Contact network-track coordinated comprehensive inspection system and method

Abstract

The present application discloses a contact network-track coordinated comprehensive inspection system and method thereof, which analyzes the time series data of the vehicle body vibration amplitude value by using a data processing and analysis algorithm based on artificial intelligence and deep learning, so as to capture and characterize the local variation characteristics of the vehicle body vibration amplitude in the time series and the vibration mode in the whole time domain as a whole, so as to calculate the vibration compensation parameter results to perform parameter compensation on the measurement parameter values ​​affected by vibration in the measurement system, such as the measurement values ​​of the contact network geometric parameters, and correct the contact network coordinates and geometric parameters through the vibration compensation parameters, so that the measurement results are consistent with the measurement values ​​when the vehicle is stationary. In this way, the data correction based on the vibration influence can be automatically performed on the contact network geometric parameters, thereby providing more accurate measurement data, providing a basis for subsequent data analysis and detection, and helping the railway management department to more effectively manage and maintain rail transit infrastructure.

Description

Technical Field

[0001] The present application relates to the field of track measurement, and more specifically, to a contact network-track coordinated integrated inspection system and method thereof. Background Art

[0002] As key components of the rail transit system, the good working condition of the contact network and track is the basis for ensuring the safe operation of the train. The contact network is responsible for providing power to the train, while the track bears the weight of the train and guides it along the correct path. Any fault in the contact network or track may lead to serious safety accidents, so parameter detection of the contact network and track is extremely important. Through the detection of contact network geometric parameters and track detection, defects and problems of the contact network and track can be discovered in time, such as whether the contact line height, pull-out value, offset, etc. meet the design requirements, as well as track unevenness, rail wear, gauge deviation and other problems.

[0003] At present, the detection of contact network and track parameters is mainly carried out manually by carrying handheld measuring devices to the line to measure the parameters. The detection efficiency is low, and the contact network and track belong to different professional fields and are managed by staff of different professions. Measurement personnel need to be arranged to measure the track and contact network separately. Not only is the process cumbersome, but it also requires a lot of manpower and time costs. Therefore, realizing automated integrated network and track detection is of great significance to improving maintenance efficiency.

[0004] Therefore, a coordinated integrated inspection scheme for contact network and track is desired. Summary of the invention

[0005] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a contact network-track coordinated comprehensive inspection system and method thereof, which collects the time series data of the vehicle body vibration amplitude value through a vibration sensor, and introduces a data processing and analysis algorithm based on artificial intelligence and deep learning at the back end to analyze the time series data of the vehicle body vibration amplitude value, so as to capture and characterize the local change characteristics of the vehicle body vibration amplitude in the time series and the vibration mode of the whole time domain as a whole, so as to calculate the vibration compensation parameter result to perform parameter compensation on the measurement parameter value affected by the vibration in the measurement system, such as the measurement value of the contact network geometric parameter, and correct the contact network coordinates and geometric parameters through the vibration compensation parameter, so that the measurement result is consistent with the measurement value when the vehicle is stationary. In this way, the contact network geometric parameters can be automatically corrected based on the data of the vibration impact, thereby providing more accurate measurement data, providing a basis for subsequent data analysis and detection, and helping the railway management department to more effectively manage and maintain rail transit infrastructure.

[0006] According to one aspect of the present application, a contact network-track coordinated comprehensive inspection method is provided, which includes:

[0007] Calibrate the catenary detection system and the laser profiler system to convert the catenary line scanning camera coordinate system of the catenary detection system and the left and right laser profiler coordinate systems of the laser profiler system into a unified measurement coordinate system, so that the measurement coordinate system coincides with the track center coordinate system when the vehicle is static;

[0008] The vehicle is driven to move, and the measured values ​​of the contact network geometric parameters are obtained, and the time sequence of the vehicle body vibration amplitude values ​​of the vehicle is collected through a vibration sensor, and the vibration compensation parameters are calculated based on the time sequence of the vehicle body vibration amplitude values;

[0009] Correcting the measured values ​​of the contact network geometric parameters using the vibration compensation parameters to obtain corrected measured values ​​of the contact network geometric parameters;

[0010] The measured values ​​of the corrected contact network geometric parameters are displayed in real time on the vehicle control screen.

[0011] According to another aspect of the present application, a contact network-track coordinated comprehensive inspection system is provided, which includes:

[0012] A calibration module, used for calibrating the catenary detection system and the laser profiler system to convert the catenary line scanning camera coordinate system of the catenary detection system and the left and right laser profiler coordinate systems of the laser profiler system into a unified measurement coordinate system, so that the measurement coordinate system of the vehicle when it is static coincides with the track center coordinate system;

[0013] A data acquisition module, used for driving the vehicle to move, acquiring the measured values ​​of the contact network geometric parameters, collecting the time queue of the vehicle body vibration amplitude value through a vibration sensor, and calculating the vibration compensation parameters based on the time queue of the vehicle body vibration amplitude value;

[0014] A correction module, used for correcting the measured value of the contact network geometric parameter by using the vibration compensation parameter to obtain the corrected measured value of the contact network geometric parameter;

[0015] The measurement value visualization module is used to display the measured values ​​of the corrected contact network geometric parameters in real time on the vehicle control screen.

[0016] Compared with the prior art, the present application provides a contact network-track coordinated comprehensive inspection system and method thereof, which collects the time series data of the vehicle body vibration amplitude value through a vibration sensor, and introduces a data processing and analysis algorithm based on artificial intelligence and deep learning at the back end to analyze the time series data of the vehicle body vibration amplitude value, so as to capture and characterize the local change characteristics of the vehicle body vibration amplitude in the time series and the vibration mode of the whole time domain as a whole, so as to calculate the vibration compensation parameter results to perform parameter compensation on the measurement parameter values ​​affected by vibration in the measurement system, such as the measurement values ​​of the contact network geometric parameters, and correct the contact network coordinates and geometric parameters through the vibration compensation parameters, so that the measurement results are consistent with the measurement values ​​when the vehicle is stationary. In this way, the contact network geometric parameters can be automatically corrected based on the data of the vibration impact, thereby providing more accurate measurement data, providing a basis for subsequent data analysis and detection, and helping the railway management department to more effectively manage and maintain rail transit infrastructure. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0018] Figure 1 Flow chart of the contact network-track coordinated comprehensive inspection method according to an embodiment of the present application;

