Channel revetment safety early warning method and system

Abstract

The invention discloses a channel revetment safety early warning method and system, and relates to the technical field of channel infrastructure operation monitoring. The method comprises the following steps: acquiring multi-source monitoring data of a channel revetment, and preprocessing the multi-source monitoring data; performing damage identification and classification on the preprocessed multi-source monitoring data, extracting damage features through a damage identification algorithm, generating a multi-source identification result, and performing fusion comparison on the multi-source identification result to generate a fusion damage result; based on the fusion damage result, performing grade evaluation on the safety state of the channel revetment to obtain a quantitative risk grade; and when the bank protection safety state reaches a preset risk level, a safety alarm mechanism is triggered, and an alarm prompt and a disposal suggestion are pushed to the management terminal. According to the method, an automatic process of sensing, diagnosis, early warning and disposal is constructed, the accuracy and reliability of risk identification are improved, and a complete technical solution is provided for operation and maintenance and precise management and maintenance of the channel infrastructure.

Description

A method and system for early warning of safety on waterway bank protection Technical Field

[0001] This invention relates to a method and system for early warning of waterway bank protection safety, belonging to the field of waterway infrastructure operation monitoring technology. Background Technology

[0002] The complex operating environment of waterways and the high technical difficulty of monitoring and sensing technologies result in insufficient monitoring capabilities for waterway infrastructure. Due to the lack of effective monitoring of waterway infrastructure operations, problems such as revetment erosion and collapse frequently occur, along with untimely dredging leading to low navigational clearance rates. In severe cases, this can cause significant impacts such as waterway blockage and temporary navigation closures. The operating environment of waterway infrastructure inherently presents technical challenges for monitoring. Revetments typically stretch for tens of kilometers or even longer, making static data measurements insufficient to reflect true operating conditions. Existing revetment early warning technologies often rely on single, fixed-point monitoring models. While these methods improve monitoring efficiency and data automation to some extent, they still have many problems. For example, abnormal data from a single measuring point may be caused by instrument malfunction or other factors, making identification difficult. Traditional monitoring instruments often require fixed installation, are complex to deploy, and are costly. Furthermore, different monitoring instruments generate independent data streams, lacking an effective collaborative fusion mechanism. Summary of the Invention

[0003] The purpose of this invention is to provide a method and system for early warning of safety of waterway revetments. By using a damage identification algorithm to independently diagnose the collected revetment monitoring data, and then verifying the correlation of different identification results, damage to the revetment structure can be detected in a timely manner, and its safety can be assessed. This allows for targeted early warning and maintenance decisions, achieving accurate perception and early warning of the revetment status and providing a new solution for waterway revetment monitoring.

[0004] To achieve the above objectives, the present invention is implemented using the following technical solution.

[0005] On the one hand, the present invention provides a method for early warning of safety on waterway bank protection, comprising:

[0006] Acquire multi-source monitoring data of the waterway revetment and preprocess the multi-source monitoring data;

[0007] Damage identification and classification are performed on the preprocessed multi-source monitoring data. Damage features are extracted using a damage identification algorithm to generate multi-source identification results. The multi-source identification results are then fused and compared to generate fused damage results.

[0008] Based on the fusion damage results, the safety status of the channel bank protection is assessed to obtain a quantitative risk level.

[0009] When the safety status of the riverbank reaches the preset risk level, a safety alarm mechanism is triggered, and an alarm prompt and handling suggestions are pushed to the management terminal.

[0010] Optionally, preprocessing the multi-source monitoring data includes at least:

[0011] Image enhancement and image denoising processing are performed on the monitoring data of waterway revetment damage. The image enhancement methods include histogram equalization, adaptive brightness enhancement or Retinex processing. The image denoising is processed by filters or denoising algorithms.

[0012] Data cleaning and processing were performed on the monitoring data of the channel revetment structure performance, including removing noise components, eliminating measurement outliers, separating long-term trend terms, and interpolating missing sampling points.

[0013] Optionally, after preprocessing the multi-source monitoring data, the method further includes storing the preprocessed multi-source monitoring data in a pre-built hierarchical database. The database construction includes:

[0014] Define the data structure schema, create the basic data tables, configure index fields, set data timestamps and associated revetment structure codes, and establish the data call interface;

[0015] The database includes at least a basic database for waterway revetments, a health monitoring database, a revetment maintenance and management database, and an information service system management database. Each database is linked and fused through revetment structure codes and data timestamps.

