亚热带典型区域水稻土氧化铁高光谱反演——以珠江三角洲为例

发布时间：2018-05-14 14:30

  本文选题：土壤氧化铁 + 遥感　； 参考：《应用生态学报》2017年11期

【摘要】：氧化铁是土壤中铁元素的主要存在形式,亚热带土壤中高含量的氧化铁形成了该区域重要的土壤附色成分针铁矿和赤铁矿等矿物,使得土壤颜色明显区别于其他气候带.以亚热带典型地区珠江三角洲为例,分析不同光谱形式与土壤氧化铁含量的相关性,提取特征光谱波段建立土壤氧化铁的反演模型.结果表明:土壤氧化铁含量与反射光谱之间呈负相关,且敏感波段主要位于404、574、784、854和1204 nm等可见近红外区域.微分处理后的光谱与土壤氧化铁的相关性明显提高.在相关性显著波段的基础上采用逐步多元线性回归以及主成分分析剔除共线性波段,最后选择特征光谱波段作为反演模型的输入参数.比较反演模型的结果,得出该地区土壤氧化铁含量的最佳反演模型为BP人工神经网络(RMSEC=0.22,RMSEP=0.81,R~2=0.93,RPD=12.20),该模型具有非常好的稳定性,适用于快速估测土壤中氧化铁含量,并且能够为测度土壤氧化铁的空间分布提供研究基础.
[Abstract]:Iron oxide is the main form of iron element in soil. The high content of iron oxide in subtropical soil forms important minerals such as goethite and hematite, which make the soil color distinct from other climatic zones. Taking the Pearl River Delta in the subtropical region as an example, the correlation between the different spectral forms and the content of iron oxide in soil was analyzed, and the inversion model of soil iron oxide was established by extracting the characteristic spectral bands. The results showed that there was a negative correlation between the ferric oxide content and the reflectance spectrum, and the sensitive bands were mainly located in the visible near infrared regions such as 404574784854 nm and 1204 nm. The correlation between the spectra after differential treatment and the soil iron oxide is obviously improved. On the basis of significant correlation band, stepwise multivariate linear regression and principal component analysis (PCA) were used to eliminate the collinear band. Finally, the characteristic spectral band was selected as the input parameter of the inversion model. By comparing the results of the inversion model, it is concluded that the best inversion model of the soil iron oxide content in this area is BP artificial neural network RMSEC 0.22 ~ 0.22 ~ (0.21) RMSEP0.91 ~ 0.93 RPD-12.200.The model has very good stability and is suitable for rapid estimation of iron oxide content in soil. And it can provide the research basis for measuring the spatial distribution of soil iron oxide.
【作者单位】： 山西农业大学资源环境学院;广东生态环境技术研究所/广东省农业环境综合治理重点实验室;广州地理研究所;中国科学院地球环境研究所;
【基金】：广东省科技计划项目(2015B070701017,2017A040406021) 国家自然科学青年科学基金项目(41601558) 广州市科技计划项目(201709010010) 广东省科学院创新平台建设专项资助~~
【分类号】：S153
，

本文编号：1888268