Method for confirming identity of catenary parts based on multi-source consistency checking

Abstract

The application discloses a catenary component identity confirmation method based on multi-source consistency verification, and solves the problem of error binding caused by misreading of on-site labels. The method comprises the following steps: S1, reading an identity code, and acquiring space position, mileage, inspection plan, component type and topological relationship information; S2, constructing a candidate identity set by searching from a device account according to the code or state; S3, performing hard constraint screening to exclude conflict candidates to obtain a feasible set; S4, calculating five types of consistency scores of space, mileage, inspection plan, component type and topology; S5, calculating a comprehensive score and sorting to determine a first candidate; and S6, performing hierarchical binding determination according to the first candidate data. The method has the beneficial effect that the on-site reading code is only one of the candidates but not the final binding result, and error binding caused by misreading, damage or abnormal analysis of adjacent labels is avoided.

Description

Technical Field

[0001] This invention relates to a method for identifying contact network components based on multi-source consistency verification. Background Technology

[0002] In the field of overhead contact line operation and maintenance management, with the continuous expansion of the operation scale of high-speed railways and urban rail transit, traditional manual inspection and paper-based record-keeping methods have gradually shifted to digital inspection. Currently, some overhead contact line operation and maintenance systems have begun to use information technology such as QR codes, electronic tags, or radio frequency identification (RFID) tags to manage component information. In specific applications, electronic tags storing unique identification information are usually installed on overhead contact line poles or components. Inspection personnel read the tags using handheld terminals or mobile inspection equipment to obtain the asset information of the inspected equipment. At the same time, the system backend configures an equipment ledger, establishing a correspondence between the tag ID and the detailed information of the overhead contact line pole (such as basic data, component composition, maintenance items, etc.), thereby realizing the electronic entry of equipment information queries and inspection records.

[0003] However, most existing methods remain at the level of equipment identification or ledger information display, primarily addressing the issue of whether component information can be quickly read. They do not adequately consider whether the read results match the current inspection target or whether the inspection record is accurately linked to the corresponding component. First, the identification information read on-site is not necessarily equivalent to the actual inspection target. Influenced by factors such as the working environment and dense equipment layout, misreading of adjacent components or incorrect pole number selection may occur. Existing methods lack a mechanism to comprehensively verify the read results with information such as on-site location, mileage, and inspection plan. Second, inspection data (such as on-site photos, inspection parameters, and defect descriptions) are often archived coarsely according to line sections or pole numbers, failing to achieve precise binding with the unique component identity. Furthermore, when a component is replaced, the new component usually creates a new equipment record. If a clear association is not established with the original component's identity and historical defect records, it can easily lead to a disconnect between the information before and after the replacement. Summary of the Invention

[0004] To address the technical problems in existing technologies, such as insufficient reliability of contact network inspection identity reading results, difficulty in accurately binding inspection data to specific components, and broken records after replacement, this invention provides a contact network component identity verification method based on multi-source consistency verification.

[0005] To achieve the above objectives, the technical solution of the present invention is as follows: A method for identifying overhead contact line components based on multi-source consistency verification is implemented according to the following steps: S1, Read the identification code of the on-site parts to obtain on-site information, which includes at least spatial location information, mileage information, inspection plan information, part type information and topology relationship information; S2, Construct a candidate identity set. Based on the identity code or on-site status information read on-site, retrieve and construct a candidate identity set from the overhead contact line component equipment ledger. The overhead contact line component equipment ledger records the identity code, theoretical installation location, theoretical line mileage, component type, topology relationship and service status of each component. S3, perform hard constraint screening. For each candidate component in the candidate identity set, screen according to the hard constraints enabled by the current inspection task type and application scenario. Eliminate candidate components with any explicit conflict that has enabled hard constraints from the automatic binding candidate range to obtain a feasible candidate identity set. S4, calculate the multi-dimensional consistency score. For each candidate component that passes the hard constraint screening, calculate the consistency score for spatial location, mileage, inspection plan, component type, and topological relationship: S5, calculate the overall consistency score of each candidate component and rank them by candidate identity to obtain the first candidate identity; S6, a hierarchical binding system, performs hierarchical binding determination based on various data of the first candidate identity.

[0006] The beneficial effects of this invention are as follows: This method uses the identity code read on-site as only one of the candidate identities rather than the final binding result, effectively avoiding incorrect binding caused by misreading adjacent tags, damaged tags, or abnormal code parsing. Through a hard constraint screening mechanism, candidate objects with clear conflicts, such as invalid ledger records, inconsistent track alignments, or significantly excessive positional deviations, are excluded before comprehensive scoring, preventing unreasonable compensation for obvious errors by high scores in other dimensions. By calculating consistency scores for candidate components that pass the hard constraints from five dimensions—spatial location, mileage, inspection plan, component type, and topological relationship—the matching degree of each candidate object with the current inspection scenario is comprehensively evaluated. The introduction of the topological relationship dimension allows for the effective differentiation of components that are close in position but have different structural affiliations, such as adjacent droppers within the same span or different locators under the same cantilever arm. The hierarchical binding judgment mechanism processes automatic binding, manual confirmation, and rejection results in a hierarchical manner, ensuring that data integrity and identity uniqueness are not entered into the formal record. Furthermore, this method allows for not only querying the final binding result during subsequent data tracing but also reconstructing the verification process and judgment basis at the time, providing a complete chain of evidence for data credibility assessment.

[0007] Furthermore, in step S1, the contact network component identification code includes at least the following fields: line number, section number, pole number, span number, component type, installation side, relative installation position, parent component number, and version number, and establishes a unique mapping relationship with the contact network component equipment ledger.