[0016] Optionally, the step of performing damage identification and classification on the preprocessed multi-source monitoring data, and extracting damage features through a damage identification algorithm, includes:

[0017] Based on the image recognition algorithm for damaged waterway revetments, key parameters are extracted from the monitoring data of damaged waterway revetments. Feature recognition is performed by combining a deep convolutional neural network model, and the damaged waterway revetments are classified by calculating the confidence level.

[0018] Optionally, the damage identification and classification further includes:

[0019] Based on the channel revetment structure damage detection algorithm, the dynamic fingerprint features of the structure represented by the frequency offset rate are extracted, and the degree of structural change is calculated according to the frequency offset rate.

[0020] Structural damage is classified based on the range of frequency offset.

[0021] Optionally, in step S2, generating multi-source identification results and fusing and comparing the multi-source identification results to generate fusion impairment results includes:

[0022] Based on the preset image recognition confidence threshold and structural frequency offset rate threshold, independent damage judgment is performed on the damage monitoring data and structural performance monitoring data, and single-source damage labels are generated respectively. The labeling results are written into the database.

[0023] The image recognition results are fused and compared with the structural response changes characterized by the structural frequency shift rate, including:

[0024] When the confidence level of image recognition in the same area Exceeding its set threshold, and the corresponding structural frequency offset rate When the threshold is exceeded, a comprehensive damage marker is generated and the marker result is written to the database;

[0025] The fusion comparison methods include rule-based threshold judgment and weighted fusion.

[0026] Optionally, the assessment of the safety status of the channel revetment includes:

[0027] A comprehensive analysis of the appearance damage and structural response indicators of the revetment was conducted to identify potential damage areas and abnormal trends.

[0028] A weighted scoring model is used to calculate risks, and combined with historical records, structural parameters and maintenance data from the database, the risk level of the revetment status is classified and a diagnostic report is automatically generated.

[0029] Optionally, the risk levels include safe, alert, and dangerous;

[0030] When the risk level reaches the warning or danger level, the alarm mechanism will be automatically activated.

[0031] In a second aspect, the present invention provides a waterway bank protection safety early warning system for implementing the waterway bank protection safety early warning method described in any one of the first aspects, comprising:

[0032] The data acquisition and preprocessing module is used to acquire multi-source monitoring data of the waterway bank protection and to preprocess the multi-source monitoring data.

[0033] The damage identification and diagnosis module is used to identify and classify damage from preprocessed multi-source monitoring data. It extracts damage features through damage identification algorithms, generates multi-source identification results, and fuses and compares the multi-source identification results to generate fused damage results.

[0034] The safety assessment module is used to assess the safety status of the waterway revetment based on the fused damage results, and obtain a quantitative risk level.

[0035] The early warning and maintenance decision-making module is used to trigger a safety alarm mechanism when the safety status of the revetment reaches a preset risk level, and push alarm prompts and handling suggestions to the management terminal.

[0036] Optionally, the system also includes a database module.

[0037] The database module is used to store basic information of the revetment structure, inspection data, maintenance records and model parameters, realize unified management of multi-source monitoring data, and support data fusion and related queries.

[0038] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0039] This invention constructs an automated "perception-diagnosis-early warning-response" process, forming an efficient and closed-loop operation and maintenance management model from anomaly detection to response suggestion delivery. Employing a multi-source information fusion mechanism overcomes potential errors in single-sensor data. Through independent diagnosis and fusion comparison of image and structural data, it performs safety status level assessment and issues early warnings, improving the accuracy and reliability of risk identification. This provides a complete technical solution for the operation, maintenance, and precise management of waterway infrastructure. Attached Figure Description

[0040] Figure 1 is a schematic flowchart of the waterway bank protection safety early warning method provided in an embodiment of the present invention;

[0041] Figure 2 is a schematic diagram of the crack surface provided in an embodiment of the present invention; Detailed Implementation

[0042] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0043] It should be noted that the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. Additionally, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0044] Example 1

