How Accurate Is the SHARE SLAM S20? A 40-Measurement Building Accuracy Study

Study Finding

In a field study of an unfinished three-story building in Dongguan, Guangdong, researchers completed one scan with the SHARE SLAM S20 and manually measured wall-target dimensions in SHARE PointClouds Studio V2.5.1. Across 40 paired comparisons between point-cloud dimensions and the mean of three laser distance meter readings, the mean absolute error (MAE) was 3.42 mm, the root mean square error (RMSE) was 4.17 mm, and the maximum absolute difference was 10 mm. Of the 40 pairs, 33 had absolute differences no greater than 5 mm.

These results demonstrate good dimensional measurement performance for the SHARE SLAM S20 in the tested unfinished three-story building and workflow. The scan also produced a navigable 3D point cloud that could be reviewed and measured again after fieldwork. This gives the method practical value for renovation remeasurement, existing-building documentation, and as-built dimension verification. On a specific project, control points and independent check points can be defined in the quality plan, with the deliverable confirmed through an independent check report.

The digital object identifier (DOI) for the research record is 10.5281/zenodo.21821068.

Test Subject and Method

The study site was an unfinished three-story building containing a void space, bedrooms, corridors, a stairwell, and a terrace. The method was:

1. Mount quadrant targets on walls to provide common endpoints for both measurement methods.
2. Aim a Deli DL331040D laser distance meter independently three times at each target pair and calculate the mean reading.
3. Complete one continuous scan of the building with the SHARE SLAM S20.
4. Identify the same target positions in SHARE PointClouds Studio V2.5.1 and manually pick each dimension three times.
5. Calculate differences and summary statistics for 40 matched dimension pairs. The sample spans approximately 1,008–11,640 mm.

Figure 1: The paper’s experimental-method composite includes the SHARE SLAM S20, repeated aiming with the laser distance meter, and wall-mounted quadrant targets.

Figure 2: Building point-cloud review and manual distance extraction in SHARE PointClouds Studio V2.5.1.

Overall Results and Floor-Level Data

The essential floor-level results are:

Figure 3: Absolute differences for the 40 sample pairs shown by floor as boxes, whiskers, and individual points.

The MAE for every floor was in the 3 mm range, while every floor-level RMSE remained below 5 mm; the overall maximum absolute difference was 10 mm. Because the comparison values were means from repeated laser distance meter readings, the statistics directly describe the agreement between the two measurement workflows for the same set of wall dimensions.

Value for Renovation and Building Documentation

Renovation Remeasurement

A single scan can record the spatial relationships among rooms, corridors, stairwells, and voids. After leaving the site, teams can revisit the point cloud and obtain additional dimensions from captured areas, supporting plan development, material ordering, and pre-construction dimensional review. Critical installation dimensions can be included in a project inspection checklist and confirmed by independent check points or field spot checks.

Existing-Building Documentation

A point cloud retains more than individual distance values: it provides a visual 3D record that can be revisited to examine wall relationships, clear distances, openings, and spatial connections. For existing buildings with incomplete drawings, this supports the creation of traceable existing-condition records and downstream computer-aided design (CAD) preparation.

As-Built Dimension Verification

Point-cloud measurements can be combined with project datums, control points, and independent check points to inspect room clear dimensions, corridor widths, and other agreed items systematically. The project method statement should define capture routes, point-picking rules, inspection locations, and tolerances. Check points excluded from the coordinate-transformation solution can then be used to generate an independent check report for deliverable confirmation.

Quality Control and Delivery in SHARE PointClouds Studio V2.6

The original study used SHARE PointClouds Studio V2.5.1. SHARE PointClouds Studio V2.6 can support processing, checking, and delivery in subsequent projects, within the following capability boundaries:

• Manual registration and error preview: Operators manually select corresponding points, preview the registration error, and save the result separately; the software does not automatically select those corresponding points.
• Coordinate transformation and independent checking: Control points are used to solve the coordinate transformation, while check points excluded from the solution are used to generate a 3D or elevation accuracy check report.
• External post-processed kinematic (PPK) trajectory input: An RTKLIB POS trajectory file generated by an external PPK tool can be selected as the Trajectory Source in global navigation satellite system (GNSS) fusion through GNSS Fusion; the software does not perform the complete raw-observation PPK solution.
• Batch sections: In Section Extraction, set the direction and the width required by the project; then choose one of two options: in Batch Cutting, set Number of Sections and Batch Interval, or in Custom Interval, set custom positions; preview and then extract the sections.
• Trajectory and photo linkage: Trajectories, photo points, undistorted photos, and the 3D view can be reviewed together to check capture coverage.
• 3D Gaussian splatting (3DGS) viewing: This supports photorealistic viewing and orientation; it does not replace engineering measurement in the point cloud.
• CAD drawing export: After section- or plan-assisted manual drafting and checking, the export produces a ZIP package containing both drawing exchange format (DXF) and DWG files. RVT, IFC, and complete building information modeling (BIM) models remain downstream professional deliverables.
• Connectivity boundary: Without an available network, login may be unavailable, some advanced functions such as 3DGS may not work normally, and logs cannot be uploaded.

Project Workflow and Independent Confirmation

A project can translate the research method into a quality-controlled deliverable through the following workflow:

1. Define measurement items and inspection ratios for renovation remeasurement, existing-building documentation, or as-built dimension verification.
2. Plan the scanning route and target positions, and record the device, software version, site conditions, and processing parameters.
3. Use control points for coordinate transformation while reserving separate independent check points; do not use the same point for both the solution and the check.
4. Measure critical dimensions using a consistent point-picking rule, and revisit the point cloud or field-check any anomalous location.
5. Issue an independent check report recording 3D or elevation differences at check points, inspected dimensions, reviewer, and conclusion.
6. Archive the point cloud, check report, and—where required—the exported ZIP package containing DXF and DWG files for traceability by designers, contractors, and owners.

Research Reference

• Device: SHARE SLAM S20
• Study software: SHARE PointClouds Studio V2.5.1
• Setting: one unfinished three-story building
• Sample: 40 paired wall dimensions
• Results: MAE 3.42 mm, RMSE 4.17 mm, maximum absolute difference 10 mm
• DOI: 10.5281/zenodo.21821068
• Project use: confirm deliverables with project-defined independent check points and an independent check report