Solar panel drone inspections

What you actually get

• Radiometric thermal imagery of every panel — actual temperature data, not a pretty rainbow JPEG
• A defect map flagging hotspots, failed cells, dead strings and bypass-diode faults, panel by panel
• Matching 4K visual imagery so real faults are separated from soiling and shading
• A report your installer or O&M contractor can act on, with each fault located on the array
• Raw radiometric files on request, for your own records or a second opinion

What it costs

Rooftop and small commercial arrays are usually priced as thermal survey work — typically £300–£800, depending on array size and reporting depth. Ground-mount farms are quoted per site: state the capacity and layout in your request and pilots will price the actual job.

What moves the price

• Array size — a domestic roof vs. a field of ground-mounted strings
• Reporting depth — a flagged-fault summary vs. a panel-by-panel radiometric report
• Access and airspace — open ground-mount sites are straightforward; rooftop arrays in controlled airspace add planning time

How the inspection is actually flown

A solar inspection is a planned grid flight, not a wander with a camera. The pilot maps the array, sets a flight height that gives the thermal sensor enough resolution to see individual cells, and flies overlapping passes with the camera held near-vertical to the panels. Thermal and 4K visual imagery are captured together, so every frame of temperature data has a matching photograph.

Time on site is short. A domestic or small commercial rooftop is a single flight, usually under an hour including setup. Ground-mount sites scale with area — a drone covers in an afternoon what a handheld thermal camera would take days to walk.

The weather window matters more than the diary. Faults appear as temperature differences, and those need strong, stable sun — around 600 W/m² of irradiance, in practice a clear day between late morning and mid-afternoon. High wind cools the panels and flattens the contrast; patchy cloud makes readings unstable. Good pilots wait for the window, because the data is worthless without it.

Reading the thermal signatures

Radiometric imagery turns electrical faults into visible patterns, and the shapes matter — each failure mode has a recognisable signature. A competent report names them rather than circling warm blobs:

• Single hot cell — often a cracked cell, a poor solder joint or persistent localised shading. Hotspots degrade the panel and tend to worsen, so they're worth tracking even when output looks fine
• One hot band across a panel — a stripe covering roughly a third of the module is the classic bypass-diode signature: the diode has failed or is stuck conducting, and that substring is producing nothing
• Whole panel warm — a module running uniformly hotter than its row usually isn't producing at all: disconnected, open-circuit or failed
• Whole string warm — every panel in one string slightly hotter than the rest of the array means the string is offline. This is the dead string that site-level monitoring can miss
• Scattered warm cells near string ends — a possible sign of potential-induced degradation (PID), worth catching early because it spreads
• Heat at the junction box — points to a connection fault rather than a cell problem

Soiling or fault? Why the visual imagery matters

Bird mess, leaves, lichen and shading all show warm on thermal, and none of them needs a replacement panel. That's why visual imagery is captured in the same pass: a leaf has an obvious photographic explanation, a bypass-diode failure doesn't. A good report uses both to separate "clean these panels" from "replace this one" — that distinction is most of its value, and a quote offering thermal-only imagery can't make it.

Standards and accuracy — what a proper inspection looks like

There is a technical standard for this work: IEC TS 62446-3 covers outdoor thermographic inspection of PV systems, including from the air. You don't need to read it — you need a pilot who works to something like it. The practical expectations it sets:

• Minimum irradiance — strong, stable sunlight, so faults develop a measurable temperature difference
• A genuinely radiometric camera — one that records a temperature value for every pixel, not a colourised video feed
• Enough resolution per panel — flying low enough that individual cells are distinguishable; too high and a diode fault smears into nothing
• Sensible geometry — camera angle close to perpendicular to the panels, to limit reflections of the sky and the drone itself
• Low wind — strong wind cools the glass and hides the contrast between healthy and faulty cells
• The system energised and under load — switched-off panels all look the same temperature

Rooftop arrays vs ground-mount farms

The physics is identical; the logistics aren't. Ground-mount sites are the easy case: open ground, predictable row layout, usually nobody about, and airspace that's rarely complicated. Flights are repeatable grids, which also makes year-on-year comparison clean.

