摘要
有好几种信号处理技术被应用于超声多普勒信号的分析.为了在平均频率估计的准确性和频谱展宽检测的敏感性上对各种方法加以优化,并比较这些不同处理方法的性能,该文研究了短时快速傅立叶交换、快速傅立叶变换结合平稳性提高的方法、改进的协方差估计法以及包括伪Wigner-Ville分布和Choi-Williams分布的时间-频率分析方法.通过模拟实验,得出了时间-频率分析方法可适用于多普勒信号的分析,参数估计方法和快速傅立叶变换结合平稳性提高的方法也有较好的分析性能.同时发现,当使用短时快速傅立叶变换分析多普勒信号时,以选用10ms的Hanning窗较为合适.
Several signal processing techniques can be used to analyse Doppler signals.To optimize each method for maximum accuracy of estimation of mean frequency,maximum sensitivity of detection of vortex and small degrees of spectral broadening and compare these performance of different methods,short term Fourier transform (STFT),stationarising plus Fast Fourier Transform (FFT),modified covaraince estimator and timefrequency distribution such as Pseudo--Wigner--Ville analysis and Choi--Williams analysisare studied in this paper' From the simulation study,it is found that time--frequency distribution are very suitable for Doppler signal while parameter estimators and stationarising plus FFT method also have a high performance. When STFT is used,a Hanningwindow with 10ms time duration is recommended.
出处
《复旦学报(自然科学版)》
CAS
CSCD
北大核心
1996年第1期53-61,共9页
Journal of Fudan University:Natural Science
基金
英国SERC基金
关键词
超声
多普勒
信号分析
ultrasound
Doppler
signal analysis