基于单体扩张的端元提取算法-红外技术.PDF

第38 卷 第11 期 红 外 技 术 Vol.38 No.11 2016 年11 月 Infrared Technology Nov. 2016 基于单体扩张的端元提取算法 董安国，韩 雪，龚文娟 （长安大学理学院，陕西 西安 710064 ） 摘要：端元提取是高光谱混合像元分解的重要环节。为了提取高光谱图像的端元，本文基于线性表示 理论与凸锥模型理论，论证了：与单体共面的单体外向量被单体的顶点向量线性表示时，表示系数必 有负值，从而给出了理想情形下判别端元的充要条件，并在此基础上，针对非理想情形提出了一种提 取端元的迭代算法。实验结果表明，算法提取端元的精度优于VCA 算法、效率高于搜索算法，算法 稳定性好，对噪声的敏感性低。 关键词：高光谱；凸锥模型；端元提取；单体 中图分类号：TP7 文献标识码：A 文章编号：1001-8891(2016)11-0947-06 Endmember Extraction Algorithm Based on the Simplex Expansion DONG Anguo ，HAN Xue ，GONG Wenjuan （School of Science, Chang ’an University, Xi ’an 710064, China） Abstract ：Endmember extraction is a key step in unmixing hyperspectral mixed pixels. In order to extract endmembers of hyperspectral image, this paper proves that if a vector in vitro and vivo is represented by a vertex vector, there will be a negative coefficient based on the theory of linear representation and the theory of convex cone model. The theory gives a necessary and sufficient condition for the endmembers identification. A new iterative algorithm is proposed under non-ideal situation. The experimental results show that the precision of the endmember extraction using the algorithm proposed in this paper is better than VCA algorithm. And this algorithm has better efficiency than the search algorithm. It has good stability and a low sensitivity to noise. Key words ：hyperspectral ，convex cone，endmember extraction，simplex N-FINDR 算法，尽管无参数选择效果好，但因其搜索 0 引言 凸面体最大体积需遍历所有可能像元，计算量巨大。 高光谱遥感图