Seven Steps to Automatically Assign Coordinates to Photos and Point Clouds Without Markers

Table of Contents

• Step 1: Confirm and prepare the reference coordinate system

• Step 2: Prepare for high-precision GNSS positioning

• Step 3: Plan photography and measurement

• Step 4: On-site shooting and point cloud acquisition without markers

• Step 5: Process acquired data and automatically assign coordinates

• Step 6: Accuracy verification and correction of results

• Step 7: Use and share the deliverable data

• Summary

In construction and civil engineering surveying and inspection sites, there is an increasing use of photos and 3D point cloud data with location coordinates assigned for practical use. When recording sites for as-built management (post-construction shape verification) or infrastructure inspection, it was conventional to place ground control points (markers), measure their coordinates, and then perform photogrammetry or align point clouds. However, placing and surveying markers is time-consuming and burdensome on site. For example, surveying on steep slopes required risking personnel to install targets, and road surveys required lane closures to set up signage, creating additional effort.

In response to these efficiency needs on site, i-Construction advocated by the Ministry of Land, Infrastructure, Transport and Tourism encourages the use of photogrammetry and point cloud data, and these practices are rapidly spreading. Advances in digital technology in recent years have made practical methods available to automatically assign coordinates to photos and point clouds without using markers. Thanks to high-precision GNSS positioning technology, improvements in photogrammetry software, and the use of smartphones, it is now possible to obtain accurate location information while omitting cumbersome preparatory work. This article explains concrete procedures for automatically assigning coordinates without markers in seven steps. By appropriately combining technologies according to site conditions and needs, even without a specialist surveying team, a single person can efficiently complete measurements and immediately utilize the acquired data.

Step 1: Confirm and prepare the reference coordinate system

First, confirm and set the reference coordinate system that will tie the data together. To overlay surveying data on other drawings or maps, it is important to unify the coordinate system used. For wide-area terrain or infrastructure inspections, using global geodetic coordinates (latitude/longitude and other global coordinates) allows direct display on electronic maps. On the other hand, measurements inside buildings or plant facilities may use proprietary local coordinate systems (based on gridlines or reference lines) from design drawings. In such cases, matching the acquired data to the same reference makes later management easier.

If known reference points (previously surveyed control points) exist on site, check their coordinate values in advance. If necessary, measure control points around the target with GNSS or a total station to understand the site’s coordinate system. This makes it easier to correct offsets when integrating data later. By deciding the reference coordinate system clearly before measurement, the coordinates assigned to photos and point cloud data will be stable, facilitating smooth cross-referencing and comparison with other materials. Also, because GNSS positioning uses ellipsoidal heights, it is advisable to perform elevation conversion using a geoid model as needed to reconcile with known bench marks.

Step 2: Prepare for high-precision GNSS positioning

To obtain accurate coordinates without markers, utilizing high-precision positioning with GNSS (Global Navigation Satellite Systems) is essential. Commercial cameras or smartphones with built-in GPS can add location tags, but their accuracy is limited to about 5–10 m (16.4–32.8 ft). To meet the accuracy required for surveying or as-built management, technologies such as RTK-GNSS that can position to within a few centimeters (a few inches) are necessary.

To perform high-precision GNSS positioning on site, prepare the necessary equipment and settings in advance. Specifically, you need an RTK-capable GNSS receiver (rover) and a connection to a base station or network that provides correction data. If you set up your own base station, place it in an open area and establish it as a known point. Alternatively, connecting to a public GNSS correction service over the Internet to receive virtual reference station correction data is also common. In any case, power up the GNSS receiver before starting measurements and wait for several tens of seconds until you obtain an RTK “Fix” (fixed solution) and confirm that positioning accuracy stabilizes. In an open-sky environment, stable positioning with centimeter-level errors can usually be achieved in a relatively short time. Accuracy decreases as the distance to the base station increases, so it is preferable to use correction information as close to the site as possible.

Note that in environments where GNSS signals are obstructed, such as under elevated structures or in mountainous areas, high-precision positioning may not be achievable. In such cases, take measures like moving the measurement position slightly to secure a view of the sky, or temporarily pausing positioning and re-measuring under clear conditions. In locations where GNSS cannot be received at all, such as inside tunnels or indoors, automatic coordinate assignment by this method is not possible. Alternative measures—such as measuring relative distances with a laser rangefinder or manually recording positions based on known building grids—are required.

Also, even if you cannot receive correction data in real time, the PPK (Post-Processing Kinematic) method, which records GNSS observation data and corrects it later, can achieve similar centimeter-level positioning.

