Cable tunnel intelligent inspection abnormity identification method, system and device and medium

Abstract

The invention discloses a cable tunnel intelligent inspection abnormity identification method, system and device and a medium, and belongs to the technical field of data processing, and the method comprises the steps: synchronously collecting the operation data of a cable tunnel through a heterogeneous sensor, obtaining multi-mode operation data including space structure information, gas component information, sound signals, electromagnetic signals and environment state information; the method comprises the following steps: respectively carrying out feature extraction and anomaly modeling on modal data, generating binary anomaly flag data, carrying out feature level fusion through a secondary algorithm, judging whether equipment is abnormal or not based on fused features, and completing equipment anomaly detection. According to the method, the problem of high misjudgment rate of a single sensor index is solved, unified multi-modal fusion feature parameters are generated, the accuracy and reliability of equipment anomaly identification in a complex environment are improved, and the limitations of isolated analysis and lack of intelligent fusion of a traditional monitoring method are overcome.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a method, system, equipment, and medium for intelligent inspection and anomaly identification of cable tunnels. Background Technology

[0002] Cable tunnels are typically located underground in complex and variable environments, including confined spaces, limited lighting, and high temperatures and humidity. Harmful gases such as methane and hydrogen sulfide may be present within the tunnels, posing a serious safety threat to inspection personnel. The tunnels also contain numerous facilities, including cables, supports, lighting, and ventilation systems, making comprehensive coverage difficult with traditional manual inspections. Defects in cables and tunnel facilities are often well-hidden, such as internal cracks or insulation peeling, and are difficult to detect using traditional inspection methods.

[0003] Traditional monitoring technologies typically can only detect single types of anomalies, making it difficult to comprehensively reflect the equipment status. They also lack intelligent analysis capabilities and struggle to integrate and process complex data. Therefore, determining whether an equipment is malfunctioning requires a comprehensive assessment of multiple data points. Traditional methods, which rely on single data indicators, often result in high false positive rates and inaccurate assessments. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by this invention is that traditional monitoring technologies can usually only detect a single type of anomaly, making it difficult to comprehensively reflect the equipment status. Traditional monitoring technologies lack intelligent analysis capabilities and struggle to integrate and process complex data. Furthermore, traditional monitoring technologies rely on single data indicators for judgment, often resulting in high false alarm rates and inaccurate assessments.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for intelligent inspection and anomaly identification of cable tunnels, comprising, By synchronously collecting cable tunnel operation data through heterogeneous sensors, multimodal operation data including spatial structure information, gas composition information, sound signals, electromagnetic signals and environmental status information can be obtained. Based on the collected multimodal operational data, feature extraction and anomaly modeling are performed respectively to generate binary anomaly flag data; The binary anomaly flag data is fused at the feature level using a two-level algorithm, and the device is judged to have anomalies based on the fused features, thus completing the device anomaly detection.

[0007] As a preferred embodiment of the intelligent inspection and anomaly identification method for cable tunnels described in this invention, the step of synchronously collecting operational data of the cable tunnel through heterogeneous sensors to obtain multimodal operational data including spatial structure information, gas composition information, sound signals, electromagnetic signals, and environmental state information includes... Space structure information is collected using lidar, gas state information is collected using gas sensors, and sound information generated during operation is collected using acoustic sensors. Electromagnetic radiation information is collected by an electromagnetic radiation sensor, and environmental status information is collected by a temperature and humidity sensor. Multimodal operational data is generated based on the collected spatial structure information, gas composition information, sound signals, electromagnetic signals, and environmental state information.

[0008] As a preferred embodiment of the intelligent inspection and anomaly identification method for cable tunnels described in this invention, the step of performing feature extraction and anomaly modeling based on collected multimodal operational data to generate binary anomaly flag data includes: Based on the collected multimodal operation data, the electric cloud registration feature parameters are calculated through a first-level algorithm, gas features are extracted, and gas concentration feature parameters are calculated. Extract acoustic features, electromagnetic radiation features, and environmental state features, and calculate state-environment feature parameters; Based on the electric cloud registration feature parameters, gas concentration feature parameters, acoustic features, electromagnetic radiation features, environmental state features, and state-environment feature parameters, preliminary anomaly detection is performed to generate binary anomaly marker data.

