| Remote sensing,with its rapid,accurate,non-destructive and wide coverage,has been widely used in crop growth monitoring,growth,yield and grain quality research and early warning and prevention of various natural disasters.However,at present,there are few studies on using remote sensing technology to study crop sowing date and winter wheat growth,yield and grain quality under different sowing dates at home and abroad.In Jiangsu Donghai,GanYu,Guanyun,Huaian,Jiangyin,Jiangyan,Jinhu,Suqian,Yizheng,Kunshan and Suining regions such as the winter wheat as the research object,obtain winter wheat sowed at different time,looks like the key,quality parameters and winter wheat yield data,through the domestic environment star HJ-1A/1B and high marks a satellite image fusion to extract the corresponding GF-1 remote sensing variables and their combination to combine agriculture data,set up winter wheat sowing period,growth,quality and output of remote sensing monitoring models and evaluation.The comprehensive change rules of growth,quality and yield of winter wheat under different sowing dates were studied to explore the feasibility of remote sensing monitoring of winter wheat sowing dates with different combinations of remote sensing variables.The main research results are as follows:(1)The relationship between different sowing dates of winter wheat and satellite remote sensing variables was analyzed,and the correlation between sowing date information of winter wheat at jointing stage and booting stage and remote sensing parameters and their combinations was analyzed.It is feasible to use NDVI/GNDVI combination to monitor the sowing date in jointing period.It is also feasible to use GNDVI/RVI combination to monitor the sowing date at booting stage.The R2 and RMSE values of the remote sensing monitoring model for winter wheat sowing at jointing stage were 0.742 and 10.820,respectively.The R2 and RMSE values of the winter wheat remote sensing monitoring model at the booting stage were 0.774 and 11.702,respectively.The model accuracy established at the booting stage was 4.3%higher than that at the jointing stage.The precision of the model established at the booting stage was improved compared with the winter wheat sowing stage model established at the jointing stage.(2)The relationship between main growth parameters of winter wheat and satellite remote sensing variables was analyzed,and it was found that biomass,SPAD value,leaf area index(LAI)and leaf nitrogen content(LNC)at jointing stage and booting stage were well correlated with remote sensing parameters and their combinations.In the jointing stage,it is feasible to use(SIPI-PSRI)/(SIPI+PSRI)combination,NRI/SIPI,SIPI-PSRI and NDVI combination to monitor the nitrogen content and biomass of SPAD,LAI and leaf,respectively,with the fitting degree R2 reaching 0.612,0.511,0.568,0.533,RMSE 6.790,0.858,0.344 and 672.290,respectively.It was also feasible to use(DVI-RVI)/(DVI+RVI)combination,(NDVI-PSRI)/(NDVI+PSRI)combination,NDVI/NRI and(GNDVI-EVI)/(GNDVI+EVI)combination to monitor the nitrogen content and biomass of SPAD,LAI and leaf at the pregnant stage.The fitting degree R2 reached 0.588,0.575,0.607,0.563,RMSE 9.527,0.763,0.425 and 1182.802,respectively.At the same time,it is found that the combined monitoring precision of remote sensing variables is generally better than that of a single variable,indicating that it is feasible to monitor the growth of winter wheat by using the combination of remote sensing variables.(3)Analyzing the relationship between winter wheat grain quality and satellite remote sensing variables and the correlation between main winter wheat quality parameters and remote sensing parameters and their combinations.The results showed that the main winter wheat quality parameters had good correlation with remote sensing parameters and their combination.It is feasible to use the combination of(SIPI-PSRI)/(SIPI+PSRI),NDVI/GNDVI and NRI/GNDVI to monitor the protein content,wet gluten content and starch content,respectively,with the fitting degree R2 reaching 0.665,0.624,0.655,RMSE 0.999,1.742 and 1.061,respectively.(4)The relationship between winter wheat yield and satellite remote sensing variables was analyzed.The results showed that there was a good correlation between winter wheat yield and remote sensing parameters and their combination.It is feasible to use the combination of GNDVI/EVI and(GNDVI-EVI)/(GNDVI+EVI)to monitor the theoretical yield and the actual yield respectively,with the fitting degree R2 reaching 0.697 and 0.630,and RMSE reaching 84.760 and 71.087,respectively.Therefore,it is feasible to monitor winter wheat yield by using remote sensing variable combination.(5)After analyzing relationship between winter wheat sowing date and winter wheat yield and the correlation between actual winter wheat yield and sowing date,it was found that there was a good correlation between actual winter wheat yield and sowing date.Through the analysis of the correlation between sowing date and the production of 2018 winter wheat can be determined the optimum sowing date range in between October 28 and November 6,the R2 and RMSE value was 0.734 and 114.403,respectively.According to the comparison between the remote sensing monitoring map of the winter wheat sowing period in Jiangsu province in 2018 and the remote sensing monitoring map of the actual winter wheat yield in Jiangsu province in 2018,the preliminary results show that the best sowing period in south Jiangsu,middle Jiangsu and north Jiangsu is from November 2 to November 6,from October 31 to November 4 and from October 28 to November 1,respectively.(6)The hyperspectral change rule of winter wheat at different sowing dates was analyzed,and the correlation between different sowing dates of winter wheat and first-order differential hyperspectral reflectance was analyzed.It was found that at jointing stage,the hyperspectral reflectance with the highest correlation with sowing date was 717nm.At booting stage,the hyperspectral reflectance with the highest correlation with sowing stage was 731nm.At the same time,correlation analysis was conducted between sowing date and hyperspectral characteristic variables,and it was found that the fitting degree R2 of first-order differential sum(SDr)within the red edge was the highest.It was found by comparison that(SDr-SDb)/(SDr+SDb)had the highest fitting degree(R2),so it could be selected to construct the model,and it was found that using the combination of hyperspectral characteristic variables(SDr-SDb)/(SDr+SDb)and sowing date modeling was feasible.(7)According to winter wheat sowing date and the relations between variables and their combination of remote sensing is obtained by modeling for winter wheat sowing date figure,remote sensing monitoring in Jiangsu province in 2018 and divided into different sowing date the growth,yield and quality of winter wheat,end up with winter wheat growth,quality and output of different sowing date thematic map and remote sensing monitoring can be timely access to winter wheat growth information under the conditions of different sowing date,the field production with high yield and high quality was used to achieve the guiding significance. |
PDF Full Text Request