The Research On Estimation Of Winter Wheat Yield Based On Multi-temporal Remote Sensing Images

The technology of Satellite remote sensing can speedily,accurately and wide monitor the growth of crops.Multi-temporal remote sensing monitoring can overcome the deficiency of single phase remote sensing monitoring,and realize the dynamic monitoring for crop growth.In this paper,Dafeng city in Jiangsu Province as the study region,it analyzed the objects spectral information and researched the extraction of planting area with multi-units of HJ satellite remote sensing images,research and analysis of the identification and extraction of different spectral information,and combining biomass model with multi-temporal on winter wheat yield estimation.1 Research and analysis of the identification and extraction of different spectral information.Firstly,according to the spectral information features of images,the normalized difference vegetation index(NDVI)was extracted in images.The NDVI images made into plus synthesizing and subtract synthesizing operation,acquired the other two units of synthetic NDVI images.And than,through comparing and analyzing the spectral information for four units of NDVI images,the paper selected the plus synthesizing image which had obvious spectral information feature of vegetation and was applied for winter wheat planting area.Finally,based on the classification of NDVI threshold value and the superposition information of GPS samples,the data of winter wheat planting area and information of space distribution was extracted in Dafeng city.The results showed that the winter wheat area extracted by the synthetic image was 78712.13 ha,the accuracy up to 92.51%.It was proved that the NDVI synthetic images by remote sensing synthesis method can significantly enhance the difference between the winter wheat spectral information and the rest of the vegetation characteristics,and it can be conducive to realization of high precision extraction of winter wheat planting area.The feasibility of method based on extracting planted area,should be used for drawing the Figure that the remote sensing of winter wheat yield estimation2 Research and analysis of combining biomass model with multi-temporal on winter wheat yield estimation.Firstly,extracted out respective images of the normalized difference vegetation index(NDVI),ratio vegetation index(RVI)and difference vegetation index(DVI)by the spectral information features of images.Established linear relationships with the measured result of winter wheat LAI,selected the NDVI of sensitive vegetation index and applied to reversal development LAI of winter wheat growth period,it showed the best effect of linear relationship on flowering stage NDVI and LAI.Secondly,parameter adjustment and operation the extinction coefficient “k”,the test coefficient “m,n” and other parameters on the biomass model by LAI of winter wheat growth period.Operational result through harvest index of winter wheat to convert the predict yield of remote sensing.Finally,the empirical linear fitting accuracy reached 85.38% by the forecast output and the actual output based on the biomass model of winter wheat.The results showed that the better applicable properties in the method of winter wheat yield estimation by remote sensing based on the selection from different-temporal of vegetation index and parameters optimized biomass model which can provide a better method for the Agriculture Department that combined point with surface in the crop yield estimation.