Research On Henan Province Winter Wheat Yield Estimation Model Based On MODIS-NDVI

Wheat is one of the major food crops in China. The sown area accounted forone-fifth of the total grain crops. The overall wheat production of Henan Provinceaccounted for approximately50%of the country’s total output. The use of advancedtechnology to monitor wheat acreage, growth condition and timely and accuratepre-estimate for wheat production， can make a contribute to wheat productionmanagement, optimize the pattern of planting space, further play its productionpotential. At the same time, it provides an important technical support for the country toadjust the grain reserves, the government departments making macroeconomic plansand scientific and rational food policy, as well as futures market commodity tradingvaluation. It has important value and significance.In recent years, with the continuous advancement of space technology andcomputer image processing technology gradually increase. satellite remote sensingtechnology is developed quickly and widely applied in various fields。Remote sensinghas also become an important information-gathering, processing means of agriculturalsciences. Estimation using remote sensing technology is based on spectral reflectancecharacteristics of different crops. The spectral data obtained by the use of remotesensing satellites carry sensors, combined with the different type of data processingmethod make up of vegetation index. Then using vegetation index and yield establishregression model.In this paper, the normalized difference vegetation index NDVI has been used forestimating yield research. By analysis and using the best estimate of the productionperiod, in accordance with the growth phase synthesis NDVI, and ultimately theoptimal yield estimation model using principal component regression method. Specificresearch work and achieved results are as follows：Remote sensing database expansion: I got the Winter wheat yield data from2009to2011of each county in Henan Province by investigate. At the same time, downloadthe MODIS satellite remote sensing images from NASA database website.Remote sensing satellite image processing research: Then use remote sensingimage processing software, such as ENVI to get one-day vegetation index remotesensing image of Henan Province by image scaling, calibration, projection work.Studies on the synthesis of NDVI: From study of the regression model set up by different synthesis methods, we know15days NDVI synthesis method has moreadvantages than the10days cycle method for yield estimation.Correlation research: From the Correlation Analysis between NDVI value andthe yield of winter wheat. We found that the best estimate period of winter wheat yieldestimation is late March, period of seedling establishment.Yield estimation model building and analysis: Establish simple and multivariateregression model for comparison analysis. Then use stepwise regression and principalcomponent regression method to optimize the multiple regression equation. We got theoptimal winter wheat yield regression model. The goodness of the equation on theactual2012annual output value was90.72%, satisfied the accuracy requirements of thepractical application of winter wheat yield. The equation has practical significance.