A method, device, equipment and medium for determining a defect detection strategy of a cable

Abstract

The application discloses a method, device, equipment and medium for determining a defect detection strategy of a cable, obtains historical detection data of historical defect detection of the cable, combines any defect detection accuracy and a voltage grade of any cable to obtain at least one first detection label, extracts to-be-processed historical detection data of the first detection label, takes the first detection label as a first target detection label if the number of the to-be-processed historical detection data of any first detection label is not less than a first detection data quantity threshold, takes a defect detection strategy involved in the to-be-processed historical detection data of the first target detection label as a first defect detection strategy, extracts first to-be-processed historical detection data of the first defect detection strategy from the historical detection data, determines a first detection accuracy value of the first defect detection strategy based on the first to-be-processed historical detection data, and determines a target defect detection strategy based on the first detection accuracy value of the first defect detection strategy, thereby improving detection accuracy.

Description

Technical Field

[0001] This application relates to the field of cables, and more particularly to a method, apparatus, equipment, and medium for determining a defect detection strategy for cables. Background Technology

[0002] Non-contact technologies, such as infrared imaging and laser ultrasonic testing, can be used for cable defect detection. However, in practical applications, these technologies often use a single method for uniform detection of cables in different application scenarios, making it difficult to adapt to varying scenarios and resulting in insufficient accuracy of cable defect detection results. Summary of the Invention

[0003] This application provides a method, apparatus, equipment, and medium for determining a cable defect detection strategy, in order to solve the problem that using a single method for uniform detection of cable defects in different application scenarios is difficult to adapt to the differences in scenarios, resulting in insufficient accuracy of cable defect detection results.

[0004] To address the aforementioned technical problems, embodiments of this application provide a method for determining a defect detection strategy for cables, comprising: Acquire historical inspection data of at least one historical defect detection of the cable; the historical defect detection involves at least one defect detection accuracy and at least one defect detection strategy; the cable has at least one voltage level; By combining any of the aforementioned defect detection accuracy and any of the aforementioned voltage levels, at least one first detection tag is obtained, and the unprocessed historical detection data of any of the first detection tags is extracted from the historical detection data. For any of the first detection tags, if the number of historical detection data to be processed for the first detection tag is not less than a preset first detection data number threshold, then the first detection tag is used as the first target detection tag. Take at least one defect detection strategy involved in the historical detection data to be processed of the first target detection label as the first defect detection strategy, and extract the first historical detection data to be processed of any of the first defect detection strategies from the historical detection data. Based on the first historical detection data to be processed, a first detection accuracy value of any of the first defect detection strategies is determined, and based on the first detection accuracy value of at least one of the first defect detection strategies, a target defect detection strategy for the first target detection tag is determined.

[0005] In this embodiment, by acquiring historical detection data, the defect detection accuracy and voltage level are combined to obtain a first detection label, and the historical detection data to be processed corresponding to each first detection label is extracted. Based on the amount of data, the first target detection label with a sufficient number of historical detection data to be processed is selected. Based on the first detection accuracy value of each defect detection strategy under the first target detection label, the target defect detection strategy is determined. This realizes a detection strategy quantitative screening mechanism driven by historical detection data and based on voltage level and detection accuracy as the basis for scenario division. It can determine the appropriate detection strategy for scenarios with different voltage levels and accuracy requirements, avoiding the use of a single method for uniform detection of cable defects in different application scenarios, thereby improving the accuracy and scenario adaptability of cable defect detection results.

[0006] Optionally, the method for determining the defect detection strategy for the cable further includes: For any of the first detection tags, if the number of unprocessed historical detection data for the first detection tag is less than the first detection data quantity threshold, then the first detection tag is taken as the unprocessed detection tag, and the unprocessed historical detection data of at least one of the unprocessed detection tags is merged according to the voltage level of the unprocessed detection tag to obtain the unprocessed historical detection data of at least one second target detection tag; the number of unprocessed historical detection data of the second target detection tag is greater than the first detection data quantity threshold. The first defect detection strategy involved in the historical detection data to be processed of the second target detection label is used as the second defect detection strategy, and the second historical detection data to be processed of any second defect detection strategy is extracted from the historical detection data. Based on the second historical detection data to be processed, a second detection accuracy value of any second defect detection strategy is determined, and based on the second detection accuracy value of at least one second defect detection strategy, a target defect detection strategy for the second target detection label is determined.

[0007] In this embodiment, by performing voltage level adjacency merging on the first detection label whose number of historical detection data to be processed is less than the first detection data quantity threshold, a second target detection label with sufficient data volume is generated. Based on the second detection accuracy value of each defect detection strategy under the merged second target detection label, the target defect detection strategy is determined. In scenarios with sparse historical data, by effectively aggregating data from adjacent voltage levels, the sample capacity is expanded to support the reliable calculation of detection accuracy values. This avoids the problem of statistical distortion due to insufficient data volume and ensures the continuity and physical rationality of the merged scenario in terms of voltage levels, thereby improving the robustness and applicability of the detection strategy determination method in scenarios with unbalanced data.

[0008] Optionally, the method for determining the defect detection strategy for the cable further includes: Using the unprocessed historical detection data of the first target detection tag, the target defect detection strategy of the first target detection tag, the unprocessed historical detection data of the second target detection tag, and the target defect detection strategy of the second target detection tag, a preset machine learning model is trained to obtain a definite model of the cable defect detection strategy.

[0009] In this embodiment, a preset machine learning model is trained using historical detection data to be processed with first and second target detection labels and their corresponding target defect detection strategies to obtain a model for determining cable defect detection strategies. High-quality labeled data selected based on detection accuracy values ​​are used as training samples, enabling the machine learning model to learn the complex mapping relationship between detection accuracy and detection strategies in different scenarios. This provides a model foundation for rapid reasoning and real-time recommendation of detection strategies in new scenarios, further improving the intelligence and generalization ability of detection strategy determination.

[0010] Optionally, the at least one historical defect detection targets cables with different temperature information and / or different humidity information; the method for determining the defect detection strategy for the cable further includes: If the number of unprocessed historical detection data of the first target detection label is not less than the preset second detection data number threshold, then the first target detection label is used as the first processable detection label. If the number of unprocessed historical detection data for the second target detection label is not less than the second detection data number threshold, then the second target detection label is used as the second processable detection label. Combine any of the temperature information and any of the humidity information involved in the historical detection data to be processed of the first processable detection tag, and combine any of the temperature information and any of the humidity information involved in the historical detection data to be processed of the second processable detection tag to obtain at least one second detection tag. Extract the processable historical detection data of any second detection tag from the unprocessed historical detection data of the first processable detection tag and the unprocessed historical detection data of the second processable detection tag, and determine the target defect detection strategy of the second detection tag based on the processable historical detection data of the second detection tag.

[0011] In this embodiment, under the condition that the amount of historical detection data to be processed meets the second detection data quantity threshold, temperature and humidity information are further introduced into the first target detection label and the second target detection label to obtain the second detection label, and the corresponding processable historical detection data is extracted to determine the target defect detection strategy. Based on the scenario division of voltage level and detection accuracy, the temperature and humidity environment dimension is further introduced according to the condition of sufficient data to refine the scenario, so that the determination of the detection strategy can take into account the impact of the temperature and humidity difference of the on-site environment on the sensitivity and accuracy of the detection method, thereby improving the fine adaptation capability of the detection strategy under different environmental conditions.

[0012] Optionally, the at least one historical defect detection targets cables with different cable attribute information and / or different environmental operating condition information; the step of determining the target defect detection strategy for the second detection tag based on the processable historical detection data of the second detection tag includes: For any of the second detection tags, if the processable historical detection data of the second detection tag is less than the preset third detection data quantity threshold, then the target defect detection strategy of the second detection tag is determined according to the combination process of the second detection tags. If the processable historical detection data of the second detection tag is not less than the preset fourth detection data quantity threshold, then the cable attribute information and environmental condition information of the cable are used as the feature information to be processed of the cable, and the processable historical detection data of the second detection tag is clustered according to at least one feature information to be processed involved in the processable historical detection data of the second detection tag to obtain at least one first processable historical detection data set; the fourth detection data quantity threshold is greater than the third detection data quantity threshold. Based on the at least one first processable historical detection data set, determine the target defect detection strategy for the second detection label; If the processable historical detection data of the second detection tag is less than the fourth detection data quantity threshold and not less than the third detection data quantity threshold, then the processable historical detection data of the second detection tag is input into the determination model of the cable defect detection strategy to obtain the target defect detection strategy of the second detection tag.

[0013] In this embodiment, by setting a third and a fourth threshold for the amount of historical detection data that can be processed for the second detection tag, the data volume is divided into three levels: insufficient, moderate, and sufficient. Rollback processing, direct model calculation processing, and clustering processing based on cable attribute information and environmental conditions are then performed respectively. This achieves adaptive hierarchical decision-making for determining detection strategies in detailed scenarios: when data is insufficient, rollback to existing results at the previous level ensures reliability; when data is moderate, direct model calculation ensures efficiency; and when data is sufficient, deep clustering mining based on cable attributes and environmental conditions is introduced to improve accuracy. This approach balances the accuracy and stability of strategy determination under different data volume conditions.

