Plain RGB drone imagery — ordinary colour photographs from the air — is genuinely useful for inspection and mapping. It is also routinely oversold for rehabilitation monitoring, usually by people who haven't looked closely at what a normal camera can and cannot see. Three traps come up again and again.
1. A colour camera is blind to most of the vegetation signal
Healthy vegetation has a distinctive spectral fingerprint: chlorophyll absorbs strongly in the red and blue, reflects a little in the green (which is why leaves look green), and then reflects very strongly in the near-infrared just beyond what the eye can see. The steepest, most information-rich feature in that fingerprint is the red edge — a sharp rise in reflectance around 700–750 nm whose position and slope shift with plant stress before any change is visible.
A standard RGB camera records none of it. It captures red, green and blue and stops at the edge of human vision. So the signals that actually indicate vigour, stress and biomass — near-infrared and red-edge — simply aren't in the file. A reassuringly green photograph can sit over stressed, sparse, or wrong-species cover. Real vegetation indices (NDVI and its relatives) exist because they use the near-infrared band; without it, "it looks green" is an impression, not a measurement. My own research took this to its logical end, using hundreds of narrow hyperspectral bands to detect nutrient deficiencies an RGB image could never show.
2. One flight is a snapshot; rehabilitation is a trajectory
Rehabilitation is a claim about change over years — disturbed ground becoming stable, vegetated land. A single drone flight, however sharp, is one point on that curve. Establishing recovery needs a consistent time series measured the same way at each step, and often it needs to reach back before monitoring began. That history usually exists only in the satellite archive, which has imaged most sites every few days since 2015 (and, in coarser form, for decades). A one-off flight cannot evidence a trend it never observed — and stitching flights from different days, cameras and light into a trend is a radiometric problem most operators never address.
3. Resolution is not accuracy, and neither is defensibility
A crisp two-centimetre image feels authoritative, and that detail has real value for inspecting a specific feature. But spatial resolution is not radiometric accuracy, and sharpness is not validation. Evidence that holds up needs ground-truth points, a documented and repeatable method, and a stated uncertainty — the same things a peer reviewer would demand. Without them, a beautiful high-resolution image is exactly that: a picture, not something a regulator or auditor can test.
Where drones belong — and where they don't
None of this means drones have no place. For fine detail, hard-to-reach features and targeted checks, they earn their keep, and a drone carrying a proper multispectral sensor is a serious instrument. The point is narrower: the backbone of rehabilitation monitoring is a documented, validated, multi-year spectral time series, and the value a client is really buying is the interpretation behind it — not the aircraft, and not the megapixels.
That backbone is what our rehabilitation monitoring service is built on. If you're weighing up how to evidence recovery on a footprint, tell us the site and the deadline.