Rail state comprehensive evaluation method and device based on multi-source associated data

Abstract

The present application relates to a kind of steel rail state comprehensive evaluation method and device based on multi-source associated data, the method first collects multi-source steel rail detection data, obtains unified line mileage coordinate system under multi-source detection associated data by mileage-time double reference dynamic alignment;According to the dynamic generation evaluation section set of line characteristic parameter and detection abnormal distribution;For each evaluation section, generate a set of multi-dimensional sub-item state index;Based on modal physical correlation rule base, perform multi-modal anomaly consistency check, generate the abnormal consistency correction factor of corresponding section;Finally, the multi-dimensional sub-item index is weighted and fused and is revised by correction factor, obtain the corrected comprehensive state index and divide steel rail state grade.The present application realizes the high-precision alignment of multi-source detection data, the consistency check of abnormal risk and adaptive granularity evaluation, can effectively improve the accuracy and explainability of steel rail state evaluation, suitable for various railway lines steel rail detection, maintenance and repair decision and disease early warning scene.

Description

Technical Field

[0001] This invention belongs to the field of intelligent detection of rail transit, condition evaluation of railway lines and intelligent operation and maintenance technology, and in particular relates to a method and device for comprehensive evaluation of rail condition based on multi-source correlation data. Background Technology

[0002] Rails are critical load-bearing components in railway systems, and their condition directly affects train operation safety, smoothness, and track maintenance costs. With the continuous increase in the operational intensity of high-speed railways, heavy-haul railways, and urban rail transit, the deterioration modes of rail service conditions are becoming increasingly complex, and relying solely on single test data is no longer sufficient to fully reflect the true condition of the rails.

[0003] Currently, the field of rail maintenance has widely adopted technologies such as track geometry inspection, rail surface image inspection, rail wear inspection, vehicle vibration response inspection, and rail flaw detection. However, in practical applications, the above-mentioned inspection data are usually collected independently by different devices, resulting in problems such as different sampling frequencies, different installation locations, different acquisition delays, and different mileage benchmarks. This makes it difficult to accurately correspond between multi-source data at the same location on the line. Summary of the Invention

[0004] To address the problems existing in the prior art, this specification provides a method and apparatus for comprehensive evaluation of rail condition based on multi-source correlation data.

[0005] The specific technical solutions of the embodiments in this specification are as follows: On one hand, embodiments of the present invention provide a comprehensive evaluation method for rail condition based on multi-source correlation data, the method comprising: Collect multi-source rail inspection data of the target line, and perform mileage and time dual-reference alignment on the multi-source rail inspection data to obtain multi-source inspection correlation data under a unified line mileage coordinate system; Based on the line characteristic parameters and the abnormal distribution in the multi-source detection correlation data, an evaluation section set of the target line is generated; For each evaluation segment in the set of evaluation segments, a sub-state index of the corresponding detection mode is generated based on the multi-source detection association data, resulting in a multi-dimensional sub-state index set for each evaluation segment. Based on the pre-built modal physical association rule library, a multimodal anomaly consistency test is performed on the multidimensional sub-item state index set of each evaluation segment to generate an anomaly consistency correction factor corresponding to each evaluation segment. A weighted fusion is performed on the multidimensional sub-item state index set of each evaluation section to obtain the initial comprehensive state index of the corresponding evaluation section. The initial comprehensive state index is then corrected by combining the anomaly consistency correction factor of the corresponding evaluation section to obtain the corrected comprehensive state index. The rail condition level of the corresponding evaluation section is then determined based on the corrected comprehensive state index.

[0006] Furthermore, the multi-source rail detection data is collected by various types of detection equipment mounted on the detection platform. The detection platform is also equipped with an odometer and an inertial measurement unit (IMU). Each detection device, odometer, and IMU corresponds to an independent sensor channel. The multi-source rail inspection data is aligned using both mileage and time references to obtain multi-source inspection correlation data in a unified line mileage coordinate system, including: By performing timestamp calibration on each detection channel using a unified time synchronization benchmark, multi-source detection data with synchronized time can be obtained. A dynamic mileage offset compensation model is constructed based on the mileage values ​​measured by the odometer and the mileage values ​​calculated by the IMU orbit, and mileage offset correction is performed on the time-synchronized multi-source detection data. By using the rotation matrix and translation vector calibrated by the sensor extrinsic parameters, the various detection data after mileage offset correction are mapped to a unified line mileage coordinate system to obtain multi-source detection correlation data.

[0007] Furthermore, the dynamic mileage offset compensation model is as follows:

[0008] in, This indicates the time-synchronized detection data corresponding to a certain type of detection equipment at time [time]. Dynamic mileage compensation amount, This represents the acceleration signal acquired by the IMU. This indicates the mileage measured by the odometer. This represents the mileage calculated based on IMU tracks. This indicates the sensor acquisition delay of this type of detection equipment. Indicates the operating speed of the testing platform. This represents the adaptive coefficient for operating conditions, determined based on line type, operating speed, acquisition frequency, or historical calibration data.

[0009] Furthermore, by mapping the mileage offset-corrected detection data to a unified line mileage coordinate system, the formula for obtaining multi-source detection correlation data is as follows:

[0010] in, This represents the mileage offset corrected detection data in the sensor coordinate system of a certain type of detection equipment. This indicates the detection-related data corresponding to this type of detection equipment in the rail coordinate system or track mileage coordinate system. Represents the rotation matrix. This represents the translation vector.

[0011] Furthermore, the line characteristic parameters include the preset basic evaluation unit length, curve radius, historical defect density, and annual total weight passing through; Based on the line characteristic parameters and the anomaly distribution in the multi-source detection correlation data, a set of evaluation sections for the target line is generated, including: Using the preset basic evaluation unit length as the evaluation unit benchmark, the preset basic evaluation unit length is adjusted according to the curve radius adjustment coefficient, historical disease density adjustment coefficient, annual total weight adjustment coefficient, and current abnormal density adjustment coefficient to obtain the length of each evaluation segment; The evaluation section set is obtained by dividing the target route into a unified mileage based on the length of each evaluation section.

[0012] Furthermore, using the preset basic evaluation unit length as the evaluation unit benchmark, the preset basic evaluation unit length is adjusted according to the curve radius adjustment coefficient, historical disease density adjustment coefficient, annual total weight adjustment coefficient, and current detected abnormal density adjustment coefficient. The formula for obtaining the length of each evaluation segment is as follows:

[0013] in, To evaluate the segment length, To preset the length of the basic evaluation unit, This represents the curve radius adjustment coefficient determined based on the curve radius of the line. This represents the historical defect density adjustment coefficient determined based on the historical defect density of the line. This represents the annual total throughput adjustment coefficient determined based on the annual total throughput of the line. This represents the current anomaly density adjustment coefficient determined based on the anomaly distribution in the multi-source detection correlation data. A higher anomaly density results in a higher adjustment coefficient. The smaller the value, the shorter the corresponding evaluation segment. The shorter.

[0014] Furthermore, the multidimensional sub-item state index set includes track geometry state index, rail surface light stripe state index, rail wear state index, rail top surface scratch state index, vehicle vibration response state index, and rail internal damage state index. The value range of each sub-item state index is 0 to 100. The lower the value, the higher the degree of abnormality of the corresponding detection mode.

