An ultrasonic signal sparse decomposition method and a signal noise reduction and defect detection method thereof

Abstract

An ultrasonic signal sparse decomposition method and a signal noise reduction and defect detection method are disclosed, relating to the field of ultrasonic nondestructive testing. The method includes: step 1, adopting a Gabor model to approximate fit the ultrasonic signal, selecting a Gabor dictionary parameter set, combining to form dictionary matrix, and utilizing sparse Bayesian learning algorithm to solve the optimal approximate solution of coefficient vector under sparse constraint condition. The number of defects is the number of non-zero terms in the coefficient vector. The atomic vector parameters are used to eliminate the corresponding noise terms in the coefficient vector and the energy threshold is set to further denoise. The invention can realize noise reduction of ultrasonicnondestructive testing signal containing complex noise, recovery of defect reflected signal and defect detection and location with low signal-to-noise ratio.

Description

technical field [0001] The invention relates to the field of ultrasonic nondestructive testing, and relates to a sparse decomposition method of ultrasonic nondestructive testing signals and a noise reduction and defect detection method. Background technique [0002] In the field of civil engineering, ultrasonic nondestructive testing has been widely used in structural damage detection of building structures, bridges, pipelines and so on. With the help of ultrasonic non-destructive testing technology, it is possible to effectively detect whether there is a defect or damage in the structure, locate the defect and measure the size of the defect, so as to evaluate the safety status of the structure in time. Compared with the overall monitoring and a small number of local monitoring technologies of structural health monitoring, ultrasonic nondestructive testing technology can better detect local micro damage of structures, and is a useful supplement to structural health monitorin...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to solve the problem that the dictionary design method and the signal sparse decomposition algorithm existing in the ultrasonic nondestructive testing signal processing of the existing signal sparse representation are not perfect enough, which leads to the inaccurate results of the ultrasonic signal analysis, and proposes a Sparse Bayesian learning (SBL)-based signal sparse decomposition, recovery and defect detection method, suitable for contact pulse reflection ultrasonic testing

Images