Leaf area index (LAI) is the important parameter for characterizing vegetation canopy structure and photosynthetic area size, as well as an important indicator to judge the status and trends of vegetation growth. Therefore, it is of great significance for monitoring the vegetation LAI real-time and dynamic. An optimal estimation model for the East Dongting Lake Carex LAI was designed by mathematical transformation of hyper-spectral data and feature analysis. Through analyzing the spectral reflectance and spectral signature of 52 Carex samples (bands ranging from 640 ~ 780 nm) which were collected from the east Dongtin lake, selecting characteristic bands, an estimation model was established by adopting multiple element step by step linear regression and partial least squares regression. The results show that the estimated characteristic band of the wavelength range were from 707~755 nm, the mathematical transformation of the spectral reflectance was good for the characteristic band bands choice; the SMLR model and PLSR model determination coefficient of the three types spectral datat were 0.526,0.815,0.565, and the root mean square error were 0.320,0.269,0.273, while the PLSR model determination coefficient were all more than 0.9, the root mean square error were respectively 0.189, 0.262, 0.134. It is suggested that partial least squares regression method is better than stepwise multiple linear regression, it can estimate the LAI of Carex quickly and efficiently.%叶面积指数(leaf area index,LAI)是表征植被冠层结构和光合面积的重要参数.实时动态的植被LAI监测对于诊断植被生长状况及趋势具有重要作用.本研究旨在通过对高光谱数据的数学变换与特征分析,构建东洞庭湖苔草LAI的最佳估算模型.通过分析52组苔草样本640～780 nm波段范围内的反射率、一阶微分及倒数的对数的光谱特征,选择特征波段,运用多元逐步线性回归法与偏最小二乘回归法建立估算模型.研究发现特征波段为707～758nm,以上3种光谱数据的多元逐步线性回归模型决定系数分别为0.526、0.815、0.565,均方根误差分别为0.320、0.269、0.273,3者偏最小二乘回归模型的决定系数均高达0.9以上,均方根误差分别为0.189、0.262、0.134.结果表明:偏最小二乘回归法优于多元逐步线性回归法,该估算模型可有效估算苔草LAI.
展开▼
