Convert GNSS Receiver Data Directly into Point Clouds! High-Precision 3D Surveying with a Smartphone

In recent years, combining smartphones with GNSS receivers has made 3D surveying, which previously required specialized equipment, accessible to anyone. By attaching a compact GNSS receiver to a smartphone, you can scan the surroundings on site and acquire point cloud data. 3D surveying, which was traditionally handled by expensive laser scanners and drones, can now be realized with the familiar smartphone + GNSS combination and high-accuracy results. This article explains the mechanisms and benefits of high-accuracy 3D surveying using smartphones and GNSS receivers, focusing on point cloud generation by photogrammetry (SfM).

What is a GNSS receiver? How it differs from a smartphone's built-in GPS

First, let's clarify what a GNSS receiver is. GNSS stands for "Global Navigation Satellite System" and is a collective term for satellite positioning systems of various countries such as GPS, GLONASS, Galileo, and Michibiki (QZSS). GNSS receiver refers to a dedicated device that receives signals from these satellites and calculates its position. Smartphones also have GPS receivers built in, but the typical positioning error of a smartphone GPS is about 5–10 m (16.4–32.8 ft). For applications that require centimeter-level accuracy (cm level accuracy (half-inch accuracy)), such as construction surveying, this is insufficient.

That's where high-precision GNSS receivers come in. In the representative high-precision positioning method, RTK (Real Time Kinematic), positioning errors are corrected in real time using correction information from a base station. This can reduce position errors to within a few centimeters (a few inches). For example, by using the centimeter-level augmentation service (CLAS) provided by Japan's Quasi-Zenith Satellite Michibiki or correction data from reference-station networks delivered over the Internet, positioning accuracy on smartphones can be dramatically improved. By using a small, smartphone-mounted RTK-GNSS receiver (GNSS receiver), you can obtain your current position with a level of positioning accuracy that a smartphone’s built-in GPS cannot achieve.

Point Cloud Generation with Photogrammetry (SfM): A Smartphone Camera Becomes a 3D Scanner

Next, let's look at point cloud generation by photogrammetry using a smartphone camera. SfM (Structure from Motion) is a technique for reconstructing three-dimensional shapes from multiple photographic images. It analyzes many photos taken with digital cameras or smartphones and automatically matches common features of the depicted objects to estimate camera positions and the objects' 3D shapes. As a result, overlapping photos generate the object's point cloud data and 3D models (meshes). Because 3D models of the surroundings can be built using only a camera without special laser equipment, photogrammetry (photogrammetry) has attracted significant attention in recent years.

In the field of civil engineering and construction, initiatives to create high-precision point clouds and orthophotos in a short time by applying SfM processing to images captured with smartphones and drones are spreading. For example, with the momentum of the Ministry of Land, Infrastructure, Transport and Tourism’s *i-Construction*, cases of using photogrammetry for as-built management and construction progress records have been increasing. Below are the main use cases of point cloud data obtained by photogrammetry.

• As-built management and quality inspection: Record post-construction terrain and structures as point cloud data and verify geometric dimensions by comparing them with design data. Deviations on the millimeter level (mm, ≈0.04 in) can be detected, and this is also useful for checking whether embankments and slope gradients match the design. In addition, 3D models generated from photographs include color data, which helps visually inspect concrete placement and other appearance-related quality aspects.

• Progress management: By photographing the site at each stage and creating 3D models, you can understand construction progress three-dimensionally. For example, by weekly drone aerial photography of an earthworks site and converting it to a point cloud, you can intuitively check from a remote location "how far excavation has progressed" and "how much fill has been placed." Details that are hard to grasp on 2D drawings become immediately clear on 3D models, making it easier to share progress between the field and the office.

• Current-condition surveying and quantity management: Even complex terrain can be surveyed over a wide area in a short time using photogrammetry. Hilly terrain that tends to have measurement gaps with traditional total station or GPS surveying can be captured with SfM point cloud measurement, recording countless points that would be difficult to obtain manually in one pass. Ground and structure distortions and crack locations can also be modeled in detail, allowing arbitrary dimensions and cross-sections to be extracted later.

