One survey. Multiple structural answers.

OUTCOME:  A single 351.14 m² dataset captured the slab, the occupied space below and the lower deck - then carried the interpretation through to 3D and CAD-ready outputs.

‍

‍

‍

THE PROJECT

The challenge: resolve a layered structure from the top deck

The subject was a suspended slab forming the upper level of a two-storey car park. The investigation needed to do more than locate reinforcement. It had to connect plan-view patterns with cross-sectional depth, distinguish responses inside the concrete from objects in the air space below, and package the results for engineering and CAD workflows.

‍

‍

The required outputs were deliberately varied:

• A plan view of the top reinforcement and areas of denser negative reinforcement near column supports.
• A relative concrete-cover map and an exportable record of detected rebar points.
• A qualitative deterioration screening map for the asphalt-covered deck.
• A slab-thickness map supported by horizon interpretation in the B-scans.
• A compact 3D interpretation that could move into CAD.

WHY THE FULL DATASET MATTERED:  The same georeferenced acquisition could be revisited at different depths and in different views, rather than treating each deliverable as a separate survey.

‍

FIELD ACQUISITION

High-resolution array data with RTK-referenced positioning

The dataset was collected with the GS9000 pushcart fitted with the GX1 array, paired with an MA8000 receiver using RTK corrections. The project draft records a 1 cm spatial interval, equal to the system's maximum specification of up to 100 scans per metre.

‍

‍

The GS9000 and GX1 combination provided the technical basis for the survey:

• Stepped-frequency GPR across 500-3,000 MHz for high-resolution work in the first metre.
• Thirty-five VV and 15 HH channels recorded simultaneously across a nominal 0.85 m scan width.
• Free Path acquisition for a georeferenced C-scan at real positions and dimensions.
• MA8000 correction options through NTRIP RTK or SSR augmentation; achieved accuracy remains subject to field conditions.
• Direct transfer through Workspace into GPR Insights for processing and interpretation.

_{Technical basis: GS9000 User Manual v1.0, pp. 7-9, 31-34, 40-44, 51 and 70. The 1 cm interval and project configuration are taken from the source case-study draft}

‍

CROSS-SECTIONAL INTERPRETATION

The B-scan exposed the full vertical story

A representative HH-polarised radargram showed far more than the first layer of reinforcement. It resolved the concrete section, a second reinforcement layer, the slab underside, large targets within the air space and a deeper interface consistent with the lower deck.

‍

‍

The interpreted vertical sequence was:

• Top reinforcement, visible as a regular sequence of hyperbolic responses.
• A second reinforcement layer within the concrete section.
• The slab underside at approximately 4.25 ns, equating to roughly 240 mm after the project velocity interpretation.
• Strong targets in the approximately 2.25 m air space, subsequently confirmed as vehicles.
• A deeper interface consistent with the slab of the lower car park level.

‍

INTERPRETATION PRINCIPLE: Travel time becomes depth only after an appropriate velocity or dielectric calibration. The dimensions reported here are project-specific interpreted values, not generic system limits.

‍

REINFORCEMENT

From individual reflections to a structural plan

Plan-view slicing combined the multi-channel information into a readable reinforcement image. The top grid is visible across the survey, while local square patterns and denser responses align with additional negative reinforcement around column locations.

‍

‍

‍

_{GPR Insights defines a slice as a horizontal plan view extracted from a 3D representation of the survey data; linked B-scan and slice views allow locations to be checked in both plan and section.}

‍

AI-ASSISTED ANALYTICS

Turning the first rebar layer into a relative cover map

GPR Insights' Analytics workflow uses AI-assisted hyperbola detection to identify the first rebar layer. For GX1 data, the software uses the HH input for these detections. The points can then be reviewed, edited and filtered by confidence before the map is accepted.

‍

‍

‍

IMPORTANT LIMITATION: The manual describes the cover map as an approximation based on a constant velocity or dielectric. It is best used to compare relative cover across the deck and calibrated against another cover method where required.

‍

QUALITATIVE CONDITION SCREENING

Cover and deterioration did not tell the same story

The same AI-detected first-rebar layer was used to generate a qualitative deterioration map. In the ASTM D6087 workflow for asphalt-covered concrete decks, lower reflection amplitudes indicate a higher likelihood of deterioration at or above the first reinforcement layer.

