Cable lead sealing internal defect positioning method and system based on multi-view TFM

Abstract

The invention provides a cable lead sealing internal defect positioning method based on a multi-view TFM. The method comprises the following steps: acquiring an original ultrasonic signal when viewpoint scanning is carried out on a cable lead sealing every time; processing the original ultrasonic signal into a signal intensity distribution image by using a full-focus algorithm, and performing three-dimensional conversion on the signal intensity distribution image; simulating propagation paths and scattering intensity of ultrasonic waves in various defects to calculate theoretical reflection intensity data of various defects; obtaining effective actual measurement signals of the defects and extracting related data; a Bayesian joint probability distribution function is constructed for estimation, and defect parameters are obtained in combination with theoretical reflection intensity data and extraction data of defects; a three-dimensional model of cable lead sealing is simulated, and automatic defect positioning and marking are carried out according to the original ultrasonic signals and defect parameters. By implementing the method, the defects existing in the aspects of acoustics, signal processing, algorithm models, defect classification and the like in an existing mode can be overcome, and the detection probability and depth measurement precision of the tiny pores are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power cable detection, and particularly to a method and system for locating internal defects of cable sealing lead based on multi-viewpoint TFM (Total Focusing Method). Background Art

[0002] In the field of power cable accessory sealing detection, as the core component for waterproof sealing of cross-linked polyethylene insulated joints, the accurate detection of minute internal defects of cable sealing lead is directly related to the long-term reliability of cable lines. Therefore, it is necessary to accurately locate the defects of cable sealing lead.

[0003] Currently, the ultrasonic detection method based on single-viewpoint TFM is used to locate the defects of cable sealing lead, mainly by analyzing the echo flight time and amplitude characteristics to achieve defect location.

[0004] However, the above detection method has the following deficiencies: (1) At the acoustic property level, the anisotropy caused by cold working hardening of the sealing lead layer results in uneven sound velocity distribution, and the single-probe incident angle design used in single-viewpoint TFM is difficult to effectively compensate for the sound beam refraction distortion; (2) At the signal processing level, the defect determination method based on time-domain threshold has a high missed detection rate for micro-cracks in low signal-to-noise ratio signals; (3) At the algorithm model level, the mode coupling effect of surface waves and shear waves is not fully considered, resulting in obvious limitations in the axial resolution of near-surface defects; in particular, the detection probability of sub-millimeter pores in the sealing lead layer is significantly insufficient, and there are significant errors in defect depth measurement; (4) At the defect classification level, defect classification depends on empirical thresholds, and there are large deviations in the orientation judgment of irregular cracks, which cannot meet the quantitative rating requirements of industry standards for sealing defects.

[0005] Therefore, it is necessary to improve the existing method of using the ultrasonic detection method based on single-viewpoint TFM to locate the defects of cable sealing lead, which can solve the defects existing in the existing method in terms of acoustics, signal processing, algorithm model, and defect classification of cables, and significantly improve the detection probability of minute pores and the depth measurement accuracy. Summary of the Invention

[0006] The technical problem to be solved by the embodiments of the present invention is to provide a method and system for locating internal defects of cable sealing lead based on multi-viewpoint TFM, which can solve the defects existing in the existing method in terms of acoustics, signal processing, algorithm model, and defect classification, and significantly improve the detection probability of minute pores and the depth measurement accuracy.

[0007] To solve the above technical problem, the embodiments of the present invention provide a method for locating internal defects of cable sealing lead based on multi-viewpoint TFM, and the method includes the following steps:

[0008] S1. When the ultrasonic probe array performs multi-viewpoint scanning on the cable sealing lead based on the spatial positions of multiple probes, obtain the original ultrasonic signals obtained from each viewpoint scan of the ultrasonic probe array; wherein, each probe spatial position is assigned as a viewpoint for scanning the cable sealing lead.

[0009] S2. Utilize the full-focusing algorithm to perform phase compensation and signal superposition on each original ultrasonic signal to generate a three-dimensional signal intensity distribution image, and through coordinate system registration, convert each signal intensity distribution image into a three-dimensional data set with a unified format.

[0010] S3. Based on the acoustic characteristics of the cable sealing lead, simulate the propagation paths and scattering intensities of ultrasonic waves in various defects, and combine the multiple probe spatial positions to calculate the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions.

[0011] S4. Extract the defect reflection signals from each signal intensity distribution image respectively, and combine the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions and the pre-defined intensity measurement model with noise to denoise and obtain the effective measured defect reflection signals, and further extract the three-dimensional data of each measured defect reflection signal in the corresponding signal intensity distribution image from the three-dimensional data set converted from each signal intensity distribution image.

[0012] S5. Based on the defect parameters, construct a Bayesian joint probability distribution function, and further estimate the Bayesian joint probability distribution function according to the three-dimensional data of each measured defect reflection signal in the corresponding signal intensity distribution image and the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions to obtain the defect parameters corresponding to each measured defect reflection signal.