[0014] Optionally, the process of determining the target defect detection strategy for the second detection tags based on the combination of the second detection tags includes: If the second detection label is obtained based on the first processable detection label, then the target defect detection strategy corresponding to the first processable detection label is used as the target defect detection strategy of the second detection label. If the second detection label is obtained based on the second processable detection label, then the target defect detection strategy corresponding to the second processable detection label is used as the target defect detection strategy of the second detection label.

[0015] In this embodiment, the target defect detection strategy is determined based on the combined source of the second detection tag. If the second detection tag is obtained based on the first processable detection tag, the target defect detection strategy corresponding to the first processable detection tag is directly reused; if it is obtained based on the second processable detection tag, the target defect detection strategy corresponding to the second processable detection tag is reused. Therefore, when the amount of historical detection data that can be processed in a detailed scene is insufficient, it can automatically backtrack to the target defect detection strategy already determined in the previous level scene to which the detailed scene belongs. Strategy inheritance can be completed without additional calculation, ensuring the integrity and reliability of the detection strategy determination in data-sparse scenes and achieving a smooth transition between broad and detailed scenes.

[0016] Optionally, determining the target defect detection strategy for the second detection tag based on the at least one first processable historical detection data set includes: If the number of processable historical detection data in at least one of the first processable historical detection datasets is not less than the third detection data quantity threshold, then the first processable historical detection datasets are combined into a target processable historical detection dataset. If at least one of the first processable historical detection data sets contains a first processable historical detection data set whose number of processable historical detection data is less than the third detection data quantity threshold, then based on the quantity of the first processable historical detection data set, a reference quantity for the clustering process is determined, and the processable historical detection data of the second detection label is clustered according to the reference quantity to obtain at least one second processable historical detection data set. If the number of processable historical detection data in at least one of the second processable historical detection datasets is not less than the third detection data quantity threshold, then the second processable historical detection datasets are combined into the target processable historical detection dataset. If the number of processable historical detection data in any of the second processable historical detection data sets is less than the third detection data quantity threshold, then the second processable historical detection data set is merged with the second processable historical detection data set that is not less than the third detection data quantity threshold, so that the number of processable historical detection data in each of the second processable historical detection data sets is not less than the third detection data quantity threshold, thus obtaining the target processable historical detection data set. For any of the target processable historical detection data sets, the processable historical detection data in the target processable historical detection data set is input into the determination model of the cable defect detection strategy to obtain the target defect detection strategy for the target processable historical detection data set under the second detection label.

[0017] In this embodiment, the data volume of the first processable historical detection data set obtained by clustering is checked. Sets with insufficient data volume are processed by reducing the number of clusters and re-clustering or merging with the nearest neighbor set until the data volume of all target processable historical detection data sets meets the third detection data volume threshold. Then, the target defect detection strategy for each set is obtained by inputting it into the cable defect detection strategy determination model. In the process of clustering refinement based on cable attributes and environmental conditions, the iterative clustering number adjustment or nearest neighbor merging mechanism ensures that each data set that finally participates in the model calculation has sufficient sample size to support it. This avoids the problem that excessive clustering leads to sparse data volume in some sets, which affects the reliability of the model output. At the same time, the feature correlation between cluster sets is preserved by nearest neighbor merging, which improves the accuracy and statistical stability of detection strategy determination in deep refinement scenarios.

[0018] This application also discloses an apparatus for determining a defect detection strategy for cables, comprising: A historical inspection data acquisition module is used to acquire historical inspection data of at least one historical defect inspection of the cable; the historical defect inspection involves at least one defect detection accuracy and at least one defect detection strategy; the cable has at least one voltage level; The first detection tag obtaining module is used to combine any of the defect detection accuracy and any of the voltage levels to obtain at least one first detection tag, and to extract the unprocessed historical detection data of any of the first detection tags from the historical detection data. The first target detection label is used as a module to, for any first detection label, if the number of unprocessed historical detection data of the first detection label is not less than a preset first detection data number threshold, then the first detection label is used as the first target detection label. The first pending historical detection data extraction module is used to take at least one defect detection strategy involved in the pending historical detection data of the first target detection tag as the first defect detection strategy, and extract the first pending historical detection data of any of the first defect detection strategies from the historical detection data. The first target defect detection strategy determination module is used to determine a first detection accuracy value of any of the first defect detection strategies based on the first historical detection data to be processed, and to determine the target defect detection strategy of the first target detection tag based on the first detection accuracy value of at least one of the first defect detection strategies.

[0019] Optionally, the device for determining the defect detection strategy for the cable further includes: The merging processing module is configured to, for any first detection tag, if the number of unprocessed historical detection data for the first detection tag is less than the first detection data quantity threshold, then take the first detection tag as the unprocessed detection tag, and merge the unprocessed historical detection data of at least one of the unprocessed detection tags according to the voltage level of the unprocessed detection tag to obtain unprocessed historical detection data of at least one second target detection tag; the number of unprocessed historical detection data of the second target detection tag is greater than the first detection data quantity threshold. The second pending historical detection data extraction module is used to take at least one defect detection strategy involved in the pending historical detection data of the second target detection tag as the second defect detection strategy, and extract the second pending historical detection data of any second defect detection strategy from the historical detection data. The second target defect detection strategy determination module is used to determine the second detection accuracy value of any second defect detection strategy based on the second historical detection data to be processed, and to determine the target defect detection strategy of the second target detection tag based on the second detection accuracy value of at least one second defect detection strategy.

[0020] Optionally, the device for determining the defect detection strategy for the cable further includes: The model acquisition module is used to train a preset machine learning model using the unprocessed historical detection data of the first target detection label, the target defect detection strategy of the first target detection label, the unprocessed historical detection data of the second target detection label, and the target defect detection strategy of the second target detection label, so as to obtain a determination model of the cable defect detection strategy.

[0021] Optionally, the at least one historical defect detection targets cables with different temperature information and / or different humidity information; the cable defect detection strategy determination device further includes: The first processable detection tag is used as a module to use the first target detection tag as the first processable detection tag if the number of unprocessed historical detection data of the first target detection tag is not less than a preset second detection data number threshold. The second processable detection tag is used as a module to use the second target detection tag as the second processable detection tag if the number of unprocessed historical detection data of the second target detection tag is not less than the second detection data number threshold. The second detection tag obtaining module is used to combine any of the temperature information and any of the humidity information involved in the historical detection data to be processed of the first processable detection tag, and to combine any of the temperature information and any of the humidity information involved in the historical detection data to be processed of the second processable detection tag, to obtain at least one second detection tag. The third target defect detection strategy determination module is used to extract the processable historical detection data of any second detection tag from the unprocessed historical detection data of the first processable detection tag and the unprocessed historical detection data of the second processable detection tag, and determine the target defect detection strategy of the second detection tag based on the processable historical detection data of the second detection tag.

[0022] Optionally, the at least one historical defect detection targets cables with different cable attribute information and / or different environmental operating condition information; the third target defect detection strategy determination module includes: The first target defect detection strategy determination submodule is used to determine the target defect detection strategy of the second detection tag based on the combination process of the second detection tags if the processable historical detection data of the second detection tag is less than the preset third detection data quantity threshold. The clustering processing submodule is used to, if the processable historical detection data of the second detection tag is not less than a preset fourth detection data quantity threshold, then take the cable attribute information and environmental condition information of the cable as the feature information to be processed of the cable, and perform clustering processing on the processable historical detection data of the second detection tag according to at least one feature information to be processed involved in the processable historical detection data of the second detection tag to obtain at least one first processable historical detection data set; the fourth detection data quantity threshold is greater than the third detection data quantity threshold. The second target defect detection strategy determination submodule is used to determine the target defect detection strategy of the second detection tag based on the at least one first processable historical detection data set. The target defect detection strategy acquisition submodule is used to input the processable historical detection data of the second detection tag into the determination model of the cable defect detection strategy if the processable historical detection data of the second detection tag is less than the fourth detection data quantity threshold and not less than the third detection data quantity threshold, so as to obtain the target defect detection strategy of the second detection tag.

[0023] Optionally, the first target defect detection strategy determination submodule includes: The first target defect detection strategy is used as a unit to use the target defect detection strategy corresponding to the first processable detection label as the target defect detection strategy of the second detection label if the second detection label is obtained based on the first processable detection label. The second target defect detection strategy is used as a unit to use the target defect detection strategy corresponding to the second processable detection label as the target defect detection strategy of the second detection label if the second detection label is obtained based on the second processable detection label.

