Make Photogrammetry Even More Accurate: Record Photo Positions with cm-Level Accuracy (half-inch accuracy) Using LRTK

To further improve the accuracy of 3D models obtained by photogrammetry, the key is to replace the per-photo position information from the conventional several-meter accuracy with cm-level accuracy (half-inch accuracy). This article explains in detail how to use the new LRTK technology to record photo positions at cm-level accuracy (half-inch accuracy) and efficiently obtain high-precision point cloud data by a single operator.

Table of Contents

• Why high-precision positioning is needed in photogrammetry

• What RTK is: centimeter-level positioning technology

• What LRTK is: turning a smartphone into a surveying tool

• How to replace photo position information with cm-level accuracy

• Benefits of cm-level photogrammetry

• Simple surveying anyone can do with smartphone LRTK

• Summary

• Frequently Asked Questions (FAQ)

Why high-precision positioning is needed in photogrammetry

Photogrammetry generates three-dimensional models and point cloud data of objects from multiple photographic images. With aerial photography by drones or ground-based shooting, terrain and structures at construction sites can be digitized in a short time, and use of this technology has rapidly expanded in recent years. In civil engineering and construction, point cloud data obtained by photogrammetry are actively used for as-built management (post-construction shape verification), earthwork volume calculations, and infrastructure maintenance. However, to properly utilize point cloud models generated by photogrammetry, you cannot avoid the issue of the "accuracy of position information."

Usually, photos taken by a camera are tagged with position coordinates from GPS, but the positioning accuracy of smartphones and common GPS cameras is typically only on the order of several meters. Even if a 3D model is reconstructed in photogrammetry software, if the original position information is inaccurate, the entire model may be offset from the real-world coordinate system and the scale may be slightly distorted. Left as is, this can cause large errors when overlaying as-built data on drawings or calculating precise volumes. Height (elevation) errors are particularly serious. For example, altitude information obtained from ordinary GPS often has errors of several meters. If a generated point cloud is 50 cm (19.7 in) higher than the actual site, for a flat 100 m (328.1 ft) square area this could lead to volume calculation errors of several hundred cubic meters. Therefore, to use 3D data from photogrammetry accurately, it is important to raise each photo's position information to centimeter-level accuracy (half-inch accuracy) so the model has the correct scale and coordinate reference.

Traditionally, improving photogrammetry accuracy required setting up known points called GCPs (Ground Control Points). Surveying multiple control points at key locations on site and inputting those coordinates into photogrammetry software corrects the overall distortion and offset of the point cloud model. However, placing and measuring a sufficient number of GCPs is time-consuming, labor-intensive, and requires surveying expertise. The larger the area surveyed, the greater the risk that parts with insufficient control points will cause the model to warp or sink (so-called "potato-chipping" deformation). A decisive solution to these issues is directly providing high-precision position information to photos. If each photo's capture location can be recorded with centimeter-level accuracy (half-inch accuracy) at the time of shooting, the model can be aligned to site coordinates with minimal effort. The enabling technology for this is RTK (Real-Time Kinematic) positioning.

What RTK is: centimeter-level positioning technology

RTK is a technique that dramatically improves positioning accuracy by applying real-time corrections to satellite positioning (GPS, GLONASS, etc.) errors. Ordinary GPS positioning contains errors on the order of several meters, but using RTK can reduce that to below a few centimeters. Specifically, a base station (reference station) and a rover exchange correction data to cancel out slight satellite signal biases and atmospheric effects. Traditionally, RTK surveying required installing a base station on site and using expensive GNSS receivers and radios. However, recent technological advances have made correction data more easily available.

In Japan, the Ministry of Land, Infrastructure, Transport and Tourism provides services such as the Continuously Operating Reference Stations network and the Quasi-Zenith Satellite System "Michibiki" offering a centimeter-class positioning augmentation service (CLAS), allowing high-precision correction data to be obtained via the Internet or satellite without a dedicated radio base station. Public surveying VRS (Virtual Reference Station) services can also be contracted and used. In short, today even a small device capable of receiving corrections over a communication link can perform real-time cm-level positioning. With the miniaturization and simplification of RTK technology, solutions that utilize RTK positioning with smartphones have emerged. A representative example is LRTK.

What LRTK is: turning a smartphone into a surveying tool

LRTK (LRTK) is a simple surveying system consisting of an integrated high-precision GNSS receiver device that attaches to a smartphone and a dedicated app. Developed by a startup from the Tokyo Institute of Technology, it is designed to meet Japanese field needs. By attaching an LRTK device ("LRTK Phone") to an iPhone or iPad with a dedicated case or adapter and turning it on, the device receives satellite positioning signals and begins applying real-time corrections. In open-sky conditions, positioning accuracy rapidly improves in about 30 seconds to 1 minute after startup, and the RTK "Fix" solution can be achieved. In Fix status, the current position is recorded on the smartphone with extremely high accuracy of about ±2 cm (±0.8 in).

