Abstract For the low spatial resolution, mixed pixels are prevalent in the Hyperspectral Image(HSI), which seriously affects the feature identification and classification accuracy. To solve the problem, Spectral Unmixing proposed to identify feature types and estimate abundance fraction is a quantitative remote sensing technology. The raditional spectral unmixing method devided into endmember extraction for and pixel unmixing is based on accuracy endmembers that construct mixed pixel, on the contrary, blind unmixing directly access endmember and abundance without the information. This paper analyzes pares the pros and cons of traditional spectral unmixing and blind unmixing, proposes a Hyperspectral image unmixing method base on iterative spectral mixture analysis bination of the tow methods. At first, mon spectral mixture model and a pixel unmixing method with excessive endmembers are introduced in this paper, then the mixed pixel position method based on iterative spectral mixture analysis is validated with real hyperspectral data; Second, non-negative matrix factorization is analyzed, and weighted non-negative matrix factorization whose weighted matrix is settled is proposed as a Blind Unmixing algorithm; Finally, the non-negative matrix factorization algorithm conbines the mixed pixel position method based on iterative spectral mixture analysis algorithm, the endmember extraction algorithm called iterative error analysis, and the endmember update method based on statistical endmember to a blind unmixing method which is validated with real and simulation hyperspectral data. Keywords: Hyperspectral, Pixel unmixing, Non-negative Matrix Factorization, Iterative Spectral Mixture Analysis 目录 摘要........................................................................................................... I ABSTRACT........................................................................................... II 1 绪论 研究背景及意义........................................
