Foliage density and spatial distribution affect droplet dispersion and deposition inside the plant canopy in pesticide spraying. Quantitative expression of foliage density and spatial distribution would contribute to improving pesticide spraying effect. Stratification porosity was proposed to describe density and spatial distribution of cotton foliage in this paper, and an estimation model of stratification porosity for cotton canopy based on morphological characteristics was put forward. Besides the randomness of local branches and leaves resulting from competitive growth, there was traceable statistical regularity about density and spatial distribution of cotton foliage based on growth law and morphological characteristics. Accordingly, an estimation model of stratification porosity was presented, which was mainly driven by growing-degree days (GDD), incorporating row spacing, plant spacing and the randomness of crop azimuth. To validate the model, a laser ranging bench was designed to measure stratification porosity. The cotton canopy was stratified into three layers in the vertical direction. Estimated stratification porosity of each layer was compared with measured values. For flowering stage, the root mean square errors (RMSE) of lower layer, middle layer and upper layer were 3.97%, 2.08% and 6.50% respectively. The mean relative errors (MRE) were 5.6%, 5.3% and 10.9% respectively. For bud stage, RMSE and MRE were 9.2% and 12.6% respectively. The estimation model was considered to be an acceptable approach for porosity acquisition with adequate accuracy. Spatial variation of stratification porosity indicated inhomogeneity of foliage distribution. Therefore, the model could provide quantitative information of cotton foliage density and spatial distribution, which would be helpful for pesticide spraying.

中文翻译：

---

基于形态特征的冠层分层孔隙度估计模型——以棉花为例

叶子密度和空间分布影响农药喷洒中植物冠层内的液滴扩散和沉积。叶片密度和空间分布的定量表达将有助于提高农药喷洒效果。提出用分层孔隙度来描述棉花叶片的密度和空间分布，并提出一种基于形态特征的棉花冠层分层孔隙度估计模型。除了竞争性生长导致局部枝叶的随机性外，棉花叶片的密度和空间分布根据生长规律和形态特征存在可溯源的统计规律。据此，提出了分层孔隙度的估计模型，主要受生长度日（GDD）驱动，包括行距、株距和作物方位角的随机性。为了验证模型，设计了一个激光测距台来测量分层孔隙度。棉花冠层在垂直方向上分为三层。将每一层的估计分层孔隙率与测量值进行比较。对于开花期，下层、中层和上层的均方根误差(RMSE)分别为3.97%、2.08%和6.50%。平均相对误差 (MRE) 分别为 5.6%、5.3% 和 10.9%。对于芽期，RMSE 和 MRE 分别为 9.2% 和 12.6%。估计模型被认为是具有足够精度的孔隙度采集的可接受方法。分层孔隙度的空间变化表明叶分布的不均匀性。因此，该模型可以提供棉花叶片密度和空间分布的定量信息，有助于农药喷洒。