[0045] This embodiment uses a section of an inland waterway, numbered GZ-1-K010, as the monitoring target. It details the specific process of the method of this invention. This embodiment introduces a waterway revetment safety early warning method, as shown in Figure 1, which specifically includes the following steps:

[0046] Step S1: Obtain monitoring data on damage to the waterway revetment and monitoring data on the structural performance of the waterway revetment, and perform data preprocessing:

[0047] High-definition video surveillance (or drone patrol video or high-definition remote sensing images) deployed on this section of the revetment automatically captured a transverse crack at the bottom of the revetment wall. At the same time, fiber optic strain sensors deployed in the same area transmitted back monitoring data reflecting the stress and vibration state of the revetment, and the data showed abnormal fluctuations.

[0048] S11, Image Enhancement and Noise Reduction:

[0049] First, the color image is converted to grayscale for processing. The Retinex algorithm is then used to enhance the image and improve the contrast. Subsequently, Gaussian filtering is used to reduce noise in the image, such as blemishes and dots, making the cracks appear clearer.

[0050] S12, Data Cleaning:

[0051] The monitoring data transmitted from the fiber Bragg grating strain sensor is cleaned. Data noise is processed using methods such as Kalman filtering and wavelet denoising. Problems such as outliers, trend terms, and desensitization are identified using supervised machine learning algorithms. Missing values ​​are imputed using methods such as mean interpolation and Lagrange interpolation.

[0052] The preprocessed crack image keyframes and the cleaned monitoring data are automatically stored in a pre-built hierarchical database. The original images are stored in the health monitoring database, and the extracted crack feature parameters and the processed strain data are associated with the revetment code GZ-1-K010 and timestamp and stored in the information service system management database.

[0053] Specifically, the database is constructed as follows:

[0054] By defining a data structure schema, establishing basic data tables, configuring index fields, setting data timestamps, associating with revetment structure codes, and establishing data access interfaces, this embodiment includes four hierarchical databases:

[0055] Basic database for waterway revetment: Stores design documents, as-built drawings, acceptance data, structural displacements or deformations and related information during the construction management process for waterway revetment.

[0056] Health monitoring database: Stores online monitoring data of revetment structures, structural damage identification information, safety status assessment information, safety diagnosis reports, early warning information, etc.

[0057] Bank protection maintenance and management database: Stores maintenance work plans, inspection records, testing and evaluation reports, maintenance engineering technical data, facility and equipment management ledgers and other relevant materials for waterway bank protection.

[0058] The information service system manages the database, which stores data on revetment health monitoring and statistical analysis, real-time monitoring and early warning data, and 3D model data of the revetment structure.

[0059] Step S2, Damage Identification and Classification:

[0060] S21. Image damage recognition:

[0061] The system invokes a deep convolutional neural network model to identify the enhanced crack image. The model extracts key parameters such as crack edge, width, length, and distribution density for feature recognition, as shown below:

[0062] ;

[0063] in:

[0064] This represents the convolution result (output feature value) of the k-th output channel at position (i,j) on the output feature map.

[0065] This represents the pixel value of the m-th channel of the input feature map at position (i+p, j+q);

[0066] This represents the weight coefficients of the convolution kernel at relative positions (p, q) that connect the input channel m to the output channel k.

[0067] Indices p and q represent the displacement range of the convolution kernel in the spatial dimension, and index m traverses the input channels.

[0068] Next, the damage is classified, and the confidence score of the classification output is calculated using Softmax, as shown below:

[0069] ;

[0070] in, This represents the confidence (probability) that a sample is classified as category c. This represents the unnormalized output of the classifier for class c; summation. Performed on all candidate category i.

[0071] The model identifies the crack as a "structural transverse crack" and outputs its width and identification confidence level. According to the preset rules: when the crack width is greater than 15mm and less than 30mm or the extension length exceeds 5% of the length of the revetment unit, the system determines it as a minor damage; when the crack width is greater than 30mm or the crack penetration rate exceeds 10%, it is determined as a severe damage.

[0072] As shown in Figure 2, in this example, the current crack width is calculated to be 23mm, with a confidence level of 80%, and is judged to be minor damage.