[0019] Figure 2 A schematic diagram of data flow of a contact network-track coordinated comprehensive inspection method according to an embodiment of the present application;

[0020] Figure 3 A block diagram of a network-rail integrated detection system according to an embodiment of the present application;

[0021] Figure 4 A schematic diagram of calibration of a detection system according to an embodiment of the present application;

[0022] Figure 5 is a schematic diagram of a rail center coordinate system according to an embodiment of the present application;

[0023] Figure 6 It is a flowchart of sub-step S2 of the contact network-track coordinated comprehensive inspection method according to an embodiment of the present application;

[0024] Figure 7 is a flow chart of network track parameter detection according to an embodiment of the present application;

[0025] Figure 8Schematic diagram of track horizontal deviation detection according to an embodiment of the present application;

[0026] Fig. 9 A schematic diagram of triangular pit detection according to an embodiment of the present application;

[0027] Fig.10 A schematic diagram of height difference detection according to an embodiment of the present application;

[0028] Fig.11 A schematic diagram of ultra-high detection according to an embodiment of the present application;

[0029] Fig.12 It is a schematic diagram of the conductor height pull-out value detection according to an embodiment of the present application;

[0030] Fig.13 A flow chart of track direction and track wear detection according to an embodiment of the present application;

[0031] Fig.14 is a schematic diagram of track direction detection according to an embodiment of the present application;

[0032] Fig.15 A schematic diagram of rail wear detection according to an embodiment of the present application;

[0033] Fig.16 A block diagram of a contact network-track coordinated comprehensive inspection system according to an embodiment of the present application;

[0034] Fig.17 is a schematic diagram of a vehicle body vibration state according to an embodiment of the present application;

[0035] Fig.18 A flow chart of the calculation of the vehicle body vibration roll angle and track gauge according to an embodiment of the present application;

[0036] Fig.19 A schematic diagram of calculating the vehicle body vibration roll angle and track gauge according to an embodiment of the present application;

[0037] Fig. 20 is a flow chart of vehicle body vibration compensation parameter calculation according to an embodiment of the present application;

[0038] Fig.21 Schematic diagram of vehicle body vibration compensation parameters according to an embodiment of the present application. DETAILED DESCRIPTION

[0039] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0040] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0041] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0042] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. Meanwhile, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0043] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0044] At present, the detection of contact network and track parameters is mainly carried out by manual measurement devices on the line, which has low detection efficiency. In addition, the contact network and track belong to different professional fields and are managed by different professional staff. It is necessary to arrange measurement personnel to go down the track and measure the contact network separately. Not only is the process cumbersome, but it also consumes a lot of manpower and time. Therefore, the realization of automated network-track integrated detection is of great significance to the improvement of maintenance efficiency. Therefore, a contact network-track coordinated comprehensive inspection solution is desired.

[0045] In view of the above-mentioned technical problems, in order to improve the efficiency of network track maintenance, solve the problem of high manpower and time costs for separate network track inspection, realize integrated network track inspection and troubleshooting, and provide guidance for network track maintenance, in the technical solution of this application, a contact network-track collaborative comprehensive inspection method is proposed, which can realize the synchronous measurement of track and contact network parameters. The hardware uses a high-frame rate line scan camera and a laser profiler, which can simultaneously meet the real-time detection needs under different driving speeds, and consider the overall correlation of the detection system. The track measurement parameters are introduced to study the impact of the complex vibration posture of the vehicle body on the measurement system and perform error calibration. It not only has high measurement accuracy, but also has the advantages of simple hardware, lower cost, and a wide range of applications. Ordinary technicians in this field should know that in actual work, track detection parameters and contact network parameters are closely related.

[0046] Any vehicle-mounted automated detection equipment with any parameters is affected by vehicle body vibration during dynamic operation. Vehicle body vibration causes the measurement system to move relative to the contact network and track at the same time, affecting the accuracy of track and contact network detection data. In order to improve the efficiency of network and track maintenance, solve the problem of high manpower and time costs for separate network and track detection, realize integrated network and track maintenance, achieve all-round detection and line hidden danger inspection and fault prediction, provide guidance for network and track maintenance, and improve the accuracy of dynamic automated measurement, this paper invents a vehicle-mounted integrated network and track detection device that can realize synchronous measurement of track and contact network parameters. The hardware uses a high-frame rate line scan camera and a laser profiler, which can simultaneously meet the real-time detection requirements under different vehicle speeds, and consider the overall correlation of the detection system. The track measurement parameters are introduced to study the impact of the complex vibration posture of the vehicle body on the measurement system and perform error calibration. It not only has high measurement accuracy, but also has the advantages of simple hardware, lower cost, and a wide range of applications.

[0047] This application belongs to the field of subway track measurement technology, involving the measurement of contact network geometric parameters, track parameters and vehicle body vibration. Figure 3 As shown, the detection system of the present invention mainly includes a vehicle body posture sensor, a laser profiler system, a contact network detection system, a data processing terminal, a data display and recording terminal, and the detection steps of the present invention are: calibration of the contact network detection system and the laser profiler system, dynamic collection of track point cloud data to calculate the vehicle body composite vibration posture, collection of contact network data and vehicle body posture sensor data, and compensation calculation of the network track measurement parameters according to the vehicle body vibration posture.

[0048] In the technical solution of the present application, a contact network-track coordinated comprehensive inspection method is proposed. Figure 1 It is a flow chart of the contact network-track coordinated comprehensive inspection method according to an embodiment of the present application. Figure 2 Schematic diagram of data flow of the contact network-track coordinated comprehensive inspection method according to an embodiment of the present application. Figure 1 and Figure 2 As shown, the contact network-track coordinated comprehensive inspection method according to the embodiment of the present application includes the steps of: S1, calibrating the contact network detection system and the laser profiler system to convert the contact network line scanning camera coordinate system of the contact network detection system and the left and right laser profiler coordinate systems of the laser profiler system into a unified measurement coordinate system, so that the measurement coordinate system of the vehicle when it is static coincides with the track center coordinate system; S2, driving the vehicle to move, obtaining the measurement values ​​of the contact network geometric parameters, and collecting the time queue of the vehicle body vibration amplitude value through the vibration sensor, and calculating the vibration compensation parameters based on the time queue of the vehicle body vibration amplitude value; S3, correcting the measurement values ​​of the contact network geometric parameters using the vibration compensation parameters to obtain the corrected measurement values ​​of the contact network geometric parameters; S4, displaying the corrected measurement values ​​of the contact network geometric parameters on the vehicle control screen in real time.

