摘要
生物量是反映大田小麦生长的重要指标之一,孕穗期是小麦关键生长时期。为准确获取小麦孕穗期生物量,利用HJ-A数据,结合偏最小二乘法(PLS)和主成分分析(PCA)分别建立及评价小麦孕穗期生物量的多变量预估模型。结果表明:利用HJ-A数据获取小麦孕穗期生物量可行,且由HJ-A数据提取的植被指数之间存在显著的多重相关性。该PLS模型和PCA模型的最优主成分数分别为5和3个,并利用降维分析法可消除植被指数之间的多重共线性。利用3种样本数据进行评价,该PLS模型的决定系数(R^(2))均大于0.70,均方根误差(RMSE)分别为326.95、341.27和373.35 kg·hm^(-2);PCA模型的R^(2)均在0.60左右,RMSE分别为439.50、451.32和486.53 kg·hm^(-2)。PLS模型精度约为90%,比PCA模型和线性模型分别提高12%、22%以上。综上所述,综合HJ-A数据和PLS建立的小麦孕穗期生物量预估模型应用效果好。
Biomass is one of the vital group indicators reflecting field-grown wheat growth.Booting stage is the key growth stage of wheat.To accurately extract wheat biomass at booting stage by HJ-A data,this study established and validated the estimation model of wheat biomass based on partial least squares(PLS)and principal component analysis(PCA),respectively.The results showed that it was feasible to extract wheat biomass at booting stage using HJ-A data,and there were significant multiple correlations between vegetation indices.The optimal principal component number of PLS model and PCA model were 5 and 3,respectively,and the multi-collinearity between vegetation indices was eliminated by dimension reduction analysis method.Using three different sample data for evaluation,the determination coefficient(R^(2))of the PLS model was greater than 0.70,and the root mean square error(RMSE)were 326.95,341.27 and 373.35 kg·hm^(-2) respectively;The R^(2) of PCA model was about 0.60,and the RMSE were 439.50,451.32 and 6.53 kg·hm^(-2)espectively.The accuracy of PLS model was about 90%,which was more than 12%and 22%higher than PCA model and linear model respectively.Therefore,it was feasible to establishthe estimation model of wheat biomass at booting stage by using HJ-A data and PLS.
作者
马顺圣
毛伟
周欣兴
张鹏鹏
李文西
MA Shunsheng;MAO Wei;ZHOU Xinxing;ZHANG Pengpeng;LI Wenxi(Collaborative Innovation Center for Geographic Information Collection,Processing and Application of Ministry of Education/Jiangsu Safety&Environment Technology and Equipment for Planting and Breeding Industry Engineering Research Center,Yangzhou Polytechnic College,Yangzhou 225009,China;Station of Land Protection of Yangzhou City,Yangzhou 225001,China;College of Agriculture,Yangzhou University,Yangzhou 225009,China)
出处
《扬州大学学报（农业与生命科学版）》
CAS
北大核心
2021年第6期29-35,共7页
Journal of Yangzhou University：Agricultural and Life Science Edition
基金
国家重点研发计划项目子课题(2016YFD0201303)
国家自然科学基金资助项目(32071902)
江苏省重点研发计划项目(BE2018362)
江苏省高等学校优秀科技创新团队项目[苏教科(2021)1号]。
