Pipeline inner wall defect detection method based on intelligent sensor pipeline robot and related device

Abstract

This application discloses a method and related device for detecting defects in the inner wall of a pipeline based on an intelligent sensor pipeline robot, relating to the field of pipeline inner wall defect detection technology. The method includes: firstly, acquiring and preprocessing a multi-dimensional sensor data set of the target pipeline to obtain standardized sensor data; then, constructing a first adjustment model based on the inherent attribute parameters of the pipeline and the usage environment parameters to determine the initial parameter configuration of the defect recognition model; next, inputting the standardized sensor data into the defect recognition model, extracting target defect features and classifying them to obtain the defect detection result at the current moment; finally, calculating the gradient of the loss function based on the detection result at the current moment and the true label, and using the gradient to iteratively update the model parameters to obtain the model parameters for data processing at the next moment, forming a continuous self-optimizing closed-loop mechanism, providing reliable technical support for the safe operation of pipelines.

Description

Technical Field

[0001] This application relates to the field of pipeline internal wall defect detection technology, and in particular to a pipeline internal wall defect detection method and related device based on an intelligent sensor pipeline robot. Background Technology

[0002] In industrial production and infrastructure construction, pipelines serve as the core carriers for fluid transportation, widely used in various fields such as petroleum, natural gas, chemical industry, and municipal water supply. Pipelines operate in complex environments for extended periods, exposed to factors such as media corrosion, temperature variations, pressure fluctuations, and mechanical wear. Consequently, the inner walls of pipelines are highly susceptible to various defects, including cracks, corrosion, pitting, and wear. If these defects are not detected and addressed promptly, they can easily lead to pipeline leaks, ruptures, and other safety accidents, causing significant economic losses and posing a serious threat to the ecological environment and human safety.

[0003] Currently, pipeline internal wall defect detection mainly relies on manual inspection combined with conventional testing equipment. This approach suffers from low efficiency, limited coverage, and insufficient accuracy. For pipeline sections in concealed, confined, or high-risk environments, manual inspection is difficult to implement effectively. With the development of sensor and robotics technologies, pipeline robots equipped with intelligent sensors are increasingly being applied to pipeline inspection scenarios. These robots can enter the pipeline interior to collect multi-dimensional data on the internal wall, providing a new technical means for pipeline defect detection. However, existing pipeline robot-based inspection methods are susceptible to interference from environmental noise, and the defect recognition algorithms lack adaptive capabilities, failing to dynamically adjust model parameters based on changes in pipeline materials and service environment. This results in insufficient accuracy and reliability of defect detection under complex operating conditions, making it difficult to meet the high-precision inspection requirements of industrial pipelines. Summary of the Invention

[0004] The purpose of this application is to provide a method and related device for detecting defects in the inner wall of pipelines based on intelligent sensor pipeline robots, which can improve the accuracy and adaptability of pipeline inner wall defect detection.

[0005] To achieve the above objectives, this application provides the following solution: In a first aspect, this application provides a method for detecting defects in the inner wall of a pipeline based on an intelligent sensor pipeline robot, comprising the following steps: The system acquires a set of sensor data collected by the intelligent sensor pipeline robot from the target pipeline, and preprocesses the sensor data set to obtain standardized sensor data.

[0006] Obtain the inherent property parameters and usage environment parameters of the target pipeline, and construct a first adjustment model based on the inherent property parameters and usage environment parameters to determine the initial parameter configuration for defect identification.

[0007] Standardized sensor data is input into the defect identification model to extract target defect features and classify defects, thereby obtaining the defect detection results at the current moment; the initial parameters of the defect identification model are determined according to the initial parameter configuration.

[0008] Based on the defect detection results and the true labels at the current moment, the gradient of the loss function is calculated, and the parameters of the defect recognition model are iteratively updated using the gradient of the loss function to obtain the model parameters for data processing at the next moment.