[0073] S22. Structural performance identification:

[0074] Vibration signals and stress change data acquired using distributed optical fibers are used to extract structural dynamic fingerprint features, including dominant frequency, energy distribution, and damping ratio, through an improved Kalman filter and wavelet packet energy analysis method. Using the frequency fingerprint in the database under non-destructive conditions as a reference (compared to historical health data), the frequency offset rate is calculated to determine structural changes. The calculation formula is as follows:

[0075] ;

[0076] in:

[0077] This represents the frequency offset rate, which indicates the relative change of the monitored frequency relative to the reference frequency.

[0078] Indicates the current monitoring frequency;

[0079] The reference frequency is the frequency determined by the health status records of the monitored object or by historical statistics.

[0080] Structural damage is classified according to the frequency offset range: if Δf > 3%, the system identifies it as potential structural damage; if Δf > 5% and the duration exceeds 1 hour, it is determined to be significant structural damage.

[0081] In this example, the current frequency offset rate was calculated to be 8% and the duration was 70 minutes, which was determined to be significant structural damage.

[0082] S23. Fusion comparison and correlation verification:

[0083] The data fusion engine correlates and verifies the two recognition results, with preset thresholds of 80% confidence in image recognition and 5% frequency offset.

[0084] The current image recognition confidence level is 80%, the frequency offset rate is 8%, and the spatiotemporal labels of the two are consistent. Therefore, a comprehensive damage label is generated and written into the database.

[0085] Step S3: Based on the fusion damage results, the safety status of the waterway revetment is assessed to obtain a quantitative risk level. This is a process of using specific information to analyze the reliability of the existing revetment and make corresponding engineering decisions to maintain a certain level of reliability for the revetment.

[0086] After receiving the "comprehensive damage marker," the assessment model, combining a vast database of historical cases and standards, uses a multi-source data fusion model to comprehensively analyze revetment damage (such as cracks, collapses, and scour) and structural response indicators (such as stress, vibration frequency, and displacement changes) to identify potential damage areas and abnormal trends. A quantitative assessment of the revetment's safety status is then performed using a weighted scoring model for risk calculation, with the following formula:

[0087] ;

[0088] in, This represents the calculated risk value (scalar), which is used for subsequent risk level classification. , The weight coefficients represent the importance of the image recognition confidence term and the structural response term, respectively. They can be determined by expert experience or data fitting, and satisfy W1≥0, W2≥0. The image recognition confidence score represented by Softmax ranges from [0,1]; Δf represents the frequency offset rate, taken as the absolute value. To reflect the magnitude of the offset;

[0089] In this example, W1=0.4 and W2=0.8, emphasizing the weight of the structure's intrinsic performance. Substituting the data: .

[0090] The risk levels are divided into three levels:

[0091] Level I (Safe): R < 0.2, structural condition is stable, and monitoring indicators are within the normal fluctuation range;

[0092] Level II (Alert): 0.2≤R<0.6, indicating a slight anomaly in monitoring data or damage identification results, requiring enhanced patrols and focused monitoring;

[0093] Level III (Hazard): R ≥ 0.6, the structure shows significant abnormalities or signs of continuous deterioration, and there is a risk of instability or damage. Emergency response should be initiated immediately.

[0094] The system determined that the area was a potential core area of ​​damage, and the damage trend was rapidly developing. Through the above quantitative assessment, the risk level was finally classified as "alert".

[0095] Based on the assessment results, and combined with the database's historical records, structural parameters, and maintenance data, the risk level of the revetment is classified, and a diagnostic report is automatically generated. The report includes the risk level, abnormal characteristics, influencing factors, and development trends, providing a basis for decision-making in maintenance, repair, and emergency management.

[0096] Step S4: When the safety status of the revetment reaches the preset risk level, trigger the safety alarm mechanism and push an alarm prompt and disposal suggestions to the management terminal.

[0097] Based on the safety assessment of the channel revetment, comprehensively judge the anomalies in the structural performance of the revetment (such as exceeding the limit of indicators, sudden change of response or expansion of damage), and trigger the linkage of hierarchical early warning and disposal. When the system detects that the risk level reaches the warning (Level II) or danger (Level III), automatically start the alarm mechanism, and push warning prompts and disposal suggestions to the management department through the information platform, including measures such as strengthening inspections, implementing reinforcement, temporarily closing the navigation or emergency risk elimination, etc., to achieve closed-loop management from anomaly identification to decision execution, and improve the operation safety and maintenance controllability of the channel revetment.