[0024] Optionally, the second target defect detection strategy determination submodule includes: The first target processable historical detection dataset is combined into a unit, which is used to combine the first processable historical detection dataset into a target processable historical detection dataset if the number of processable historical detection data in at least one of the first processable historical detection datasets is not less than the third detection data number threshold. A clustering processing unit is configured to determine a reference number for clustering processing based on the number of the first processable historical detection data sets if at least one of the first processable historical detection data sets contains a first processable historical detection data set whose number of processable historical detection data is less than the third detection data quantity threshold, and to perform clustering processing on the processable historical detection data of the second detection label according to the reference number to obtain at least one second processable historical detection data set. The second target processable historical detection dataset is combined into a unit, which is used to combine the second processable historical detection dataset into the target processable historical detection dataset if the number of processable historical detection data in at least one of the second processable historical detection datasets is not less than the third detection data number threshold. The target processable historical detection data set obtaining unit is used to merge the second processable historical detection data set with the second processable historical detection data set that is not less than the third detection data quantity threshold if the number of processable historical detection data in any second processable historical detection data set is less than the third detection data quantity threshold, so that the number of processable historical detection data in each of the second processable historical detection data sets is not less than the third detection data quantity threshold, thereby obtaining the target processable historical detection data set; The target defect detection strategy obtaining unit is used to input the processable historical detection data in the target processable historical detection data set into the determination model of the cable defect detection strategy for any target processable historical detection data set, so as to obtain the target defect detection strategy for the target processable historical detection data set under the second detection label.

[0025] This application also discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the method described in this application.

[0026] This application also discloses a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the method described in this application. Attached Figure Description

[0027] Figure 1 This is a flowchart illustrating the steps of a method for determining a defect detection strategy for cables provided in an embodiment of this application. Figure 2 This is a flowchart illustrating the steps of a method for determining a defect detection strategy for cables, as provided in another embodiment of this application. Figure 3 This is a structural block diagram of a device for determining a defect detection strategy for cables, provided in an embodiment of this application. Detailed Implementation

[0028] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0029] It should be understood that the step numbers used in the text are for ease of description only and are not intended to limit the order in which the steps are performed.

[0030] To facilitate understanding of the technical solutions and effects of the embodiments of this application, the relevant technologies of this application will be briefly described below.

[0031] In related technologies, cable defects can be detected through manual inspection or contact instruments. However, these methods are not only inefficient but also have limited ability to identify internal cable defects. Furthermore, contact instruments require power outages before testing, severely impacting power supply continuity.

[0032] Non-contact technologies, such as infrared imaging and laser ultrasonic testing, can also be used for cable defect detection. However, in practical applications, these technologies often use a single method for uniform detection of cables in different application scenarios, making it difficult to adapt to varying scenarios and resulting in insufficient accuracy of cable defect detection results.

[0033] Reference Figure 1 The diagram illustrates a flowchart of the steps involved in determining a defect detection strategy for a cable according to an embodiment of this application. Specifically, the method may include the following steps: Step 101: Obtain historical detection data for at least one historical defect detection of the cable; the historical defect detection involves at least one defect detection accuracy and at least one defect detection strategy; the cable has at least one voltage level; In this embodiment, at least one historical defect detection can be a non-contact cable defect detection. Historical detection data for any given historical defect detection can include: length information of the specific segment of the cable being tested, defect detection strategy, defect detection accuracy, cable grade, temperature and humidity information of the cable during detection, environmental conditions of the cable, cable attribute information, defect characteristic information, and verification defect data. The defect detection strategy can be a single detection method or a combination of two or more detection methods. Defect detection accuracy can include three different levels: low, medium, and high. The cable's environmental conditions information includes various environmental parameters that affect defect detection. Cable attribute information records various attribute characteristics of the cable itself. Defect characteristic information includes the location of the detected defect, the specific type of the detected defect, and the severity level of the defect determined by the detection. Verification defect data, serving as verification data for the detection results, includes the actual location of the defect, the actual type of the defect, and the actual severity level of the defect.

[0034] Defect detection strategies can include pulsed laser ultrasound, laser-induced infrared thermography, laser photoluminescence imaging, infrared thermography, ultraviolet imaging, Laser-Ultrasonic+UHF (laser ultrasound combined with ultra-high frequency), LU+millimeter-wave radar (laser ultrasound combined with millimeter-wave radar), IR+UV+HFCT (infrared thermography combined with ultraviolet imaging combined with high-frequency current transformer), LIP+IR (laser-induced photoluminescence combined with infrared thermography), high-frequency TDR+ECA (high-frequency time domain reflectometer combined with eddy current detection), etc.

[0035] Temperature information can include high temperature, medium temperature, and low temperature; humidity information can include dry and wet; and voltage level values ​​can be 35kV, 64kV, 110kV, 220kV, etc.

[0036] Environmental operating condition information includes atmospheric pressure, wind speed, solar irradiance, electromagnetic interference level, etc., while cable attribute information includes insulation material type, service life, total section length, metal sheath material, laying method, etc.

[0037] Defect types can include partial discharge, metal corrosion, insulation scratches, water trees, and joint abnormalities. Both the detected severity level and the actual severity level include Level 1, Level 2, Level 3, and Level 4, which represent minor, moderate, severe, and critical, respectively.

[0038] Step 102: Combine any of the aforementioned defect detection accuracy and any of the aforementioned voltage levels to obtain at least one first detection tag, and extract the unprocessed historical detection data of any of the first detection tags from the historical detection data; In this embodiment, the defect detection accuracy and voltage level contained in the historical detection data of at least one historical defect detection are used as the core classification criteria to systematically classify and organize the historical detection data. Through this classification operation, any defect detection accuracy and any voltage level are combined to obtain at least one first detection tag, and the historical detection data to be processed for any first detection tag is extracted.

[0039] Step 103: For any of the first detection tags, if the number of historical detection data to be processed for the first detection tag is not less than the preset first detection data number threshold, then the first detection tag is used as the first target detection tag. In this embodiment, the amount of unprocessed historical detection data for each first detection tag is statistically analyzed, and the statistically obtained quantity is compared with a pre-set first detection data quantity threshold. If the amount of unprocessed historical detection data for a certain first detection tag is not less than the first detection data quantity threshold, it means that the amount of unprocessed historical detection data for that first detection tag is sufficient to meet the needs of subsequent analysis applications. The first detection tag is then used as the first target detection tag, that is, the first detection tag is marked as non-unpopular to clarify its data quantity compliance status. The value of the first detection data quantity threshold can be 30.

[0040] Step 104: Take at least one defect detection strategy involved in the historical detection data to be processed of the first target detection tag as the first defect detection strategy, and extract the first historical detection data to be processed of any of the first defect detection strategies from the historical detection data. In this embodiment of the application, for the historical detection data to be processed of the first target detection label, the defect detection strategy used for each historical defect detection is extracted from the historical detection data to be processed. All the extracted defect detection strategies are summarized and sorted, and after removing duplicates, the first defect detection strategy of the first target detection label is obtained, which includes multiple different defect detection strategies that have been used in the scenario corresponding to the first target detection label.

[0041] In this embodiment of the application, for each first defect detection strategy, all the first historical detection data corresponding to historical defect detection using the first defect detection strategy are selected from the unprocessed historical detection data of the first target detection label.

[0042] Step 105: Based on the first historical detection data to be processed, determine the first detection accuracy value of any of the first defect detection strategies, and based on the first detection accuracy value of at least one of the first defect detection strategies, determine the target defect detection strategy of the first target detection tag.

[0043] In this embodiment, based on the first historical detection data to be processed by the first defect detection strategy, the first detection accuracy value of each first defect detection strategy is calculated using the calculation formula of the detection accuracy value.

[0044] In this embodiment, the first detection accuracy values ​​of all first defect detection strategies corresponding to the first target detection label can be sorted from largest to smallest. According to a preset specified number, several first defect detection strategies corresponding to the top-ranked first detection accuracy values ​​are selected. These selected first defect detection strategies are integrated to form the set of optimal detection methods for the wide scene corresponding to the first target detection label, which is used as the target defect detection strategy of the first target detection label.

[0045] In this embodiment, by acquiring historical detection data, the defect detection accuracy and voltage level are combined to obtain a first detection label, and the historical detection data to be processed corresponding to each first detection label is extracted. Based on the amount of data, the first target detection label with a sufficient number of historical detection data to be processed is selected. Based on the first detection accuracy value of each defect detection strategy under the first target detection label, the target defect detection strategy is determined. This realizes a detection strategy quantitative screening mechanism driven by historical detection data and based on voltage level and detection accuracy as the basis for scenario division. It can determine the appropriate detection strategy for scenarios with different voltage levels and accuracy requirements, avoiding the use of a single method for uniform detection of cable defects in different application scenarios, thereby improving the accuracy and scenario adaptability of cable defect detection results.