Step 3: Plan photography and measurement

Once the equipment and methods are decided, plan the actual photography and measurement approach. To acquire coordinate-attached data without markers, a shooting plan that sufficiently covers the subject and site is important. First, confirm the target area and terrain, and consider from which positions to take photos or scans to eliminate blind spots. For large development sites or entire structures, using a drone to shoot from above is an option (flight permission and safety measures are required). For narrow areas or locations requiring detail, shoot and scan from the ground with a camera or smartphone.

When performing photogrammetry (SfM), plan shooting positions and the number of images to ensure sufficient overlap. Shoot from various angles with short intervals so that adjacent photos overlap by 60–80% or more. Consider distance to the subject and the camera’s field of view so that both near and far scenes are evenly captured. It is also important to set focus and exposure appropriately and take sharp, blur-free photos. Because you will rely on feature points automatically detected by software without markers, confirm that the subject has sufficient texture or relief (if necessary, consider temporarily placing adhesive tape or other makeshift markers).

When acquiring point clouds with a laser scanner or smartphone LiDAR, visualize the scanning route in advance. Identify areas easily overlooked, such as behind obstacles or at height, and vary standing positions and scan angles to collect an all-around point cloud. When scanning while walking with a mobile device, move slowly at a constant speed to avoid breaking sensor tracking. Preview the data as needed and check onsite for omissions to prevent additional shooting later. By shooting and measuring according to a meticulous plan, you lay the foundation for acquiring reliable data even without markers.

Step 4: On-site shooting and point cloud acquisition without markers

Execute on-site shooting and measurement according to the plan. With the preparations above, you can acquire data without installing markers. When using high-precision GNSS, image capture will automatically record the measured coordinates as geotags each time the shutter is released. In point cloud scanning, global coordinates can be assigned to acquired points in real time. Because the traditional tasks of “installing and measuring markers” are unnecessary, a single operator can start measurements immediately and make efficient use of limited work time.

For example, in a landslide site topographic survey, this method allowed one person to complete the work in about 10 minutes, whereas a conventional pole survey required three people and took approximately 60 minutes in total. This illustrates how preparation through to measurement can be streamlined.

During shooting, collect data while changing camera angles and positions as planned. When flying a drone, pay attention to wind, radio conditions, and safety, and use automated flight plans where possible to execute aerial photography systematically. When shooting with a smartphone or handheld camera, continuously check the operation of the GNSS receiver or smartphone app and ensure positioning does not drop out. Especially when scanning point clouds while walking with a smartphone + GNSS, the device screen will display current accuracy and acquisition status, so proceed while confirming that a Fix solution is maintained. If location information becomes unstable, stop and wait for positioning to recover before resuming to maintain quality.

Although markerless measurement automatically aligns data, be sure to perform on-the-spot checks. For example, review several captured photos on site to confirm adequate overlap and absence of blur. For scans, display the acquired point cloud in real time and inspect for missed areas. Performing these checks on site prevents situations like “when processed back at the office, some areas were missing.” By saving the time that would have been spent placing markers and instead focusing on data quality control during acquisition, you can achieve highly reliable deliverables.

Step 5: Process acquired data and automatically assign coordinates

Bring the photos and point cloud data acquired on site back and process them using dedicated software or cloud services. For photo data, import images into photogrammetry (SfM) software to perform feature matching between images and camera pose estimation (image alignment). The coordinates attached to images at the time of shooting (geotags) are utilized here, and the software automatically assigns scale and absolute coordinates to the entire model. With simple settings, the resulting point cloud or 3D model can be output already aligned to the specified surveying coordinate system. This omits or greatly simplifies the traditional process of inputting multiple target (control point) coordinates after feature point computation to align the model. Recently, cloud-based services that process photo data have appeared; by uploading from the field, they can automatically generate point clouds and orthophotos, enabling results even without a high-performance PC on hand.

For point cloud scanning, if positions were recorded by GNSS at acquisition, the exported point cloud data will already include absolute (world) coordinates for each point. Therefore, when loading multiple scans into software, they will automatically overlap in the correct positional relationship. For example, even if a structure was scanned in two sessions (front and back), each point cloud is recorded in a common coordinate space and will align perfectly when merged. This is because all data are assigned a unified reference coordinate, eliminating the need for manual alignment.

After processing, 3D point cloud data and orthophotos (top-down photographic maps) are generated. Since these deliverables already contain absolute coordinates, they can be imported into CAD or GIS software and directly overlaid on other terrain maps or design data. If your project uses a proprietary local coordinate system, processing software can easily perform coordinate transformations to match that reference. Utilizing automatic coordinate assignment streamlines the entire flow from data acquisition to deliverable creation and enables immediate use of field measurements.