[0009] As a preferred embodiment of the intelligent inspection and anomaly identification method for cable tunnels described in this invention, the step of fusing binary anomaly flag data at the feature level using a two-level algorithm, and determining whether the equipment has an anomaly based on the fused features, thereby completing the equipment anomaly detection, includes: Based on binary anomaly marker data, feature-level fusion is performed through a two-level algorithm to output multimodal data fusion feature parameters; Based on the feature parameters fused from multimodal data, the device status is determined, and abnormal alarm information is output according to the device status. Based on the abnormal alarm information, output suggested measures, and conduct anomaly investigation based on the suggested measures.

[0010] As a preferred embodiment of the intelligent inspection and anomaly identification method for cable tunnels described in this invention, the steps include: calculating electrical cloud registration feature parameters, extracting gas features, and calculating gas concentration feature parameters based on collected multimodal operating data using a first-level algorithm. Three-dimensional point cloud data is acquired using LiDAR, expressed as: in, For vector data registration error, This represents the coordinate vector of the i-th 3D point acquired by the LiDAR during the current scanning cycle. Here, N represents the reference point cloud for the point cloud model, and N is the Nth data point in the 3D point cloud data. , For registering feature parameters of electric clouds, anomaly thresholds are set. The expression for the gas concentration detection vector is: in, Let the gas concentration vector be... This represents the measured concentration of sulfur dioxide gas. This is a measured value for the concentration of hydrogen sulfide gas. This is a measured value of carbon monoxide gas concentration. This represents the measured concentration of sulfur hexafluoride gas. For transpose; The gas concentration change rate feature is extracted, and the expression is: in, The rate of change of gas concentration. , Let be the concentration value of the k-th gas measured at time t. In time The measured concentration value of the kth gas, For time intervals, , Thresholds are used to determine the characteristic parameters of gas concentration.

[0011] As a preferred embodiment of the intelligent inspection and anomaly identification method for cable tunnels described in this invention, the extraction of acoustic features, electromagnetic radiation features, and environmental state features, and the calculation of environmental state feature parameters include: The characteristic expression for abnormal sounds is: in, Here, D represents the sound pressure level characteristic data, and D represents the analysis frame duration. For sound feature functions, , Thresholds are determined based on acoustic features; Partial discharge detection is performed using an electromagnetic radiation sensor, and the expression is: in, This is partial discharge detection data. For indicator functions, This represents the original signal amplitude collected by the electromagnetic radiation sensor at time t. The discharge detection threshold, , Thresholds for determining electromagnetic radiation characteristics; The temperature and humidity data are collected by a temperature and humidity sensor, and features are extracted using the following expression: in, , These are the first weighting coefficient and the second weighting coefficient, respectively. It is a comprehensive index of temperature and humidity. This is the temperature value. This is the humidity value. , The threshold for judging the environmental state.

[0012] This invention quantifies abnormal sounds by sound pressure level, identifies partial discharges by pulse counting, and assesses environmental conditions by temperature and humidity comprehensive index, thereby achieving standardized characterization of multimodal feature information and enabling unified modeling and fusion analysis of data from different sensors.

[0013] As a preferred embodiment of the intelligent inspection anomaly identification method for cable tunnels described in this invention, the step of feature-level fusion based on binary anomaly flag data through a two-level algorithm includes: The expression for fusing outlier data features with multimodal data is as follows: in, Feature data for multimodal data fusion , , The first, second, and third weights are respectively. For indicator functions, For vector data registration error, , To register the abnormal threshold of the feature parameters for electric clouds, The rate of change of gas concentration , Thresholds are used to determine the characteristic parameters of gas concentration. , The threshold for judging the environmental state. , Thresholds for feature data judgment in multimodal data fusion.

