Underground drainage pipeline defect identification method based on depth surveying and mapping data
By using equipment such as lidar mounted on drones and mobile surveying vehicles to perform non-contact scanning of underground drainage pipes, combined with multi-view point cloud fusion algorithms and real-time data analysis, the problems of low efficiency and safety risks of traditional detection methods have been solved, achieving efficient and safe pipe defect identification.
Patent Information
- Application Number
- CN202511023984.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-24
- Publication Date
- 2025-11-11
Smart Images
Figure CN120929909A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of underground drainage pipeline defect identification technology, specifically a method for identifying underground drainage pipeline defects based on depth mapping data. Background Technology
[0002] Underground drainage pipes, as a key component of urban infrastructure, undertake important functions such as sewage discharge and rainwater diversion. Their operational status directly affects the normal functioning of the city and the quality of life of residents. However, due to the complex and harsh environment in which underground drainage pipes are exposed to various factors such as sewage corrosion, geological changes, tree root intrusion, and human damage, pipe defects are becoming increasingly prominent, including cracks, corrosion, leakage, blockages, and deformation. If these defects are not detected and addressed in a timely manner, they will not only lead to pipe failure and cause disasters such as urban flooding and sewage overflow, but also cause serious pollution to groundwater resources and the soil environment.
[0003] Traditional methods for detecting defects in underground drainage pipes mainly rely on manual visual inspection and simple instrument measurements. Manual visual inspection requires workers to enter the pipes and observe using tools such as flashlights and mirrors. This method is not only labor-intensive and inefficient, but also poses certain safety risks. Simple instrument measurements, such as using a mud measuring bucket to detect sediment thickness and a laser rangefinder to measure pipe deformation, can assist in detection to some extent, but these methods usually only obtain localized information about the pipes and cannot comprehensively and accurately reflect the overall condition and defect characteristics of the pipes. For some well-hidden defects, such as early-stage micro-cracks and internal corrosion, it is difficult to detect them in a timely and accurate manner. Summary of the Invention
[0004] To address the problems of the above-mentioned solutions, this invention provides a method for identifying defects in underground drainage pipes based on depth mapping data.
[0005] The objective of this invention can be achieved through the following technical solutions:
[0006] A method for identifying defects in underground drainage pipes based on depth mapping data includes:
[0007] Step 1: Obtain pipeline information for the underground drainage pipelines that require defect monitoring, and obtain the user's real-time requirements for pipeline monitoring; the platform will then build a defect identification model for the user to identify pipeline defects based on the real-time requirements and pipeline information.
[0008] Furthermore, the platform establishes a defect identification model for identifying pipeline defects based on real-time requirements and pipeline information, including:
[0009] The platform provider establishes a resource library, which stores the identification efficiency and implementation cost of different defect identification and modeling methods under various in-depth mapping data backgrounds.
[0010] Identify the background of the depth mapping data of the underground drainage pipeline based on the pipeline information, and match the corresponding defect identification modeling method from the resource library based on the background of the depth mapping data and the real-time requirements.
[0011] Prioritize the defect identification modeling methods to obtain a recommendation list, display the recommendation list to the platform, and have the platform and users negotiate to determine the target modeling method. Then, establish a defect identification model based on the target modeling method.
[0012] Furthermore, the establishment of the resource repository includes:
[0013] Obtain various depth mapping data backgrounds of underground drainage pipelines, perform simulation modeling analysis on each depth mapping data background, and obtain the recognition efficiency of the corresponding depth mapping data background under different defect recognition modeling methods;
[0014] The obtained in-depth mapping data background, defect identification modeling methods, and identification efficiency are integrated, and a corresponding resource library is established based on the integrated data.
[0015] Furthermore, the priority evaluation of the matching defect identification modeling methods includes:
[0016] Identify the identification efficiency of each defect identification modeling method and label it as the initial identification efficiency; estimate the modeling cost, annual maintenance cost, expected service life, and identification efficiency change curve for each defect identification modeling method. The horizontal axis of the identification efficiency change curve is time, and the vertical axis is identification efficiency. The identification efficiency at t=0 is the initial identification efficiency; fit the identification efficiency change curve to obtain the corresponding identification efficiency function, and label the identification efficiency function as SH(t), where t is time;
[0017] Determine the proportional coefficients corresponding to recognition efficiency and implementation cost based on user needs;
[0018] The priority value of the corresponding defect identification modeling method is calculated according to the priority formula, which is as follows:
[0019]
[0020] In the formula: YQ is the priority value; L is the expected service life; E0 is the modeling cost; E1 is the annual maintenance cost;
[0021] The defect identification modeling methods are sorted in descending order of priority to obtain a recommended list.