[0015] Furthermore, the formula for calculating the orbital geometric state index is as follows:

[0016] in, These represent the abnormal lengths of mild, moderate, and severe cases, respectively. These represent the lengths of minor, moderate, and severe track anomalies, respectively. These represent the lengths of mild, moderate, and severe horizontal abnormalities, respectively. These represent the lengths of minor, moderate, and severe gauge anomalies, respectively. These represent the lengths of the slight, moderate, and severe vertical acceleration anomalies, respectively. These represent the lengths of mild, moderate, and severe lateral acceleration anomalies, respectively. This represents the anomaly weight coefficient corresponding to the anomaly level.

[0017] Furthermore, the track surface light strip state index determines the anomaly level based on the intersection-exchange ratio of the actual light strip area and the standard light strip area, and generates a weighted average of the number of light strip anomalies of different levels within the evaluation section. The rail top surface abrasion status index is generated by dividing the abrasion depth abnormality level based on the allowable speed of the line and by weighting the number of abrasion abnormalities of different levels in the evaluation section. The vehicle vibration response state index is generated by weighting the abnormal lengths of vertical acceleration and lateral acceleration; the rail internal damage state index is generated by weighting the number of damages of different levels within the evaluation section based on the flaw detection B-image morphology and alarm level classification.

[0018] Furthermore, the rail wear state index is calculated separately for curved sections and straight sections. The comprehensive wear quantification value is obtained by weighting the wear area state quantity and the wear excess length state quantity. The number of wear anomalies of different line types and different levels within the evaluation section is generated by weighting. The wear area state quantity adopts different piecewise linear calculation models for curved sections and straight sections.

[0019] Furthermore, based on the pre-built modal physical association rule base, a multimodal anomaly consistency test is performed on the multidimensional sub-item state index set of each evaluation segment to generate anomaly consistency correction factors corresponding to each evaluation segment, including: Anomaly labels for each detection mode are extracted from the multidimensional sub-state index set of each evaluation segment. The anomaly labels include anomaly type, anomaly level, and anomaly mileage range. The modal-physical association rule base pre-sets multiple modal pairs with physical associations and the association strength corresponding to each modal pair. For an evaluation segment, each modal pair in the modal-physical association rule base is traversed, and a consistency index for the modal pair is calculated based on whether the anomaly labels of the two modes in the modal pair overlap within the evaluation segment. ,in i and j These are the indexes of the two modes in the mode pair; The consistency index The rule for determining the value is as follows: if both modes in the mode pair exist within the evaluation segment and the abnormal mileage ranges overlap, then... If two modes are physically related but only one of the modes has an anomalous label within the evaluation segment, then Other cases ; Based on the consistency index of each mode pair within the evaluation segment, the mode association weight corresponding to the association strength, and the combination coefficient corresponding to the anomaly level of the two modes in the mode pair, the anomaly consistency correction factor for the evaluation segment is generated by summing all mode pairs within the evaluation segment. The calculation formula is as follows:

[0020] in, This indicates the anomaly consistency correction factor for the evaluation section. This represents the set of all mode pairs within the evaluation segment that meet the correlation strength requirements. Representing mode pairs Modal association weights corresponding to the association strength Representing mode pairs Consistency indicators Representing mode pairs intermediate mode i and modality j The combination coefficients corresponding to the abnormality level.

[0021] Furthermore, a weighted fusion is performed on the multidimensional sub-item state index set for each evaluation segment to obtain the initial comprehensive state index for the corresponding evaluation segment, including: The initial comprehensive state index is obtained by weighting and summing the state indices of each component in the multidimensional component state index set according to their corresponding fusion weights.

[0022] Furthermore, the initial comprehensive state index is corrected by incorporating the anomaly consistency correction factor for the corresponding evaluation segment, resulting in the formula for the corrected comprehensive state index:

[0023] in, This represents the corrected overall state index. This represents the initial composite state index. This represents the abnormal consistency correction factor.

[0024] Furthermore, the rail condition level of the corresponding evaluation section is obtained based on the corrected comprehensive condition index, including: When the corrected comprehensive state index is greater than or equal to the first threshold, the rail condition level is normal. When the corrected comprehensive condition index is greater than or equal to the second threshold and less than the first threshold, the rail condition level is a state of concern. When the corrected comprehensive state index is greater than or equal to the third threshold and less than the second threshold, the rail condition level is a warning state. When the corrected comprehensive condition index is less than the third threshold, the rail condition level is a risky state.

[0025] On the other hand, embodiments of the present invention also provide a comprehensive rail condition evaluation device based on multi-source correlation data, the device comprising: The multi-source rail inspection data acquisition and processing unit is used to acquire multi-source rail inspection data of the target line, and perform mileage and time dual-reference alignment on the multi-source rail inspection data to obtain multi-source inspection correlation data under a unified line mileage coordinate system. The evaluation section generation unit is used to generate a set of evaluation sections for the target line based on the line characteristic parameters and the abnormal distribution in the multi-source detection correlation data. The sub-item status index calculation unit is used to generate a sub-item status index corresponding to the detection mode for each evaluation segment in the set of evaluation segments based on the multi-source detection association data, so as to obtain a multi-dimensional sub-item status index set corresponding to each evaluation segment. The consistency verification unit is used to perform multimodal anomaly consistency verification on the multidimensional sub-item state index set of each evaluation segment based on the pre-built modal physical association rule base, and generate anomaly consistency correction factor corresponding to each evaluation segment. The correction unit is used to perform weighted fusion on the multi-dimensional sub-item state index set of each evaluation section to obtain the initial comprehensive state index of the corresponding evaluation section. The initial comprehensive state index is corrected by combining the abnormal consistency correction factor of the corresponding evaluation section to obtain the corrected comprehensive state index. The rail condition level of the corresponding evaluation section is obtained according to the corrected comprehensive state index.

[0026] On the other hand, embodiments of this specification also provide a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the above-described method.

[0027] On the other hand, embodiments of this specification also provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0028] On the other hand, embodiments of this specification also provide a computer program product, which includes a computer program that, when executed by a processor, implements the above-described method.

[0029] The methods described in this specification have at least the following technical effects: Compared with the prior art, the present invention has at least the following beneficial effects: 1. Improve the spatial alignment accuracy of multi-source detection data: By unifying time synchronization, FPGA time calibration, odometer correction, IMU track estimation and spatial extrinsic parameter mapping, dynamic alignment of multi-source detection data under unified route mileage coordinates is achieved, reducing spatial offset caused by changes in vehicle operating speed, acquisition delay and installation location differences.

[0030] 2. Improve the reliability of the comprehensive state assessment: Instead of simply linearly fusing the six types of test results, a multimodal anomaly consistency test is performed after generating the sub-state indices, and a correction factor is generated based on the consistency test results, thereby improving the reliability of the comprehensive state index.

[0031] 3. Reduce the impact of false alarms from a single sensor: When an anomaly occurs in a certain detection mode but no corresponding anomaly occurs in other modes that are strongly physically related to it, the system can generate a verification correction result, indicating that there may be detection error, alignment error or single-source false alarm.

[0032] 4. Enhanced ability to identify high-risk sections: When track geometry anomalies, wear anomalies, light band anomalies, vibration anomalies, or internal damage anomalies coexist in the same section, the system can generate a risk-enhancing correction factor, making high-risk sections easier to identify.