[0014] This invention achieves effective integration and complementarity of heterogeneous anomaly information by weighted fusion and unification into a comprehensive index S, reducing the complexity of direct processing of multi-source data. The weight design reflects the differences in the contribution of different modal anomalies to the overall risk, improving the engineering applicability of the final decision.

[0015] This invention provides a system for intelligent inspection and anomaly identification in cable tunnels.

[0016] To address the aforementioned technical problems, this invention provides the following technical solution: a system for intelligent inspection and anomaly identification of cable tunnels, comprising: a data acquisition module, a feature extraction module, and a data fusion and decision-making module. The data acquisition module synchronously collects the operating data of the cable tunnel through heterogeneous sensors, and obtains multimodal operating data including spatial structure information, gas composition information, sound signals, electromagnetic signals and environmental status information. The feature extraction module performs feature extraction and anomaly modeling based on the collected multimodal operation data, and generates binary anomaly flag data. The data fusion and decision module performs feature-level fusion of binary anomaly flag data through a two-level algorithm, and determines whether the device has an anomaly based on the fused features, thus completing the device anomaly detection.

[0017] The present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the intelligent inspection anomaly identification method for cable tunnels.

[0018] The present invention provides a computer-readable storage medium having a computer program stored thereon, characterized in that the computer program, when executed by a processor, implements the steps of the intelligent inspection anomaly identification method for cable tunnels.

[0019] The beneficial effects of this invention are as follows: This invention utilizes a first-level algorithm to extract features and model independent anomalies in each modal data, generating preliminary binary anomaly markers, thus solving the problem of high misjudgment rate of single sensor indicators. It introduces a second-level feature-level fusion algorithm to weight and synthesize each anomaly marker, generating unified multimodal fusion feature parameters, realizing cross-validation and collaborative judgment of multi-source anomaly information, thereby improving the accuracy and reliability of equipment anomaly identification in complex environments and overcoming the limitations of traditional monitoring methods that rely on isolated analysis and lack intelligent fusion. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.

[0021] Figure 1 The above is a flowchart of an intelligent inspection anomaly identification method for cable tunnels provided in one embodiment of the present invention. Detailed Implementation

[0022] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0023] Example 1, referring to Figure 1 This is one embodiment of the present invention, which provides a method for intelligent inspection and anomaly identification of cable tunnels, including: To address the limitations of traditional monitoring technologies, which typically detect only single types of anomalies and fail to comprehensively reflect equipment status, and the lack of intelligent analysis capabilities to process complex data, traditional monitoring technologies rely on single data indicators for judgment, often resulting in high false positive rates and inaccurate assessments. This invention provides an intelligent anomaly identification method for cable tunnel inspections.

[0024] S1: Synchronously collect cable tunnel operation data through heterogeneous sensors to obtain multimodal operation data including spatial structure information, gas composition information, sound signals, electromagnetic signals and environmental status information; S2: Based on the collected multimodal operational data, feature extraction and anomaly modeling are performed respectively to generate binary anomaly flag data; S3: The binary anomaly flag data is fused at the feature level through a two-level algorithm, and the device is judged to have an anomaly based on the fused features, thus completing the device anomaly detection.

[0025] Therefore, the synchronous acquisition of heterogeneous sensors forms a multi-dimensional data coverage of the cable tunnel's operating status. Dedicated feature extraction and independent anomaly modeling are designed for each modal data, generating preliminary binary anomaly flags to reduce misjudgment interference from single-source data. A feature-level fusion algorithm integrates these independent flags into a unified decision, enabling cross-validation and collaborative judgment of anomaly information from different dimensions, thereby enhancing the accuracy and reliability of the overall identification system.

[0026] Example 2, an embodiment of the present invention, provides a method for intelligent inspection and anomaly identification of cable tunnels based on the previous embodiment, including: In this embodiment of the application, step S1 involves synchronously collecting operational data of the cable tunnel using heterogeneous sensors to obtain multimodal operational data that includes spatial structure information, gas composition information, sound signals, electromagnetic signals, and environmental state information. This includes the following steps A1-A3: A1: Spatial structure information is collected through lidar, gas state information is collected through gas sensors, and sound information generated during operation is collected through acoustic sensors.