[0022] Step 2: The platform provider establishes a platform defect identification model within the platform and labels the platform defect identification model with corresponding user tags;
[0023] Furthermore, the platform's defect identification model can analyze homomorphically encrypted depth mapping data, obtain corresponding encrypted defect identification results, and send the defect identification results to the corresponding user terminal for decryption.
[0024] Step 3: Conduct real-time monitoring of underground drainage pipes to obtain corresponding depth mapping data. Based on the depth mapping data, establish a pipe model of the underground drainage pipes. Analyze the depth mapping data through a defect identification model to obtain corresponding defect identification results and confidence levels. Display the defect identification results and confidence levels in the pipe model accordingly.
[0025] Furthermore, a pipeline model of the underground drainage pipeline is established based on the depth mapping data, including:
[0026] A pipeline basic model is established based on the initially obtained depth mapping data; dynamic update data corresponding to the depth mapping data is acquired in real time, and the pipeline basic model is dynamically adjusted according to the dynamic update data, and the current pipeline basic model is marked as the pipeline model.
[0027] Step 4: Perform real-time confidence analysis based on the pipeline model. When it is determined that calibration analysis is required, send the corresponding depth mapping data to the platform for calibration analysis to obtain the corresponding calibrated defect identification results. Adjust the display of the pipeline model according to the defect identification results.
[0028] Furthermore, real-time confidence analysis is performed based on the pipeline model, including:
[0029] Establish a confidence analysis model. The expression for the confidence analysis model is as follows:
[0030]
[0031] In the formula: (s, BC) are the input data, s is the confidence level of the corresponding defect identification result; BC is the confidence standard of the corresponding defect identification result; s→BC indicates that the confidence level of the corresponding defect identification result meets the confidence standard; the output data is the confidence analysis value ZX(s, BC), and the confidence analysis value is 1 or 0.
[0032] The confidence level of each defect identification result is identified in real time based on the pipeline model. The confidence standard corresponding to the defect identification result is obtained. The confidence standard and confidence level are used as input data to be input into the confidence analysis model for analysis to obtain the confidence analysis value of the defect identification result.
[0033] When the confidence analysis value is 1, no corresponding operation is performed;
[0034] When the confidence analysis value is 0, the assessment requires calibration analysis of the defect identification results.
[0035] Furthermore, the confidence criteria are dynamically updated based on pipeline location and pipeline anomalies.
[0036] Furthermore, the depth mapping data is sent to the platform for calibration and analysis, including:
[0037] The depth mapping data is sent to the platform. The platform calls the corresponding platform defect identification model according to the user corresponding to the depth mapping data. The platform defect identification model analyzes the depth mapping data to obtain the corresponding defect identification results, and feeds back the defect identification results to the corresponding user.
[0038] Furthermore, before sending the depth mapping data to the platform, the depth mapping data is homomorphically encrypted, and the encrypted depth mapping data is then sent to the platform.
[0039] Furthermore, real-time pipeline anomaly early warning processing is performed based on the defect identification results displayed in the pipeline model.
[0040] Compared with the prior art, the beneficial effects of the present invention are:
[0041] By employing drones and mobile surveying vehicles equipped with lidar, ground-penetrating radar, and multispectral sensors, non-contact, fully automated scanning of external and shallow defects in pipelines is achieved, replacing the high-risk operation mode of manual visual inspection that requires entering the pipeline interior. A high-precision three-dimensional digital model of the pipeline is constructed through a multi-view point cloud fusion algorithm, fully presenting the pipeline's orientation, branch structure, and deformation characteristics. Real-time data from flow meters, pressure sensors, and settlement monitoring instruments are integrated. Furthermore, through cooperation between the platform provider and the user, efficient analysis is achieved at the user's location. Attached Figure Description
[0042] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0043] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation
[0044] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.
