Pipeline defect detection method and device, electronic equipment and storage medium

Abstract

The invention relates to the field of pipeline detection, in particular to a pipeline defect detection method, a pipeline defect detection device, electronic equipment and a storage medium, and the pipeline defect detection method comprises the following steps: 1, obtaining a pipeline image; 2, reconstructing the pipeline image based on the image patch model, and generating a reconstructed image corresponding to the pipeline image; 3, segmenting the reconstructed image, and determining a background information matrix and a defect information matrix corresponding to the reconstructed image; 4, optimizing the background information matrix and the defect information matrix based on an accelerated near-end gradient algorithm to obtain a pipeline background image and a pipeline defect image; and 5, fusing the pipeline background image and the pipeline defect image to generate a target pipeline image. The pipeline image is reconstructed through the image patch model, the signal-to-noise ratio of the pipeline image is improved, the pipeline image is optimized through the accelerated near-end gradient algorithm subsequently, the signal-to-noise ratio of the defect in the image is remarkably improved, and the detection capacity of the pipeline defect is improved.

Description

Technical Field

[0001] The present invention relates to the field of pipeline detection, and in particular to a pipeline defect detection method, a pipeline defect detection device, an electronic device and a storage medium. Background Art

[0002] In the petroleum, chemical, electric power, metallurgy and other industrial fields, metal pipes are usually used to transport high-pressure, high-temperature and corrosive gas or liquid media. Due to medium wear and fluid accelerated corrosion, the pipe wall will be corroded and thinned, or even perforated, which may easily cause accidents such as medium leakage and explosion. Pipe wall corrosion will lead to a decrease in the pressure bearing performance of the pipeline, which seriously threatens production safety. Therefore, it is necessary to conduct non-destructive testing and evaluation of pipe wall corrosion on a regular basis to ensure the safe operation of the pipeline.

[0003] At present, the industry usually uses pulse eddy current detection technology to detect pipe walls. By turning off the current at the moment of probe loading, a rapidly decaying pulse magnetic field is stimulated. This magnetic field can pass through a certain thickness of protective layer and insulation layer to induce eddy currents on the surface of the inspected component. The induced eddy currents will diffuse from the upper surface to the lower surface. At the same time, during the eddy current diffusion process, a secondary magnetic field in the opposite direction of the exciting magnetic field will be generated, and this induced voltage will be output in the receiving sensor of the probe. If there are defects on the pipeline, it will affect the pulse eddy current condition on the loading pipeline, and then affect the induced voltage on the receiving sensor. Through algorithm analysis, the specific location and severity of the corrosion defects can be scanned.

[0004] The probe can be divided into point probe and array probe according to the different scanning methods during detection. Among them, the working mode of the array probe is the transmission / reception mode, and the transmitting coil and the receiving coil are electromagnetically coupled. The transmitting coil is driven by AC excitation to induce eddy currents in the pipe wall to be detected. The receiving coil works in passive mode, and its induced voltage is generated by the change of magnetic flux passing through the coil. Defects in the pipe wall will change the distribution of the induced eddy currents, and then change the magnetic field generated by the induced eddy currents, causing changes in the magnetic flux. This change will be captured and recorded as a signal by the receiving coil.

[0005] During the signal acquisition process, the detection system is affected by factors such as lift-off changes, environmental noise, and its own components. Therefore, noise will inevitably be mixed into the detection signal, affecting the identification and quantification of defects.

[0006] Therefore, it is necessary to process the original signal. Some classic filtering algorithms, such as mean filtering, low-pass filtering, high-pass filtering, median filtering, etc., are simple and efficient, but they are difficult to apply to all types of noise. When processing data with complex backgrounds, the current filtering algorithms are not widely applicable, have poor detection capabilities, and have certain limitations. Summary of the invention

[0007] In view of this, the purpose of the present invention is to provide a pipeline defect detection method, device, electronic device and storage medium, which are used to suppress background noise, improve the signal-to-noise ratio of the detection image, enhance the image characteristics of the target signal, and improve the detection capability of tiny defects.

[0008] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0009] In a first aspect, the present application provides a pipeline defect detection method, which is applied to array probe pipeline detection, including: step 1: acquiring a pipeline image; step 2: reconstructing the pipeline image based on an image patch model to generate a reconstructed image corresponding to the pipeline image; step 3: segmenting the reconstructed image to determine a background information matrix and a defect information matrix corresponding to the reconstructed image; step 4: optimizing the background information matrix and the defect information matrix based on an accelerated proximal gradient algorithm to obtain a pipeline background image and a pipeline defect image; step 5: fusing the pipeline background image and the pipeline defect image to generate a target pipeline image.