[0027] A2: Electromagnetic radiation information is collected through an electromagnetic radiation sensor, and environmental status information is collected through a temperature and humidity sensor.

[0028] A3: Generate multimodal operating data based on the collected spatial structure information, gas composition information, sound signals, electromagnetic signals and environmental state information.

[0029] In this embodiment of the application, the synchronous acquisition of cable tunnel operation data through heterogeneous sensors in step S1 is specifically manifested as follows: Three-dimensional point cloud data is acquired through lidar; gas concentration vector data is acquired through gas sensors; sound pressure signals and sampling frequencies are acquired through acoustic sensors; ultra-high frequency and bandwidth signals of electromagnetic radiation are acquired through electromagnetic radiation sensors; and temperature and humidity data are acquired through temperature and humidity sensors.

[0030] In an optional implementation, the synchronous acquisition of cable tunnel operation data by heterogeneous sensors in step S1 can also employ fusion or smart sensors, using integrated multi-element environmental sensors (such as temperature, humidity, air pressure, and light intensity) to simplify deployment. Smart sensors with edge computing capabilities are used to preprocess the raw signals (such as sound and vibration) at the acquisition end before uploading the feature data.

[0031] In another optional implementation, the synchronous acquisition of cable tunnel operation data via heterogeneous sensors in step S1 can also employ a different data acquisition and characterization method. For spatial structure information, in addition to lidar point clouds, at specific inspection nodes or fixed monitoring points, three-dimensional structured light scanning or high-precision depth cameras can be used to acquire surface three-dimensional models and texture information. For electromagnetic signals, in addition to ultra-high frequency signals, very low frequency / power frequency electromagnetic field strength data can be acquired simultaneously for monitoring power frequency interference and grounding status.

[0032] In this embodiment of the application, the synchronous acquisition of cable tunnel operation data by heterogeneous sensors in step S1 establishes a clear data interface and processing foundation for the unified aggregation and feature extraction of multimodal data, ensuring the accuracy and consistency of the conversion from physical perception to digital information.

[0033] In this embodiment of the application, step S2 involves feature extraction and anomaly modeling based on the collected multimodal operational data to generate binary anomaly flag data, including the following steps B1-B3: B1: Based on the collected multimodal operation data, the electric cloud registration feature parameters are calculated through a first-level algorithm, gas features are extracted, and gas concentration feature parameters are calculated.

[0034] The expression for the gas concentration detection vector is: in, Let the gas concentration vector be... This represents the measured concentration of sulfur dioxide gas. This is a measured value for the concentration of hydrogen sulfide gas. This is a measured value of carbon monoxide gas concentration. This represents the measured concentration of sulfur hexafluoride gas. For transpose; The gas concentration change rate feature is extracted, and the expression is: in, The rate of change of gas concentration. , Let be the concentration value of the k-th gas measured at time t. In time The measured concentration value of the kth gas, For time intervals, , The threshold for judging gas concentration characteristic parameters is the threshold for whether there is normal or abnormal data in the rate of change database.

[0035] B2: Extract acoustic features, electromagnetic radiation features, and environmental state features, and calculate state environmental feature parameters.

[0036] The characteristic expression for abnormal sounds is: in, Here, D represents the sound pressure level characteristic data, and D represents the analysis frame duration. For sound feature functions, , The threshold for judging acoustic features is the threshold for whether there is normal or abnormal data in the rate of change database. Partial discharge detection is performed using an electromagnetic radiation sensor, and the expression is: in, This is partial discharge detection data. For indicator functions, This represents the original signal amplitude collected by the electromagnetic radiation sensor at time t. The discharge detection threshold, , The threshold for judging electromagnetic radiation characteristics is the threshold for whether there is normal or abnormal data in the rate of change database.

[0037] B3: Based on the electric cloud registration feature parameters, gas concentration feature parameters, acoustic features, electromagnetic radiation features, environmental state features, and state-environment feature parameters, preliminary anomaly detection is performed to generate binary anomaly flag data.

