PCCP broken wire signal detection method based on improved YOLOv8

Abstract

The invention discloses a PCCP broken wire signal detection method based on improved YOLOv8, and belongs to the technical field of hydraulic engineering detection. The method comprises the following steps: S1, carrying out a PCCP broken wire prototype test by adopting a distributed optical fiber sensor, collecting sound wave signals generated in the test process, and sorting and screening the collected data to obtain a PCCP broken wire one-dimensional sound wave signal data set; s2, processing and converting the one-dimensional sound wave signal into a two-dimensional spectrogram by adopting synchronous extrusion wavelet transform, labeling, and constructing a data set; s3, training the improved YOLOv8 model by using the data set in the S2, and carrying out broken wire signal identification detection by using the improved YOLOv8 model; performance evaluation is carried out, and the model meeting the evaluation requirement serves as a final target detection model and is used for detecting the wire breakage signal under the PCCP body running state. According to the invention, the problem of insufficient detection and identification of the broken wire signal in the operation process of the PCCP body can be solved, and accurate detection is realized.

Description

Technical Field

[0001] The invention belongs to the technical field of water conservancy project detection, and particularly relates to a PCCP broken wire signal detection method based on improved YOLOv8. Background Art

[0002] PCCP (Pre-stressed Concrete Cylinder Pipe) is widely used in long-distance water supply projects due to its excellent pressure-bearing capacity and durability. Prestressed steel wire is the main pressure-bearing structure of PCCP. Once the wire is broken, the internal and external pressure will increase, leading to tensile failure of the steel wire. The breakage of prestressed steel wire is the main cause of PCCP failure. If the wire breakage develops further, it may cause cracks in the pipe body or even burst the pipe, causing the interruption of water supply projects and resulting in property and life losses. Therefore, it is necessary to monitor the broken wire signal in real time to provide reliable information for the safety assessment of PCCP. However, the reconstruction and analysis of PCCP broken wire signals is still an engineering problem.

[0003] The monitoring system based on distributed optical fiber technology can filter data through preliminary sorting, identify noise similar to prestressed steel wire breakage, and capture and record possible wire breakage signals, which require further analysis and confirmation. The YOLO target detection model has the advantage of strong real-time performance, and can simultaneously complete target detection and classification in one forward propagation, achieving fast end-to-end processing. The YOLOv8 target detection model further improves the detection accuracy and speed. Thanks to its optimized network architecture and more efficient feature extraction method, it can better balance lightweight and performance. Secondly, YOLOv8 supports flexible model fine-tuning and adaptive anchor frame mechanism, making it more versatile and adaptable in different scenarios. Therefore, the YOLOv8 detection model can be used to process the acoustic wave signals monitored during the operation of the PCCP, which is of great significance for realizing the detection of wire breakage signals under the operation state of the PCCP pipe body. Summary of the invention

[0004] The present invention aims to solve one of the technical problems in the above-mentioned related art at least to a certain extent.

[0005] To this end, the purpose of the present invention is to provide a PCCP broken wire signal detection method based on improved YOLOv8, which can solve the problem of insufficient detection and identification of broken wire signals during the operation of the PCCP pipe body and realize accurate detection of PCCP broken wire signals.

[0006] In order to solve the above-mentioned technical problems, the present invention is achieved as follows:

[0007] The embodiment of the present invention provides a PCCP broken wire signal detection method based on improved YOLOv8, and the steps of the method include:

[0008] S1: Use distributed optical fiber sensors to conduct PCCP broken wire prototype tests, collect acoustic wave signals generated during the PCCP broken wire prototype tests, sort and filter the collected data, and obtain a one-dimensional acoustic wave signal data set for PCCP broken wires;

[0009] S2: Use synchronous squeezing wavelet transform to process the obtained one-dimensional sound wave signal data, convert it into a two-dimensional spectrum graph, annotate the two-dimensional spectrum graph, and construct a broken wire signal data set that can be input into the YOLOv8 model;

[0010] S3: Use the data set of S2 to train the improved YOLOv8 model to obtain the PCCP broken wire signal detection model, and use the PCCP broken wire signal detection model for PCCP broken wire signal recognition and detection; then use a variety of evaluation indicators to evaluate the detection performance of the PCCP broken wire signal detection model. If it does not meet the evaluation requirements, continue training to obtain a model that meets the evaluation requirements as the final target detection model, which is used for the detection of broken wire signals under the operation state of the PCCP pipe body.

