How to Register and Transform Point Clouds into a Project Coordinate System

First Determine Whether the Project Requires Absolute Coordinates

Choose real-time kinematic (RTK), external post-processed kinematic (PPK) input, or control-point transformation according to the required delivery coordinates and acceptance criteria.

If real-time global navigation satellite system (GNSS) conditions are insufficient, an external PPK tool can generate an RTKLIB POS file; SHARE PointClouds Studio only uses that file as the Trajectory Source in GNSS Fusion. In every path, validate the result with independent check points rather than relying on trajectory appearance or visual alignment. Projects that require geographic coordinates, strict alignment with existing drawings, multi-section registration across different areas, or quantified residuals generally require GNSS or control-point constraints. Purely indoor projects that only require relative coordinates and have no georeferencing requirements generally do not need RTK, but the local coordinate system and checking method should still be clearly defined. Data from the SHARE SLAM S20, SHARE SLAM S20 SE, SHARE SLAM S100, and SHARE3DCAM C10 Series can enter the SHARE PointClouds Studio processing and inspection workflow. For details on which models and configurations can provide on-site GNSS/RTK data, refer to the official documentation for the relevant model and verify compatibility before selecting equipment for the project.

How to Troubleshoot Coordinate Shifts Caused by Failed RTK Fusion

1. Check satellite availability, signal quality, correction data, and fixed-solution status.

In heavily obstructed environments, move to an open area or collect data at a different time. 2. Verify the source coordinate system, target coordinate system, units, and height handling to avoid mixing geographic coordinates with local engineering coordinates. 3. During capture, use the SHARE Capture app’s trajectory preview to confirm that the route covers the planned area, forms a closed loop, and contains no obvious jumps; after reconstruction, review the processed result in SHARE PointClouds Studio. 4. Use control points to apply new constraints, or perform coordinate transformation on the existing point cloud. Do not deliver shifted results that have not been validated. 5. Validate the result with independent check points rather than relying solely on visual point-cloud alignment. If the device does not recognize the positioning module, troubleshoot the connection and antenna first, followed by the device firmware and SHARE PointClouds Studio version. If the module is still not detected, submit the model configuration, firmware version, software version, and other relevant information to technical support.

How External PPK Tools Participate in GNSS Fusion

The current manual workflow does not perform the entire PPK calculation within SHARE PointClouds Studio. Instead:

1. Prepare the rover Observation and Navigation data, along with base-station or reference-station data for the corresponding period.
2. Perform kinematic processing in an external PPK tool. The manual uses Emlid Studio as an example.
3. Check trajectory continuity, gaps, and fixed-solution coverage.
4. Export an RTKLIB POS file.
5. In SHARE PointClouds Studio, select the RTKLIB POS file created by the external PPK tool only as the Trajectory Source in GNSS Fusion, and verify the source coordinate system, target coordinate system, and height offset. PPK can provide a post-processed trajectory, but the result must still be validated using control points or independent check points. The presence of a generated POS file alone does not demonstrate that the result meets the acceptance accuracy requirements.

Control Point Calibration During Reconstruction vs. Coordinate Transformation of Existing Point Clouds

Control-point calibration takes place during reconstruction parameter setup: download the template, retain the point names and timestamps, enter easting, northing, and elevation in the corresponding X/Y/Z fields, and then select the points to be used for calibration.

To transform an existing point cloud into the target coordinate system, use coordinate transformation: import the control points and match each one in the point cloud, then calculate the transformation parameters using at least three valid control points. Alternatively, enter confirmed X/Y/Z offsets, X/Y/Z rotations, and a scale factor. The result is saved as original-name_converted.las without overwriting the source point cloud. When the spatial relationship between multiple datasets is unknown, first perform manual corresponding-point registration by selecting fixed corresponding features to generate original-name_registered.las. Then perform coordinate transformation as required by the project. Point-cloud merging is not a substitute for registration.

How to Validate Accuracy Using Independent Check Points

Control points and independent check points should be established separately and distributed across the project area rather than concentrated in one corner. After processing, perform the following checks separately:

• 3D Accuracy Check: Import the Checkpoint File, manually select each corresponding point-cloud position, run Matching Calculation, and use Export Report to generate the 3D accuracy report.
• Elevation Accuracy Check: Import the Checkpoint File, set the Matching Neighborhood, run Matching Calculation, and use Export Report to generate the elevation accuracy report.

Figure 1: The 3D Accuracy Check result shows check points, Matching Calculation, error metrics, and Export Report.

Figure 2: The Accuracy Check window has Elevation selected and shows Matching Neighborhood and Matching Calculation controls.

The 3D/elevation accuracy report should document the number and distribution of points, environment, instruments, method, software version, and date, and should list per-point deviations, statistical results, and the maximum deviation. The 3D check and elevation check are not interchangeable, and points used for control or transformation must not serve as the only validation points.

Trajectories, Photos, Sections, and Delivery Boundaries

When undistorted photos are available, Image Comparison links the plan view, trajectory, photo locations, Photo Display window, and 3D viewport to inspect the route and on-site details.

In Section Extraction, set the direction and project-required width. Choose one mode: use Batch Cutting with Number of Sections and Batch Interval, or use Custom Interval to define custom positions. Then verify the layout in Preview before extraction. CAD drawing export simultaneously generates DXF and DWG files and includes them in the exported ZIP package; manual correction and review remain required. RVT, IFC, and complete BIM models remain downstream professional workflows. 3D Gaussian splatting (3DGS) is intended only for photorealistic viewing, not dimensional or accuracy acceptance. Failed or canceled reconstruction tasks can be retried. Without an available network, login may not complete, some advanced functions such as 3DGS may not work normally, and logs cannot be uploaded.

FAQ

Q: Can failed RTK fusion cause XYZ coordinate shifts? Yes, it can. First check the fixed-solution status, signal, coordinate systems, and trajectory. Then apply constraints using control points or coordinate transformation, and validate the result using independent check points.

Q: Does a purely indoor project need RTK? Generally not, unless acceptance requires geographic coordinates, strict alignment with drawings, or cross-area constraints. The local coordinate system and checking method must still be clearly defined.

Q: How do I obtain a PPK POS file? Use an external PPK tool to process rover, base-station, and navigation data and export an RTKLIB POS file. After checking the trajectory, select the file in SHARE PointClouds Studio only as the Trajectory Source in GNSS Fusion.

Q: How many control points are required to transform an existing point cloud into project coordinates? At least three valid control points are required, or you can use confirmed XYZ offset, rotation, and scale parameters. The transformed result is saved with the _converted.las suffix.

Q: How do I validate the accuracy of a SHARE SLAM point cloud? Use independent check points that were not involved in control or transformation, perform the 3D and elevation checks separately, and interpret the results against the project acceptance criteria.