[0010] Furthermore, the pipeline image includes a pipeline axial channel image and a pipeline circumferential channel image.

[0011] Furthermore, step 2 includes: step 2-1: performing two-dimensional median filtering on the pipeline image; step 2-2: reconstructing the pipeline image after filtering based on the image patch model to generate a reconstructed image corresponding to the pipeline image.

[0012] Furthermore, step 3 includes: step 3-1: based on a stable principal component analysis algorithm, dividing the reconstructed image into a background information component, a defect information component and a noise information component; step 3-2: determining the background information matrix and the defect information matrix corresponding to the reconstructed image according to the background information component, the defect information component and the noise information component.

[0013] Furthermore, step 4 includes: step 4-1: converting the background information matrix and the defect information matrix into a convex optimization problem; step 4-2: solving the convex optimization problem by accelerating the proximal gradient algorithm to obtain the pipeline background image and the pipeline defect image.

[0014] Furthermore, step 5 includes: step 5-1: filtering the pipeline background image and the pipeline defect image; step 5-2: fusing the filtered pipeline background image and the pipeline defect image to generate a target pipeline image.

[0015] Furthermore, step 5-1 includes: step 5-1-1: processing the pipeline background image and the pipeline defect image through a one-dimensional median filtering algorithm to obtain a first filtered image corresponding to the pipeline background image and the pipeline defect image; step 5-1-2: processing the first filtered image through a two-dimensional mean filtering algorithm to obtain a target filtered image corresponding to the pipeline background image and the pipeline defect image.

[0016] In a second aspect, the present application provides a pipeline defect detection device, which is applied to array probe pipeline detection, including: an acquisition module, used to acquire a pipeline image; a reconstruction module, used to reconstruct the pipeline image based on an image patch model, and generate a reconstructed image corresponding to the pipeline image; a segmentation module, used to segment the reconstructed image, and determine the background information matrix and defect information matrix corresponding to the reconstructed image; a processing module, used to optimize the background information matrix and the defect information matrix based on an accelerated proximal gradient algorithm to obtain a pipeline background image and a pipeline defect image; a generation module, used to fuse the pipeline background image and the pipeline defect image to generate a target pipeline image.

[0017] In a third aspect, the present application provides an electronic device, comprising: a processor; and a memory for storing processor executable instructions; wherein the processor implements the pipeline defect detection method implemented in the first aspect as described above by running the executable instructions.

[0018] In a fourth aspect, the present application provides a non-volatile readable storage medium storing computer-readable instructions. When the computer-readable instructions are executed by a processor, the processor executes the pipeline defect detection method implemented in the first aspect or realizes the function of the pipeline defect detection device implemented in the second aspect.

[0019] It can be seen from the above technical solution that the advantages and positive effects of the pipeline defect detection method proposed by the present invention are:

[0020] The pipeline image is reconstructed through the image patch model to improve the signal-to-noise ratio of the pipeline image. After the image is segmented, the pipeline image is further optimized by using the segmented background information matrix and defect information matrix as parameters through the accelerated proximal gradient algorithm to process the data of the complex background. Finally, the optimized image is fused to generate the target pipeline image, which significantly improves the signal-to-noise ratio of defects in the image and improves the detection capability of pipeline defects. The pipeline image is first subjected to two-dimensional median filtering to suppress high-frequency noise in the pipeline image and achieve preliminary noise reduction. Since the detection object is a tiny defect, there are factors such as environmental noise and circuit noise, and the defect signal is relatively small compared to the overall image. The size is small, the contour features of the background image tend to be blurred, and there is non-local autocorrelation. Therefore, the image is divided into background information component, defect information component and noise information component through the stable principal component analysis algorithm to determine the corresponding background information matrix and defect information matrix, which is convenient for subsequent optimization processing; the background information matrix and defect information matrix are converted into convex optimization problems, and the convex optimization problem is solved in combination with the accelerated proximal gradient algorithm to suppress the noise signal and enhance the edge details of the defect signal; the pipeline background image and pipeline defect image are processed by the one-dimensional median filtering algorithm and the two-dimensional mean filtering algorithm respectively to suppress the noise in the image and further improve the signal-to-noise ratio of the target filtered image. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The above content and the following specific embodiments of the present invention will be better understood when read in conjunction with the accompanying drawings. It should be noted that the accompanying drawings are only examples of the technical solutions claimed for protection.