‍

‍

Approximately 5.75 m², or 1.6% of the scanned area, fell within the red category. The pattern did not show a strong spatial correlation with the lowest-cover areas. That is an important practical result: reduced cover and lower reflection amplitude are different indicators and should not be treated as interchangeable.

‍

USE WITH ENGINEERING JUDGEMENT:  This is a qualitative screening output. The D6087 approach is intended for asphalt-covered bridge decks, requires suitable scan orientation relative to the top reinforcement, and may be affected by deck construction and material conditions. Targeted verification should follow where decisions depend on the result.

_{Technical basis: GPR Insights User Manual v3.6, pp. 57-61.}

‍

DIGITAL TWIN

Interpretation objects gave the radar data physical form

For a representative area, the reinforcement was interpreted manually and a separate object was created for the dense negative reinforcement around a support. GPR Insights objects carry x, y and z positions and can be assigned cross-sections, dimensions, comments and visual textures.

‍

‍

‍

• Objects preserve spatial location rather than becoming detached screen annotations.
• The radar slice can be shown or hidden to move between evidence and interpreted geometry.
• Comments and object categories provide a place to carry interpretation context forward.

_{Technical basis: GPR Insights User Manual v3.6, pp. 55-57.}

‍

DELIVERABLES

The interpretation moved into CAD

The interpreted geometry was then opened in CAD. This step matters because the value of the survey is not limited to a software screenshot: interpretation items can be exported from GPR Insights as DXF, SHP, KML, CSV or XYZ depending on the active view and object type.

‍

‍

That export pathway supports a practical chain of custody for the interpretation:

• Acquire and georeference the multi-channel dataset in the field.
• Process, slice and interpret the data in GPR Insights.
• Review automatic detections and manually add project-specific geometry.
• Export the resulting objects and point data for engineering and CAD workflows.

COLLABORATION BENEFIT:  The exported interpretation can be combined with drawings and other survey information without forcing every stakeholder to work inside the GPR analysis environment.

_{Technical basis: GPR Insights User Manual v3.6, pp. 64-66.}

‍

SLAB THICKNESS

Horizon mapping exposed variation across the deck

The slab underside was followed with the horizon-mapping workflow. The algorithm traces a selected positive or negative amplitude peak through a depth window in each B-scan, after which the analyst can adjust the line and generate depth, thickness or amplitude maps.

‍

‍

‍

The interpreted range was approximately 245-285 mm. Only 0.18% of the scanned area fell in the red class, while the majority of the area was in the green class. The map therefore highlighted localised thin zones without losing the broader context of the deck.

‍

QUALITY CONTROL

The map remained traceable to the B-scan pick

A mapped result is more useful when the reader can trace it back to the underlying signal. Here, the horizon line is visible directly on the B-scan, allowing the analyst to check whether the automated trace follows the intended slab interface and to edit it where required.

‍

‍

• The search depth can be constrained to the relevant interface.
• Positive or negative peaks can be selected to match the target horizon.
• Smoothing, distance limits and manual line editing support review of difficult sections.
• The horizon map can be exported as a PNG with a CSV of x-y coordinates and depth values.

MARKETING VALUE WITHOUT LOSING TECHNICAL TRUST:  The colour map communicates the result quickly; the retained B-scan horizon shows how that result was derived.

_{Technical basis: GPR Insights User Manual v3.6, pp. 62-64.}

‍

BEYOND THE SLAB

The deepest slice explained the mystery objects

A deeper slice showed large targets beneath the upper slab. Their position within the interpreted air-filled zone, where relative permittivity approaches 1, separated them from responses inside the concrete. Site observation confirmed the targets as vehicles on the lower level.

‍

‍

THE RESULT: One survey produced a coherent structural narrative: reinforcement layout, relative cover, qualitative condition, slab-thickness variation, support-zone interpretation, deep-object context and CAD-ready geometry.

For asset owners and engineers, the practical advantage is continuity. The interpretation can move from a georeferenced plan, to the exact B-scan behind it, to a 3D object and finally into a wider CAD workflow. That makes the dataset useful both for immediate review and for future project coordination.

‍

‍

‍