Drone inspection replaces ladders, lifts, rope teams and shutdowns with a stabilised camera at standoff — and the discipline is the same across every asset class: right sensor, repeatable capture, comparable evidence, actionable report.
- 3 modesVisual, thermal, combined EO/IR
- 4 stepsThe workflow that scales
- ProcessWhere evidence quality comes from
- DecisionsWhat asset owners pay for
Key takeaways
- Three inspection modes cover everything: visual zoom for surface condition, thermal for what the eye cannot see, and combined EO/IR for detection plus cause in one pass.
- Evidence quality is a process property, not a camera property — consistent altitude, angle and route matter more than sensor grade.
- Log the conditions. Load, weather and irradiance are what make one survey comparable with the next.
- Asset owners pay for decisions, not photographs. Price against the access equipment you replace.
On this page
The Three Inspection Modes
Visual zoom inspection documents surface condition from a safe distance. Cracks, corrosion, erosion, missing hardware, vegetation encroachment — anything an inspector would look at, resolved from standoff instead of from a ladder.
Thermal inspection reveals electrical, moisture and process anomalies invisible to the eye. A hot connection, wet insulation under a membrane, a blocked radiator section: these have no visual signature until they have already failed.
Combined EO/IR pairs them in one pass — detection on thermal, cause confirmation on visible. This is where a multi-sensor payload earns its price, because the two channels are boresight-aligned and the finding and its evidence come from the same moment.
Most mature programmes run the third mode by default and fall back to the first two only where mass or budget forces it.
Asset-by-Asset Entry Points
Each asset class has its own defect vocabulary, timing constraints and payload priorities. These are the deep dives.
| Asset | Primary mode | Key timing constraint | Guide |
|---|---|---|---|
| Power lines | Thermal + zoom | Survey at meaningful load | Power line payloads |
| Substations | Radiometric thermal | Load must be logged | Substation workflow |
| Solar farms | Radiometric thermal | Irradiance threshold, clear sky | Solar farm inspection |
| Wind turbines | Zoom + thermal | Morning thermal transition | Blade inspection |
| Industrial roofs | Thermal | 30–90 min after sunset | Roof inspection |
| Building envelopes | Thermal | 10 °C indoor–outdoor delta | Envelope surveys |
| Bridges | Zoom + thermal | Midday for deck delamination | Bridge inspection |
| Pipelines | Thermal + zoom | Corridor-length coverage | Pipeline monitoring |
The pattern worth noticing: almost every thermal entry has a timing constraint, and almost every one of those constraints is non-negotiable. Fly outside the window and the survey looks professional and diagnoses nothing.
The Workflow That Separates Programmes From Pilots
Consistent altitude, angle and route. Logged conditions — load, weather, irradiance. Radiometric capture where numbers matter. Baseline comparison flight over flight. That is the whole difference, and none of it is about the camera.
Evidence quality is a process property, not a camera property. Two operators flying identical hardware over the same asset produce deliverables worth very different amounts, and the gap is entirely in whether the second survey can be compared with the first.
- Plan a repeatable route with fixed altitude, standoff and viewing angle, and fly it the same way every time.
- Record the conditions that affect interpretation — load for electrical, irradiance for solar, delta-T for envelope, wind for everything.
- Capture radiometric imagery wherever the report will quote a temperature; the decision cannot be reversed later.
- Grade findings against comparable peers — the other two phases, the adjacent panel, the same bay last quarter.
- Compare against the previous flight. A finding that is new or growing outranks a finding that has been stable for two years.
The comparison step is what turns inspection into monitoring, and monitoring is what asset owners actually want. A single survey tells you the current state; a series tells you the trajectory, and maintenance planning runs on trajectories.