[0008] The beneficial effects of the above technical solution are as follows: The structured design of the identification code enables the code itself to carry the spatial location information and structural subordinate relationships of components in the overhead contact system. Line number, section number, pole number, and span number together constitute the basic spatial positioning of components, providing a benchmark for position consistency verification; the component type field is used to distinguish different equipment objects such as droppers, positioners, and insulators, avoiding confusion between different types of components in specialized maintenance tasks; the installation side and relative installation position fields are used to distinguish multiple similar components in the same pole number or span, for example, multiple droppers arranged sequentially in the same span can be distinguished by the relative installation position field; the parent component number field establishes the installation subordinate relationship between components, allowing the identification process to trace whether the structural affiliation of the current candidate component is correct. Through the unique mapping relationship between the structured code and the equipment ledger, the identification code becomes a bridge connecting the physical entities on site and the background historical data, providing a unified data source for subsequent consistency verification in various dimensions.

[0009] Furthermore, in step S3, the hard constraint screening conditions include at least one of the following: the candidate component has a valid record in the equipment ledger and is in valid service status; the line, section, and track to which the candidate component belongs are within the allowable range of the current inspection task; the positional deviation between the candidate component and the on-site reading position does not exceed the maximum allowable value; the deviation between the on-site mileage and the theoretical mileage of the candidate component does not exceed the maximum allowable value; the candidate component belongs to the target range or allowable buffer range of the current inspection plan; for strict type matching tasks, the type of the candidate component is consistent with the task target type or belongs to an allowable substitution type; for continuous inspection tasks, the candidate component and the previous formally bound component meet the topological connectivity and inspection direction requirements.

[0010] The beneficial effects of the above technical solution are as follows: Hard constraint screening, as a pre-screening step before comprehensive scoring, effectively reduces the size of the candidate identity set and lowers the computational complexity of subsequent comprehensive scoring. It also targets explicit conflicts that cannot be compensated for by higher scores in any other dimension. Ledger validity checks ensure that candidate identities correspond to real and currently usable physical entities; track and line checks set the first spatial filtering boundary at the task level; position deviation and mileage deviation are used to double-verify the positions of candidate components from two independent sources—spatial coordinates and mileage—with an extremely low probability of both occurring in the same direction, resulting in high reliability through joint verification; plan scope checks avoid cross-task misreading in scenarios with multiple people working simultaneously or multiple tasks running in parallel; strict type matching ensures that specialized maintenance tasks are not incorrectly matched due to the closer spatial positions of other types of components; topological connectivity and inspection direction checks ensure the logical correctness of the reading order in continuous inspection scenarios. Excluded candidate identities all have clear and irreparable reasons for conflict, ensuring the reliability of the screening results and providing clear exclusion criteria for manual verification.

[0011] Furthermore, in step S4: S41, the spatial position consistency score is determined according to the following rules: it is determined based on the deviation between the theoretical installation position of the candidate component and the on-site reading position in the longitudinal, lateral and height directions. When the deviation in each direction is within the allowable range, the score is calculated by a preset scoring function. S42, the mileage consistency score is determined according to the following rules: it is determined based on the deviation between the on-site mileage and the theoretical mileage of the candidate parts. When the mileage deviation is within the allowable range, the score is calculated by a preset scoring function.

[0012] The beneficial effects of the above technical solution are as follows: Decomposing spatial position deviation into three independent directions—longitudinal, lateral, and height—for calculation effectively distinguishes the impact of different types of positioning errors on identity verification. A relatively lenient allowable deviation range is set in the longitudinal direction to tolerate normal positioning drift. In the lateral direction, in scenarios such as station areas and sections with multiple parallel tracks, lateral deviation means that inspection personnel misread adjacent track labels or that track-level positioning errors occur, and its severity is significantly higher than longitudinal drift; therefore, the allowable lateral range should be more stringent. Height deviation is used to distinguish components at different installation levels in the same location, preventing incorrect matching between the bottom device of the support column and the top structure. Mileage consistency verification, as a data source independent of spatial positioning, belongs to a different measurement system and has its own independent error sources. When spatial positioning drifts due to signal obstruction, mileage information can still maintain a stable longitudinal reference, and vice versa. Joint verification of the two can form cross-validation, effectively identifying abnormal situations such as cross-track misreading, misreading of adjacent components, and long-distance positioning jumps.

[0013] Furthermore, in step S4: S43, the consistency score of the inspection plan is determined according to the following rules: the highest value is taken when the candidate part belongs to the core target range of the current inspection plan, the reduced value is taken when it belongs to the buffer range near the task boundary, and the lowest value is taken when it belongs to neither the core target range nor the buffer range. S44, the component type consistency score is determined according to the following rules: the highest value is taken when the candidate component type is completely consistent with the task target type, a higher reduction value is taken when it belongs to the allowed substitution type, a lower reduction value is taken when it belongs to the task allowed association type, and the lowest value is taken when there is no valid correspondence.

[0014] The beneficial effects of the above technical solution are as follows: The inspection plan consistency scoring, by distinguishing between the core target range and the buffer range, ensures that components within the planned range receive high matching scores while preserving the possibility of matching for components near the task boundary. This avoids the unreasonable exclusion of components at the boundary due to minor positional deviations caused by excessively narrow task boundaries. In actual inspections, components located near the task area boundary are spatially close to the planned boundary. By setting a buffer range and using a reduced value rather than the lowest value in the scoring, both the accuracy and fault tolerance of boundary identification are considered. The component type consistency scoring refines the degree of type matching into three levels: completely consistent, allowable substitution, and related types. In strict type matching tasks, only completely consistent or allowable substitution types have the basic conditions to participate in subsequent scoring; in general type matching tasks, the same superior type category or related types allowed by the task can still obtain lower reduced values, giving them the opportunity to achieve identity confirmation through high scores in other dimensions. This ensures both the type accuracy requirements in strict scenarios and the matching adaptability in flexible scenarios.

