首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于高光谱的核桃叶片铁元素含量反演模型研究
引用本文:瞿余红,王振锡,丁雅,刘玉霞,董淼,李园.基于高光谱的核桃叶片铁元素含量反演模型研究[J].新疆农业科学,2018,55(7):1264-1273.
作者姓名:瞿余红  王振锡  丁雅  刘玉霞  董淼  李园
作者单位:新疆农业大学林学与园艺学院/新疆教育厅干旱区林业生态与产业技术重点实验室,乌鲁木齐 830052
基金项目:新疆维吾尔自治区高校科研计划项目(XJEDU2017M013);中国博士后科学基金项目(2015 M572668 XB)
摘    要:【目的】铁元素是评价植物长势的重要指标,监测营养状况是果树栽培管理过程中的主要指标之一。分析核桃叶片Fe素含量与叶片光谱之间的关系,监测核桃活体叶片Fe素含量,为核桃叶片Fe含量快速诊断提供理论依据和技术支撑。【方法】对阿克苏红旗坡地区8个核桃生产园进行随机采样,测定其叶片光谱反射率及Fe素含量,对经过数据变换后的光谱采用逐步线性回归和主成分回归两种方法构建Fe素含量估测模型。用随机分组法选取检验样本并进行精度评价。【结果】经过数据转换的光谱与Fe素含量的相关性明显高于原始光谱,其中以对数二阶光谱725~730 nm、1 140、1 141、1 443、1 444和1 445 nm作为敏感波段构建的主成分回归模型精度最佳,构建模型为y=558.648+113.024 PC13-31.514 PC12+71.519 PC1,R2为0.870,RMSE为0.304,RPO为2.39。【结论】光谱技术在开展阿克苏红旗坡地区核桃叶片Fe素含量估算方面具有一定的应用潜力。

关 键 词:核桃  光谱反射率  主成分回归  逐步回归  
收稿时间:2018-05-21

Research on the Inversion Model for Predicting Fe Content in Juglans regia L. Leaves Based on Hyperspectral Method
QU Yu-hong,WANG Zhen-xi,DING Ya,LIU Yu-xia,DONG Miao,LI Yuan.Research on the Inversion Model for Predicting Fe Content in Juglans regia L. Leaves Based on Hyperspectral Method[J].Xinjiang Agricultural Sciences,2018,55(7):1264-1273.
Authors:QU Yu-hong  WANG Zhen-xi  DING Ya  LIU Yu-xia  DONG Miao  LI Yuan
Institution:Collage of Forestry and Horticulture, Xinjiang Agricultural University / Key Laboratory of Forestry Ecology and Industry Technology in Arid Region, Education Department of Xinjiang, Urumqi 830052, China
Abstract:【Objective】 Fe element is an important index to evaluate plant growth, and monitoring nutritional status is one of the main contents in the process of fruit tree cultivation and management. This project aims to realize monitoring of foliar Fe content in Juglans regia L.in vivo by analyzing the relationship between foliar Fe content in Juglans regia L. and foliar spectra in the hope of providing theoretical basis and technical support for rapid diagnosis of Fe content.【Method】Eight Juglans regia L. production garden were selected in Aksu Hongqipo area for random sampling, the spectral reflectance and Fe content of the foliar were measured. The method of stepwise linear regression and principal component regression was used to construct the estimation model of Fe content. Random grouping method was used to select test samples and evaluate the accuracy.【Result】The correlation between the content of Fe in Juglans regia L. foliar and spectra transformed by differential, logarithm and so on was significantly higher than that of the original spectrum. In these data conversions, the highest accuracy of the model was principal component regression model, which was based on the logarithmic second-order spectra of 725-730 nm, 1,140 nm, 1,141 nm, 1,443 nm, 1,444 nm and 1,445 nm. So the optimal model for monitoring foliar Fe element of Juglans regia L. was y=558.648+113.024 PC13-31.514 PC12 +71.519 PC1, R2 was 0.870, RMSE was 0.304, RPO was 2.39.【Conclusion】Spectral technique has a certain application potential in estimating the Fe content of Juglans regia L. in Hongqipo area of Aksu.
Keywords:Juglans regia L    spectral reflectance  principal component regression  gradual regression  
本文献已被 CNKI 等数据库收录!
点击此处可从《新疆农业科学》浏览原始摘要信息
点击此处可从《新疆农业科学》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号