Choosing the Payload
Three tiers cover the range, and the discriminator is standoff rather than asset type.
| Programme type | Payload class | Example | Why |
|---|---|---|---|
| Close-range single-asset work | Compact dual-sensor | MV-2P — 130 g, 640×512, <40 mK | Fits light aircraft; most sensitive published NETD in the line |
| Corridor and standoff work | Zoom + thermal + LRF | OP-90A — 10x optical, 640×512, 1,200 m ranging | Distance-tagged findings from safe standoff |
| Long-range and large-site | Flagship EO/IR | OP-125A / LX-9B — 30x optical, 2,000 m ranging | One hover covers multiple bays or spans |
Specify the thermometry version wherever the deliverable quotes temperatures. Radiometric vs non-radiometric covers why that is a purchase-time decision, and what drives payload price covers the six cost drivers behind the quote.
Do not economise on stabilisation. A gimbal that cannot hold its pointing specification wastes every dollar spent on resolution and zoom above it — at 1,000 m the difference between ±0.01° and ±0.1° is 17 cm versus 1.7 m of wander.
The Two-Stage Reality
Aerial inspection finds problems; it does not always characterise them. Mature programmes fly the site to produce an anomaly map and then verify the flagged points at close range — with a handheld thermal camera, a chain drag, a core sample or a hands-on inspection, depending on the asset.
This is not a weakness of the aerial survey. It is what makes the aerial survey valuable: a whole site examined in one flight, with expensive verification confined to the handful of places that need it. Drone thermal versus handheld covers the division of labour, and thermal reflections and false hot spots covers the findings that ground verification correctly dismisses.
Programmes that skip stage two issue reports full of anomalies a ground check would have rejected, and they lose credibility with the engineers who have to act on them. Programmes that skip stage one inspect everything at ground-verification cost. The value is in the ratio between them.
The Commercial Side
Asset owners pay for decisions, not photos. A folder of thermal imagery is a cost to them; a ranked list of findings with severity, location and recommended action is something they can put into a maintenance system.
Deliver classified findings with severity and location, and price against the access equipment you replace rather than against your own hours. A survey that displaces a snooper truck, a lane closure and a traffic management plan is worth what that costs, not what your flight time costs — what a report should contain covers the structure.
And budget for the analyst, not just the aircraft. At scale the flying is the cheap part; interpretation and reporting dominate the hours, and untrained interpretation is where confident nonsense enters the record — see thermography training and certification and starting a drone thermal inspection business.
Related reading
- Radiometric Temperature Measurement on UAV Payloads
- Gimbal Stabilization Technology
- Substation Thermal Inspection by Drone
- Wind Turbine Blade Inspection
- Bridge and Infrastructure Inspection
- Building Envelope Thermal Surveys
- What a Thermal Inspection Report Should Contain
- Drone Thermal vs Handheld Thermal Camera
- MV-2P — 640×512 thermal at 130 g
- OP-90A — zoom, thermal and ranging
- OP-125A — flagship EO/IR pod
- Power Line Inspection Drone Payloads
FAQ
Which inspections require radiometric thermal?
Any that feed maintenance decisions with temperature numbers: electrical inspection, solar classification, industrial trend monitoring, building envelope audits. Detection-only missions can skip it — search, security patrol and wildlife work need contrast, not calibration. The decision must be made at purchase, because radiometry depends on factory calibration of the specific core and cannot be added by firmware afterwards.
How do I price inspection services?
Asset owners pay for decisions, not photos. Deliver classified findings with severity and location, and price against the access equipment you replace — a snooper truck plus a lane closure plus traffic management is a five-figure day, and a survey that displaces it is worth a fraction of that rather than a fraction of your flight hours. Pricing against your own time systematically undervalues the work.
What is the single most common programme mistake?
Treating evidence quality as a camera property. Two operators with identical hardware produce very different deliverables, and the difference is process: consistent altitude and angle, logged conditions, radiometric capture where numbers matter, and comparison against the previous flight. A programme that cannot compare this quarter with last quarter is producing surveys rather than monitoring, and monitoring is what asset owners actually buy.
Questions about the technology? Talk to our engineers — we reply within 2 business days.