[0046] In some embodiments of this application, the method for determining the defect detection strategy for the cable further includes: For any of the first detection tags, if the number of unprocessed historical detection data for the first detection tag is less than the first detection data quantity threshold, then the first detection tag is taken as the unprocessed detection tag, and the unprocessed historical detection data of at least one of the unprocessed detection tags is merged according to the voltage level of the unprocessed detection tag to obtain the unprocessed historical detection data of at least one second target detection tag; the number of unprocessed historical detection data of the second target detection tag is greater than the first detection data quantity threshold. The first defect detection strategy involved in the historical detection data to be processed of the second target detection label is used as the second defect detection strategy, and the second historical detection data to be processed of any second defect detection strategy is extracted from the historical detection data. Based on the second historical detection data to be processed, a second detection accuracy value for any of the second defect detection strategies is determined, and based on the second detection accuracy value of at least one of the second defect detection strategies, a target defect detection strategy for the second target detection label is determined.

[0047] In this embodiment, for any first detection tag, if the amount of historical detection data to be processed for the first detection tag is less than the first detection data quantity threshold, then the first detection tag is considered as a first detection tag with insufficient data and is treated as a detection tag to be processed. A specific merging strategy is adopted for the detection tags to be processed. During the merging process, the defect detection accuracy corresponding to these detection tags to be processed is kept unchanged, and only adjacent voltage levels are merged. This merging operation continues until the total amount of historical detection data corresponding to the merged combination is greater than or equal to the first detection data quantity threshold, and the number of first detection tags participating in the merging is two or more. After the merging is completed, the two or more first detection tags participating in the merging are marked as merged. At the same time, a second target detection tag is determined based on these merged first detection tags, and the second target detection tag is marked as non-unpopular, indicating that the amount of scene data corresponding to it after merging has reached the standard. For example, the first detection labels 110kV-low and 132kV-low, whose number of historical detection data to be processed is less than the first detection data number threshold, are merged, and the determined merged second target detection label is 110–132kV-low.

[0048] In this application embodiment, for those first detection labels that have neither obtained the non-unpopular label nor the merge label, that is, the first detection labels that still have insufficient historical detection data after screening and merging, will be marked as unpopular to indicate that the scene data corresponding to this type of label is insufficient and does not meet the conditions for direct use in subsequent in-depth analysis.

[0049] In this embodiment, for the unprocessed historical detection data of the second target detection label, the defect detection strategy used for each historical defect detection is extracted from the unprocessed historical detection data. All extracted defect detection strategies are summarized and sorted, and duplicates are removed to obtain the second defect detection strategy for the second target detection label. In the unprocessed historical detection data of two or more first detection labels corresponding to the second target detection label, all second unprocessed historical detection data corresponding to historical defect detections completed using the second defect detection strategy are searched. All of these second unprocessed historical data are collected, while retaining the original first detection label to which each historical defect detection belonged.

[0050] In this embodiment, based on the second historical detection data to be processed by the second defect detection strategy, the second detection accuracy value of each second defect detection strategy is calculated using the calculation formula of the detection accuracy value.

[0051] In this embodiment, the second detection accuracy values ​​of all second defect detection strategies corresponding to the second target detection label can be sorted from largest to smallest. A predetermined number of second defect detection strategies corresponding to the top-ranked second detection accuracy values ​​are selected. These selected second defect detection strategies are then integrated to form the set of optimal detection methods for the wide-scene corresponding to the second target detection label, which serves as the target defect detection strategy for the second target detection label.

[0052] In this embodiment of the application, the formula for calculating the detection precision value is expressed as follows: ; ; in, These represent the cable segment length and defect detection accuracy in the historical detection data of the m-th historical defect detection in the historical detection data of the k-th defect detection strategy corresponding to the i-th first target detection label or the second target detection label; the k-th defect detection strategy is either the first defect detection strategy or the second defect detection strategy; the historical detection data to be processed is either the first historical detection data to be processed or the second historical detection data to be processed. This represents the error allowable factor of the historical detection data of the mth historical defect detection in the historical detection data to be processed in the historical detection data of the kth defect detection strategy corresponding to the ith first target detection label or the second target detection label; N1 represents the precise location, type, and severity values ​​of the m-th historical defect detection in the historical detection data to be processed for the k-th defect detection strategy corresponding to the i-th first target detection label or the k-th target detection label; N1 represents the number of actual severity levels. This represents the number of historical detection data to be processed for the k-th defect detection strategy corresponding to the i-th first target detection label or the second target detection label; These represent the location, type, and severity level of the detected defect in the historical data of the m-th historical defect detection in the historical data of the k-th defect detection strategy corresponding to the i-th first target detection label or the k-th target detection label, respectively. These represent the actual defect location, actual defect type, and actual severity level in the historical detection data of the m-th historical defect detection in the historical detection data of the k-th defect detection strategy corresponding to the i-th first target detection label or the second target detection label.

[0053] This represents the weight of the nth first detection label when the i-th first target detection label or the second target detection label is the second target detection label, after merging. This represents the priority factor for merging the nth first detection label when the i-th first target detection label or the second target detection label is the second target detection label. Let represent the first detection label of the m-th historical defect detection in the second historical detection data of the k-th second defect detection strategy, where the i-th first target detection label or the second target detection label is the second target detection label; and let represent the number of first detection labels that are merged, where the i-th first target detection label or the second target detection label is the second target detection label. This indicates that the nth first detection label, which is either the i-th first target detection label or the second target detection label, is merged with the second target detection label. This represents the first detection label of the m-th historical defect detection in the second historical detection data to be processed, where the i-th first target detection label or the second target detection label is the second target detection label, and the k-th second defect detection strategy is used. This represents the i-th first object detection label or the second object detection label. This represents the detection accuracy value of the k-th defect detection strategy for the i-th first target detection label or the second target detection label. This represents the detection accuracy sub-value of the k-th first defect detection strategy when the i-th first target detection label or the second target detection label is the first target detection label. This represents the detection accuracy sub-value of the k-th second defect detection strategy when the i-th first target detection label or the second target detection label is the second target detection label. This represents the i-th first target detection label or the second target detection label.

[0054] The first detection label represents all historical defect detections in the second historical detection data of the second defect detection strategy, where the first target detection label or the second target detection label is the second target detection label. The number of the nth first detection labels to be merged is also included.

[0055] In this embodiment, by performing voltage level adjacency merging on the first detection label whose number of historical detection data to be processed is less than the first detection data quantity threshold, a second target detection label with sufficient data volume is generated. Based on the second detection accuracy value of each defect detection strategy under the merged second target detection label, the target defect detection strategy is determined. In scenarios with sparse historical data, by effectively aggregating data from adjacent voltage levels, the sample capacity is expanded to support the reliable calculation of detection accuracy values. This avoids the problem of statistical distortion due to insufficient data volume and ensures the continuity and physical rationality of the merged scenario in terms of voltage levels, thereby improving the robustness and applicability of the detection strategy determination method in scenarios with unbalanced data.

[0056] In some embodiments of this application, the method for determining the defect detection strategy for the cable further includes: Using the unprocessed historical detection data of the first target detection tag, the target defect detection strategy of the first target detection tag, the unprocessed historical detection data of the second target detection tag, and the target defect detection strategy of the second target detection tag, a preset machine learning model is trained to obtain a definite model of the cable defect detection strategy.

[0057] In this embodiment, a model training set can be constructed using the unprocessed historical detection data of the first target detection label, the target defect detection strategy of the first target detection label, the unprocessed historical detection data of the second target detection label, and the target defect detection strategy of the second target detection label. The target defect detection strategy of the first target detection label is the label of the unprocessed historical detection data of the first target detection label. The target defect detection strategy of the second target detection label is the label of the unprocessed historical detection data of the second target detection label.

[0058] In the embodiments of this application, a model training set can be used to train the machine learning model to obtain a definite model for the cable defect detection strategy.

[0059] In this embodiment, a pre-defined machine learning model is trained using historical detection data with first and second target detection labels and their corresponding target defect detection strategies. This yields a model for determining cable defect detection strategies. High-quality labeled data selected based on detection accuracy is used as training samples. This allows the machine learning model to learn the complex mapping relationship between detection accuracy and detection strategies in different scenarios. This provides a model foundation for rapid inference and real-time recommendation of detection strategies in new scenarios, further enhancing the intelligence and generalization ability of detection strategy determination. The model training set provides high-quality data support for training the machine learning model, improving the reliability and adaptability of detection strategies under different defect detection accuracies and voltage levels.

[0060] In some embodiments of this application, the cables targeted by the at least one historical defect detection have different temperature information and / or different humidity information; the method for determining the defect detection strategy for the cable further includes: If the number of unprocessed historical detection data of the first target detection label is not less than the preset second detection data number threshold, then the first target detection label is used as the first processable detection label. If the number of unprocessed historical detection data for the second target detection label is not less than the second detection data number threshold, then the second target detection label is used as the second processable detection label. Combine any of the temperature information and any of the humidity information involved in the historical detection data to be processed of the first processable detection tag, and combine any of the temperature information and any of the humidity information involved in the historical detection data to be processed of the second processable detection tag to obtain at least one second detection tag. Extract the processable historical detection data of any second detection tag from the unprocessed historical detection data of the first processable detection tag and the unprocessed historical detection data of the second processable detection tag, and determine the target defect detection strategy of the second detection tag based on the processable historical detection data of the second detection tag.

