A SVD-Based Signal De-Noising Method With Fitting Threshold for EMAT

The electromagnetic acoustic transducer (EMAT) is a powerful and useful non-destructive testing technology for structural health monitoring. However, EMAT has an issue of low efficiency in conversion its signal easily affected by noise, which make it difficult to accurately identify evaluate defects. Thereby, de-noising preprocessing essential the evaluation In this paper, we proposed improved singular value decomposition (SVD) method based on fitting threshold signal. For SVD method, key point determine reconstructing We applied segmented regression model find appropriate approach. To investigate efficacy simulation signals experimental are used verification respectively. A comparative analysis been under-taken confirm that advantages over other methods noise reduction, more accurate