欢迎访问中国科学院大学学报,今天是

中国科学院大学学报 ›› 2011, Vol. 28 ›› Issue (2): 187-194.DOI: 10.7523/j.issn.2095-6134.2011.2.008

• 论文 • 上一篇    下一篇

松嫩平原土壤有机质含量高光谱反演研究

武彦清1,2, 张柏1, 宋开山1, 刘焕军3, 王宗明1, 刘殿伟1   

  1. 1. 中国科学院东北地理与农业生态研究所, 长春 130012;
    2. 中国科学院研究生院, 北京 100049;
    3. 东北农业大学, 哈尔滨 150030
  • 收稿日期:2010-04-26 修回日期:2010-05-31 发布日期:2011-03-15
  • 基金资助:

    中国科学院知识创新工程重要方向项目(KZCX2-YW-340)资助 

Retrieval of soil organic matter content from hyper-spectra in Songnen Plain

WU Yan-Qing1,2, ZHANG Bai1, SONG Kai-Shang1, LIU Huan-Jun3, WANG Zong-Ming1, LIU Dian-Wei1   

  1. 1. The Northeast Institute of Geography and Agricultural Ecology, Chinese Academy of Sciences, Changchun 130012, China;
    2. Graduate University, Chinese Academy of Sciences, Beijing 100049, China;
    3. Northeast Agricultural University, Harbin 150030, China
  • Received:2010-04-26 Revised:2010-05-31 Published:2011-03-15

摘要:

采集松嫩平原典型类型土壤样本252个.室内条件下测定了风干后土样的光谱.对原始光谱数据作预处理后,分别运用多元线性逐步回归法和偏最小二乘法建立土壤有机质含量高光谱预测模型.结果表明,采用2种方法建立的模型均可满足有机质含量速测要求,但偏最小二乘法得到的模型更具稳健性.

关键词: 高光谱, 有机质, 多元线性逐步回归, 偏最小二乘回归

Abstract:

A total of 252 soil samples were collected in Songnen Plain. Reflectance was measured in a controlled laboratory environment. After pre-processing of the primitive spectra, hyper-spectral models for predicting soil organic matter content were built up by using the methods of Stepwise Multiple Linear Regression (SMLR) and Partial Least Squares Regression(PLSR). The results show that the models using the two methods are capable of predicting SOM content and the model using PLSR is more robust.

Key words: hyper-spectral, soil organic matter, stepwise multiple linear regression, partial least squares regression

中图分类号: