城市混合像元分解中土壤与不透水面纯像元选取方法的对比研究——以厦门为例

近年来混合像元分解在城市地表组分监测与分析中的应用逐渐成为城市遥感的一个热点。纯像元的选取是混合像元分解过程中的重点和关键所在。以沿海城市厦门为研究对象,根据不同的土壤和不透水面纯像元选取规则,使用2组12种不同的纯像元选取方法对2007年1月8日TM影像进行混合像元分解,对分解结果的模型适宜度进行了比较,并使用2006年12月25日SPOT5高分辨率影像对分解结果的精度进行了比较和评估。结果表明:混合像元分解在纯像元选取时,S端元选取兼顾低反射率裸土和高反射率裸露基岩的纯像元可以整体提高分解的模型适宜度和分解精度;适度提高I分量纯像元中高反射率纯像元的比例有助于改善整体尤其是S、W分量的分解效果。

Pure Pixel Collection Methods for Soil and Imprevious Surface in Urban Spectral Mixture Comparative Analysis

Spectral mixture analysis has been used to monitor and analyze urban surface composition in recent years. Pure pixel collection is the key procedure of spectral mixture analysis. Taking Xiamen as an example,this paper compared twelve pure pixels collecting methods with different ratios of low to high reflectance pure pixels in both soil and impervious surface endmembers. Liner Spectral Model Analysis was applied in Landsat TM image acquired on January 8,2007 with four endmembers,vegetation,impervious surface,soil and water.Then this paper analyzed the model fitness of those twelve methods and assessed their accuracy through visual interpretation of SPOT 5 image acquired on December 25,2006. Results showed that,when selecting pure pixels of soil,it will improve model fitness and unmixing accuracy by taking both bare soil of low reflectance and bare rock of high reflectance into consideration;it will generally improve the unmixing results of all endmembers by raising the ratio of high reflectance pure pixels for impervious surface,especially for soil and water endmembers.