[0033] 5. Avoid averaging fixed sections to mask local defects: Through an adaptive evaluation section generation mechanism, finer-grained evaluation units can be used in curved sections, historically densely affected sections, and heavily loaded sections, while longer evaluation units can be used in stable sections, thus balancing evaluation accuracy and computational efficiency.

[0034] 6. Improve the interpretability of operation and maintenance decisions: Through multi-source linkage index charts, six-dimensional status index radar charts, and three-dimensional hierarchical visualization charts, not only is a comprehensive status index output, but also the source of anomalies, alignment relationships, and correction factor sources are displayed, which helps operation and maintenance personnel to quickly locate problems and formulate maintenance plans. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in the embodiments of this specification or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the embodiments of this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0036] Figure 1 The diagram shown is a flowchart illustrating a comprehensive rail condition evaluation method based on multi-source correlation data in an embodiment of this specification. Figure 2 The diagram shown is a schematic diagram of the multi-source detection data synchronous acquisition and dynamic alignment process in an embodiment of the present invention; Figure 3 The diagram shown is a schematic diagram of the installation and coordinate mapping relationship of the multi-source detection device in an embodiment of the present invention; Figure 4 The diagram shown is a flowchart illustrating the process of performing mileage and time dual-reference alignment on the multi-source rail detection data to obtain multi-source detection correlation data under a unified line mileage coordinate system in an embodiment of the present invention. Figure 5 The diagram shown is a schematic representation of the multi-source detection result linkage index and association display in an embodiment of the present invention; Figure 6 The diagram shows a flowchart of generating a set of evaluation sections for a target line based on line characteristic parameters and the abnormal distribution in the multi-source detection associated data in an embodiment of the present invention. Figure 7 The diagram shown is a schematic diagram of adaptive evaluation segment generation in an embodiment of the present invention; Figure 8 The diagram shown is a schematic diagram of a six-dimensional state index radar in an embodiment of the present invention; Figure 9 The diagram shown is a flowchart of the anomaly consistency test and correction factor generation in an embodiment of the present invention. Figure 10 The image shown is a schematic diagram of a full-depth orbital scene in an embodiment of the present invention; Figure 11 The diagram shown is a schematic diagram of the sleeper and fastener area in an embodiment of the present invention; Figure 12 The diagram shown is a schematic representation of the rail area in an embodiment of the present invention. Figure 13The diagram shown is a structural schematic of a rail condition comprehensive evaluation device based on multi-source correlation data in an embodiment of this specification. Figure 14 The diagram shown is a structural schematic of the computer device according to an embodiment of this specification.

[0037] [Explanation of Figure Markers]: 1301. Multi-source rail inspection data acquisition and processing unit; 1302. Evaluation section generation unit; 1303, Calculation unit for sub-item status index; 1304, Consistency Verification Unit; 1305, Correction Unit; 1402. Computer equipment; 1404. Processing equipment; 1406. Storage resources; 1408. Drive mechanism; 1410. Input / Output Module; 1412. Input devices; 1414. Output devices; 1416. Presentation equipment; 1418. Graphical User Interface; 1420. Network interface; 1422. Communication link; 1424. Communication bus. Detailed Implementation

[0038] The technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the embodiments of this specification, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the embodiments of this specification.

[0039] It should be noted that the terms "first," "second," etc., in the description, claims, and accompanying drawings of the embodiments in this specification are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the data can be used interchangeably where appropriate so that the embodiments described herein can be implemented in an order different from that illustrated or described in the specification. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or devices.

[0040] It should be noted that the acquisition, storage, use, and processing of data in the technical solutions of the embodiments of this specification all comply with the relevant provisions of national laws and regulations.

[0041] It should be noted that in the embodiments of this specification, certain software, components, models, and other existing solutions in the industry may be mentioned. These should be considered as exemplary, and their purpose is only to illustrate the feasibility of implementing the technical solution of this application, but does not mean that the applicant has already used or necessarily used such a solution. Similarly, the descriptions of prediction algorithms, table organization forms, and hardware structures are also exemplary.

[0042] The inventors discovered in their research that while some existing technologies can automate the processing of single-type detection data such as internal rail damage, they still have the following shortcomings: 1. Lack of a dynamic compensation mechanism for high-speed operation: Existing solutions typically use a unified timestamp, mileage coding, or simple mileage matching to correlate different detection data, without considering factors such as vehicle speed changes, wheel diameter errors, odometer cumulative errors, IMU drift, sensor installation extrinsic parameter deviations, and acquisition delays. This leads to inconsistencies in the mileage location of the same rail defect across multiple data sources, especially in curved sections and acceleration / deceleration sections where the error is further amplified. 2. Lack of a multimodal anomaly consistency verification mechanism: Most existing technologies use separate identification and quantification of each detection mode, followed by linear fusion or hierarchical summarization to generate a comprehensive conclusion. This fails to adequately determine whether anomalies between different detection modes satisfy physical correlations, and cannot distinguish between real high-risk defects co-occurring in multiple modes and false alarms or alignment errors in single modes, resulting in insufficient reliability of the comprehensive evaluation results. 3. Fixed-granularity evaluation sections are difficult to adapt to track variations: Existing rail condition evaluations are usually based on fixed-length sections, which cannot simultaneously address the fine-grained evaluation needs of small-radius curves, turnout proximity areas, areas with dense defects, and heavy-load sections, as well as the efficiency requirements of stable sections. This can easily lead to local high-risk defects being masked by averaging over long sections. 4. Insufficient integration of visualization with alignment and correction processes: Existing detection systems display multi-source data in a scattered manner, lacking a unified mileage-linked index. They also fail to display dynamic alignment results, anomaly consistency verification processes, correction factor sources, and the generation process of section-level comprehensive indices, resulting in insufficient support for operation and maintenance decision-making.

[0043] The purpose of this invention is to provide a comprehensive evaluation method and system for the multi-source condition of rails based on dynamic alignment and anomaly consistency verification, in order to solve the problems in the prior art such as inaccurate spatial correspondence of multi-source detection data, lack of physical consistency judgment between modes in comprehensive evaluation, local risks being masked due to fixed section granularity, and insufficient interpretability of comprehensive results.

[0044] Figure 1 The diagram shown is a flowchart illustrating a comprehensive rail condition evaluation method based on multi-source correlation data, as described in this specification. The order of steps listed in the embodiment is merely one possible execution order among many and does not represent the only possible order. In actual system or device products, the methods shown in the embodiment or accompanying drawings can be executed sequentially or in parallel. Specifically, as shown... Figure 1 As shown, the method may include: Step 101: Collect multi-source rail inspection data of the target line, and perform mileage and time dual-reference alignment on the multi-source rail inspection data to obtain multi-source inspection correlation data under a unified line mileage coordinate system; Step 102: Generate a set of evaluation sections for the target line based on the line characteristic parameters and the abnormal distribution in the multi-source detection correlation data; Step 103: For each evaluation segment in the set of evaluation segments, generate the corresponding sub-state index of the detection mode based on the multi-source detection association data, and obtain the multi-dimensional sub-state index set corresponding to each evaluation segment; Step 104: Based on the pre-built modal physical association rule library, perform multimodal anomaly consistency test on the multidimensional sub-item state index set of each evaluation segment, and generate anomaly consistency correction factor corresponding to each evaluation segment; Step 105: Perform weighted fusion on the multidimensional sub-item status index set of each evaluation section to obtain the initial comprehensive status index of the corresponding evaluation section. Combine the anomaly consistency correction factor of the corresponding evaluation section to correct the initial comprehensive status index to obtain the corrected comprehensive status index. Based on the corrected comprehensive status index, the rail status level of the corresponding evaluation section is determined.

