基于快速独立分量分析的多分辨率遥感图像融合算法

The Fusion Arithmetic of Multi-Resolution Remote Sense Image Based on M-FastICA

独立分量分析作为盲源信号分离的一种有效的方法在许多方面获得成功应用。讨论了独立分量分析的基本原理、目标函数选择和算法,并在此基础上,对快速独立分析算法FastICA的核心迭代过程进行改进,得到M-FastICA算法,改进算法减少了独立分量分析的迭代次数,从而提高了算法的收敛速度。最后将M-FastICA算法应用到遥感图像的融合上,实验结果表明,改进算法在融合效果相当的前提下,收敛速度更快。

Independent Component Analysis （ICA） is an effective approach to the separation of blind signal, and has attracted broad attention and has been successfully used in many fields. The fundamental, target functions and practical algorithm of Independent Component Analysis are discussed. Then, by modifying the performance of FastICA algorithm are merged into one iteration of M-FastICA. The M-Fas- tICA algorithm achieves the correspondent effect of FastICA. So the convergence of ICA will be accelerated. Finally, M-FastICA is applied to images fusion. The experiment results show that the modified algorithm reduces iterations with the correspondent fusion performance.