[0013] S6. According to the actual physical structure of the cable sealing lead, simulate a three-dimensional solid model of the cable sealing lead, and perform automatic defect location and marking on the three-dimensional solid model of the cable sealing lead according to the original ultrasonic signals traced back from each measured defect reflection signal to the corresponding signal intensity distribution image and their corresponding defect parameters.

[0014] Wherein, before performing the step of each viewpoint scan of the cable sealing lead by the ultrasonic probe array each time, the following steps are further included:

[0015] The ultrasonic probe array rotates around the cable sealing lead and adjusts the probes provided thereon to the corresponding one probe spatial position.

[0016] Wherein, the step S2 specifically includes:

[0017] Obtain each original ultrasonic signal s ij (t); where i is the index of the transmitting unit, and i = 1, 2, ..., N; j is the index of the receiving unit, and j = 1, 2, ..., N; t is the time variable;

[0018] Determine the phase time τ ij compensated for each original ultrasonic signal s ij (x, z), and coherently superimpose the amplitudes of each original ultrasonic signal s ij (t) at the corresponding compensated phase time τ ij (x, z) to generate a three-dimensional signal intensity distribution image I m (x, z); where, s ij (τ ij (x, z)) is the amplitude of the original ultrasonic signal s ij (t) at the corresponding compensated phase time τ ij (x, z); m is the viewpoint index, and m = 1, 2, ..., M; is the round-trip propagation time of the ultrasonic wave from the transmitting unit to the pixel point and then back to the receiving unit; x i is the spatial coordinate of the i-th transmitting unit, x j is the spatial coordinate of the j-th receiving unit, c is the propagation speed of ultrasonic waves in the material, ||·|| is the Euclidean norm, and (x, z) is the spatial coordinate of the pixel point;

[0019] Perform spatial registration on all the signal intensity distribution images I m (x, z) to form a multi-dimensional data set And through multi-angle data fusion, each signal intensity distribution image I m (x, z) is converted into a three-dimensional data set with a unified format where, I m ={I m (x1, z1), ..., I m (x Nx , z Nz )}; N x is the number of pixels of the image grid in the horizontal x-axis direction, and N z is the number of pixels of the image grid in the depth z-axis direction.

[0020] Among them, the specific steps of step S3 include:

[0021] Determine the acoustic characteristics of the cable sealing lead, and through the elastic wave equation, model the propagation path of ultrasonic waves in various defects, and further combine the scattering characteristics of various defects to model the scattering intensity of ultrasonic waves in various defects A(φ); where, is the wave number, and f is the ultrasonic frequency; is the directivity function, which reflects the change of the scattering amplitude with the orientation angle θ; is the attenuation factor, which represents the energy attenuation when the ultrasonic wave propagates to the depth z d ; φ is the defect parameter, and φ = (a, b, θ); a is the major axis, and a ∈ [a min , a max ; b is the minor axis, and θ ∈ [0°, 180°];

[0022] According to the spatial positions of the multiple probes, a scattering intensity data set of various defects in the corresponding probe spatial positions is formed and further mapped with the defect parameters to determine the three-dimensional theoretical reflection intensity data D = {(a, b, θ) → A1(φ), A2(φ), A3(φ),..., A M (φ)}; where N a is the discretization number of the major axis a, and N b is the discretization number of the minor axis ratio b / a, and N q is the discretization number of the orientation angle θ.

[0023] Among them, the step S4 specifically includes:

[0024] In each signal intensity distribution image I m (x, z), the defect region is located by the threshold segmentation method, and in the defect region located in each signal intensity distribution image, the signal with the maximum intensity is extracted as the defect reflection signal respectively;

[0025] Based on the intensities of the defect reflection signals, a defect reflection signal intensity vector Y = [Y1, Y2,..., Y M T is constructed, and combined with the intensity measurement model with noise and the three-dimensional theoretical reflection intensity data of various defects in the corresponding probe spatial positions, the defect reflection signals with intensities meeting the predetermined conditions are selected as the effective defect measured reflection signals; among them, the intensity measurement model with noise is constructed by using the noise obeying the zero-mean multivariate Gaussian distribution function, and its expression is e = [e1, e2,..., e M T is the noise vector; is the multivariate Gaussian distribution with a zero-mean vector and a covariance matrix of Σ; is the noise covariance matrix, which is estimated by the maximum likelihood estimation;

[0026] ​​From the three-dimensional data sets converted from the signal intensity distribution images, extract the three-dimensional data of each measured reflection signal of the defect in the corresponding signal intensity distribution image.

[0027] Among them, the step S5 specifically includes:

[0028] Based on the defect parameters, construct the Bayesian joint probability distribution function as Among them, φ is the defect parameter, and φ = (a, b, θ); ∑ -1 The inverse matrix of the noise covariance matrix; exp(·) is the Gaussian likelihood function; is the matching degree; P(φ) is the prior distribution of the encoded defect parameters;

[0029] By the Markov chain Monte Carlo method, draw samples from the posterior distribution P(φ|Y) and import them into the Bayesian joint probability distribution function to estimate the joint probability distribution of the defect parameters in each measured reflection signal of the defect. Further, the defect parameter with the maximum joint probability distribution is output as the final defect parameter of each measured reflection signal of the defect; among them, Y in the sample comes from the three-dimensional data of each measured reflection signal of the defect in the corresponding signal intensity distribution image; φ in the sample comes from the three-dimensional theoretical reflection intensity data of various types of defects at the corresponding probe spatial positions.

