基于S变换和时频谱空间滤波的超声缺陷回波检测方法

A flaw echo detection method based on S-transformation and time-frequency spectrum spatial filtering

缺陷回波的检测是超声探伤的一项重要内容,为减弱噪声的影响准确检测缺陷回波,提出基于S变换时频分析和时频谱空间滤波的信号处理方法。讨论了高斯回波模型下的到达时间和中心频率与S变换时频谱的关系,说明了利用S变换时频谱幅值矩阵的极值提取回波到达时间和中心频率的合理性;为检测回波,首先对原始信号作S变换,然后对得到的时频谱幅值矩阵应用最大熵法自适应选择去噪阈值,对S变换时频谱作空间滤波完成降噪;从降噪后的区域中提取反映缺陷的到达时间和中心频率;对降噪后的时频谱作S逆变换,获得缺陷回波明显的时域信号。仿真研究表明,基于S变换和时频谱空间滤波的方法能够有效去除噪声,检测回波。棒材试块的实验结果同样表明了该方法在缺陷检测上的有效性。

Detection of flaw echoes is an important task in ultrasonic testing. In order to eliminate the effects of noise and detect echoes correctly,a signal processing method based on the S-transformation（ ST） and the time-frequency spectrum（ TFS） spatial filtering was proposed. Based on the Gaussian echo model,the relationships between the time of arrival（ TOA） together with the center frequency（ CF） and ST TFS were discussed. The reasonableness of the method using the extremum of ST TFS amplitude matrix for TOA and CF extraction were illustrated. To detect the echoes,firstly the ST was performed on the original signal. Then the dual-threshold were determined by the Maximum Entropy Thresholding method adaptively,and the spatial filtering was performed on ST TFS for denoising. The TOA and CF could be extracted from the denoised region. Finally,the inverse ST was performed on the denoised TFS and the signal with clear flaw echoes were gained. Simulation results show that using ST and TFS spatial filtering can remove noise and detect echoes effectively. And the experimental results of bar specimen also show the effectiveness of the method in flaw detection.