[0015] Furthermore, in step S4: S45, the topology consistency score is determined according to the following rules: a topology model of the contact network components is established, which includes at least the spatial adjacency relationship between component nodes, the parent-child installation subordination relationship, and the sequential relationship along the inspection direction; the topology consistency score is calculated based on the deviation between the actual topology distance and the expected topology distance between the previous formally bound component and the current candidate component in the topology model, combined with the structural relationship matching degree of the candidate component. The degree of structural relationship matching is determined based on the subordinate relationship between the candidate component and the current task target structural unit.

[0016] The beneficial effects of the above technical solution are: topology consistency scoring upgrades the identity verification process from simple physical location matching to structural relationship matching. Contact network components are organized into a network structure with clear adjacency, installation subordination, and inspection sequence relationships according to installation logic and structural hierarchy. These structural relationships are difficult to effectively express and utilize in traditional location or mileage matching methods. By establishing a topology model and incorporating topology distance deviation and structural relationship matching degree into the scoring, the system can make correct judgments based on the differences in topology subordination when two candidate components are physically close. For example, if two locators are installed on opposite sides of the same support column, their spatial positions differ by only a few meters, but they belong to different cantilever structure units and have different installation sides, the topology relationship scoring can effectively distinguish them through the joint judgment of parent-child subordination. During continuous inspections, when inspectors skip a dropper and directly read the next dropper, the topology distance deviation reflects the difference between the actual and expected topology distance, thus indicating possible missed inspections or skipped readings. When the first read does not have a previous officially bound component, the system determines whether the candidate is in the correct starting position based on the starting node specified by the task. If no starting node is specified, the topology data will not participate in the scoring temporarily to avoid incorrect low scores due to lack of topology reference.

[0017] Furthermore, in step S45, the degree of structural relationship matching is determined according to the following rules: the highest value is taken when the structural relationship of the candidate component fully meets the requirements of the current task, the reduced value is taken when it belongs to the pre-allowed association relationship of the task, and the lowest value is taken when there is no effective association between it and the target structure; the expected topology distance is determined according to the current inspection mode and the expected access order. When the candidate component should be the next adjacent object of the previous component, the expected topology distance is set to the adjacent step size.

[0018] The beneficial effects of the above technical solution are: the grading rules for the degree of structural relationship matching further refine the accuracy of topology relationship scoring. Direct installation of subordinate relationships means that candidate components are directly installed on the target structural unit, such as a dropper directly belonging to a cantilever. In this case, the structural relationship is completely consistent with the task requirements and should receive the highest score. Pre-allowed associations appear in comprehensive inspection or cross-structure inspection tasks, such as requiring the inspection of a certain support but allowing the simultaneous inspection of all cantilever, locator, and dropper attached to that support. In this case, although these components are not core targets, they have an allowed structural association with the target support and should receive a reduced value rather than the lowest value. The setting of the expected topology distance is closely related to the inspection mode. In the fine inspection mode of one pole and one component, the expected topology distance is the adjacent step size. In the rapid inspection or sampling inspection mode, the task allows skipping some intermediate nodes, and the expected topology distance increases accordingly. By quantifying the degree of structural relationship matching and the expected topology distance into calculable scoring factors, the topology consistency score can adapt to different types of inspection tasks and different sampling densities.

[0019] Furthermore, in step S5: S51, the comprehensive consistency score is obtained by weighting the five types of individual consistency scores according to the basic importance of each consistency indicator, data reliability, and data availability; S52, the data coverage is determined based on the comprehensive results of the basic importance of various consistency indicators, data reliability, and data availability; when the data coverage is lower than the preset minimum threshold, automatic binding is not performed and the process is switched to manual confirmation.

[0020] The beneficial effects of the above technical solution are as follows: The weighted calculation of the overall consistency score fully considers the differences in importance of each verification dimension in identity verification and the actual quality of the on-site data. Basic importance reflects the relative contribution of each dimension to identity judgment under ideal data conditions; data reliability reflects the credibility of each dimension's data during the collection process. For example, spatial location data has higher reliability when satellite signals are good, but its reliability decreases in signal-blocked areas such as tunnels or stations; mileage data has higher reliability after recent calibration, but its reliability decreases when it has not been calibrated for a long time; data availability indicates whether the data for that dimension is normally included in the calculation, and data that is missing or damaged is not included in the score. Data coverage comprehensively characterizes the sufficiency of information on which the current judgment is based. When all five types of data are valid and have high reliability, the coverage is close to 1; when some data is missing or has low reliability, the coverage decreases accordingly. By setting a minimum data coverage threshold, the system refuses to perform automatic binding under conditions of insufficient information, avoiding unreliable identity verification decisions based solely on a few high-scoring indicators, and reducing the risk of misbinding under incomplete information conditions.