[0061] In this embodiment, the cables targeted by at least one historical defect detection may have different temperature information and / or different humidity information. The historical detection data to be processed for the first target detection tag completely covers the actual application data of different defect detection strategies under that tag, and the historical detection data to be processed for the second target detection tag completely covers the actual application data of different defect detection strategies under that tag.

[0062] In this embodiment, the amount of unprocessed historical detection data for each first target detection label or second target detection label can be statistically analyzed. If the amount of unprocessed historical detection data for a first target detection label is not less than a preset second detection data quantity threshold, then the amount of historical detection data corresponding to that label is sufficiently rich and meets the conditions for further subdivision. At this time, the first target detection label is further subdivided and marked, that is, the first target detection label is designated as the first processable detection label. The value of the second detection data quantity threshold can be 200.

[0063] In this embodiment of the application, if the number of unprocessed historical detection data of the second target detection label is not less than the second detection data number threshold, the second target detection label is further subdivided and labeled, and the second target detection label is used as the second processable detection label.

[0064] In this embodiment of the application, the selected first target detection labels and second target detection labels with subdivision markers can be integrated and collected to finally form a subdivision scene set. This set is a refined extension of the scene set of the first detection labels and includes all scene labels that can be split according to more detailed dimensions.

[0065] In this embodiment, temperature and humidity information contained in each historical detection data can be extracted, and these temperature and humidity information can be classified, organized, and summarized. Any temperature and humidity information involved in the historical detection data to be processed of the first processable detection tag can be combined, and any temperature and humidity information involved in the historical detection data to be processed of the second processable detection tag can be combined. Duplicate combinations are removed, and finally multiple uniform and standardized second detection tags are determined. These second detection tags cover all temperature and humidity combinations that have occurred in the historical defect detection process, such as high temperature-dry, low temperature-humid, etc.

[0066] In this embodiment, processable historical detection data of any second detection tag can be extracted from the unprocessed historical detection data of the first processable detection tag and the unprocessed historical detection data of the second processable detection tag. Specifically, each determined second detection tag can be used as a classification basis, and for the unprocessed historical detection data of each first or second processable detection tag in the subdivided scenario set, the historical detection data of all sub-historical defect detections contained therein can be screened and divided one by one, and collected to form the processable historical detection data specific to the second detection tag under the first or second processable detection tag.

[0067] In this embodiment of the application, the target defect detection strategy of the second detection tag can be determined based on the processable historical detection data of the second detection tag.

[0068] In this embodiment, under the condition that the amount of historical detection data to be processed meets the second detection data quantity threshold, temperature and humidity information are further introduced into the first target detection label and the second target detection label to obtain the second detection label, and the corresponding processable historical detection data is extracted to determine the target defect detection strategy. Based on the scenario division of voltage level and detection accuracy, the temperature and humidity environment dimension is further introduced according to the condition of sufficient data to refine the scenario, so that the determination of the detection strategy can take into account the impact of the temperature and humidity difference of the on-site environment on the sensitivity and accuracy of the detection method, thereby improving the fine adaptation capability of the detection strategy under different environmental conditions.

[0069] In some embodiments of this application, the at least one historical defect detection targets cables with different cable attribute information and / or different environmental operating condition information; the step of determining the target defect detection strategy for the second detection tag based on the processable historical detection data of the second detection tag includes: For any of the second detection tags, if the processable historical detection data of the second detection tag is less than the preset third detection data quantity threshold, then the target defect detection strategy of the second detection tag is determined according to the combination process of the second detection tags. If the processable historical detection data of the second detection tag is not less than the preset fourth detection data quantity threshold, then the cable attribute information and environmental condition information of the cable are used as the feature information to be processed of the cable, and the processable historical detection data of the second detection tag is clustered according to at least one feature information to be processed involved in the processable historical detection data of the second detection tag to obtain at least one first processable historical detection data set; the fourth detection data quantity threshold is greater than the third detection data quantity threshold. Based on the at least one first processable historical detection data set, determine the target defect detection strategy for the second detection label; If the processable historical detection data of the second detection tag is less than the fourth detection data quantity threshold and not less than the third detection data quantity threshold, then the processable historical detection data of the second detection tag is input into the determination model of the cable defect detection strategy to obtain the target defect detection strategy of the second detection tag.

[0070] In the embodiments of this application, at least one historical defect detection may target cables with different cable attribute information and / or different environmental operating conditions.

[0071] In this embodiment, for each first target detection tag or any second detection tag under a second target detection tag, if the processable historical detection data of the second detection tag is less than a preset third detection data quantity threshold, then the target defect detection strategy of the second detection tag is determined according to the combination process of the second detection tags. The value of the third detection data quantity threshold can be 50.

[0072] For each first target detection label or any second detection label under a second target detection label, if the processable historical detection data of the second detection label is not less than a preset fourth detection data quantity threshold, then the amount of detailed scene data corresponding to the second detection label is sufficient and has the conditions for further subdivision. At this time, the first mark of the second detection label is determined as further subdivision. This mark clarifies that the detailed scene can be deeply split based on more dimensions. The value of the fourth detection data quantity threshold can be 500.

[0073] Specifically, cable attribute information and environmental condition information are used as the cable's unprocessed feature information. Based on at least one unprocessed feature information involved in the processable historical detection data of the second detection tag, the processable historical detection data of the second detection tag is clustered, grouping data with similar unprocessed feature information into the same group to obtain at least one first processable historical detection data set. The fourth detection data quantity threshold is greater than the third detection data quantity threshold. The historical detection data within the set exhibits high consistency in environmental conditions and cable attributes.

[0074] Then, based on at least one first processable historical detection data set, the target defect detection strategy for the second detection label is determined.

[0075] In this embodiment of the application, if the processable historical detection data of the second detection label is less than the fourth detection data quantity threshold and not less than the third detection data quantity threshold, then the amount of detailed scene data corresponding to the second detection label is at a moderate level, which is sufficient to support the model to perform direct analysis, and does not require complex clustering subdivision processing. At this time, the first label of the second detection label is determined to be calculated.

[0076] The processable historical detection data of the second detection tag is input into the determination model of the cable defect detection strategy to obtain the target defect detection strategy of the second detection tag. Specifically, processable historical detection data corresponding to the calculated second detection tag can be collected for each first target detection tag or each first label of the second target detection tag in the subdivided scenario set. This data contains complete historical detection data of multiple historical defect detections in that scenario. The processable historical detection data corresponding to the second detection tag is input into the determination model of the cable defect detection strategy. Based on the calculation output of the model, the optimal detection method set for the subdivided scenario corresponding to the second detection tag of the first target detection tag or the second target detection tag is determined. This set can accurately match the temperature and humidity conditions of the scenario with the cable detection requirements.

[0077] In this embodiment, by setting a third and a fourth threshold for the amount of historical detection data that can be processed for the second detection tag, the data volume is divided into three levels: insufficient, moderate, and sufficient. Rollback processing, direct model calculation processing, and clustering processing based on cable attribute information and environmental conditions are then performed respectively. This achieves adaptive hierarchical decision-making for determining detection strategies in detailed scenarios: when data is insufficient, rollback to existing results at the previous level ensures reliability; when data is moderate, direct model calculation ensures efficiency; and when data is sufficient, deep clustering mining based on cable attributes and environmental conditions is introduced to improve accuracy. This approach balances the accuracy and stability of strategy determination under different data volume conditions.

[0078] In some embodiments of this application, the process of determining the target defect detection strategy for the second detection tags based on the combination of the second detection tags includes: If the second detection label is obtained based on the first processable detection label, then the target defect detection strategy corresponding to the first processable detection label is used as the target defect detection strategy of the second detection label. If the second detection label is obtained based on the second processable detection label, then the target defect detection strategy corresponding to the second processable detection label is used as the target defect detection strategy of the second detection label.

[0079] In this embodiment of the application, for each first target detection label or any second detection label under the second target detection label, if the processable historical detection data of the second detection label is less than the preset third detection data quantity threshold, it means that the amount of data in the detailed scene corresponding to the second detection label is insufficient to support the selection of a reliable optimal detection method for the detailed scene. At this time, the first mark of the second detection label is determined to be a rollback.

[0080] In this embodiment, all first and second target detection labels in the subdivided scene set are comprehensively reviewed, and second detection labels with each first label set as fallback are selected. For these fallback second detection labels, instead of attempting to mine the optimal detection method for the subdivided scene from insufficient data, the optimal detection method set for the broad scene to which the second detection label belongs is directly called. This verified and well-supported set of optimal broad scene detection methods is directly determined as the optimal detection method set for the subdivided scene corresponding to the fallback second detection label.