[0009] Optionally, the sensor dataset is a multi-dimensional sensor dataset; the sensor dataset of the target pipeline collected by the intelligent sensor pipeline robot is obtained, and the sensor dataset is preprocessed to obtain standardized sensor data, specifically including the following steps: Acquire sensing data of the inner wall of the pipe from at least two different types of sensors mounted on an intelligent sensor pipeline robot, forming a multi-dimensional sensing data set.

[0010] The multi-dimensional sensor data set is denoised to obtain denoised sensor data.

[0011] The noise reduction sensor data is standardized by standardizing the data units to obtain standardized sensor data.

[0012] Optionally, the inherent property parameters include pipe material parameters, and the usage environment parameters include environmental state parameters; a first adjustment model is constructed based on the inherent property parameters and the usage environment parameters to determine the initial parameter configuration for defect identification, specifically including the following steps: Based on pipeline material parameters and environmental condition parameters, a parameter mapping function is determined.

[0013] An ensemble learning algorithm is used to learn the influence of pipeline material parameters and environmental state parameters on defect identification, and a parameter adjustment weight vector is obtained through a parameter mapping function.

[0014] Adjust the weight vector according to the parameters and initialize at least one of the feature extraction threshold, classifier penalty coefficient, and learning rate of the defect identification model.

[0015] Optionally, the defect identification model includes a feature extraction network and a classification network; standardized sensor data is input into the defect identification model to extract target defect features and classify defects to obtain the defect detection result at the current time, specifically including the following steps: Spatial and temporal features are extracted from standardized sensor data using a feature extraction network, and then the spatial and temporal features are fused to obtain a comprehensive feature vector.

[0016] The comprehensive feature vector is input into the classification network to obtain the predicted probability distribution of various defects. Based on the predicted probability distribution of various defects, the defect category with the highest predicted probability is determined as the defect detection result.

[0017] Optionally, the feature extraction network includes a convolutional neural network layer for extracting spatial features and a long short-term memory network layer for extracting temporal correlation features.

[0018] Optionally, the loss function for training the defect identification model consists of a weighted sum of classification loss terms and regression loss terms.

[0019] Optionally, after obtaining the defect detection results, the method further includes the following steps: Obtain the defect type, location coordinates, severity level, and confidence level corresponding to the defect detection results.

[0020] A structured defect detection report is generated based on the defect type, location coordinates, severity level, and confidence level.

[0021] Optionally, after obtaining the defect detection results, the method further includes the following steps: The manually verified defect detection results, along with the corresponding sensor data and model parameters, are stored in the adaptive learning database.

[0022] The defect identification model is incrementally trained periodically using data from the adaptive learning database to update its parameters. During incremental training, a sliding window mechanism is used to filter training samples and remove abnormal data.

[0023] Secondly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the pipeline inner wall defect detection method based on intelligent sensor pipeline robot described above.

[0024] Thirdly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the pipe inner wall defect detection method based on an intelligent sensor pipe robot described above.

[0025] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application provides a method and related apparatus for detecting defects in the inner wall of a pipeline based on an intelligent sensor pipeline robot. The method first acquires a set of sensor data collected by the intelligent sensor pipeline robot from the target pipeline, and preprocesses the sensor data to obtain standardized sensor data, effectively reducing noise interference and improving data quality. Then, it acquires the inherent attribute parameters and usage environment parameters of the target pipeline, and constructs a first adjustment model based on these parameters to determine the initial parameter configuration for defect identification, ensuring that the detection initially matches the actual pipeline condition and enhancing the targeting and adaptability of the detection. Next, the standardized sensor data is input into the defect identification model to extract target defect features and classify defects, obtaining the defect detection result at the current moment. The initial parameters of the defect identification model are determined according to the initial parameter configuration. Finally, based on the defect detection result at the current moment and the true label, the gradient of the loss function is calculated, and the parameters of the defect identification model are iteratively updated using the gradient of the loss function to obtain the model parameters for data processing at the next moment. This forms a continuously self-optimizing closed-loop mechanism during real-time analysis, enabling the model to dynamically adapt to the complex and changing environment inside the pipeline and continuously improving the accuracy of defect identification. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 This is a flowchart illustrating a pipeline inner wall defect detection method based on an intelligent sensor pipeline robot, as provided in one embodiment of this application.

