基于小波变换的近红外光谱预测土壤有机质  

Prediction of Soil Organic Matter with Near-infrared Spectroscopy Treated by Wavelet Transformation

作　　者：杨苗[1] 左月明[2] 杨萍果[3] 

机构地区：[1]山西农业大学文理学院,山西太谷030801 [2]山西农业大学工程技术学院,山西太谷030801 [3]山西师范大学生命科学院,山西临汾041000

出　　处：《山西农业大学学报（自然科学版）》2010年第2期154-158,共5页Journal of Shanxi Agricultural University(Natural Science Edition)

基　　金：山西农业大学科技创新基金(2009011);山西省基础研究项目(2010011044-3)

摘　　要：为了探索快速检测土壤有机质含量的方法,试验采用不同分解水平的Coiflet函数的小波（wavelet）分析方法,对山西关帝山土壤样品的近红外光谱信号进行了消噪处理,来快速获取土壤中有机质含量。结果表明：对有机质敏感波段的为450-600 nm,810-935 nm,1 030-1 315 nm,1 380-1 400 nm;有机质NIRS法与实验室标准法测定值之间的相关系数R2为0.9818;说明通过小波变换滤波,选择敏感波段,用偏最小二乘回归方法预测土壤有机质含量是可行的。In order to explore the rapid detection method of soil organic matter content,this paper adopts a different decomposition levels of wavelet Coiflet function（wavelet） analysis method.Soil samples from Shanxi Guandi Mountain were carried out near-infrared spectroscopy signal de-noising processing,to quickly obtain the soil organic matter content.The results showed that sensitive bands to the organic matter were 450 nm-600 nm,810 nm-935 nm,1 030 nm-1 315 nm and 1 380 nm-1 400 nm;The correlation coefficient between the value of organic matter NIRS method and the standard laboratory method was 0.9818;Through wavelet filtering,sensitive band selection employing partial least-squares regression method to predict soil organic matter content is feasible.

关 键 词：小波变换 土壤光谱 有机质 偏最小二乘回归 

分 类 号：S159[农业科学—土壤学]