[0081] In this embodiment of the application, the optimal set of detection methods for detailed scenes is determined by setting the first label of the first target detection label or the second target detection label in the subdivided scene set as the second detection label of the backtracking. This can avoid the unreliability of detailed scene methods caused by small samples, realize the dynamic connection between wide scenes and detailed scenes, take into account the reliability of detection methods and the integrity of scene coverage, and reduce the impact of data sparsity on the accuracy of detection and recommendation.

[0082] In this embodiment, the target defect detection strategy is determined based on the combination source of the second detection tag. If the second detection tag is obtained based on the first processable detection tag, the target defect detection strategy corresponding to the first processable detection tag is directly reused; if it is obtained based on the second processable detection tag, the target defect detection strategy corresponding to the second processable detection tag is reused. Therefore, when the amount of historical detection data that can be processed in a specific scene is insufficient, the target defect detection strategy already determined in the previous level scene to which the specific scene belongs can be automatically traced back. Strategy inheritance can be completed without additional calculation, ensuring the integrity and reliability of the detection strategy determination in data-sparse scenes and achieving a smooth transition between broad and specific scenes.

[0083] In some embodiments of this application, determining the target defect detection strategy for the second detection tag based on the at least one first processable historical detection data set includes: If the number of processable historical detection data in at least one of the first processable historical detection datasets is not less than the third detection data quantity threshold, then the first processable historical detection datasets are combined into a target processable historical detection dataset. If at least one of the first processable historical detection data sets contains a first processable historical detection data set whose number of processable historical detection data is less than the third detection data quantity threshold, then based on the quantity of the first processable historical detection data set, a reference quantity for the clustering process is determined, and the processable historical detection data of the second detection label is clustered according to the reference quantity to obtain at least one second processable historical detection data set. If the number of processable historical detection data in at least one of the second processable historical detection datasets is not less than the third detection data quantity threshold, then the second processable historical detection datasets are combined into the target processable historical detection dataset. If the number of processable historical detection data in any of the second processable historical detection data sets is less than the third detection data quantity threshold, then the second processable historical detection data set is merged with the second processable historical detection data set that is not less than the third detection data quantity threshold, so that the number of processable historical detection data in each of the second processable historical detection data sets is not less than the third detection data quantity threshold, thus obtaining the target processable historical detection data set. For any of the target processable historical detection data sets, the processable historical detection data in the target processable historical detection data set is input into the determination model of the cable defect detection strategy to obtain the target defect detection strategy for the target processable historical detection data set under the second detection label.

[0084] In this embodiment of the application, if the number of processable historical detection data in at least one first processable historical detection data set is not less than the third detection data quantity threshold, then the first processable historical detection data set is combined into a target processable historical detection data set.

[0085] If at least one first processable historical detection dataset contains a set of first processable historical detection datasets with a number of processable historical detection data points less than the third threshold for the number of detection data points, then based on the number of first processable historical detection datasets, a reference number for clustering is determined, and the processable historical detection data of the second detection label is clustered according to the reference number to obtain at least one second processable historical detection dataset. Specifically, if at least one first processable historical detection dataset contains a set of first processable historical detection datasets with a number of processable historical detection data points less than the third threshold for the number of detection data points, it indicates that the dataset has insufficient data volume to support subsequent accurate analysis. In this case, clustering analysis needs to be performed again, and during the re-clustering, the number of first processable historical detection datasets is reduced by one to form a second processable historical detection dataset.

[0086] If the number of processable historical detection data in at least one second processable historical detection dataset is not less than the third detection data quantity threshold, then the second processable historical detection dataset is combined into the target processable historical detection dataset.

[0087] If the number of processable historical detection data in any second processable historical detection data set is less than the third detection data quantity threshold, then the second processable historical detection data set is merged with any second processable historical detection data set whose number of processable historical detection data is not less than the third detection data quantity threshold. This ensures that the number of processable historical detection data in each second processable historical detection data set is not less than the third detection data quantity threshold, resulting in a target processable historical detection data set. Specifically, each second processable historical detection data set whose data volume is less than the third detection data quantity threshold is merged with the nearest neighbor (i.e., the set with the smallest inter-cluster cluster) among all other second processable historical detection data sets whose data volume is greater than or equal to the third detection data quantity threshold. After merging, a new target processable historical detection data set is formed, ensuring that the number of historical detection data in each merged set is greater than or equal to the third detection data quantity threshold.

[0088] For any target's processable historical inspection data set, the processable historical inspection data in the target's processable historical inspection data set is input into the determination model of the cable defect detection strategy to obtain the target defect detection strategy for the target's processable historical inspection data set under the second inspection label.

[0089] In this embodiment, all temperature and humidity clusters for each first demand voltage tag or each first label of the second demand voltage tag in the subdivided scenario set, designated as a second temperature and humidity tag, are input into the constructed learning model. The learning model performs in-depth analysis on the historical detection sub-data in each temperature and humidity cluster set. Through model processing, it outputs the detection method adaptability analysis results for each temperature and humidity cluster set. Based on these output results, the detection method with the best detection effect in each temperature and humidity cluster set is selected. These optimal detection methods are integrated and summarized to form a set of optimal detection methods for each subdivided scenario, which can accurately match the environmental conditions and cable attribute characteristics of the corresponding temperature and humidity cluster set.

[0090] In this embodiment, a set of optimal detection methods for further subdivided scenarios is determined by identifying multiple temperature and humidity clusters within a subdivided scenario set, where the first label of the first or second demand voltage label is a further subdivided second temperature and humidity label. This allows for deep adaptation of the environment, cable attributes, and detection methods, overcoming the limitations of a single temperature and humidity classification, further improving the accuracy and relevance of detection recommendations in subdivided scenarios, and effectively uncovering the optimal adaptation scheme for subdivided operating conditions.

[0091] In this embodiment, the data volume of the first processable historical detection data set obtained by clustering is checked. Sets with insufficient data volume are processed by reducing the number of clusters and re-clustering or merging with the nearest neighbor set until the data volume of all target processable historical detection data sets meets the third detection data volume threshold. Then, the target defect detection strategy for each set is obtained by inputting it into the cable defect detection strategy determination model. In the process of clustering refinement based on cable attributes and environmental conditions, the iterative clustering number adjustment or nearest neighbor merging mechanism ensures that each data set that finally participates in the model calculation has sufficient sample size to support it. This avoids the problem that excessive clustering leads to sparse data volume in some sets, which affects the reliability of the model output. At the same time, the feature correlation between cluster sets is preserved by nearest neighbor merging, which improves the accuracy and statistical stability of detection strategy determination in deep refinement scenarios.

[0092] Reference Figure 2 This document illustrates a flowchart of the steps involved in determining a defect detection strategy for cables according to an embodiment of this application. Specifically, the method may include the following steps: Step 201: Obtain historical detection data for non-contact cable defect detection, and determine multiple first demand voltage labels, multiple second demand voltage labels, and a first scenario set based on the historical detection data.

[0093] Step 202: Based on the scene detection data of all first demand voltage labels and all second demand voltage labels in the first scene set, construct a learning model to determine the set of wide scene optimal detection methods for each first demand voltage label or second demand voltage label.

[0094] Step 203: Based on historical detection data and the first scene set, determine the subdivided scene set, multiple second temperature and humidity labels, and the temperature and humidity detection data of each temperature and humidity label.

[0095] Step 204: Based on the subdivided scenario set, the temperature and humidity detection data of all temperature and humidity labels, and the set of optimal detection methods for all wide scenarios, determine the recommended detection method data for non-contact cable defect detection.

[0096] Step 205: Collect real-time environmental data and cable attribute data for non-contact cable defect detection, and determine the optimal detection method for real-time detection based on the real-time environmental data, cable attribute data, and recommended detection method data.

[0097] In this application embodiment, addressing the issues that non-contact detection methods in related technologies often employ a single method to address different scenarios, failing to adapt to factors such as voltage level and defect detection accuracy, and lacking effective utilization of scattered historical detection data, as well as a lack of systematic scenario segmentation and data-driven recommendation mechanisms, this application acquires and analyzes historical detection data for non-contact cable defect detection. Based on the voltage level and defect detection accuracy in all historical detection data, the historical detection data is segmented to determine multiple first target detection labels and the historical detection data to be processed for each first target detection label. Furthermore, by merging adjacent voltage level labels for first detection labels with insufficient data volume, second target detection labels and the historical detection data to be processed for each second target detection label are determined. This leads to the construction of a determination model for cable defect detection strategies, identifying the optimal set of detection methods for each first or second target detection label across a wide range of scenarios, i.e., the target defect detection strategy. By systematically dividing historical detection data into scenarios and merging sparse labels, the problem of scattered storage and uneven data volume leading to ineffective utilization of historical detection data is solved. This ensures sufficient data volume in each scenario, enabling quantitative evaluation of the adaptability of detection methods in each scenario. It achieves precise matching between detection methods and voltage levels and defect detection accuracy, improves the scientificity and efficiency of detection strategy selection in a wide range of scenarios, and reduces detection deviations caused by improper strategy adaptation.