[0045] This embodiment uses the multi-source detection data output by the intelligent railway operating environment detection system as the starting point for processing and the section-level rail comprehensive condition evaluation result as the processing endpoint. The overall process is as follows: Figure 2 As shown in the figure. The multi-source detection data is collected by various detection devices mounted on the detection platform. The detection platform is also equipped with an odometer and an IMU (Inertial Measurement Unit). Each detection device, odometer, and IMU corresponds to an independent sensor acquisition channel. The installation position of each device on the detection platform and its coordinate system mapping relationship are shown in the figure. Figure 3 As shown.

[0046] First, six types of core inspection data are collected from the target line section using track inspection vehicles, integrated inspection platforms, or ground inspection equipment: 1. Track geometry detection data: including parameters such as elevation, track alignment, horizontality, track gauge, vertical acceleration, and lateral acceleration; 2. Rail surface light strip inspection data: including the light strip area on the top surface of the rail, light strip width, light strip position, light strip depth value, intersection-over-interference ratio of the standard light strip area and the actual light strip area, etc.; 3. Rail wear detection data: including vertical wear, side wear, total wear, wear area, wear distribution on curved sections, wear distribution on straight sections, etc. 4. Rail top surface scratch detection data: including scratch location, scratch depth, scratch length, and depth of wavy wear valleys, etc.; 5. Vehicle vibration response data: including vertical acceleration, lateral acceleration of the vehicle body, and track irregularities reflected by the vehicle response; 6. Rail internal damage detection data: including B-mode images generated by ultrasonic testing, electromagnetic testing or other testing methods, damage alarm level, damage location and damage morphology characteristics.

[0047] Simultaneously, auxiliary positioning data such as IMU inertial measurement data, odometer / wheel speed sensor data, vehicle operating speed, acquisition delay parameters, and sensor extrinsic parameter calibration data are collected.

[0048] Dynamic alignment of the collected multi-source rail inspection data with mileage and time dual benchmarks is performed. Specifically, for example... Figure 4 The diagram illustrates the process of aligning the multi-source rail inspection data with both mileage and time references in an embodiment of the present invention to obtain multi-source inspection correlation data under a unified line mileage coordinate system, including: Step 401: Perform timestamp calibration on each detection channel using a unified timing reference to obtain time-synchronized multi-source detection data; Step 402: Construct a dynamic mileage offset compensation model based on the mileage value measured by the odometer and the mileage value calculated by the IMU track, and perform mileage offset correction on the time-synchronized multi-source detection data; Step 403: Using the rotation matrix and translation vector calibrated by the sensor extrinsic parameters, map the mileage offset corrected detection data to a unified line mileage coordinate system to obtain multi-source detection correlation data.

[0049] In this embodiment of the invention, time synchronization is first performed by providing a unified time reference to each detection device through a unified time synchronization host. The FPGA time calibration module timestamps the acquisition channels of each sensor, enabling the track geometry detection unit, rail surface detection unit, rail wear detection unit, vehicle vibration response detection unit, rail internal damage detection unit, IMU inertial measurement unit, and odometer unit to output raw data under a unified time reference. Then, mileage synchronization is performed by acquiring the raw mileage information of the detection platform through the odometer, wheel speed sensor, or track positioning system, while simultaneously acquiring vehicle running attitude, speed change, and acceleration information using the IMU inertial measurement unit. Finally, dynamic mileage offset compensation is performed by constructing a dynamic mileage offset compensation model using the IMU track estimation value and the odometer measurement value, and then correcting the mileage offset for the multi-source detection data after time synchronization.

[0050] Specifically, the dynamic mileage offset compensation model is as follows:

[0051] in, This indicates the time-synchronized detection data corresponding to a certain type of detection equipment at time [time]. Dynamic mileage compensation amount, This represents the acceleration signal acquired by the IMU. This indicates the mileage measured by the odometer. This represents the mileage calculated based on IMU tracks. This indicates the sensor acquisition delay of this type of detection equipment. Indicates the operating speed of the testing platform. This represents the adaptive coefficient for operating conditions, determined based on line type, operating speed, acquisition frequency, or historical calibration data.

[0052] Then, spatial extrinsic parameter mapping is performed. For the differences in installation positions between the 3D camera, line scan camera, laser measurement module, flaw detection module, IMU, and vehicle coordinate system, spatial mapping is performed using extrinsic parameter calibration parameters. This maps the mileage offset-corrected detection data to a unified route mileage coordinate system. The mapping relationship is as follows:

[0053] in, This represents the mileage offset corrected detection data in the sensor coordinate system of a certain type of detection equipment. This indicates the detection-related data corresponding to this type of detection equipment in the rail coordinate system or track mileage coordinate system. Represents the rotation matrix. This represents the translation vector. Figure 3 In That is, in the mapping relationship , Figure 3 In That is, in the mapping relationship .

[0054] After dynamic mileage compensation and spatial extrinsic parameter mapping, all detection data are transformed into a unified line mileage coordinate system, forming a multi-source detection data association index table. Under the unified mileage coordinate system, various detection waveforms, status indices, and abnormal results can be displayed in a linked manner, such as... Figure 5 As shown, the vertical vernier line at the same mileage position can be used to quickly locate the correspondence between different modal detection results.

[0055] In this embodiment of the invention, instead of using fixed-length evaluation units, a set of evaluation sections for the target line is dynamically generated based on line characteristic parameters and the abnormal distribution in multi-source detection correlation data. The line characteristic parameters include the preset basic evaluation unit length, curve radius, historical defect density, and annual total weight, while also receiving input factors such as the location of special structures such as turnouts, bridges, and tunnels, historical maintenance records, and current detection anomaly density.

[0056] In embodiments of the present invention, such as Figure 6The diagram illustrates the process of generating a set of evaluation sections for a target line based on line characteristic parameters and the anomaly distribution in the multi-source detection correlation data, as described in this embodiment of the invention. The process includes: Step 601: Using the preset basic evaluation unit length as the evaluation unit benchmark, adjust the preset basic evaluation unit length according to the curve radius adjustment coefficient, historical disease density adjustment coefficient, annual total weight adjustment coefficient, and current abnormal density adjustment coefficient to obtain the length of each evaluation segment; Step 602: Divide the evaluation section set along the unified route mileage of the target route according to the length of each evaluation section.

[0057] In this embodiment of the invention, a preset basic evaluation unit length is used as the evaluation unit benchmark. The preset basic evaluation unit length is adjusted according to the curve radius adjustment coefficient, historical disease density adjustment coefficient, annual total weight adjustment coefficient, and current detected abnormal density adjustment coefficient. The formula for obtaining the length of each evaluation segment is as follows:

[0058] in, To evaluate the segment length, To preset the length of the basic evaluation unit, This represents the curve radius adjustment coefficient determined based on the curve radius of the line. This represents the historical defect density adjustment coefficient determined based on the historical defect density of the line. This represents the annual total throughput adjustment coefficient determined based on the annual total throughput of the line. This represents the current anomaly density adjustment coefficient determined based on the anomaly distribution in the multi-source detection correlation data. A higher anomaly density results in a higher adjustment coefficient. The smaller the value, the shorter the corresponding evaluation segment. The shorter.