[0038] In this embodiment of the application, the calculation of electric cloud registration feature parameters by a first-level algorithm in step B1 is specifically manifested as follows: Three-dimensional point cloud data is acquired using LiDAR, expressed as: in, For vector data registration error, This represents the coordinate vector of the i-th 3D point acquired by the LiDAR during the current scanning cycle. Here, N represents the reference point cloud for the point cloud model, and N is the Nth data point in the 3D point cloud data. , The abnormal threshold for the registration feature parameters of the electric cloud is the threshold for whether there is normal or abnormal data in the rate of change database.

[0039] In an optional implementation, the calculation of the electric cloud registration feature parameters in step B1 using the first-level algorithm can also employ local registration error analysis of key areas. Instead of calculating the average error of the overall point cloud, the registration error between the point cloud and the reference point cloud is calculated separately for the point cloud subsets of predefined key equipment areas (such as around cable joints, supports, and grounding boxes) and high-risk sections (such as settlement-sensitive zones) within the tunnel.

[0040] In another alternative implementation, the calculation of the electrical cloud registration feature parameters by the first-level algorithm in step B1 can also be carried out by segmented or layered registration comparison, dividing the tunnel into several continuous segments along the longitudinal direction, or layering the point cloud by height (such as focusing on the arch, sidewall, and cable layer), and calculating the registration error of each segment or layer separately.

[0041] In this embodiment of the application, in step B1, the electric cloud registration feature parameters are calculated by a first-level algorithm and the complex geometric comparison problem is transformed into a clear binary anomaly judgment by passing a preset threshold, thereby improving the objectivity and automation level of structural anomaly detection.

[0042] In this embodiment of the application, the calculation of state environment characteristic parameters in step B2 is specifically manifested as follows: The temperature and humidity data are collected by a temperature and humidity sensor, and features are extracted using the following expression: in, , These are the first and second weighting coefficients, which can be set manually based on experience. It is a comprehensive index of temperature and humidity. This is the temperature value. This is the humidity value. , The threshold for judging the environmental state.

[0043] In an optional implementation, the calculation of environmental characteristic parameters in step B2 can also employ the construction of a multi-dimensional state mapping based on the comfort zone or equipment operating condition zone. This mapping is not limited to a single index; a two-dimensional state plane with temperature and humidity as coordinate axes can be established. Based on equipment operating standards or historical safety data, safe zones, warning zones, and danger zones can be divided on the plane. Characteristic parameters can be defined as the distance of the current monitoring point (T, H) relative to the boundary of the safe zone or the area label it belongs to.

[0044] In another optional implementation, the calculation of environmental characteristic parameters in step B2 can also use dew point temperature or absolute humidity as core parameters. In key sections where high humidity easily causes condensation and leads to insulation failure, the dew point temperature can be directly calculated and monitored, or absolute humidity (the mass of water vapor in a unit volume of air) can be used as a characteristic parameter. By comparing it with the current ambient temperature, the risk of condensation can be directly determined.

[0045] In the embodiments of this application, the calculation of environmental characteristic parameters in step B2 provides a unified and adjustable judgment benchmark, making the judgment of abnormal environmental conditions more direct, and can flexibly adapt to the differences in environmental sensitivity of different tunnel sections or seasons by adjusting the weights.

[0046] It should be noted that the structural deformation is quantified by point cloud registration error, and the dynamics of gas components are monitored by dual indicators of concentration vector and rate of change. The acoustic vibration and electromagnetic states are characterized by sound pressure level integral and partial discharge frequency, respectively. A temperature and humidity weighted comprehensive index is constructed to reflect environmental conditions. After the quantified features are judged by preset thresholds, standardized binary anomaly flags can be generated, overcoming the misjudgment caused by traditional methods relying on a single indicator or subjective experience.

[0047] In the embodiment of the present application, in step S3, the binary abnormal flag data is subjected to feature-level fusion through a secondary algorithm, and based on the fused features, it is determined whether the device is abnormal, completing the device abnormal detection, including the following steps C1 - C3: C1: Based on the binary abnormal flag data, perform feature-level fusion through a secondary algorithm, and output the multi-modal data fusion feature parameters.