[0045] like Figure 1 As shown, a method for identifying defects in underground drainage pipes based on depth mapping data includes:
[0046] Step 1: Obtain pipeline information for the underground drainage pipelines requiring defect monitoring, such as geometric structure, material properties, operating environment, historical maintenance, monitoring methods, and other relevant information; obtain the user's real-time requirements for pipeline monitoring, as this invention allows users to have higher real-time requirements, thus enabling them to propose corresponding real-time requirements, rather than sacrificing real-time performance for the sake of accurate identification as is currently the case; the platform establishes a defect identification model for identifying pipeline defects based on the real-time requirements; the defect identification model has a confidence assessment function, meaning that the output defect identification results have a corresponding confidence level.
[0047] The platform establishes a defect identification model for identifying pipeline defects based on real-time requirements. That is, while meeting real-time requirements, the defect identification model is set up in a targeted manner, such as reducing model complexity, simplifying data preprocessing, and optimizing algorithm selection. Specifically, it is built using existing deep learning technology and has self-learning capabilities, and can be trained and optimized based on the corrected identification results. The defect identification model is deployed on the user's client.
[0048] In one embodiment, the platform establishes a defect identification model for identifying pipeline defects based on real-time requirements, including:
[0049] This involves acquiring various depth mapping data backgrounds for underground drainage pipelines, including monitoring methods that users might use to monitor these pipelines, resulting in different backgrounds for subsequent depth mapping data processing and variations in the real-time performance of the data. Simulation modeling analysis is then performed on each depth mapping data background to obtain the recognition efficiency under different defect identification modeling methods. Historical defect identification data of the underground drainage pipelines is statistically analyzed, such as calculating the recognition efficiency of various techniques used in existing defect identification models for underground drainage pipelines under the corresponding depth mapping data backgrounds. Since the platform's resource aggregation phase can be completed manually by platform staff, this process can be integrated to form a resource library.
[0050] Based on the pipeline information, the corresponding depth mapping data background is determined. Based on the depth mapping data background and real-time requirements, the corresponding defect identification modeling method is matched from the resource library. That is, the identification efficiency of the matched defect identification modeling method meets the real-time requirements and can be implemented in the context of the depth mapping data.
[0051] Prioritize the matched defect identification modeling methods to obtain a corresponding recommendation list. Display the recommendation list to the platform, and the platform and users will negotiate to determine the target modeling method. Based on the target modeling method, a defect identification model will be established.
[0052] In one embodiment, the priority evaluation of the matching defect identification modeling method can be based on existing priority evaluation methods, prioritizing based on identification efficiency and implementation cost.
[0053] In one embodiment, prioritizing the matching defect identification modeling methods includes:
[0054] Identify the recognition efficiency of each defect recognition modeling method; estimate the modeling cost, annual maintenance cost, expected service life, and recognition efficiency change curve for each defect recognition modeling method. The horizontal axis of the recognition efficiency change curve represents time, and the vertical axis represents recognition efficiency. The recognition efficiency corresponding to t=0 is the matched recognition efficiency, i.e., the initial recognition efficiency. Subsequently, the corresponding recognition efficiency changes are statistically analyzed based on the historical recognition efficiency data of the defect recognition modeling method, thereby generating the corresponding recognition efficiency change curve; fit the recognition efficiency change curve to obtain the corresponding recognition efficiency function, which is denoted as SH(t), where t is time;
[0055] Determine the proportional coefficients corresponding to recognition efficiency and implementation cost based on user needs;
[0056] The priority value of the corresponding defect identification modeling method is calculated according to the priority formula, which is as follows:
[0057]
[0058] In the formula: YQ is the priority value; L is the expected service life, i.e. the service life of the defect identification model; E0 is the modeling cost; E1 is the annual maintenance cost;
[0059] The defect identification modeling methods are sorted in descending order of priority to obtain a recommended list.
[0060] Step 2: The platform provider establishes a platform defect identification model applicable to all user pipeline information within the platform. The platform defect identification model is a defect identification model with high generalization and high accuracy. The platform defect identification model is also labeled with corresponding user tags, indicating that the platform defect identification model can analyze the in-depth mapping data of the corresponding user.
[0061] In one embodiment, the platform defect identification model can be established simultaneously by the platform when establishing the defect identification model for the corresponding user. However, it is mainly established based on standards such as high generalization and high accuracy, and is used to compensate for the defect identification model at the user's location and for subsequent optimization learning.
[0062] In one embodiment, the platform defect identification model can be established by the platform provider by summarizing the continuously accumulated training data to create a platform defect identification model that can cover a variety of user situations.