[0030] Among them, the step S6 specifically includes:

[0031] According to the actual physical structure of the cable sealing lead, use simulation software to simulate the three-dimensional solid model of the cable sealing lead;

[0032] According to the original ultrasonic signals traced back from the measured reflection signals of each defect by the corresponding signal intensity distribution images, determine the three-dimensional coordinate positions of each defect on the cable sealing lead, and the measured reflection signals of each defect fed back in the corresponding signal intensity distribution images;

[0033] On the three-dimensional solid model of the cable sealing lead, according to the three-dimensional coordinate positions of each defect and combined with the defect parameters of the measured reflection signals of each defect, automatically perform defect positioning and marking.

[0034] The embodiment of the present invention also provides a cable sealing lead internal defect positioning system based on multi-viewpoint TFM, including:

[0035] A multi-viewpoint scanning signal acquisition unit, configured to acquire the original ultrasonic signals obtained by each viewpoint scan of the ultrasonic probe array when the ultrasonic probe array performs multi-viewpoint scanning on the cable sealing lead based on multiple probe spatial positions; among them, each probe spatial position is assigned as a viewpoint for scanning the cable sealing lead;

[0036] A multi-viewpoint TFM imaging data processing unit is used to perform phase compensation and signal superposition on each original ultrasonic signal by using the full focusing algorithm to generate a three-dimensional signal intensity distribution image, and convert each signal intensity distribution image into a three-dimensional data set with a unified format through coordinate system registration;

[0037] A defect theory data simulation unit is used to simulate the propagation path and scattering intensity of ultrasonic waves in various defects based on the acoustic characteristics of the cable sealing lead, and calculate the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions in combination with the spatial positions of the multiple probes;

[0038] A multi-viewpoint TFM imaging data denoising unit is used to separately extract defect reflection signals in each signal intensity distribution image, and denoise to obtain effective measured defect reflection signals in combination with the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions and a predefined intensity measurement model with noise, and further extract the three-dimensional data of each measured defect reflection signal in the corresponding signal intensity distribution image from the three-dimensional data set converted from each signal intensity distribution image;

[0039] A defect parameter probability matching unit is used to construct a Bayesian joint probability distribution function based on defect parameters, and further estimate the Bayesian joint probability distribution function according to the three-dimensional data of each measured defect reflection signal in the corresponding signal intensity distribution image and the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions to obtain the defect parameters corresponding to each measured defect reflection signal;

[0040] A defect location marking unit is used to simulate a three-dimensional solid model of the cable sealing lead according to the actual physical structure of the cable sealing lead, and perform automatic defect location and marking on the three-dimensional solid model of the cable sealing lead according to the original ultrasonic signal traced back from each measured defect reflection signal in the corresponding signal intensity distribution image and its corresponding defect parameters.

[0041] Implementing the embodiments of the present invention has the following beneficial effects:

[0042] The present invention obtains multi-dimensional acoustic response data of internal defects of the cable sealing lead through ultrasonic array multi-viewpoint scanning. After constructing a multi-viewpoint TFM image according to the full focusing algorithm, the Bayesian theorem is used to infer the probability distribution of defect parameters, and an automatic defect location and marking are performed using the three-dimensional solid model of the cable sealing lead, thereby being able to solve the defects existing in the aspects of acoustics, signal processing, algorithm models, and defect classification in the existing methods, and significantly improving the detection probability of micro air holes and the depth measurement accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, obtaining other drawings based on these drawings still falls within the scope of the present invention.

[0044] Figure 1 It is a flowchart of a method for locating internal defects in cable sealing lead based on multi-viewpoint TFM provided by an embodiment of the present invention;

[0045] Figure 2 It is a schematic structural diagram of a system for locating internal defects in cable sealing lead based on multi-viewpoint TFM provided by an embodiment of the present invention. Detailed implementation manners

[0046] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the drawings.

[0047] As Figure 1 shown, in an embodiment of the present invention, a method for locating internal defects in cable sealing lead based on multi-viewpoint TFM is provided. The method includes the following steps:

[0048] Step S1: When the ultrasonic probe array performs multi-viewpoint scanning on the cable sealing lead based on the spatial positions of multiple probes, obtain the original ultrasonic signals obtained by each viewpoint scanning of the ultrasonic probe array; wherein, each probe spatial position is assigned as a viewpoint for scanning the cable sealing lead.

[0049] The specific process is as follows. First, before each step of the ultrasonic probe array performing viewpoint scanning on the cable sealing lead, it further includes: the ultrasonic probe array rotates around the cable sealing lead, and the probes provided thereon are adjusted to corresponding one probe spatial positions.

