Multi-dimensional signal reconstruction-based circumferential weld ultrasonic C scanning image quality improvement method

Abstract

The invention discloses a circumferential weld ultrasonic C scanning image quality improvement method based on multi-dimensional signal reconstruction, and relates to the field of detection signal processing of ultrasonic nondestructive detection and imaging based on detection signals. Comprising the following steps: S1, performing four-layer wavelet packet decomposition on a detection signal to obtain a wavelet packet coefficient of a decomposed signal; s2, screening and reserving two specific wavelet packet coefficients in the fourth layer of wavelet packet decomposition signals; s3, reconstructing and normalizing the wavelet packet coefficients of the two reserved decomposition signals, and reconstructing a detection image based on the reconstructed signals; and S4, all the two reserved decomposition signals are subjected to superposition reconstruction after being normalized, and a C scanning image is reconstructed based on the superposed signals. The method has the advantages of strong anti-noise capability, good reservation of detail signals, high image reconstruction precision and the like.

Description

Technical Field

[0001] The present invention mainly relates to the field of girth weld ultrasonic detection signal processing, and in particular to a girth weld ultrasonic C-scan image quality improvement method based on multi-dimensional signal reconstruction. Background Art

[0002] With the rapid development of modern industry, the quality requirements for parts produced in industrial manufacturing are getting higher and higher. Welding technology, as a commonly used technical means in industrial applications, is widely used in various industrial fields. The girth weld structure produced by welding is prone to defects such as lack of fusion and pores. In order to ensure its welding quality, the weld structure needs to be non-destructively tested and evaluated. At present, ultrasonic testing technology is the main means to evaluate the quality of girth welds, but for tiny defects in the girth weld structure, the ultrasonic detection echo signal is weak and easily aliased with the high-frequency electrical noise of the system, resulting in reduced detection resolution. Wavelet packet decomposition is a signal analysis method based on the concept of wavelet transform, but provides a more flexible and comprehensive signal frequency domain characteristic analysis. It has the characteristics of decomposing the signal at any time-frequency resolution. After decomposition, the signal in the required frequency band can be reconstructed as needed, which plays a filtering role and can accurately analyze the signal locally. Therefore, a girth weld ultrasonic C-scan image quality improvement method based on multi-dimensional signal reconstruction is proposed, which effectively improves the detection resolution. Summary of the invention

[0003] In view of the technical problems existing in the prior art, the present invention proposes a method for improving the quality of circumferential weld ultrasonic C-scan images based on multi-dimensional signal reconstruction, which is safe, reliable, accurate and efficient.

[0004] In order to solve the above technical problems, the technical solution proposed by the present invention is:

[0005] The method for improving the quality of girth weld ultrasonic C-scan images based on multi-dimensional signal reconstruction comprises the following steps:

[0006] S1, performing four-layer wavelet packet decomposition on the detection signal to obtain the wavelet packet coefficients of the decomposed signal;

[0007] S2, screening and retaining two specific wavelet packet coefficients in the fourth layer wavelet packet decomposition signal;

[0008] S3, reconstructing and normalizing the wavelet packet coefficients of the two retained decomposition signals respectively, and then reconstructing the detection image based on the reconstructed signals;

[0009] S4. All two retained decomposed signals are normalized and then superimposed and reconstructed, and a C-scan image is reconstructed based on the superimposed signals.

[0010] Preferably, the specific process of step S1 is:

[0011] (1) An ultrasonic probe with a center frequency of 15 MHz is used to extract the original detection signal.

[0012] (2) Using the Sym8 wavelet basis, after completing the first layer of decomposition, only the high-frequency signal of the first layer of decomposition is continuously decomposed to the fourth layer to obtain the decomposed wavelet coefficients of 8 detection signals.

[0013] Preferably, in step S2, the second decomposition signal wavelet coefficient and the fourth decomposition signal wavelet coefficient of the fourth layer wavelet packet decomposition are retained according to the obtained wavelet packet decomposition signal, and the frequency ranges of the reconstructed signals are 5-15 MHz and 15-25 MHz respectively.

[0014] Preferably, the specific process of step S3 is:

[0015] The wavelet packet coefficients of the two retained decomposed signals are reconstructed into two time domain reconstructed signals and normalized respectively, and then the signal pulse amplitude is extracted near the arrival time of the weld bottom echo for C-scan imaging.

