中国农业科学2017，Vol.50Issue(5)：871-880,10.DOI:10.3864/j.issn.0578-1752.2017.05.010
基于高光谱遥感的冬小麦叶水势估算模型
Leaf Water Potential Estimating Models of Winter Wheat Based on Hyperspectral Remote Sensing
摘要
Abstract
[Objective] A model for fast,non-destructive and accurately monitoring leaf water potential of winter wheat was established with hyperspectra technology,it will provide a scientific basis for the precision irrigation management of winter wheat.[Method] Using the field trials of different water treatments,the canopy spectral reflectance,leaf water potential and soil moisture were synchronously determined in the growth period of winter wheat.Then the correlation between the hyperspectral vegetation indices and leaf water potential was analyzed.Using correlation analysis,regression analysis and other methods,four inversion models were constructed for estimating leaf water potential based on different water treatments.[Result] The canopy spectral reflectance of winter wheat had significant change characteristics in different water treatments and growth periods.In the visible wave band,the canopy reflectance of winter wheat was reduced gradually along with the increase of soil water content.But,in the near-infrared wave band,the canopy reflectance was increased with the increase of soil water content.With the development of wheat growth period,in the near-infrared wave band,the canopy reflectance at heading stage was higher than the reflectance at jointing stage.And after the filling stage,the reflectance of red and blue band was rose faster.The correlation between four vegetation indices and leaf water potential was all reached the significant level (P＜0.05),and its absolute values of correlation coefficient were all above 0.711.Four vegetation indices could be used for quantitative monitoring leaf water potential of winter wheat.Under the field capacity of 70％,the absolute values of correlation coefficient |r| between the vegetation indexes of OSAVI and EVI and the leaf water potential were 0.75 and 0.771,respectively,they were lower than the |r| between the vegetation indexes of NDVI and RVI and leaf water potential,which the values of |r| were 0.808 and 0.896,respectively.But,under the field capacity of 50％,the results were just the opposite.The |r| between the vegetation indexes of OSAVI and EVI2 and the leaf water potential were 0.857 and 0.853,respectively,which were higher than the |r| between the vegetation indexes of NDVI and RVI and the leaf water potential,which the values of |r| were 0.711 and 0.792,respectively.The estimation values of 45 samples in prediction set were close to the measured values,the range of R2,MRE,and RMSE were 0.616-0.616,-17.50％-12.52％ and 0.102-0.133,respectively.Under the 70％ FC water treatment,the estimating model of leaf water potential based on EVI2 had the highest R2,the value of R2 was 0.922,and under the 60％ FC and 50％ FC water treatments,because of considering the influence of soil background,the inversion models of leaf water potential based on OSAVI had the highest R2,the values of R2 were 0.922 and 0.856,respectively.[Conclusion] All the four vegetation indices could be used for quantitative monitoring leaf water potential of winter wheat.But,when the leaf water potential estimating models were built for different water treatments,the influence of soil background on canopy spectral should be considered.The research results could provide a technical basis for wheat precision irrigation management and also provide supporting models for the parametric inversion of the onboard data.关键词
高光谱遥感/冬小麦/植被指数/叶水势/估算Key words
hyperspectral remote sensing/winter wheat/vegetation index/leaf water potential/estimation引用本文复制引用
陈智芳,宋妮,王景雷,孙景生..基于高光谱遥感的冬小麦叶水势估算模型[J].中国农业科学,2017,50(5):871-880,10.基金项目
国家自然科学基金(51609245)、水利部公益性行业科研专项经费项目(201501016-2)、国家现代农业产业技术体系建设专项(CARS-3-1-30)、国家公益性行业(农业)科研专项(201203077) （51609245）