[0011] In addition, the PCCP broken wire signal detection method based on improved YOLOv8 according to the present invention may also have the following additional technical features:

[0012] In some of the embodiments, the improved YOLOv8 model includes:

[0013] A dual-level routing mechanism is introduced in the backbone part of YOLOv8 to replace the original attention mechanism, so that the model can capture both global and local feature information.

[0014] In some of the embodiments, the improved YOLOv8 model includes:

[0015] In the neck part of YOLOv8, the original standard convolution module is replaced by a lightweight convolution module, and the original C2f module is replaced by a VoV-GSCSP module.

[0016] In some of the embodiments, the improved YOLOv8 model includes:

[0017] The detection head of YOLOv8 is simplified to build a lightweight detection head.

[0018] In some implementations, the synchronous squeezing wavelet transform formula in step S2 is:

[0019]

[0020]

[0021]

[0022]

[0023]

[0024]

[0025] Among them, W s (m,n) is the wavelet coefficient, s(t) is the time domain signal, is the wavelet mother function, is the wavelet function The conjugate complex number of , m>0, is the scale parameter, and n∈R is the time parameter. and They are respectively wavelet mother functions and the Fourier transform of the time domain signal s(t), ω is the frequency, ε is the angular frequency, j is the imaginary unit, j 2 = -1, A is the amplitude, ω s (m,n) is the instantaneous frequency, T s (ω l , n) is the synchronous squeezing transform value; δ(ε) is the Dirac delta function, which is an ideal impulse function in the frequency domain; ω l is the center frequency; δ(ε-ω) and δ(ε+ω) represent the spectral components at the angular frequencies ω and -ω, is the partial derivative of the time parameter, Δω=ω l -ω l-1 is the frequency conversion value, m k is the discrete scale, Δm k =m k -m k-1 .

[0026] In some implementations, in step S2, a labeling tool labelImg is used to label the location where the broken wire signal appears in the two-dimensional image.

[0027] In some of the embodiments, simplifying the detection head of YOLOv8 is specifically simplifying the detection head into a 1×1 convolution layer and a 3×3 convolution layer.

[0028] In some implementations, in step S3, the mean average precision (mAP) at an IOU threshold of 0.5 is used as an evaluation indicator of the model detection performance.

[0029] In some of the embodiments, in step S3, an inter-module ablation experiment is performed on the PCCP broken wire signal detection model and compared to evaluate the performance of each module.

[0030] In some embodiments, the contents of the PCCP broken wire prototype test using the distributed optical fiber sensor in step S1 include:

[0031] Clean the inner concrete wall of the PCCP to make the sensor fit better on the pipe;

[0032] The optical fiber sensor is placed on the pipe wall in a spiral winding manner to collect the acoustic wave signals generated during the test. The optical fiber sensor is initially fixed with self-spray paint and quick-drying glue, and then epoxy resin and carbon fiber impregnation glue are applied to the optical fiber surface for hardening protection.

[0033] A hydrogen embrittlement wire breaking test is carried out at a position of 180° circumferentially and 2-3m longitudinally on the PCCP pipe body by manually cutting the wires under different internal pressure conditions. The test ends when a large number of abnormal cracks appear in the outermost mortar or the pipeline leaks and cannot stabilize the pressure.

[0034] Compared with the prior art, the present invention has at least the following beneficial effects:

[0035] In the embodiment of the present invention, the PCCP broken wire signal detection method based on improved YOLOv8 is provided, which reduces parameter redundancy and improves calculation time through the improvement of YOLOv8 and the replacement of modules. The accurate broken wire signal detection result ensures the safety and reliability of pipeline operation.

