Combining LiDAR and GPR in One Coordinate System

# Combining LiDAR and GPR: When Geometry and Subsurface Data Need to Sit in One Coordinate System

Structural investigation increasingly demands that surface geometry and subsurface condition data occupy the same spatial reference frame. When LiDAR point clouds and GPR scan lines are processed in isolation, the result is two datasets that describe the same asset from different perspectives but cannot be directly interrogated together. Federated scan data, where both datasets share a common coordinate system, allows engineers to query rebar position relative to as-built geometry, identify cover depth anomalies in three-dimensional context, and produce coordinated deliverables that support design, remediation, and asset management decisions.

The technical challenge is not simply one of data volume or software compatibility. It is a registration problem: GPR produces time-domain reflection data that must be converted to depth using a dielectric constant, then georeferenced to a physical coordinate system. LiDAR produces dense point clouds referenced to scanner positions. Merging these requires deliberate survey control, consistent datum management, and a clear understanding of the error budget each method introduces. When done correctly, the combined dataset is substantially more useful than either survey in isolation.

This approach is increasingly relevant on complex infrastructure projects, heritage building investigations, and high-rise concrete assessments where design teams require coordinated as-built information. The following sections address the technical basis for integration, the workflow requirements, the limitations that govern confidence in the output, and the conditions under which additional testing is warranted.

Why Geometry and Subsurface Data Must Share a Coordinate System

A GPR scan line processed without spatial context produces a radargram: a two-dimensional cross-section showing reflection events at depth. The engineer can identify a rebar reflection and estimate its depth, but without georeferencing, that position cannot be reliably located on a drawing or related to adjacent structural elements. Similarly, a LiDAR point cloud captures surface geometry with millimetre-level precision but carries no information about what lies beneath the surface.

When both datasets are registered to a common coordinate system, each GPR reflection event can be assigned X, Y, and Z coordinates consistent with the point cloud. A rebar detected at 45 mm depth on a soffit can be placed precisely within the three-dimensional model of that soffit, including its curvature, deflection, and relationship to adjacent elements. This is the core value of LiDAR and GPR integration: the ability to answer spatial questions that neither dataset can answer alone.

For structural engineers assessing concrete condition, this means cover depth measurements are no longer abstracted from geometry. For architects working on adaptive reuse, it means services and reinforcement can be located in the same model used for design. For project leads managing complex investigations, it means a single coordinated deliverable rather than multiple disconnected reports.

Survey Control and Registration Workflow

The foundation of any integrated LiDAR-GPR survey is a shared control network. Targets used for LiDAR registration must also be used to georeference GPR scan lines. In practice, this means establishing a site control network using total station or GNSS, placing retro-reflective targets at known coordinates, and ensuring GPR scan paths are recorded against the same datum.

GPR scan path recording is typically achieved using one of three methods:

• Odometer wheel encoding: with a known start point referenced to the control network
• GNSS-tagged GPR: where satellite positioning is logged alongside the radar trace, suitable for open external areas
• Robotic total station tracking: of the GPR antenna, providing the highest positional accuracy in enclosed or GPS-denied environments

For internal concrete scanning on structures such as multi-storey car parks, podium slabs, or basement walls, robotic total station tracking is generally the most reliable method. Positional accuracy of the GPR scan path directly governs the spatial accuracy of any detected feature. An uncontrolled scan path introduces lateral positional error that can exceed 50 mm, which is unacceptable when locating reinforcement prior to coring or anchor installation.

LiDAR registration follows standard survey practice: multiple scanner setups are registered using common targets, producing a unified point cloud with a defined coordinate origin. The same target coordinates used in LiDAR registration are used to georeference the GPR scan network, ensuring both datasets share the same spatial reference frame.

Dielectric Calibration and Depth Accuracy

GPR depth estimates depend on the velocity of the electromagnetic wave through the material being scanned. Velocity is governed by the dielectric constant of the material, which varies with concrete mix design, moisture content, and aggregate type. An assumed dielectric constant of 6.0 to 9.0 is commonly applied to concrete, but this range produces meaningful depth variation. At a two-way travel time corresponding to 50 mm depth, the difference between a dielectric of 6 and 9 is approximately 8 mm, which is significant when assessing minimum cover to AS 3600 requirements.

Calibration is performed by scanning over a known feature, typically a rebar at a confirmed depth established by direct measurement at a breakout point, or by using a calibration block. Where calibration is not possible, depth estimates should be reported with an explicit uncertainty range. In federated datasets, depth uncertainty propagates into the Z-coordinate of every detected feature, and this must be documented in the investigation report.

Ground penetrating radar operating at higher frequencies, typically 1.6 GHz to 2.6 GHz, provides better depth resolution in concrete but reduced penetration depth. For slabs up to 300 mm thick, high-frequency antennas are appropriate. For deeper elements such as transfer slabs, pile caps, or thick walls, lower frequency antennas in the 400 MHz to 900 MHz range are required, with corresponding reduction in resolution. The choice of antenna frequency affects both the detectability of features and the accuracy of depth estimates in the integrated dataset.