A major feature of LRTK is its ease of use without the need for a dedicated base station or complicated cabling. Carrying only a smartphone and LRTK device, the system automatically receives CLAS or reference station data over mobile communication, allowing centimeter-level positioning without special operations. The device itself is compact enough to fit in the palm of the hand (weighing a few hundred grams or less), making it easy to transport to the field. A smartphone app lets you check positioning status and accuracy in real time, showing statuses such as "no RTK" (no augmentation, ± several meters), Float (corrections in progress, about ±1 m (±3.3 ft)), and Fix (corrections complete, about ±2 cm (±0.8 in)). Even non-experts can follow on-screen instructions to operate it, making it a solution that truly embodies "RTK anyone can use."

Moreover, LRTK is designed as a multi-purpose field tool, not just a positioning device. It integrates with smartphone cameras and LiDAR sensors to become an all-in-one surveying instrument that handles positioning, photography, point cloud scanning, surveying measurements, and AR navigation. This allows photogrammetry and RTK surveying—previously requiring separate equipment—to be accomplished on a single platform. Next, let's look at the specific method of using LRTK to record the position of each photo with cm-level accuracy (half-inch accuracy).

How to replace photo position information with cm-level accuracy

Using LRTK, you can link high-precision coordinate data to each individual photo you shoot. In other words, you can replace the photo position information, which normally has meter-level errors, with RTK-provided centimeter-level accuracy (half-inch accuracy). The basic procedure is as follows.

1. Preparation: Attach the LRTK device to the smartphone, launch the dedicated app, and start positioning. Confirm that RTK correction information (Michibiki CLAS signals or reference station data via the Internet) is being received in the app. When the positioning mode becomes Fix and high accuracy is established, you are ready. If needed, measure a few reference points on site at this stage for peace of mind (this also helps reduce the number of GCPs required as described later).

2. Shooting: Use the smartphone camera (such as the camera mode inside the LRTK app) to photograph the target or site from various angles. When photographing buildings or terrain, follow photogrammetry best practices, ensuring sufficient overlap between adjacent photos. With LRTK, each time you press the shutter, the capture position (latitude, longitude, altitude) of that photo is automatically recorded with RTK cm-level accuracy (half-inch accuracy) and saved in the photo file's EXIF data. In other words, when you later analyze those photos, each image already has an extremely accurate geotag.

3. Processing: Import the many photos you shot into photogrammetry software to generate a 3D point cloud model. Structure from Motion (SfM) algorithms perform feature matching between photos and compute camera positions and the point cloud. Having high-precision geotags from LRTK allows the model to be reconstructed with position and scale close to real-world coordinates. The software will internally refine camera positions, but with accurate initial coordinates it converges with fewer errors. As a result, the output point cloud data will already be aligned to the site’s survey coordinates (for example, a plane rectangular coordinate system), making additional coordinate transformation work largely unnecessary.

4. Verification & use: Check that the generated point cloud model matches surrounding known points and design drawings. If small offsets in position or elevation are observed, you can fine-tune with auxiliary GCPs. However, if the position information added to each photo by LRTK is accurate, the model will in most cases be within a few centimeters of error. The completed high-precision point cloud can be immediately used for accurate earthwork volume calculations, construction quality checks, and creation of as-built drawings. You can also easily compare point clouds or overlay CAD data to verify consistency.

Through this process, a photo set shot with LRTK yields a 3D model with high-precision coordinate values from the start. The significant advantage is that the many GCP installations and cumbersome coordinate alignment tasks that were previously indispensable can be greatly reduced.

Benefits of cm-level photogrammetry

Raising photo position information to cm-level accuracy (half-inch accuracy) with LRTK brings various practical benefits to photogrammetry. The main advantages are summarized below.

• Improved point cloud accuracy: The absolute accuracy (positional and dimensional correctness) of models improves dramatically. Previously, scale errors of a few percent or positional shifts of several tens of centimeters could occur, but using photos with cm-level geotags (half-inch accuracy) greatly reduces point cloud errors. This ensures high reliability for dimensional checks in as-built management and for earthwork volume calculations.

• Reduced additional surveying and control points: Because high-precision coordinates are already embedded in photos, the number of GCPs required for model alignment can be greatly reduced. In some cases, measuring only a few known points for checking is sufficient, eliminating the need to install nearly 10 control points as before. Pre- and post-surveying tasks are simplified and the overall workflow becomes more efficient.

• Immediate model usability: Point cloud data generated immediately after shooting can be used for various measurements in the site's coordinate system as-is. Since there is no need to repeatedly move or rotate models to match existing drawings, you can analyze and make decisions on-site right away. For example, you could photograph and model in the morning and prepare an as-built report by the afternoon.

• Labor savings and improved safety: High-precision surveying completed with a smartphone and a small device makes "one-person surveying" realistic. The need for multiple personnel carrying heavy equipment or staff to operate rods is reduced, helping address labor shortages and lowering personnel costs. It also minimizes entry into hazardous areas or work near heavy machinery, reducing safety risks.