[0063] In one embodiment, the platform defect identification model can analyze homomorphically encrypted depth mapping data, obtain corresponding defect identification results, send the defect identification results to the corresponding user terminal, and decrypt the received defect identification results on the user terminal.
[0064] For example, choose a model that supports homomorphic encryption:
[0065] Linear model:
[0066] Linear regression: suitable for predicting continuous values of pipeline defects (such as defect depth).
[0067] Logistic regression: Suitable for binary classification problems (such as whether a defect exists).
[0068] Lightweight Neural Networks: Simple fully connected networks implemented using fully homomorphic encryption (such as CKKS), suitable for complex feature extraction.
[0069] Decision trees / random forests: implemented through secure comparison protocols (such as obfuscated circuits), but with high computational complexity.
[0070] Model training:
[0071] Local user training: Users train the model locally using plaintext data, encrypt the trained model parameters (weights), and then upload them to the platform. The platform directly uses the encrypted parameters to perform inference on the encrypted data.
[0072] Joint training (MPC): Multiple users collaboratively train the model, with data encrypted throughout the process, sharing only gradients or model updates (requiring homomorphic encryption or differential privacy protection).
[0073] Model optimization:
[0074] Quantization and approximate computation: Homomorphic encryption computation is relatively inefficient, but it can be accelerated by model quantization (such as converting floating-point numbers to integers) or low-precision computation.
[0075] Feature engineering: Performing dimensionality reduction or feature extraction (such as PCA) on the user side of the original data to reduce the amount of computation required for encrypted data.
[0076] Step 3: Conduct real-time monitoring of underground drainage pipes to obtain corresponding depth mapping data. Establish a pipe model of the underground drainage pipes based on the depth mapping data. Analyze the depth mapping data through a defect identification model to obtain corresponding defect identification results and confidence levels. Display the defect identification results and confidence levels in the pipe model accordingly.
[0077] In one embodiment, drones or mobile surveying vehicles equipped with lidar, ground-penetrating radar, and multispectral sensors can be used to monitor underground drainage pipes in real time.
[0078] In one embodiment, a pipeline model of the underground drainage pipeline is established based on depth mapping data, using existing modeling techniques.
[0079] In one embodiment, establishing a pipeline model for underground drainage pipes based on depth mapping data includes:
[0080] A basic pipeline model is established based on the initially obtained depth mapping data. This involves using high-precision 3D point cloud data, pipeline location data, pipeline direction, and depth data corresponding to the depth mapping data. Data from different sources (such as point clouds, GPS tracks, and sensor readings) are unified to the same coordinate system using spatiotemporal alignment algorithms (such as ICP point cloud registration). The pipeline contour is extracted using edge detection (such as the Canny algorithm) or deep learning models (such as U-Net). Invalid data is filtered by combining pipeline design specifications (such as diameter and slope). A parametric model is generated based on the pipeline centerline (fitted by B-splines or cubic splines) and cross-section (circular or elliptical). The generated model is then calibrated and adjusted to obtain the basic pipeline model.
[0081] Real-time acquisition of dynamic update data corresponding to depth mapping data generally refers to monitoring data from flow meters, pressure sensors, and inclinometers, used to dynamically update the corresponding data of the pipeline foundation model; such as water flow velocity and flow rate in the pipeline, abnormal pressure data caused by blockage or leakage, and pipeline settlement or deformation data; dynamic adjustment of the pipeline foundation model based on the dynamic update data, and marking the current pipeline foundation model as the pipeline model.
[0082] Step 4: Perform real-time confidence analysis based on the pipeline model. When it is determined that calibration analysis is required, send the corresponding depth mapping data to the platform for calibration analysis to obtain the corresponding calibrated defect identification results. Adjust the display of the pipeline model according to the defect identification results.
[0083] In one embodiment, real-time confidence analysis is performed based on the pipeline model to determine whether the corresponding confidence level meets the confidence criteria. If the confidence criteria are not met, a calibration analysis is required. Specifically, this is done using existing evaluation methods.