[0021] Furthermore, in step S6: S61, calculate the score difference between the first candidate identity with the highest comprehensive score and the second candidate identity with the second highest comprehensive score, and use this difference as the candidate identity discrimination index; S62, when the overall consistency score of the first candidate identity reaches a preset automatic binding threshold, the data coverage reaches a preset minimum requirement, and the candidate identity distinguishability reaches a preset minimum distinguishability threshold, automatic binding is performed; S63, when the candidate identity distinguishability is less than the preset minimum distinguishability threshold, even if the comprehensive consistency score of the first candidate identity reaches the automatic binding threshold, the process will still be transferred to manual confirmation. S64. When there are no feasible candidate identities or the overall consistency score of the first candidate identity is lower than the preset minimum candidate threshold, the binding is rejected and an abnormal record is generated.

[0022] The beneficial effects of the above technical solution are as follows: The introduction of candidate identity differentiation solves the problem of misjudgment in traditional scoring methods when the scores of candidate objects are close, as the "highest scorer wins" strategy is prone to misjudgment. Multiple similar components arranged adjacently within the same span often have very similar individual scores in dimensions such as spatial location, mileage, planned range, and component type, resulting in a small difference in the overall score. In this case, the advantage of the first candidate over the second candidate is not obvious, and the confidence of automatic binding is insufficient. By calculating the score difference between the first and second candidates as a differentiation index, the system can quantitatively evaluate the uniqueness of the first candidate relative to other similar candidates. When the differentiation is high, the first candidate is significantly better than other candidates in multiple dimensions, the matching advantage is statistically significant, and the confidence of automatic binding is high. When the differentiation is low, multiple candidates are close in overall matching degree, and the information available to the system is insufficient to reliably identify the unique and correct identity. In this case, it is more reliable to entrust the decision-making power to manual confirmation. This mechanism is particularly suitable for the differentiation problem between densely installed and similarly characterized components such as multiple droppers within the same span or multiple positioners on the same support column.

[0023] Furthermore, in step S2: S21, when the identity code format read on site is complete and can be mapped to a valid record in the equipment ledger, the component corresponding to the identity code is taken as the priority candidate identity. S22, when the on-site identity code reading fails, the code is incomplete, or it cannot be mapped to a valid ledger record, the candidate identity that may be corresponding is retrieved from the equipment ledger according to the on-site location, mileage, inspection plan, component type, and topology relationship; components that have been removed, discontinued, or whose identity version has expired in the equipment ledger are not included in the candidate identity set.

[0024] The beneficial effects of the above technical solution are as follows: The branching processing mechanism for constructing the candidate identity set ensures that the system can form a valid candidate identity set under various on-site reading conditions, while filtering out invalid identities. When an identity code is successfully read and can be mapped to the ledger, the corresponding component has the highest matching probability, and using it as a priority candidate identity is in line with operational norms. However, this priority candidate is only a member of the candidate set and not the final binding result. It still needs to undergo subsequent hard constraint screening and consistency scoring, reflecting the design principle of "prioritizing but not blindly trusting". When the identity code reading fails or cannot be mapped, the system does not simply report an error and terminate the operation. Instead, it actively retrieves the target object by making full use of information such as spatial location, mileage, inspection plan, component type, and topology relationship collected on-site. This allows the inspection operation to continue under non-ideal conditions such as tag damage and communication abnormalities, improving the system's on-site adaptability and fault tolerance. Components that have been removed, discontinued, or whose identity versions have expired are excluded during the candidate construction stage to prevent historical data from interfering with the current inspection judgment and to avoid incorrectly classifying the defect records of old components into the history of newly installed components. Attached Figure Description

[0025] Figure 1 This is a flowchart illustrating an embodiment of the present invention; Figure 2 This is a flowchart illustrating the inspection history binding method according to Embodiment 2 of the present invention. Detailed Implementation

[0026] An embodiment of the present invention provides a method for verifying the identity of overhead contact line components based on multi-source consistency verification, which is implemented according to the following steps.

[0027] This invention first constructs a component identification code based on the structural characteristics of the overhead contact system. The identification code of the i-th component is represented as follows: in For line numbering, For interval numbering, The pole number, For span numbering, For component type, For installation on the side, For relative installation positions, Number the parent component. This refers to the version number. The identification code must contain at least the above fields and establish a unique mapping relationship with the overhead contact line component equipment ledger. After generating the identification code, the system writes it into information carriers such as QR codes, electronic tags, or RFID tags. These information carriers are installed in prominent locations on the overhead contact line components so that inspection personnel can scan and read them using mobile terminals. The equipment ledger records the identification code of each component, its theoretical installation location (which can be represented by line coordinates, geographical coordinates, or local track coordinates based on the line centerline), theoretical line mileage, component type, topological relationship, and service status (including in service, out of service, dismantled, or pending confirmation).

[0028] During on-site inspections or maintenance, inspectors use mobile inspection terminals to read QR codes, electronic tags, or RFID tags on components to obtain the original identification code. Simultaneously, the mobile inspection terminal collects on-site status information, including at least the identification code reading time and the current spatial location of the inspection terminal. The system displays the following information: current route mileage, current route, section, track and pole number, current inspection plan or maintenance work order, target component type required by the current task, and the identity of the previously officially bound component. The previously officially bound component identity refers to the component identity that was automatically bound or manually confirmed before this data reading. This identity is used to determine the topological connectivity between the current candidate component and the previous component, and the sequential reading order. For inspection data such as on-site images, continuous video, detection parameters, and defect descriptions, the system establishes a correlation with the identity reading behavior according to a unified time base. The identity reading time, image capture time, video acquisition time, and detection parameter recording time are kept synchronized. When there is a clock deviation between different acquisition devices, time correction is performed before the inspection data is bound.