[0048] C2: Based on the multi-modal data fusion feature parameters, judge the device state, and output abnormal alarm information according to the device state.

[0049] Judge whether it is abnormal according to the abnormal data threshold of each modality in the historical database. S > 0.5 is a serious abnormality, 0.3 < S ≤ 0.5 is an abnormality, and S ≤ 0.3 is normal.

[0050] C3: According to the abnormal alarm information, output the recommended measures, and conduct abnormal troubleshooting according to the recommended measures.

[0051] In the embodiment of the present application, the specific manifestation of performing feature-level fusion through a secondary algorithm in step S3 is as follows: Perform multi-modal data fusion on the abnormal data features, and the expression is: Among them, is the feature data of multi-modal data fusion, , , are the first, second, and third weights respectively. The statistical weights of each modality in each state are allocated historically, and the weights can be manually allocated by humans. is the indicator function, is the vector data registration error, , is the abnormal threshold of the electro-cloud registration feature parameter, is the gas concentration change rate, , is the gas concentration feature parameter judgment threshold, , is the environmental state judgment threshold, , is the judgment threshold of the feature data of multi-modal data fusion, which is the data threshold of normal or abnormal data in the change rate database.

[0052] In an optional implementation, the feature-level fusion performed in step S3 using a two-level algorithm can also employ rule-based logical decision tree fusion. First, it is determined whether there is direct evidence of high risk of structural displacement. If not, it is further checked whether there is a combination of indirect evidence such as simultaneous occurrence of gas concentration abrupt change and partial discharge. Finally, the comprehensive state level is output through preset Boolean logic rules.

[0053] In another optional implementation, the feature-level fusion in step S3 using the two-level algorithm can also employ a feature fusion method based on anomaly probability weighting. First, the anomaly judgment result is converted into the corresponding anomaly occurrence probability value. Based on the reliability of each modality under different device states in the historical database, corresponding weights are assigned to each anomaly probability. Finally, the anomaly probability of each modality is comprehensively calculated to obtain the overall anomaly degree of the device, and compared with the anomaly classification threshold to determine whether the device is abnormal and trigger an alarm.

[0054] In this embodiment of the application, feature-level fusion is performed in step S3 using a two-level algorithm, which improves the accuracy and reliability of judging the overall operating status of cable tunnels under multi-source information while ensuring computational efficiency and model interpretability.

[0055] In summary, by employing a multimodal sensing data acquisition and feature processing mechanism, the problems of incomplete information and high misjudgment rate in traditional single monitoring methods are solved. A multi-source heterogeneous data synchronous acquisition system covering structural deformation, gas composition, acoustic vibration, electromagnetic radiation, and environmental parameters is constructed. A first-level algorithm is used to apply independent threshold criteria to each modal feature to generate binary anomaly markers. Then, a second-level feature fusion algorithm is used to weight and synthesize the scattered binary markers into a unified state score, realizing collaborative diagnosis and hierarchical early warning of abnormal events.

[0056] Example 3 is an embodiment of the present invention, which provides a method for intelligent inspection and anomaly identification of cable tunnels. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through experiments.

[0057] Table 1 Monitoring Data Table

[0058] The data from each sensor has been preprocessed into standardized feature values. The point cloud registration error ΔP = 3.2 mm was calculated, and the SO2 concentration change rate was calculated. =1.5ppm / s, calculate the sound pressure level =65dB, statistical pulse count =8 times / second, calculate the comprehensive index =48.

[0059] Apply a threshold comparison to each feature: ΔP=3.2mm>ϵP=2.0mm→Abnormal =1.5ppm / s>ϵCSO2=1.0ppm / s→Abnormal =65dB>ϵSPL=60dB→Abnormal =8 times / second > ϵPD=5 times / second → Abnormal =48<ϵTH=50→Normal.

[0060] Perform weighted fusion, and set the weights as follows: =0.3, =0.2, =0.2, =0.15, =0.1, =0.05.

