基于近地高光谱的土壤氯离子含量估测  

Estimating Content of Soil Chloride Based on Hyperspectral Data

作　　者：王明宽 莫宏伟 陈红艳[2] 

机构地区：[1]湖南科技大学资源环境与安全工程学院地理科学系,湖南湘潭411201 [2]山东农业大学,山东泰安271000

出　　处：《水土保持通报》2017年第6期214-219,共6页Bulletin of Soil and Water Conservation

基　　金：湖南省教育厅科学研究项目"湘江流域土地利用变化及土壤潜力转移图谱研究"(14C0467);地理空间信息技术国家地方联合工程实验室开放基金项目(2014GISNELJJ004)

摘　　要：[目的]建立土壤氯离子与高光谱波段的多元线性回归模型,获取盐渍化信息,为盐分的高精度提取提供更有效的方法,为农业生态环境重建工作提供科学依据。[方法]以山东省垦利县作为研究区,于2014年10月5—7日野外采集93个土壤样本,利用ASD高光谱仪野外采集土样高光谱数据并进行预处理,然后采用多元回归和主成分分析方法建立估测氯离子含量的高光谱模型,以快速估测氯离子含量。[结果]氯离子在近红外749,830,987,1301,1 432,1 486nm较为敏感,在土壤光谱分析的基础上,得到室内风干土壤氯离子含量预测最优模型,模型均通过T检验和F检验,能较好地预测土壤氯离子含量。[结论]研究区土壤各组分盐离子中,阳离子以钠离子为主,阴离子以氯离子为主,该模型使间接得到土壤盐分含量具有较好可行性。[Objective]Constructing a multiple linear regression model of describing soil chloride ion with high spectral bands as independent variables to get soil salinization information,in order to provide a more effective method for the high precision extraction of salt,and to provide scientific basis for the reconstruction of agricultural ecological environment.[Methods]93 field soil samples were collected and processed by high ASD spectrometer in Kenli County of Shandong Province on October 5 to 7,2014.A estimating model was built using multiple regression and principal component analysis method to evaluating the content of chloride ion quickly.[Results]The chloride ion is sensitive in the 749,830,987,1 301,1 432 and 1 486 nm of spectral bands.An optimum model for predicting chlorine ion content in indoor dried soil was obtained on the basis of soil spectral analysis,this model was verified by Student＇s t test.[Conclusion]The cations in the soil components of the study area are mainly sodium ions,the anions are mainly chloride ions,and this model makes it possible to obtain soil salinity indirectly.

关 键 词：氯离子 高光谱 反演 多元线性回归 主成分分析 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置] S153.6[自动化与计算机技术—控制科学与工程]