Study On Estimating Model Of Cotton Yield Based On Satellite Remote Sensing

The studying on forecasting models of cotton yield by remote sensing of cotton yield is at developing stage in China, and the methods and technologies for estimation of cotton yield are not mature. Forecasting models of cotton yield on regional scale of Xinjiang by space information technology (such as remote sensing) were thought to be significant. In the present study, based on the data of TM Image in Shihezi of North Xinjing on 2007 and 2008, the correlation analysis among normalized difference vegetation index (NDVI), ratio vegetation index (RVI), deferent vegetable index (DVI) and perpendicular vegetation index(PVI) were performed combining with the data of cotton yield by five linear equations with one unknown and nonlinearity (index, logarithm, quadratic polynomial, cubic polynomial and power function), to establish the optimum forecasting models of cotton yield. Results:(1)In the present study, we fit the cotton yield and four VI of three cotton varieties in research region. And there is close correlation between VI and cotton yield of different cotton varieties at bolling stage and dehiscence stage, thus it is feasible to evaluate the cotton yield.(2)Obtained the optimum forecasting models of cotton yield With a regression analysis of the NDVI and cotton yield, the PVI linear model of multitemporal TM Image at bud stage and bolling stage fit well, and so does the NDVI exponent models of single phase TM Image.(3)The measured values are close to predicting values and the accuracy of fitting reach above 0.8, which show that forecasting models of yield based on NDVI and PVI can meet the need of general supervision of cotton yield. In addition, it is a simple and effective method for evaluating the cotton yield.