基于光谱反射特征的柑橘叶片含水率模型

Model for Citrus Leaves Water Content Based on Spectral Signature

水分是作物生长发育的主要限制因子之一。为了明确柑橘叶片含水率与光谱之间的定量关系,为果树遥感监测提供基础,本研究采集了46个柑橘叶片信息,其中23片用于模型构建,23片用于模型验证,采用光谱逐步回归分析和构造光谱指数2种方法分析了叶片光谱(380～2500nm)反射率与含水率之间的关系并构建柑橘叶片含水率—光谱反射率模型。结果表明,柑橘叶片含水率与反射光谱之间有较强的相关性。2种方法所得模型预测值与实测值的相对误差都小于10%,说明模型具有较好的预测效果,而且光谱逐步回归分析所得模型精度高于后者,但从物理含义和逻辑性方面考虑,推荐光谱指数法构建的模型。

The moisture content is one of main factors which limit the growth and development of crops.In order to identify the quantitative relation between leaf water content and spectra,also provide a basis for remote sensing monitoring of fruits,this experiment collected data of 46 citrus leaves,half of which were for establishing model,and other half of which were for verifying model.The paper studied the relationship between leaf water content and leaf reflectance,and established models of leaves water content based on spectra with stepwise regression method and constructing spectral indexes method respectively.The results showed there was good relevance between water content and reflection spectrum of Citrus leaves.The relative errors between predictive and observed models with two methods were all less than 10%.It made clear that models obtained better forecasting results,and the model established with stepwise regression method had higher precision than the model with another method.Considering the physical meaning and logic,the model established with constructing spectral indexes method was recommended.