Estimation of Tobacco Yield Based on the Radarsat-2 in Karst Mountainous Area

In order to realize the quantitative monitoring of modern tobacco cultivation in karst mountainous area, This paper take the Liuchang modern tobacco agriculture base in Qingzhen City Guizhou as the study area, According to the tobacco planting characteristics in karst mountainous area, with Radarsat-2 data source, choose radar image of tobacco in different growth period. After image preprocessing, calculated the tobacco scattering coefficient in different periods with different polarization, combined the field investigation data, utilized the coupling relationship of different backscattering of polarization tobacco radar data and the corresponding period fresh tobacco of weight establish three polynomial estimation model to estimate the yield of tobacco in different growth periods and verified the accuracy of the modelby the measured data. The results shown that: the overall accuracy of tobacco vigorous, mature period estimation models were respectively above 97 % and 94 %. Establish three polynomial estimation model with SAR backscatter coefficient and tobacco fresh weight can reflect the weight of tobacco well and meet the tobacco yield of estimation requirements in karst mountainous area, provide technical support for modern tobacco monitor widely.