• Earthwork volume calculation: Using terrain models obtained by SfM can streamline volume calculations for fills and excavations. On-site practices include comparing pre- and post-construction point cloud data to calculate volume differences and measuring stockyard material piles by photogrammetry. If software automatically computes the volume of the required area on the point cloud, it enables faster and more accurate earthwork volume estimation compared to traditional cross-sectional area calculations.

Photogrammetry-based 3D point cloud generation is useful across a wide range of tasks, from quality control and progress monitoring to quantity estimation. The fact that 3D data can be obtained without hiring specialist contractors—using only photos taken with a digital camera or a smartphone—is revolutionary, and adoption is progressing even at small- to medium-sized sites.

Challenges of SfM Surveying: Difficulty of Alignment and Ensuring Accuracy

SfM methods for generating point clouds from photographs are convenient and powerful, but there are also challenges in ensuring surveying accuracy. To align the generated 3D models and point cloud data with the correct real-world coordinate system, certain measures have traditionally been necessary. A representative example is the placement of GCPs (Ground Control Points, ground reference points).

In photogrammetry, several targets (markers) with known coordinates are placed on the ground to match the model's scale and position to reality. For example, when aerially photographing a large development site, about 5–10 ground control points (GCPs) are pre-surveyed and installed at the four corners and the center of the area, and photos are taken so that they appear in the images. During SfM processing, GCP targets are marked on each photo and assigned their corresponding real-world coordinates to correct the entire model to a geographic coordinate system. This allows the model to be scaled and to be assigned absolute coordinates such as latitude/longitude and elevation.

However, GCP installation requires effort and manpower. Surveying each point to high precision is laborious, and for large sites or routine flights it is unrealistic to place and measure GCPs every time. If you cannot deploy a sufficient number of GCPs, the generated point cloud will remain in a local relative coordinate system, risking scale errors and overall positional shifts. Also, in areas with poor footing or hazardous locations it may be physically impossible to install targets, which has been a bottleneck for photogrammetry.

In this way, "how to assign accurate coordinates to models derived from photographs" has long been a challenge, but in recent years a solution has emerged. That is direct georeferencing using high-precision GNSS, i.e., a method of directly embedding location information into photos at the time of capture.

Directly Assign Coordinates to Photos with GNSS: The RTK Revolution in Automatic Geotagging

The combination of a smartphone and a GNSS receiver has made it possible to record precise positional coordinates for each photo in real time. By attaching a compact RTK-capable GNSS receiver to the smartphone and shooting while receiving satellite correction information via a dedicated app, the capture position (latitude, longitude, altitude) of each photo is automatically recorded with centimeter-level accuracy (half-inch accuracy). In other words, the idea is to rapidly take photos in the field with high-precision geotags attached.

When the photographs acquired by this method are subjected to SfM analysis, each image already has precise coordinates assigned, so the scale and positional relationships between photos automatically align to real-world space during the initial stages of model generation. Without the post-processing by GCPs that used to be necessary, you can obtain a point cloud model with accurate dimensions and coordinates from the start. By also using data from the smartphone's built-in tilt sensor (IMU) and electronic compass, the camera orientation can be estimated, allowing efficient, high-precision 3D reconstruction with a minimum of measurements.

This "direct coordinate assignment with GNSS" approach is a groundbreaking technology that dramatically lowers the barriers to ensuring accuracy in photogrammetry. For example, the solution called LRTK offered by Refixia Co., Ltd. allows anyone to perform centimeter-level positioning (half-inch-level positioning) easily by using a small smartphone-mounted RTK-GNSS receiver and a dedicated app. The previously difficult high-precision surveying with smartphones has become readily and affordably achievable thanks to LRTK.

3D surveying with just a smartphone: On-site measurement procedures

Now, let's take a look at the basic workflow for on-site 3D surveying using a smartphone and a GNSS receiver. Even without specialized training, anyone can obtain high-accuracy point cloud data by following the steps below.

• Attaching the GNSS receiver: Attach a compact GNSS receiver compatible with the smartphone (e.g., a magnet-mounted RTK receiver). Pair the smartphone and receiver via Bluetooth and configure them so that positioning data is transmitted to the phone.

• Configuring positioning corrections: Launch the GNSS receiver app and configure it to receive RTK correction information. By obtaining correction data from a regional base station over the network or receiving augmentation signals from satellites, real-time high-precision positioning on the smartphone becomes possible.