[0036] In the embodiment of the present invention, the PCCP broken wire signal detection method based on improved YOLOv8 can accurately predict the broken wire situation during the operation of the PCCP pipe body structure, maintain high accuracy at a low training time, and does not increase the training time due to the change of the module structure;

[0037] In the embodiments of the present invention, the PCCP broken wire signal detection method based on improved YOLOv8 can effectively grasp the safety status of the PCCP pipe body during operation, and provide a new theory and method for the operation and maintenance of the actual pipeline structure.

[0038] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A flow chart of a PCCP broken wire signal detection method based on improved YOLOv8 disclosed in one embodiment of the present invention;

[0040] Figure 2A distributed optical fiber sensor disclosed in one embodiment of the present invention; wherein (a) is a sensing optical fiber with a diameter of 900 μm wrapped by glass fiber cloth; (b) is a schematic diagram of an optical fiber cross section;

[0041] Figure 3 A layout diagram of a distributed optical fiber sensor disclosed in one embodiment of the present invention;

[0042] Figure 4 A diagram of a test process of manually cutting broken wires disclosed in one embodiment of the present invention;

[0043] Figure 5 An effective one-dimensional sound wave signal diagram collected and sorted according to an embodiment of the present invention;

[0044] Figure 6 A spectrum diagram is formed by synchronous squeezing wavelet transform disclosed in one embodiment of the present invention;

[0045] Figure 7 This is a diagram of labeling a two-dimensional image using labelImg according to an embodiment of the present invention; wherein (a) is before labeling, and (b) is after labeling;

[0046] Figure 8 This is a screenshot of the exported .txt file list disclosed in one embodiment of the present invention;

[0047] Fig. 9 A schematic diagram of the overall structure of a model disclosed in one embodiment of the present invention;

[0048] Fig.10 A schematic diagram of the structure of a double-layer routing attention mechanism disclosed in one embodiment of the present invention;

[0049] Fig.11 A schematic diagram of the structure of a GSConv module disclosed in one embodiment of the present invention;

[0050] Fig.12 A schematic diagram of the VoV-GSCSP module structure disclosed in one embodiment of the present invention;

[0051] Fig.13 A schematic diagram of a model performance evaluation index disclosed in one embodiment of the present invention;

[0052] Fig.14 This is a diagram of the output of a broken wire signal detection model disclosed in an embodiment of the present invention; wherein (a) is the result of manual annotation, and (b) is the result of model detection.

[0053] Description of reference numerals:

[0054] 1. Fiber core; 2. Coating layer; 3. Inner protective layer; 4. Outer protective layer; 5. Polyelastic sheath; 6. Fiber optic sensor; 7. Glass fiber cloth. DETAILED DESCRIPTION

[0055] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] The embodiments of the present invention are described in detail below through specific embodiments and application scenarios in conjunction with the accompanying drawings.

[0057] See also Figure 1 As shown, in some embodiments of the present invention, in view of the lack of existing PCCP pipe body broken wire signal detection, especially the insufficiency of broken wire signal detection and identification during the operation of the PCCP pipe body, a PCCP broken wire signal detection method based on improved YOLOv8 is provided, and the detection and identification of broken wire signals during the operation of the PCCP pipe body are realized by establishing a PCCP broken wire signal detection model based on YOLOv8, and a model comparison and evaluation is performed, so as to provide an advanced and reliable broken wire signal detection method for the operation of the PCCP pipeline and ensure the safe operation of the pipeline.

[0058] In this embodiment, the PCCP broken wire signal detection method includes the following steps:

[0059] Step 1: Use distributed optical fiber sensors to conduct PCCP broken wire prototype tests to monitor the acoustic wave signals generated during the operation of PCCP. Based on the PCCP prototype test, collect the acoustic wave signals generated during the test, sort and filter the data, and construct a one-dimensional acoustic wave signal data set of PCCP broken wires.