[0098] To address the issue that non-contact detection methods in related technologies are not adapted to environmental factors such as temperature and humidity, and are prone to insufficient sensitivity and low positioning accuracy due to improper environmental adaptation, this application, based on historical detection data and a first scene set under a first target detection label and a second target detection label, determines a subdivided scene set, multiple second detection labels, and the processable historical detection data for each second detection label. By introducing a temperature and humidity dimension to further subdivide scenes with sufficient data, it ensures that each subdivided scene has sufficient data, achieves accurate matching between the detection method and temperature and humidity conditions, avoids detection deviations caused by temperature and humidity differences, improves the adaptation accuracy of the detection method under different voltage levels and temperature and humidity conditions, and provides refined scene support for subsequent accurate recommendations.

[0099] To address the lack of data-driven recommendation mechanisms and difficulty in adapting to diverse scenarios in related technologies, which leads to insufficient accuracy in cable defect detection results, this application determines recommended detection strategies for non-contact cable defect detection based on a set of subdivided scenarios, all processable historical detection data for second detection tags, and a set of optimal detection methods for all broad scenarios. It also collects real-time environmental data and cable attribute data for non-contact cable defect detection and determines the optimal detection method for real-time detection based on this data. By forming multi-level recommended data, accurate matching to diverse scenarios is achieved, improving the accuracy of detection strategy recommendations under different voltage levels, environmental conditions, and cable attributes. Multi-level scenario decomposition and data reuse balance accuracy and reliability. Dynamically integrating adaptable method sets for broad and detailed scenarios to form recommended data achieves end-to-end method optimization from historical data to real-time detection, further reducing detection errors caused by inappropriate environmental adaptation and improving the scientific rigor, adaptability, and efficiency of detection methods in complex field environments.

[0100] It should be noted that, for the sake of simplicity, the method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of this application are not limited to the described order of actions, because according to the embodiments of this application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of this application.

[0101] Reference Figure 3 This application also discloses a structural block diagram of a device for determining a defect detection strategy for cables, comprising: The historical inspection data acquisition module 301 is used to acquire historical inspection data of at least one historical defect inspection of the cable; the historical defect inspection involves at least one defect detection accuracy and at least one defect detection strategy; the cable has at least one voltage level; The first detection tag obtaining module 302 is used to combine any of the defect detection accuracy and any of the voltage levels to obtain at least one first detection tag, and to extract the unprocessed historical detection data of any of the first detection tags from the historical detection data. The first target detection label is module 303, which is used to, for any first detection label, if the number of unprocessed historical detection data of the first detection label is not less than a preset first detection data number threshold, then the first detection label is used as the first target detection label. The first pending historical detection data extraction module 304 is used to take at least one defect detection strategy involved in the pending historical detection data of the first target detection tag as the first defect detection strategy, and extract the first pending historical detection data of any of the first defect detection strategies from the historical detection data. The first target defect detection strategy determination module 305 is used to determine a first detection accuracy value of any first defect detection strategy based on the first historical detection data to be processed, and to determine the target defect detection strategy of the first target detection tag based on at least one first defect detection strategy.

[0102] In one optional embodiment of this application, the device for determining the defect detection strategy for the cable further includes: The merging processing module is configured to, for any first detection tag, if the number of unprocessed historical detection data for the first detection tag is less than the first detection data quantity threshold, then take the first detection tag as the unprocessed detection tag, and merge the unprocessed historical detection data of at least one of the unprocessed detection tags according to the voltage level of the unprocessed detection tag to obtain unprocessed historical detection data of at least one second target detection tag; the number of unprocessed historical detection data of the second target detection tag is greater than the first detection data quantity threshold. The second pending historical detection data extraction module is used to take at least one defect detection strategy involved in the pending historical detection data of the second target detection tag as the second defect detection strategy, and extract the second pending historical detection data of any second defect detection strategy from the historical detection data. The second target defect detection strategy determination module is used to determine the second detection accuracy value of any second defect detection strategy based on the second historical detection data to be processed, and to determine the target defect detection strategy of the second target detection tag based on the second detection accuracy value of at least one second defect detection strategy.

[0103] In one optional embodiment of this application, the device for determining the defect detection strategy for the cable further includes: The model acquisition module is used to train a preset machine learning model using the unprocessed historical detection data of the first target detection label, the target defect detection strategy of the first target detection label, the unprocessed historical detection data of the second target detection label, and the target defect detection strategy of the second target detection label, so as to obtain a determination model of the cable defect detection strategy.

[0104] In one optional embodiment of this application, the cables targeted by the at least one historical defect detection have different temperature information and / or different humidity information; the device for determining the defect detection strategy for the cable further includes: The first processable detection tag is used as a module to use the first target detection tag as the first processable detection tag if the number of unprocessed historical detection data of the first target detection tag is not less than a preset second detection data number threshold. The second processable detection tag is used as a module to use the second target detection tag as the second processable detection tag if the number of unprocessed historical detection data of the second target detection tag is not less than the second detection data number threshold. The second detection tag obtaining module is used to combine any of the temperature information and any of the humidity information involved in the historical detection data to be processed of the first processable detection tag, and to combine any of the temperature information and any of the humidity information involved in the historical detection data to be processed of the second processable detection tag, to obtain at least one second detection tag. The third target defect detection strategy determination module is used to extract the processable historical detection data of any second detection tag from the unprocessed historical detection data of the first processable detection tag and the unprocessed historical detection data of the second processable detection tag, and determine the target defect detection strategy of the second detection tag based on the processable historical detection data of the second detection tag.

[0105] In one optional embodiment of this application, the at least one historical defect detection targets cables with different cable attribute information and / or different environmental operating condition information; the third target defect detection strategy determination module includes: The first target defect detection strategy determination submodule is used to determine the target defect detection strategy of the second detection tag based on the combination process of the second detection tags if the processable historical detection data of the second detection tag is less than the preset third detection data quantity threshold. The clustering processing submodule is used to, if the processable historical detection data of the second detection tag is not less than a preset fourth detection data quantity threshold, then take the cable attribute information and environmental condition information of the cable as the feature information to be processed of the cable, and perform clustering processing on the processable historical detection data of the second detection tag according to at least one feature information to be processed involved in the processable historical detection data of the second detection tag to obtain at least one first processable historical detection data set; the fourth detection data quantity threshold is greater than the third detection data quantity threshold. The second target defect detection strategy determination submodule is used to determine the target defect detection strategy of the second detection tag based on the at least one first processable historical detection data set. The target defect detection strategy acquisition submodule is used to input the processable historical detection data of the second detection tag into the determination model of the cable defect detection strategy if the processable historical detection data of the second detection tag is less than the fourth detection data quantity threshold and not less than the third detection data quantity threshold, so as to obtain the target defect detection strategy of the second detection tag.

[0106] In one optional embodiment of this application, the first target defect detection strategy determination submodule includes: The first target defect detection strategy is used as a unit to use the target defect detection strategy corresponding to the first processable detection label as the target defect detection strategy of the second detection label if the second detection label is obtained based on the first processable detection label. The second target defect detection strategy is used as a unit to use the target defect detection strategy corresponding to the second processable detection label as the target defect detection strategy of the second detection label if the second detection label is obtained based on the second processable detection label.

[0107] In one optional embodiment of this application, the second target defect detection strategy determination submodule includes: The first target processable historical detection dataset is combined into a unit, which is used to combine the first processable historical detection dataset into a target processable historical detection dataset if the number of processable historical detection data in at least one of the first processable historical detection datasets is not less than the third detection data number threshold. A clustering processing unit is configured to determine a reference number for clustering processing based on the number of the first processable historical detection data sets if at least one of the first processable historical detection data sets contains a first processable historical detection data set whose number of processable historical detection data is less than the third detection data quantity threshold, and to perform clustering processing on the processable historical detection data of the second detection label according to the reference number to obtain at least one second processable historical detection data set. The second target processable historical detection dataset is combined into a unit, which is used to combine the second processable historical detection dataset into the target processable historical detection dataset if the number of processable historical detection data in at least one of the second processable historical detection datasets is not less than the third detection data number threshold. The target processable historical detection data set obtaining unit is used to merge the second processable historical detection data set with the second processable historical detection data set that is not less than the third detection data quantity threshold if the number of processable historical detection data in any second processable historical detection data set is less than the third detection data quantity threshold, so that the number of processable historical detection data in each of the second processable historical detection data sets is not less than the third detection data quantity threshold, thereby obtaining the target processable historical detection data set; The target defect detection strategy obtaining unit is used to input the processable historical detection data in the target processable historical detection data set into the determination model of the cable defect detection strategy for any target processable historical detection data set, so as to obtain the target defect detection strategy for the target processable historical detection data set under the second detection label.