[0022] Figure 1 is a schematic diagram of an array probe structure provided by an embodiment of the present application;

[0023] Figure 2 is a flow chart of a pipeline defect detection method provided by an embodiment of the present application;

[0024] Figure 3 It is a schematic diagram of a pipeline axial channel and a pipeline circumferential channel provided in an embodiment of the present application;

[0025] Figure 4 This is a schematic diagram of a process of reconstructing an image using an image patch model provided by an embodiment of the present application;

[0026] Figure 5 It is a pipeline defect size diagram provided in an embodiment of the present application;

[0027] Figure 6 is a schematic diagram of a pipeline image before processing provided by an embodiment of the present application;

[0028] Figure 7is a schematic diagram of a processed pipeline image provided by an embodiment of the present application;

[0029] Figure 8 It is a structural block diagram of a pipeline defect detection device provided in one embodiment of the present application.

[0030] The reference numerals are described as follows:

[0031] Pipeline defect detection device 100;

[0032] Acquisition module 10;

[0033] Reconstruction module 20;

[0034] Segmentation module 30;

[0035] Processing module 40;

[0036] Generate module 50. DETAILED DESCRIPTION

[0037] The detailed features and advantages of the present invention are described in detail in the specific implementation modes below, and the contents are sufficient to enable any person skilled in the art to understand the technical contents of the present invention and implement them accordingly. Moreover, according to the description, claims and drawings disclosed in this specification, those skilled in the art can easily understand the relevant objects and advantages of the present invention.

[0038] It should be noted that in this specification, similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings.

[0039] In the description of this embodiment, it should be noted that the terms "upper", "lower", "inner", "bottom", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, or are the orientations or positional relationships in which the product is usually placed when in use. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0040] In order to make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0041] At present, electromagnetic probes can be divided into point probes and array probes according to the different scanning methods during detection. Among them, the point probe contains a single receiving sensor. The output of the point probe is line scanning data. If a point probe is used to detect and image a surface to be detected, the probe needs to perform two-dimensional scanning, and the detection efficiency is low.

[0042] like Figure 1 As shown, compared with point probes, the detection system composed of array probes has the advantages of large detection coverage area and simple scanning steps, which can improve the detection efficiency of pipeline defects.

[0043] It should be noted that the pipeline image in the present application refers to the C-scan image output when the array probe detects the pipeline.

[0044] Please refer to Figure 2 The present application provides a pipeline defect detection method, which is applied to array probe pipeline detection, and includes:

[0045] Step 1: Get the pipeline image.

[0046] Step 2: Reconstruct the pipeline image based on the image patch model to generate a reconstructed image corresponding to the pipeline image.

[0047] Step 3: Segment the reconstructed image and determine the background information matrix and defect information matrix corresponding to the reconstructed image.

[0048] Step 4: Optimize the background information matrix and defect information matrix based on the accelerated proximal gradient algorithm to obtain the pipeline background image and pipeline defect image.

[0049] Step 5: Fuse the pipeline background image and the pipeline defect image to generate the target pipeline image.

[0050] It can be understood that the pipeline image is reconstructed through the image patch model to improve the signal-to-noise ratio of the pipeline image. After the image is segmented, the pipeline image is further optimized by using the accelerated proximal gradient algorithm and taking the segmented background information matrix and defect information matrix as parameters to process the data of the complex background. Finally, the optimized image is fused to generate the target pipeline image, which significantly improves the signal-to-noise ratio of defects in the image, thereby improving the detection capability of pipeline defects.

[0051] Please refer to Figure 3 In one embodiment, the pipeline image may include a pipeline axial channel image and a pipeline circumferential image.

[0052] It can be understood that different types of pipeline channel images have different defects, and pipeline axial channel images may have axial defects, and pipeline circumferential channel images may have circumferential defects. By performing image processing on different types of pipeline channel images, the detection effect of pipeline defects can be effectively improved.

[0053] In one embodiment, step 2 may include:

[0054] Step 2-1: Perform two-dimensional median filtering on the pipeline image.

[0055] Step 2-2: Reconstruct the filtered pipeline image based on the image patch model to generate a reconstructed image corresponding to the pipeline image.