[0059] Typical adjustment rules are as follows: When a heavy load passes through the total weight section... For a capacity of one million tons per year, a 50m evaluation unit is used; when the historical disease density... When the radius of curvature is 800m / km, a 100m evaluation unit is used; when the curve radius R ≤ 800m, a 200m evaluation unit is used; for straight and stable sections, a 1000m evaluation unit can be used. The logic for dividing adaptive evaluation sections is as follows: Figure 7 As shown, this mechanism can prevent local high-risk diseases from being masked by long-segment averaging, while also taking into account data processing efficiency.

[0060] The evaluation section set is obtained by dividing the target route along a unified mileage based on the length of each evaluation section.

[0061] For each evaluation segment in the evaluation segment set, a sub-item state index for the corresponding detection mode is generated based on multi-source detection correlation data, resulting in a multi-dimensional sub-item state index set for each evaluation segment. Each sub-item state index is represented in a normalized form, with a value range of 0 to 100. The lower the state index, the higher the degree of anomaly of the corresponding mode.

[0062] In this embodiment of the invention, the track geometry state index is used to characterize the degree of irregularity of track geometric parameters within the target evaluation section. Track geometric parameters may include elevation, alignment, horizontality, gauge, vertical acceleration, and lateral acceleration, and can be classified into mild, moderate, and severe anomalies according to their degree of abnormality. The spatial distribution lengths of different levels of anomalies of various geometric parameters within the target evaluation section are statistically analyzed, and the track geometry state index is generated based on anomaly weighting coefficients. :

[0063] in, These represent the abnormal lengths of mild, moderate, and severe cases, respectively. These represent the lengths of minor, moderate, and severe track anomalies, respectively. These represent the lengths of mild, moderate, and severe horizontal abnormalities, respectively. These represent the lengths of minor, moderate, and severe gauge anomalies, respectively. These represent the lengths of the slight, moderate, and severe vertical acceleration anomalies, respectively. These represent the lengths of mild, moderate, and severe lateral acceleration anomalies, respectively. This represents the anomaly weight coefficient corresponding to the anomaly level.

[0064] In this embodiment of the invention, the rail surface light strip state index is used to characterize the light strip state of the rail top surface and the wheel-rail contact quality. This is achieved by calculating the intersection-exchange ratio (IoG) between the actual detected light strip area and the standard light strip area. IOU :

[0065] in, This refers to the actual detected light band region. This is the standard light band region.

[0066] Track surface light band anomalies can be classified as: minor anomalies ( ), moderate abnormality ( ), severe abnormalities ( The number of mild anomalies within the target evaluation segment. Quantity of moderately abnormal Number of severely abnormal Generate the orbital light band state index :

[0067] in, , , These are the weighting coefficients for mild, moderate, and severe light band anomalies, respectively. If the calculated RSLBS value is less than 0, it is normalized to 0.

[0068] The rail wear condition index is used to characterize the comprehensive state of vertical rail wear, lateral wear, and wear area within the target evaluation section. Different weights can be assigned to curved and straight sections respectively. The comprehensive quantitative value of rail wear can be expressed as:

[0069] in, This refers to the wear area state quantity based on the average wear area or the average total wear. This refers to the wear length state quantity based on minor injury, severe injury, or excessive length. , For the corresponding weights.

[0070] For the curved section, the wear area state quantity is as follows:

[0071] For straight sections, the wear area state measurement is as follows:

[0072] The wear-exceeding length state quantity can be expressed as:

[0073] in, This represents the average wear area. This is an abnormal length due to slight wear. The abnormal length for severe wear can be set by staff based on their experience.

[0074] Further statistics were compiled on the number of curves showing slight wear within the target evaluation section. , curve of moderate wear quantity , curve of severe wear quantity Linear light wear quantity Quantity of moderate wear in a straight line Quantity of linear heavy wear Generate rail wear condition index :

[0075] in, , , , , , These are the anomaly weighting coefficients corresponding to different line types and different wear levels. If the calculated RWS value is less than 0, it is normalized to 0.

[0076] The rail top surface abrasion condition index is used to characterize the degree of rail top surface abrasion within the target evaluation section. Anomaly levels are classified according to the permissible speed and abrasion depth: For lines with a permissible speed greater than 120 km / h, a minor anomaly is defined as an abrasion depth greater than 0.5 mm but not greater than 1 mm, and a severe anomaly is defined as an abrasion depth greater than 1 mm; for lines with a permissible speed not greater than 120 km / h, a minor anomaly is defined as an abrasion depth greater than 1 mm but not greater than 2 mm, and a severe anomaly is defined as an abrasion depth greater than 2 mm.

[0077] The number of abnormal minor abrasions within the target evaluation section is counted. and abnormal number of severe abrasions Generate a rail top surface scratch condition index :

[0078] in, The abnormal weighting coefficient for minor abrasions is... This is the abnormal weighting coefficient for severe abrasions. If the calculated RSSS value is less than 0, it is normalized to 0.

[0079] The vehicle vibration response state index is used to characterize the degree of vibration response anomalies caused by track conditions during vehicle operation. Vehicle vibration response data includes vertical acceleration and lateral acceleration. Vertical acceleration anomalies can be classified as: mild anomaly (absolute value greater than 0.1 and not greater than 0.15), moderate anomaly (absolute value greater than 0.15 and not greater than 0.2), and severe anomaly (absolute value greater than 0.2); lateral acceleration anomalies can be classified as: mild anomaly (absolute value greater than 0.06 and not greater than 0.09), moderate anomaly (absolute value greater than 0.09 and not greater than 0.15), and severe anomaly (absolute value greater than 0.15).

[0080] Length of slight vertical anomalies within the statistical target evaluation segment Vertical moderately abnormal length Vertical severe abnormal length Horizontal slight abnormal length Horizontal moderate abnormal length Horizontal severe abnormal length Generate vehicle vibration response state index :

[0081] in, , , , , , These are the anomaly weighting coefficients for different vibration response anomaly levels. If the calculated TVRS value is less than 0, it is normalized to 0.

[0082] The rail internal damage state index is used to characterize the internal damage state of the rail within the target evaluation section. Based on the flaw detection results, the rail internal damage can be classified as follows: Level 1 injury (B-view image does not show a clear injury pattern or trend, or is confirmed as a low-level alarm upon review); Level 2 injury (B-mode image shows a relatively obvious injury pattern, or is confirmed as a medium-level alarm upon review); Level 3 injury (B-screen shows severe injury, or is confirmed as a high-level alarm upon review).

[0083] The number of level 1 injuries within the target evaluation section is counted. Number of second-degree injuries Number of third-degree injuries Generate an index of the internal damage status of the rail. :

[0084] in, , , These are the abnormal injury weighting coefficients for grades one, two, and three, respectively. If the calculated RIDS value is less than 0, it is normalized to 0.

[0085] The aforementioned six-dimensional sub-item state indices can be visually displayed in the form of a radar chart, while simultaneously overlaying the standard state boundary, the warning threshold boundary, and the current segment state boundary, such as... Figure 8 As shown, it can quickly identify the weak modes in the current segment.

[0086] After generating the six-dimensional sub-item state indices, based on the pre-built modal physical association rule base, a multimodal anomaly consistency test is performed on the multi-dimensional sub-item state index set for each evaluation segment, generating an anomaly consistency correction factor for each evaluation segment. The inspection process is as follows: Figure 9 As shown.

