一种双模噪声模型的统计特性及其去噪研究  

The Study on Statistics of One Category of Dual-Mode Noise and Resolve It

作　　者：孙万麟[1] 杨莲红[1] 山拜.达拉拜 

机构地区：[1]昌吉学院物理系,新疆昌吉831100 [2]新疆大学信息科学与工程学院,新疆乌鲁木齐830046

出　　处：《江西师范大学学报（自然科学版）》2010年第5期506-510,共5页Journal of Jiangxi Normal University(Natural Science Edition)

基　　金：国家自然科学基金(60062001);昌吉学院科研基金(2010YJYB009)资助项目

摘　　要：利用多尺度小波包变换良好的时频局部分析能力,对一类非高斯噪声——双模噪声的统计特性进行研究.理论分析和仿真表明,小波包检测具有计算量小、算法比较简单和实时性较强的特点,不仅实现了将非高斯噪声——双模噪声转化为简单的高斯噪声来处理,而且比传统的高阶统计量和经典检测的检测性能都优越.On the basis of time-frequency analysis capabilities of multi-scale wavelet packet transform,statistics of one special of non-Gaussian noise——Bimodal Noise is be studied.Theoretical analysis and simulation results show that the wavelet packet inspection has a small amount of calculation,the simple algorithm and the strong real-time characteristics,not only brings out turning non-Gaussian noise（Bimodal Noise）out to be easy Gaussian noise to deal with,but also is better than the traditional higher-order statistics and classical test detection performance.

关 键 词：小波包变换 非高斯噪声 双模噪声 检测性能 

分 类 号：TN911[电子电信—通信与信息系统] TN918[电子电信—信息与通信工程]