Using Remote Sensing Technology To Monitor Key Growth Diagnosis Parameters Of Winter Wheat Following Rice At Main Growth Stages

The use of satellite remote sensing technology in monitoring the growth and predicting grain yield and quality of wheat following rice has been extensively studied and gained some achievements in the production practice. The research is conducted in Jiangsu Taixing, jiangyan, Xinghua, Dafeng and Yizheng with rice stubble wheat as the research object, obtaining key diagnosis parameters and mature period of wheat grain yield data, such as LAI biomass, SPAD, leaf nitrogen content, leaf nitrogen accumulation and leaf nitrogen density in regreening stage, jointing stage and booting stage, blooming stage,15days after blooming stage. Based on the remote sensing variables of the corresponding environmental mitigation satellite HJ-1A/1B imaging, the research analyzes the agronomy datas and remote sensing variables, establishes and evaluates the model used in key diagnosis parameters monitoring and wheat yield remote sensing forecasting, discussing wheat’s diagnosis in main growth period and grain yield forecasting feasibility of the remote sensing monitoring, the main research results are as follows:(1) By the analysis of the single-factor remote sensing model of main growth diagnosis parameters at the growth stages of regreening, jointing, booting, blooming and15days after blooming stage, there is a certain correlation between LAI, biomass, SPAD, leaf nitrogen content, leaf nitrogen accumulation, leaf nitrogen density and different remote sensing parameters at main growth stages. At regreening stages, NDVI, RVI,B1, RVI, NRI and B2could be used to monitor LAI, biomass, SPAD, leaf nitrogen content, leaf nitrogen accumulation and leaf nitrogen density. At jointing stages, B3could be used to monitor LAI and leaf nitrogen content, B4could be used to monitor biomass, leaf nitrogen accumulation and leaf nitrogen density, NRI could be used to monitor SPAD. At booting stages, RVI could be used to monitor LAI and leaf nitrogen density, NRI, B2, NDVI and DVI could be used to monitor biomass, SPAD, leaf nitrogen content and leaf nitrogen accumulation. At blooming stages, RVI, B4, B3, and GNDVI could be used to monitor LAI, biomass, SPAD and leaf nitrogen density, NDVI could be used to monitor leaf nitrogen content and leaf nitrogen accumulation. At15days after blooming stage, SIPI、B4、GNDVI and B1could be used to monitor LAI, biomass, SPAD and leaf nitrogen density.(2)By the analysis of the multi-factor model of main growth diagnosis parameters at one stage:At regreening stage, we can use B1and B2to establish the model to monitor SPAD, and the model’s RMSE changes from4.15to2.19; At jointing stage, we can use B4, DVI and NDVI to establish the model to monitor biomass, and the model’s RMSE changes from617.45kg/ha to442.5kg/ha; At booting stage we can use B2and NRI to establish the model to monitor leaf nitrogen accumulation, increased prediction efficiency of model.(3) Analysis of multi-factor monitoring model of main growth diagnosis parameters at different stages, the accuracy of the multi-factor model is much better than the single-factor model. At jointing stage, the correlation of the monitoring model using the B4(jointing stage) and NRI (regreening stage) to monitor leaf nitrogen accumulation is slightly lower,but the RMSE (g/m2) is reduced from1.73to1.09, indicating that we can use the the B4(jointing stage) and NRI (regreening stage) to monitor leaf nitrogen accumulation; At booting stage, the RMSE (g/m2) of the monitoring model using the RVI (booting stage) and B3(jointing stage) to monitor LAI changes from1.07to0.61, indicating that the combination of remote sensing variables to monitor LAI is feasible; At blooming stage, the accuracy of combination model using the RVI (blooming stage) and NRI (booting stage) to monitor LAI greatly improved, RMSE decreased from0.92to0.31. The RMSE (g/m2) of the monitoring model using the the GNDVI, DVI ((blooming stage) and B4(booting stage) to monitor leaf nitrogen density changes from1.07to0.61; At15days after blooming stage, we can use the NDVI (blooming stage) and GNDVI (15days after blooming stage) to monitor leaf nitrogen content, we also can use the NDVI (blooming stage) and GNDVI and PSRI (15days after blooming stage) to monitor leaf nitrogen accumulation; So the multi-factor monitoring model of main growth diagnosis parameters at different stages is feasible, and the correlation or the accuracy of the model is better than the single-factor model.(4) By the analysis of the correlation between the grain yield and remote sensing variables at blooming stage and15days after blooming stage, the results show that:at blooming stage the, we can select DVI and RVI to build a multi-factor yield prediction model, then the accuracy of the model significantly improve. At15days after blooming stage, we can select NDVI and GNDVI to build a multi-factor yield prediction model, then the accuracy of the model is higher than the NDVI single-factor yield prediction model.(5) Based on the quantitative relationship between growth diagnosis parameters and remote sensing variables, the monitoring and forecasting model can be set up to obtain the spatial distribution of wheat growth diagnosis parameters on remote sensing image. According to the different level’s diagnosis parameters and production division, images of remote sensing monitoring projects in different diagnosis parameters and output forecasting projects can be formulated. The wheat’s diagnosis is timely available which can be effectively controlled.thus we can reasonably guide production, finally realizing the good quality and high yield.