[0108] It is understood that the above-described device embodiments correspond to the method embodiments of this application, and can implement the method provided by any of the above-described method embodiments of this application.

[0109] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.

[0110] It should be noted that the device embodiments described above are merely illustrative, and some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided in this application, the connection relationships between modules indicate that they have communication connections, which can specifically be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0111] Based on the embodiments of the above methods, another embodiment of this application provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the method described in any embodiment of this application.

[0112] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete this application. The one or more module units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the terminal device.

[0113] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0114] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0115] Based on the above-described method embodiments, another embodiment of this application provides a computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the method described in any of the above-described method embodiments of this application.

[0116] The modules / units integrated in the device / terminal equipment, if implemented as software functional units and sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0117] Based on the above-described method embodiments, another embodiment of this application provides a computer program product, including a computer program or instructions, which, when executed by a communication device, implements the method described in any of the above-described method embodiments of this application.

[0118] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above descriptions are merely specific embodiments of this application and are not intended to limit the scope of protection of this application. In particular, it should be noted that any modifications, equivalent substitutions, or improvements made by those skilled in the art within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for determining a defect detection strategy for cables, characterized in that, include: Acquire historical inspection data of at least one historical defect detection of the cable; the historical defect detection involves at least one defect detection accuracy and at least one defect detection strategy; The cable has at least one voltage rating; By combining any of the aforementioned defect detection accuracy and any of the aforementioned voltage levels, at least one first detection tag is obtained, and the unprocessed historical detection data of any of the first detection tags is extracted from the historical detection data. For any of the first detection tags, if the number of historical detection data to be processed for the first detection tag is not less than a preset first detection data number threshold, then the first detection tag is used as the first target detection tag. Take at least one defect detection strategy involved in the historical detection data to be processed of the first target detection label as the first defect detection strategy, and extract the first historical detection data to be processed of any of the first defect detection strategies from the historical detection data. Based on the first historical detection data to be processed, a first detection accuracy value of any of the first defect detection strategies is determined, and based on the first detection accuracy value of at least one of the first defect detection strategies, a target defect detection strategy for the first target detection tag is determined.

2. The method for determining the defect detection strategy for cables according to claim 1, characterized in that, The method for determining the defect detection strategy for the cable further includes: For any of the first detection tags, if the number of unprocessed historical detection data for the first detection tag is less than the first detection data quantity threshold, then the first detection tag is taken as the unprocessed detection tag, and the unprocessed historical detection data of at least one of the unprocessed detection tags is merged according to the voltage level of the unprocessed detection tag to obtain the unprocessed historical detection data of at least one second target detection tag; the number of unprocessed historical detection data of the second target detection tag is greater than the first detection data quantity threshold. The first defect detection strategy involved in the historical detection data to be processed of the second target detection label is used as the second defect detection strategy, and the second historical detection data to be processed of any second defect detection strategy is extracted from the historical detection data. Based on the second historical detection data to be processed, a second detection accuracy value for any of the second defect detection strategies is determined, and based on the second detection accuracy value of at least one of the second defect detection strategies, a target defect detection strategy for the second target detection label is determined.

3. The method for determining the defect detection strategy for cables according to claim 2, characterized in that, The method for determining the defect detection strategy for the cable further includes: Using the unprocessed historical detection data of the first target detection tag, the target defect detection strategy of the first target detection tag, the unprocessed historical detection data of the second target detection tag, and the target defect detection strategy of the second target detection tag, a preset machine learning model is trained to obtain a definite model of the cable defect detection strategy.

4. The method for determining the defect detection strategy for cables according to claim 3, characterized in that, The at least one historical defect detection targets cables with different temperature information and / or different humidity information; The method for determining the defect detection strategy for the cable further includes: If the number of unprocessed historical detection data of the first target detection label is not less than the preset second detection data number threshold, then the first target detection label is used as the first processable detection label. If the number of unprocessed historical detection data for the second target detection label is not less than the second detection data number threshold, then the second target detection label is used as the second processable detection label. Combine any of the temperature information and any of the humidity information involved in the historical detection data to be processed of the first processable detection tag, and combine any of the temperature information and any of the humidity information involved in the historical detection data to be processed of the second processable detection tag to obtain at least one second detection tag. Extract the processable historical detection data of any second detection tag from the unprocessed historical detection data of the first processable detection tag and the unprocessed historical detection data of the second processable detection tag, and determine the target defect detection strategy of the second detection tag based on the processable historical detection data of the second detection tag.

5. The method for determining the defect detection strategy for cables according to claim 4, characterized in that, The at least one historical defect detection targets cables with different cable attribute information and / or different environmental operating condition information. The process of determining the target defect detection strategy for the second detection tag based on the processable historical detection data of the second detection tag includes: For any of the second detection tags, if the processable historical detection data of the second detection tag is less than the preset third detection data quantity threshold, then the target defect detection strategy of the second detection tag is determined according to the combination process of the second detection tags. If the processable historical detection data of the second detection tag is not less than the preset fourth detection data quantity threshold, then the cable attribute information and environmental condition information of the cable are used as the feature information to be processed of the cable, and the processable historical detection data of the second detection tag is clustered according to at least one feature information to be processed involved in the processable historical detection data of the second detection tag to obtain at least one first processable historical detection data set; the fourth detection data quantity threshold is greater than the third detection data quantity threshold. Based on the at least one first processable historical detection data set, determine the target defect detection strategy for the second detection label; If the processable historical detection data of the second detection tag is less than the fourth detection data quantity threshold and not less than the third detection data quantity threshold, then the processable historical detection data of the second detection tag is input into the determination model of the cable defect detection strategy to obtain the target defect detection strategy of the second detection tag.

6. The method for determining the defect detection strategy for cables according to claim 5, characterized in that, The process of determining the target defect detection strategy for the second detection tags based on the combination of the second detection tags includes: If the second detection label is obtained based on the first processable detection label, then the target defect detection strategy corresponding to the first processable detection label is used as the target defect detection strategy of the second detection label. If the second detection label is obtained based on the second processable detection label, then the target defect detection strategy corresponding to the second processable detection label is used as the target defect detection strategy of the second detection label.

7. The method for determining the defect detection strategy for cables according to claim 5, characterized in that, The step of determining the target defect detection strategy for the second detection tag based on the at least one first processable historical detection data set includes: If the number of processable historical detection data in at least one of the first processable historical detection datasets is not less than the third detection data quantity threshold, then the first processable historical detection datasets are combined into a target processable historical detection dataset. If at least one of the first processable historical detection data sets contains a first processable historical detection data set whose number of processable historical detection data is less than the third detection data quantity threshold, then based on the quantity of the first processable historical detection data set, a reference quantity for the clustering process is determined, and the processable historical detection data of the second detection label is clustered according to the reference quantity to obtain at least one second processable historical detection data set. If the number of processable historical detection data in at least one of the second processable historical detection datasets is not less than the third detection data quantity threshold, then the second processable historical detection datasets are combined into the target processable historical detection dataset. If the number of processable historical detection data in any of the second processable historical detection data sets is less than the third detection data quantity threshold, then the second processable historical detection data set is merged with the second processable historical detection data set that is not less than the third detection data quantity threshold, so that the number of processable historical detection data in each of the second processable historical detection data sets is not less than the third detection data quantity threshold, thus obtaining the target processable historical detection data set. For any of the target processable historical detection data sets, the processable historical detection data in the target processable historical detection data set is input into the determination model of the cable defect detection strategy to obtain the target defect detection strategy for the target processable historical detection data set under the second detection label.

8. A device for determining a defect detection strategy for cables, characterized in that, include: A historical inspection data acquisition module is used to acquire historical inspection data of at least one historical defect inspection of the cable; the historical defect inspection involves at least one defect detection accuracy and at least one defect detection strategy; the cable has at least one voltage level; The first detection tag obtaining module is used to combine any of the defect detection accuracy and any of the voltage levels to obtain at least one first detection tag, and to extract the unprocessed historical detection data of any of the first detection tags from the historical detection data. The first target detection label is used as a module to, for any first detection label, if the number of unprocessed historical detection data of the first detection label is not less than a preset first detection data number threshold, then the first detection label is used as the first target detection label. The first pending historical detection data extraction module is used to take at least one defect detection strategy involved in the pending historical detection data of the first target detection tag as the first defect detection strategy, and extract the first pending historical detection data of any of the first defect detection strategies from the historical detection data. The first target defect detection strategy determination module is used to determine a first detection accuracy value of any of the first defect detection strategies based on the first historical detection data to be processed, and to determine the target defect detection strategy of the first target detection tag based on the first detection accuracy value of at least one of the first defect detection strategies.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in 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 a processor, it implements the method as described in any one of claims 1 to 7.

Images