The Application and Evaluation of Two Methods Based on LSMM Model——A Case Study in Guangzhou
The Application and Evaluation of Two Methods Based on LSMM Model——A Case Study in Guangzhou
复制标题
DOI：
--
复制
发表时间：
2008
期刊：
影响因子：
--
通讯作者：
Fan Feng-lei
中科院分区：
文献类型：
--
作者：
Fan Feng-lei
目前，城市不透水面的研究方法很多，如线性光谱分解法、非线性光谱分解法、匹配滤波法等，其中，线性光谱解混方法得到了广泛的应用，基于线性光谱混合模型（LSMM）的线性光谱解混方法本文采用像元纯度指数（PPI）和人工选择方法进行端元选择，利用线性光谱解混方法得到了每个像元中各端元的丰度图像和RSME图像，并利用RSME图像对各端元的丰度图像进行了分析。最后，对两种端元选取方法进行了分析，并从定性的角度对结果的精度进行了评价，结果表明，该方法能够有效地获取广州市8个地区的不透水面权重图像，人工选取方法获取的权重图像精度优于PPI方法。
Currently,there were lots of methods on urban impervious surface researching,such as Linear Spectral Un-mixing method,Non-linear Spectral Un-mixing method,Matched Filtering method and so on,however, among all these methods the Linear Spectral Un-mixing method was applied popularly and wildly by the researchers.The Linear Spectral Un-mixing method based on Linear Spectral Mixing Model (LSMM) and obtained the proportion of each land types in each pixel through un-mixing each pixel in the remote sensing image.This paper used Purity Pixel Index (PPI) and manual selection method to select end-member and gained the abundance images of each end-member in each pixel and the RSME image by using linear spectral un-mixing method.At last,it obtained the impervious surface weight image of 8 areas in Guangzhou.Besides,this paper analyzed the two end-member selection methods and evaluated the precision of the result from the qualitative standpoint.The result revealed that the method could abtain the impervious surface weight image of 8 areas in Guangzhou efficiently and the manual selection method could gain the weight image with more precision than the PPI method.