[0061] Calculate the fusion score: S=0.3*1+0.2*1+0.2*1+0.15*1+0.1*1+0.05*0=0.3+0.2+0.2+0.15+0.1+0=0.95 If the feature data judgment threshold ϵS=0.5 for multimodal data fusion is 0.95>0.5, then it is judged as a serious anomaly.

[0062] The system determines that the device has a serious malfunction and generates an alarm message: Alarm type: Equipment damage (may be caused by multiple factors such as deformation, overheating, mechanical failure, etc.).

[0063] Location: Coordinates of the anomaly point located by the lidar (x=10.5m, y=3.2m, z=1.8m).

[0064] Recommended measures: Immediately stop the machine for inspection, prioritizing the investigation of equipment deformation, gas leaks, and mechanical failures.

[0065] Example 4 is an embodiment of the present invention, which provides a system for intelligent inspection and anomaly identification of cable tunnels, including a data acquisition module, a feature extraction module, and a data fusion and decision-making module. The data acquisition module synchronously collects the operating data of the cable tunnel through heterogeneous sensors, and obtains multimodal operating data including spatial structure information, gas composition information, sound signals, electromagnetic signals and environmental status information. The feature extraction module performs feature extraction and anomaly modeling based on the collected multimodal operational data, generating binary anomaly flag data. The data fusion and decision module performs feature-level fusion of binary anomaly flag data and determines whether the device has an anomaly based on the fused features, thus completing the device anomaly detection.

[0066] This embodiment also provides an electronic device applicable to a method for intelligent inspection and anomaly identification of cable tunnels, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the method for intelligent inspection and anomaly identification of cable tunnels as proposed in the above embodiment.

[0067] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, it implements the intelligent inspection anomaly identification method for cable tunnels as proposed in the above embodiments.

[0068] The storage medium proposed in this embodiment belongs to the same inventive concept as the method for intelligent inspection and anomaly identification of cable tunnels proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0069] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0070] It should be noted that the above embodiments are only used to illustrate the technical solutions 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 solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for intelligent inspection and anomaly identification in cable tunnels, characterized in that: include, By synchronously collecting cable tunnel operation data through heterogeneous sensors, multimodal operation data including spatial structure information, gas composition information, sound signals, electromagnetic signals and environmental status information can be obtained. Based on the collected multimodal operational data, feature extraction and anomaly modeling are performed respectively to generate binary anomaly flag data; The binary anomaly flag data is fused at the feature level using a two-level algorithm, and the device is judged to have anomalies based on the fused features, thus completing the device anomaly detection.

2. The intelligent inspection anomaly identification method for cable tunnels as described in claim 1, characterized in that: The method of synchronously acquiring cable tunnel operation data through heterogeneous sensors to obtain multimodal operation data including spatial structure information, gas composition information, sound signals, electromagnetic signals, and environmental state information includes... Space structure information is collected using lidar, gas state information is collected using gas sensors, and sound information generated during operation is collected using acoustic sensors. Electromagnetic radiation information is collected by an electromagnetic radiation sensor, and environmental status information is collected by a temperature and humidity sensor. Multimodal operational data is generated based on the collected spatial structure information, gas composition information, sound signals, electromagnetic signals, and environmental state information.

3. The intelligent inspection anomaly identification method for cable tunnels as described in claim 2, characterized in that: The collected multimodal operational data is used for feature extraction and anomaly modeling to generate binary anomaly flag data, including: Based on the collected multimodal operation data, the electric cloud registration feature parameters are calculated through a first-level algorithm, gas features are extracted, and gas concentration feature parameters are calculated. Extract acoustic features, electromagnetic radiation features, and environmental state features, and calculate state-environment feature parameters; Based on the electric cloud registration feature parameters, gas concentration feature parameters, acoustic features, electromagnetic radiation features, environmental state features, and state-environment feature parameters, preliminary anomaly detection is performed to generate binary anomaly marker data.

