Pipeline internal defect detection method and device based on ultrasonic waves

Abstract

The invention discloses a pipeline internal defect detection method and device based on ultrasonic waves, and relates to the technical field of ultrasonic detection. The method comprises the steps that when a target internal pipeline is detected, a water cavity device sprays a preset coupling liquid content to an ultrasonic phase control aging probe and the inner wall of the target internal pipeline; receiving state information; substituting the state information for calculation to obtain a transmission effective score; if the transmission effective score is smaller than the preset transmission score threshold value, the coupling liquid with the content is adjusted again; if the transmission effective score is greater than or equal to a preset transmission score threshold value, detecting the target internal pipeline according to the ultrasonic phased array to obtain a pipeline image; and importing the pipeline image into a target identification model to obtain a pipeline defect result. Coupling liquid is jetted through the water cavity equipment, energy loss of ultrasonic signals is reduced, and stable transmission of the signals is guaranteed; and through the target identification model, the pipeline interior identification capability is improved.

Description

Technical Field

[0001] This invention belongs to the field of ultrasonic testing technology, specifically relating to a method and apparatus for detecting internal defects in pipelines based on ultrasonic waves. Background Technology

[0002] With the development of ultrasonic testing technology, it is used in the field of nondestructive testing to detect defects and analyze the internal structure of various objects (such as industrial parts) using ultrasonic technology, ensuring that these objects meet safety, quality, and other relevant requirements during use. By fusing ultrasonic signals with photographic image information, the internal structural features of the object, the shape and location of potential defects (such as cracks, holes, impurities, etc.) are clearly presented. Through precise data analysis and judgment, a basis is provided for subsequent quality assessment, safety inspection, or fault repair.

[0003] The prior art (publication number: CN119245741A) discloses a quality inspection method and system for water conservancy engineering pipelines. The method involves acquiring water pressure data inside the pipeline and ultrasonic testing data of the pipeline wall; determining water pressure parameters of the pipeline based on water pressure data under different water flow rates, whereby the water pressure parameters characterize the degree to which the quality of the pipeline is affected by the water hammer effect; determining ultrasonic parameters of the pipeline based on the ultrasonic testing data, whereby the ultrasonic parameters characterize the defect information of the pipeline; determining load parameters of the pipeline based on the water pressure parameters, ultrasonic parameters, and the damage value of the pipeline; and determining the quality grade of the pipeline corresponding to the load parameters based on the pre-set correspondence between the load parameters and the quality grade of the pipeline.

[0004] The aforementioned patent uses ultrasonic parameters to inspect pipeline defects. However, in actual pipeline internal ultrasonic testing, factors such as oil stains and scale inside the pipeline can affect the effectiveness of ultrasonic transmission, thereby reducing the accuracy of defect detection. Furthermore, there is no dedicated identification model to identify pipeline defects, which reduces the identification capability. Summary of the Invention

[0005] The purpose of this invention is to address the problems that, in actual pipeline internal ultrasonic testing, factors such as oil stains and scale inside the pipeline can affect the effectiveness of ultrasonic transmission, thereby reducing the accuracy of defect detection; and the lack of a dedicated identification model for pipeline defects reduces the identification capability. Therefore, this invention proposes an ultrasonic-based method and device for detecting internal pipeline defects.

[0006] In a first aspect of this invention, a method for detecting internal defects in pipes based on ultrasound is first proposed, the method comprising:

[0007] A water cavity device is installed in front of the ultrasonic phased array probe. When detecting the internal pipe of the target, the water cavity device sprays a preset coupling fluid into the ultrasonic phased array probe and the inner wall of the internal pipe of the target; and receives the status information of the ultrasonic phased array probe.

[0008] The transmission validity score is obtained by calculating the status information and the content of the preset coupling fluid.

[0009] If the effective transmission score is less than the preset transmission score threshold, the content of the preset coupling fluid will be readjusted.

[0010] If the effective transmission score is greater than or equal to the preset transmission score threshold, the pipe image is obtained by detecting the internal pipe of the target using an ultrasonic phased array; the pipe image is then imported into the target recognition model to obtain the pipe defect result; the defect result includes the pipe crack shape, wear depth, and location.

