北京生物医学工程

基于FPGA的超声射频信号自适应滤波器

A field programmable gate array-based adaptive filter for ultrasound RF signals

作者：林一聪郑政赵智慧张彦

单位：上海理工大学医疗器械与食品学院（上海 200093）

关键词：医学超声；射频信号；滤波；自适应滤波器；现场可编程门阵列；延时最小均方误差算法

分类号：

出版年·卷·期（页码）：2012·31·1（21-25）

摘要：

目的医学超声的射频回波信号具有明显的时变特性，传统滤波器难以获得理想的滤波效果，本文提出一种基于现场可编程门阵列(field programmable gate array，FPGA)的实时自适应滤波器以解决这一问题。方法根据双通道自适应滤波器原理，设计双脉冲激励以获得双通道输入，滤波器集成在一个Altera Cyclone II FPGA上，采用延时最小均方误差(delayed least-mean-square，DLMS)控制算法以提高速度。结果对超声体模和人体组织回波进行的仿真及硬件滤波实验表明，该滤波器具有明显的信号增强作用，对不同深度信号的处理效果没有明显的差别，并且可以达到50MS/s的数据吞吐量，满足实时处理要求。结论 FPGA速度快，体积小，耗电低，在超声扫描仪中应用广泛，本文提出的基于FPGA超声射频信号的自适应滤波方案滤波效果明显，具有很强的实用性。

Objective The medical ultrasonic radio frequency(RF)echo signals are apparently time-varying，can not be filtered effectively with the traditional fixed-band filters. This paper describes a two-channel adaptive filter based on the field programmable gate array(FPGA)to resolve this problem. Methods A double-pulse excitation method is designed to retrieve the two inputs which are required by the adaptive filter. The filter is integrated on an Altera Cyclone II FPGA，and the delayed least-mean-square(DLMS) control algorithm is used to increase the speed of process. Results Simulation and hardware filtering experiments are conducted on the echo signals in ultrasound phantom and human tissues，the results show that the adaptive filter enhances the signal significantly，and no difference is observed among echoes from different depths. The throughput rate of the filter is up to 50 MS/s，which meets the requirements of real-time processing. Conclusions The proposed adaptive filter based on the FPGA is an excellent solution to ultrasound echo signal filtering and is easy to be widely used in ultrasound scanners because of FPGA with the features of high speed，small size，and low power consumption.

参考文献：

［1］Flax SW，Pele NJ，Clover GH et al．Spectral characterization and attenuation measurements in ultrasound［J］．Ultrasonic Imaging，1983,5(2):95-116．
［2］Brenner AR，Eck K，Wilhelm W. Improved resolution and dynamic range in medical ultrasonic imaging using depth-dependent mismatched filtering［J］.Ultrasonics Symposium，1997，2:1475-1480.
［3］Yi Y， Woods R. High speed FPGA-based implementations of DLMS filter［J］. The Journal of VLSI Signal Processing，2005，39(1-2):113-131.
［4］杨荣峰，魏义祥.改进的非线性自适应滤波器用于ECG信号去噪［J］.北京生物医学工程,2007,26(1):57-59.
［5］许宏吉，彭玉华，刘琚.基于DSP的雷达信号自适应滤波器［J］.计算机工程，2004，30(5):173-191.
［6］Ferrara ER Jr，Widrow B. Multichannel adaptive filtering for signal enhancement［J］. IEEE Transactions on Acoustics，Speech and Signal Processing，1981，29(3):766-770.
［7］Chan FHY，Lam FK. Detection of brainstem auditory evoked potential by adaptive filtering［J］. Medical and Biological Engineering and Computing,1995, 33(1): 69-75.
［8］Li DD. The Enhancement of Pulse-echo Ultrasound Signal Using Adaptive Filter with NLMS，BMEI，2010.
［9］Long G. The LMS algorithm with delayed coefficient adaptation［J］. Acoustics，Speech and Signal Processing，1989，37(9):1397-1405.
［10］杨东，王建业.DLMS高速自适应滤波器的FPGA实现［J］.电声技术，2010，34(5)：41-43
［11］陈文博，耿相铭，马伟敏.高速自适应DLMS算法及其硬件实现［J］.信息技术，2007(10):59-62.

服务与反馈：

【文章下载】【加入收藏】

提示：您还未登录，请登录！点此登录