彩色多普勒血流系统中血流参数的算法分析

The Arithmetic Analysis of Flow Parameter in the Color Flow Mapping System

目的本文通过讨论处理多普勒信号的常用算法:自相关法,AR模型法,希望找到一种性能较好的算法。方法利用仿真数据以及取自人体的真实数据对算法进行仿真比较。结果通过几组仿真数据的比较可看出,二者的中心频率估计差别不大,但AR模型法的分辨率及方差性能明显优于自相关法。由于CFM计算时数据较短(通常为8～16个),因此用自相关法估计方差时,谱展宽现象严重。结论经典功率谱估计方法的方差性能较差,分辨率较低。

Objective To explore a better arithmetic of Doppler signal by discussing lag correlation method, AR model method. Methods Compare the two methods by simulation data and truthful data of human body. Results Center frequency of two methods doesn＇t have obvious difference, but resolving capability and skedasticity of AR model method was better than lag correlation methods. Because the data was shorter calculated by CFM （usual 8-16）, spectrum widening was badly estimated by lag correlation method. Conclusion Scedasticity of classic power spectrum estimation was bad,resolving capability was bad.