[0060] In some embodiments of the present invention, step 1 further comprises:

[0061] Step 11: Before laying the optical fiber, simply clean the inner concrete pipe wall to make the sensor fit better on the pipe body;

[0062] Step 12: Select the NEN-DSS-C09 optical fiber sensor to monitor the broken wire signal. The layout of the optical fiber sensor includes: using self-spray paint and 502 quick-drying glue to initially fix the optical fiber, and then using epoxy resin and carbon fiber impregnation glue to brush on the surface of the optical fiber for hardening protection; the selected optical fiber structure is shown in the attached figure. Figure 2 As shown, the optical fiber is cylindrical and includes a core 1, a coating layer 2, an inner protective layer 3, an outer protective layer 4 and a polyelastic sheath 5 from the inside to the outside. When in use, a cylindrical optical fiber 6 is wrapped with a glass fiber cloth 7 and laid on the concrete pipe wall in a spiral winding manner. A total of 7 annular sections are selected, as shown in the attached Figure 3 As shown, they are located at 1.0m, 2.0m, 2.25m, 2.5m, 2.75m, 3.0m and 4.0m from the bottom of the pipe (the socket end);

[0063] Step 13: A hydrogen embrittlement wire breaking test is conducted at a position of 180° in the circumferential direction and 2-3m in the longitudinal direction of the pipe body. This is achieved by manually cutting the wires under different internal pressure conditions. During the test, the internal water pressure is increased in increments of 0.1MPa. When the internal water pressure reaches 0.1MPa, 0.4MPa, 0.8MPa and 1.15MPa respectively and the pressure is stabilized for 5 minutes, the wires are cut from the middle 2.5m to both ends, and 5 wires are cut each time. The test ends when a large number of abnormal cracks appear in the outermost mortar, or the pipeline leaks and cannot stabilize the pressure. The manually cut wires are as shown in the attached figure. Figure 4 As shown, the corresponding steel wire is cut off by using a cutting tool;

[0064] Step 14: During the prototype test, the MS-DAS2000II device was used to monitor the broken wire signal during the test, with a sampling frequency of 1kHz and a spatial resolution of 2m. The selected values ​​of the sampling frequency and spatial resolution are currently the most suitable parameters for this test. The sampling frequency of 1kHz means that 1000 data points are collected per second, and the time resolution reaches 1 millisecond, which has a high capture capability for fast transient signals (such as broken wire events) and can accurately record the specific time and characteristics of the broken wire event. The spatial resolution selected the lowest spatial resolution that can be selected by general instruments. Because the tube length is 5 meters, selecting too large a spatial distribution rate will increase the error.

[0065] Step 15: Screen and sort the signals collected during the test, remove the messy and irregular signals, select the signals with obvious characteristics, and place them separately in an Excel table. The acoustic wave signals with obvious broken wire signal characteristics are shown in the attached Figure 5 As shown in Figure 2, a one-dimensional acoustic wave signal dataset of PCCP broken wires was constructed.

[0066] Step 2: Use synchronous squeezing wavelet transform to process the one-dimensional sound wave signal dataset and convert it into a two-dimensional spectrum graph. Annotate and divide the two-dimensional broken wire image dataset to construct a dataset that can be input into the YOLO broken wire signal model.

[0067] In some embodiments of the present invention, step 2 further comprises:

[0068] Step 21: Use synchronous squeezing wavelet transform to transform the one-dimensional acoustic wave signal data of PCCP broken wire into a spectrum diagram, and construct a two-dimensional image data set of PCCP standardized broken wire signal. The schematic diagram of synchronous squeezing wavelet transform is shown in the attached figure. Figure 6 As shown; the synchronous squeezing wavelet transform formula is as follows:

[0069]

[0070]

[0071]

[0072]

[0073]

[0074]

