01 / What you are seeing
A deterministic, synthetic 3.84 ha maize parcel. Red, near-infrared, red-edge and canopy-temperature patterns represent plausible spatial responses. No imagery, live satellite feed or farm measurements are being claimed.
Vegetation indices are computed from simulated reflectance. The 10 m analysis cells aggregate observations; they are not individual drone image pixels. Robot, crop and buildings are illustrative geometry.
02 / Why multiple sensors?
NDVI = (NIR − red) / (NIR + red). It responds to vegetation amount and greenness, but can saturate in dense canopy. NDRE substitutes the red-edge band and adds sensitivity to canopy chlorophyll variation. Neither index identifies a unique cause.
A hot canopy may support a water-stress hypothesis. Temperature also depends on weather, canopy cover, soil background, timing and sensor calibration. This demo does not compute a calibrated crop water stress index.
03 / Measurement quality
Ground sampling distance is illustrative: GSD = altitude × 3.45 μm pixel pitch / 8.8 mm focal length. The thermal camera has different optics and resolution. Flight path and sensor footprint are explanatory, not a flight-planning product.
Variable illumination lowers the synthetic quality score and increases measurement uncertainty. Real surveys require reflectance calibration, suitable illumination, georeferencing, overlap and independent validation. Confidence is a demonstration score, not a calibrated probability.
04 / From map to management
Check soil moisture, plant tissue and field conditions before deciding. A sample here supports only the selected 10 m cell in this simulation. Approved responses are investigation or management tasks, not agronomically calibrated fertilizer doses or autonomous machine commands.
Drone imaging offers flexible, detailed surveys; satellite observations can provide regional coverage and repeat context. Neither replaces representative field sampling. Area prioritisation is not a measured saving in cost, water, emissions or yield.