Rooftops add the constraints. Domestic and commercial arrays sit near people, parked cars and neighbouring property, and urban sites can fall in controlled airspace — all manageable for a qualified pilot, but it shapes the flight plan and sometimes the lead time. In return, a drone inspects a rooftop array with nobody walking the roof, no scaffold, no harness plan and no panels isolated. For steep, fragile or high roofs, that alone justifies the method.

Drone vs handheld thermography — and when not to fly

Handheld thermography still has a place. A technician on a walkway gets closer to a suspect module than a drone reasonably can, and for diagnosing one known-faulty panel it's often the better tool. What it can't do is scale: walking a large site with a handheld camera takes days and samples the array rather than covering it.

Be honest about when a drone is the wrong call. If inverter monitoring has already isolated a fault to one string, an electrician with an I-V curve tracer answers the question directly — thermography would only tell you where to point the tracer. A very small array may be cheaper to check during a routine electrical service. And no aerial thermal inspection works under flat overcast skies, however good the pilot. Thermography is a screening tool: out of thousands of panels, it tells you which handful deserve electrical testing. It points the testing; it doesn't replace it.

How often to inspect

For routine O&M, an annual thermal flyover in the high-irradiance months is the common cadence. Some owners stretch to every two years on small, young arrays; sites with performance guarantees tend to fly yearly without much debate. Beyond the routine, four moments earn an extra flight:

• At commissioning — a thermal pass before sign-off catches manufacturing and installation defects while putting them right is still the contractor's problem
• Before warranty expiry — the last cheap chance to put panel defects on the manufacturer's account
• After extreme weather — hail, storms and lightning leave thermal evidence before visible damage shows
• When monitoring shows a shortfall you can't explain — the flyover turns a vague underperformance into a list of named panels

From report to repair

The report is the start of a process, not the end of one. A usable defect map keys every fault to the array's as-built layout — row and module position — so a contractor can walk to the right panel without re-surveying, and classifies faults by severity: act now, watch, cosmetic.

From there, your O&M contractor or installer verifies the flagged panels electrically, repairs or replaces, and submits warranty claims where the module itself is at fault — manufacturers generally expect radiometric evidence, which is exactly what the survey produced. After repairs, a re-fly of the affected sections confirms the fix; some pilots will price that follow-up in the original quote, and it's worth asking.

CAA rules and insurance, in plain English

The aviation compliance is the pilot's job, but here's the shape of it. Every commercial drone pilot needs a CAA Operator ID and a recognised qualification — typically an A2 CofC, or a GVC with an Operational Authorisation for work closer to people and property not involved in the job. How close a drone may legally fly depends on the aircraft and the authorisation held: over your own ground-mount site with nobody about, the rules are easily satisfied; a rooftop on a busy street needs more planning; controlled airspace near an aerodrome adds lead time, not impossibility.

Two things to check rather than assume: that the qualification is current, and that the pilot carries public liability insurance appropriate for commercial work. Every pilot on Sober Pilots has both verified, and the checks are dated — but whoever you hire, ask. A serious operator expects the question.

Briefing a pilot — and what a good quote includes

A quote worth accepting names the camera (radiometric, and its resolution), states the irradiance conditions the pilot will fly in, lists the deliverables, and includes the pilot's CAA credentials and insurance. If any of those are missing, ask.

In return, quotes are only as good as the brief. Tell pilots:

• System size — capacity in kWp or MW, or a panel count if that's easier
• Mounting — rooftop (and what kind of roof) or ground-mount
• Location and access — postcode, plus anything awkward about getting a pilot and kit to the array
• What prompted the inspection — routine O&M, a monitoring alert, storm damage, a pre-purchase check
• The output you need — a flagged-fault summary, or a panel-by-panel report your contractor can work from

Aftercare: files, ownership and retention

Three questions to settle before the flight, not after. Formats: you should receive the report as a PDF, the defect list in a form your contractor can actually use (commonly a spreadsheet or CSV), and — if you want them — the raw radiometric files, which are what a second opinion or a warranty claim works from. Ownership: the deliverables should be yours; make sure the quote says so. Retention: pilots typically keep flight data for a period for their own records — confirm how long, because next year's inspection compared against this year's baseline tells you whether a fault is growing, and a one-off snapshot can't.

Pilots offering this service

Browse the full directory >