4. The intelligent inspection anomaly identification method for cable tunnels as described in claim 3, characterized in that: The step of fusing binary anomaly flag data at the feature level using a two-level algorithm, and determining whether the device has an anomaly based on the fused features, to complete the device anomaly detection includes: Based on binary anomaly marker data, feature-level fusion is performed through a two-level algorithm to output multimodal data fusion feature parameters; Based on the feature parameters fused from multimodal data, the device status is determined, and abnormal alarm information is output according to the device status. Based on the abnormal alarm information, output suggested measures, and conduct anomaly investigation based on the suggested measures.

5. The intelligent inspection anomaly identification method for cable tunnels as described in claim 4, characterized in that: The collected multimodal operational data is used to calculate the electric cloud registration feature parameters using a first-level algorithm, extract gas features, and calculate gas concentration feature parameters, including... Three-dimensional point cloud data is acquired using LiDAR, expressed as: in, For vector data registration error, This represents the coordinate vector of the i-th 3D point acquired by the LiDAR during the current scanning cycle. Here, N represents the reference point cloud for the point cloud model, and N is the Nth data point in the 3D point cloud data. , For registering feature parameters of electric clouds, anomaly thresholds are set. The expression for the gas concentration detection vector is: in, Let the gas concentration vector be... This represents the measured concentration of sulfur dioxide gas. This is a measured value for the concentration of hydrogen sulfide gas. This is a measured value of carbon monoxide gas concentration. This represents the measured concentration of sulfur hexafluoride gas. For transpose; The gas concentration change rate feature is extracted, and the expression is: in, The rate of change of gas concentration. , Let be the concentration value of the k-th gas measured at time t. In time The measured concentration value of the kth gas, For time intervals, , Thresholds are used to determine the characteristic parameters of gas concentration.

6. The intelligent inspection anomaly identification method for cable tunnels as described in claim 5, characterized in that: The extraction of acoustic features, electromagnetic radiation features, and environmental state features, and the calculation of environmental state feature parameters include... The characteristic expression for abnormal sounds is: in, Here, D represents the sound pressure level characteristic data, and D represents the analysis frame duration. For sound feature functions, , Thresholds are determined based on acoustic features; Partial discharge detection is performed using an electromagnetic radiation sensor, and the expression is: in, This is partial discharge detection data. For indicator functions, This represents the original signal amplitude collected by the electromagnetic radiation sensor at time t. The discharge detection threshold, , Thresholds for determining electromagnetic radiation characteristics; The temperature and humidity data are collected by a temperature and humidity sensor, and features are extracted using the following expression: in, , These are the first weighting coefficient and the second weighting coefficient, respectively. It is a comprehensive index of temperature and humidity. This is the temperature value. This is the humidity value. , The threshold for judging the environmental state.

7. The intelligent inspection anomaly identification method for cable tunnels as described in claim 6, characterized in that: The feature-level fusion based on binary anomaly flag data using a two-level algorithm includes, The expression for fusing outlier data features with multimodal data is as follows: in, Feature data for multimodal data fusion , , The first, second, and third weights are respectively. For indicator functions, For vector data registration error, , To register the abnormal threshold of the feature parameters for electric clouds, The rate of change of gas concentration. , Thresholds are used to determine the characteristic parameters of gas concentration. , The threshold for judging the environmental state. , Thresholds for feature data judgment in multimodal data fusion.

8. A system for intelligent inspection and anomaly identification of cable tunnels, employing the intelligent inspection and anomaly identification method for cable tunnels as described in any one of claims 1 to 7, characterized in that... It includes a data acquisition module, a feature extraction module, and a data fusion and decision-making module. The data acquisition module synchronously collects the operating data of the cable tunnel through heterogeneous sensors, and obtains multimodal operating data including spatial structure information, gas composition information, sound signals, electromagnetic signals and environmental status information. The feature extraction module performs feature extraction and anomaly modeling based on the collected multimodal operation data, and generates binary anomaly flag data. The data fusion and decision module performs feature-level fusion of binary anomaly flag data through a two-level algorithm, and determines whether the device has an anomaly based on the fused features, thus completing the device anomaly detection.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the intelligent inspection anomaly identification method for cable tunnels according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the intelligent inspection anomaly identification method for cable tunnels according to any one of claims 1 to 7.

Images