• Low-cost adoption: By using LRTK, high-precision 3D measurements that previously required specialized surveying firms or expensive equipment can be introduced in-house at relatively low cost. Since it can be combined with commercial drones and smartphones, it enables surveying DX while keeping initial investment down.

Thus, combining photogrammetry and RTK delivers significant gains in both accuracy and efficiency and has the potential to transform future surveying practices.

Simple surveying anyone can do with smartphone LRTK

Introducing centimeter-level positioning with LRTK opens a new stage where "non-surveying professionals can perform high-precision surveys." The ability to perform precise surveying easily with just a smartphone has the potential to dramatically change field workflows.

For example, instead of waiting for a surveying team, a construction manager can walk the site, take photos, generate a 3D model on the spot, and calculate earthwork volumes. Quick PDCA cycles—shooting key locations during morning rounds and producing quantity reports or as-built drawings by noon—become feasible. Intuitive smartphone apps make LRTK easy to operate even for young engineers with limited surveying experience or staff from other departments.

If everyone has a smartphone LRTK, small teams can cover large sites and chronic shortages of survey personnel can be alleviated. Digital tools can supplement tasks that relied on elderly experienced surveyors, supporting generational transition. High-precision data collected can be shared and stored in the cloud in real time, allowing office staff to grasp 3D site conditions remotely and issue instructions. The LRTK series promotes such on-site DX (digital transformation) and aligns with the Ministry of Land’s "i-Construction" initiative as an innovative approach.

In the world of simple surveying enabled by LRTK, "anyone, anytime, anywhere" can perform high-precision surveys. Once you experience this convenience and accuracy, returning to meter-level ambiguous positioning will be hard. Consider introducing LRTK at the front lines of your site to achieve both higher accuracy and efficiency.

Summary

The key to improving photogrammetry accuracy is the precision of the position information attached to photos. Using LRTK, centimeter-level accuracy (half-inch accuracy) that was once beyond ordinary GPS can be easily achieved with a smartphone. By using photos recorded with cm-level coordinates (half-inch accuracy) at the time of shooting, point cloud models are produced from the start with correct coordinate systems and scale, making them immediately usable for as-built management and earthwork calculations.

High-precision photogrammetry with LRTK is a step toward DX in surveying fields. As a compact, low-cost solution to staffing and efficiency challenges, it is receiving attention. For those who want to "make photogrammetry more accurate" or "replace photo position information with cm-level accuracy (half-inch accuracy)," LRTK is a strong option. The era in which anyone can perform centimeter-level surveying has begun. Take this opportunity to expand the possibilities of photogrammetry with high-precision positioning.

Frequently Asked Questions (FAQ)

Q: Smartphones tag photos with GPS—why is RTK necessary? A: Typical smartphone GPS accuracy is on the order of several meters, which is insufficient for obtaining precise 3D models in photogrammetry. If position information is off by several meters, the entire model will be shifted and you cannot measure distances or volumes accurately. Height errors in particular introduce large volume calculation errors. Using RTK reduces errors to a few centimeters or less, allowing photogrammetry results to be used reliably as-is.

Q: Does using LRTK mean I don't need to set GCPs (control points)? A: With high-precision geotags on each photo from LRTK, you can greatly reduce the number of GCPs needed for model alignment. Theoretically you can model without GCPs, but in practice it is reassuring to set a few known verification points. Even when shooting with a drone that lacks RTK, measuring about 3–5 reference points beforehand with LRTK makes it easy to align models to site coordinates in photogrammetry software. In short, by using the high-precision coordinates obtained with LRTK as anchors, you can minimize model distortion and offsets.

Q: Do I need special equipment or qualifications to use LRTK? A: No, basically you do not. LRTK consists of a small integrated GNSS receiver and a smartphone app, and you do not need a dedicated base station if you have a communication environment. Satellite augmentation signals and reference station data can be received via the Internet, so you don’t need a radio license to operate your own base station as in the past. All you need is a compatible smartphone, the LRTK device, and a communication line to receive correction data. There is some initial setup, but following the manual allows use without specialized knowledge.

Q: Can people with little surveying experience operate LRTK? A: Yes. The LRTK system is designed to be operated via an intuitive smartphone app. Pressing buttons like "Start Positioning" or "Shoot" performs the necessary processing automatically, so you don’t need specialized GNSS knowledge to achieve high-precision positioning. The app displays accuracy and procedural guidance for on-site use, so users typically get the hang of it after a few uses. There are growing examples of on-site personnel themselves using LRTK to obtain high-precision point cloud data for business use.

Q: What applications are suitable for LRTK high-precision positioning? A: In addition to photogrammetry-based 3D model creation, LRTK is suitable for as-built management and quantity calculations in civil engineering, infrastructure inspections, terrain measurement in forestry and agriculture, and many other uses. LRTK devices also integrate with smartphone AR features to project design drawings or construction lines on site for AR surveying. In short, any scene requiring accurate positioning can benefit beyond construction and civil engineering. Because it is small and portable, it is effective in mountain surveys, disaster site recording, cultural heritage preservation, and other situations where conventional equipment was impractical.