[0049] In particular, the S1 calibrates the catenary detection system and the laser profiler system to convert the catenary line scan camera coordinate system of the catenary detection system and the left and right laser profiler coordinate systems of the laser profiler system into a unified measurement coordinate system, so that the measurement coordinate system coincides with the track center coordinate system when the vehicle is static. The catenary detection system is mainly composed of a high frame rate line scan camera and a laser, and the laser profiler system is mainly composed of two left and right laser profilers and a network controller. When the vehicle is stationary and the track gauge is the standard track gauge, the calibration diagram of the detection system is as follows: Figure 4 As shown in the figure, the spatial transformation relationship between the contact network line scanning camera coordinate system and the left and right laser profiler coordinate systems is converted to the same measurement coordinate system through calibration. At this time, the measurement coordinate system coincides with the track center coordinate system. The schematic diagram of the track center coordinate system is shown in Figure 5 shown.

[0050] In particular, the S2 drives the vehicle to move, obtains the measured values ​​of the contact network geometric parameters, and collects the time queue of the vehicle body vibration amplitude value through the vibration sensor, and calculates the vibration compensation parameters based on the time queue of the vehicle body vibration amplitude value. Wherein, the measured values ​​of the contact network geometric parameters include the guide height value, the pull-out value, the height difference, the distance between the two lines, the span and the positioning point. It should be understood that in actual work, the vehicle generates complex vibrations while collecting data during operation. At the same time, the vibration state of the vehicle is a state of superposition of multiple vibrations, that is, multiple vibrations occur simultaneously. The on-board automated detection equipment for track detection parameters and contact network geometric parameters is affected by the vibration of the vehicle body during dynamic operation. At this time, the camera follows the vibration of the vehicle body. The vibration of the vehicle body causes the measurement system to produce relative movement with the contact network and the track at the same time. The measurement value of the geometric parameter detection camera is affected by the vibration of the vehicle body and produces errors, which affects the accuracy of the track and contact network detection data. In particular, in a specific example of the present application, Figure 6 As shown, the S2 includes: S21, sampling the time queue of the vehicle body vibration amplitude value based on a predetermined sampling window and with a preset step size to obtain a set of local time queues of the vehicle body vibration amplitude value; S22, performing feature extraction and feature significant aggregation processing on the set of local time queues of the vehicle body vibration amplitude value to obtain a time series significant aggregation representation of the vehicle body vibration mode; S23, generating the vibration compensation parameter based on the time series significant aggregation representation of the vehicle body vibration mode.

[0051] Specifically, in S21, based on a predetermined sampling window and with a preset step size, the time queue of the vehicle body vibration amplitude value is sampled to obtain a set of local time queues of the vehicle body vibration amplitude value. It should be understood that by sampling the time queue of the vehicle body vibration amplitude value, the main vibration characteristics can be retained, which is helpful for subsequent processing.

[0052] Specifically, the S22 performs feature extraction and feature significant aggregation processing on the set of local time queues of the vehicle body vibration amplitude values ​​to obtain a significant aggregation representation of the vehicle body vibration mode time series. In a specific example of the present application, first, the set of local time queues of the vehicle body vibration amplitude values ​​is time-series encoded to obtain a set of implicit correlation features of the local time series of vehicle body vibration; in an embodiment of the present application, the local time queues of each vehicle body vibration amplitude value in the set of local time queues of the vehicle body vibration amplitude values ​​are respectively input into a vehicle body vibration time series sequence encoder based on a Bi-LSTM model to obtain a set of implicit correlation feature vectors of the local time series of vehicle body vibration as the set of implicit correlation features of the local time series of vehicle body vibration. It should be understood that in order to extract local time series features related to the vehicle body vibration amplitude from the set of local time series of the vehicle body vibration amplitude values ​​to extract features of local time series changes of vibration and global vibration modes, so as to calculate vibration compensation parameters, in the technical solution of the present application, each local time series of the vehicle body vibration amplitude value in the set of local time series of the vehicle body vibration amplitude value is further input into the vehicle body vibration time series sequence encoder based on the Bi-LSTM model for encoding, so as to extract the local time series correlation feature information of the vehicle body vibration amplitude in each local time period, thereby obtaining a set of implicit correlation feature vectors of the local time series of vehicle body vibration. Subsequently, the set of implicit correlation features of the local time series of the vehicle body vibration is subjected to feature significant aggregation processing based on cluster analysis to obtain the significant aggregation representation of the time series of the vehicle body vibration mode. It should be understood that, since the set of implicitly associated features of the local time series of vehicle body vibration contains local time series features related to the vibration amplitude of the vehicle body in each local time period, there is an overall correlation relationship between the local time series features of the vibration amplitude based on the full time domain. At the same time, since different local time series features of the vehicle body vibration amplitude have different importance in the overall vibration full time domain aggregation characterization process, some are important for the subsequent vibration time series pattern feature extraction and vibration compensation parameter calculation tasks, while some have a smaller contribution to the subsequent tasks. Therefore, in order to enhance the expression effect of the local time series features of vibration related to the subsequent tasks in the aggregation process, in the technical solution of the present application, the set of implicitly associated features of the local time series of vehicle body vibration is further subjected to feature significant aggregation processing based on cluster analysis to obtain the significant aggregation representation of the vehicle body vibration mode time series. In this way, the most representative features can be extracted from the numerous implicitly associated features of the local time series of vehicle body vibration, which helps to more comprehensively and accurately describe the vehicle body vibration mode time series characteristics, so as to perform vibration compensation calculation and contact network geometric parameter correction.Specifically, the process of feature significant aggregation processing based on cluster analysis includes the following steps: first, cluster analysis is performed on the set of implicitly associated feature vectors of the local time series of body vibration using an unsupervised learning method, so as to automatically discover the natural patterns and structures in the local time series characteristics of the body vibration amplitude, and generate a self-supervised cluster representation vector of the local time series associated features of the body vibration. This method can reveal the inherent structure of the set of implicitly associated feature vectors of the local time series of body vibration without labeled data, which helps the model learn meaningful local time series feature representations of vibration. Next, the implicit clustering contribution factors between each implicitly associated feature vector of the local time series of body vibration and the self-supervised cluster representation vector of the local time series associated features of the body vibration are calculated, and these factors reflect the degree of contribution of each implicitly associated feature vector of the local time series of body vibration to the cluster to which it belongs. This step helps to identify which feature vectors represent the local time series features of body vibration that play a more important role in the clustering process. Then, the sequence of implicit clustering contribution factors of the local time series correlation feature of the vehicle body vibration is vectorized and arranged into the distribution vector of the clustering contribution field of the local time series correlation feature of the vehicle body vibration, that is, the sequence of implicit clustering contribution factors is integrated into the vector field using data integration technology to construct the clustering contribution field. This field represents the relative importance of each local time series implicit correlation feature vector of the vehicle body vibration in the entire cluster. Then, the distribution vector of the clustering contribution field of the local time series correlation feature of the vehicle body vibration is explicitly modeled using the self-attention mechanism to generate the modulation weight vector of the clustering contribution field of the local time series correlation feature of the vehicle body vibration. The self-attention mechanism allows the model to focus on the local time series features of the vibration amplitude that contribute greatly to the clustering, while suppressing the expression of unimportant features, thereby improving the quality of feature representation and the accuracy of vibration compensation. Finally, the set of the local time series implicit correlation feature vectors of the vehicle body vibration is weightedly fused using the modulation weight vector of the clustering contribution field of the local time series correlation feature of the vehicle body vibration to generate a significant aggregation representation vector of the time series of the vehicle body vibration mode. This cluster representation vector combines the original vibration timing feature information and the knowledge learned in the clustering process, enabling the model to more accurately capture the vibration mode timing changes and aggregate representation key features, thereby improving the performance and generalization ability of the model.

