引用本文:王琦,宋晓宇,杨贵军,李振海,冯海宽.区域冬小麦籽粒蛋白含量遥感预测研究[J].中国农业信息,2018,30(6):38-56
【打印本页】   【HTML】   【下载PDF全文】   查看/发表评论  【EndNote】   【RefMan】   【BibTex】
←前一篇|后一篇→ 过刊浏览    高级检索
本文已被:浏览 75次   下载 91 本文二维码信息
码上扫一扫!
分享到: 微信 更多
区域冬小麦籽粒蛋白含量遥感预测研究
王琦1,2,3,4,宋晓宇1,3,4,杨贵军1,3,4,李振海1,3,4,冯海宽1,3,4
1.农业部农业遥感机理与定量遥感重点实验室,北京农业信息技术研究中心 北京 100097;2.山东农业大学,泰安 271018;3.国家农业信息化工程技术研究中心,北京 100097;4.北京市农业物联网工程技术研究中心,北京 100097
摘要:
目的 籽粒蛋白含量是衡量小麦品质优劣的重要标准,快速准确预测小麦GPC有利于其品质评价和分级管理。方法 文章分别以卫星光谱参数、农学氮素参数以及气象因子为影响因素,并运用多元线性回归模型、极限学习机算法、地理加权回归模型3种方法实现对冬小麦GPC的预测,最终构建及评价基于不同自变量和不同方法的GPC预测模型。结果 (1)小麦开花期氮素参数,小麦冠层光谱参数与小麦籽粒蛋白品质的关系显著相关,影响小麦籽粒蛋白品质的关键性气象因子包括5月26—30日降雨和、5月中旬至6月上旬日照和、3月上旬至6月上旬积温和;(2)以卫星光谱参数、农学氮素参数和气象因子为自变量,分别采用多元线性回归、极限学习机和地理加权回归3种方法构建小麦GPC的预测模型;其中,基于多元线性回归模型构建的GPC模型决定系数R2为0.598,验证集标准均方根误差nRMSE和平均绝对误差MAE分别为10.36%、1.091,验证结果较稳定;基于ELM构建的GPC模型R2为0.483,验证nRMSE和MAE分别为10.895、1.111;基于GWR的GPC模型建模精度及验证精度相对最优,其建模R2为0.616,验证nRMSE及MAE分别为8.58%、0.956,为最优选择。结论 综合分析模型的精度评价指标可知,考虑空间数据不稳定性构建的地理加权回归模型的预测精度最好,能更加准确地预测冬小麦籽粒蛋白含量,为精确反演冬小麦GPC区域间和年际间的预测提供可靠依据,具有广泛的应用前景。
关键词:  遥感;模型;籽粒蛋白含量;极限学习机;地理加权回归;冬小麦
DOI:10.12105/j.issn.1672-0423.20180604
分类号:
基金项目:国家重点研发计划2016YFD0300603,2016YFD0700303国家重点研发计划(2016YFD0300603,2016YFD0700303),国家自然科学基金项目(41371349;41471285)
Remote sensing prediction of grain protein content in regional winter wheat
Wang Qi1,2,3,4,Song Xiaoyu1,3,4,Yang Guijun1,3,4,Li Zhenghai1,3,4,Feng Haikuan1,3,4
1.Key Laboratory of Quantitative Remote Sensing in Agriculture of Ministry of Agriculture,Beijing Research Center for Information Technology in Agriculture Beijing 100097,China;2.Shandong Agricultural University,Taian 271018,China;3.National Engineering Research Center for Information Technology in Agriculture,Beijing 100097,China;4.Beijing Engineering Research Center for Agriculture Internet of Things,Beijing 100097,China
Abstract:
Purpose Grain protein content (GPC) is an important indicator to evaluate the quality of wheat.Predicting the GPC quickly and accurately is beneficial to quality evaluation and hierarchical management.Method In this study,several factors that correlated with GPC evaluation,including meteorological factors in study area,wheat plant nitrogen parameters in flowering stage,and satellite spectral parameters for wheat samples,were analyzed using multi-linear regression (MLR),extreme learning machine algorithm(ELM) and geographical weighted regression (GWR) methods.Then the GPC prediction models based on different independent variables and methods were built and evaluated.Result The results showed that:(1) The nitrogen parameters at wheat flowering stage and the wheat canopy spectral parameters have significant correlations with wheat GPC.The key meteorological factors affecting the GPC of wheat include rainfall from 26 May to 30 May,sunshine time from mid-May to early June and accumulated temperature from early March to early June. (2) The coefficient of determination (R2) of MLR model is 0.598,while the accuracies of Normalized root mean squared error (nRMSE) and Mean absolute error (MAE) are 10.36% and 1.091 respectively,which are stable.The R2 of the GPC model based on ELM is 0.483,while the standard nRMSE and MAE are 10.895% and 1.111 respectively.The R2 of the GPC GWR model is 0.616,while the standard nRMSE and MAE are 8.58% and 0.956 respectively,which are the optimal.Conclusion According to the precision evaluation indicator of the comprehensive analysis model,the multivariate parameter model is superior to the univariate parameter model.The multivariate parameter GWR model that takes the instability of spatial data into account has the best prediction accuracy and can predict the GPC more accurately.This study provides a reliable basis for accurately predicting the GPC in different regions and years,which has a broad application prospect in the future.
Key words:  remote sensing;models;grain protein content;extreme learning machine;geographically weighted regression model;winter wheat