Lopez Villaverde Eduardo, Robert Sébastien, Prada Claire
CEA, LIST, F-91191 Gif-sur-Yvette, France.
Institut Langevin, ESPCI Paris Tech, CNRS, PSL Research University, 1 rue Jussieu, 75238 Paris, France.
J Acoust Soc Am. 2016 Jul;140(1):541. doi: 10.1121/1.4958683.
In the present work, the Synthetic Transmit Aperture (STA) imaging is combined with the Decomposition of the Time Reversal Operator (DORT) method to image a coarse grained austenitic-ferritic steel using a contact transducer array. The highly heterogeneous structure of this material produces a strong scattering noise in ultrasound images. Furthermore, the surface waves guided along the array interfere with the bulk waves backscattered by defects. In order to overcome these problems, the DORT method is applied before calculating images with the STA algorithm. The method consists in analyzing in the frequency domain the singular values and singular vectors of the full array transfer matrix. This paper first presents an analysis of the singular values of different waves contained in the data acquisition, which facilitates the identification of the subspace associated with the surface guided waves for filtering operations. Then, a filtered matrix is defined where the contribution of structural noise and guided waves are reduced. Finally, in the time domain, the STA algorithm is applied to this matrix in order to calculate an image with reduced structural noise. Experiments demonstrate that this filtering improves the signal-to-noise ratio by more than 12 dB in comparison with the STA image before filtering.
在本研究中，合成发射孔径（STA）成像与时间反转算子分解（DORT）方法相结合，使用接触式换能器阵列对粗晶奥氏体 - 铁素体钢进行成像。这种材料高度不均匀的结构在超声图像中产生强烈的散射噪声。此外，沿阵列传播的表面波会干扰由缺陷背散射的体波。为了克服这些问题，在使用STA算法计算图像之前应用DORT方法。该方法包括在频域中分析全阵列传输矩阵的奇异值和奇异向量。本文首先对数据采集中包含的不同波的奇异值进行分析，这有助于识别与表面导波相关的子空间以进行滤波操作。然后，定义一个滤波矩阵，其中结构噪声和导波的贡献被降低。最后，在时域中，将STA算法应用于该矩阵以计算具有降低结构噪声的图像。实验表明，与滤波前的STA图像相比，这种滤波将信噪比提高了超过12 dB。