[0028] Figure 2 This is a flowchart of step S2 in a pipeline inner wall defect detection method based on an intelligent sensor pipeline robot provided in an embodiment of this application.

[0029] Figure 3 This is a schematic diagram of the defect identification model in a pipeline inner wall defect detection method based on an intelligent sensor pipeline robot, provided in an embodiment of this application.

[0030] Figure 4 This is a functional block diagram of a pipeline inner wall defect detection system based on an intelligent sensor pipeline robot, provided in one embodiment of this application.

[0031] Figure 5 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0032] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0033] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0034] This application provides a method for detecting defects in the inner wall of a pipeline based on an intelligent sensor pipeline robot. In one exemplary embodiment, such as... Figure 1 As shown, it includes the following steps: S1. Obtain the set of sensor data collected by the intelligent sensor pipeline robot on the target pipeline, and preprocess the set of sensor data to obtain standardized sensor data.

[0035] In this embodiment, the sensor data set specifically refers to the multi-dimensional sensor data set collected by the intelligent sensor pipeline robot. Specifically, the intelligent sensor pipeline robot can be equipped with various types of sensors, such as visual sensors, ultrasonic sensors, and electromagnetic sensors, to collect image data, thickness data, and structural integrity data of the pipeline's inner wall, thereby forming a comprehensive multi-dimensional sensor data set, denoted as... , Where K represents the number of sensor data dimensions (i.e., the number of sensor types), and M represents the number of data acquisition samples.

[0036] Specifically, in this embodiment, the preprocessing of the multi-dimensional sensing data set includes: Denoising is performed on the multi-dimensional sensor data set to remove environmental interference noise, resulting in denoised sensor data. The denoising process can employ an adaptive median filtering algorithm, the calculation formula of which is as follows: .

[0037] in, This represents the m-th sample data in the k-th dimension after noise reduction. Indicates the size of the filtering window. This represents the original k-th dimension m+t sample data.

[0038] The noise-reduced sensor data is standardized to unify its data units, resulting in the standardized sensor data. The expression for the standardization process is: .

[0039] in, That is, the standardized k-th dimension m-th sample data. and Let represent the minimum and maximum values ​​of all sensor data in the k-th dimension after noise reduction, respectively.

[0040] Implementing the design described in step S1 involves acquiring multi-dimensional data using various sensors to form a comprehensive dataset that accurately reflects the condition of the pipe's inner wall. An adaptive median filtering algorithm is employed to reduce noise, effectively removing environmental interference and improving data quality. This reduces noise interference in subsequent analysis. Standardized processing unifies data dimensions, making different types of data comparable and avoiding analytical biases caused by differences in dimensions. This provides a high-quality, standardized data foundation for subsequent defect detection, helping to improve the accuracy and reliability of defect identification and ensuring that the detection results truly reflect the actual defect situation of the pipe's inner wall.

[0041] S2. Obtain the inherent attribute parameters and usage environment parameters of the target pipeline, and construct a first adjustment model based on the inherent attribute parameters and usage environment parameters to determine the initial parameter configuration for defect identification.

[0042] Among them, the inherent property parameters mainly refer to the pipe material parameters. For example, it can be specified by the material hardness. Corrosion resistance rating thermal conductivity The vector formed Use environmental parameters It can be specified as being determined by temperature. ,humidity Corrosiveness of the medium The vector formed .

[0043] like Figure 2 As shown, step S2 specifically includes the following steps: S21, Based on pipe material parameters and usage environment parameters Determine a parameter mapping function .

[0044] S22. An ensemble learning algorithm is used to learn the influence weights of these parameters on defect identification. In this embodiment, the ensemble learning algorithm is specifically the gradient boosting tree algorithm, and the parameter mapping function learned by this algorithm is: .

[0045] in, This represents the parameter adjustment weight vector. T Indicates the number of weak learners. Indicates the first i The weight coefficients of a weak learner Indicates the first i The output of a weak learner.

[0046] S23. Adjust the weight vector according to the calculated parameters. This is used to initialize the core parameters of the subsequent defect identification model. These core parameters include, but are not limited to, feature extraction threshold, classifier penalty coefficient, and learning rate.

