Study On Remote Sensing Inversion Of Growth Parameters Of Oilseed Rape At Seedling Stage

Oilseed rape is one of the main oil crops in China,with large sowing area and wide distribution area.Accurate and timely monitoring of regional oilseed rape growth can provide important decision information and technical support for the government and farmers.Leaf area index,chlorophyll content and biomass of the critical growth period can accurately reflect the nutritional status and growth trend of oilseed rape,but the traditional field monitoring methods are time-consuming and difficult to monitor in real time.Remote sensing technology can quickly estimate crop growth parameters through abundant spectral information,and is widely used in modern agriculture.Based on field experiment data,this study researched the experience model and PROSAIL model method for the inversion of the growth parameters of oilseed rape at seeding stage.Based on GF-1 image data,we obtained the chlorophyll content distribution at over-wintering stage of oilseed rape with both two inversion methods.By comparing with ground measured data,this paper analyzes the feasibility of two kinds of inversion methods for large scale growth monitoring of oilseed rape.Around the above research content,there are mainly the following results:(1)According to spectral response function of GF-1 WFV3 satellite,the oilseed rape canopy reflectance spectral of experimental field were simulated to the broadband of blue,green,red and near-infrared,and 15 broadband vegetation indexes were calculated.Then we analyzed the correlation relationships of vegetation indexes and growth parameter of oilseed rape.In this fifteen vegetation indexes,Enhanced vegetation index(EVI),Optimized soil adjusted vegetation index(OSAVI),Green ratio vegetation index(GRVI),Reclassified vegetation index(RDVI),Green normalized vegetation index(GNDVI)and Blue-green normalized vegetation index(GBNDVI)have strong relationships with Leaf area index(LAI),Ratio vegetation index(RVI),Green ratio vegetation index(GRVI)and Green normalized vegetation index(GNDVI)have strong relationships with biomass,Green ratio vegetation index(GRVI)has the best relationship with chlorophyll,then is Ratio vegetation index(RVI).The empirical statistical models of leaf area index,chlorophyll content and biomass of oilseed rape at seeding stage were constructed respectively,and the vegetation index of optimal model was GNDVI,GRVI and GNDVI.(2)The sensitive bands corresponding to the input parameters of the PROSAIL model are different,and the chlorophyll content Cab is the most sensitive parameter in the visible band,and the leaf area index is the most sensitive parameter in the near infrared band.Parameter uncertainty and spectral data selection will affect the PROSAIL model inversion accuracy,for the 2 sets of parameters,P1 witch has eight unknown parameters,and P2keeps five based on real field condition,as well as three spectral data band,B1 are all the visible-near infrared hyperspectral bands within 400-1040nm,B2 are four simulated bands of GF-1,B3 is based on the correlation analysis and stepwise regression optimization of filter band.The inversion accuracy of both leaf area index and chlorophyll content with PROSAIL model have regular of P1<P2,S3>S2>S1.For the inversion of leaf area index,except for the low inversion accuracy of P1-S1,of which the decision coefficient of the predicted value and the measured value is only 0.42,and the inversion accuracy of several other inversion schemes is better(R~2>0.5).For inversion of chlorophyll content,the decision coefficient of the predicted values of several inversion schemes and measured values are higher than 0.5,and the decision coefficient is higher than 0.6 for broadband and optimal band inversion.(3)The chlorophyll content of the oilseed rape in Shayang from 2015 to 2016 at over-winter stage was reflected by the optimal prediction model of GRVI-chlorophyll content and the P2-B2 PROSAIL inversion scheme,and the chlorophyll content distribution map was obtained.By comparing with the ground survey data,empirical model and PROSAIL model inversion results are lower than the measured values,but PROSAIL model inversion result is better than empirical model,which show that PROSAIL model inversion has higher precision stability.