[0075] Among them, W s (m,n) is the wavelet coefficient, s(t) is the time domain signal, is the wavelet mother function, is the wavelet function The conjugate complex number of , m>0, is the scale parameter, and n∈R is the time parameter. and They are respectively wavelet mother functions and the Fourier transform of the time domain signal s(t), ω is the frequency, ε is the angular frequency, j is the imaginary unit, j 2 = -1, A is the amplitude, ω s (m,n) is the instantaneous frequency, T s (ω l ,n) is the synchronous squeezing transform value; δ(ε) is the Dirac delta function, which is an ideal "impulse" function in the frequency domain; ω l is the center frequency, δ(ε-ω) and δ(ε+ω) represent the spectral components at the angular frequencies ω and -ω, is the partial derivative of the time parameter, Δω=ω l -ω l-1 is the frequency conversion value, m k is the discrete scale, Δm k =m k -m k-1 ;

[0076] Step 22: Use the labelImg labeling tool to mark the position where the broken wire signal appears in the two-dimensional image obtained by synchronous extrusion wavelet transform with a labeling box, as shown in the attached figure. Figure 7 As shown in the figure, select the YOLO format file from the labeling tool labelImg to save and export the marked file. The resulting file is in .txt format. Each broken wire signal data has its corresponding .txt file. Construct a broken wire signal data set that can be input into the YOLO model, as shown in the attached figure. Figure 8 As shown, attached Figure 8 The 1-16 annotation files obtained after the broken wire signal annotation collected at 0.1MPa are shown;

[0077] Step 23: The broken wire signal data set finally generated is divided into two-dimensional spectrograms of all acoustic wave signals, and 80% of them are randomly selected as the test set, and the remaining 20% ​​are used as the validation set.

[0078] Step 3: Construct a PCCP broken wire signal detection model based on YOLOv8, use the constructed target model for PCCP broken wire signal recognition and detection, use a variety of evaluation indicators to evaluate the detection performance of the model, and obtain a target detection model that can accurately detect PCCP broken wire signals, thereby achieving accurate detection of broken wire signals under the operation state of the PCCP pipe body.

[0079] In some embodiments of the present invention, step 3 further comprises:

[0080] Step 31: A novel dual-level routing mechanism (BiLevelRoutingAttention) is introduced after the SPPF module at the end of the backbone network of YOLOv8, as shown in the attached figure. Fig. 9 The dual-level routing mechanism uses sparsity to save computation and memory, and only involves GPU-friendly dense matrix multiplication, which enables the model to capture global and local feature information at the same time, thereby improving the performance of the model in visual tasks. The schematic diagram of the attention mechanism structure is shown in the attached figure. Fig.10 As shown;

[0081] Step 32: In the neck (feature enhancement network) part of YOLOv8, a lightweight convolution method (GSConv) is used to replace the original standard convolution (Conv). GSConv convolution is a mixed convolution of standard convolution (SC), depthwise separable convolution (DSC) and shuffle. DSC greatly improves the speed of the detector, but with lower accuracy. The cooperation of SC and DSC makes up for the shortcomings of DSC and improves the accuracy of the model. The module structure diagram of GSConv is shown in the attached figure. Fig.11 In addition, the VoV-GSCSP module is used to replace the original C2f module to reduce the computational complexity and reasoning time. The module VoV-GSCSP is obtained by introducing the GS bottleneck on the basis of GSConv and then using the one-time aggregation method to design a cross-stage partial network (GSCSP). The schematic diagram of the VoV-GSCSP structure is shown in the attached figure. Fig.12 The network structure after replacement is shown in the attached Fig. 9 As shown;

[0082] Step 33: Simplify the detection head of YOLOv8. The head uses the previously extracted features to identify them and is responsible for the final target detection and classification tasks. In the present invention, the original head structure is replaced with a 1×1 convolution layer and a 3×3 convolution layer to construct a lightweight detection head Detect_Efficient, which maintains a high detection accuracy and reduces the model calculation time while reducing the model parameters.