[0029] The system retrieves and constructs a candidate identity set from the equipment ledger based on the identity codes and on-site status information read on-site. This is divided into two scenarios: S21, when the identity code format read on site is complete and can be mapped to the valid record in the equipment ledger, the system uses the identity code as an index to retrieve the theoretical installation location, theoretical mileage, line section, track, pole number, span, component type, parent structure, service status and historical record of the corresponding component from the equipment ledger, and adds the component corresponding to the identity code as a priority candidate identity to the candidate identity set. S22. When the on-site identification code reading fails, the code is incomplete, or a mapping cannot be established with a valid ledger record, the system does not terminate the inspection process. Instead, it retrieves possible candidate identities from the equipment ledger based on the on-site location, mileage, inspection plan, component type, and topology. Components that have been removed, deactivated, or whose identification version has expired in the equipment ledger are not included in the candidate identity set, but their historical records are retained for subsequent queries.

[0030] The system performs hard constraint screening on each candidate component in the candidate identity set. Hard constraint screening conditions include at least one of the following: the candidate component has a valid record in the equipment ledger and is in valid service status; the line, section, and track to which the candidate component belongs are within the allowable range of the current inspection task; the positional deviation between the candidate component and the field reading location does not exceed the maximum allowable value; the deviation between the field mileage and the theoretical mileage of the candidate component does not exceed the maximum allowable value; the candidate component belongs to the target range or allowable buffer range of the current inspection plan; for strict type matching tasks, the candidate component type is consistent with the task target type or belongs to an allowable substitution type; for continuous inspection tasks, the candidate component and the previously formally bound component meet the topological connectivity and inspection direction requirements. The system activates corresponding hard constraint conditions according to the current inspection task type and application scenario. When a candidate component has any activated explicit conflict, it is excluded from the automatic binding candidate range, obtaining a feasible candidate identity set. Data missing and explicit conflict should be handled differently. When a certain type of data is missing, invalid, or unreliable, the candidate component is not directly identified as conflicting. Instead, the corresponding consistency index is marked as unavailable. When data that is required by the current task is missing, automatic binding is not performed, and the process is switched to manual confirmation.

[0031] For each candidate component that enters the set of feasible candidate identities after passing the hard constraint screening, the system calculates a consistency score for five categories: spatial location, mileage, inspection plan, component type, and topological relationship.

[0032] The spatial position consistency score is calculated as follows. The system transforms the on-site position and the theoretical position of the candidate components to a local track coordinate system based on the track centerline, and calculates the positional deviations along the track direction (longitudinal), lateral, and height directions respectively: The system sets maximum permissible deviations in the longitudinal, lateral, and vertical directions. When the positional deviation in any direction exceeds the corresponding maximum permissible value, it is determined that there is a significant spatial conflict.

[0033] For candidate components that do not exceed the maximum permissible deviation, calculate the spatial position consistency score: in , , These are the positional deviation characteristic scales in the longitudinal, lateral, and height directions, respectively. Lateral deviation typically indicates misreading across tracks, lines, or adjacent overhead contact line structures, and its impact is greater than small-scale positioning errors along the track direction. Therefore, the lateral positional deviation characteristic scale can be set smaller than the longitudinal positional deviation characteristic scale. Mileage consistency scoring is calculated as follows: The system converts the on-site mileage and the theoretical mileage of candidate components to the same track mileage coordinate system and calculates the mileage deviation. When the mileage deviation exceeds the maximum allowable value, it is determined that there is a significant mileage inconsistency.

[0034] For candidate components whose mileage deviation is within the allowable range, calculate the mileage consistency score: in This is the characteristic scale for mileage deviation.

[0035] The consistency score for the inspection plan is determined according to the following rules: 1 is assigned when the candidate component falls within the core target range of the current inspection plan; a preset reduction value is assigned when it falls within the buffer range near the task boundary. ,and If the value is neither within the core target range nor the buffer range, it should be set to 0. The on-site reading time should also be within the operating time allowed by the inspection plan. If it exceeds the allowed range, it should be marked as an abnormal planned time.

[0036] The component type consistency score is determined according to the following rules: 1 is awarded when the candidate component type is completely consistent with the task target type; the higher reduction value is awarded when it belongs to a replacement model, upgrade model, or compatible type explicitly permitted by the work order or equipment technical standards. When inspection tasks belong to the same superior category or are associated with other tasks within the same permitted task category, the lower reduction value shall be applied. ; take 0 when there is no valid correspondence, where For tasks requiring strict type matching, if a candidate component type is neither equal to the task target type nor belongs to an allowed alternative type, a clear type conflict is determined, and it is directly excluded from the automatic binding candidate range.

[0037] The topology consistency score is calculated as follows. The system establishes a topology model of the overhead contact line components: in For a set of component nodes, It is a set of spatial adjacency relationships. Install a set of dependency relationships for the parent-child hierarchy. This is the set of sequential relationships established along the inspection direction. The system calculates the actual topological distance based on the shortest path in the topology model between the previously formally bound component and the current candidate component. Simultaneously, the expected topology distance is determined based on the current inspection mode and the expected access sequence. The deviation between the two is: The system also determines whether the candidate component has established a valid adjacency, installation, or parent-child relationship with the support, cantilever, span, or installation position corresponding to the current task target. For tasks with a specified fixed inspection direction, it also needs to determine whether the topological direction from the previous formally bound component to the current candidate component is consistent with the task's specified direction.

