艾比湖流域盐渍土含水量光谱特征分析与建模

发布时间：2018-07-09 18:34

本文选题：光谱特征 + 含水量　；参考：《中国水土保持科学》2017年01期

【摘要】：盐渍土表层盐分累积与土壤含水量有着直接的关系。为了建立干旱区盐渍土含水量高光谱遥感监测模型,本文以艾比湖流域不同含水量的盐渍土为研究对象,采用光谱反射率变换和多元统计分析(MSAM)方法,对土壤含水量的光谱特征进行分析和建模。结果表明:随着土壤含水量的增加,土壤反射率呈下降趋势;在一定范围内,波长越长,土壤光谱反射率与含水量的相关性越高,其中1 937 nm处的土壤光谱反射率与含水量具有最高的相关性(r=-0.636)。对土壤光谱反射率进行8种光谱反射率变换后,在此基础上利用多元统计方法分析盐渍土的不同含水量与光谱之间的相关性,筛选敏感波段,建立关系模型。得出对数一阶微分(Logarithm First Order Differential)在波长2 024和2 357 nm建立的模型以及均方根一阶微分(Root Mean Square First Order Differential)在波长1 972和2 357 nm建立的模型最优,相关系数r分别为0.894和0.865。基于上述模型作者构建了一种耦合模型,其相关性r=0.926比对数一阶微分模型提高了0.032,比均方根一阶模型提高了0.061;因此,所构建的盐渍土含水量估算模型是可行的,可以为遥感反演提供理论参考,对高光谱遥感反演具有一定意义。
[Abstract]:The accumulation of salt in the surface layer of saline soil is directly related to the soil moisture content. In order to establish a hyperspectral remote sensing monitoring model for water content of saline soil in arid area, this paper takes the saline soil with different water content in Ebinur Lake Basin as the research object, adopts spectral reflectance transform and multivariate statistical analysis (MSAM) method. The spectral characteristics of soil moisture content were analyzed and modeled. The results showed that the soil reflectance decreased with the increase of soil moisture content, and the correlation between soil spectral reflectance and water content increased with the increase of wavelength in a certain range. There was the highest correlation between soil spectral reflectance and water content at 1 937 nm (r-0.636). On the basis of eight spectral reflectance transformation of soil spectral reflectance, the correlation between different water content and spectrum of salinized soil was analyzed by multivariate statistical method, sensitive bands were screened, and a relational model was established. The model of logarithm first order differential at 2 024 nm and 2 357 nm and the model of Root mean Square first order differential at wavelength 1 972 and 2 357 nm are obtained. The correlation coefficients are 0.894 and 0.865, respectively. Based on the above model, a coupling model is constructed, whose correlation rn 0.926 is 0.032 higher than that of the first order differential model and 0.061 higher than that of the first order RMS model, so it is feasible to estimate the water content of salinized soil. It can provide theoretical reference for remote sensing inversion and has certain significance for hyperspectral remote sensing inversion.
【作者单位】：新疆大学资源与环境科学学院;新疆大学绿洲生态教育部重点实验室;新疆智慧城市与环境建模普通高校重点实验室;
【基金】：国家自然科学基金“新疆联合基金本地优秀青年人才培养专项”(U1503302)
【分类号】：S127;S156.41

【相似文献】