基于植被指数与连续小波变换的玉米叶片Cu^(2+)含量反演  

Inversion of Cu^(2+) content in corn leaves based on vegetation index and continuous wavelet transform

作　　者：郭辉 戴志林[2,3] 石海 GUO Hui;DAI Zhilin;SHI Hai(Key Laboratory of Mine Spatial Information Technology,State Administration of Surveying,Mapping and Geographic Information,Jiaozuo 454003,China;Henan Provincial Key Laboratory of Mine Spatial Information Technology,Jiaozuo 454003,China;School of Spatial Information and Surveying Engineering,Anhui University of Science and Technology,Huainan 232001,China)

机构地区：[1]矿山空间信息技术国家测绘地理信息局重点实验室,河南焦作454003 [2]矿山空间信息技术河南省重点实验室,河南焦作454003 [3]安徽理工大学空间信息与测绘工程学院,安徽淮南232001

出　　处：《安徽科技学院学报》2024年第1期24-31,共8页Journal of Anhui Science and Technology University

基　　金：矿山空间信息技术国家测绘地理信息局重点实验室(KLM201801)。

摘　　要：目的:确定Cu^(2+)污染胁迫下玉米叶片Cu^(2+)含量最优反演模型。方法:以室内盆栽玉米为研究对象,在采集不同胁迫梯度下玉米叶片光谱以及同期叶片Cu^(2+)含量的基础上,通过遍历计算出两波段原始光谱植被指数,并将其与叶片Cu^(2+)含量进行相关性分析;利用0.1~0.9阶、1.1~1.9阶与1~4阶共22种光谱微分预处理重采样后的光谱数据,经连续小波变换后分析小波系数与叶片Cu^(2+)含量之间的相关性;根据相关性分析提取最优植被指数与最优小波系数,建立反演模型。结果:植被指数与叶片Cu^(2+)含量显著相关,最优波段组合分别为:DI(621.5 nm, 1 889.2 nm)、RI(482.2 nm, 1 418.5 nm)、NDVI(666.3 nm, 1 917.2 nm)、RNDVI(621.5 nm, 1 889.2 nm),其光谱特征均集中在可见光与近红外波段附近;小波系数也与叶片Cu^(2+)含量之间具有良好的相关性,其敏感波段位于400、600、900、1 200、2 400 nm附近,与最优植被指数敏感波段一致。经0.9阶光谱微分预处理后得到的小波系数与叶片Cu^(2+)含量相关系数最大,为0.88;通过相关性分析提取最优植被指数和最优小波系数,以植被指数与不同阶微分变换的连续小波变换提取的小波系数为自变量,建立线性反演模型,其中利用最优植被指数建立的反演模型精度最高,模型最稳定,RMSE为4.97μg/g。结论:植被指数和连续小波变换两种方法在农作物重金属污染监测方面具有重要的参考价值,应用前景广阔。Objective:To determine the optimal inversion model of Cu^(2+)content in corn leaves under Cu^(2+)pollution stress.Methods:Indoor potted corn was taken as the research object.On the basis of collecting spectra of corn leaves under different stress gradients and Cu^(2+)content in corn leaves at the same period,and the original spectral vegetation index of two bands was traversed and the correlation between the vegetation index and Cu^(2+)content in leaves was analyzed.Twenty-two kinds of spectral differential pretreated resampling spectral data of order 0.1-0.9,order 1.1-1.9 and order 1-4 were used to analyze the correlation between the wavelet coefficients and Cu^(2+)content in leaves by continuous wavelet transform.Based on correlation analysis,the optimal vegetation index and wavelet coefficients were extracted and the inversion model was established.Results:There was a significant correlation between vegetation index and Cu^(2+)content in leaves.The optimal band combinations with the spectral features of DI(621.5 nm,1889.2 nm),RI(482.2 nm,1418.5 nm),NDVI(666.3 nm,1917.2 nm),RNDVI(621.5 nm,1889.2 nm)were concentrated in the visible and near infrared bands.The wavelet coefficient also had a good correlation with Cu^(2+)content in leaves,and its sensitive bands were located around 400,600,900,1200,2400 nm,which was consistent with the sensitive bands of optimal vegetation index.The correlation coefficient between the Cu^(2+)content and the wavelet coefficients obtained by the 0.9 order spectral differentiation pretreatment was 0.88 at most.The optimal vegetation index and the optimal wavelet coefficients were extracted by correlation analysis,and the wavelet coefficients extracted by the vegetation index and the continuous wavelet coefficients of different differential transforms were used as independent variables to establish a linear inversion model.The inversion model established by using the optimal vegetation index had the highest accuracy and the most stable model,with a RMSE of 4.97μg/g.Conclusion:The resu

关 键 词：高光谱遥感 铜污染胁迫 植被指数 连续小波变换 反演模型 

分 类 号：X87[环境科学与工程—环境工程]