[0050] In an example, a circular ultrasonic probe array is used to perform omnidirectional scanning on the cable sealing lead from different spatial angles. At this time, the ultrasonic probe array configuration uses a combination of high-frequency probes (5 - 10 MHz) and low-frequency probes (1 - 5 MHz), which are respectively used for surface high-resolution imaging and deep penetration detection.

[0051] Secondly, the ultrasonic probe array will cyclically traverse the transmit-receive element combinations at each probe spatial position (such as different spatial positions on the left side, top, etc. of the cable sealing lead) to obtain the original ultrasonic signal s obtained by each corresponding viewpoint scanning. ij(t). Wherein, i is the index of the transmitting unit, and i = 1, 2,..., N; j is the index of the receiving unit, and j = 1, 2,..., N; t is the time variable. It should be noted that the original ultrasonic signal s ij (t) is characterized as the time-domain signal recorded by the j-th receiving unit after the i-th transmitting unit excites ultrasonic waves. And after each viewpoint scan, there will be N 2 *N 2 original ultrasonic signals.

[0052] S2. Using the full focusing algorithm, perform phase compensation and signal superposition on each original ultrasonic signal to generate a three-dimensional signal intensity distribution image, and through coordinate system registration, convert each signal intensity distribution image into a three-dimensional data set with a unified format;

[0053] The specific process is as follows. First, obtain each original ultrasonic signal s ij (t).

[0054] Secondly, determine the phase time τ ij (x, z) to be compensated for each original ultrasonic signal s ij (t), and perform coherent superposition on the amplitudes of each original ultrasonic signal s ij (t) at the corresponding compensated phase time τ ij (x, z) to generate a three-dimensional signal intensity distribution image I m (x, z); where s ij (τ ij (x, z)) is the amplitude of the original ultrasonic signal s ij (t) at the corresponding compensated phase time τ ij (x, z); m is the viewpoint index, and m = 1, 2,..., M; is the two-way propagation time of ultrasonic waves from the transmitting unit to the pixel point and then back to the receiving unit; x i is the spatial coordinate of the i-th transmitting unit, x j is the spatial coordinate of the j-th receiving unit, c is the propagation speed of ultrasonic waves in the material, ||·|| is the Euclidean norm, and (x, z) is the spatial coordinate of the pixel point.

[0055] It should be noted that by compensating the phase time to align the propagation times of each channel signal at the target point, coherent superposition is achieved, and the signal intensity is enhanced through the coherent superposition of all transmit-receive channel signals to suppress noise.

[0056] Finally, perform spatial registration on all signal intensity distribution images I m (x, z) to form a multi-dimensional data set Through multi - angle data fusion, obtain the signal intensity distribution images I m (x, z) are both converted into a three - dimensional data set with a unified format Among them, I m ={I m (x1, z1),..., I m (x Nx , z Nz )}; N x is the number of pixels of the image grid in the horizontal x - axis direction, and N z is the number of pixels of the image grid in the depth z - axis direction.

[0057] S3. Based on the acoustic characteristics of the cable sealing lead, simulate the propagation paths and scattering intensities of ultrasonic waves in various defects, and combine the spatial positions of the multiple probes to calculate the three - dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions;

[0058] The specific process is as follows. First, determine the acoustic characteristics of the cable sealing lead, and model the propagation paths of ultrasonic waves in various defects through the elastic wave equation Navier, and further combine the scattering characteristics of various defects to model the scattering intensity of ultrasonic waves in various defects A(φ); where, is the wave number, and f is the ultrasonic frequency; is the directivity function, reflecting the change of the scattering amplitude with the orientation angle θ; is the attenuation factor, indicating the energy attenuation when the ultrasonic wave propagates to the depth z d ; φ is the defect parameter, and φ=(a, b, θ); a is the major axis, a ∈[a min , a max ; b is the minor axis, and θ ∈[0°, 180°];

[0059] In an example, first, the input acoustic characteristics of the cable sealing lead include material parameters such as the sound velocity c L , c S (longitudinal / transverse wave), density ρ, attenuation coefficient α, etc.; second, through the ray - tracing method, simulate the propagation path of ultrasonic waves from the transmitting unit to the defect and then back to the receiving unit, and calculate its propagation time and amplitude attenuation using the elastic wave equation Navier; then, define the boundary conditions of typical defects (holes, cracks, sand holes, etc.), including the major axis a, minor axis b, orientation angle θ and position (x d , z d ), so as to model the scattering intensity of ultrasonic waves in various defects, and this model is

[0060] Finally, according to the spatial positions of multiple probes, a dataset of the scattering intensities of various defects at the corresponding probe spatial positions is formed. And further map it with the defect parameters to determine the three-dimensional theoretical reflection intensity data D = {(a, b, θ) → A1(φ), A2(φ), A3(φ),..., A M (φ)} at the corresponding probe spatial positions for various defects; where N a is the number of discretization points of the major axis a, N b is the number of discretization points of the minor axis ratio b / a, N q is the number of discretization points of the orientation angle θ.

[0061] It can be understood that D is the set of theoretical scattering intensities corresponding to M different viewpoints for each defect parameter combination (a, b, θ).

