Drone and satellite crop monitoring

Aerial crop monitoring is the most photogenic part of precision agriculture and the one with the widest gap between capability and realised value. The capability is real: stressed crops change their reflectance before they change their appearance, so imagery can find a developing problem days before a walk-through would.

The gap is that a map is not an action.

What an index is and is not

Vegetation indices combine reflectance in different wavelength bands into a single number that correlates with canopy condition. NDVI is the best known: healthy vegetation reflects strongly in the near infrared and absorbs strongly in the red, so the normalised difference between them tracks green biomass.

This is genuinely informative and routinely over-interpreted. An index is a symptom, not a diagnosis. A low patch might be water stress, nitrogen shortage, disease, compaction, pest damage, a shallow soil, or the place where a machine turned. The imagery has narrowed a large field to a small area worth walking. That is its actual function, and it is a valuable one.

The workflow that works is: image, identify anomalies, go and look, then decide. The workflow that fails is: image, generate prescription, apply. The second skips the only step that establishes cause.

Satellite first, almost always

For most growers the honest recommendation is to exhaust free satellite imagery before buying anything.

Public Earth observation now provides multispectral imagery at around ten metre resolution every few days, free, with a multi-year archive. For broadacre crops that resolution defines management zones perfectly adequately, and the archive is arguably more valuable than any single image, because it shows which parts of a field underperform consistently rather than just this season.

Persistent underperformance is a soil or drainage problem worth investigating properly. A one-off patch is an event. Distinguishing them requires history, and the satellite archive supplies it at no cost.

Where drones win

Four situations justify the equipment. Plots too small for ten metre pixels, which covers most horticulture and all research trials. Perennial crops where individual trees or vines matter. Situations where timing is critical and a fixed revisit cycle is too slow. And regions where cloud cover makes satellite coverage unreliable during the months that matter - flying under cloud is a genuine structural advantage.

There is also a strong case for a cheap ordinary camera drone regardless. Blocked emitters, broken sprinklers, drainage failures, storm damage and flooding are all obvious in plain colour imagery and expensive to find on foot.

The costs nobody quotes

The aircraft is the small part. Pilot certification and airspace compliance apply in most jurisdictions and take time. Multispectral data requires radiometric calibration to compare flights, which means a reference panel and disciplined procedure. Processing is computationally heavy and the data volumes are large. And someone has to interpret the output every time, in season, when they are already busy.

Farms that get value from drones almost always either have a dedicated person or buy the flight as a service. Farms that buy the aircraft and add it to an existing workload generally fly it enthusiastically for one season.

Related

Imagery locates variability; explaining it usually needs ground measurement, which is where soil moisture sensors come in. Storing and comparing imagery across seasons is a job for farm management software.