[0056] It can be understood that the pipeline image is processed by a two-dimensional median filtering algorithm to suppress high-frequency noise in the original pipeline image, thereby achieving the purpose of preliminary noise reduction of the pipeline image.

[0057] Please refer to Figure 4 , the image patch model can be implemented by using a sliding window of length m and width n. For example, starting from the upper left corner of the pipeline image, the image is moved from left to right and from top to bottom at equal distances, which can be set according to the size of the sliding window and the moving distance. The number b is the total number of times the sliding window can be moved in the matrix corresponding to the pipeline image. The pipeline image can be reconstructed by moving the sliding window to obtain a reconstructed image with m×n rows and b columns and a corresponding two-dimensional matrix.

[0058] It should be noted that the quality of the reconstructed image is not only affected by the original size of the pipeline image, but also depends on the size of the sliding window set in the image patch model and the distance the window slides. The appropriate sliding window size and window sliding distance can be selected according to different problem analyses and specific circumstances.

[0059] In one embodiment, step 3 may include:

[0060] Step 3-1: Based on the stable principal component analysis algorithm, the reconstructed image is divided into background information component, defect information component and noise information component.

[0061] In one embodiment, the reconstructed image is divided into three components by a stable principal component tracking algorithm, as shown below.

[0062] E=L+D+N…(1)

[0063] Among them, L is a low-rank matrix, represented as the background information component, D is a sparse matrix containing defect information, represented as the defect information component, and N is a noise matrix, represented as the noise information component.

[0064] It can be understood that in the C-scan image of the array probe eddy current detection, the defect information has obvious graphic features and high image edge contrast relative to the background information component. In addition, since the detection object is a tiny defect, the defect signal occupies a small size relative to the entire image, so the defect information component in the detection image can be represented by a sparse matrix. Furthermore, due to factors such as environmental noise and circuit noise, the contour features of the original pipeline image background image tend to be blurred and there is non-local autocorrelation, so the background information component can be represented by a low-rank matrix. The noise information component corresponding to the noise matrix can be understood as the presence of noise in the image.

[0065] Step 3-2: Determine the background information matrix and defect information matrix corresponding to the reconstructed image according to the background information component, the defect information component and the noise information component.

[0066] In one embodiment, step 4 may include:

[0067] Step 4-1: Convert the background information matrix and defect information matrix into a convex optimization problem.

[0068] It can be understood that, assuming that the random noise in the image is independent and identically distributed, the background information matrix and the defect information matrix can be converted into a convex optimization problem.

[0069] Step 4-2: Solve the convex optimization problem by accelerating the proximal gradient algorithm to obtain the pipeline background image and the pipeline defect image.

[0070] Assuming that the random noise is independent and identically distributed, the convex optimization problem can be solved by using the accelerated proximal gradient algorithm as follows.

[0071] min L,D ‖L‖ * +λ‖D‖1s.t.‖ELD‖ F ≤δ…(2)

[0072] Among them, λ is a positive weight constant, which can be set according to the specific situation.

[0073] It can be understood that by converting the background information matrix and the defect information matrix into a convex optimization problem and solving the convex optimization problem in combination with the accelerated proximal gradient algorithm, the noise signal can be suppressed and the edge details of the defect signal can be enhanced.

[0074] In one embodiment, step 5 may include:

[0075] Step 5-1: Filter the pipeline background image and pipeline defect image.

[0076] Step 5-2: Fuse the filtered pipeline background image and pipeline defect image to generate a target pipeline image.

[0077] In one embodiment, step 5-1 may include:

[0078] Step 5-1-1: Process the pipeline background image and the pipeline defect image by a one-dimensional median filtering algorithm to obtain a first filtered image corresponding to the pipeline background image and the pipeline defect image;

[0079] Step 5-1-2: Process the first filtered image using a two-dimensional mean filtering algorithm to obtain a target filtered image corresponding to the pipeline background image and the pipeline defect image.

[0080] It can be understood that the pipeline background image and the pipeline defect image are processed by the one-dimensional median filtering algorithm and the two-dimensional mean filtering algorithm respectively to suppress the noise in the image and further improve the signal-to-noise ratio of the target filtered image.

[0081] Please refer to Figures 5 to 7 , the defect detection effect of this application will be fully described below.

[0082] Defects of different sizes were machined into the metal pipe samples. Figure 5 It can be seen from the parameters that these defects are very small, with the smallest defect being only 1*0.1*0.3mm 3 ,It is difficult to detect such small pipeline defects using current methods.