[0084] In one embodiment, real-time confidence analysis based on a pipeline model includes:
[0085] Establish a confidence analysis model. The expression for the confidence analysis model is as follows:
[0086]
[0087] In the formula: (s, BC) represents the input data, s represents the confidence level of the corresponding defect identification result, BC represents the confidence standard of the corresponding defect identification result, s→BC indicates that the confidence level of the corresponding defect identification result meets the confidence standard, the output data is the confidence analysis value ZX(s, BC), the confidence analysis value is 1 or 0; training is performed using the corresponding training set of historical data, the confidence standard is the corresponding preset confidence level, not lower than the confidence standard is considered to meet the confidence standard, and then the training set is set for training;
[0088] The confidence level of each defect identification result is identified in real time based on the pipeline model. The confidence standard corresponding to the defect identification result is obtained. The confidence standard and confidence level are used as input data to be input into the confidence analysis model for analysis to obtain the corresponding confidence analysis value.
[0089] When the confidence analysis value is 1, no corresponding operation is performed;
[0090] When the confidence analysis value is 0, the assessment requires calibration analysis of the corresponding defect identification results.
[0091] In one embodiment, the confidence standard is set based on the pipeline location and anomaly corresponding to the defect identification result. Generally, it is initially set by the platform based on historical data of similar pipelines, and subsequently adjusted based on the historical accuracy of defect identification of the underground drainage pipeline. For example, a confidence analysis model can be established based on machine learning, deep learning algorithms, etc., and the platform sets a corresponding training set for training. The training set includes input data and output data. The input data is the historical defect identification result data of the relevant pipeline location at the corresponding pipeline anomaly, including the accuracy corresponding to the corresponding confidence level. The output data is the corresponding confidence standard. Alternatively, the confidence standard can also be set or adjusted directly by the user as needed.
[0092] The confidence standard is dynamically updated based on the pipeline location and pipeline anomalies.
[0093] In one embodiment, a simple way to set the confidence standard is to set a uniform confidence level for various pipeline anomalies, and then the user can dynamically adjust the confidence standard for the corresponding pipeline location and pipeline anomaly based on the user experience.
[0094] In one embodiment, sending depth mapping data to the platform for calibration analysis includes:
[0095] The depth mapping data is sent to the platform. The platform calls the corresponding platform defect identification model based on the user corresponding to the depth mapping data. The platform defect identification model analyzes the depth mapping data to obtain the corresponding defect identification results, and then feeds the defect identification results back to the corresponding user.
[0096] In one embodiment, before sending the depth mapping data to the platform, the depth mapping data is homomorphically encrypted, and the encrypted depth mapping data is then sent to the platform.
[0097] In one embodiment, real-time pipeline anomaly early warning processing is performed based on the defect identification results displayed in the pipeline model. That is, when the defect identification result indicates a pipeline anomaly, an early warning is issued to the user, and the user can handle the situation according to preset emergency handling measures.
[0098] The above formulas are all numerical calculations after removing dimensions. The formulas are obtained by software simulation based on a large amount of data and are closest to the real situation. The preset parameters and preset thresholds in the formulas are set by those skilled in the art according to the actual situation or obtained by simulation based on a large amount of data.
[0099] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.
Claims
1. A method for identifying defects in underground drainage pipes based on depth mapping data, characterized in that, include: Obtain pipeline information for underground drainage pipelines that require defect monitoring, and obtain user requirements for the real-time performance of pipeline monitoring; The platform provider establishes a defect identification model for users to identify pipeline defects based on real-time requirements and pipeline information; The platform provider establishes a platform defect identification model within the platform and labels the platform defect identification model with corresponding user tags; Real-time monitoring of underground drainage pipes is conducted to obtain corresponding depth mapping data. Based on the depth mapping data, a pipe model of the underground drainage pipe is established. The depth mapping data is analyzed through a defect identification model to obtain corresponding defect identification results and confidence levels. Based on the defect identification results and confidence levels, the pipe model is displayed accordingly. Real-time confidence analysis is performed based on the pipeline model. When it is determined that calibration analysis is needed, the corresponding depth mapping data is sent to the platform for calibration analysis to obtain the corresponding calibrated defect identification results. The pipeline model is then adjusted for display based on the defect identification results.
2. The method for identifying defects in underground drainage pipelines based on depth mapping data according to claim 1, characterized in that, The platform provider establishes a defect identification model for identifying pipeline defects based on real-time requirements and pipeline information, including: The platform provider establishes a resource library, which stores the identification efficiency and implementation cost of different defect identification and modeling methods under various in-depth mapping data backgrounds. Identify the background of the depth mapping data of the underground drainage pipeline based on the pipeline information, and match the corresponding defect identification modeling method from the resource library based on the background of the depth mapping data and the real-time requirements. Prioritize the defect identification modeling methods to obtain a recommendation list, display the recommendation list to the platform, and have the platform and users negotiate to determine the target modeling method. Then, establish a defect identification model based on the target modeling method.