• Imaging and scanning: Use the smartphone camera to photograph the area to be surveyed. Walk slowly while capturing the subject and terrain from various angles. Blind spots that are difficult for drones to capture, such as narrow spaces or the backs of structures, can be fully covered if photographed handheld by a person. While shooting, the app records the position coordinates obtained from GNSS at each shutter moment and saves them as geotags in the photo files.

• Data processing: After shooting is complete, upload the photo data to cloud or PC SfM software for analysis. If using a dedicated cloud service, point cloud generation can be carried out rapidly simply by sending photos directly from the field. By using photos with high-accuracy coordinates, the resulting point cloud will already conform to the geodetic reference frame.

• Reviewing results: Check the generated 3D point cloud and model on the smartphone. Rotate and zoom the point cloud in a cloud viewer or app, and measure distances, areas, and volumes as needed. Because you can immediately verify as-built conditions on-site and check for any missed captures, you can utilize measurement results on the spot. Also, since data is automatically shared via the cloud, you can share information in real time with staff who are not on site.

As a result of the above process, 3D surveying that can be completed with just a smartphone becomes possible. Because a single person can perform all measurement tasks, it also helps reduce staffing needs significantly. There is no need to carry heavy equipment or remain in hazardous locations for extended periods, making it safe and efficient.

Benefits of Smartphone × GNSS Surveying: Convenient, High Accuracy, Low Cost

Surveying methods that use smartphones and GNSS receivers offer various advantages not found in traditional surveying instruments. Finally, let’s summarize their main benefits.

• High accuracy: By using an RTK-capable GNSS receiver, positioning can achieve a horizontal accuracy of a few centimeters (a few in). You can accurately capture terrain undulations and structural dimensions, offering a level of precision that separates it from standalone GPS positioning. The resulting point clouds can also achieve positional accuracy of 20 mm (0.79 in) or better, ensuring quality comparable to laser scanners.

• Ease of use: Measurement is possible with just a smartphone and a small receiver, making equipment transport easy. Sites that traditionally required tripods or stationary instruments can be surveyed by simply walking around with lightweight gear. The app operation is intuitive, making it easy to use even for those without specialized knowledge.

• Rapid result acquisition: The cycle from photography to point cloud generation is short, allowing quick results on site. With cloud processing, point cloud models can be ready immediately after capture, enabling on-site verification and measurements. This eliminates the need for later office analysis and supports immediate decision-making in the field.

• Cost reduction: Initial costs are significantly lower compared to conventional laser scanners, surveying instruments, or introducing drones. The combination of a high-performance smartphone and a GNSS receiver is more affordable than dedicated equipment, and maintenance costs are minimal. This makes it well suited for small projects and budget-constrained sites.

• Versatility: Handheld capture and photogrammetry can be applied widely, from outdoor terrain surveys to indoor structural documentation. In places where drones cannot fly or GPS reception is weak—under forest canopy, beneath overpasses, etc.—a person can enter and capture images to generate 3D data. Combining drone aerial data with ground-based smartphone captures can produce point cloud models that cover both large areas and fine details.

• Improved safety: Tasks that formerly required climbing to heights or setting equipment in roadways posed safety risks. Smartphone surveying allows zoomed shots from a distance or short-duration data collection, contributing to worker safety. Eliminating long-term setups or observations also reduces physical strain in extreme heat or cold, improving working conditions.

As noted above, point cloud surveying using smartphones + GNSS brings significant advantages in accuracy, efficiency, and cost. The door is opening for 3D measurement—which was previously left entirely to specialist technicians—to be used routinely by anyone on site.

Try a simple survey using LRTK

The LRTK technology that combines a smartphone with a high-precision GNSS is a representative solution that enables this kind of easy 3D surveying. LRTK consists of an RTK-GNSS receiver attachable to a smartphone and a dedicated app. Even without complicated setup or specialized knowledge, using LRTK turns the smartphone on-site into a versatile surveying instrument, allowing anyone to obtain surveying data with centimeter-level accuracy (half-inch accuracy).

If you're tempted to try "high-precision point cloud generation with a smartphone," be sure to try simplified surveying with LRTK. With a level of ease and accuracy that overturns conventional expectations, your on-site operations should become dramatically more efficient. The future of 3D surveying is just around the corner — grab a GNSS receiver and a smartphone and start a new way of surveying.