论文部分内容阅读
研究表明,血液循环系统的发病率呈逐年上升的趋势,当血管壁因为某些因素发生病变时,靠近管壁附近的血流信号也会随之发生明显的变化,这种变化可由其时频分布谱表示出来[1][2]。2001年,Durka提出了基于随机字典的Matching Pursuit算法(MP算法)。本论文利用MP算法对血流信号的时频分布谱进行分析,并且从时频分布谱中提取出平均频率曲线。结果表明:用基于随机字典的MP算法对血流信号进行时频分布谱估计时,信号不同的分段数以及截取窗口,计算出的时频分布谱、平均频率曲线,以及同理论谱、曲线之间的均方误差均有所差异。
Studies have shown that the incidence of the blood circulatory system showed a trend of rising year by year, when the vascular wall lesions due to certain factors, near the wall of the blood flow signal will also be significant changes occur, this change can be its time-frequency The distribution spectrum is shown [1] [2]. In 2001 Durka proposed a Matching Pursuit algorithm (MP algorithm) based on a random dictionary. In this thesis, the MP algorithm is used to analyze the time-frequency distribution of blood flow signal, and the average frequency curve is extracted from the time-frequency distribution spectrum. The results show that when using the MP algorithm based on random dictionary to estimate the time-frequency spectrum of blood flow signal, the number of different segments and interception windows, the calculated time-frequency distribution spectrum, the average frequency curve and the same theory spectrum, the curve The mean square error between the different.