[0098] Since the risk level in this embodiment is "warning", on the large-screen monitoring system of the channel management center, this section of the revetment becomes a yellow flashing state on the map and emits a warning sound.

[0099] The system automatically pushes a structured alarm message to the terminals (APP, computer workstation) of the person in charge of the maintenance department and technical personnel, as follows:

[0100] [Channel Revetment Safety Warning]

[0101] Location: Revetment Code GZ-1-K010;

[0102] Time: 2025 - 12 - 1 11:30:05;

[0103] Level: Warning

[0104] Details: Through the fusion of images and structural data, it is determined that there are structural horizontal cracks (width 23mm) in this section of the revetment and a significant decrease in local stiffness (frequency deviation 8%) is accompanied, and there is a risk of further development.

[0105] At the same time, based on the knowledge base and preset rules, the system automatically generates and pushes preliminary suggestions:

[0106] [Suggested Disposal Measures]

[0107] Act immediately: Set up physical warning lines in this area to restrict the entry of personnel and heavy equipment;

[0108] Strengthen monitoring: Increase the video monitoring and sensor data collection frequency to once every 5 minutes;

[0109] Solution: According to the maintenance regulations, carry out local reinforcement of the crack area.

[0110] Embodiment 2

[0111] Based on the same inventive concept as Embodiment 1, this embodiment introduces a waterway revetment safety early warning system for implementing the waterway revetment safety early warning method described in any one of Embodiment 1, comprising:

[0112] The data acquisition and preprocessing module is used to acquire multi-source monitoring data of the waterway bank protection and to preprocess the multi-source monitoring data.

[0113] The damage identification and diagnosis module is used to identify and classify damage from preprocessed multi-source monitoring data. It extracts damage features through damage identification algorithms, generates multi-source identification results, and fuses and compares the multi-source identification results to generate fused damage results.

[0114] The safety assessment module is used to assess the safety status of the waterway revetment based on the fused damage results, and obtain a quantitative risk level.

[0115] The early warning and maintenance decision-making module is used to trigger a safety alarm mechanism when the safety status of the revetment reaches a preset risk level, and automatically push alarm prompts and handling suggestions to the management terminal.

[0116] The specific functional implementation of each of the above modules is described in the relevant content of the method in Embodiment 1, and will not be repeated here. It should be noted that:

[0117] The system also includes a database module for storing basic information about the revetment structure, inspection data, maintenance records, and model parameters, enabling unified management of multi-source monitoring data and supporting data fusion and related queries.

[0118] In summary, this invention effectively fuses, compares, and deeply analyzes isolated data sources (images, sensors), realizing an integrated data "collection-diagnosis-assessment-decision" process. When risks first emerge, the system can issue timely and accurate alarms and provide guiding disposal suggestions, greatly improving the intelligent level of waterway bank protection safety management and emergency response efficiency, and providing a complete technical solution for the operation, maintenance, and precise management of waterway infrastructure.

[0119] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0120] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in one or more blocks of the flowchart illustrations and / or one or more blocks of the block diagrams.

[0121] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement the functions specified in one or more flowcharts and / or one or more block diagrams.

[0122] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, such that the instructions, which execute on the computer or other programmable apparatus, provide steps for implementing the functions specified in one or more flowcharts and / or one or more block diagrams.

[0123] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A method for early warning of safety on waterway bank protection, characterized in that, include: Acquire multi-source monitoring data of the waterway revetment and preprocess the multi-source monitoring data; Damage identification and classification are performed on the preprocessed multi-source monitoring data. Damage features are extracted using a damage identification algorithm to generate multi-source identification results. The multi-source identification results are then fused and compared to generate fused damage results. Based on the fusion damage results, the safety status of the channel bank protection is assessed to obtain a quantitative risk level. When the safety status of the riverbank reaches the preset risk level, a safety alarm mechanism is triggered, and an alarm prompt and handling suggestions are pushed to the management terminal.