[0083] The overall structure diagram of the improved PCCP broken wire signal detection model based on YOLOv8 is shown in the attached figure. Fig. 9 As shown, its backbone network includes two levels of first Conv module, first C2f module, second Conv module, second C2f module, third Conv module, third C2f module, fourth Conv module, fourth C2f module, SPPF module and double-level routing mechanism module arranged in sequence; the feature enhancement network fuses feature maps of different scales through bottom-up path and top-down path, thereby realizing cross-scale transmission of information; the bottom-up path, also called the first path in this embodiment, includes a first Upsample module, a first Concat module, a first VoVGSCSP module, a second Upsample module, a second Concat module and a second VoVGSCSP module arranged in sequence; the top-down path, also called the second path in this embodiment, includes a first GSConv module, a third Concat module, a third VoVGSCSP module, a second GSConv module, a fourth Concat module and a fourth VoVGSCSP module arranged in sequence. The dual-level routing mechanism module of the backbone network is connected to the first Upsample module and the fourth VoVGSCSP module in two ways; the third C2f module is connected to the first Concat module; the second C2f module is connected to the second Concat module; the second VoVGSCSP module is connected to the first GSConv module; the first VoVGSCSP module is connected to the third Concat module. The second VoVGSCSP module, the third VoVGSCSP module and the fourth VoVGSCSP module are each connected to a lightweight detection head Detect_Efficient.

[0084] Step 34: Input the broken wire signal dataset into the model to evaluate the performance of the model. The average precision mAP with an IOU threshold of 0.5-0.9 is used. The evaluation index formula is as follows:

[0085]

[0086]

[0087]

[0088] Among them, IOU is an indicator to measure the degree of overlap between the predicted box and the true box. The original target (true box) is in range A, and the detected result (predicted box) is in range B. The value of IOU is between 0 and 1. IOU of 1 means that the predicted bounding box and the ground truth bounding box completely overlap. AP is the average precision, and mAP averages the AP values ​​of all categories. In the formula, N is the number of samples.

[0089] Step 35: Compare the proposed model with the modules through ablation experiments, evaluate the performance and function of each module, eliminate and replace the inappropriate modules, and verify the superiority of the model. The model detection results are shown in the attached figure. Fig.13 and attached Fig.14 shown.

[0090] Table 1 is a comparison table of the results of the ablation experiment and the same series of detection models. GFLOPs is a model performance parameter. The smaller the GFLOPs, the lower the memory usage and the faster the inference time. Training time refers to the training time of the model. The model comparison evaluation results show that the detection model proposed in the present invention can accurately predict the broken wire situation during the operation of the PCCP pipe body structure, and still maintains a high accuracy at a lower training time, and does not increase the training time due to the structural change of the module. The replacement of the module reduces parameter redundancy and improves the calculation time. The accurate broken wire signal detection results ensure the safety and reliability of pipeline operation.

[0091] Table 1 Comparison of ablation experiment and detection model results in the same series

[0092]

[0093]

[0094] The present invention improves the target detection model YOLOv8, realizes the accurate detection of broken wire signals during the operation of PCCP pipe body, and proposes a PCCP broken wire signal detection method based on improved YOLOv8. Based on the broken wire signal detection result, the safety status of the PCCP pipe body during the operation of the pipe body can be effectively grasped, which provides a new theory and method for the operation and maintenance of actual pipeline structures.

[0095] For parts of the present invention that are not described in detail, reference may be made to the prior art in the art or technologies known to those skilled in the art.

[0096] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

Claims

1. A PCCP broken wire signal detection method based on improved YOLOv8, characterized in that: The steps of the method include: S1: Use distributed optical fiber sensors to conduct PCCP broken wire prototype tests, collect acoustic wave signals generated during the PCCP broken wire prototype tests, sort and filter the collected data, and obtain a one-dimensional acoustic wave signal data set for PCCP broken wires; S2: Use synchronous squeezing wavelet transform to process the obtained one-dimensional sound wave signal data, convert it into a two-dimensional spectrum graph, annotate the two-dimensional spectrum graph, and construct a broken wire signal data set that can be input into the YOLOv8 model; S3: Use the data set of S2 to train the improved YOLOv8 model to obtain the PCCP broken wire signal detection model, and use the PCCP broken wire signal detection model for PCCP broken wire signal recognition and detection; then use a variety of evaluation indicators to evaluate the detection performance of the PCCP broken wire signal detection model. If it does not meet the evaluation requirements, continue training to obtain a model that meets the evaluation requirements as the final target detection model, which is used for the detection of broken wire signals under the operation state of the PCCP pipe body.

