6 Checkpoints to Verify Automatic Coordinate Assignment to Prevent LiDAR Positional Drift

Accurate alignment of the acquired data coordinates with real-world space is critically important in point cloud surveying using LiDAR. In recent years, methods combining positioning devices with "automatic coordinate assignment" that can assign world coordinates to point clouds in real time have been becoming widespread. This allows omitting the traditional work of matching to control points (registration), offering the benefit of conducting three-dimensional surveying more efficiently. However, if automatic coordinate assignment does not function properly, there is a risk of a "positional shift" in which the entire point cloud is displaced from its true location. If the point cloud is misaligned with actual features, overlaying it with design drawings or other survey data will fail and measurements of dimensions and as-built (volume) management will incur errors. For example, when comparing pre- and post-construction terrain, positional shifts in the point cloud can lead to mistakes in calculating earthfill and excavation volumes. To prevent such LiDAR coordinate positional shifts in advance, it is essential to check several important points before and after surveying. In this article, we introduce six checkpoints to keep in mind when acquiring point clouds using automatic coordinate assignment. By checking each point beforehand and taking countermeasures, you can georeference LiDAR survey results with high accuracy.

Table of Contents

• Checkpoint 1: Unification of coordinate systems and vertical datum

• Checkpoint 2: Calibration of measuring instruments and mounting accuracy

• Checkpoint 3: Time synchronization of LiDAR and GNSS data

• Checkpoint 4: Utilization of high-precision positioning and accuracy management

• Checkpoint 5: Preparation of the measurement environment and measurement planning

• Checkpoint 6: Accuracy verification using check points

• Summary

Confirmation Point 1: Unification of Coordinate Systems and Vertical Datum

To properly align point clouds obtained by automatic coordinate assignment with other data or design drawings, it is necessary to standardize the coordinate system being used. Raw point clouds from LiDAR sensors are often in a local coordinate system that takes the device itself as the origin, whereas civil engineering design drawings and other survey data may be managed in a public coordinate system (for example, geodetic latitude/longitude in a global geodetic system or plane rectangular coordinates) or a project-specific arbitrary coordinate system. If these coordinate systems do not match, large positional discrepancies on the order of tens of meters (tens of ft) can occur when comparing the point cloud and the drawings at the same location. For example, if the point cloud remains in the instrument coordinate system (an arbitrary origin) while the drawings are expressed in the national reference coordinates, it is obvious they will not align. Also, because Japan’s plane rectangular coordinate system has different origins for each region, choosing the wrong zone number can cause the entire dataset to be displayed with a large shift.

Additionally, attention is required regarding the vertical reference for elevation. The height automatically assigned by GNSS positioning is usually the reference called "ellipsoidal height" (height based on the Earth's ellipsoid model). However, the elevations used in civil engineering drawings are generally "elevation (orthometric height)" referenced to mean sea level. Ellipsoidal height and orthometric height vary by region, but near Japan there is approximately a 30-40 m (98.4-131.2 ft) difference, and failure to correct for this can cause large discrepancies in the vertical direction. For example, if you compare point cloud data heights as-is to the elevation values on design drawings, you may see an offset such as being tens of meters (tens of feet) higher (or lower) than the actual.

As a precaution, before measurement always confirm the coordinate system and vertical datum to be used for the project, and configure the point cloud acquisition settings accordingly. When using GNSS, select the appropriate geodetic datum and coordinate system in the receiver or measurement app’s coordinate output settings (e.g., geographic coordinates of JGD2011 or the corresponding plane rectangular coordinate system). Apply a geoid model as needed so that correction from ellipsoidal height to orthometric height can be performed automatically. Also check for unit system differences (e.g., meters (m (ft)) vs. feet (ft), or meters (m) vs. millimeters (mm (in))). If the acquired point cloud is recorded in a different reference system, perform a coordinate transformation (localization) using known control points during post-processing to align it to the unified coordinate system. By unifying the coordinate system and datum in advance, point cloud data with automatically assigned coordinates can be used without discrepancies with other materials.