[0011] Optionally, the calculation process for the transmission validity score includes:

[0012]

[0013] Where TC represents the transmission validity score, This indicates the content of the preset coupling fluid. This indicates the maximum value of the coupling fluid content. This indicates the signal amplitude value received by the target pipe. Indicates the optimal signal amplitude value. This represents the signal-to-noise ratio value received by the target pipe. This indicates the optimal signal-to-noise ratio value.

[0014] Optionally, the principle process of the target recognition model includes:

[0015] The target recognition model is an improvement upon the YOLOv8 model, specifically including:

[0016] In the neck structure, the C2f module is replaced with the C2f_CV module;

[0017] Replace all Bottlelneck modules in the C2f module with the CV_Bottlneck module;

[0018] The workflow of the CV_Bottlneck module specifically includes:

[0019] Use the output features of the second layer in the C2f_CV module as the input features;

[0020] The input features are successively substituted into the deep convolution (1×1) module and the deep convolution (3×3) module to obtain the first convolution feature and the second convolution feature;

[0021] Substitute the first convolutional feature and the second convolutional feature into the SE module to obtain the third feature;

[0022] Substitute the third feature into the Econv(1×1) module to obtain the output feature.

[0023] Optionally, the target recognition model can be further improved by adding a corresponding feature fusion module to the neck network.

[0024] Modify the input-output relationship between the sixteenth and seventeenth layers, using the output of the sixteenth layer as the input of the feature fusion module, and using the output of the feature fusion module as the input of the sixteenth layer;

[0025] The computational expressions of the feature fusion module include:

[0026]

[0027] Where T1 is the output feature map of the sixteenth layer, C3 is the output feature map of the third layer, Y1, Y2 and Y3 are feature maps generated during the calculation process, Y4 represents the output of the fusion feature module, and Y4 is used as the predicted feature map of the Detect_1 module; This indicates the operation of the C2f_CV module, upsample means upsampling, and concat means channel concatenation.

[0028] Optionally, obtaining the pipe image based on ultrasonic phased array detection of the internal pipes of the target includes:

[0029] A preset angle range and a preset step angle are obtained based on an ultrasonic phased array, and ultrasonic waves are emitted in a fan shape. The preset angle range is divided into multiple emission angles according to the preset step angle. The beam at the target emission angle is emitted, and the corresponding signal circuit diagram is obtained by receiving it through a probe. The signal circuit diagram is the signal amplitude value at different depths corresponding to the target emission angle. The target emission angle is any one of the multiple emission angles. The signal circuit diagrams corresponding to the multiple emission angles are obtained sequentially.

[0030] A two-dimensional image is constructed based on the multiple emission angles and the corresponding signal circuit diagrams, and the signal amplitude value is converted into a grayscale value; the emission angle is taken as the horizontal direction and the pipe depth is taken as the vertical direction; the two-dimensional image is determined as a pipe image.

[0031] In a second aspect of the invention, an ultrasonic-based pipe internal defect detection device is provided, comprising:

[0032] Data acquisition module: A water cavity device is installed in front of the ultrasonic phased array probe. When detecting the internal pipe of the target, the water cavity device sprays the content of a preset coupling fluid into the ultrasonic phased array probe and the inner wall of the target internal pipe; and receives the status information of the ultrasonic phased array probe.

[0033] Transmission validity module: Calculates the transmission validity score by combining the status information and the content of the preset coupling fluid;

[0034] Transmission adjustment module: If the effective transmission score is less than the preset transmission score threshold, the content of the preset coupling fluid will be readjusted;

[0035] Defect Result Module: If the effective transmission score is greater than or equal to the preset transmission score threshold, the pipe image is obtained by detecting the internal pipe of the target using an ultrasonic phased array; the pipe image is then imported into the target recognition model to obtain the pipe defect result; the defect result includes the pipe crack shape, wear depth, and location.

[0036] Optionally, the transmission validity module is further configured to transmit the calculation process of the validity score, including:

[0037]

[0038] Where TC represents the transmission validity score, This indicates the content of the preset coupling fluid. This indicates the maximum value of the coupling fluid content. This indicates the signal amplitude value received by the target pipe. Indicates the optimal signal amplitude value. This represents the signal-to-noise ratio value received by the target pipe. This indicates the optimal signal-to-noise ratio value.

