论文部分内容阅读
利用MODIS增强型植被指数(EVI)时序数据,基于中国陆地生态系统55种植被类型上的468个测试点和一个测试区进行了实验,综合比较欧氏距离、光谱信息离散度、光谱角余弦、核光谱角余弦、相关系数、光谱角余弦—欧氏距离6种距离测度方法对遥感植被指数时序数据聚类精度的影响,结果表明:相关系数方法的聚类精度最差;光谱角余弦—欧氏距离方法充分利用了植被指数时序数据的曲线幅度和形状特征,在这6种距离测度方法中表现出了最优的聚类效果;只对光谱亮度敏感的欧氏距离方法或只对曲线形状敏感的光谱角余弦方法,无论是在区分地物类型方面,还是在区域应用上,表现效果均较差;核光谱角余弦虽然在点数据测试上表现较差,但在区域应用上却有较好的表现;光谱信息离散度无论是在点数据测试上还是在区域应用上均表现出了较为适中的效果。
Based on MODIS enhanced vegetation index (EVI) time series data, 468 test sites and a test area of 55 types of terrestrial ecosystems in China were tested. Euclidean distance, spectral information dispersion, spectral cosine, The results show that the correlation coefficient method has the poorest clustering accuracy. The spectral angle cosine - Euclidean distance - Euclidean distance is used to measure the clustering accuracy of remote sensing vegetation index time series data. The Euclidean distance method takes full advantage of the curve amplitude and shape characteristics of the vegetation index time series data and shows the optimal clustering effect in the six distance measurement methods; the Euclidean distance method only sensitive to spectral brightness or the curve shape Sensitive spectral cosine method, whether it is in the distinction of feature types, or in the regional application, the performance is poor; nuclear spectral cosine Although the point data test performed poor, but there are more applications in the region Good performance; spectral information dispersion no matter in the point of data testing or in the regional application showed a more modest effect.