Research On Ultrasonic Testing Method Of Metal Internal Defects Based On Spatial Correlation Denoising And Semi-blind Image Restoration

Metal materials have a series of excellent properties and are widely used in high-end industrial fields such as aerospace,nuclear industry and national defense.At present,the requirements for in-service performance testing of metal materials are becoming more and more stringent.Various types of internal defects occur in metal materials during the process of casting,forging,welding,etc.These defects lead to the decline of material performance,and ultimately affect the safety and reliability of the equipment.Therefore,it is of great significance to research on the detection technology of internal defects of metal materials to ensure the safety and reliability of the equipment.Ultrasonic testing technology is a widely used non-destructive testing method in industry.Compared with other non-destructive testing methods,it is easier to operate,more sensitive and safer,and is suitable for many types of testing objects and testing scenes.Ultrasonic testing is mainly used to identify and analyze the defects inside materials,so as to evaluate whether the quality of the tested objects reaches the standard.Ultrasonic signal de-noising and ultrasonic image restoration are important research contents in the field of ultrasonic detection.In this paper,an ultrasonic testing method of signal de-noising based on spatial correlation and semi-blind image restoration based on parameterized PSF optimization is proposed.The main contents of this paper include the following four aspects:(1)The principle of ultrasonic testing method and ultrasonic imaging method are studied.The composition of the ultrasonic signal is analyzed,and an ultrasonic signal model is constructed.Combined with the analysis of the ultrasonic signal model and case,it is found that for coarse-grained metal materials,the ultrasonic signal contains random noise and strong structural noise,and the echo of the defect is easily covered by the noise,leading to the defect difficult to identify.The diffusion characteristics of the sound beam of the ultrasound probe and the process of ultrasonic C-scan imaging are studied.A measured C-scan image is analyzed and it is found that the ultrasonic C-scan images have the problems of low image resolution and blurred edge,and accurate size of the defect cannot be obtained from the original images.The original C-scan images need to be restored.(2)Aiming at the problem that the noise in the ultrasonic signal conceals the defect echo and makes it difficult to identify the defect,a threshold de-noising method of the ultrasonic signal based on CEEMDAN and spatial correlation is proposed.According to the high correlation between the defect echoes in the ultrasonic signals of adjacent detection positions,the defect echoes are identified and noise components are suppressed.By CEEMDAN,a pair of adjacent ultrasonic signals are decomposed into two sets of intrinsic mode functions(IMFs).A sliding window function is used to calculate the spatial correlation coefficient distribution function of the "parallel modes",and then the soft threshold processing is performed on the spatial correlation coefficient distribution function to calculate the reconstruction coefficient.Finally,a de-noised ultrasonic signal is obtained.Simulation signal analysis verifies the advantages and stability of the proposed method in noise reduction of ultrasonic signals.(3)Aiming at the problem of blurry and low resolution of ultrasonic C-scan images,a semi-blind restoration method of ultrasonic images based on Point Spreading Function(PSF)parameters optimization is proposed.The distribution of the sound field emitted by the ultrasonic probe based on the multivariate Gaussian beam model is studied,and then the parameterized representation of the PSF of the ultrasonic imaging system is given.The PSF multi-parameter optimization method based on the PSO algorithm is studied,to optimize the main parameters in the multivariate Gaussian beam model.The RL-TV algorithm and the idea of PSF parameter optimization are introduced into the framework of alternative minimization.Based on these studies,a semi-blind restoration method of ultrasonic images based on PSF parameters optimization is proposed.The simulation analysis shows that the defects in the restored image obtained by the proposed method are more complete and clear,and the image SNR is improved significantly,which verifies its effectiveness in ultrasonic image restoration.(4)The ultrasonic testing experiment of 304 stainless steel blocks is carried out to verify the proposed methods in this paper.The noise reduction result of the ultrasonic signals verifies the superiority and stability of the threshold de-noising method of the ultrasonic signal based on CEEMDAN and spatial correlation,which eliminates the noise in the ultrasonic signals more effectively,and retains the weak echo better.The restoration result C-scan images of the flatbottom holes verifies the effectiveness of the semi-blind restoration method of ultrasonic images based on PSF parameters optimization,which improves the clarity of the ultrasonic image.The accuracy of the defect size obtained from the restored result of the proposed method is improved significantly,which shows the superiority of the proposed method in ultrasonic image restoration and the quantitative analysis.