基于自回归谱外推的全聚焦成像分辨力提升

针对相邻缺陷全聚焦超声成像混叠问题，结合低阶、宽有效频带自回归谱外推方法，压缩超声波时域脉冲宽度，实现亚波长级全聚焦（Total focusing method，TFM）成像分辨力。建立碳钢试块模型，设置两个中心间距1.8 mm，直径1.3 mm圆孔，选用中心频率2.25 MHz，32阵元相控阵探头采集全矩阵数据。针对全矩阵数据，选择自回归阶数为2，信号频谱最大幅值下降14 dB为有效频带，建立自回归模型并外推有效频带外的高频与低频成分，随后对全矩阵数据进行延迟叠加处理和TFM成像。仿真结果表明，低阶、宽有效频带自回归谱外推处理方法具有较高的鲁棒性和准确性，TFM成像后可有效分离中心间距0.7λ（λ为超声波长）圆孔，保留缺陷横向位置信息的同时，定位误差不超过0.73%。对碳钢试块中相同位置及尺寸的圆孔进行试验验证，定位误差不超过1.06%，有效地提高TFM成像分辨力。

Resolution Enhancement of Total Focusing Imaging Combining with Autoregressive Spectral Extrapolation

The defects with close spacing are difficult to be distinguished by ultrasonic imaging with total focusing method(TFM). Autoregressive spectrum extrapolation with low-order and wide effective frequency band is introduced to compress the pulse width in time domain, achieving subwavelength resolution imaging of TFM. Two preset side-drilled holes having 1.8 mm center distance and 1.3 mm diameter in the simulated model of carbon steel are detected by array probe with 2.25 MHz center frequency and 32 elements. Focusing on full matrix capture, taking 2-order and the width of frequency window represented a 14 dB drop from the maximum amplitude of spectrum, and the autoregressive model is established to extrapolate the high-frequency and low-frequency components beyond effective frequency band. Then, the delay-and-sum algorithm is implemented on the processed full matrix capture data to obtain the TFM image. The simulated results indicated that the combination of TFM and autoregressive spectrum extrapolation with low-order and wide effective frequency band had high robustness and accuracy, so as to effectively separate adjacent defects with center distance of 0.7λ (λ is ultrasonic wavelength). The positioning error is no more than 0.73% while retaining the lateral resolution. Finally, side-drilled holes having 1.8 mm center distance are calculated with the same testing parameters in experiment, and the positioning error is no more than 1.06%, effectively improving imaging resolution of TFM.