[0062] S4. Extract the defect reflection signals from each signal intensity distribution image respectively, and combine the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions and the pre-defined intensity measurement model with noise to denoise and obtain the effective measured defect reflection signals. Moreover, further extract the three-dimensional data of each measured defect reflection signal in the corresponding signal intensity distribution image from the three-dimensional dataset converted from each signal intensity distribution image;

[0063] The specific process is as follows. First, in each signal intensity distribution image I m (x, z), the defect regions are located by the threshold segmentation method, and in the defect regions located in each signal intensity distribution image, the signal with the maximum intensity is extracted as the defect reflection signal respectively.

[0064] Secondly, based on the intensities of the defect reflection signals, construct the intensity vector Y = [Y1, Y2,..., Y M T , and combine the intensity measurement model with noise and the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions to screen out the defect reflection signals whose intensities meet the predetermined conditions as the effective measured defect reflection signals; where the intensity measurement model with noise is constructed by using a multivariate Gaussian distribution function with zero mean for the noise, and its expression is e = [e1, e2,..., e M T is the noise vector; is a multivariate Gaussian distribution with a zero mean vector and a covariance matrix Σ; is the noise covariance matrix, which is estimated by the maximum likelihood estimation.

[0065] ​​It should be noted that Y is composed of the theoretical response A(θ) and the random noise ε, and the non - diagonal elements of the covariance matrix Σ reflect the noise correlation between different viewpoints.

[0066] Estimate the noise covariance matrix from historical data through maximum likelihood estimation (MLE):

[0067]

[0068] where N data is the number of samples in the historical dataset, Y (n) is the measured amplitude vector of the n - th sample, and f (n) is the known defect parameter corresponding to the n - th sample.

[0069] Finally, extract the three - dimensional data of each measured defect reflection signal in the corresponding signal intensity distribution image from the three - dimensional dataset converted from each signal intensity distribution image.

[0070] In an example, in the intensity vector Y = [Y1, Y2,..., Y M T composed of the measured defect reflection signal and the intensity vector A(φ) = [A1(φ), A2(φ),..., A M (φ)] T of the theoretical reflection intensities of various defects, extract the noise ε by calculating the noise between Y m and A m (φ) corresponding to the viewpoint m. If it is determined that the signal - to - noise ratio of the noise ε m is lower than a certain decibel, it is considered to meet the conditions and retain the intensity Y m corresponding to the defect reflection signal and output it as an effective measured defect reflection signal; otherwise, discard it. m Finally, from the dataset I

[0071] = {I m (x1, z1),..., I m (x m (x Nx , z Nz )}, extract the three - dimensional data I m corresponding to the intensity Y m (x Ny , z Ny ).

[0072] ​S5. Based on the defect parameters, construct a Bayesian joint probability distribution function, and further estimate the Bayesian joint probability distribution function according to the three-dimensional data of the measured reflection signals of each defect in the corresponding signal intensity distribution image and the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions, so as to obtain the defect parameters corresponding to the measured reflection signals of each defect;

[0073] Specifically, first, based on the defect parameters, construct a Bayesian joint probability distribution function as where φ is the defect parameter, and φ = (a, b, θ); ∑ -1 The inverse matrix of the noise covariance matrix; exp(·) is the Gaussian likelihood function; is the matching degree; P(φ) is the prior distribution of the encoded defect parameters;

[0074] Finally, through the Markov chain Monte Carlo method, samples are drawn from the posterior distribution P(φ|Y) and imported into the Bayesian joint probability distribution function to estimate the joint probability distribution of the defect parameters in the measured reflection signals of each defect. Furthermore, the defect parameters with the largest joint probability distribution are output as the final defect parameters of the measured reflection signals of each defect; where Y in the samples comes from the three-dimensional data of the measured reflection signals of each defect in the corresponding signal intensity distribution image; φ in the samples comes from the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions.

[0075] In an example, the specific steps of drawing samples from the posterior distribution by the Markov chain Monte Carlo (MCMC) method are as follows:

[0076] First, perform initialization: randomly select the initial parameter φ(0);

[0077] Then, perform iterative sampling: (k = 1, 2,..., K);

[0078] Generate a candidate parameter φ' from the Gaussian distribution q(φ'|φ (k) );

[0079] Next, calculate the acceptance probability:

[0080] Accept φ' with probability α, that is, φ(k + 1) = φ'; otherwise, retain the current parameter φ(k + 1) = φ(k); where q(φ'|φ (k) ) is the Gaussian distribution, centered on the current parameter φ'(k), and K is the number of sampling times.

[0081] The output signal at this time is: the posterior distribution sample set is where a (k) , b (k) , θ (k)The major axis, minor axis, and orientation angle obtained from the k-th sampling;

[0082] At this time, the sample set represents the posterior probability distribution of the defect parameters and can be used to calculate the defect parameter estimation:

[0083] Then, calculate the mean and confidence interval, including the mean estimation and the 95% confidence interval where σ a : The sample standard deviation of the major axis a. 1.96: The quantile coefficient corresponding to the 95% confidence interval of the normal distribution.

