Precision farming runs a continuous sense → analyse → act loop: sensors and imagery measure the field, models flag problems, and actuators (irrigation, sprayers) respond automatically.
NDVI = (NIR − Red) / (NIR + Red)
NDVI ≈ 0.6–0.9 → dense, healthy canopy
NDVI ≈ 0.2–0.4 → stressed / sparse crop
NDVI < 0.1 → bare soil or water
Soil-moisture control loop:
if moisture(cell) < threshold: irrigate(cell) // IoT sensor + actuator
Pest spread (untreated):
P(infect neighbour) grows each tick while undetected
drone overflight → detected = true → pest decays, NDVI recovers
- View — Visual shows the crop canopy as a farmer would see it; NDVI shows the same field as a false-colour vegetation-index map, the way a multispectral drone or satellite reports it.
- Auto-irrigate threshold — the soil-moisture level that triggers each plot's sprinkler; lower it and the field runs drier but saves water, raise it and yield stabilises at higher cost.
- Scan speed — how fast the drone sweeps its boustrophedon (back-and-forth) survey path; faster scans detect outbreaks sooner but each pass covers a plot for less time.
- Trigger outbreak — seeds a pest/disease patch that spreads to neighbouring plots until the drone's next overflight detects and treats it, mirroring real early-warning computer-vision systems.