潍坊北部地区土壤盐分光谱特征分析

Analysis on Spectral Characteristics of Soil Salinity in Northern Weifang City

摘要:

为提高海水入侵地区土壤盐分的监测效率，以潍坊北部地区作为研究区，利用高光谱测量技术野外实地测量，获得高光谱数据，并对该区域土壤盐分高光谱数据进行分析，获得土样盐分含量相关数据，与光谱数据结合分析。通过分析各类离子与全盐量之间的相关关系，分析潍北地区土壤盐分的离子构成与离子的相关来源。对高光谱原始测量数据进行包络线去除、变换，选取特征波段、一阶二阶微分变换等提高模型精度，建立高光谱数据与土壤盐分数据之间的逐步回归模型，并对模型进行精度检验，对比各类模型检验结果，分析各类离子对高光谱数据的响应效果。其中，光谱数据一阶微分变换后的逐步回归模型对土壤中K+&Na+估算精度较高，R2=0.894且均方根误差最小，精度较高，可为该地区的高光谱土壤盐分监测提供有力的数据分析结果。

Abstract:

In order to improve the monitoring efficiency of soil salinity in seawater intrusion area, setting northern region in Weifang city as the study area, by using hyperspectral measurement technology, hyperspectral data of soil salinity in the region has been gained and analyzed, and relevant data of soil salt content has also been gained. Through analysis on the relationship between various ions and total salt contents, ion composition and ion source of soil salinity in Northern Weifang have been analyzed. Through removing and transforming of original hyperspectral data, choosing characteristic bands, first and second order differential transformation, the accuracy of the model can be improved. Stepwise regression model between hyperspectral data and soil salinity data can be established, and the accuracy of the model has been checked, various types of model test results have been compared, and the response effect of various ions on hyperspectral data has been analyzed. Among them, by using regression model of first order differential transform, the estimation accuracy of K+ and Na+ in soil is high, R2=0.894, the error of root mean square is small, and the accuracy is high. It can provide useful data for hyperspectral monitoring of soil salinity.