025-83701931 设为首页 加入收藏
文章摘要
基于地理加权回归模型的葡萄园土壤砷含量高光谱反演
Hyperspectral Inversion of Arsenic Content in Vineyard Soil Based on Geographically Weighted Regression Model
  
DOI:
中文关键词:   高光谱反演  地理加权回归模型  葡萄园土壤
英文关键词: Arsenic  Hyperspectral inversion  Geographically weighted regression model  Vineyard soil
基金项目:国家自然科学基金资助项目(41867067)
作者单位
阿依努尔·麦提努日 新疆师范大学地理科学与旅游学院 
麦麦提吐尔逊·艾则孜 新疆师范大学地理科学与旅游学院新疆大学化学与化工学院 
李新国 新疆师范大学地理科学与旅游学院 
摘要点击次数: 68
全文下载次数: 0
中文摘要:
      以吐鲁番盆地葡萄园土壤中砷(As)为研究对象,分析15种光谱变换下的土壤光谱反射率数据与土壤As含量的相关性,构建土壤As含量预测的偏最小二乘回归(PLSR)模型和地理加权回归(GWR)模型。结果表明:葡萄园土壤原始光谱率(R)经一阶微分(FD)、平方根一阶微分(SRFD)、平方根二阶微分(SRSD)、倒数二阶微分(RTSD)、对数一阶微分(LTFD)、倒对数一阶微分(ATFD)变换对As光谱特征的增强作用最突出。模型预测结果表明,采用基于LTFD变换的GWR模型可有效提高葡萄园土壤As含量的预测精度。
英文摘要:
      Taking arsenic(As) in vineyard soil in Turpan Basin as the research object, the correlation between soil spectral reflectance data and As content in soil was analyzed under 15 spectral transformations, partial least squares regression(PLSR) model and geography weighted regression(GWR) model for As content prediction were established. The results showed that the original spectral rate of vineyard soil transformed by the first order differential(FD), square root first order differential(SRFD), square root second order differential(SRSD), reciprocal second order differential(RTSD), logarithmic first order differential(LTFD) and inverse logarithmic first order differential(ATFD) had a significant enhancement effect on the spectral characteristics of As. According to model prediction, GWR model based on LTFD transformation could effectively improve the prediction accuracy of As content in vineyard soil.
查看全文   查看/发表评论  下载PDF阅读器
关闭