[0039] Optionally, the defect result module is also used to define the principle process of the target recognition model:

[0040] The target recognition model is an improvement upon the YOLOv8 model, specifically including:

[0041] In the neck structure, the C2f module is replaced with the C2f_CV module;

[0042] Replace all Bottlelneck modules in the C2f module with the CV_Bottlneck module;

[0043] The workflow of the CV_Bottlneck module specifically includes:

[0044] Use the output features of the second layer in the C2f_CV module as the input features;

[0045] The input features are successively substituted into the deep convolution (1×1) module and the deep convolution (3×3) module to obtain the first convolution feature and the second convolution feature;

[0046] Substitute the first convolutional feature and the second convolutional feature into the SE module to obtain the third feature;

[0047] Substitute the third feature into the Econv(1×1) module to obtain the output feature.

[0048] Optionally, the target recognition model can be further improved by adding a corresponding feature fusion module to the neck network.

[0049] Modify the input-output relationship between the sixteenth and seventeenth layers, using the output of the sixteenth layer as the input of the feature fusion module, and using the output of the feature fusion module as the input of the sixteenth layer;

[0050] The computational expressions of the feature fusion module include:

[0051]

[0052] Where T1 is the output feature map of the sixteenth layer, C3 is the output feature map of the third layer, Y1, Y2 and Y3 are feature maps generated during the calculation process, Y4 represents the output of the fusion feature module, and Y4 is used as the predicted feature map of the Detect_1 module; This indicates the operation of the C2f_CV module, upsample means upsampling, and concat means channel concatenation.

[0053] Optionally, obtaining the pipe image based on ultrasonic phased array detection of the internal pipes of the target includes:

[0054] A preset angle range and a preset step angle are obtained based on an ultrasonic phased array, and ultrasonic waves are emitted in a fan shape. The preset angle range is divided into multiple emission angles according to the preset step angle. The beam at the target emission angle is emitted, and the corresponding signal circuit diagram is obtained by receiving it through a probe. The signal circuit diagram is the signal amplitude value at different depths corresponding to the target emission angle. The target emission angle is any one of the multiple emission angles. The signal circuit diagrams corresponding to the multiple emission angles are obtained sequentially.

[0055] A two-dimensional image is constructed based on the multiple emission angles and the corresponding signal circuit diagrams, and the signal amplitude value is converted into a grayscale value; the emission angle is taken as the horizontal direction and the pipe depth is taken as the vertical direction; the two-dimensional image is determined as a pipe image.

[0056] The beneficial effects of this invention are:

[0057] This invention proposes an ultrasonic-based method for detecting internal defects in pipes. A water cavity device is installed in front of an ultrasonic phased array probe. When detecting the internal pipe of a target, the water cavity device sprays a predetermined amount of coupling fluid onto the ultrasonic phased array probe and the inner wall of the target pipe. The method receives the status information of the ultrasonic phased array probe; calculates a transmission validity score based on the status information and the predetermined coupling fluid content; if the transmission validity score is less than a predetermined transmission score threshold, the predetermined coupling fluid content is readjusted; if the transmission validity score is greater than or equal to the predetermined transmission score threshold, a pipe image is obtained by detecting the internal pipe of the target using an ultrasonic phased array; the pipe image is then imported into a target recognition model to obtain the pipe defect results; the defect results include the shape, wear depth, and location of pipe cracks. This invention reduces ultrasonic signal energy loss and ensures stable signal transmission by spraying coupling fluid through a water cavity device, laying the foundation for detection. Simultaneously, it can receive probe status information and promptly detect anomalies. By calculating a transmission effectiveness score, the amount of coupling fluid can be dynamically adjusted, reducing costs and improving the accuracy of pipeline status information. When transmission quality meets standards, a target recognition model can accurately identify pipeline crack shapes, wear depths, and locations, thereby improving the accuracy of identifying internal pipeline defects. Attached Figure Description

[0058] The invention will now be further described with reference to the accompanying drawings.

[0059] Figure 1 A flowchart illustrating an ultrasonic-based method for detecting internal defects in pipelines, provided as an embodiment of the present invention;

[0060] Figure 2 This is a schematic diagram of the structure of a target recognition model provided in an embodiment of the present invention;

[0061] Figure 3 This is a schematic diagram of the structure of a YOLOv8 model provided in an embodiment of the present invention;

[0062] Figure 4 This is a structural schematic diagram illustrating the working principle of a C2f_CV module provided in an embodiment of the present invention;

[0063] Figure 5 This is a framework diagram of an ultrasonic-based pipe internal defect detection device provided in an embodiment of the present invention. Detailed Implementation

[0064] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The term "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and B can represent: A alone, A and B simultaneously, and B alone. Furthermore, descriptions involving "first," "second," etc., in this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" or "second" can explicitly or implicitly include at least one of those features. Additionally, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention.

