Characterizing leaf geometry for grass and crop canopies from hotspot observations: A simulation study

The potential of canopy reflectance distributions in the hotspot region for characterizing leaf geometry (leaf size and shape) of grass and crop canopies is explored with computer simulations. In this article, a computer graphics method—Lindenmayer-systems (L-systems)—is used to render a series of leaf (grass) and architecturally realistic row-planted crop (corn-like) canopies that have a variety of geometrical structures. A radiosity–graphics combined model is then employed to calculate the radiation regime in a canopy, including canopy directional reflectance. An effectiveness ratio (E ratio) is proposed, which is able to evaluate the performance of a given measure or index in estimation of the parameter of interest under the influence of a number of “noise” factors (other geometric and optical parameters of the canopy) at various noise levels. This E ratio is then applied to evaluate reflectance and normalized reflectance in the hotspot region for leaf geometry characterization. The result from simulated hotspot reflectance demonstrates that for both canopies, leaf geometry is estimable by using normalized reflectance within ±4–8° (or ±2–4°) around the hotspot direction in the principal cone (or principal plane). However, the center position and angular width of the optimal sampling region are affected by the number of noise factors [such as leaf area index, leaf angle distribution for leaf canopies, plus row structure for row-plant crop canopies] and their variation ranges. In most cases, normalized spectral reflectance in the near-infrared at a high solar zenith angle in the PC produces the most reliable results. The reason for better estimation of leaf geometry for grass and crop canopies than forests from hotspot observations is also discussed in this article.