Point Cloud Concrete Scanning: Data Federation in Practice

Data federation refers to the process of combining datasets from different sources into a single queryable environment without necessarily merging them into a single file format. In structural investigation, this typically means linking a registered LiDAR point cloud with georeferenced GPR interpretation data within a common coordinate system, accessible through GIS platforms, BIM environments, or specialist inspection software.

A reinforced concrete car park investigation in Sydney's inner west demonstrated the value of this approach. The structure, a 1970s-era post-tensioned flat plate system, required investigation prior to a change of use application. LiDAR scanning captured the as-built geometry of all soffit surfaces, columns, and walls across four levels. GPR scanning was conducted on a 200 mm grid across all soffit panels, with scan paths recorded using robotic total station. The federated dataset revealed that post-tensioning tendon profiles deviated from the original drawings by up to 120 mm in plan on several panels, and that cover to mild steel reinforcement in edge beams was below the 20 mm minimum in multiple locations. Neither finding would have been spatially locatable without the coordinated dataset. The outcome was a targeted remediation scope that avoided unnecessary concrete removal and supported the structural engineer's assessment with spatially accurate evidence.

Limitations and Error Budget Considerations

Federated LiDAR-GPR datasets carry cumulative error from multiple sources. Engineers and project leads should understand the principal contributors:

• LiDAR point cloud accuracy: is typically ±2 mm to ±5 mm for close-range terrestrial scanning, dependent on target geometry and registration quality
• GPR scan path positional accuracy: ranges from ±5 mm with robotic total station tracking to ±50 mm or greater with odometer-only methods
• Depth estimation uncertainty: from dielectric calibration is typically ±5 mm to ±15 mm depending on calibration method and material variability
• GPR lateral resolution: is governed by antenna frequency and depth; at 50 mm depth with a 2 GHz antenna, lateral resolution is approximately 20 mm to 30 mm

The combined positional uncertainty of a detected feature in three-dimensional space is the vector sum of these contributions. For most structural investigation purposes, a combined uncertainty of ±15 mm to ±25 mm is achievable with controlled survey methods. This is sufficient for locating reinforcement prior to coring, assessing tendon profiles, and producing coordinated as-built drawings. It is not sufficient for applications requiring sub-millimetre accuracy.

GPR cannot detect features below its penetration depth, cannot reliably resolve closely spaced elements where reflections overlap, and cannot characterise material properties beyond dielectric contrast. Where the investigation requires confirmation of rebar diameter, concrete strength, or carbonation depth, supplementary testing is required. Half-cell potential mapping per ASTM C876 is appropriate for corrosion activity assessment. Compressive strength assessment should follow AS 1012.14 for core testing or be supplemented by Schmidt Hammer rebound index testing calibrated to core results. Carbonation depth requires phenolphthalein indicator testing on freshly broken or drilled surfaces.

When Engineering Review and Further Testing Are Required

A coordinated LiDAR-GPR dataset is an investigation tool, not a structural assessment. The data requires interpretation by a structural engineer with knowledge of the relevant design standards, construction era, and structural system. Conditions that warrant escalation to detailed engineering review include:

• Cover deficiencies: below the minimums specified in AS 3600 Table 4.10.3 for the exposure classification
• Tendon profile deviations: that alter the effective eccentricity and therefore the flexural capacity of post-tensioned elements
• Anomalous reflections: that may indicate voids, delamination, or honeycombing within the concrete section
• Geometric deformations: in the point cloud that suggest deflection, settlement, or distortion beyond serviceability limits
• Inconsistencies: between the detected reinforcement layout and the original structural drawings

In each of these cases, the federated dataset provides the spatial context for targeted investigation, not the conclusion. Coring, endoscopy, load testing, or detailed finite element analysis may be required depending on the findings.

Coordinated As-Built Survey Deliverables

The output of an integrated LiDAR-GPR investigation should be structured to serve multiple downstream uses. Standard deliverables include a registered point cloud in LAS or E57 format, georeferenced GPR interpretation data in a format compatible with the project's BIM or GIS environment, plan and section drawings showing detected features at nominated levels, and a technical report documenting survey control, calibration, uncertainty, and interpretation methodology.

Where the project involves BIM, detected reinforcement and services can be modelled as objects within the point cloud environment, allowing clash detection and design coordination. This is particularly relevant for retrofit and adaptive reuse projects where existing structure must be integrated with new design elements.

Conclusion

Integrating LiDAR and GPR within a shared coordinate system is a technically demanding but well-established investigation methodology. The value lies in the ability to answer spatial questions that neither technology can address independently: where exactly is the reinforcement relative to the as-built geometry, and what does that mean for the structural assessment or design decision at hand? Achieving reliable results requires disciplined survey control, calibrated depth estimation, and a clear understanding of the error budget. When these conditions are met, federated scan data provides a level of spatial intelligence that supports faster, more accurate engineering decisions across the full project lifecycle.