[0065] Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0066] This invention provides a method for detecting internal defects in pipelines based on ultrasonic waves. See also... Figure 1 , Figure 1 A flowchart illustrating an ultrasonic-based method for detecting internal defects in pipelines, provided as an embodiment of the present invention. The method includes the following steps:

[0067] A water cavity device is installed in front of the ultrasonic phased array probe. When detecting the internal pipe of the target, the water cavity device sprays a preset coupling fluid into the ultrasonic phased array probe and the inner wall of the internal pipe of the target; and receives the status information of the ultrasonic phased array probe.

[0068] The effective transmission score is obtained by calculating the status information and the content of the preset coupling fluid.

[0069] If the effective transmission score is less than the preset transmission score threshold, the content of the preset coupling fluid will be readjusted.

[0070] If the effective transmission score is greater than or equal to the preset transmission score threshold, the pipe image is obtained by detecting the internal pipe of the target using ultrasonic phased array; the pipe image is then imported into the target recognition model to obtain the pipe defect result; the defect result includes the pipe crack shape, wear depth and location.

[0071] The present invention provides an ultrasonic-based method for detecting internal defects in pipelines. This method utilizes a water-cavity device to spray coupling fluid, reducing ultrasonic signal energy loss and ensuring stable signal transmission, thus laying the foundation for detection. Simultaneously, it can receive probe status information to promptly detect anomalies. By calculating the effective transmission score, the amount of coupling fluid can be dynamically adjusted, reducing costs and improving the accuracy of pipeline status information. When the transmission quality meets the standards, a target recognition model can accurately identify defect information such as the shape, wear depth, and location of pipeline cracks, thereby improving the accuracy of detecting internal pipeline defects.

[0072] Specifically, the preset coupling fluid content is set based on the staff's historical experience (the coupling fluid is usually water, or additives, lubricants, etc.), the specific scanning method of the ultrasonic phased array is a sector scanning mode; the water cavity device can be a water immersion tank, etc.; the status information includes signal amplitude value and signal-to-noise ratio value;

[0073] In one implementation, the calculation process for transmitting a valid score includes:

[0074]

[0075] Where TC represents the transmission validity score, This indicates the content of the preset coupling fluid. This indicates the maximum value of the coupling fluid content. This indicates the signal amplitude value received by the target pipe. Indicates the optimal signal amplitude value. This represents the signal-to-noise ratio value received by the target pipe. This indicates the optimal signal-to-noise ratio value.

[0076] In one implementation, the signal amplitude value represents the intensity of the reflected signal received by the probe. A larger amplitude indicates less sufficient coupling, more severe attenuation of ultrasonic energy, and a lower effective transmission score. The signal-to-noise ratio (SNR) represents the ratio of the effective signal to the background noise. A low SNR means greater noise interference, which may indicate insufficient coupling fluid, directly reducing detection accuracy and resulting in a lower effective transmission score. The signal propagation time represents the time it takes for the ultrasonic wave to travel from emission to reception. A deviation of the propagation time from the preset theoretical value may be due to uneven coupling fluid thickness, causing changes in the sound wave path and consequently a lower effective transmission score.

[0077] In one implementation, see [link to implementation details]. Figure 2 , Figure 2 This is a schematic diagram of the structure of a target recognition model provided in an embodiment of the present invention. The principle process of the target recognition model includes:

[0078] See Figure 3 , Figure 3This is a schematic diagram of the structure of a YOLOv8 model provided in an embodiment of the present invention. The target recognition model is obtained by improving the YOLOv8 model, and the specific improvements include:

[0079] The target recognition model is an improvement upon the YOLOv8 model, specifically including:

[0080] In the neck structure, the C2f module is replaced with the C2f_CV module;

[0081] Replace all Bottlelneck modules in the C2f module with the CV_Bottlneck module;

[0082] The workflow of the CV_Bottlneck module specifically includes:

[0083] Use the output features of the second layer in the C2f_CV module as the input features;

[0084] The input features are successively substituted into the deep convolution (1×1) module and the deep convolution (3×3) module to obtain the first convolution feature and the second convolution feature;

[0085] Substitute the first and second convolutional features into the SE module to obtain the third feature;

[0086] Substitute the third feature into the Econv(1×1) module to obtain the output feature.

