Difference between Statistical Data, Actual Data and Prediction Data of Crop Area by Using Remote Sensor Imagery in Honghu City, Hubei Province
湖北省洪湖市作物播种面积的三种数据差异分析

Crop area estimation is a key technique for crop yield prediction. Today remotely sensed data have been widely applied to crop area estimation, but the result often differs from the statistics of crop area in the Statistical Yearbook, therefore the data of crop yield estimation by using remote sensing data is less adopted in the government's agricultural decision. Moreover, agricultural subsidy policy has been adopted by many countries, which aimed at the protection and development of agriculture. In China, the grain subsidy was provided according to taxed farmland area, a kind of statistics of crop area and usually smaller than the actual crop area, which caused part of peasant profits to be lost. Why the difference between the statistics of crop area in the Statistical Yearbook differ from predicted result by using remote sensing technique? Which is more close to the actual crop area? For answering the above questions and preventing peasant profits from being damaged, this paper studied the middle-season rice area in Honghu city, Hubei Province by using SPOT satellite imagery with a 10-m spatial resolution and data from field sampling survey. The result shows that the statistical area of middle-season rice from the Statistical Yearbook of Honghu city in 2006 was remarkably smaller than the actual and the predicted data of 58788.09 hm2 by remote sensing technique, or 52.156% lower than the latter. The main reason is due to the high ratio of customary unit of crop area adopted in estimation to standard unit(1.507). Besides, some local government concealed the actual rice planting area for evasion of agricultural taxation, which also made the statistical area of the middle-season rice differ from the actual and the predicted data by using remote sensing technique to a certain extent.