AgTech - Science Lecture 2

True Colour Composites:
- Use visible Red, Green, and Blue bands to resemble natural colour appearance.
- Purpose: provide a natural look for general interpretation.
False Colour Composites:
- Map a non-visible band (commonly Near-Infrared or NIR) into the visible spectrum (often Red) to enhance contrast.
- Benefit: vegetation stands out from soil and other features, improving discrimination.

Indices reduce multi-spectral reflectance to a single numerical value to indicate vegetation status.
Primary target: photosynthetically-active biomass (PAB).
Most popular indices: Normalised Difference Vegetation Index (NDVI) and Simple Ratio (SR).

Simple Ratio (SR): SR = \frac{NIR}{R}
- Often interpreted as a measure related to Plant Cell Density (PCD).
Normalised Difference Vegetation Index (NDVI): NDVI = \frac{NIR - R}{NIR + R}
- Range: -1 \le NDVI \le 1.
- Higher values indicate more vegetation activity, typically positive for vegetation (0 to ~1).
- Lower values indicate non-vegetated surfaces.
- Used as a proxy for PAB and, with calibration, for Leaf Area Index (LAI).
Both SR and NDVI are useful indicators of PAB; NDVI is generally more robust across varying conditions.

Common band pair examples: NIR = 753\,\text{nm} and Red = 682\,\text{nm}.
Interpretation: higher NIR reflectance relative to Red indicates healthier vegetation.
NDVI sensitivity to canopy density:
- Increasing leaf layers (pseudo-LAI) raises NDVI values:
  - 1 leaf: NDVI ~ 0.78
  - 2 leaves: NDVI ~ 0.83
  - 3 leaves: NDVI ~ 0.93

SAVI (Soil-Adjusted Vegetation Index):
SAVI = \frac{(1 + L)(NIR - R)}{NIR + R + L}, where L\in[0,1] (L is a soil brightness correction factor).
Additional non-linear variants include NLI, MNLI, and MSR.
Note: Many indices exist, primarily relying on Red and NIR bands.

LAI (Leaf Area Index): defined as the total leaf area per unit ground area (e.g., two leaves stacked implies LAI = 2).
Normalised SR (nSR): SR normalized by SR at a reference LAI (e.g., LAI = 5).
Both NDVI and SR are valuable for mapping and monitoring crop vigor, biomass, and health.
Display and interpretation of indices rely on accurate radiance reflectance data from sensors.

Yield forecasting: Demonstrated with crops like carrots, where canopy-derived Vegetation Index (VI) correlates with root yield.
Block-level VI-based predictions: Analysis of mango blocks for fruit counts and size estimations.
Data farming platforms: Utilization of VI time-series data to support management decisions, integrating area-wide imaging and cloud platforms.

Cloud cover: Significantly limits image availability.
Image processing requirements: Involves extensive preprocessing and calibration.
Costs: Associated with data capture and intricate processing.
2D information: Provides limited insights into depth or volume.
Daytime imagery only: Restricts monitoring to daylight hours.