[0087] First, anomaly labels for each detection mode are extracted from the multidimensional sub-state index set of each evaluation segment. The anomaly labels include anomaly type, anomaly level, and anomaly mileage range.

[0088] The modal physical association rule base contains multiple pre-defined modal pairs that are physically associated, as well as the association strength corresponding to each modal pair. Typical association rules are shown in Table 1 below: Table 1

[0089] For the current evaluation segment, traverse each associated mode pair in the modal physical association rule base, and calculate the consistency index of the mode pair based on whether the anomaly labels of the two modes in the mode pair overlap within the current evaluation segment. ,in , These are the two mode numbers in the mode pair. The consistency index... The value of is determined as follows: if the abnormal labels of both modes exist in the evaluation area and the abnormal mileage ranges overlap, the value is 1; if the two modes have a strong physical relationship but only one of the modes has an abnormal label in the evaluation area, the value is -1; otherwise, the value is 0.

[0090] Based on the consistency index of each associated mode pair within the current evaluation segment, the mode association weight corresponding to the association strength, and the anomaly level combination coefficient, the anomaly consistency correction factor for the current evaluation segment is generated by summing all associated mode pairs within the current evaluation segment. The calculation formula is as follows:

[0091] in, (Right now Figure 9 In This indicates the anomaly consistency correction factor for the evaluation section. This represents the set of all mode pairs within the evaluation segment that meet the correlation strength requirements. Representing mode pairs Modal association weights corresponding to the association strength Representing mode pairs Consistency indicators Representing mode pairs intermediate mode i and modality j The combination coefficients corresponding to the abnormality level.

[0092] When multiple strongly correlated modes simultaneously exhibit anomalies within the same segment A negative value further reduces the comprehensive state index to enhance the identification of high-risk sections; when the system determines that an anomaly has an alignment error or detection inconsistency, a verification correction result is generated and the corresponding verification flag is output.

[0093] Then, a comprehensive state index is generated and a level is assigned.

[0094] In this embodiment of the invention, a weighted fusion is performed on the multidimensional sub-item state index set of each evaluation segment to obtain the initial comprehensive state index of the corresponding evaluation segment, including: The initial comprehensive state index is obtained by weighting and summing the state indices of each component in the multidimensional component state index set according to their corresponding fusion weights.

[0095] in, The initial composite state index, i.e. Figure 9 In , For the first The fusion weights of the category-specific state indices, For the first The category-specific status index has a total weight of 1. The weights can be dynamically configured based on line type, operating speed, historical fault characteristics, or maintenance strategy.

[0096] The initial comprehensive state index is corrected by incorporating the anomaly consistency correction factor for the corresponding evaluation section, resulting in the corrected comprehensive state index:

[0097] in, (Right now Figure 9 In this context, G represents the corrected composite state index. This is an abnormal consistency correction factor.

[0098] Based on the corrected comprehensive state index The rail condition is divided into multiple levels. Specifically, the rail condition level of the corresponding evaluation section is obtained based on the corrected comprehensive condition index, including: When the corrected comprehensive state index is greater than or equal to the first threshold, the rail condition level is normal. When the corrected comprehensive condition index is greater than or equal to the second threshold and less than the first threshold, the rail condition level is a state of concern. When the corrected comprehensive state index is greater than or equal to the third threshold and less than the second threshold, the rail condition level is a warning state. When the corrected comprehensive condition index is less than the third threshold, the rail condition level is a risky state.

[0099] For example, typical partitioning rules are shown in Table 2 below: Table 2

[0100] The aforementioned thresholds can be configured based on line type, speed rating, maintenance standards, and historical data. Each segment record in the final output will include at least the segment's start and end mileage, comprehensive condition index, condition level, sub-indices for each dimension, source of anomalies, source of correction factors, review recommendations, and maintenance priority.

[0101] The following uses the section from K218+000 to K219+000 of a certain railway line as an example to illustrate the implementation process of this invention: 1. Data Acquisition and Dynamic Alignment: The detection platform operates along the track, simultaneously acquiring data on track geometry, rail surface light bands, rail wear, vehicle vibration, internal rail damage, as well as IMU and odometer data. After unified time synchronization, dynamic mileage compensation, and spatial extrinsic parameter mapping, the system identifies geometric anomalies, wear anomalies, and vibration anomalies at K218+512 that correspond to the same track location, establishing a multi-source anomaly correlation.

[0102] 2. Adaptive section division: The system reads the line parameters of this section and finds that K218+400 to K218+700 is a small radius curve section with a curve radius of 450m and a high density of historical defects. Therefore, the evaluation unit of this section is shortened from the conventional 200m to 100m.

[0103] 3. Generation of six-dimensional state indices: For the section from K218+500 to K218+600, the system generates the following sub-state indices: TGS=68, RSLBS=52, RWS=41, RSSS=77, TVRS=63, RIDS=89.

[0104] 4. Anomaly Consistency Check: Based on the modal physical association rule base, the system determines that: track geometric anomalies and vehicle vibration response anomalies coexist, indicating a strong correlation; rail surface light band anomalies and rail wear anomalies coexist, indicating a strong correlation; the rail internal damage state index is high, showing no serious internal damage, but surface and wear anomalies are obvious, thus generating a review prompt. Based on this, the system generates an anomaly consistency correction factor, correcting the comprehensive state index of this section from the initial value of 61 to 54, and marking it as a risk state.

[0105] 5. Visualization Output: The system outputs a multi-source detection result linkage index chart, a six-dimensional state index radar chart, and a three-dimensional hierarchical detection visualization chart (such as...). Figure 10 , Figure 11 and Figure 12 As shown, the ballast, sleepers, fasteners, and rails can be displayed in layers according to depth. Figure 10 The image shown is a schematic diagram of a full-depth orbital scene in an embodiment of the present invention. Figure 11 The diagram shown is a schematic representation of the sleeper and fastener area in an embodiment of the present invention. Figure 12 The diagram shown is a schematic diagram of the rail area in an embodiment of the present invention. The list of high-risk sections, the explanation of the source of the anomaly, the review and maintenance suggestions directly support maintenance decisions.

[0106] This invention can be integrated into various railway inspection systems as an independent rail condition evaluation module, and is applicable to rail condition inspection and maintenance decision-making for high-speed railways, conventional railways, heavy-haul railways and urban rail transit lines.

[0107] Based on the same inventive concept, embodiments of this specification also provide a comprehensive rail condition evaluation device based on multi-source correlated data, such as... Figure 13 The diagram shown is a structural schematic of a rail condition comprehensive evaluation device based on multi-source correlation data according to an embodiment of this specification. The device includes: The multi-source rail inspection data acquisition and processing unit 1301 is used to acquire multi-source rail inspection data of the target line, perform mileage and time dual-reference alignment on the multi-source rail inspection data, and obtain multi-source inspection correlation data under a unified line mileage coordinate system. Evaluation segment generation unit 1302 is used to generate a set of evaluation segments for the target line based on line characteristic parameters and the abnormal distribution in the multi-source detection correlation data. The sub-item status index calculation unit 1303 is used to generate a sub-item status index corresponding to the detection mode for each evaluation segment in the set of evaluation segments based on the multi-source detection association data, so as to obtain a multi-dimensional sub-item status index set corresponding to each evaluation segment. The consistency verification unit 1304 is used to perform multimodal anomaly consistency verification on the multidimensional sub-item state index set of each evaluation segment based on the pre-built modal physical association rule base, and generate anomaly consistency correction factor corresponding to each evaluation segment. The correction unit 1305 is used to perform weighted fusion on the multi-dimensional sub-item state index set of each evaluation section to obtain the initial comprehensive state index of the corresponding evaluation section, and to correct the initial comprehensive state index by combining the abnormal consistency correction factor of the corresponding evaluation section to obtain the corrected comprehensive state index. The rail condition level of the corresponding evaluation section is obtained according to the corrected comprehensive state index.

