基于自适应滤波的TOFD D扫描图像侧向波抑制

Direct Wave Suppression of TOFD D Scanning Image Based on Adaptive Filtering

针对超声波时差衍射法(TOFD)D扫描图像,提出一种新的侧向波抑制算法。该方法首先将图像分解为一维信号,对一维信号进行小波去噪后重构以去除图像噪声。通过讨论参数对滤波结果的影响来比较RLS自适应滤波算法和变步长LMS自适应滤波算法,从而选择对参数不敏感的变步长LMS算法。最后在变步长LMS算法的基础上提出一种改进的自适应滤波方法,以去除侧向波与缺陷波未混叠的缺陷图像的残余部分。试验结果表明提出的算法是有效的。

In order to be able to suppress direct wave of TOFD （time of flight diffraction） D scanning image, a new method was introduced. In order to remove noise, firstly the TOFD D scanning image was decomposed into a series of one-dimensional signals. Then wavelet denoising was used to remove noise of one-dimensional signal. Finally, with a series of processed signals a new image was reconstructed. The effect of parameters on the filtering results would he discussed for the comparison of RLS （recursive least square） adaptive filtering algorithm and variable stepsize LMS （least mean square） adaptive filtering algorithm, and variable step-size LMS adaptive filtering algorithm would be used. Moreover, an improved algorithm was introduced to suppress direct wave of a non-near surface defect image. The result demonstrated that the algorithm was available.