| In recent years, Doppler ultrasound technology has developed rapidly and it is widely used in clinical diagnosis of vascular disease due to its non-destructive detection of blood flow of blood vessels. A lot of research has shown that vascular disease can be reflected by the change of the max frequency of Doppler signal. Therefore, it is very important for the diagnosis of vascular disease whether the max frequency of Doppler signal (the envelope) can be extracted accurately.Firstly, taking the research project of the extraction of max frequency curve of Doppler blood flow as the background, the experiments of spectrum interpolation are done in both time domain and frequency domain and it is known that the missing spectrum can be interpolated completely by the one-dimensional linear interpolation method in frequency domain. Therefore, it can be determined that denosing should be carried out in time domain and data interpolation in frequency domain. Thus, the new scheme of Ultrasonic Doppler blood flow signal estimation is given.Secondly, in order to reduce the influence of noises in practical application, The Doppler signal denoising based on neighborhood shrinkage method which should be carried out before spectrum analysis and envelope extraction is proposed. By analyzing the selection principle of wavelet basis function and the character of Doppler signal, the wavelet basis function which is suitable for Doppler signal denoising is selected. Thus, the proposed method and wavelet threshold method are adopted to denoise the practical Doppler signal. In simulation, the influence on denoising effect of such factors as decomposition level, wavelet basis function and window size of neighbor coefficients are analyzed. Simulation results have shown that the proposed denoising method based on neighborhood shrinkage outperforms the traditional wavelet threshold method in denoising performance and robustness.Finally, by improving the envelope extraction of ultrasonic Doppler spectrogram based on three-line fitting method, the maximum distance method is proposed, and the performance of the percentage method and the maximum distance method are compared in accuracy, instantaneity and robustness. Simulation results have shown that the proposed method is superior to the percentage method in accuracy and robustness, and has a good clinical application. |