[0083] The pipeline image detected by the array probe is as follows Figure 6 As shown in the figure, the signals of larger defects (#1-#3) are more obvious, while the signals of smaller defects (#4-#6) are weaker and close to the noise level, and the defect pattern features are not obvious in the original image. In addition, the amplitude of some noise signals is similar to that of micro-defect signals, which greatly affects the recognition of defects.

[0084] The image processed by the above algorithm is as follows Figure 7 As shown in the image results, it can be seen that the processed image suppresses noise and makes the defect signal more obvious.

[0085] Furthermore, if an appropriate threshold is applied to the fused data, for example, the threshold is one-tenth of the maximum defect signal in the image, and the signal below the threshold is set to zero, all defects can be reliably detected from the fused image. The signal-to-noise ratio of defect 5 is improved from 7.69dB to 50.10dB, and the algorithm significantly improves the signal-to-noise ratio of the defect.

[0086] It should be noted that the shape and material of the coils used in the array probe are exactly the same, and the dimensions are unified as 0.5 mm inner diameter and 1.6 mm outer diameter. The coil height is set to 1 mm, the diameter of the winding copper wire is 0.04 mm, and the number of turns of the coil is 240.

[0087] The array probe is composed of coils, and the distance between the centers of two adjacent coils is 2 mm. The host controls the scanning platform to realize the scanning imaging of the probe. A signal generator (NIPXIe5753) can be used to generate a sine wave signal to provide an excitation signal to the probe. The output signal of the probe is connected and digitized by a data acquisition unit (NIPXIe5753). The amplitude of the excitation voltage is 3V and the frequency is 600kHz. The scanning step size is set to 0.5mm and the scanning speed is set to 10mm / s. The sliding window size in the algorithm is 10×10 and the moving step size is 3. The above image reconstruction algorithm can be implemented by using MATLAB software.

[0088] like Figure 8 As shown, the present application also provides a pipeline defect detection device 100 , which includes: an acquisition module 10 , a reconstruction module 20 , a segmentation module 30 , a processing module 40 and a generation module 50 .

[0089] Among them, the acquisition module 10 is used to acquire the pipeline image. The reconstruction module 20 is used to reconstruct the pipeline image based on the image patch model and generate a reconstructed image corresponding to the pipeline image. The segmentation module 30 is used to segment the reconstructed image and determine the background information matrix and defect information matrix corresponding to the reconstructed image. The processing module 40 is used to optimize the background information matrix and the defect information matrix based on the accelerated proximal gradient algorithm to obtain the pipeline background image and the pipeline defect image. The generation module 50 is used to fuse the pipeline background image and the pipeline defect image to generate the target pipeline image.

[0090] In one embodiment, the processing module 40 is further used to perform two-dimensional median filtering on the pipeline image, reconstruct the pipeline image after filtering based on the image patch model, and generate a reconstructed image corresponding to the pipeline image.

[0091] In one embodiment, the processing module 40 is also used to divide the reconstructed image into a background information component, a defect information component and a noise information component based on a stable principal component analysis algorithm, and determine the background information matrix and the defect information matrix corresponding to the reconstructed image according to the background information component, the defect information component and the noise information component.

[0092] In one embodiment, the processing module 40 is further used to convert the background information matrix and the defect information matrix into a convex optimization problem, and solve the convex optimization problem by an accelerated proximal gradient algorithm to obtain the pipeline background image and the pipeline defect image.

[0093] In one embodiment, the processing module 40 is further used to filter the pipeline background image and the pipeline defect image, and fuse the filtered pipeline background image and the pipeline defect image to generate a target pipeline image.

[0094] In one embodiment, the processing module 40 is also used to process the pipeline background image and the pipeline defect image through a one-dimensional median filtering algorithm to obtain a first filtered image corresponding to the pipeline background image and the pipeline defect image, and to process the first filtered image through a two-dimensional mean filtering algorithm to obtain a target filtered image corresponding to the pipeline background image and the pipeline defect image.

[0095] The present application also provides an electronic device, comprising: a processor; and a memory for storing processor executable instructions; wherein the processor implements the above-mentioned pipeline defect detection method by running the executable instructions.