[0084] Next, determine the defect type according to the aspect ratio threshold: When α < 0.1, it is a crack; when α ≥ 0.1, it is a hole.

[0085] S6. According to the actual physical structure of the cable sealing lead, simulate a three-dimensional solid model of the cable sealing lead, and based on the original ultrasonic signals traced from the measured reflection signals of each defect by the corresponding signal intensity distribution images and their corresponding defect parameters, perform automatic defect positioning and annotation on the three-dimensional solid model of the cable sealing lead.

[0086] The specific process is as follows. First, according to the actual physical structure of the cable sealing lead, use simulation software to simulate a three-dimensional solid model of the cable sealing lead.

[0087] Secondly, according to the original ultrasonic signals traced from the measured reflection signals of each defect by the corresponding signal intensity distribution images, determine the three-dimensional coordinate positions of each defect on the cable sealing lead, as well as the measured reflection signals of each defect reflected in the corresponding signal intensity distribution images. It should be noted that by the spatial positions of the probes in the ultrasonic probe array, infer the three-dimensional coordinate position when the original ultrasonic signal s ij (t) is the measured reflection signal of the defect, and combined with the signal propagation time of the original ultrasonic signal s ij (t), further calculate the distance from the original ultrasonic signal to the defect to deduce the three-dimensional coordinate position of the defect, which are all common technical means in the art and will not be elaborated here.

[0088] Finally, on the three-dimensional solid model of the cable sealing lead, according to the three-dimensional coordinate positions of each defect and combined with the defect parameters of the measured reflection signals of each defect, automatically perform defect positioning and annotation. It should be noted that on the three-dimensional solid model of the cable sealing lead, find the three-dimensional coordinate position of the defect, and at this position, highlight the defect parameters such as the type, size, and orientation angle of the defect by high brightness to clearly present the defect.

[0089] Such asFigure 2 As shown in the figure, in an embodiment of the present invention, a cable lead sealing internal defect location system based on multi-viewpoint TFM is provided, including:

[0090] A multi-viewpoint scanning signal acquisition unit 110, configured to obtain the original ultrasonic signals obtained by each viewpoint scan of the ultrasonic probe array each time when the ultrasonic probe array performs multi-viewpoint scanning on the cable lead sealing based on multiple probe spatial positions; wherein, each probe spatial position is assigned as a viewpoint for scanning the cable lead sealing;

[0091] A multi-viewpoint TFM imaging data processing unit 120, configured to use the full focus algorithm to perform phase compensation and signal superposition on each original ultrasonic signal to generate a three-dimensional signal intensity distribution image, and convert each signal intensity distribution image into a three-dimensional data set with a unified format through coordinate system registration;

[0092] A defect theory data simulation unit 130, configured to simulate the propagation path and scattering intensity of ultrasonic waves in various defects based on the acoustic characteristics of the cable lead sealing, and calculate the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions in combination with the multiple probe spatial positions;

[0093] A multi-viewpoint TFM imaging data denoising unit 140, configured to separately extract defect reflection signals in each signal intensity distribution image, and combine the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions and a predefined intensity measurement model with noise to denoise and obtain effective defect measured reflection signals, and further extract the three-dimensional data of each defect measured reflection signal in the corresponding signal intensity distribution image from the three-dimensional data set converted from each signal intensity distribution image;

[0094] A defect parameter probability matching unit 150, configured to construct a Bayesian joint probability distribution function based on defect parameters, and further estimate the Bayesian joint probability distribution function according to the three-dimensional data of each defect measured reflection signal in the corresponding signal intensity distribution image and the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions to obtain the defect parameters corresponding to each defect measured reflection signal;

[0095] A defect location marking unit 160, configured to simulate a three-dimensional solid model of the cable lead sealing according to the actual physical structure of the cable lead sealing, and perform automatic defect location and marking on the three-dimensional solid model of the cable lead sealing according to the original ultrasonic signals traced back from each defect measured reflection signal to the corresponding signal intensity distribution image and their corresponding defect parameters.

[0096] Implementing the embodiments of the present invention has the following beneficial effects:

[0097] The present invention obtains multi-dimensional acoustic response data of the internal defects of cable sealing lead through ultrasonic array multi-viewpoint scanning. After the multi-viewpoint TFM image construction is performed according to the full-focusing algorithm, the Bayesian theorem is used to infer the probability distribution of defect parameters, and the three-dimensional solid model of the cable sealing lead is used for automatic defect positioning and annotation, so as to solve the defects existing in the aspects of acoustics, signal processing, algorithm models, and defect classification in the existing methods, and significantly improve the detection probability of micro air holes and the depth measurement accuracy.

[0098] It should be noted that in the above system embodiments, the included system modules are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific names of the functional modules are only for the convenience of mutual distinction and do not limit the protection scope of the present invention.

[0099] Those of ordinary skill in the art can understand that all or part of the steps in implementing the methods of the above embodiments can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable storage medium, such as ROM / RAM, disk, optical disc, etc.

[0100] The above-disclosed are only the preferred embodiments of the present invention, and of course, the scope of the rights of the present invention cannot be limited thereby. Therefore, equivalent changes made according to the claims of the present invention still fall within the scope covered by the present invention.