[0047] By designing step S2, pipe material property parameters and usage environment parameters are collected, fully considering various factors affecting pipe defects. The gradient boosting tree algorithm is used to learn the influence weight of parameters on defect identification, accurately grasp the degree of influence of different factors, and initialize the core parameters of the algorithm based on the weight vector. This makes the initial parameter configuration more in line with the actual situation, improves the adaptability of the algorithm to different pipe and environmental conditions, enhances the pertinence and accuracy of defect identification, avoids misjudgment or missed judgment due to unreasonable parameter settings, and effectively improves the practicality and effectiveness of the detection method.

[0048] S3. Input standardized sensor data into the defect identification model to extract target defect features and classify defects to obtain the defect detection results at the current moment; the initial parameters of the defect identification model are determined according to the initial parameter configuration.

[0049] In this embodiment, the defect identification model is a complete end-to-end deep neural network, such as... Figure 3 As shown, it consists of two main parts: a feature extraction network and a classification network. The initial parameters of the defect recognition model (including the weights of the feature extraction network and the classification network) are set according to the initial parameter configuration determined in step S2.

[0050] This step is an iterative process that unfolds continuously over time to process the standardized sensor data at time t. For example, the specific process is as follows: First, the standardized sensor data at time t Input the defect identification model. The data will flow through the following network layers in sequence: (1) Extract the defect features of the target through a feature extraction network.

[0051] Specifically, this feature extraction network consists of Convolutional Neural Network (CNN) layers and Long Short-Term Memory (LSTM) layers. The CNN layers are used to extract spatial features from the data, i.e. LSTM layers are used to capture temporal correlation features in the data, i.e. Subsequently, the two features are fused to obtain a comprehensive feature vector. .

[0052] (2) The comprehensive feature vector Input into the subsequent classification network (i.e., classifier).

[0053] Classification networks are based on input The calculations are performed to output the predicted probability distributions for various defects. The classification network can use an improved Softmax function as the output layer. .

[0054] in, This represents the predicted probability of the i-th type of defect. Let represent the score of the i-th type of defect, and C represent the total number of defect categories (which may include, but are not limited to, corrosion, cracks, wear, etc.). The defect category with the highest predicted probability is determined as the defect detection result. .

[0055] After obtaining the defect detection results, the system can also obtain the defect type, location coordinates, severity level, and confidence level corresponding to the defect detection results, and generate a structured defect detection report based on the defect type, location coordinates, severity level, and confidence level.

[0056] S4. Based on the defect detection results and the true labels at the current moment, calculate the gradient of the loss function, and use the gradient of the loss function to iteratively update the parameters of the defect recognition model to obtain the model parameters for data processing at the next moment.

[0057] The defect detection result at time t is obtained in step S3. Then, it is compared with the true label of the data at that moment. Perform comparisons and calculate the loss function. The value of . In this embodiment, the loss function Taking into account both classification accuracy and prediction error, the loss is a weighted sum of the classification loss and regression loss, which can be represented as, for example, as cross-entropy loss. ) and mean square error loss ( The weighted sum of ) .

[0058] in, This is the loss weighting coefficient.

[0059] Subsequently, the error signal is propagated back from the output of the classification network to the input of the feature extraction network using the backpropagation algorithm, thereby calculating the loss function with respect to the current model parameters. The gradient, i.e. .

[0060] Next, the calculated gradients are used to update the model parameters for use in the next time step. It's important to clarify that the gradients calculated here... Its purpose is not to modify the parameters at the current time that have already been calculated. Instead, it is used to calculate the model parameters at the next time step (t+1). The parameter update formula is: .

[0061] in, This represents the updated model parameters that will be used to process the data at time t+1. This represents the model parameters at time t. This represents the adaptive learning rate, which adjusts the weight vector based on the parameters obtained in step S2. Make dynamic adjustments.

[0062] New data at time t+1 When it arrives, the model will use the updated parameters. Repeat steps one through three above to achieve dynamic self-optimization of the model during continuous real-time analysis.

