Rosnitskiy Pavel B, Sapozhnikov Oleg A, Khokhlova Vera A, Kreider Wayne, Tsysar Sergey A, Thomas Gilles P L, Contreras Kaizer, Khokhlova Tatiana D
IEEE Trans Ultrason Ferroelectr Freq Control. 2025 May;72(5):564-580. doi: 10.1109/TUFFC.2025.3542405. Epub 2025 May 7.
Transient acoustic holography is a useful technique for characterization of ultrasound transducers. It involves hydrophone measurements of the 2-D distribution of acoustic pressure waveforms in a transverse plane in front of the transducer-a hologram-and subsequent numerical forward projection (FP) or backward projection of the ultrasound field. This approach enables full spatiotemporal reconstruction of the acoustic field, including the vibrational velocity at the transducer surface. This allows identification of transducer defects as well as structural details of the radiated acoustic field such as sidelobes and hot spots. However, numerical projections may be time-consuming ( $10^{{10}}$ - $10^{{11}}$ operations with complex exponents). Moreover, backprojection from the measurement plane to the transducer surface is sensitive to misalignment between the axes of the positioning system and the axes associated with the transducer. This article presents an open-access transducer characterization toolbox for use in MATLAB or Octave on Windows computers (https://github.com/pavrosni/xDDx/releases). The core algorithm is based on the Rayleigh integral implemented in C++ executables for graphics and central processing units (GPUs and CPUs). The toolbox includes an automated procedure for correcting axes misalignments to optimize the visualization of transducer surface vibrations. Beyond using measured holograms, the toolbox can also simulate the fields radiated by user-defined transducers. Measurements from two focused 1.25-MHz 12-element sector transducers (apertures of 87 mm and focal distances of 65 and 87 mm) were used with the toolbox for demonstration purposes. Simulation speed tests for different computational devices showed a range of 0.2 s-3 min for GPUs and 1.6 s-57 min for CPUs.
瞬态声全息术是一种用于超声换能器表征的有用技术。它涉及在换能器前方横向平面内对声压波形的二维分布进行水听器测量——即全息图——以及随后对超声场进行数值正向投影(FP)或反向投影。这种方法能够对声场进行完整的时空重建,包括换能器表面的振动速度。这有助于识别换能器缺陷以及辐射声场的结构细节,如旁瓣和热点。然而,数值投影可能很耗时(涉及复指数的操作次数为(10^{10}) - (10^{11})次)。此外,从测量平面到换能器表面的反向投影对定位系统的轴与换能器相关轴之间的未对准很敏感。本文介绍了一个可在Windows计算机上的MATLAB或Octave中使用的开放获取换能器表征工具箱(https://github.com/pavrosni/xDDx/releases)。核心算法基于在C++可执行文件中为图形处理器和中央处理器(GPU和CPU)实现的瑞利积分。该工具箱包括一个用于校正轴未对准以优化换能器表面振动可视化的自动化程序。除了使用测量的全息图外,该工具箱还可以模拟用户定义换能器辐射的场。为了演示目的,使用了两个聚焦的1.25 MHz 12元件扇形换能器(孔径为87 mm,焦距分别为65 mm和87 mm)的测量数据与该工具箱一起进行测试。针对不同计算设备的模拟速度测试表明,GPU的模拟时间范围为0.2秒至3分钟,CPU的模拟时间范围为1.6秒至57分钟。