3. The method for identifying defects in underground drainage pipelines based on depth mapping data according to claim 2, characterized in that, The establishment of the resource repository includes: Obtain various depth mapping data backgrounds of underground drainage pipelines, perform simulation modeling analysis on each depth mapping data background, and obtain the recognition efficiency of the corresponding depth mapping data background under different defect recognition modeling methods; The obtained in-depth mapping data background, defect identification modeling methods, and identification efficiency are integrated, and a corresponding resource library is established based on the integrated data.
4. The method for identifying defects in underground drainage pipelines based on depth mapping data according to claim 2, characterized in that, Prioritize the matching defect identification modeling methods, including: Identify the identification efficiency of each defect identification modeling method and label it as the initial identification efficiency; estimate the modeling cost, annual maintenance cost, expected service life, and identification efficiency change curve for each defect identification modeling method. The horizontal axis of the identification efficiency change curve is time, and the vertical axis is identification efficiency. The identification efficiency at t=0 is the initial identification efficiency; fit the identification efficiency change curve to obtain the corresponding identification efficiency function, and label the identification efficiency function as SH(t), where t is time; Determine the proportional coefficients corresponding to recognition efficiency and implementation cost based on user needs; The priority value of the corresponding defect identification modeling method is calculated according to the priority formula, which is as follows: In the formula: YQ is the priority value; L is the expected service life; E0 is the modeling cost; E1 is the annual maintenance cost; The defect identification modeling methods are sorted in descending order of priority to obtain a recommended list.
5. The method for identifying defects in underground drainage pipelines based on depth mapping data according to claim 1, characterized in that, The platform's defect identification model can analyze homomorphically encrypted depth mapping data, obtain corresponding encrypted defect identification results, and send the defect identification results to the corresponding user terminal for decryption.
6. The method for identifying defects in underground drainage pipelines based on depth mapping data according to claim 1, characterized in that, Based on depth mapping data, a pipeline model of the underground drainage system is established, including: A pipeline basic model is established based on the initially obtained depth mapping data; dynamic update data corresponding to the depth mapping data is acquired in real time, and the pipeline basic model is dynamically adjusted according to the dynamic update data, and the current pipeline basic model is marked as the pipeline model.
7. The method for identifying defects in underground drainage pipelines based on depth mapping data according to claim 1, characterized in that, Real-time confidence analysis based on pipeline model includes: Establish a confidence analysis model. The expression for the confidence analysis model is as follows: In the formula: (s, BC) are the input data, s is the confidence level of the corresponding defect identification result; BC is the confidence standard of the corresponding defect identification result; s→BC indicates that the confidence level of the corresponding defect identification result meets the confidence standard; the output data is the confidence analysis value ZX(s, BC), and the confidence analysis value is 1 or 0. The confidence level of each defect identification result is identified in real time based on the pipeline model. The confidence standard corresponding to the defect identification result is obtained. The confidence standard and confidence level are used as input data to be input into the confidence analysis model for analysis to obtain the confidence analysis value of the defect identification result. When the confidence analysis value is 1, no corresponding operation is performed; When the confidence analysis value is 0, the assessment requires calibration analysis of the defect identification results.
8. A method for identifying defects in underground drainage pipelines based on depth mapping data according to claim 7, characterized in that, The confidence level is dynamically updated based on the pipeline location and pipeline anomalies.
9. A method for identifying defects in underground drainage pipelines based on depth mapping data according to claim 5, characterized in that, Before sending the depth mapping data to the platform, the depth mapping data is homomorphically encrypted, and the encrypted depth mapping data is then sent to the platform.
10. A method for identifying defects in underground drainage pipelines based on depth mapping data according to claim 1, characterized in that, The depth mapping data is sent to the platform for calibration and analysis, including: The depth mapping data is sent to the platform. The platform calls the corresponding platform defect identification model according to the user corresponding to the depth mapping data. The platform defect identification model analyzes the depth mapping data to obtain the corresponding defect identification results, and feeds back the defect identification results to the corresponding user.