Research On Hyperspectral Remote Sensing Monitoring Model Of Rice Growth Parameters Under Gradient Warming Condition

Taking the north japonica rice NO.2 as the research object and through field experiment of Free Air Temperature Increase system(Free Air Temperature Increase,FATI),this research determined the different temperature condition(CK-0w,T1-500 w,T2-1000 w,T3-1500 w,T4-3000w)of rice jointing stage,booting stage,heading stage,filling stage and mature stage crown height spectrum data and its growth parameters(LAI,biomass,canopy leaf nitrogen content in rice).The relationship between the original spectral reflectance and vegetation index and the corresponding growth parameters was analyzed by using the techniques and methods of mathematical statistics analysis,the monitoring model was established,and the model was evaluated and tested to select the most suitable hyperspectral monitoring model for the estimation of rice growth parameters at different growth stages.The research results can provide a theoretical basis for remote sensing monitoring of rice in northern China under the background of climate warming.The main results are as follows:(1)Analysis of spectral characteristics of rice canopy.The spectral changes of rice canopy had typical spectral characteristics of green vegetation,and the spectral reflectance was different at different growth periods.In the visible region,the spectral reflectance of the canopy during grouting and maturity was significantly higher than that during other growth periods.In the near infrared region,the spectral reflectance of the canopy in booting and heading stages was significantly higher than that in other growing stages.The effects of different warming conditions on the canopy spectra of rice were basically the same in different growth periods.The increase of canopy temperature could not change the spectral reflection characteristics,but changed the values of the reflectance,and the "green peak" and "red valley" appeared in advance to different degrees.(2)Study on monitoring model of rice LAI under gradient warming condition.The monitoring model of rice LAI was established by transforming soil adjusted vegetation index(TSAVI)at jointing stage and grouting stage and perpendicular vegetation index(PVI)at booting stage and heading stage that the correlation coefficient between predicted value and real value was greater than 0.8(P<0.01).Among them,the model established by PVI(698,960)at heading stage can monitor LAI during the whole growth period of rice,and the correlation coefficient between the predicted value and the measured value is 0.7277**.(3)Study on monitoring model of rice biomass under gradient heating condition.The monitoring model of rice biomass was established by perpendicular vegetation index(PVI)at jointing stage and maturity stage,transforming soil adjusted vegetation index(TSAVI)at booting stage and modifying soil adjustment vegetation index(m SAVI)at heading that the correlation coefficient between predicted value and real value reached 0.9004**,and the monitoring model of rice biomass was not established in the grouting stage.Among them,the models of PVI(696,944)at the jointing stage and PVI(625,947)at the maturity stage could be used to monitor the biomass of rice during the whole growth period,and the correlation coefficients of predicted value and measured values were 0.7037** and 0.8541**,respectively.(4)The monitoring model of nitrogen content in rice canopy leaves under gradient warming condition was studied.The monitoring model of nitrogen content in rice canopy leaves was established by regression difference vegetation index(RDVI)at jointing stage and grouting stage,perpendicular vegetation index(PVI)at heading stage,ratio of vegetation index(RVI)at booting stage and maturity stage and new vegetation index(NVI)at booting stage that the correlation coefficients between predicted and true values were all greater than 0.75(P<0.01).Among them,the models of booting stage RVI(718,996)and NVI(718,996)had the same accuracy as those of the model established in the mature stage RVI(727,784)and could monitor the nitrogen content in the canopy leaves of rice during the whole growth period.The correlation coefficients of predicted value and measured value were 0.9246** and 0.9220**,respectively,which were higher than those of the monitoring model during the partial growth period.