This is the same NDVI remote-sensing physics as the 3D pushbroom scanner, but computed a genuinely different way: instead of blending two discrete bands, it builds a full continuous reflectance spectrum R(λ) from 400–1000nm from first principles, integrates real sensor bands out of that spectrum, and numerically detects the red-edge inflection point — while a logistic canopy-growth ODE actually integrates a growing season instead of a static parameter sweep.
R_leaf(λ) = visible pigment absorption (λ<680nm)
⤷ logistic red-edge rise centered at λ₀(Cab) (λ≥680nm)
R_pixel(λ) = (1−f)·R_soil(λ) + f·R_leaf(λ)
Red, NIR = mean of R_pixel(λ) over the sensor's real bandwidth
NDVI = (NIR − Red) / (NIR + Red)
λ_red-edge = argmax dR_pixel/dλ over 680–760nm [numerical]
df/dt = r·f·(1 − f/f_max) [canopy growth ODE]
- Chlorophyll content (Cab) — deepens the red-absorption dip and shifts the red-edge inflection to longer wavelengths, a real, well-documented hyperspectral chlorophyll indicator distinct from the raw NDVI number.
- Mesophyll structure health — sets the height of the near-infrared plateau; degraded internal leaf structure scatters less NIR light even at unchanged chlorophyll, which is exactly why some stress signatures show up in NIR before the visible bands change.
- Temperature & water availability — drive the rate and carrying-capacity of a logistic canopy-growth equation integrated live, frame by frame, so vegetation cover f — and the NDVI it produces — genuinely emerges over a simulated growing season instead of being set by a slider.
- Red-edge λ — found numerically each frame as the wavelength of steepest reflectance rise (finite-difference slope search across 680–760nm), the same first-derivative technique used on real hyperspectral imagery.