Application of basis pursuit in signal denoising of ultrasonic testing flaw

In ultrasonic nondestructive testing of weld defects, the background noise generated by the grain boundary and the micro structure of the material and the electric noise generated by the detecting instrument can interfere with the detection of the defect. Those even submerge defect signals, so that it is difficult to accurately extract the defect signal. Therefore, it is needed to use advanced signal processing tools to deal with the defect signals. The de-noising method based on sparse decomposition is superior to the extraction of ultrasonic flaw signals. In this paper, we propose a method Basis Pursuit (BP) algorithm to extract the ultrasonic defect signal. In this method, the best atoms matching the signal are found from the over complete Gabor atom library, so the defect signal can be recovered accurately and the noise can be filtered out. Basis Pursuit compared with the wavelet de-noising algorithm. Experimental results show that the proposed method has higher signal to noise ratio and can be more accurate than the wavelet algorithm.