论文部分内容阅读
利用SPOT4 VEGETATION的遥感数据产品生成的NDVI与NDWI植被指数时间序列图像集 ,通过ISODATA非监督分类方法 ,编制中国西北地区土地覆盖图。以TM影像人工解译结果作为真实值 ,通过对西北五省共计 47个均匀分布且异质性强的 2 5km× 2 5km样本区内的土地覆盖类型及其面积进行统计分析 ,修正了SPOT4 VEGETATION的土地覆盖分类系统 ,建立了各省验证结果的样本统计直方图并计算其回归系数。结果表明SPOT4 VEGETATION中国西北地区土地覆盖图在总体上具有较高的准确性。影响遥感数据自动分类精度 ,造成土地覆盖误判的原因主要来源于两个方面 :即异物同谱和混合像元问题。对于前者通过叠加各种辅助数据如DEM等可以降低误判的机率 ;对于后者运用混合像元分解的各种算法可以提高分类精度
Using NDVI and NDWI vegetation index time series images generated by remote sensing data products of SPOT4 VEGETATION, the land cover map of northwestern China was prepared by using ISODATA unsupervised classification method. Based on the artificial interpretation results of TM images as real values, the land cover types and their areas in a total of 47 samples with a uniform distribution and heterogeneity of 25 km × 25 km in the five northwestern provinces were statistically analyzed, and the SPOT4 VEGETATION Of the land cover classification system, established a sample of the provinces verify the results of statistical histograms and calculate the regression coefficients. The results show that the land cover map of the SPOT4 VEGETATION Northwest China is generally of high accuracy. Affect the automatic classification accuracy of remote sensing data, resulting in misjudgment of land cover mainly due to two aspects: the same foreign matter spectrum and mixed pixel problem. For the former by superimposing a variety of auxiliary data such as DEM and so can reduce the chance of misjudgment; for the latter using mixed pixel decomposition of various algorithms can improve the classification accuracy