Claims

1. A method for locating internal defects in cable sealing lead based on multi-viewpoint TFM, characterized in that, The method includes the following steps: S1. When the ultrasonic probe array performs multi-viewpoint scanning on the cable sealing lead based on multiple probe spatial positions, obtain the original ultrasonic signals obtained from each viewpoint scan of the ultrasonic probe array; wherein, each probe spatial position is assigned as a viewpoint for scanning the cable sealing lead; S2. Use the full focusing algorithm to perform phase compensation and signal superposition on each original ultrasonic signal to generate a three-dimensional signal intensity distribution image, and convert each signal intensity distribution image into a three-dimensional data set with a unified format through coordinate system registration; S3. Based on the acoustic characteristics of the cable sealing lead, simulate the propagation paths and scattering intensities of ultrasonic waves in various defects, and combine the multiple probe spatial positions to calculate the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions; S4. Extract defect reflection signals from each signal intensity distribution image respectively, and combine the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions and a predefined intensity measurement model with noise to denoise and obtain effective measured defect reflection signals, and further extract the three-dimensional data of each measured defect reflection signal in the corresponding signal intensity distribution image from the three-dimensional data set converted from each signal intensity distribution image; S5. Based on the defect parameters, construct a Bayesian joint probability distribution function, and further estimate the Bayesian joint probability distribution function according to the three-dimensional data of each measured defect reflection signal in the corresponding signal intensity distribution image and the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions to obtain the defect parameters corresponding to each measured defect reflection signal; S6. According to the actual physical structure of the cable sealing lead, simulate a three-dimensional solid model of the cable sealing lead, and perform automatic defect positioning and annotation on the three-dimensional solid model of the cable sealing lead according to the original ultrasonic signals traced back from each measured defect reflection signal to the corresponding signal intensity distribution image and their corresponding defect parameters.

2. The method for locating internal defects of cable sealing lead based on multi-viewpoint TFM according to claim 1, wherein Before performing the step of each viewpoint scan of the cable sealing lead by the ultrasonic probe array, the following steps are further included: The ultrasonic probe array rotates around the cable sealing lead, and the probes provided thereon are adjusted to corresponding probe spatial positions.

3. The method for locating internal defects of cable sealing lead based on multi-viewpoint TFM according to claim 1, wherein The specific content of step S2 includes: Obtain each original ultrasonic signal s ij (t); where i is the transmitting unit index, and i = 1, 2, ..., N; j is the receiving unit index, and j = 1, 2, ..., N; t is the time variable; Determine each original ultrasonic signal s ij (t) compensated phase time τ ij (x, z), and coherently superimpose the amplitudes of each original ultrasonic signal s ij (t) at the corresponding compensated phase time τ ij (x, z) to generate a three-dimensional signal intensity distribution image I m (x, z); where s ij (τ ij (x, z)) is the amplitude of the original ultrasonic signal s ij (t) at the corresponding compensated phase time τ ij (x, z); m is the viewpoint index, and m = 1, 2,..., M; is the round-trip propagation time of the ultrasonic wave from the transmitting unit to the pixel point and then back to the receiving unit; x i is the spatial coordinate of the i-th transmitting unit, x j is the spatial coordinate of the j-th receiving unit, c is the propagation speed of ultrasonic waves in the material, ||·|| is the Euclidean norm, and (x, z) is the spatial coordinate of the pixel point; For all signal intensity distribution images I m (x, z) perform spatial registration to form a multi-dimensional dataset And through multi-angle data fusion, obtain each signal intensity distribution image I m (x, z) are all converted into a three-dimensional dataset with a unified format Among them, I m ={I m (x1, z1),..., I m (x Nx , z Nz )}; N x is the number of pixels of the image grid in the horizontal x-axis direction, and N z is the number of pixels of the image grid in the depth z-axis direction.

4. The method for locating internal defects of cable sealing lead based on multi-viewpoint TFM according to claim 3, wherein The specific content of step S3 includes: Determine the acoustic properties of the cable sealing lead, model the propagation path of ultrasonic waves in various defects through the elastic wave equation, and further combine the scattering characteristics of various defects to model the scattering intensity A(φ) of ultrasonic waves in various defects; where is the wave number, and f is the ultrasonic frequency; is the directivity function, which reflects the change of the scattering amplitude with the orientation angle θ; is the attenuation factor, indicating the energy attenuation when the ultrasonic wave propagates to the depth z d ; φ is the defect parameter, and φ = (a, b, θ); a is the major axis, a ∈ [a min , a max ; b is the minor axis, and θ ∈ [0°, 180°]; Based on the spatial positions of the multiple probes, a dataset of the scattering intensities of various types of defects at the corresponding probe spatial positions is formed. And further map it with the defect parameters to determine the three-dimensional theoretical reflection intensity data D = {(a, b, θ) → A1(φ), A2(φ), A3(φ),..., A M (φ)}; where N a is the number of discretization points of the major axis a, N b is the number of discretization points of the minor axis ratio b / a, N q is the number of discretization points of the orientation angle θ.