2. The method for early warning of safety for waterway revetments according to claim 1, characterized in that, The preprocessing of the multi-source monitoring data includes at least the following: image enhancement and image denoising of the channel revetment damage monitoring data, wherein the image enhancement methods include histogram equalization, adaptive brightness enhancement or Retinex processing, and the image denoising is performed by a filter or denoising algorithm; and data cleaning of the channel revetment structure performance monitoring data, including removing noise components, eliminating measurement outliers and separating long-term trend terms and interpolating missing sampling points.

3. The method for early warning of safety for waterway revetments according to claim 2, characterized in that, After preprocessing the multi-source monitoring data, the process also includes storing the preprocessed multi-source monitoring data in a pre-constructed hierarchical database. The database construction includes: defining a data structure schema, establishing basic data tables, configuring index fields, setting data timestamps and associated revetment structure codes, and establishing a data call interface. The database includes at least a basic revetment database, a health monitoring database, a revetment maintenance and management database, and an information service system management database. Each database is associated and fused through revetment structure codes and data timestamps.

4. The waterway bank protection safety early warning method according to claim 3, characterized in that, The process involves damage identification and classification of the preprocessed multi-source monitoring data, and the extraction of damage features through damage identification algorithms. This includes: extracting key parameters from the monitoring data of waterway revetment damage based on the waterway revetment damage image recognition algorithm, performing feature recognition by combining a deep convolutional neural network model, and classifying the waterway revetment damage by calculating confidence levels.

5. The method for early warning of safety for waterway revetments according to claim 4, characterized in that, The damage identification and classification also includes: extracting dynamic fingerprint features of the structure represented by the frequency offset rate based on the channel revetment structure damage detection algorithm, and calculating the degree of structural change based on the frequency offset rate; and classifying the structural damage according to the range of change of the frequency offset rate.

6. The method for early warning of safety for waterway revetments according to claim 5, characterized in that, In step S2, generating multi-source recognition results and fusing and comparing them to generate fused damage results includes: based on preset image recognition confidence thresholds and structural frequency offset rate thresholds, performing independent damage judgments on damage monitoring data and structural performance monitoring data, generating single-source damage markers respectively, and writing the marker results into the database; fusing and comparing the image recognition results with the structural response changes represented by the structural frequency offset rate, including: when the image recognition confidence level of the same area is high... Exceeding its set threshold, and the corresponding structural frequency offset rate When the threshold is exceeded, a comprehensive damage label is generated and the labeling result is written into the database; the fusion comparison method includes rule threshold judgment and weighted fusion.

7. The method for early warning of safety for waterway revetments according to claim 6, characterized in that, The assessment of the safety status of the waterway revetment includes: a comprehensive analysis of the revetment's appearance damage and structural response indicators to identify potential damage areas and abnormal trends; risk calculation using a weighted scoring model; and, in conjunction with historical records, structural parameters, and maintenance data from the database, classifying the revetment's status into risk levels and automatically generating a diagnostic report.

8. The method for early warning of safety for waterway revetments according to claim 7, characterized in that, The risk levels include safe, alert, and dangerous; when the risk level reaches alert or dangerous, an alarm mechanism is automatically activated.

9. A waterway revetment safety early warning system, used to implement the waterway revetment safety early warning method according to any one of claims 1-8, characterized in that, include: The data acquisition and preprocessing module is used to acquire multi-source monitoring data of the waterway bank protection and to preprocess the multi-source monitoring data. The damage identification and diagnosis module is used to identify and classify damage from preprocessed multi-source monitoring data. It extracts damage features through damage identification algorithms, generates multi-source identification results, and fuses and compares the multi-source identification results to generate fused damage results. The safety assessment module is used to assess the safety status of the waterway revetment based on the fused damage results, and obtain a quantitative risk level. The early warning and maintenance decision-making module is used to trigger a safety alarm mechanism when the safety status of the revetment reaches a preset risk level, and push alarm prompts and handling suggestions to the management terminal.

10. The waterway bank protection safety early warning system according to claim 9, characterized in that, The system also includes a database module, which stores basic information of the revetment structure, detection data, maintenance records and model parameters, realizes unified management of multi-source monitoring data, and supports data fusion and correlation query.

Images