[0053] In an embodiment of the present application, a feature significant aggregation processing based on cluster analysis is performed on the set of local temporal implicit correlation features of the vehicle body vibration to obtain the temporal significant aggregation representation of the vehicle body vibration mode, including: cluster analysis is performed on the set of local temporal implicit correlation feature vectors of the vehicle body vibration to obtain a self-supervised cluster representation vector of local temporal correlation features of the vehicle body vibration; based on the self-supervised cluster representation vector of local temporal correlation features of the vehicle body vibration, feature clustering field modulation aggregation is performed on the set of local temporal implicit correlation feature vectors of the vehicle body vibration to obtain a temporal significant aggregation representation vector of the vehicle body vibration mode as the temporal significant aggregation representation of the vehicle body vibration mode.

[0054] Among them, the process of performing cluster analysis on the set of implicit correlation feature vectors of local time series of vehicle body vibration to obtain a self-supervised clustering representation vector of local time series correlation features of vehicle body vibration includes: calculating the mean vector of the set of implicit correlation feature vectors of local time series of vehicle body vibration to obtain a self-supervised clustering representation vector of local time series correlation features of vehicle body vibration.

[0055] More specifically, based on the self-supervised clustering representation vector of the local time series correlation feature of the vehicle body vibration, the set of the local time series implicit correlation feature vectors of the vehicle body vibration is subjected to feature clustering field modulation aggregation to obtain a vehicle body vibration mode time series significant aggregation representation vector as the vehicle body vibration mode time series significant aggregation representation process, including: calculating the implicit clustering contribution factor of each local time series implicit correlation feature vector of the vehicle body vibration in the set of the local time series implicit correlation feature vector of the vehicle body vibration relative to the self-supervised clustering representation vector of the local time series correlation feature of the vehicle body vibration to obtain a sequence of implicit clustering contribution factors of the local time series correlation feature of the vehicle body vibration; The sequence of implicit clustering contribution factors of the local temporal correlation characteristics of vibration is vectorized to obtain the clustering contribution field distribution vector of the local temporal correlation characteristics of vehicle body vibration; the clustering contribution field distribution vector of the local temporal correlation characteristics of vehicle body vibration is input into the clustering contribution field explicit modeling module based on the self-attention mechanism to obtain the clustering contribution field modulation weight vector of the local temporal correlation characteristics of vehicle body vibration; each eigenvalue in the clustering contribution field modulation weight vector of the local temporal correlation characteristics of vehicle body vibration is used as a weight to calculate the position-weighted sum of the set of the local temporal implicit correlation feature vectors of the vehicle body vibration to obtain the temporal significant aggregation representation vector of the vehicle body vibration mode. Among them, calculating the implicit clustering contribution factor of each local time series implicit correlation feature vector of the vehicle body vibration in the set of the local time series implicit correlation feature vector of the vehicle body vibration relative to the local time series correlation feature self-supervised clustering representation vector of the vehicle body vibration to obtain a sequence of implicit clustering contribution factors of the local time series correlation feature of the vehicle body vibration, including: calculating the absolute value of the positional division between the local time series implicit correlation feature vector of the vehicle body vibration and the local time series correlation feature self-supervised clustering representation vector of the vehicle body vibration to obtain the local time series self-supervised clustering interactive correlation feature vector of the vehicle body vibration; calculating the local time series self-supervised clustering representation vector of the vehicle body vibration with a base of 2. The logarithmic function value of each position feature value in the self-supervised clustering interaction correlation feature vector is used to obtain the self-supervised clustering interaction logarithmic representation vector of the local time series of vehicle body vibration; the weighted sum of the feature values ​​at each position in the local time series implicit correlation feature vector of the vehicle body vibration is calculated using the feature values ​​at each position in the self-supervised clustering interaction logarithmic representation vector of the local time series of vehicle body vibration as weighting coefficients to obtain the clustering contribution representation factor of the local time series correlation feature of vehicle body vibration; the exponential function value of the clustering contribution representation factor of the local time series correlation feature of vehicle body vibration is calculated with the natural constant e as the base to obtain the implicit clustering contribution factor of the local time series correlation feature of vehicle body vibration.More specifically, the body vibration local temporal correlation feature cluster contribution field distribution vector is input into the cluster contribution field explicit modeling module based on the self-attention mechanism to obtain the body vibration local temporal correlation feature cluster contribution field domain modulation weight vector, including: calculating the matrix multiplication between the body vibration local temporal correlation feature cluster contribution field distribution vector and the query weight matrix, key weight matrix and value weight matrix respectively to obtain the body vibration cluster contribution field distribution query vector, the body vibration cluster contribution field distribution key vector and the body vibration cluster contribution field distribution value vector; after calculating the multiplication between the body vibration cluster contribution field distribution query vector and the transposed vector of the body vibration cluster contribution field distribution key vector, the obtained matrix is ​​divided by the square root of the length of the body vibration cluster contribution field distribution key vector by position to obtain the query-key body vibration cluster contribution field semantic interaction representation matrix; utilizing. The function normalizes the query-key vehicle body vibration cluster contribution field semantic interaction representation matrix to obtain a normalized query-key vehicle body vibration cluster contribution field semantic interaction representation matrix; calculates the multiplication between the vehicle body vibration cluster contribution field distribution value vector and the query-key vehicle body vibration cluster contribution field semantic interaction representation matrix to obtain the vehicle body vibration local temporal correlation feature cluster contribution field domain modulation weight vector.

