基于小波变换的非平稳信号分析与处理

Non-Stationary Signals Analysis and Processing Based on Wavelet Transform

针对傅里叶变换在非平稳信号分析中的难点问题,将小波变换和短时傅里叶变换应用于典型的非平稳信号,进行时频分析研究,所获得的特性比普通的傅里叶变换构成的频谱分析结果更清晰、准确,表明这两种时频分析方法独有的优越性、正确性、实用性.小波的多分辨特性更使得分析结果优于短时傅里叶变换.在此基础上用该方法对小波时频分析方法的抗噪性能进行研究,并从小波基函数的性质和实际仿真效果两方面设定参数,得出最优分析结果,改善了小波时频分析方法在高频处频率不易确定的问题.

For the limitation of the Fourier transform, the wavelet transform and short-time Fourier transform are applied to typical non-stationary signals. The time-frequency characteristics obtained is more clear and accurate than traditional FFT spectrum analyze, indicating that the two time-frequency analysis methods are practical and advanced. And the multiresolution feature makes the analysis results better than short time Fourier transform. On the basis, study the anti-noise performance and choose appropriate parameter to get optimal results, improving the problem of poor frequency resolution at high frequencies of wavelet time-frequency analysis method.