摘要
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.
出处
《江南大学学报（自然科学版）》
CAS
2012年第2期178-181,共4页
Joural of Jiangnan University (Natural Science Edition)
关键词
振动与波
小波包
消噪
多阈值
频段排序
vibration and wave
wavelet packet
de-noising
multi-threshold
frequency band sort