[0056] In summary, in the above embodiment, the set of implicitly associated features of the local time series of the vehicle body vibration is subjected to a feature significant aggregation process based on cluster analysis to obtain a time series significant aggregation representation of the vehicle body vibration mode, including: performing a feature significant aggregation process based on cluster analysis on the set of implicitly associated features of the local time series of the vehicle body vibration to obtain a time series significant aggregation representation of the vehicle body vibration mode using the following feature aggregation formula; wherein the feature aggregation formula is:

[0057]

[0058]

[0059]

[0060]

[0061]

[0062]

[0063]

[0064]

[0065]

[0066] in, is the set of implicitly associated eigenvectors of the local time series of the vehicle body vibration, are the first, second, and third implicit correlation feature vectors in the set of the local time series of the vehicle body vibration. and The implicit correlation eigenvector of the local time series of vehicle body vibration, is the first in the set of implicitly associated eigenvectors of the local time series of vehicle body vibration The implicit correlation eigenvector of the local time series of vehicle body vibration, is the number of eigenvectors in the set of implicitly associated eigenvectors of the local time series of vehicle body vibration, is the self-supervised clustering representation vector of the local time series correlation features of vehicle body vibration, is the first in the set of implicitly associated eigenvectors of the local time series of vehicle body vibration The implicit correlation eigenvector of the local time series of vehicle body vibration, For the said The first implicit correlation feature vector of the local time series of vehicle body vibration The eigenvalues ​​at the positions, is the first in the self-supervised clustering representation vector of the local time series correlation feature of the vehicle body vibration The eigenvalues ​​at the positions, represents the logarithmic function value with base 2, represents the value of the natural exponential function, For the said The implicit clustering contribution factor of the local time series correlation feature of the vehicle body vibration corresponding to the implicit correlation feature vector of the local time series of the vehicle body vibration, The cluster contribution field distribution vector of the local time series correlation feature of the vehicle body vibration is obtained by quantizing the implicit cluster contribution factor vectors of the local time series correlation feature of the vehicle body vibration. , and They are query weight matrix, key weight matrix and value weight matrix respectively. , and are the query vector of the vehicle body vibration cluster contribution field distribution, the key vector of the vehicle body vibration cluster contribution field distribution, and the value vector of the vehicle body vibration cluster contribution field distribution, respectively. is the length of the vehicle body vibration cluster contribution field distribution bond vector, for function, is vector multiplication, Contribute to the field modulation weight vector for clustering the local temporal correlation features of vehicle body vibration, A significant aggregation representation vector for the vehicle body vibration mode time series.

[0067] Specifically, the S23 generates the vibration compensation parameters based on the time-series significant aggregation representation of the vehicle body vibration mode. In a specific example of the present application, the time-series significant aggregation representation vector of the vehicle body vibration mode is input into a compensation module based on a decoder to obtain a decoded value of the vibration compensation parameter. That is, the time-series aggregation representation feature of the vehicle body vibration mode is used to perform decoding regression, so as to calculate the vibration compensation parameters to perform parameter compensation on the contact network geometric parameters affected by the vibration in the measurement system, and the contact network coordinates and geometric parameters are corrected by the vibration compensation parameters so that the measurement results are consistent with the measurement values ​​when the vehicle is stationary. In this way, the contact network geometric parameters can be automatically corrected based on the data of the vibration impact, thereby providing more accurate measurement data, providing a basis for subsequent data analysis and detection, and helping the railway management department to more effectively manage and maintain rail transit infrastructure.

[0068] Preferably, the vehicle body vibration mode time series significant aggregation representation vector is input into a decoder-based compensation module to obtain a decoded value of the vibration compensation parameter, including:

[0069] Based on the association value and distance value of the eigenvalue of the vehicle body vibration mode time series significant aggregation representation vector, the vehicle body vibration mode time series significant aggregation association response matrix and the vehicle body vibration mode time series significant aggregation distance response matrix are determined, that is:

[0070]

[0071]

[0072] in, and respectively represent the first The eigenvalues ​​and eigenvalues, represents the significant aggregation representation vector of the vehicle body vibration mode time series, represents the set of real numbers, The length of the vector representing the significant aggregation of the vehicle body vibration mode time series, Indicates the significant aggregation correlation response matrix of the vehicle body vibration mode time series The eigenvalues ​​of the positions, Indicates the significant aggregation distance response matrix of the vehicle body vibration mode time series The eigenvalue of the position;

[0073] Performing matrix multiplication of the vehicle body vibration mode time series significant aggregation representation vector and the vehicle body vibration mode time series significant aggregation association response matrix to obtain a vehicle body vibration mode time series significant aggregation association response vector;

[0074] Performing matrix multiplication of the vehicle body vibration mode time series significant aggregation distance response matrix and the transposed vector of the vehicle body vibration mode time series significant aggregation representation vector to obtain a vehicle body vibration mode time series significant aggregation distance response vector;

[0075] After performing a dot product on the vehicle body vibration mode time series significant aggregation association response matrix and the vehicle body vibration mode time series significant aggregation distance response matrix, matrix multiplication is performed on the transposed vector of the vehicle body vibration mode time series significant aggregation representation vector to obtain a vehicle body vibration mode time series significant aggregation logic bias vector;

[0076] The vehicle body vibration mode time series significant aggregation associated response vector, the vehicle body vibration mode time series significant aggregation distance response vector and the vehicle body vibration mode time series significant aggregation logic bias vector are added to obtain an optimized vehicle body vibration mode time series significant aggregation representation vector;

[0077] The optimized vehicle body vibration mode time series significant aggregation representation vector is input into a decoder-based compensation module to obtain a decoded value of the vibration compensation parameter.

