基于快速曲波变换和独立分量分析的有噪图像盲分离算法  

Noisy Image Blind Separation Algorithm Based on The Fast Curvelet Transform and Independent Component Analysis

作　　者：毕杨[1,2] 肖军[1] 

机构地区：[1]西安航空技术高等专科学校电气系,陕西西安710077 [2]西北工业大学航海学院,陕西西安710072

出　　处：《自动化技术与应用》2010年第1期53-56,共4页Techniques of Automation and Applications

摘　　要：曲波(Curvelet)变换有两种数值实现方法,一种是基于非均匀采样的快速Fourier变换即usfft算法;另一种是基于特殊选择的Fourier采样的卷绕即wrapping算法。将这两种实现方法针对图像去噪性能进行了比较,并且提出将曲波变换的wrapping算法和FastICA算法结合起来,对含有噪声的混合图像进行了盲分离。仿真研究结果表明,对于含有加性观测噪声的混合图像,该方法有更好的去噪分离效果,并且运算速度显著提高。The There are two digital implementations： for curvelet transform the first digital transformation is based on unequallyspaced fast Fourier transform while the second is based on the wrapping of specially selected Fourier samples. The realization of these two methods for image denoising performance is compared, and a new method of noisy image blind separation is proposed using Fast Discrete Curvelet Transforms via wrapping algorithm and the algorithm of Fast ICA. The simulation results show that this method has more effective performance and the higher running speed in mixed image de-noising and separation for the mixed images with additive white Gaussian noise. Therefore, this algorithm is more suited for sophisticated noisy image blind separation.

关 键 词：快速离散曲波变换 usfft算法 wrapping算法 FASTICA算法 图像盲分离 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]