小波包多阈值去噪的一种改进方法

本资料包含pdf文件1个，下载需要1积分

针对含噪信号中信号与噪声在频域内分布的差异性,通过小波包分解,将得到的小波包按照频率顺序重新排列,经过能量的比较将小波包分为3段进行阈值处理,实现多阈值去噪。仿真结果表明,与常用的Stein无偏似然估计阈值、极大极小准则阈值和固定形式阈值等方法比较,改进方法去噪效果更佳。 According to the differences of signal and noise frequency distribution,the signal and noise were decomposed by the wavelet packet.The resulting wavelet packets were rearranged in terms of the frequency order and were divided into three sections for threshold processing by comparing wavelet packet coefficients energy.Finally,the multi-threshold denoising was achieved.Comparing with the commonly used threshold methods,such as Stein,Minimaxi,Sqtwolog and etc.,the simulation results showed that the improved de-noising method was more effective.