[0096] It should be noted that the systems, devices or modules described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer, and the specific form of the computer can be a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email transceiver, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0097] In a typical configuration, a computer includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0098] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0099] Computer-readable media include permanent and non-permanent, removable and non-removable media that can be used to store information by any method or technology. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, disk storage, quantum memory, graphene-based storage media or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include temporary computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0100] The terms and expressions used herein are for descriptive purposes only, and the present invention should not be limited to these terms and expressions. The use of these terms and expressions does not mean to exclude any equivalent features of the illustrations and descriptions (or parts thereof), and it should be recognized that various modifications that may exist should also be included in the scope of the claims. Other modifications, variations and substitutions may also exist. Accordingly, the claims should be deemed to cover all such equivalents.

[0101] Similarly, it should be pointed out that although the present invention has been described with reference to the current specific embodiments, ordinary technicians in this technical field should realize that the above embodiments are only used to illustrate the present invention, and various equivalent changes or substitutions may be made without departing from the spirit of the present invention. Therefore, as long as the changes and modifications to the above embodiments are within the scope of the essential spirit of the present invention, they will fall within the scope of the claims of the present invention.

Claims

1. A pipeline defect detection method, applied to array probe pipeline detection, characterized in that: include: Step 1: Get pipeline image; Step 2: reconstructing the pipeline image based on the image patch model to generate a reconstructed image corresponding to the pipeline image; Step 3: Segment the reconstructed image to determine the background information matrix and defect information matrix corresponding to the reconstructed image; Step 4: Optimizing the background information matrix and the defect information matrix based on the accelerated proximal gradient algorithm to obtain a pipeline background image and a pipeline defect image; Step 5: Fusing the pipeline background image and the pipeline defect image to generate a target pipeline image.

2. The pipeline defect detection method according to claim 1, characterized in that: The pipeline image includes a pipeline axial channel image and a pipeline circumferential channel image.

3. The pipeline defect detection method according to claim 1, characterized in that: The step 2 comprises: Step 2-1: performing two-dimensional median filtering on the pipeline image; Step 2-2: reconstruct the pipeline image after filtering based on the image patch model to generate a reconstructed image corresponding to the pipeline image.

4. The pipeline defect detection method according to claim 1, characterized in that: The step 3 comprises: Step 3-1: Based on a stable principal component analysis algorithm, the reconstructed image is divided into a background information component, a defect information component and a noise information component; Step 3-2: Determine a background information matrix and a defect information matrix corresponding to the reconstructed image according to the background information component, the defect information component and the noise information component.

5. The pipeline defect detection method according to claim 1, characterized in that: The step 4 comprises: Step 4-1: Convert the background information matrix and the defect information matrix into a convex optimization problem; Step 4-2: Solve the convex optimization problem by using the accelerated proximal gradient algorithm to obtain the pipeline background image and the pipeline defect image.

6. The pipeline defect detection method according to claim 1, characterized in that: The step 5 comprises: Step 5-1: performing filtering processing on the pipeline background image and the pipeline defect image; Step 5-2: Fusing the filtered pipeline background image and the pipeline defect image to generate the target pipeline image.

7. The pipeline defect detection method according to claim 6, characterized in that: The step 5-1 comprises: Step 5-1-1: Processing the pipeline background image and the pipeline defect image by a one-dimensional median filtering algorithm to obtain a first filtered image corresponding to the pipeline background image and the pipeline defect image; Step 5-1-2: Process the first filtered image using a two-dimensional mean filtering algorithm to obtain a target filtered image corresponding to the pipeline background image and the pipeline defect image.

8. A pipeline defect detection device, applied to array probe pipeline detection, characterized in that: include: An acquisition module, used to acquire pipeline images; A reconstruction module, used to reconstruct the pipeline image based on an image patch model to generate a reconstructed image corresponding to the pipeline image; A segmentation module, used to segment the reconstructed image and determine a background information matrix and a defect information matrix corresponding to the reconstructed image; A processing module, used for optimizing the background information matrix and the defect information matrix based on an accelerated proximal gradient algorithm to obtain a pipeline background image and a pipeline defect image; The generation module is used to fuse the pipeline background image and the pipeline defect image to generate a target pipeline image.

9. An electronic device, comprising: processor; A memory for storing processor executable instructions; wherein the processor implements the pipeline defect detection method according to any one of claims 1 to 7 by running the executable instructions.

10. A non-volatile readable storage medium storing computer-readable instructions, wherein when the computer-readable instructions are executed by a processor, the processor executes a method configured for pipeline defect detection as described in any one of claims 1 to 7 or implements the function of an apparatus configured for pipeline defect detection as described in claim 8.

Images