[0038] For candidate components that pass the basic constraints, the topological consistency score is as follows: in This is the topological distance deviation attenuation coefficient. This represents the structural relationship matching coefficient. The degree of structural relationship matching is determined according to the following rules: when the structural relationship of the candidate components fully meets the requirements of the current task... The value is set to 1; a reduction of less than 1 is taken when the relationship belongs to the pre-allowed association of the task; and 0 is taken when there is no valid association with the target structure. For the first identity reading in the inspection task, since there is no previous formally bound component, its topological consistency is judged based on the starting node, current pole number, span, and target structural unit specified by the task. When the task does not specify a starting node and no valid topological basis can be obtained, the topological data is marked as temporarily unavailable and does not participate in the first comprehensive consistency score.

[0039] The system sets a basic importance for each consistency indicator. (The sum of the five basic importance levels is 1), and the reliability of each type of data is assessed based on positioning accuracy, mileage measurement error, inspection plan completeness, component type field validity, and topology database update status. The more reliable the data, the greater its contribution to the overall evaluation result. For data that can be used normally in the calculation, its usability... Availability is denoted as 1 for missing, corrupted, or unparseable data, and 0 for other types of data. The system calculates data coverage as follows: The above formula is used to determine whether the information on which the current consistency evaluation is based is sufficient.

[0040] For each candidate component that enters the set of feasible candidate identities, the system calculates a comprehensive consistency score: in Assign a single consistency score to the candidate component on the k-th category of indicators. Candidate components with clear hard conflicts have been excluded and will no longer participate in the candidate ranking. When the data coverage is lower than the preset minimum threshold, it indicates that the data participating in the consistency judgment is insufficient or the overall reliability is low. The system will not perform automatic binding and will transfer to the manual confirmation process.

[0041] The system sorts all candidate components in the feasible candidate identity set from highest to lowest according to the overall consistency score: The candidate with the highest overall score is selected as the first candidate, and the candidate with the second highest score is selected as the second candidate. The system calculates the score difference between the first and second candidates. As a candidate identity distinguishability, the system automatically binds the current inspection data to the first candidate identity when the overall consistency score of the first candidate identity reaches the preset automatic binding threshold, the data coverage reaches the preset minimum requirement, and the candidate identity distinguishability reaches the preset minimum distinguishability threshold. If the candidate identity distinguishability is less than the preset minimum distinguishability threshold, the system will still proceed to manual confirmation even if the overall consistency score of the first candidate identity reaches the automatic binding threshold. The binding results are categorized into three types: automatic binding, manual confirmation, and binding rejection. During manual confirmation, the system displays the original on-site identity, the first candidate and other candidate identities, the individual consistency scores of each candidate, the overall consistency score, the deviation between the on-site location and the theoretical location, the deviation between the on-site mileage and the theoretical mileage, the candidate component type, pole number, span and topological relationship, and the specific reason for triggering manual confirmation to the inspection personnel or back-end reviewers. Manual confirmation personnel can select the first candidate, select other candidates, reread the identity tag, or mark the current data as unconfirmable. When the set of feasible candidate identities is empty or the overall consistency score of the first candidate is lower than the minimum candidate threshold, the system refuses to bind and generates an abnormal binding record. The abnormal record saves the original reading result, the on-site status, the candidate screening process and the reason for the refusal to bind.

[0042] For continuously acquired video data, the system reads the time using an identity code. Based on this, the corresponding video segments are automatically extracted. ,in and These are the preset capture durations before and after data retrieval. Inspection data, automatically bound or manually confirmed, is written into the corresponding component's official entity history. Inspection, testing, defect, repair, and re-inspection records generated during the component's service life from installation and commissioning to retirement are saved chronologically. Simultaneously, the consistency verification information used for this identity determination is saved, including the original on-site identity, candidate identity set, individual consistency score for each candidate identity, overall consistency score, candidate identity discrimination, data coverage, and the final binding determination result. Inspection data whose identity cannot be determined is temporarily stored in the pending confirmation record area or abnormal record area and is not directly entered into the component's official history.

[0043] The following describes the implementation of the present invention in conjunction with a specific inspection scenario, which also includes the following two steps.

[0044] S7, Maintenance Work Order Closed-Loop Consistency Verification. When a maintenance work order is associated with the current inspection or maintenance operation, the system compares the final on-site identity with the target identity specified in the work order, the on-site installation location (including line, section, track, pole number, span) with the location specified in the work order, the component type with the work order requirements, and verifies whether the on-site operation time is within the work order's allowable range and whether the maintenance result and re-inspection result meet the work order's technical requirements. When all the above comparison conditions are met, the system updates the work order status to closed-loop completion and writes the corresponding maintenance work order, maintenance record, and re-inspection record into the component's entity history. If any comparison condition is not met, the system retains the work order in an unclosed-loop state and generates a consistency anomaly prompt for maintenance or review personnel to handle.

[0045] S8, Component Replacement, Identity Association, and Full Lifecycle History Construction. When a component is determined to need replacement, the system reads the component's identity code and confirms it as the target identity of the corresponding maintenance work order based on the aforementioned consistency check. The installation location identifier remains unchanged before and after replacement, and the system generates an independent identity code for the newly installed component. The system establishes an identity association record between the old and new components, recording the replacement reason, associated maintenance work order, operator, and re-inspection results. The system retains the original component's historical defects and maintenance records in its own entity history and updates its service status to "removed." The new component forms an independent entity history from the time of installation. The system connects the entity histories and replacement association records of all generations of components at the same installation location in chronological order to construct the full lifecycle history of that installation location. After integrity verification, once the original component identity, new component identity, installation location, replacement work order, service time, and re-inspection results all meet the consistency requirements, the system formally establishes the old and new identity association and supports querying and tracing the inspection, defect, maintenance, replacement, and re-inspection processes of components of different generations at the same installation location using either the component identity code or the installation location identifier.