[0016] Preferably, the specific process of step S4 is:

[0017] The reconstructed time-domain decomposition signal is normalized and then superimposed into a time-domain reconstructed signal. The signal pulse amplitude is extracted near the arrival time of the weld bottom echo for C-scan imaging.

[0018] The present invention has the advantages of strong anti-noise capability, good detail signal preservation, high image reconstruction accuracy, and the like. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a flow chart of the present invention;

[0020] Figure 2 This is a girth weld inspection sample diagram of the present invention;

[0021] Figure 3 It is the original ultrasonic C-scanning image diagram of the girth weld detection signal of the present invention, wherein 1 to 5 are respectively a φ0.4 mm flat bottom hole defect, a φ0.3 mm flat bottom hole defect, a φ0.2 mm flat bottom hole defect, a prefabricated natural defect and a girth weld tip;

[0022] Figure 4 It is the detection signal of the φ0.4 mm flat bottom hole defect in the original ultrasonic C-scan imaging image of the girth weld of the present invention;

[0023] Figure 5 The detection signal of the tip of the girth weld in the original ultrasonic C-scan imaging image of the girth weld of the present invention;

[0024] Figure 6 It is a detection signal of the intact weld position in the original ultrasonic C-scan imaging image of the girth weld of the present invention;

[0025] Figure 7 It is a detection signal of a prefabricated natural defect in the original ultrasonic C-scan imaging image of the girth weld of the present invention;

[0026] Figure 8 It is a reconstructed signal of the wavelet packet coefficients of the φ0.4 mm flat bottom hole defect in the original ultrasonic C-scan imaging image of the girth weld of the present invention with a frequency spectrum range of 5 to 15 MHz;

[0027] Fig. 9 It is a reconstructed signal of the wavelet packet coefficients with a frequency spectrum range of 5 to 15 MHz at the tip of the girth weld in the original ultrasonic C-scan imaging image of the girth weld of the present invention;

[0028] Fig.10 It is a reconstructed signal of wavelet packet coefficients with a frequency spectrum range of 5 to 15 MHz at the intact weld position in the original ultrasonic C-scan imaging image of the girth weld of the present invention;

[0029] Fig.11 It is a reconstructed signal of wavelet packet coefficients with a frequency spectrum range of 5 to 15 MHz of the prefabricated natural defects in the original ultrasonic C-scan imaging image of the girth weld of the present invention;

[0030] Fig.12 It is a reconstructed signal of the wavelet packet coefficient of 15-25MHz in the frequency spectrum range of the φ0.4mm flat-bottom hole defect in the original ultrasonic C-scan imaging image of the girth weld of the present invention;

[0031] Fig.13 It is a reconstructed signal of the wavelet packet coefficients with a frequency spectrum range of 15 to 25 MHz at the tip of the girth weld in the original ultrasonic C-scan imaging image of the girth weld of the present invention;

[0032] Fig.14 It is a reconstructed signal of wavelet packet coefficients with a frequency spectrum range of 15 to 25 MHz at the intact weld position in the original ultrasonic C-scan imaging image of the girth weld of the present invention;

[0033] Fig.15 The spectrum range of the prefabricated natural defect in the original ultrasonic C-scan imaging image of the girth weld of the present invention is a reconstructed signal of the wavelet packet coefficient of 15 to 25 MHz;

[0034] Fig.16 The C-scan image reconstructed based on the reconstructed signal of the wavelet packet coefficients in the frequency spectrum range of 5 to 15 MHz of the present invention;

[0035] Fig.17 The C-scan image reconstructed based on the reconstructed signal of the wavelet packet coefficients in the frequency spectrum range of 15 to 25 MHz of the present invention;

[0036] Fig.18The reconstructed signal is obtained by normalizing and superimposing the two retained decomposition signals of the φ0.4 mm flat-bottom hole defect of the present invention;

[0037] Fig.19 The reconstructed signal is a reconstructed signal obtained by superimposing and normalizing the two decomposed signals retained at the tip of the girth weld of the present invention;

[0038] Fig. 20 The reconstructed signal is obtained by normalizing and superimposing the two decomposed signals retained at the intact weld position of the present invention;

[0039] Fig.21 The reconstructed signal is a reconstructed signal obtained by normalizing and superimposing the two decomposed signals retained by the prefabricated natural defects of the present invention;

[0040] Fig. 22 It is a C-scan detection image reconstructed based on the reconstructed signal after superposition reconstruction. DETAILED DESCRIPTION

[0041] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments.

