Drone NDT Explained: Where Imaging Ends and Testing Begins

On This Page

• Match the method to the question
• Understand what changes when the drone makes contact
• Prepare the asset information and field workflow
• Specify a report that can be checked
• Keep missing data separate from clear findings
• Ask the provider to demonstrate one complete result

Match the method to the question

An inspection request should name the suspected condition: visible coating loss, a thermal anomaly, remaining wall thickness or a particular weld discontinuity. These are different questions even when they concern the same tank or pipe.

ASNT's visual testing guidance includes examination through optical instruments and cameras. Lighting, a usable view and the ability to interpret surface features all matter. A dark streak may be an observation worth recording, but its appearance alone cannot establish the wall thickness beneath it.

Use the following distinctions when reviewing a proposed service. The requested records and follow-up actions are procurement recommendations, based on the technical sources cited here, rather than a universal test procedure.

For temperature measurements, FLIR's manual identifies emissivity, reflected temperature, distance and atmospheric conditions as relevant inputs. Emissivity describes how strongly a surface emits radiation relative to an ideal reference. Reflections can contribute to the camera's reading. Treat an unexplained thermal pattern as a reason to investigate, with operating conditions and alternative causes considered before assigning a diagnosis.

For linear assets, our pipeline inspection workflows explain why visual, thermal and methane observations need different deliverables. A sensor label is too broad to serve as the inspection specification.

Understand what changes when the drone makes contact

In conventional pulse-echo ultrasonic testing, a probe sends sound into a material and receives returning echoes. Their timing helps locate reflecting boundaries; suitable echoes can support thickness measurement. An A-scan displays the returned signal against time, giving the inspector information beyond the displayed number. ASNT's ultrasonic testing overview explains these principles and the effects of geometry, material and discontinuity orientation.

Mounting the probe on a drone adds a positioning problem. The aircraft must bring the sensor to a usable location and maintain the interaction the measurement needs. Reaching the surface is only one part of obtaining a valid reading.

Flyability's Elios 3 UT documentation provides a commercial example: thickness points are associated with locations, and the system records A-scans that can be reviewed in its software. These are manufacturer-described capabilities. Their suitability for a buyer's asset still needs demonstration. A thickness-gauging demonstration does not establish the ability to find and size a specified weld flaw; that requires a technique designed for that examination.

The manufacturer's UT quick start guide specifies correct probe and test-block selection, calibration and material sound velocity. It also identifies surface condition, sufficient couplant, probe stability and contact time as factors. Couplant is the interface medium used to help ultrasound enter the material. Qualification to interpret the UT data remains necessary; learning the aircraft controls does not establish that competence.

Do not assume that all ultrasonic systems use the same interface. In original aerial NDT research by Hui and colleagues, an electromagnetic acoustic transducer, or EMAT, was integrated into a sliding payload tested on a Voliro T. The researchers investigated contact conditions and system dynamics for scanning a flat surface. That experiment shows why payload mechanics and control deserve scrutiny; it does not establish suitability for every curved, coated or operating asset.

Ask the supplier to demonstrate the proposed sensor on representative material and geometry. Define how an unstable or ambiguous signal will be rejected, and how the missed location will be handled. A successful flight demonstration without interpretable test data answers a different question.

Prepare the asset information and field workflow

Begin with the asset owner's inspection objective and the person responsible for evaluating the results. The drone team, NDT specialist and asset engineer may have different responsibilities. Put those responsibilities in writing before agreeing the field scope.

For procurement, we recommend supplying or resolving these inputs:

• Asset identifiers, drawings, material and relevant thickness or construction records.
• The suspected damage mechanism, previous findings and locations that must be revisited.
• Required examination areas or measurement points, plus a method for locating them again.
• Coatings, deposits, surface temperature, geometry and other conditions relevant to the selected technique.
• Access openings, obstructions, operating restrictions and the site's permitted work conditions.
• The procedure and evaluation criteria the responsible specialist will apply, including what happens when the result is inconclusive.

For refinery work, address operational coordination through a specific refinery inspection plan. A test instrument's suitability for a material does not by itself settle whether the proposed flight and access arrangement are acceptable at that site.

Organize the field work around an early demonstration on an agreed representative area. Check the data while the team can still revisit the asset. Log completed coverage, unusable observations and departures from the planned method. Then have the appropriate analyst interpret the results and return the agreed handover package.

This sequence is a recommended contracting approach, not a substitute for the applicable inspection procedure. Its purpose is to discover a method or access mismatch before the provider produces a large folder of unusable data. Record the handover checks and rework responsibilities in the drone inspection scope of work.

Specify a report that can be checked

A useful report should let a second reviewer move from a finding to its asset location and original data. Agree the export formats, identifiers and review access before capture. A screenshot in a PDF can help explain a finding, but it should not be the only record available when the interpretation depends on underlying signal data.

The following handover checklist is our recommended way to connect a commercial deliverable to the examination that produced it.

Set project-specific requirements for coverage and measurement quality with the responsible specialist. There is no single image count, point spacing or decimal precision that makes every drone NDT job complete. If a provider proposes a coverage percentage, ask what the denominator includes and how inaccessible regions are counted.

Keep missing data separate from clear findings

Consider a hypothetical tank-wall job. The buyer requests a visual survey and ultrasonic readings at identified points. The provider delivers usable photographs and valid readings at some points, while several locations cannot produce an interpretable signal.

The correct handover distinguishes three things: the surface that was visible, the locations with valid measurements and the locations still unresolved. It does not substitute a nearby reading for a missed point without an explicit technical justification. It also does not label an inaccessible region clear because no defect was recorded there.

Likewise, the lowest accepted reading in the delivered set is only the lowest measured value in that set. Inferring the minimum over an entire wall requires a sampling or scanning strategy justified for the inspection objective. A digital model with measurement pins makes the results easier to locate, but its visual completeness cannot fill gaps in the examination.

Agree who decides whether another access method, further surface preparation or a different test is needed. Keep this follow-up distinct from ordinary image recapture: recovering a blurred photograph and obtaining a valid measurement on a difficult surface can involve different work and personnel.

Ask the provider to demonstrate one complete result

Before buying a full campaign, request one representative result with its location, original data, method record, interpretation and proposed action. Ask who can judge signal validity, what conditions make the method unsuitable, how missed coverage is reported and who owns follow-up work.

Choose the approach that can answer the asset question with reviewable records. Visual screening may be sufficient to direct a closer examination. A thickness decision needs suitable measurements and coverage. A service that openly identifies unresolved areas gives the buyer a usable basis for the next inspection step.

Source notes

• ASNT: Visual Testing. Institutional explanation of surface examination, remote viewing and method limitations.
• ASNT: Ultrasonic Testing. Institutional explanation of echoes, thickness measurement, A-scans and technique limits.
• FLIR: Measuring temperatures. Manufacturer manual explaining temperature-measurement inputs; not a drone performance study.
• Flyability: UT Payload Quick Start Guide. Manufacturer instructions covering setup, interpretation competence and measurement conditions.
• Flyability: Elios 3 UT Payload. Manufacturer documentation for located thickness measurements and A-scan review.
• Hui et al.: Versatile Airborne Ultrasonic NDT Technologies via Active Omni-Sliding with Over-Actuated Aerial Vehicles. Original research abstract describing an experimental sliding EMAT payload; submitted November 8, 2023.

Last checked: September 10, 2026.