[0046] The inspection terminal reads a dropper identification tag within a certain section of the line. The system executes step S1, obtaining the on-site spatial coordinates, on-site mileage, the current inspection plan requiring the inspection of droppers within the range of supports 120 to 125, the target component type being droppers, and the last formally bound component being the dropper at support 119. The system executes step S2; the identification code format is complete and can be mapped to the equipment ledger. The system adds this identification as a priority candidate to the candidate identification set. Simultaneously, based on the on-site location, mileage, planned range, component type, and topology, the system retrieves other nearby candidate droppers from the ledger, ultimately selecting two candidate components. Both candidates have valid ledger records, and their respective lines, sections, and tracks meet the current task requirements. The system executes step S3, performing hard constraint screening on the two candidates. Neither has exceeded the limits for position deviation, mileage deviation, planned range conflict, type conflict, or topology conflict, and both are added to the feasible candidate identification set. The system executes step S4, calculating the five-category single-item consistency scores for the two candidates. The basic importance of spatial location, mileage, inspection plan, component type, and topology relationship are set to 0.25, 0.20, 0.20, 0.15, and 0.20, respectively, corresponding to data reliability of 0.85, 0.90, 1.00, 0.95, and 0.80. The system executes step S5, calculating the data coverage as follows: This value is higher than the minimum data coverage threshold of 0.75. The scores for the five categories of the first candidate are 0.94, 0.92, 1.00, 1.00, and 0.90, respectively, while the scores for the five categories of the second candidate are 0.82, 0.88, 1.00, 1.00, and 0.72, respectively.

[0047] The overall score of the first candidate: The second candidate's overall score: The system executes step S6 to calculate the candidate identity discrimination score: This value is higher than the preset minimum discrimination threshold of 0.05. The first candidate's comprehensive score of 0.852 is higher than the automatic binding threshold of 0.82, the data coverage of 0.895 is higher than 0.75, and the discrimination of 0.062 is higher than 0.05. All three conditions are met, so the system performs automatic binding, binding the current inspection data to the first candidate identity and writing it into the formal entity history of the hanging string.

[0048] However, if the overall consistency score of the second candidate increases to 0.830, the discrimination between the two will be 0.022. The discrimination score is lower than the preset minimum discrimination score threshold. At this point, even if the overall consistency score of the first candidate reaches the automatic binding threshold, the system will not perform automatic binding, but will instead transfer the current inspection data to the manual confirmation process.

[0049] This embodiment illustrates that, after excluding candidate components with clear conflicts, the present invention uses five types of consistency scores, data coverage, and candidate identity differentiation to jointly determine whether the first candidate has sufficient matching degree and identity uniqueness, thereby reducing misbinding between adjacent or similar components within the same span.

[0050] The system executes step S7 (Maintenance Work Order Closed-Loop Consistency Verification): Assuming this inspection is associated with a maintenance work order that specifies the handling of wear defects in droppers within the range of supports 120 to 125, the system compares the final bound identity on-site with the target identity specified in the work order, the installation location (line, section, track, pole number, span) with the location specified in the work order, the component type with the requirements of the work order, the on-site operation time within the allowable range of the work order, and the repair results and re-inspection results meeting the technical requirements of the work order. After all the above conditions are met, the system updates the work order status to "Closed-Loop Completed" and writes the maintenance work order, maintenance record, and re-inspection record into the entity history of the dropper.

[0051] The system executes step S8 (component replacement, identity association, and full lifecycle history construction): Assuming the dropper at support No. 122 needs replacement after inspection, the system reads the identity code of the dropper to be replaced and confirms it as the target identity of the work order based on the aforementioned consistency check; the installation location identifier remains unchanged before and after replacement, and the system generates an independent identity code for the new dropper. The system establishes an identity association record between the old and new droppers, recording the replacement reason, maintenance work order, operator, and re-inspection results. The system retains the historical defects and maintenance records of the original dropper in its own entity history and updates its status to "removed." The new dropper forms an independent entity history from the time of installation. The system connects the entity histories of all generations of components and the replacement association records in chronological order according to the same installation location to construct the full lifecycle history of that installation location; after integrity verification, once the original component identity, new component identity, installation location, replacement work order, service time, and re-inspection results all meet the requirements, the old and new identity association is formally established, and the system supports querying and tracing the inspection, defect, maintenance, replacement, and re-inspection processes of components of different generations at the same installation location by either component identity code or installation location identifier.

[0052] The above embodiments are merely one preferred embodiment of the present invention. Ordinary variations and substitutions made by those skilled in the art within the scope of the technical solution of the present invention are all included within the protection scope of the present invention.

Claims

1. A method for identifying contact network components based on multi-source consistency verification, implemented according to the following steps: S1, reading the identification code of the field component to obtain field information, which includes at least spatial location information, mileage information, inspection plan information, component type information and topology relationship information; S2. Construct a candidate identity set: Based on the identity code or on-site status information read from the field, retrieve and construct a candidate identity set from the overhead contact line component equipment ledger. This ledger records the identity code, theoretical installation location, theoretical line mileage, component type, topology, and service status of each component. S3. Perform hard constraint screening: For each candidate component in the candidate identity set, screen according to the hard constraints enabled by the current inspection task type and application scenario. Exclude candidate components with any explicit conflict that has enabled hard constraints from the automatically bound candidate range to obtain a feasible candidate identity set. S4. Calculate multi-dimensional consistency scores: For each candidate component that passes the hard constraint screening, calculate the consistency scores for spatial location, mileage, inspection plan, component type, and topology. S5, calculate the overall consistency score of each candidate component and rank them by candidate identity to obtain the first candidate identity; S6, a hierarchical binding system, performs hierarchical binding determination based on various data of the first candidate identity.