[0078] Here, the optimized vehicle body vibration mode time series significant aggregation representation vector It is expressed as:

[0079]

[0080] exist ,and In the case of:

[0081]

[0082]

[0083] in, represents the significant aggregation correlation response matrix of the vehicle body vibration mode time series, Represents the significant aggregation distance response matrix of the vehicle body vibration mode time series, represents the significant aggregation representation vector of the vehicle body vibration mode time series, represents the set of real numbers, The length of the vector representing the significant aggregation of the vehicle body vibration mode time series, represents matrix multiplication, It means point multiplication by position. Indicates adding by position point. represents the transpose of a vector, Indicates the significant aggregation intermediate vector of the vehicle body vibration mode time series, Represents the significant aggregation logic bias vector of the vehicle body vibration mode timing, The optimized vehicle body vibration mode time series significant aggregation representation vector.

[0084] That is, considering that each of the local time series implicit correlation feature vectors of the vehicle body vibration in the set of the local time series implicit correlation feature vectors respectively represents the short-range-long-range bidirectional time series correlation characteristics of the vehicle body vibration amplitude value in the local time domain, when performing feature significant aggregation based on clustering contribution field analysis, the difference in time series field clustering contribution of each local time domain in the global time domain will lead to sparse aggregation distribution, resulting in large-scale simple repetition of the feature distribution of the vehicle body vibration mode time series significant aggregation representation vector, which affects the logical dependency of the vibration mode semantic aggregation and reduces the accuracy of the decoding result based on the semantic distribution.

[0085] Therefore, by using the autocorrelation matrix and self-distance matrix of the vehicle body vibration mode temporal significant aggregation representation vector as the statistical reference-free distribution response framework of the vehicle body vibration mode temporal significant aggregation representation vector, the simple repetition of feature distribution is avoided based on the retrieval-enhanced reverse response construction of the vehicle body vibration mode temporal significant aggregation representation vector, and the surface combination mapping of the vehicle body vibration mode temporal significant aggregation representation vector is avoided by ensuring the intrinsic mapping logic of the vehicle body vibration mode temporal significant aggregation representation vector based on the retrieval-response context, so as to realize the logical dependency mapping of the vehicle body vibration mode temporal significant aggregation representation vector to the target decoding domain while maintaining the intuitive response correlation, thereby improving the accuracy of the decoded value of the vibration compensation parameter obtained by the vehicle body vibration mode temporal significant aggregation representation vector input to the compensation module based on the decoder. In this way, the vibration compensation parameter can be calculated more accurately, so as to perform parameter compensation on the contact network geometric parameters affected by vibration in the measurement system, and the contact network coordinates and geometric parameters are corrected by the vibration compensation parameter, so that the measurement result is consistent with the measurement value when the vehicle is stationary.

[0086] In particular, the S3 uses the vibration compensation parameters to correct the measured values ​​of the contact network geometric parameters to obtain the corrected measured values ​​of the contact network geometric parameters. In one example, after obtaining the vibration compensation parameter results, it is necessary to perform parameter compensation on the measured parameter values ​​in the measurement system that are affected by vibration, including the measured values ​​of the outer rail superelevation, level, height, triangular pits and contact network geometric parameters. The contact network coordinates, the two-track point cloud coordinates and the measurement angle of the vehicle body posture sensor are corrected by the vibration compensation parameters so that the measurement results are consistent with the measured values ​​when the vehicle is stationary. The system detects the network and track parameters method as follows: Figure 7 After correcting the track point cloud coordinates according to the vibration parameters, the height difference between the two tracks is measured to calculate the track horizontal deviation. Figure 8The figure is a schematic diagram of track horizontal deviation detection. The horizontal deviation of the two tracks is continuously measured, and the triangular pit value is measured every 6.25m. The triangular pit detection method is as follows Fig. 9 As shown, the horizontal deviation at AA' is -1mm, the horizontal deviation at BB' is +2mm, and the horizontal deviation at CC' is -3mm. The triangular pit value of the AA'-BB' section = |-1-(+2)|=3mm, and the triangular pit value of the BB'-CC' section = |+2-(-3)|=5mm. Continuously measure the vertical coordinate value transformation of the same track to detect the height difference of the rail. The height difference detection method is as follows Fig.10 As shown. At the curve position, the track point cloud coordinates and the measurement angle of the attitude sensor are corrected by the vibration parameters, and then the superelevation is measured by trigonometric function in combination with the track gauge. Among them, the superelevation detection method is as follows Fig.11 As shown. According to the contact network coordinate values ​​measured by the vibration parameter compensation contact network detection system, the accurate conductor height pull-out value is obtained. Among them, the conductor height pull-out value detection schematic diagram is shown in Fig.12 The track direction and track wear in the track inspection parameters are obtained by continuously detecting the heading angle transformation of the vehicle body posture sensor and comparing the track point cloud data with the standard track drawing parameter difference. Among them, the track direction and track wear detection method is as follows Fig.13 shown. Fig.14 This is a schematic diagram of track detection. Fig.15 Schematic diagram of track wear detection.

[0087] In particular, the S4 displays the measured values ​​of the corrected contact network geometric parameters on the vehicle control screen in real time. By displaying the measured values ​​of the corrected contact network geometric parameters on the vehicle control screen in real time, it is possible to provide instant correction data, provide data reference for maintenance personnel, and enable operators to quickly understand the current state of the contact network.

