基于小波变换的土壤重金属铬含量定量反演  

Quantitative inversion of heavy metal chromium content in soil based on wavelet transform

作　　者：郭云开[1,2] 谢晓峰 谢琼 张思爱[1,2] GUO Yunkai;XIE Xiaofeng;XIE Qiong;ZHANG Siai(School of Traffic and Transportation Engineering, Changsha University of Science and Technology, Changsha 410076, China;School of Surveying and Mapping Remote Sensing Applied Technology, Changsha University of Science and Technology, Changsha 410076, China;Department of Surveying and Mapping Geography, Hunan Vocational College of Engineering, Changsha 410151, China;Engineering Laboratory of Spatial Information Technology of Highway Geological Disaster Early Warning, Changsha University of Science and Technology, Changsha 410114, China)

机构地区：[1]长沙理工大学交通运输工程学院,湖南长沙410076 [2]长沙理工大学测绘遥感应用技术研究所,湖南长沙410076 [3]湖南工程职业技术学院测绘地理学院,湖南长沙410151 [4]长沙理工大学公路地质灾害预警空间信息技术湖南省工程实验室,湖南长沙410114

出　　处：《测绘工程》2021年第3期66-71,共6页Engineering of Surveying and Mapping

基　　金：国家自然科学基金资助项目(41471421;41671498);公路地质灾害预警空间信息技术湖南省工程实验室开放基金资助项目(kfj190603)。

摘　　要：耕地污染日益严重,耕地土壤的重金属高光谱信息属于非线性的微弱信号。小波变换作为常用的非线性微弱信号处理手段,在保留更多微弱信号的基础上可以更好的提取出土壤重金属的微弱光谱信息。文中研究在Db4小波对土壤原始光谱进行分解与重构的基础上提取特征波段,利用特征波段与重金属含量的相关性建立偏最小二乘模型反演土壤重金属铬含量。研究表明,利用Db4小波函数对原始光谱进行分解和重构可以有效提取土壤重金属铬的特征光谱信息;利用小波分解与重构所提取的特征光谱信息与重金属铬含量之间的相关性所建立的PLS模型的决定系数明显高于基于传统一阶微分处理土壤光谱所建立的PLS模型的决定系数。The heavy metal hyper-spectral information of cultivated soil belongs to nonlinear weak signal.As a common nonlinear weak signal processing method,the wavelet transform can better extract the weak spectral information of soil heavy metals on the basis of retaining more weak signals.In this paper,based on the decomposition and reconstruction of the original spectrum of soil by DB4 wavelet,the feature band is extracted,and the partial least square model is established to retrieve the content of heavy metal chromium in soil by using the correlation between the feature band and heavy metal content.The result shows that the characteristic spectrum information of heavy metal chromium in soil can be extracted effectively by decomposing and reconstructing the original spectrum with DB4 wavelet function.The decision coefficient of PLS model based on the correlation between the feature spectrum information extracted by wavelet decomposition and reconstruction and the content of heavy metal chromium is significantly higher than that of PLS model based on traditional first-order differential treatment of soil spectrum.

关 键 词：高光谱遥感 土壤重金属 小波分解与重构 偏最小二乘 

分 类 号：X87[环境科学与工程—环境工程]