5. The method for locating internal defects of cable sealing lead based on multi-viewpoint TFM according to claim 4, wherein The specific content of step S4 includes: In each signal intensity distribution image I m (x, z), the defect region is located by the threshold segmentation method, and within the defect region located in each signal intensity distribution image, the signal with the maximum intensity is extracted respectively as the defect reflection signal; Based on the intensities of the defect reflection signals, an intensity vector Y = [Y1, Y2,..., Y M of the defect reflection signals is constructed. T Combined with the intensity measurement model with noise and the three-dimensional theoretical reflection intensity data of various defects at their corresponding probe spatial positions, the defect reflection signals whose intensities meet the predetermined conditions are selected as the effective defect measured reflection signals. Among them, the intensity measurement model with noise is constructed by using a noise that follows a zero-mean multivariate Gaussian distribution function, and its expression is e = [e1, e2,..., e M T is the noise vector; is a multivariate Gaussian distribution with a zero-mean vector and a covariance matrix Σ; is the noise covariance matrix, which is estimated by maximum likelihood estimation; Extract the three-dimensional data of each measured defect reflection signal in the corresponding signal intensity distribution image from the three-dimensional data set converted from each signal intensity distribution image.

6. The method for locating internal defects of cable sealing lead based on multi-viewpoint TFM according to claim 5, characterized in that, The specific content of step S5 includes: Based on the defect parameters, the Bayesian joint probability distribution function is constructed as where, ∑ -1 the inverse matrix of the noise covariance matrix; exp(·) is the Gaussian likelihood function; is the matching degree; P(φ) is the prior distribution of the coding for the defect parameters; Samples are drawn from the posterior distribution P(φ|Y) by the Markov chain Monte Carlo method and imported into the Bayesian joint probability distribution function to estimate the joint probability distribution of the defect parameters in the measured reflection signals of each defect. Further, the defect parameters with the maximum joint probability distribution are output as the final defect parameters of the measured reflection signals of each defect. Among them, Y in the samples comes from the three-dimensional data of the measured reflection signals of each defect in the corresponding signal intensity distribution image, and φ in the samples comes from the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions.

7. The method for locating internal defects of cable sealing lead based on multi-viewpoint TFM according to claim 6, characterized in that, The specific steps of step S6 include: According to the actual physical structure of the cable sealing lead, a three-dimensional solid model of the cable sealing lead is simulated using simulation software. Based on the original ultrasonic signals traced back from the measured reflection signals of each defect by the corresponding signal intensity distribution image, the three-dimensional coordinate positions of each defect on the cable sealing lead are determined, as well as the measured reflection signals of each defect reflected in the corresponding signal intensity distribution image. On the three-dimensional solid model of the cable sealing lead, according to the three-dimensional coordinate positions of each defect and in combination with the defect parameters of the measured reflection signals of each defect, defect positioning and labeling are automatically performed.

8. A cable lead sealing internal defect location system based on multi-viewpoint TFM, characterized in that, It includes: A multi-viewpoint scanning signal acquisition unit, which is used to acquire the original ultrasonic signals obtained by each viewpoint scan of the ultrasonic probe array when the ultrasonic probe array performs multi-viewpoint scanning on the cable sealing lead based on multiple probe spatial positions. Among them, each probe spatial position is assigned as a viewpoint for scanning the cable sealing lead. A multi-viewpoint TFM imaging data processing unit, which is used to perform phase compensation and signal superposition on each original ultrasonic signal using the full focusing algorithm to generate a three-dimensional signal intensity distribution image, and convert each signal intensity distribution image into a three-dimensional data set with a unified format through coordinate system registration. A defect theoretical data simulation unit, which is used to simulate the propagation path and scattering intensity of ultrasonic waves in various defects based on the acoustic characteristics of the cable sealing lead, and calculate the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions in combination with the multiple probe spatial positions. A multi-viewpoint TFM imaging data denoising unit, which is used to extract the defect reflection signals from each signal intensity distribution image, and denoise to obtain effective measured reflection signals of defects in combination with the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions and a predefined intensity measurement model with noise. Further, the three-dimensional data of each measured reflection signal of a defect in the corresponding signal intensity distribution image is extracted from the three-dimensional data set converted from each signal intensity distribution image. A defect parameter probability matching unit, which is used to construct a Bayesian joint probability distribution function based on the defect parameters, and further estimate the Bayesian joint probability distribution function according to the three-dimensional data of the measured reflection signals of each defect in the corresponding signal intensity distribution image and the three-dimensional theoretical reflection intensity data of various defects at the corresponding probe spatial positions to obtain the defect parameters corresponding to the measured reflection signals of each defect. The defect location and marking unit is used to simulate a three-dimensional solid model of the cable sealing lead according to the actual physical structure of the cable sealing lead, and perform automatic defect location and marking on the three-dimensional solid model of the cable sealing lead based on the original ultrasonic signals traced from the measured reflection signals of each defect and their corresponding defect parameters by the corresponding signal intensity distribution images.

Images