[0088] In summary, the contact network-track coordinated comprehensive inspection method according to the embodiment of the present application is explained, which collects the time series data of the vehicle body vibration amplitude value through a vibration sensor, and introduces a data processing and analysis algorithm based on artificial intelligence and deep learning at the back end to analyze the time series data of the vehicle body vibration amplitude value, so as to capture and characterize the local change characteristics of the vehicle body vibration amplitude in the time series and the vibration mode of the whole time domain as a whole, so as to calculate the vibration compensation parameter results to perform parameter compensation on the measurement parameter values ​​affected by vibration in the measurement system, such as the measurement values ​​of the contact network geometric parameters, and correct the contact network coordinates and geometric parameters through the vibration compensation parameters, so that the measurement results are consistent with the measurement values ​​when the vehicle is stationary. In this way, the contact network geometric parameters can be automatically corrected based on the data of the vibration impact, thereby providing more accurate measurement data, providing a basis for subsequent data analysis and detection, and helping railway management departments to more effectively manage and maintain rail transit infrastructure.

[0089] Furthermore, a contact network-track coordinated comprehensive inspection system is also provided.

[0090] Fig.16 FIG. 1 is a block diagram of a contact network-track coordinated comprehensive inspection system according to an embodiment of the present application. Fig.16 As shown, the contact network-track collaborative comprehensive inspection system 300 according to the embodiment of the present application includes: a calibration module 310, which is used to calibrate the contact network detection system and the laser profiler system to convert the contact network line scanning camera coordinate system of the contact network detection system and the left and right laser profiler coordinate systems of the laser profiler system into a unified measurement coordinate system, so that the measurement coordinate system of the vehicle when it is static coincides with the track center coordinate system; a data acquisition module 320, which is used to drive the vehicle to move, obtain the measured values ​​of the contact network geometric parameters, and collect the time queue of the vehicle body vibration amplitude value through the vibration sensor, and calculate the vibration compensation parameters based on the time queue of the vehicle body vibration amplitude value; a correction module 330, which is used to correct the measured values ​​of the contact network geometric parameters using the vibration compensation parameters to obtain the corrected measured values ​​of the contact network geometric parameters; a measurement value visualization module 340, which is used to display the corrected measured values ​​of the contact network geometric parameters on the vehicle control screen in real time.

[0091] As described above, the contact network-track collaborative comprehensive inspection system 300 according to the embodiment of the present application can be implemented in various wireless terminals, such as a server with a contact network-track collaborative comprehensive inspection algorithm. In one possible implementation, the contact network-track collaborative comprehensive inspection system 300 according to the embodiment of the present application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the contact network-track collaborative comprehensive inspection system 300 can be a software module in the operating system of the wireless terminal, or it can be an application developed for the wireless terminal; of course, the contact network-track collaborative comprehensive inspection system 300 can also be one of the many hardware modules of the wireless terminal.

[0092] Alternatively, in another example, the contact network-track collaborative integrated inspection system 300 and the wireless terminal may also be separate devices, and the contact network-track collaborative integrated inspection system 300 may be connected to the wireless terminal via a wired and / or wireless network, and transmit interactive information in accordance with an agreed data format.

[0093] In another example, the vehicle undergoes complex vibrations while collecting data during operation. At the same time, the vehicle's vibration state is a superposition of multiple vibrations, that is, multiple vibrations occur simultaneously. At this time, the camera follows the vibration of the vehicle body, and the relative position of the camera, track, and contact network changes. The geometric parameter detection camera measurement value is affected by the vibration of the vehicle body and produces errors. It is necessary to prioritize the measurement of vibration quantity and then analyze the subsequent network track data. Fig.17 As shown, x'o'y' is the track center coordinate system, and xoy is the coordinate system where the measurement system measurement value is located after the vehicle body vibration occurs.

[0094] First, the composite vibration posture of the vehicle body is evaluated through the coordinate points of the two tracks. Fig.18 As shown, the point cloud data of the left and right tracks are analyzed, the distance between the positioning points of the left and right tracks is calculated to obtain the track gauge, and the arc tangent value of the ratio of the height difference between the two tracks and the track gauge is calculated to calculate the current rolling angle of the vehicle body.

[0095] Then, if Fig.19 As shown in the figure, after the roll angle is obtained, the horizontal and vertical offset distances of the vehicle body after the vibration is generated can be calculated by comparing the coordinate transformation of the two track positioning points in the static state.

[0096] So far, the vibration compensation parameters of the vehicle body are measured as follows: vehicle vibration roll angle, horizontal offset distance and vertical offset distance of the center of the vehicle body, such as Fig. 20 and Fig.21 shown.

[0097] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A contact network-track coordinated comprehensive inspection method, characterized in that: include: Calibrate the catenary detection system and the laser profiler system to convert the catenary line scanning camera coordinate system of the catenary detection system and the left and right laser profiler coordinate systems of the laser profiler system into a unified measurement coordinate system, so that the measurement coordinate system coincides with the track center coordinate system when the vehicle is static; The vehicle is driven to move, and the measured values ​​of the contact network geometric parameters are obtained, and the time sequence of the vehicle body vibration amplitude values ​​of the vehicle is collected through a vibration sensor, and the vibration compensation parameters are calculated based on the time sequence of the vehicle body vibration amplitude values; Correcting the measured values ​​of the contact network geometric parameters using the vibration compensation parameters to obtain corrected measured values ​​of the contact network geometric parameters; Displaying the measured values ​​of the corrected contact network geometric parameters in real time on the vehicle control screen; The calculation of vibration compensation parameters based on the time queue of the vehicle body vibration amplitude value includes: Sampling the time queue of the vehicle body vibration amplitude value based on a predetermined sampling window and with a preset step size to obtain a set of local time queues of the vehicle body vibration amplitude value; The set of local time queues of the vehicle body vibration amplitude values ​​is subjected to feature extraction and feature significant aggregation processing to obtain a significant aggregation representation of the vehicle body vibration mode time series, including: performing time series encoding on the set of local time queues of the vehicle body vibration amplitude values ​​to obtain a set of implicit correlation features of the vehicle body vibration local time series; performing feature significant aggregation processing based on cluster analysis on the set of implicit correlation features of the vehicle body vibration local time series to obtain a significant aggregation representation of the vehicle body vibration mode time series; The vibration compensation parameters are generated based on the significant aggregation representation of the vehicle body vibration mode time series.

2. The contact network-track coordinated comprehensive inspection method according to claim 1 is characterized in that: The set of local time queues of the vehicle body vibration amplitude values ​​is time-series encoded to obtain a set of local time series implicit correlation features of vehicle body vibration, including: inputting each local time queue of the vehicle body vibration amplitude value in the set of local time queues of the vehicle body vibration amplitude values ​​into a vehicle body vibration time series sequence encoder based on a Bi-LSTM model to obtain a set of local time series implicit correlation feature vectors of vehicle body vibration as the set of local time series implicit correlation features of vehicle body vibration.