[0042] like Figure 1 As shown, a method for improving the quality of girth weld ultrasonic C-scan images based on multi-dimensional signal reconstruction according to an embodiment of the present invention comprises the following steps:

[0043] S1, performing four-layer wavelet packet decomposition on the detection signal to obtain the wavelet packet coefficients of the decomposed signal;

[0044] S2, screening and retaining two specific wavelet packet coefficients in the fourth layer wavelet packet decomposition signal;

[0045] S3, reconstructing and normalizing the wavelet packet coefficients of the two retained decomposition signals respectively, and then reconstructing the detection image based on the reconstructed signals;

[0046] S4. All two retained decomposed signals are normalized and then superimposed and reconstructed, and a C-scan image is reconstructed based on the superimposed signals.

[0047] (1) Girth weld specimen structure Figure 2 As shown in the figure, the penetration depth of the qualified girth weld is about 2mm. Flat-bottom holes with a depth of 2.0mm and diameters of 0.4mm, 0.3mm, and 0.2mm are processed on the weld fusion line to simulate weld defects. In addition, a slag defect is pre-buried in the center of the weld during the welding process. The water distance is adjusted so that the ultrasonic sound beam with a center frequency of 15MHz forms a focused sound field at a depth of 2mm in the weld. During the C-scan movement of the girth weld, the A-scan detection signal is extracted point by point, and the peak value of the detection signal in the time domain range of 3.24 to 3.72μs (depth range: 1.2mm to 2.6mm) is extracted, and a C-scan image is formed based on the pulse peak value in this range, as shown in the figure. Figure 3 shown.

[0048] (2) Use Sym8 wavelet basis to perform four-layer wavelet packet decomposition on the detection signal to obtain the wavelet packet decomposition coefficients of each decomposition layer of the detection signal; select two decomposition coefficients with a spectrum range of 5-15MHz and 15-25MHz to reconstruct the detection signal, and obtain two decomposition signals with a spectrum range of 5-15MHz and 15-25MHz. Select the detection signals of φ0.4mm flat-bottom holes, embedded defects, and fusion lines, and perform wavelet packet decomposition and decomposition signal reconstruction according to the above method. The detection signal and the reconstructed signal are as follows: Figures 4 to 15 As shown in the figure, the characteristic signal echo amplitudes of the detection signals on the φ0.4mm flat-bottom hole, embedded defect and fusion line are 259.8mV, 785mV and 121.6mV, respectively, and the noise signal amplitudes are 39.3mV, 58.9mV and 54.3mV, respectively; the characteristic signal echo amplitudes of the reconstructed decomposition signals in the frequency range of 5-15MHz are 243.7mV, 420.6mV and 103.1mV, respectively, and the noise signal amplitudes are 19.5mV, 18.1mV and 29.2mV, respectively; the characteristic signal echo amplitudes of the reconstructed decomposition signals in the frequency range of 15-25MHz are 136.5mV, 420.6mV and 37.6mV, respectively, and the noise signal amplitudes are 11.1mV, 18.1mV and 12.05mV, respectively. The signal-to-noise ratio of the reconstructed decomposed signal in the frequency range of 5-15MHz on the φ0.4mm flat-bottom hole, embedded defects and fusion line is increased by 1.89, 1.58 and 1.75 times compared with the signal-to-noise ratio of the original signal; the signal-to-noise ratio of the reconstructed decomposed signal in the frequency range of 15-25MHz on the φ0.4mm flat-bottom hole, embedded defects and fusion line is increased by 1.86, 1.75 and 1.39 times compared with the signal-to-noise ratio of the original signal.

