摘要
用小波分析进行信号消噪常用多尺度小波变换和小波包变换。多尺度小波变换是将信号分解成高频和低频成分,低频包含信号的主要性能,高频含较多噪声,将高频平滑后再重建即可消噪。其消噪处理有强制、默认阈值和给定软/硬阈值3种方法。小波包变换消噪是将信号的小波包分解、计算最佳小波包基、分解系数的阈值化处理及重构原来信号实现消噪。并给出了几个Matlab消噪函数:多尺度一维小波分解和重构wavedec及waverec、消噪默认阈值ddencmp和消噪函数wdencmp;小波包变换的ddencmp 和wdencmp函数。试验表明,此方法具有较高的有效性和实用性。
Based on wavelet analysis, the noise of signal was eliminated with multi-scale wavelet transform and wavelet packet transform. Multi-scale wavelet transform is that the signal was decomposed into high or low frequency elements. The low frequency elements include the main performances of the signal while the high frequency elements consist of more noises, reconstructing the signal can be achieved after flatting the high frequency elements. The three means of constraint, acquiescent threshold value, and given soft/hard threshold value were used to eliminate signal noise. Wavelet packet transform elimination noise is that wavelet packet of the signal was decomposed, and the best wavelet packet base was computed, the threshold value of decomposed coefficients was processed and the original signal was reconstructed, so, elimination noise of signal was realized. A few Matlab elimination noise functions are given such as wavedec and waverec 1-D function of decomposition and reconstruction of multi-scale wavelet, ddencmp function of acquiescent threshold value and wdencmp function of elimination noise, ddencmp and wdencmp function of wavelet packet transform. Experiments show this method has high efficiency and practicality.
出处
《兵工自动化》
2002年第6期22-24,共3页
Ordnance Industry Automation
关键词
小波分析
小波包
分解
重构
消噪
信号处理
噪声
Wavelet analysis
Wave packet
Decomposition
Reconstruction
Elimination noise