[0087] In one implementation, the target recognition model is further improved by adding a corresponding feature fusion module to the neck network.

[0088] Modify the input-output relationship between the sixteenth and seventeenth layers, using the output of the sixteenth layer as the input of the feature fusion module, and the output of the feature fusion module as the input of the sixteenth layer;

[0089] The computational expressions for the feature fusion module include:

[0090]

[0091] Where T1 is the output feature map of the sixteenth layer, C3 is the output feature map of the third layer, Y1, Y2 and Y3 are feature maps generated during the calculation process, Y4 represents the output of the fusion feature module, and Y4 is used as the predicted feature map of the Detect_1 module; This indicates the operation of the C2f_CV module, upsample means upsampling, and concat means channel concatenation.

[0092] Specifically, it should be noted that the original LOYO8 model had two upsampling operations, and a third upsampling operation was performed on top of that to more accurately identify small and medium-sized target defects inside the pipe. The pipe is mainly composed of complex background textures, lighting variations, and potential mechanical scratches, among other interfering factors.

[0093] In one implementation, see [link to implementation details]. Figure 4 , Figure 4 This is a schematic diagram illustrating the working principle of a C2f_CV module according to an embodiment of the present invention. The working principle of the C2f_CV module includes:

[0094] Replace all Bottlelneck modules in the C2f module with the CV_Bottlneck module;

[0095] The workflow of the CV_Bottlneck module specifically includes:

[0096] Use the output features of the second layer in the C2f_CV module as the input features;

[0097] The input features are successively substituted into the deep convolution (1×1) module and the deep convolution (3×3) module to obtain the first convolution feature and the second convolution feature;

[0098] Substitute the first and second convolutional features into the SE module to obtain the third feature;

[0099] Substitute the third feature into the Econv(1×1) module to obtain the output feature.

[0100] In one implementation, the CV_Bottlneck module contains n modules. Specifically, the C2f_CV module's ability to quickly capture pipeline defect information, optimize multi-scale feature extraction, and suppress background noise enables higher detection accuracy in complex scenarios. Furthermore, a more efficient feature processing mechanism reduces computational burden, thereby improving resource utilization.

[0101] In one implementation, obtaining a pipe image based on ultrasonic phased array detection of the internal pipes of a target includes:

[0102] The ultrasonic phased array is used to obtain a preset angle range and a preset step angle, and ultrasonic waves are emitted in a fan shape. The preset angle range is divided into multiple emission angles according to the preset step angle. The beam of the target emission angle is emitted and the corresponding signal circuit diagram is obtained by the probe. The signal circuit diagram is the signal amplitude value of different depths corresponding to the target emission angle. The target emission angle is any one of the multiple emission angles. The signal circuit diagrams corresponding to the multiple emission angles are obtained in sequence.

[0103] A two-dimensional image is constructed based on multiple transmission angles and corresponding signal circuit diagrams, and the signal amplitude value is converted into grayscale value; the transmission angle is taken as the horizontal direction and the pipe depth is taken as the vertical direction; the two-dimensional image is determined as the pipe image.

[0104] Specifically, the preset angle range and preset step angle are set by staff based on historical experience; for example, the preset angle range is (-30°). 0 30 0 (e.g., the preset step angle can be 0.5) 0 ,1 0 wait;

[0105] Specifically, it should be noted that the conversion of signal amplitude values ​​into grayscale values ​​and the calculation of transmission effectiveness scores by the target transmission intelligent model are interconnected. The higher the transmission effectiveness score, the higher the quality of the received signal amplitude values, and the more accurate the corresponding converted grayscale values. This indicates that the quality of the converted pipeline image is higher and the image pixels are clearer.

[0106] Based on the same inventive concept, this invention also provides an ultrasonic-based device for detecting internal defects in pipelines. See also... Figure 5 , Figure 5 A framework diagram of an ultrasonic-based pipe internal defect detection device provided for an embodiment of the present invention includes:

[0107] Data acquisition module: A water cavity device is installed in front of the ultrasonic phased array probe. When detecting the internal pipe of the target, the water cavity device sprays the content of a preset coupling fluid into the ultrasonic phased array probe and the inner wall of the target internal pipe; and receives the status information of the ultrasonic phased array probe.

