枣叶片N素质量分数高光谱估算模型  

Hyperspectral Estimation Model for Nitrogen Content of Jujube Leaf

作　　者：丁雅 王振锡[1] 瞿余红 刘玉霞 李园 Ding Ya;Wang Zhenxi;Qu Yuhong;Liu Yuxia;Li Yuan(Xinjiang Agricultural University,Urumqi 830052,P.R.China)

机构地区：[1]新疆农业大学,乌鲁木齐830052

出　　处：《东北林业大学学报》2018年第11期45-50,共6页Journal of Northeast Forestry University

基　　金：新疆维吾尔自治区高校科研计划项目(XJEDU2017M013);中国博士后科学基金项目(2015M572668XB)。

摘　　要：以新疆阿克苏地区红旗坡农场规模种植枣叶片为研究对象,通过分析枣叶片光谱反射特征,采用相关分析和逐步回归方法构建枣叶片氮素质量分数的光谱估测模型。结果表明:原始光谱的一阶微分和二阶微分变换能够显著提高与枣叶片N素质量分数的相关性,相关系数均达到0.6以上,光谱参量(B890+B970) nm与氮素质量分数构建的Y=0.791e^(-3 953.718X)指数模型对枣叶片氮素质量分数的反演效果最优,估测模型决定系数(A^2)值为0.576,验证模型拟合度(r^2)为0.847,估测精度为99.37%。该研究建立的光谱指数模型对枣叶片氮素质量分数的反演是可行的,为新疆阿克苏地区枣树营养状况及长势等监测、评价提供技术途径和理论依据。With the scale planted Jujube at Hongqipo Farm at Aksu of Xinjiang,we studied Jujube leaves by analyzing its reflection spectrum features to construct a spectrum-estimating model which estimates the mass ratio of nitrogen(N),by correlation analysis and stepwise regression analysis.The relation parameter between N mass ratio and spectral region,was significantly enhanced by first and second derivative of the raw spectrum,and reached 0.6 even higher.The equation,Y=0.791 e-3 953.718 X,is the best reversion model to describe the spectrum parameter(B 890+B 970)nm and the N mass ratio of Jujube leaves.The fitting degree(A 2)of estimation model is 0.576 with the verification fitting degree of 0.847 and the accuracy of 99.37%.Therefore,using this model to estimate N mass ratio is feasible to monitor and estimate nutrition situation of Jujube in Aksu of Xinjiang.

关 键 词：高光谱 枣叶片 氮素质量分数 估测模型 精度检验 

分 类 号：S771.8[农业科学—森林工程]