Checkpoint 2: Calibration of measuring instruments and mounting accuracy

When combining a LiDAR sensor and a GNSS positioning device to automatically assign coordinates, it is essential to verify the calibration and mounting accuracy of each instrument. If the positional relationship between the LiDAR and the GNSS antenna (an offset called the lever arm) or the misalignment between the LiDAR’s measurement axis and the IMU (inertial measurement unit) attitude axis (boresight error) are not properly corrected, a consistent positional shift will occur across the entire acquired point cloud. For example, if the GNSS antenna is mounted directly above the LiDAR sensor by 1 m (3.3 ft) but that height offset is not corrected when assigning coordinates, the point cloud data will be recorded consistently 1 m (3.3 ft) lower than their actual positions. Also, if there is a small angular misalignment between the LiDAR’s scan direction and the vehicle’s direction of travel, the farther the distance, the more the point cloud positions will be shifted obliquely.

To avoid such problems, make sure to properly calibrate the equipment you will use in advance. If the manufacturer provides calibration parameters (such as offset amounts and inter-sensor angular correction values), enter them correctly into the measurement system or application. For drone-mounted or vehicle-mounted LiDAR, calibration is carried out at shipment or during initial assembly by measuring multiple known points to correct misalignments between devices, but regular checks before operation are also recommended. Especially when used in harsh field environments, be vigilant for changes in the relative positions of sensors caused by vibration or impacts. When measuring with a small GNSS receiver attached to a smartphone, also check that the attachment has not loosened or become tilted, and ensure that the LiDAR and GNSS always maintain a stable relative position and orientation.

Additionally, calibrating the device’s built-in IMU is important. If procedures such as leaving it stationary on a level surface before starting measurements or moving it in a figure-eight pattern are recommended, be sure to follow them to initialize the sensors. This ensures correct attitude estimation in combination with GNSS and magnetic sensors, improving the accuracy of coordinate assignment. In surveys that combine multiple instruments, tiny misalignments invisible to the human eye can lead to overall accuracy degradation, so not neglecting calibration and checks of the mounting is fundamental to high-precision positioning.

Verification Point 3: Time synchronization between LiDAR and GNSS data

Time synchronization between the point cloud data acquired by LiDAR and the position and orientation data from GNSS and IMU is critically important in mobile mapping (LiDAR measurement while moving). If their timestamps are misaligned, the correspondence between the moving platform’s position information and the actual laser measurements will be thrown off, causing distortions and positional offsets in the point cloud. For example, if the timestamps between sensors differ by 0.1 seconds while a vehicle is traveling at 60 km/h, the point cloud acquired during driving would be recorded approximately 1.7 m (5.6 ft) away from the actual position. Even at a slow walking speed, insufficient synchronization can produce a blurred offset across the entire point cloud, making the positional relationships of fine details inaccurate.

To prevent such issues, ensure the LiDAR sensor and the positioning device record data to a common time reference. In dedicated surveying equipment, by sharing GPS clock pulses or timestamps, the LiDAR laser firing timing and GNSS positioning data are synchronized to the nanosecond level. Even when measuring with a smartphone + GNSS device, the app is designed to attach the same time information to both sets of data so that positions are assigned at matching times. As a user, confirm that each device’s clock is correct before starting measurements, and if possible use an external trigger or synchronization cable to achieve hardware-level synchronization for greater assurance. If you later merge LiDAR and GNSS logs that were recorded separately, merge the data using elapsed time from the same start time and, if necessary, align the time series using known markers (for example, firing the laser at a recognizable point when beginning the measurement).

If time synchronization is properly ensured, coordinates can be assigned with accuracy comparable to stationary conditions even when acquiring point clouds while moving. Conversely, if synchronization is lax, point clouds can curve and distort during high-speed travel, and object contours can blur into double images during continuous capture, so always check the system's synchronization settings.