[0063] By implementing steps S3 and S4 above, a closed-loop mechanism of "complete forward propagation → loss calculation and gradient solution → iterative update for use in the next time step" is logically and rigorously solved, addressing the technical implementation problem of dynamically adjusting model parameters. This enables the model to continuously optimize itself based on real-time processing results and adapt to the complex and ever-changing environment within the pipeline.

[0064] In another exemplary embodiment of this application, after obtaining the defect detection result in step S3, the method further updates the adaptive learning database based on the detection result to continuously optimize the algorithm's recognition accuracy, including the following steps: The manually verified defect detection results, along with the corresponding sensor data and model parameters, are stored in the adaptive learning database. Specifically, the defect detection results (preferably those that have been manually verified) and the corresponding sensor data and model parameters are stored together in the adaptive learning database.

[0065] The defect identification model is incrementally trained periodically using data from the adaptive learning database to update its parameters. The system also periodically performs batch training based on historical data from this database, updating the model parameters using an incremental learning algorithm. The update formula can be expressed as: .

[0066] in, This represents the updated model parameters. This represents the model parameters before the update. This represents the parameter retention coefficient, used to control the degree to which existing model knowledge is retained. These are new parameters obtained based on a new batch of data. To ensure the stability and accuracy of model optimization, a sliding window mechanism is used to filter valid training samples and remove outlier data before each training session. Through this step, the algorithm model has the ability to continuously learn and self-evolve, constantly improving the accuracy of defect identification during use.

[0067] Based on the same inventive concept, this application also provides a system for implementing the above-described method for detecting defects in the inner wall of a pipe based on an intelligent sensor pipe robot. The solution provided by this system is similar to the implementation described in the above method; therefore, the specific limitations in one or more system embodiments provided below can be found in the above-described limitations of the method for detecting defects in the inner wall of a pipe based on an intelligent sensor pipe robot, and will not be repeated here.

[0068] In one exemplary embodiment, such as Figure 4 As shown, a pipeline inner wall defect detection system based on an intelligent sensor pipeline robot is provided, comprising: The data acquisition and preprocessing module is used to acquire the set of sensor data collected by the intelligent sensor pipeline robot from the target pipeline, and to preprocess the set of sensor data to obtain standardized sensor data.

[0069] The model building and initialization module is used to obtain the inherent attribute parameters and usage environment parameters of the target pipeline, and build a first adjustment model based on the inherent attribute parameters and usage environment parameters to determine the initial parameter configuration for defect identification.

[0070] The pipeline inner wall defect detection module is used to input standardized sensor data into the defect identification model to extract target defect features and classify defects, thereby obtaining the defect detection results at the current moment; the initial parameters of the defect identification model are determined according to the initial parameter configuration.

[0071] The model parameter iterative update module is used to calculate the gradient of the loss function based on the defect detection results and the true labels at the current time, and to use the gradient of the loss function to iteratively update the parameters of the defect recognition model to obtain the model parameters for data processing at the next time step.

[0072] certainly, Figure 4 The architecture shown is merely exemplary; it can be omitted as needed when implementing different functionalities. Figure 4 One or at least two components of the system shown.

[0073] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 5 As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When the computer program is executed by the processor, it can implement the pipeline inner wall defect detection method based on an intelligent sensor pipeline robot provided in the previous embodiment.

[0074] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0075] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0076] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0077] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0078] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Moreover, the collection, use and processing of the relevant data are carried out in compliance with the relevant data protection laws and policies of the country where the location is located, and with the authorization granted by the owner of the corresponding device.

[0079] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0080] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0081] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0082] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for detecting defects in the inner wall of a pipeline based on an intelligent sensor pipeline robot, characterized in that, include: Acquire the set of sensor data of the target pipeline collected by the intelligent sensor pipeline robot, and preprocess the set of sensor data to obtain standardized sensor data; Obtain the inherent attribute parameters and usage environment parameters of the target pipeline, and construct a first adjustment model based on the inherent attribute parameters and usage environment parameters to determine the initial parameter configuration for defect identification; The standardized sensor data is input into the defect identification model to extract target defect features and classify defects, thereby obtaining the defect detection result at the current moment. The initial parameters of the defect identification model are determined according to the initial parameter configuration; Based on the defect detection results and the true labels at the current moment, the gradient of the loss function is calculated, and the parameters of the defect recognition model are iteratively updated using the gradient of the loss function to obtain the model parameters for data processing at the next moment.