[0108] Transmission validity module: Calculates the transmission validity score by combining the status information and the preset coupling fluid content;

[0109] Transmission adjustment module: If the effective transmission score is less than the preset transmission score threshold, the content of the preset coupling fluid will be readjusted;

[0110] Defect Result Module: If the effective transmission score is greater than or equal to the preset transmission score threshold, the pipe image is obtained by detecting the internal pipe of the target using ultrasonic phased array; the pipe image is imported into the target recognition model to obtain the pipe defect result; the defect result includes the pipe crack shape, wear depth and location.

[0111] The ultrasonic-based pipe internal defect detection device provided in this invention reduces ultrasonic signal energy loss and ensures stable signal transmission by spraying coupling fluid through a water cavity device, laying the foundation for detection. Simultaneously, it can receive probe status information to promptly detect anomalies. The calculated transmission effectiveness score allows for dynamic adjustment of the coupling fluid dosage, reducing costs and improving the accuracy of pipe status information. When transmission quality meets standards, the target recognition model can accurately identify defect information such as pipe crack shape, wear depth, and location, thereby improving the accuracy of internal pipe defect detection.

[0112] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A method for detecting internal defects in pipelines based on ultrasound, characterized in that, The method includes: A water cavity device is installed in front of the ultrasonic phased array probe. When detecting the internal pipe of the target, the water cavity device sprays a preset coupling fluid into the ultrasonic phased array probe and the inner wall of the internal pipe of the target; and receives the status information of the ultrasonic phased array probe. The transmission validity score is obtained by calculating the status information and the content of the preset coupling fluid. If the effective transmission score is less than the preset transmission score threshold, the content of the preset coupling fluid will be readjusted. If the effective transmission score is greater than or equal to the preset transmission score threshold, the pipe image is obtained by detecting the internal pipe of the target using ultrasonic phased array; the pipe image is then imported into the target recognition model to obtain the pipe defect result; the defect result includes the pipe crack shape, wear depth, and location.

2. The method for detecting internal defects in pipelines based on ultrasound according to claim 1, characterized in that, The calculation process for the transmission validity score includes: Where TC represents the transmission validity score, This indicates the content of the preset coupling fluid. This indicates the maximum value of the coupling fluid content. This indicates the signal amplitude value received by the target pipe. Indicates the optimal signal amplitude value. This represents the signal-to-noise ratio value received by the target pipe. This indicates the optimal signal-to-noise ratio value.

3. The method for detecting internal defects in pipelines based on ultrasound according to claim 1, characterized in that, The principle and process of the target recognition model include: The target recognition model is an improvement upon the YOLOv8 model, specifically including: In the neck structure, the C2f module is replaced with the C2f_CV module; Replace all Bottlelneck modules in the C2f module with the CV_Bottlneck module; The workflow of the CV_Bottlneck module specifically includes: Use the output features of the second layer in the C2f_CV module as the input features; The input features are successively substituted into the deep convolution (1×1) module and the deep convolution (3×3) module to obtain the first convolution feature and the second convolution feature; Substitute the first convolutional feature and the second convolutional feature into the SE module to obtain the third feature; Substitute the third feature into the Econv(1×1) module to obtain the output feature.

4. The method for detecting internal defects in pipelines based on ultrasound according to claim 3, characterized in that, The target recognition model is further improved by adding a corresponding feature fusion module to the neck network. Modify the input-output relationship between the sixteenth and seventeenth layers, using the output of the sixteenth layer as the input of the feature fusion module, and using the output of the feature fusion module as the input of the sixteenth layer; The computational expressions of the feature fusion module include: Where T1 is the output feature map of the sixteenth layer, C3 is the output feature map of the third layer, Y1, Y2 and Y3 are feature maps generated during the calculation process, Y4 represents the output of the fusion feature module, and Y4 is used as the predicted feature map of the Detect_1 module; This indicates the operation of the C2f_CV module, upsample means upsampling, and concat means channel concatenation.

5. The method for detecting internal defects in pipelines based on ultrasound according to claim 1, characterized in that, The process of obtaining a pipe image based on ultrasonic phased array detection of the internal pipes of a target includes: A preset angle range and a preset step angle are obtained based on an ultrasonic phased array, and ultrasonic waves are emitted in a fan shape. The preset angle range is divided into multiple emission angles according to the preset step angle. The beam at the target emission angle is emitted, and the corresponding signal circuit diagram is obtained by receiving it through a probe. The signal circuit diagram is the signal amplitude value at different depths corresponding to the target emission angle. The target emission angle is any one of the multiple emission angles. The signal circuit diagrams corresponding to the multiple emission angles are obtained sequentially. A two-dimensional image is constructed based on the multiple emission angles and the corresponding signal circuit diagrams, and the signal amplitude value is converted into a grayscale value; the emission angle is taken as the horizontal direction and the pipe depth is taken as the vertical direction; the two-dimensional image is determined as a pipe image.