Checkpoint 4: Utilization of High-Precision Positioning and Accuracy Management

When automatically assigning coordinates, the most fundamental thing is ensuring positioning accuracy by GNSS. With typical standalone positioning (standalone GPS), planar errors of several meters (several ft) occur, so this alone cannot precisely align the positions of point clouds. To obtain high-precision positions, positioning using error corrections such as RTK is indispensable. By installing a base station to transmit correction information by radio, or by using a network RTK service (distribution of correction information via the Internet), GNSS position accuracy can be improved to about a few centimeters (a few in). In recent years, satellite augmentation signals that enable high-precision real-time positioning even when used alone, such as Japan’s quasi-zenith satellite Michibiki’s centimeter-class augmentation service (CLAS), which provides cm level accuracy (half-inch accuracy), have also been made available.

Precision management is also important when utilizing high-precision positioning. In RTK surveying, always confirm that GNSS can maintain a "fixed solution (Fix)". A fixed solution is a state where the integer ambiguity of the phase differences has been successfully resolved, and only in this state is positioning accurate to a few centimeters (a few cm (about 1 in)) guaranteed. If signals are interrupted during positioning due to building shadows or trees, the fixed solution may be temporarily lost, and the point cloud captured during that time will have locally reduced positional accuracy. If a fixed solution cannot be obtained for an extended period, you should temporarily suspend measurements and wait for GNSS reinitialization, or improve the signal conditions by adjusting the base station location or antenna height. Also, when installing your own base station, take care not to make errors in its known coordinate values. If the reference point coordinates are entered incorrectly, the point clouds may match each other relatively, but from the public coordinate system's perspective the entire dataset will be recorded at an offset position. With network RTK, base station coordinates are automatically ensured, but it is advisable to reconfirm that the coordinate system and geoid model settings being used match those in the field.

Furthermore, to stabilize positioning accuracy, pay attention to the satellite configuration (geometry). Before measurement, check the GNSS receiver’s PDOP value and the number of visible satellites, and if possible choose a time when the satellites are widely distributed for surveying. Modern GNSS receivers can simultaneously use signals from multiple satellite positioning systems (not only GPS but also GLONASS, Galileo, Michibiki, etc.), so the influence of biased satellite distribution is smaller than before; however, satellite geometry can still deteriorate easily in mountainous areas and urban canyons, so caution is necessary. To fully reap the benefits of high-precision positioning, it is essential to maintain conditions that consistently provide high-accuracy position information. During measurement, periodically check the GNSS status and be mindful that the expected accuracy is being maintained.

Checkpoint 5: Preparing the measurement environment and the measurement plan

The risk of coordinate shifts is greatly affected by on-site environmental conditions and measurement methods. First, when using GNSS, it is desirable that the surrounding sky be well open. In environments where the view of the sky is limited—such as in streets of high-rise buildings or inside forests—satellite signals can be blocked or reflected by buildings or trees (multipath), making positioning accuracy unstable. As a result, the entire point cloud may be offset in a particular direction, or accuracy may vary by location. If the site is confined and GNSS is difficult to use, consider switching to a ground-based static laser scanner, or, as traditionally done, installing known control points on site and aligning the point cloud during post-processing.

Attention must also be paid to the measurement environment of the laser scanner itself. Measuring objects with extremely high (or low) reflectivity, such as mirrors, metal plates, or water surfaces, can cause the sensor to record noisy points or ghosts (false points that do not actually exist), which can interfere with alignment. Before starting measurements, survey the site and remove, as far as possible, anything that is likely to cause obvious noise. Be careful in locations with strong winds or significant vibration. When mounting LiDAR on a tripod, stabilize the footing and minimize sway caused by wind. For drone-mounted LiDAR, do not force flights on high-wind days; choose days with better conditions or use an aircraft with high stability control performance to obtain point clouds with less motion blur.

If you plan to acquire point clouds in multiple sessions, it's also important to ensure sufficient overlap between each dataset. When dividing a large area that cannot be captured at once into sections for measurement, if adjacent point clouds share few common areas it becomes difficult to verify and correct their alignment later. Set survey lines and scan positions so that there is at least 20–30% overlap, and make sure common features appear in the scans. If necessary, place several artificial targets (such as reference boards or other markers) to create reference points that are captured in every point cloud for added assurance. With this kind of preparation, even if some coordinates are offset, it will be easier to correct them later through relative adjustments.