Step 6: Accuracy verification and correction of results

Verify the accuracy of the automatically processed deliverables. Even in workflows without markers, it is important to confirm that final errors are within acceptable ranges for your work. Specifically, compare points on the point cloud or model with known points measured on site to check for coordinate offsets. For example, separately measure a few known corner points or boundary markers on site with GNSS or a total station, and compare those coordinate values with the same locations in the point cloud data. If errors are within a few centimeters (a few inches), the data is generally accurate enough for as-built management or volume calculations. If offsets exceed allowable values, consider corrections in the software.

Photogrammetry software often allows you to add reference point coordinates later and re-optimize the 3D model. If necessary, input known points obtained for verification as control points and adjust the model to fit them. This corrects slight scale errors or positional offsets and yields more reliable coordinate consistency. However, in many cases obtained without markers, large distortions are rare and no correction is needed. What matters most is substantiating that “this data is accurate” through the verification process. Keeping a record of verification results also provides assurance when the data is used downstream.

By performing such verification and correction, you can ensure quality control even without markers.

If verification finds areas that do not meet required accuracy, reinforce the data by acquiring additional photos or adding new control points, then reprocess to improve accuracy.

Combining the efficiency of automatic coordinate assignment with cross-checks using known points is the key to balancing accuracy and operational practicality on site.

Step 7: Use and share the deliverable data

Finally, apply the coordinate-attached deliverables to site operations. 3D point cloud data can be used directly for as-built drawings and quantity calculations. For example, overlaying the post-construction terrain point cloud on the design 3D model allows intuitive assessment of embankment or excavation completion and helps identify nonconforming areas for correction. Comparing pre- and post-excavation surface point clouds to calculate soil volumes enables more accurate earthwork management than before. For bridge or tunnel inspections, high-resolution point clouds allow measurements of required dimensions at a desk and marking and recording of deteriorated areas. Photos with coordinates can be plotted on electronic maps to manage shooting locations or embedded in CAD drawings, making spatial organization straightforward.

Coordinate information greatly facilitates data sharing. With multiple personnel using the same coordinate reference, everyone can clearly understand the location of information within the site. Cloud services now allow uploading point clouds and photos so teams can view and measure them simultaneously. You can share the latest 3D data acquired on site with design staff in the office the same day, immediately discuss problem areas, and issue instructions for additional investigation. Compared to handing over paper drawings or USB drives, this significantly reduces time and effort and decreases misunderstandings among staff through centralized data management.

Using coordinate-attached data smooths information exchange between field and office and streamlines the entire construction and maintenance management process. Accumulated time-series data can also be analyzed to understand long-term changes and support future planning. Collected data can be managed like an “electronic medical record” for each facility, enabling integrated oversight of inspection histories. Advanced applications—such as overlaying past inspection data on the real object using a smartphone’s AR function—are also possible, demonstrating the immense benefits of coordinate-attached data. By fully utilizing high-accuracy data obtained without marker placement and sharing it among stakeholders, you can truly drive DX (digital transformation) of site operations.

Summary

Markerless automatic coordinate assignment in photogrammetry and point cloud measurement greatly contributes to on-site productivity improvements. Omitting cumbersome target placement enables rapid data acquisition and collection of wide-area 3D information in a short time. Since the resulting deliverables are already registered to surveying coordinates, lead time to analysis and sharing is shortened, accelerating decision-making. Eliminating the need for personnel to enter hazardous locations to install targets also improves safety. Moreover, because 3D data acquisition can be completed in-house without relying on expensive specialized equipment or outsourcing, long-term cost savings are expected. Amid labor shortages and workstyle reforms, this workflow enabling few personnel to perform efficient, high-precision surveying is likely to become increasingly prevalent on sites.

Tools that make these latest technologies easy to implement are also emerging. For example, using a small high-precision GNSS receiver (such as LRTK) attached to a smartphone allows centimeter-level positioning (half-inch accuracy) with just an iPhone while taking photos or scanning point clouds. These solutions provide accuracy comparable to dedicated equipment with the ease of single-person operation, strongly supporting measurements based on the steps introduced in this article. Leveraging such solutions can greatly simplify 3D surveying that previously required specialists and extensive effort. Digital surveying centered on smartphones is expected to become mainstream in field measurements. Adopt a markerless automatic coordinate-assignment workflow and promote DX in your surveying operations.