[0049] (3) The wavelet packet coefficients of the two retained decomposition signals are reconstructed and normalized respectively, and the C-scan image is reconstructed based on the peak value in the 3.24-3.72μs range (depth range: 1.2mm-2.6mm) of the decomposition signal. Based on the decomposition signal in the frequency range of 5-15MHz, the signal-to-noise ratio of the reconstructed detection image is improved, and it can more clearly display the φ0.2-0.4mm flat-bottom hole defects, prefabricated natural defects, and the tip of the girth weld; in particular, the tip of the girth weld is enhanced; based on the decomposition signal in the frequency range of 15-25MHz, the signal-to-noise ratio of the reconstructed detection image is improved, and it can display the φ0.2-0.4mm flat-bottom hole defects, prefabricated natural defects, and the tip of the girth weld; in particular, the tip of the girth weld is enhanced; Fig.16 and Fig.17 shown.

[0050] (4) The two retained decomposition signals are normalized and then superimposed and reconstructed. The reconstructed signal is as follows: Figure 18 to Figure 21 As shown in the figure, the characteristic signal echo amplitudes of the reconstructed signals on the φ0.4mm flat-bottom hole, embedded defects, and fusion line are 0.13mV, 0.35mV, and 0.04mV, respectively, and the noise signal amplitudes are 0.01mV, 0.017mV, and 0.011mV, respectively. The signal-to-noise ratio is improved by 1.97, 1.63, and 1.55 times compared with the signal-to-noise ratio of the original signal. Based on the peak value reconstruction of the superimposed signal within the range of 3.24 to 3.72μs (depth range: 1.2mm to 2.6mm), the reconstructed C-scan image can display φ0.2 to 0.4mm flat-bottom hole defects, prefabricated natural defects, and girth weld tips; in particular, the girth weld tip is enhanced, and the quality of the reconstructed image is significantly improved compared with the original image, as shown in the figure. Fig. 22 shown.

Claims

1. A method for improving the quality of girth weld ultrasonic C-scan images based on multi-dimensional signal reconstruction, characterized in that: The following steps are involved: S1, performing four-layer wavelet packet decomposition on the detection signal to obtain the wavelet packet coefficients of the decomposed signal; S2, screening and retaining two specific wavelet packet coefficients in the fourth layer wavelet packet decomposition signal; S3, reconstructing and normalizing the wavelet packet coefficients of the two retained decomposition signals respectively, and then reconstructing the detection image based on the reconstructed signals; S4. All two retained decomposed signals are normalized and then superimposed and reconstructed, and a C-scan image is reconstructed based on the superimposed signals.

2. The method for improving the quality of girth weld ultrasonic C-scan images based on multi-dimensional signal reconstruction according to claim 1, characterized in that: The detection signal in step (1) is extracted by an ultrasonic probe with a center frequency of 15 MHz.

3. The method for improving the quality of girth weld ultrasonic C-scan images based on multi-dimensional signal reconstruction according to claim 1 or 2, characterized in that: In step (1), the specific method of wavelet packet decomposition is: The Sym8 wavelet basis is used. After the first layer of decomposition is completed, only the low-frequency signal of the first layer of decomposition is continuously decomposed to the fourth layer to obtain the decomposed wavelet coefficients of 8 detection signals.

4. The method for improving the quality of girth weld ultrasonic C-scan images based on multi-dimensional signal reconstruction according to claim 1, characterized in that: The specific method of step (2) is: The second decomposition signal wavelet coefficient and the fourth decomposition signal wavelet coefficient of the fourth layer wavelet packet decomposition are retained, and the frequency ranges of the reconstructed signals are 5-15MHz and 15-25MHz respectively.

5. The method for improving the quality of girth weld ultrasonic C-scan images based on multi-dimensional signal reconstruction according to claim 1, characterized in that: The specific method of step (3) is: The wavelet packet coefficients of the two retained decomposed signals are reconstructed into two time domain reconstructed signals and normalized respectively, and then the signal pulse amplitude is extracted near the arrival time of the weld bottom echo for C-scan imaging.

6. The method for improving the quality of girth weld ultrasonic C-scan images based on multi-dimensional signal reconstruction according to claim 1, characterized in that: The specific method of step (4) is: The reconstructed time-domain decomposition signal is normalized and then superimposed into a time-domain reconstructed signal. The signal pulse amplitude is extracted near the arrival time of the weld bottom echo for C-scan imaging.

Images