture—such as the undersides of bridge girders or shaded areas under trees—can be completely filled by LRTK ground point clouds. Multiple datasets measured on different days are always placed on the same reference coordinates, so there is no worry about shifts between surveys in staged construction. This unified coordinate-based data integration enables construction of a precise 3D model that covers the entire site without omission.

• Automatic photo tagging: Photos taken with an LRTK-linked app automatically record position information such as latitude, longitude, and height at the time of shooting. In the cloud, each photo can be pinned on drawings or maps, making it immediately clear “which location this photo is of.” Previously, it was necessary to write numbers on drawings and match them to a photo album, but LRTK eliminates that step. Human errors in photo management are reduced, and photos with spatial information remain assets for the future. It is also easy to link photos to point cloud data, so finding “which photo shows this part of the site” becomes a one-click operation. Such automatic tagging dramatically improves the systematization and sharing of site records.

• Comparison with design data: Point clouds obtained with LRTK are located in a public coordinate system, so direct comparison with design drawings or BIM/CIM models is possible. By overlaying design data (for example, a 3D model of the completed shape or elevation data) with as-built point clouds in the cloud, you can automatically determine where conditions match the design and where there are deviations. Specifically, you can create an as-built heat map that color-codes differences from the design on the point cloud, instantly revealing bumps and deficiencies. If there is no deviation between the design surface and the actual surface, it might be shown in green–blue, and deviations outside the standard are shown in red, making identification of construction defects intuitive. The volume of soil lacking or in excess can also be calculated immediately, so estimates of corrective work quantities can be made instantly. Compared to the conventional method of checking only some measurement points, point cloud comparisons allow checking across the whole site, dramatically improving the accuracy and reliability of as-built inspections.

By utilizing LRTK in this way, point clouds acquired with a smartphone can be given accuracy and added value comparable to surveying instruments. In fact, LRTK 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 of coordinates based on public surveying reference systems. Next, let’s look at concrete examples of how these technologies are actually used on sites in 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 in as-built management for various civil engineering works such as road construction, slope works, and land development. Here are some representative applications.

• Quality control in road construction: Smartphone surveying is effective for measuring roadbed and pavement thickness. Traditionally, after the work was completed, road width, thickness, and height were measured every several tens of meters and checked to be within allowable ranges. However, unevenness between points could not be fully captured, risking missing flatness irregularities. Scanning the pavement with an LRTK-equipped smartphone allows continuous acquisition of the road surface height distribution. By comparing the obtained point cloud with design elevations, you can immediately visualize on a color map where thickness is insufficient or where there are excess buildups. As a result, corrective work can be accurately undertaken before inspection, reducing rework.

• Shape confirmation in slope works: Smartphone surveying is safe and effective for checking slope gradients and shaping conditions. Measuring slope as-built manually required dangerous tasks such as entering steep slopes with a tape to measure slope length or using a total station from a distance to measure point heights, which was demanding. With an LRTK smartphone, the entire slope can be photographed from a safe distance and 3D shape obtained in a short time. Cutting arbitrary longitudinal and cross sections from the generated point cloud makes it possible to immediately judge whether slope gradients meet design. Moreover, comparing the point cloud with the design model instantly reveals shortages or excesses of embankment or excavation, enabling prompt correction without missing variances in as-built conditions.

• Earthwork volume management in land development: In housing site development and large excavation sites, tracking the volumes of soil moved and as-built confirmation is important. Traditionally, surveys before and after construction were used for volume calculations, but manual surveys yielded coarse mesh estimates with limited accuracy. Using smartphone surveying, you can scan the terrain in detail before and after construction or during progress, and calculate vast fill and excavation volumes in a few clicks by comparison. For example, on one site, daily backfill volumes were automatically calculated in the cloud based on LRTK point clouds, aiding real-time progress management. This enabled revising transport plans and streamlining preparation of submission materials for as-built inspections.

As described above, smartphone + LRTK as-built management allows one person to safely and quickly grasp site conditions, visualizing construction quality and defects with data. For municipalities and small contractors managing sites with limited personnel, this is 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 performing data processing and sharing in the cloud, role division between site and office and remote progress checks become easy.

First, photos taken with an LRTK smartphone and generated point cloud data are uploaded to the cloud for centralized management. In the cloud, not only site personnel but also supervisors or clients at remote offices can view site data in real time. For example, a model photographed and point-clouded by a field technician in the morning can be checked by headquarters or municipal staff on an office PC in the afternoon. This allows understanding and directing construction remotely without having to visit the site, reducing travel time and enabling faster decision-making.

Next, the cloud also enables division of on-site shooting and office analysis. On site, workers can focus only on quickly shooting photos with a smartphone, leaving heavy analysis and detailed drawing checks to office staff. If on-site staff collect data and headquarters technicians perform data analysis and report creation, multiple sites can be efficiently managed by a small team. For small companies and municipalities that cannot station surveyors at every site, it becomes realistic to have local staff only do shooting while central office personnel perform data validation.

Furthermore, accumulating data in the cloud promotes information sharing and reuse. As-built point clouds and photos can be shared instantly among project stakeholders and used directly for electronic delivery to clients. For long-term projects, interim data can be stored chronologically to support plan changes and as-built evaluations. In the future, when excavating the same site for another project, past 3D data stored in the cloud could be referenced to predict the locations of buried objects, among other uses. Cloud integration that combines remote management and data accumulation enables smart construction management that transcends time and place constraints.

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

The new surveying method that combines smartphones, SfM, and LRTK is transforming on-site surveying, which previously required significant effort and experience. As discussed in this article, the traditional issues of staff shortages, work burden, and accuracy variability can be addressed by smartphone surveying. The ability to perform centimeter-precision 3D surveying comparable to specialized equipment using a tool as ubiquitous as a smartphone is revolutionary for improving site productivity.

There are two major significances to introducing smartphone surveying with LRTK. First is the immediate on-site improvement effect. Labor-saving in surveying saves personnel and time, and the comprehensiveness and precision of the acquired data raise quality control. Early detection and correction of construction mistakes become easier, leading to enhanced safety and cost reduction. Second is the establishment of a future-oriented technical foundation. The construction industry is expected to increasingly standardize the use of three-dimensional data and digital construction. With initiatives like the Ministry of Land, Infrastructure, Transport and Tourism’s *i-Construction*, the administration is also promoting ICT on sites. In this context, adopting LRTK smartphone surveying now is an investment in future operational standards.

Site DX brings 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 such organizations. If smartphones proliferate as a “universal surveying device” for each person, the way surveying and as-built management are done will be fundamentally updated. Please consider introducing smartphone surveying using LRTK to contribute to next-generation site operations. It is likely to become the new norm in future construction sites.