Solar farm drone inspection

On a solar farm, thermal imaging detects faults that are invisible to the naked eye under normal operating conditions. A healthy solar cell converts sunlight into electricity. A faulty cell — whether cracked, delaminated, or affected by potential-induced degradation (PID) — does not convert that energy. It absorbs it as heat instead. That cell runs measurably hotter than its neighbours, and that temperature difference is what the thermal camera sees.

Different faults produce different thermal signatures. A single bright cell in an otherwise uniform panel suggests a cracked cell, localised shading, or physical damage. An entire string of cells running warmer than the rest of the array indicates string underperformance. A warm, uniform rectangular patch extending across a panel typically points to an activated or failed bypass diode, while irregular warm patches at the edges can indicate delamination. In each case, what the camera records is the heat differential — the contrast between a healthy panel and one that is absorbing energy rather than converting it. The drone captures every panel on the site in a single sortie — no panels removed, no array de-energised, and no ground team walking the rows.

A thermal camera can detect hotspots caused by cracked or delaminated cells, activated or failed bypass diodes, underperforming strings, and potential-induced degradation (PID). Each fault type produces a recognisable thermal signature: a single bright cell suggests a crack, a warm rectangular patch points to a bypass diode issue, and an entire string running warmer than its neighbours indicates string-level underperformance.

The key point is that these faults are invisible to the naked eye under normal operating conditions. A panel that looks perfectly normal in a visual inspection can be losing measurable output — and the thermal camera sees the heat differential that reveals it.

For a typical utility-scale solar farm, the on-site flight can be completed in a single day. The drone follows a pre-planned automated flight path that ensures every panel is captured from the correct angle and altitude, with thermal and visual imagery recorded simultaneously. Post-flight processing — stitching the imagery into a georeferenced orthomosaic and producing the annotated fault map — typically takes a few working days.

The comparison that matters is not flight time versus flight time. It is that a ground team covering the same site manually can take weeks, with panels needing to be checked one by one. A drone covers every panel in a single sortie with no downtime and no interruption to generation.

An orthophoto is an aerial image that has been geometrically corrected so that every pixel represents a true top-down view, free from the perspective distortion of a regular photograph. Unlike a standard aerial photo, you can measure true distances on an orthophoto: the distance between two panels, the area of a section of the array, the coordinates of a specific fault. When multiple orthophotos are stitched together to cover an entire site, the result is an orthomosaic. These can be huge and incredibly detailed: a single orthomosaic of a utility-scale solar farm can be hundreds of megapixels in size, with every panel visible and every fault located to within centimetres.

For a solar farm, the orthomosaic is the foundation deliverable. Every panel is accurately positioned, every defect is located by GPS coordinates, and the image can be loaded directly into CAD, GIS, or asset management software. It is the difference between “here is a photograph of your site” and “here is a measurable, georeferenced map of your site.”

Panel manufacturers typically warrant their modules for 25 years, but a warranty claim requires evidence. A thermal drone survey produces a timestamped, georeferenced dataset that records the condition of every panel on a specific date — each fault GPS-tagged, each thermal anomaly captured and classified. That dataset is an objective baseline: if a string of panels degrades below the warranted performance threshold, the before-and-after evidence is already in place.

Photogrammetry is the process of using multiple overlapping aerial photographs, taken from slightly different angles, to produce accurate 2D maps and 3D models of the ground below. Software identifies the same point on the ground across dozens of images and triangulates its true position in three-dimensional space. The result is a survey-grade orthomosaic, a Digital Elevation Model showing bare-earth terrain, and a Digital Surface Model that includes buildings, hedgerows, and vegetation.

For solar developers, photogrammetry is the survey method that feeds directly into specialist solar design CAD software. One flight delivers the orthophoto, the elevation data, and the site-feature mapping that a designer needs to assess terrain suitability, model panel placement, and run shading analysis across the full site.

A Digital Elevation Model (DEM) represents the bare ground with all surface features — trees, buildings, hedgerows — digitally removed. It shows the shape of the terrain itself, and is what a solar designer needs to assess slope, aspect, and drainage across the site. A Digital Surface Model (DSM) includes all surface features: buildings, vegetation, walls, and other structures. A designer uses the DSM to model shading from trees and hedgerows across the year.

Both are produced from the same drone flight using photogrammetry. The drone captures overlapping images across the site, and the processing software triangulates every visible point in three-dimensional space — producing the orthomosaic, DEM, and DSM as outputs of the same dataset.

A standard aerial photo can be misleading. Grass, crops, and shadows can hide the true shape of the ground — a gentle slope can look flat, and a drainage depression can disappear under vegetation. An elevation heatmap replaces colour with absolute height data: every pixel has a known Z-value, so slope angles, aspect (which way the ground faces), and natural drainage paths are unambiguous.

For a solar developer, this matters at the feasibility stage. The heatmap reveals whether a site’s slopes face south (good for yield), whether there are depressions where surface water will pool around inverter stations, and whether the natural drainage paths will work with or against the cable trench layout. It turns a site visit into a measurable dataset that feeds directly into PVsyst and solar design CAD.