Is GPR scanning accurate?

GPR scanning is a reliable non-destructive technique, but its accuracy depends on three factors: equipment, calibration, and surveyor.

Typical accuracy figures

For concrete scanning, accuracy is set by the instrument and the conditions. As an example, Hilti publishes these figures for the PS 1000 X-Scan:

• Plan position of detected targets: ±10 mm.
• Depth indication: ±10 mm below 100 mm depth, ±15% beyond, depending on how well the velocity calibration represents the host material.
• Detection: small, deep or closely spaced targets may be missed.

Hilti notes that site conditions can significantly affect these figures.

For utility detection:

• Plan position: stated on the drawing as the PAS 128 quality level. The best detection level, QL-B1, allows ±150 mm in plan and ±15% of depth.
• Detection rate: depends on ground conditions, service material, and depth. Metallic and pressurised services detect more reliably than non-metallic abandoned plant.

What “accurate” requires

GPR accuracy depends on:

• Calibrated equipment. Manufacturer calibration plus session calibration against a known target.
• Appropriate antenna frequency. Higher frequency for shallow, finer targets; lower frequency for deeper or larger features.
• Qualified interpretation. A trained surveyor reading the scan correctly, not just running the equipment.
• Honest acknowledgement of limits. The surveyor knows when the scan reaches its useful depth.

A calibrated scan with a qualified surveyor produces defensible accuracy. An uncalibrated scan with an inexperienced operator does not.

When it can be less accurate

Some conditions degrade accuracy:

• Heavily reinforced concrete (dense top mat shadows the lower layer).
• Saturated or wet concrete (changes the velocity).
• Conductive ground (clays, salt-contaminated soils, limits penetration).
• Closely spaced targets at depth (resolution limit).
• Edge effects near corners and openings.

A defensible deliverable states the conditions and the resulting accuracy. A deliverable claiming uniform high accuracy regardless of condition is over-claiming.

Verification

For applications where exact position matters more than the survey accuracy delivers, for example, the precise tie-in of a critical service, GPR detection is followed by intrusive verification (a trial pit or vacuum excavation). This is standard practice on safety-critical work and is the verification level “QL-A” in PAS 128 utility surveys.

How accuracy is reported

A defensible GPR deliverable includes:

• The accuracy class for the scan (in PAS 128 work, the QL-B sub-level).
• The depth-velocity calibration applied.
• Any conditions that affected accuracy on the day.
• A clear statement of where intrusive verification is recommended for absolute confirmation.

The bottom line

GPR scanning is accurate enough for what it does, when it is done properly. It is not a substitute for intrusive verification on safety-critical points, but it suits the routine pre-drill, pre-cut, and reinforcement-mapping work that makes up most concrete scanning. For more on reading the deliverable, see How to read a GPR scan report.