[0108] The beneficial effects obtained by the above-described device are the same as those obtained by the above-described method, and will not be described in detail in the embodiments of this specification.

[0109] like Figure 14The diagram illustrates the structure of a computer device according to an embodiment of this specification. The apparatus in this embodiment can be the computer device described above, executing the methods described in the embodiments of this specification. The computer device 1402 may include one or more processing devices 1404, such as one or more central processing units (CPUs), each of which can implement one or more hardware threads. The computer device 1402 may also include any storage resource 1406 for storing information of any kind, such as code, settings, data, etc. Without limitation, for example, the storage resource 1406 may include any one or more combinations of: any type of RAM, any type of ROM, flash memory, hard disk, optical disk, etc. More generally, any storage resource can use any technology to store information. Further, any storage resource can provide volatile or non-volatile retention of information. Further, any storage resource can represent a fixed or removable component of the computer device 1402. In one case, when the processing device 1404 executes associated instructions stored in any storage resource or combination of storage resources, the computer device 1402 can perform any operation of the associated instructions. The computer device 1402 also includes one or more drive mechanisms 1408 for interacting with any storage resource, such as a hard disk drive mechanism, an optical disk drive mechanism, etc.

[0110] Computer device 1402 may also include an input / output module 1410 (I / O) for receiving various inputs (via input device 1412) and providing various outputs (via output device 1414). A specific output mechanism may include a presentation device 1416 and an associated graphical user interface (GUI) 1418. In other embodiments, the input / output module 1410 (I / O), input device 1412, and output device 1414 may be omitted, and the device may function solely as a computer device within a network. Computer device 1402 may also include one or more network interfaces 1420 for exchanging data with other devices via one or more communication links 1422. One or more communication buses 1424 couple the components described above together.

[0111] Communication link 1422 can be implemented in any way, such as via a local area network, a wide area network (e.g., the Internet), a point-to-point connection, or any combination thereof. Communication link 1422 may include any combination of hardwired links, wireless links, routers, gateway functions, name servers, etc., governed by any protocol or combination of protocols.

[0112] This specification also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0113] This specification also provides computer-readable instructions, wherein when a processor executes the instructions, the program therein causes the processor to perform the above-described method.

[0114] It should be understood that in the various embodiments of this specification, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this specification.

[0115] It should also be understood that, in the embodiments of this specification, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Furthermore, in the embodiments of this specification, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0116] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this specification can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of the embodiments in this specification.

[0117] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0118] In the embodiments provided in this specification, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the couplings or direct couplings or communication connections shown or discussed may be indirect couplings or communication connections through some interfaces, devices, or units, or they may be electrical, mechanical, or other forms of connection.

[0119] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments described in this specification, depending on actual needs.

[0120] Furthermore, the functional units in the various embodiments of this specification can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0121] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this specification, in essence, or the parts that contribute to the prior art, or all or part of the technical solutions, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this specification. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0122] This specification describes the principles and implementation methods of the embodiments using specific examples. The above descriptions of the embodiments are only for the purpose of helping to understand the methods and core ideas of the embodiments in this specification. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the embodiments in this specification. Therefore, the content of this specification should not be construed as a limitation on the embodiments in this specification.

Claims

1. A comprehensive evaluation method for rail condition based on multi-source correlation data, characterized in that, The method includes: Collect multi-source rail inspection data of the target line, and perform mileage and time dual-reference alignment on the multi-source rail inspection data to obtain multi-source inspection correlation data under a unified line mileage coordinate system; Based on the line characteristic parameters and the abnormal distribution in the multi-source detection correlation data, an evaluation section set of the target line is generated; For each evaluation segment in the set of evaluation segments, a sub-state index of the corresponding detection mode is generated based on the multi-source detection association data, resulting in a multi-dimensional sub-state index set for each evaluation segment. Based on the pre-built modal physical association rule library, a multimodal anomaly consistency test is performed on the multidimensional sub-item state index set of each evaluation segment to generate an anomaly consistency correction factor corresponding to each evaluation segment. A weighted fusion is performed on the multidimensional sub-item state index set of each evaluation section to obtain the initial comprehensive state index of the corresponding evaluation section. The initial comprehensive state index is then corrected by combining the anomaly consistency correction factor of the corresponding evaluation section to obtain the corrected comprehensive state index. The rail condition level of the corresponding evaluation section is then determined based on the corrected comprehensive state index.

2. The method according to claim 1, characterized in that, The multi-source rail detection data is collected by various types of detection equipment mounted on the detection platform. The detection platform is also equipped with an odometer and an inertial measurement unit (IMU). Each detection device, odometer, and IMU has an independent sensor channel. The multi-source rail inspection data is aligned using both mileage and time references to obtain multi-source inspection correlation data in a unified line mileage coordinate system, including: By performing timestamp calibration on each detection channel using a unified time synchronization benchmark, multi-source detection data with synchronized time can be obtained. A dynamic mileage offset compensation model is constructed based on the mileage values ​​measured by the odometer and the mileage values ​​calculated by the IMU orbit, and mileage offset correction is performed on the time-synchronized multi-source detection data. By using the rotation matrix and translation vector calibrated by the sensor extrinsic parameters, the various detection data after mileage offset correction are mapped to a unified line mileage coordinate system to obtain multi-source detection correlation data.

3. The method according to claim 2, characterized in that, The dynamic mileage offset compensation model is as follows: in, This indicates the time-synchronized detection data corresponding to a certain type of detection equipment at time [time]. Dynamic mileage compensation amount, This represents the acceleration signal acquired by the IMU. This indicates the mileage measured by the odometer. This represents the mileage calculated based on IMU tracks. This indicates the sensor acquisition delay of this type of detection equipment. Indicates the operating speed of the testing platform. This represents the adaptive coefficient for operating conditions, determined based on line type, operating speed, acquisition frequency, or historical calibration data.

4. The method according to claim 2, characterized in that, The formula for mapping the mileage offset-corrected detection data to a unified line mileage coordinate system is as follows: in, This represents the mileage offset corrected detection data in the sensor coordinate system of a certain type of detection equipment. This indicates the detection-related data corresponding to this type of detection equipment in the rail coordinate system or track mileage coordinate system. Represents the rotation matrix. This represents the translation vector.

5. The method according to claim 1, characterized in that, The line characteristic parameters include the preset basic evaluation unit length, curve radius, historical defect density, and annual total weight passing through; Based on the line characteristic parameters and the anomaly distribution in the multi-source detection correlation data, a set of evaluation sections for the target line is generated, including: Using the preset basic evaluation unit length as the evaluation unit benchmark, the preset basic evaluation unit length is adjusted according to the curve radius adjustment coefficient, historical disease density adjustment coefficient, annual total weight adjustment coefficient, and current abnormal density adjustment coefficient to obtain the length of each evaluation segment; The evaluation section set is obtained by dividing the target route into a unified mileage based on the length of each evaluation section.