2. The method for detecting defects in the inner wall of a pipeline based on an intelligent sensor pipeline robot according to claim 1, characterized in that, The sensor data set is a multi-dimensional sensor data set; the sensor data set of the target pipeline collected by the intelligent sensor pipeline robot is obtained, and the sensor data set is preprocessed to obtain standardized sensor data, specifically including: Acquire sensing data of the inner wall of the pipe collected by at least two different types of sensors carried by the intelligent sensor pipeline robot, and form the multi-dimensional sensing data set. The multi-dimensional sensor data set is subjected to noise reduction processing to obtain noise-reduced sensor data. The noise-reduced sensing data is standardized to unify the data dimensions, resulting in the standardized sensing data.

3. The method for detecting defects in the inner wall of a pipeline based on an intelligent sensor pipeline robot according to claim 1, characterized in that, The inherent property parameters include pipe material parameters, and the usage environment parameters include environmental state parameters; a first adjustment model is constructed based on the inherent property parameters and the usage environment parameters to determine the initial parameter configuration for defect identification, specifically including: Based on the pipeline material parameters and the environmental state parameters, a parameter mapping function is determined; An ensemble learning algorithm is used to learn the influence of the pipe material parameters and the environmental state parameters on defect identification, and a parameter adjustment weight vector is obtained through the parameter mapping function. Adjust the weight vector according to the parameters, and initialize at least one of the feature extraction threshold, classifier penalty coefficient, and learning rate of the defect identification model.

4. The method for detecting defects in the inner wall of a pipeline based on an intelligent sensor pipeline robot according to claim 1, characterized in that, The defect identification model includes a feature extraction network and a classification network. The standardized sensor data is input into the defect identification model to extract target defect features and classify defects, obtaining the defect detection result at the current time. Specifically, this includes: The feature extraction network extracts spatial features and temporal correlation features from the standardized sensor data, and fuses the spatial features and temporal correlation features to obtain a comprehensive feature vector. The comprehensive feature vector is input into the classification network to obtain the predicted probability distribution of various defects. Based on the predicted probability distribution of various defects, the defect category with the highest predicted probability is determined as the defect detection result.

5. The method for detecting defects in the inner wall of a pipeline based on an intelligent sensor pipeline robot according to claim 4, characterized in that, The feature extraction network includes a convolutional neural network layer for extracting the spatial features and a long short-term memory network layer for extracting the temporal correlation features.

6. The method for detecting defects in the inner wall of a pipeline based on an intelligent sensor pipeline robot according to claim 4, characterized in that, The loss function for training the defect identification model is composed of a weighted sum of classification loss terms and regression loss terms.

7. The method for detecting defects in the inner wall of a pipeline based on an intelligent sensor pipeline robot according to claim 4, characterized in that, After obtaining the defect detection results, the method also includes: Obtain the defect type, location coordinates, severity level, and confidence level corresponding to the defect detection results; Based on the defect type, location coordinates, severity level, and confidence level, a structured defect detection report is generated.

8. The method for detecting defects in the inner wall of a pipeline based on an intelligent sensor pipeline robot according to claim 1, characterized in that, After obtaining the defect detection results, the method also includes: The manually verified defect detection results, along with the corresponding sensor data and model parameters, are stored in the adaptive learning database. The defect identification model is incrementally trained periodically using data from the adaptive learning database to update its model parameters; during the incremental training process, training samples are filtered using a sliding window mechanism to remove abnormal data.

9. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the pipeline inner wall defect detection method based on an intelligent sensor pipeline robot according to any one of claims 1-8.

10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the pipeline inner wall defect detection method based on an intelligent sensor pipeline robot as described in any one of claims 1-8.

Images