【摘要】 MODIS时间序列数据提取水稻面积是一种常见的区域尺度水稻种植面积提取方法,但对于该方法的面积提取精度的分析还不够明确。为此本文利用Landsat 8 OLI数据结合监督分类的最大似然法提取江苏省水稻面积,并以Landsat数据的提取结果作为参考值,对时间序列MODIS数据的水稻面积提取结果空间位置匹配状况进行评估。主要通过两种方案来展开,一是沿用前人的MODIS数据水稻提取算法参数,提取江苏省水稻面积,评估水稻面积提取总量精度,二是修改水稻提取算法参数,使MODIS数据水稻面积提取与统计数据持平,分析其空间位置匹配状况。研究发现:对于方案一,时间序列MODIS数据提取的水稻面积,在空间分布上有一定准确度,但总体面积偏低;对于方案二,空间匹配上存在较大误差,特别是在南京和启东等地存在误提。沿用统一水稻提取算法参数,年纪之间气象条件变化,造成提取结果的不确定性。更多还原

【Abstract】 The method utilizing time series data to extract rice area is a common one, but the analysis of extraction accuracy lacks in current research.This paper used Landsat 8 OLI data and maximum likelihood of supervised classification method to extract the rice area of Jiangsu province. The Landsat data extraction results were adopted as the reference to evaluate the matching condition of spatial location, there were two groups of data using for comparison, one applied MODIS data extraction algorithm parameters from related articles to extract the rice area of Jiangsu province, The other modified the rice extraction algorithm parameters, making the rice area of MODIS data extraction and statistical data consist to analysis the matching condition of spatial location. The results illustrated that the first method had an acceptable accuracy, while it didn’t match well on the spatial distribution of the overall area. For the second one, there was a large matching error on the overall scale, especially in Nanjing and Qidong. In both methods, unified algorithm parameters of rice area extraction and inter-annual varations of meteorological conditions caused the uncertainty of the extraction results.更多还原