2. The method for identifying contact network components based on multi-source consistency verification according to claim 1, characterized in that: In step S1, the contact network component identification code includes at least the following fields: line number, section number, pole number, span number, component type, installation side, relative installation position, parent component number, and version number, and establishes a unique mapping relationship with the contact network component equipment ledger.

3. The method for identifying contact network components based on multi-source consistency verification according to claim 1, characterized in that: In step S3, the hard constraint screening conditions include at least one of the following: the candidate component has a valid record in the equipment ledger and is in valid service status; the line, section, and track to which the candidate component belongs are within the allowable range of the current inspection task; the positional deviation between the candidate component and the on-site reading position does not exceed the maximum allowable value. The deviation between the on-site mileage and the theoretical mileage of the candidate components did not exceed the maximum allowable value; Candidate parts fall within the target range or allowable buffer range of the current inspection plan; for strict type matching tasks, the type of candidate parts is consistent with the type of the task target or belongs to an allowable alternative type. For continuous inspection tasks, the candidate component and the previous formally bound component must meet the requirements of topological connectivity and inspection direction.

4. The method for identifying contact network components based on multi-source consistency verification according to claim 1, characterized in that: In step S4: S41, the spatial position consistency score is determined according to the following rules: it is determined based on the deviation between the theoretical installation position of the candidate component and the on-site reading position in the longitudinal, lateral, and height directions. When the deviation in each direction is within the allowable range, the score is calculated by a preset scoring function; S42, the mileage consistency score is determined according to the following rules: it is determined based on the deviation between the on-site mileage and the theoretical mileage of the candidate component. When the mileage deviation is within the allowable range, the score is calculated by a preset scoring function.

5. The method for identifying contact network components based on multi-source consistency verification according to claim 1, characterized in that: In step S4: S43, the inspection plan consistency score is determined according to the following rules: the highest value is taken when the candidate component belongs to the core target range of the current inspection plan, the reduced value is taken when it belongs to the buffer range near the task boundary, and the lowest value is taken when it belongs to neither the core target range nor the buffer range; S44, the component type consistency score is determined according to the following rules: the highest value is taken when the candidate component type is completely consistent with the task target type, the higher reduced value is taken when it belongs to the allowed substitution type, the lower reduced value is taken when it belongs to the task allowed association type, and the lowest value is taken when there is no effective correspondence.

6. The method for identifying contact network components based on multi-source consistency verification according to claim 1, characterized in that: In step S4: S45, the topology consistency score is determined according to the following rules: A topology model of the contact network components is established, which includes at least the spatial adjacency relationship between component nodes, the parent-child installation subordination relationship, and the sequential relationship along the inspection direction; based on the deviation between the actual topology distance and the expected topology distance between the previous formally bound component and the current candidate component in the topology model, combined with the structural relationship matching degree of the candidate component, the topology consistency score is calculated; wherein, the structural relationship matching degree is determined based on the subordination relationship between the candidate component and the current task target structural unit.

7. The method for identifying contact network components based on multi-source consistency verification according to claim 6, characterized in that: In step S45, the degree of structural relationship matching is determined according to the following rules: the highest value is taken when the structural relationship of the candidate component fully meets the requirements of the current task, the reduced value is taken when it belongs to the pre-allowed association relationship of the task, and the lowest value is taken when there is no effective association between it and the target structure; the expected topology distance is determined according to the current inspection mode and the expected access order. When the candidate component should be the next adjacent object of the previous component, the expected topology distance is set to the adjacent step size.

8. The method for identifying contact network components based on multi-source consistency verification according to claim 1, characterized in that: In step S5: S51, the comprehensive consistency score is obtained by weighting the five types of individual consistency scores according to the basic importance, data reliability, and data availability of various consistency indicators; S52, the data coverage is determined based on the comprehensive result of the basic importance, data reliability, and data availability of various consistency indicators; when the data coverage is lower than the preset minimum threshold, automatic binding is not performed and the process is switched to manual confirmation.

9. The method for identifying contact network components based on multi-source consistency verification according to claim 1, characterized in that: In step S6: S61, calculate the score difference between the first candidate identity with the highest comprehensive score and the second candidate identity with the second highest comprehensive score, and use the difference as the candidate identity discrimination index; S62, when the overall consistency score of the first candidate identity reaches a preset automatic binding threshold, the data coverage reaches a preset minimum requirement, and the candidate identity distinguishability reaches a preset minimum distinguishability threshold, automatic binding is performed; S63, when the candidate identity distinguishability is less than the preset minimum distinguishability threshold, even if the comprehensive consistency score of the first candidate identity reaches the automatic binding threshold, the process will still be transferred to manual confirmation. S64. When there are no feasible candidate identities or the overall consistency score of the first candidate identity is lower than the preset minimum candidate threshold, the binding is rejected and an abnormal record is generated.

10. The method for identifying contact network components based on multi-source consistency verification according to claim 1, characterized in that: In step S2: S21, when the identity code format read on-site is complete and can be mapped to a valid record in the equipment ledger, the component corresponding to the identity code is taken as a priority candidate identity; S22, when the on-site identity code reading fails, the code is incomplete, or cannot be mapped to a valid ledger record, the possible corresponding candidate identities are retrieved from the equipment ledger based on the on-site location, mileage, inspection plan, component type, and topology relationship; components that have been removed, discontinued, or whose identity version has expired in the equipment ledger are not included in the candidate identity set.

Images