GSC框架下的最小均方超声波束形成算法

The Least Mean Square Ultrasonic Beamforming Algorithm Based on GSC

针对传统超声信号幅度变迹方法中主瓣宽度较宽、空间分辨率较低的问题,提出了一种基于广义旁瓣相消器框架下的最小均方超声波束形成算法。该方法基于期望信号最小方差无失真准则,构造广义旁瓣相消器,将接收到的超声信号分解为自适应与非自适应上下两个部分:上支路保留期望信号与噪声信号,且期望信号无失真响应约束得到保证;下支路阻塞期望信号,只含有噪声。将两路信号进行维纳滤波,上下支路噪声得到抵消,期望信号被无失真输出。为了使该算法在硬件上易于实现,采用最小均方算法自适应迭代求取,并给出了FPGA的详细设计过程。仿真实验表明,采用该算法加权的得到的波束与传统幅度变迹方法相比,主瓣更窄,具有抑制干扰和噪声的能力,提高了超声成像的横向分辨率与对比度分辨率。

In order to narrow the width of the main lobe and increase the spatial resolution of conventional ultrasonic apodization algorithm imaging method,a generalized sidelobe canceller under the framework of the Least Mean Square ultrasonic beamforming algorithm was proposed in this paper.The expected signal minimum variance distortionless criterion was adopted to design the generalized sidelobe canceller with which received ultrasonic signalsare decomposed into the adaptive parts(the upper branch)and the non-adaptive parts(the bottom branch).The distortionless desired signal and noise were reserved in the upper branch,and the noise was contained in the bottom branch.Wiener filtering method was used to filter the signals of the two parts by removing noise and exporting the distortionless desired signal.This algorithm was realized in hardware applied by adaptive iteration LMS algorithm,and the detailed FPGA implementation was designed.The simulation results depicted that the main lobe was narrower in this weighted apodization algorithm compared to the conventional amplitude apodization algorithm,and the noise and interference were suppressed to improve the lateral resolution and contrast resolution of ultrasound imaging.