2. According to the PCCP broken wire signal detection method based on improved YOLOv8 according to claim 1, it is characterized in that: Improvements to the improved YOLOv8 model include: A dual-level routing mechanism is introduced in the backbone part of YOLOv8 to replace the original attention mechanism, so that the model can capture both global and local feature information.

3. The PCCP broken wire signal detection method based on improved YOLOv8 according to claim 1, characterized in that: Improvements to the improved YOLOv8 model include: In the neck part of YOLOv8, the original standard convolution module is replaced by a lightweight convolution module, and the original C2f module is replaced by a VoV-GSCSP module.

4. The PCCP broken wire signal detection method based on improved YOLOv8 according to claim 1, characterized in that: Improvements to the improved YOLOv8 model include: The detection head of YOLOv8 is simplified to build a lightweight detection head.

5. The PCCP broken wire signal detection method based on improved YOLOv8 according to claim 1, characterized in that: The synchronous squeezing wavelet transform formula in step S2 is: Among them, W s (m,n) is the wavelet coefficient, s(t) is the time domain signal, is the wavelet mother function, is the wavelet function The conjugate complex number of , m>0, is the scale parameter, and n∈R is the time parameter. and They are respectively wavelet mother functions and the Fourier transform of the time domain signal s(t), ω is the frequency, ε is the angular frequency, j is the imaginary unit, j 2 = -1, A is the amplitude, ω s (m,n) is the instantaneous frequency, T s (ω l , n) is the synchronous squeezing transform value; δ(ε) is the Dirac delta function, which is an ideal impulse function in the frequency domain; ω l is the center frequency, δ(ε-ω) and δ(ε+ω) represent the spectral components at the angular frequencies ω and -ω, is the partial derivative of the time parameter, Δω=ω l -ω l-1 is the frequency conversion value, m k is the discrete scale, Δm k =m k -m k-1 .

6. The PCCP broken wire signal detection method based on improved YOLOv8 according to claim 1, characterized in that: In step S2, a labeling tool labelImg is used to mark the location where the broken wire signal appears in the two-dimensional image.

7. The PCCP broken wire signal detection method based on improved YOLOv8 according to claim 4, characterized in that: The detection head of YOLOv8 is simplified into a 1×1 convolutional layer and a 3×3 convolutional layer.

8. The PCCP broken wire signal detection method based on improved YOLOv8 according to claim 1, characterized in that: In step S3, the mean average precision (mAP) with an IOU threshold of 0.5 is used as the evaluation indicator of the model detection performance.

9. The PCCP broken wire signal detection method based on improved YOLOv8 according to claim 1, characterized in that: In step S3, an ablation experiment is performed on the modules of the PCCP broken wire signal detection model and compared to evaluate the performance of each module.

10. The PCCP broken wire signal detection method based on improved YOLOv8 according to claim 1, characterized in that: The contents of the PCCP broken wire prototype test using the distributed optical fiber sensor in step S1 include: Clean the inner concrete wall of the PCCP to make the sensor fit better on the pipe; The optical fiber sensor is placed on the pipe wall in a spiral winding manner to collect the acoustic wave signals generated during the test. The optical fiber sensor is initially fixed with self-spray paint and quick-drying glue, and then epoxy resin and carbon fiber impregnation glue are applied to the optical fiber surface for hardening protection. A hydrogen embrittlement wire breaking test is carried out at a position of 180° circumferentially and 2-3m longitudinally on the PCCP pipe body by manually cutting the wires under different internal pressure conditions. The test ends when a large number of abnormal cracks appear in the outermost mortar or the pipeline leaks and cannot stabilize the pressure.

Images