6. An ultrasonic-based device for detecting internal defects in pipelines, characterized in that, The device includes: Data acquisition module: A water cavity device is installed in front of the ultrasonic phased array probe. When detecting the internal pipe of the target, the water cavity device sprays the content of a preset coupling fluid into the ultrasonic phased array probe and the inner wall of the target internal pipe; and receives the status information of the ultrasonic phased array probe. Transmission validity module: Calculates the transmission validity score by combining the status information and the content of the preset coupling fluid; Transmission adjustment module: If the effective transmission score is less than the preset transmission score threshold, the content of the preset coupling fluid will be readjusted; Defect Result Module: If the effective transmission score is greater than or equal to the preset transmission score threshold, the pipe image is obtained by detecting the internal pipe of the target using an ultrasonic phased array; the pipe image is then imported into the target recognition model to obtain the pipe defect result; the defect result includes the pipe crack shape, wear depth, and location.

7. The ultrasonic-based pipe internal defect detection device according to claim 6, characterized in that, The transmission validity module is also used for the calculation process of the transmission validity score, including: Where TC represents the transmission validity score, This indicates the content of the preset coupling fluid. This indicates the maximum value of the coupling fluid content. This indicates the signal amplitude value received by the target pipe. Indicates the optimal signal amplitude value. This represents the signal-to-noise ratio value received by the target pipe. This indicates the optimal signal-to-noise ratio value.

8. The ultrasonic-based pipe internal defect detection device according to claim 6, characterized in that, The defect result module is also used to explain the principle process of the target recognition model: The target recognition model is an improvement upon the YOLOv8 model, specifically including: In the neck structure, the C2f module is replaced with the C2f_CV module; Replace all Bottlelneck modules in the C2f module with the CV_Bottlneck module; The workflow of the CV_Bottlneck module specifically includes: Use the output features of the second layer in the C2f_CV module as the input features; The input features are successively substituted into the deep convolution (1×1) module and the deep convolution (3×3) module to obtain the first convolution feature and the second convolution feature; Substitute the first convolutional feature and the second convolutional feature into the SE module to obtain the third feature; Substitute the third feature into the Econv(1×1) module to obtain the output feature.

9. The ultrasonic-based pipe internal defect detection device according to claim 8, characterized in that, The target recognition model is further improved by adding a corresponding feature fusion module to the neck network. Modify the input-output relationship between the sixteenth and seventeenth layers, using the output of the sixteenth layer as the input of the feature fusion module, and using the output of the feature fusion module as the input of the sixteenth layer; The computational expressions of the feature fusion module include: Where T1 is the output feature map of the sixteenth layer, C3 is the output feature map of the third layer, Y1, Y2 and Y3 are feature maps generated during the calculation process, Y4 represents the output of the fusion feature module, and Y4 is used as the predicted feature map of the Detect_1 module; This indicates the operation of the C2f_CV module, upsample means upsampling, and concat means channel concatenation.

10. The ultrasonic-based pipe internal defect detection device according to claim 6, characterized in that, The process of obtaining a pipe image based on ultrasonic phased array detection of the internal pipes of a target includes: A preset angle range and a preset step angle are obtained based on an ultrasonic phased array, and ultrasonic waves are emitted in a fan shape. The preset angle range is divided into multiple emission angles according to the preset step angle. The beam at the target emission angle is emitted, and the corresponding signal circuit diagram is obtained by receiving it through a probe. The signal circuit diagram is the signal amplitude value at different depths corresponding to the target emission angle. The target emission angle is any one of the multiple emission angles. The signal circuit diagrams corresponding to the multiple emission angles are obtained sequentially. A two-dimensional image is constructed based on the multiple emission angles and the corresponding signal circuit diagrams, and the signal amplitude value is converted into a grayscale value; the emission angle is taken as the horizontal direction and the pipe depth is taken as the vertical direction; the two-dimensional image is determined as a pipe image.

Images