PROSPECULAR: A model for simulating multi-angular spectral properties of leaves by coupling PROSPECT with a specular function
PROSPECULAR: A model for simulating multi-angular spectral properties of leaves by coupling PROSPECT with a specular function
复制标题
DOI：
10.1016/j.rse.2023.113754
复制
发表时间：
2023-11
影响因子：
13.5
通讯作者：
Xiao Li;Zhongqiu Sun;Sha Lu;K. Omasa
中科院分区：
文献类型：
--
作者：
Xiao Li;Zhongqiu Sun;Sha Lu;K. Omasa
了解叶片的光学特性对于表征叶片性状与遥感信号之间的关系具有重要意义。生物化学参数和叶片光谱反射之间的关系已经通过前景广泛建立，而表面结构参数可以通过双向反射分布函数（BRDF）模型与叶片多角度反射联系起来。然而，它仍然是具有挑战性的，以表征如何叶光谱响应变化与不同的照明和观看几何形状，在连接到生化和表面结构性状。为了解决这一挑战，我们提出了模型PROSPECULAR，它结合了前景与镜面反射功能确定的照明观看的几何参数，波长相关的折射率，和粗糙度。该模型模拟光谱和多角度BRF和检索叶片生化和表面结构参数。PROSPECULAR在镜面反射函数中考虑了叶片表面折射率的波长依赖性，并在漫反射分量计算中将镜面反射与前景分离。PROSPECULAR光谱模拟的准确性是使用在半球空间中测量的多角度数据集（在20-50°源天顶角（SZA）处的42个观察天顶角（VZA））进行评估的。然后，使用主平面上测量的多角度数据集（13个VZA和20-50° SZA）验证PROSPECULAR的反演精度。结果表明，PROSPECULAR能够模拟多角度的BRF（RRMSE = 16.27%），并利用多角度BRF反演了每个叶片样品的一组统一的生化参数叶绿素、水分、干物质、蛋白质和碳基成分的RMSE分别为11.31 μg/cm 2、0.0029 g/cm 2、0.0022 g/cm 2、0.00096 g/cm 2和0.0028 g/cm 2，分别）。对不同植物的多角度BRF的计算结果与实测结果一致，表明PROSPECULAR是一个有效的叶片光学模型。相比之下，现有的辐射传输模型（前景，PROCOSINE和PROSDM）在反演生化参数时显示出不同的准确度，这取决于VZA。PROSPECULAR的优点是它的潜力，在任意给定的照明观察几何形状和反演叶参数使用BRF数据从几个VZA在半球空间的光谱和方向特性的叶片。
Understanding the optical properties of leaves is of great importance for characterizing the relationships between leaf traits and remote sensing signals. The relationship between biochemical parameters and leaf spectral reflection has been widely established through PROSPECT, while surface structural parameters can be linked to leaf multi-angular reflection through Bidirectional Reflectance Distribution Function (BRDF) models. However, it's still challenging to characterize how leaf spectra response changes with different illumination and viewing geometries, in connection to both biochemical and surface structural traits. To address this challenge, we propose the model PROSPECULAR, which combines PROSPECT with a specular function determined by illumination-viewing geometry parameters, wavelength-dependent refractive index, and roughness. This model simulates spectral and multi-angular BRF and retrieves both leaf biochemical and surface structural parameters. The PROSPECULAR accounts for the wavelength dependence of the leaf surface refractive index in the specular function and separates the specular reflection from PROSPECT in the diffuse component calculation. The accuracy of PROSPECULAR's spectral simulation was assessed using a multi-angular dataset measured in the hemispherical space (42 viewing zenith angles (VZAs) at 20–50° source zenith angles (SZAs)). Then, the inversion accuracy of PROSPECULAR was validated using a multi-angular dataset measured in the principal plane (13 VZAs and 20–50° SZAs). The results showed that PROSPECULAR was able to simulate multi-angular BRF (RRMSE = 16.27%) when SZA was <50° and retrieve a unified set of biochemical parameters for each leaf sample using multi-angular BRF (RMSE of chlorophyll, water, dry matter, protein, and carbon-based constituents were 11.31 μg/cm2, 0.0029 g/cm2, 0.0022 g/cm2, 0.00096 g/cm2,and 0.0028 g/cm2, respectively). The agreement between the calculations and measurements of multi-angular BRF for various plant species indicated that PROSPECULAR is an effective leaf optical model. In contrast, the existing radiative transfer models (PROSPECT, PROCWT, PROCOSINE, and PROSDM) showed varying levels of accuracy in inverting biochemical parameters, depending on the VZA. The advantage of PROSPECULAR is its potential for characterizing both spectral and directional properties of leaves at arbitrary given illumination-viewing geometry and inverting leaf parameters using BRF data from several VZAs in the hemispherical space.