Research On Soybean Yield Estimation Model Based On Vegetation Index, Temperature And Precipitation

Soybean is regard as a kind of food raw material and industrial and agricultural production and it is also a kind of bulk import cargo,so it is meaningful for not only improving production control and developing productivity but also formulating macroeconomic plans and food policies to estimate yield of soybean precisely and quickly.It is significant and meaningful to estimate the yield of soybean before harvesting.Both remotely sensed imagery of GF-2 and meteorological data recorded daily about Xianghe farm are used in this paper to establish the model of estimating yield with remotely sensed imagery,the model of estimating yield with meteorological data and the model of estimating yield with both remotely sensed imagery and meteorological data.Finally,the accuracy of different models are estimated.The conclusions are displayed as follow:(1)The change of NDVI between early July and mid-August could be fitted with a parabola function and the max appeared in the early August,and it is appropriate to estimating the yield with NDVIs in the time period.(2)The influents caused by temperature appear in August and that caused by rain appear in July-August.(3)The yield grows with the temperature increasing and a linear function could be used to depict the relationship between them,and the yield reduces with the rain increasing and a linear function could be used to depict the relationship between them.(4)The minimum precision,the maximum precision and the average precision of the comprehensive model of remote sensing index and meteorological factors can reach 89.69%,95.37% and 93.92% respectively,which are higher than those of other models and can significantly improve the accuracy of the yield estimation model.