Overall, pre-measurement site preparation and appropriate planning are indispensable for preventing shifts in LiDAR coordinates. By carefully observing the surrounding conditions and conducting measurements after eliminating risk factors as much as possible, the reliability of the resulting point cloud data is markedly improved.

Please translate the following input into English.

Checkpoint 6: Accuracy verification using validation points

A final verification to ensure there are no offsets in the point cloud data obtained by automatic coordinate assignment is also indispensable. During measurement, prepare several independent verification points (checkpoints) whose positions are known, and you can confirm accuracy by matching them with the point cloud after acquisition. For example, set clear feature points (corners, markers, etc.) on structures within the site and measure their coordinate values separately with high precision. After processing the point cloud, read the coordinate values on the point cloud that correspond to those feature points and compare them with the values measured in advance. If the amount of deviation falls within a predetermined tolerance, the coordinate assignment can be considered generally acceptable. Conversely, if a clear difference of several centimeters (several in) or more is confirmed, some error factors may be present.

One possible way to deal with detected misalignments is that, if the entire point cloud has a consistent offset, you can correct it by translating it by that amount. Applying corrections such as translating it ○ cm (○ in) toward the northeast in the horizontal plane and raising it ○ cm (○ in) vertically should eliminate the discrepancy with the reference point. However, if the offset amount varies by location or if rotation or scale errors are suspected, a simple translation will not suffice. You will need to review the data-processing procedures and verify whether there were coordinate-system setting mistakes or malfunctions in some instrument data. Based on that, consider remeasuring if necessary or readjusting the whole dataset by using multiple reference points to perform a Helmert transformation (scale and rotation correction).

In addition to comparing with control points, it is also useful to overlay the resulting point cloud data onto existing drawings or other survey results to check for any inconsistencies. By overlaying it with the design model or previously measured terrain data, you can conclude that an error occurred if there is an obvious positional shift. As part of a pre-delivery checklist, be sure to perform positional comparisons at several locations to confirm that the LiDAR survey results can be trusted. By making such verification checks routine, you can detect any data shifts early and reduce the risk of using results that contain serious errors.

Summary

Above, we explained six checkpoints to prevent coordinate misalignment in LiDAR surveying. Unifying coordinate systems, equipment calibration, time synchronization, using high-precision GNSS, considering the environment and planning, and verifying the resulting data are all fundamental matters, but I hope it is clear that neglecting even one of them can lead to significant errors in the final deliverables. In particular, on construction and civil engineering sites, even a few centimeters of misalignment directly impacts quality and safety, so be sure to manage the accuracy of point-cloud data based on the points raised here. Even if alignment problems occur, following this checklist to identify causes and take corrective action should bring you closer to a highly accurate deliverable.

That said, depending on site conditions you may face challenges such as “insufficient satellite visibility in confined spaces,” “becoming too far from the reference station during large-area surveys,” or “lack of confidence in coordinate transformations or software settings.” This is where the iPhone-mountable GNSS high-precision positioning device “LRTK” comes in. LRTK consists of a compact, high-performance RTK-GNSS receiver and a smartphone app, and is a next-generation solution that enables anyone on site to easily achieve centimeter-level positioning (half-inch accuracy). Even in environments where satellite signals tend to become unstable with conventional methods, it enables stable high-precision positioning through simultaneous use of multiple satellite systems and a built-in tilt compensation function. It also directly supports the centimeter-class augmentation service (CLAS) provided by Japan’s Quasi-Zenith Satellite System, Michibiki (QZSS), enabling standalone centimeter-level positioning (half-inch accuracy) from satellite correction information even at sites where setting up a base station is not possible. Furthermore, it supports i-Construction, promoted by the Ministry of Land, Infrastructure, Transport and Tourism, and is a tool that dramatically improves the productivity and accuracy of surveying work. Because you can monitor the deviation from the reference point on the screen in real time, you can detect errors on the spot and respond quickly. If you currently find challenges or complexity in aligning point cloud data in your workflow, please consider using LRTK. It makes accurate alignment possible even without advanced surveying knowledge, and will support fieldwork as a reliable partner even in complex environments.