6. The method according to claim 5, characterized in that, Using the preset basic evaluation unit length as the evaluation unit benchmark, the preset basic evaluation unit length is adjusted according to the curve radius adjustment coefficient, historical disease density adjustment coefficient, annual total weight adjustment coefficient, and current detected abnormal density adjustment coefficient. The formula for obtaining the length of each evaluation segment is as follows: in, To evaluate the segment length, To preset the length of the basic evaluation unit, This represents the curve radius adjustment coefficient determined based on the curve radius of the line. This represents the historical defect density adjustment coefficient determined based on the historical defect density of the line. This represents the annual total throughput adjustment coefficient determined based on the annual total throughput of the line. This represents the current anomaly density adjustment coefficient determined based on the anomaly distribution in the multi-source detection correlation data. A higher anomaly density results in a higher adjustment coefficient. The smaller the value, the shorter the corresponding evaluation segment. The shorter.

7. The method according to claim 1, characterized in that, The multidimensional sub-item state index set includes track geometry state index, rail surface light stripe state index, rail wear state index, rail top surface scratch state index, vehicle vibration response state index, and rail internal damage state index. The value range of each sub-item state index is 0 to 100. The lower the value, the higher the degree of abnormality of the corresponding detection mode.

8. The method according to claim 7, characterized in that, The formula for calculating the orbital geometric state index is as follows: in, These represent the abnormal lengths of mild, moderate, and severe cases, respectively. These represent the lengths of minor, moderate, and severe track anomalies, respectively. These represent the lengths of mild, moderate, and severe horizontal abnormalities, respectively. These represent the lengths of minor, moderate, and severe gauge anomalies, respectively. These represent the lengths of the slight, moderate, and severe vertical acceleration anomalies, respectively. These represent the lengths of mild, moderate, and severe lateral acceleration anomalies, respectively. This represents the anomaly weight coefficient corresponding to the anomaly level.

9. The method according to claim 7, characterized in that, The track surface light strip state index is determined based on the intersection-exchange ratio of the actual light strip area and the standard light strip area to determine the anomaly level, and is generated by weighting the number of light strip anomalies of different levels in the evaluation section. The rail top surface abrasion status index is generated by dividing the abrasion depth abnormality level based on the allowable speed of the line and by weighting the number of abrasion abnormalities of different levels in the evaluation section. The vehicle vibration response state index is generated by weighting the abnormal lengths of vertical acceleration and lateral acceleration; the rail internal damage state index is generated by weighting the number of damages of different levels within the evaluation section based on the flaw detection B-image morphology and alarm level classification.

10. The method according to claim 7, characterized in that, The rail wear state index is calculated separately for curved sections and straight sections. The comprehensive wear quantification value is obtained by weighting the wear area state quantity and the wear excess length state quantity. The number of wear anomalies of different line types and different levels within the evaluation section is generated by weighting. The wear area state quantity adopts different piecewise linear calculation models for curved sections and straight sections.

11. The method according to claim 1, characterized in that, Based on a pre-built modal physical association rule base, a multimodal anomaly consistency test is performed on the multidimensional sub-item state index set of each evaluation segment to generate anomaly consistency correction factors for each evaluation segment, including: Anomaly labels for each detection mode are extracted from the multidimensional sub-state index set of each evaluation segment. The anomaly labels include anomaly type, anomaly level, and anomaly mileage range. The modal-physical association rule base pre-sets multiple modal pairs with physical associations and the association strength corresponding to each modal pair. For an evaluation segment, each modal pair in the modal-physical association rule base is traversed, and a consistency index for the modal pair is calculated based on whether the anomaly labels of the two modes in the modal pair overlap within the evaluation segment. ,in i and j These are the indexes of the two modes in the mode pair; The consistency index The rule for determining the value is as follows: if both modes in the mode pair exist within the evaluation segment and the abnormal mileage ranges overlap, then... If two modes are physically related but only one of the modes has an anomalous label within the evaluation segment, then Other cases ; Based on the consistency index of each mode pair within the evaluation segment, the mode association weight corresponding to the association strength, and the combination coefficient corresponding to the anomaly level of the two modes in the mode pair, the anomaly consistency correction factor for the evaluation segment is generated by summing all mode pairs within the evaluation segment. The calculation formula is as follows: in, This indicates the anomaly consistency correction factor for the evaluation section. This represents the set of all mode pairs within the evaluation segment that meet the correlation strength requirements. Representing mode pairs Modal association weights corresponding to the association strength Representing mode pairs Consistency indicators Representing mode pairs intermediate mode i and modality j The combination coefficients corresponding to the abnormality level.

12. The method according to claim 1, characterized in that, A weighted fusion is performed on the multidimensional sub-item state index set for each evaluation segment to obtain the initial comprehensive state index for the corresponding evaluation segment, including: The initial comprehensive state index is obtained by weighting and summing the state indices of each component in the multidimensional component state index set according to their corresponding fusion weights.

13. The method according to claim 1, characterized in that, The initial comprehensive state index is corrected by incorporating the anomaly consistency correction factor for the corresponding evaluation segment, resulting in the following formula for the corrected comprehensive state index: in, This represents the corrected overall state index. This represents the initial composite state index. This represents the abnormal consistency correction factor.

14. The method according to claim 1, characterized in that, The rail condition level of the corresponding evaluation section is obtained based on the corrected comprehensive condition index, including: When the corrected comprehensive state index is greater than or equal to the first threshold, the rail condition level is normal. When the corrected comprehensive condition index is greater than or equal to the second threshold and less than the first threshold, the rail condition level is a state of concern. When the corrected comprehensive state index is greater than or equal to the third threshold and less than the second threshold, the rail condition level is a warning state. When the corrected comprehensive condition index is less than the third threshold, the rail condition level is a risky state.

15. A comprehensive rail condition evaluation device based on multi-source correlation data, characterized in that, The device includes: The multi-source rail inspection data acquisition and processing unit is used to acquire multi-source rail inspection data of the target line, and perform mileage and time dual-reference alignment on the multi-source rail inspection data to obtain multi-source inspection correlation data under a unified line mileage coordinate system. The evaluation section generation unit is used to generate a set of evaluation sections for the target line based on the line characteristic parameters and the abnormal distribution in the multi-source detection correlation data. The sub-item status index calculation unit is used to generate a sub-item status index corresponding to the detection mode for each evaluation segment in the set of evaluation segments based on the multi-source detection association data, so as to obtain a multi-dimensional sub-item status index set corresponding to each evaluation segment. The consistency verification unit is used to perform multimodal anomaly consistency verification on the multidimensional sub-item state index set of each evaluation segment based on the pre-built modal physical association rule base, and generate anomaly consistency correction factor corresponding to each evaluation segment. The correction unit is used to perform weighted fusion on the multi-dimensional sub-item state index set of each evaluation section to obtain the initial comprehensive state index of the corresponding evaluation section. The initial comprehensive state index is corrected by combining the abnormal consistency correction factor of the corresponding evaluation section to obtain the corrected comprehensive state index. The rail condition level of the corresponding evaluation section is obtained according to the corrected comprehensive state index.

16. A computer device comprising a memory, a processor, and a computer program stored in the memory, characterized in that, When the processor executes the computer program, it implements the method described in any one of claims 1 to 14.

17. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method of any one of claims 1 to 14.

18. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the method of any one of claims 1 to 14.

Images