3. The contact network-track coordinated comprehensive inspection method according to claim 2 is characterized in that: The set of implicitly associated features of the local time series of the vehicle body vibration is subjected to a feature significant aggregation process based on cluster analysis to obtain a significant aggregation representation of the time series of the vehicle body vibration mode, including: Performing cluster analysis on the set of implicit correlation feature vectors of the local time series of vehicle body vibration to obtain a self-supervised cluster representation vector of the correlation feature of the local time series of vehicle body vibration; Based on the self-supervised clustering representation vector of the local time series correlation feature of the vehicle body vibration, the set of the local time series implicit correlation feature vectors of the vehicle body vibration is subjected to feature clustering field modulation aggregation to obtain a vehicle body vibration mode time series significant aggregation representation vector as the vehicle body vibration mode time series significant aggregation representation.

4. The contact network-track coordinated comprehensive inspection method according to claim 3 is characterized in that: A cluster analysis is performed on the set of implicit correlation feature vectors of the local time series of vehicle body vibration to obtain a self-supervised cluster representation vector of the local time series correlation features of the vehicle body vibration, including: calculating the mean vector of the set of implicit correlation feature vectors of the local time series of vehicle body vibration to obtain a self-supervised cluster representation vector of the local time series correlation features of the vehicle body vibration.

5. The contact network-track coordinated comprehensive inspection method according to claim 4 is characterized in that: Based on the self-supervised clustering representation vector of the local time series correlation feature of the vehicle body vibration, a set of implicit correlation feature vectors of the local time series of the vehicle body vibration is subjected to feature clustering field modulation aggregation to obtain a vehicle body vibration mode time series significant aggregation representation vector as the vehicle body vibration mode time series significant aggregation representation, including: Calculating the implicit clustering contribution factor of each vehicle body vibration local time series implicit correlation feature vector in the set of vehicle body vibration local time series implicit correlation feature vectors relative to the vehicle body vibration local time series correlation feature self-supervised clustering representation vector to obtain a sequence of vehicle body vibration local time series correlation feature implicit clustering contribution factors; Vectorizing and arranging the sequence of implicit clustering contribution factors of the local time series correlation characteristics of vehicle body vibration to obtain a clustering contribution field distribution vector of the local time series correlation characteristics of vehicle body vibration; Input the distribution vector of the cluster contribution field of the local time series correlation feature of the vehicle body vibration into the cluster contribution field explicit modeling module based on the self-attention mechanism to obtain the modulation weight vector of the cluster contribution field of the local time series correlation feature of the vehicle body vibration; Taking each eigenvalue in the field modulation weight vector of the cluster contribution of the local time series correlation feature of the vehicle body vibration as a weight, the position-weighted sum of the set of implicit correlation feature vectors of the local time series of the vehicle body vibration is calculated to obtain the significant aggregation representation vector of the time series of the vehicle body vibration mode.

6. The contact network-track coordinated comprehensive inspection method according to claim 5 is characterized in that: Calculating the implicit clustering contribution factor of each vehicle body vibration local time series implicit correlation feature vector in the set of the vehicle body vibration local time series implicit correlation feature vectors relative to the vehicle body vibration local time series correlation feature self-supervised clustering representation vector to obtain a sequence of vehicle body vibration local time series correlation feature implicit clustering contribution factors, including: Calculating the absolute value of the positional division between the vehicle body vibration local time series implicit correlation feature vector and the vehicle body vibration local time series correlation feature self-supervised clustering representation vector to obtain the vehicle body vibration local time series self-supervised clustering interactive correlation feature vector; Calculate the logarithmic function value of each position feature value in the vehicle body vibration local time series self-supervised clustering interactive correlation feature vector with base 2 to obtain the vehicle body vibration local time series self-supervised clustering interactive logarithmic representation vector; Taking each position feature value in the vehicle body vibration local time series self-supervised clustering interactive logarithm representation vector as a weighting coefficient, calculating the weighted sum of each position feature value in the vehicle body vibration local time series implicit correlation feature vector to obtain a vehicle body vibration local time series correlation feature clustering contribution degree representation factor; The exponential function value of the cluster contribution degree representation factor of the local time series correlation feature of the vehicle body vibration is calculated with the natural constant e as the base to obtain the implicit cluster contribution factor of the local time series correlation feature of the vehicle body vibration.

7. The contact network-track coordinated comprehensive inspection method according to claim 6 is characterized in that: The distribution vector of the cluster contribution field of the local time series correlation feature of the vehicle body vibration is input into the cluster contribution field explicit modeling module based on the self-attention mechanism to obtain the modulation weight vector of the cluster contribution field of the local time series correlation feature of the vehicle body vibration, including: Calculating matrix multiplications between the vehicle body vibration local time series correlation feature cluster contribution field distribution vector and the query weight matrix, the key weight matrix and the value weight matrix respectively to obtain a vehicle body vibration cluster contribution field distribution query vector, a vehicle body vibration cluster contribution field distribution key vector and a vehicle body vibration cluster contribution field distribution value vector; After calculating the multiplication between the body vibration cluster contribution field distribution query vector and the transposed vector of the body vibration cluster contribution field distribution key vector, the obtained matrix is ​​divided by the square root of the length of the body vibration cluster contribution field distribution key vector by position to obtain a query-key body vibration cluster contribution field semantic interaction representation matrix; use The function normalizes the query-key vehicle body vibration cluster contribution field semantic interaction representation matrix to obtain a normalized query-key vehicle body vibration cluster contribution field semantic interaction representation matrix; The multiplication between the vehicle body vibration cluster contribution field distribution value vector and the query-key vehicle body vibration cluster contribution field semantic interaction representation matrix is ​​calculated to obtain the vehicle body vibration local temporal correlation feature cluster contribution field domain modulation weight vector.

8. The contact network-track coordinated comprehensive inspection method according to claim 7 is characterized in that: Generating the vibration compensation parameter based on the vehicle body vibration mode temporal significant aggregation representation includes: inputting the vehicle body vibration mode temporal significant aggregation representation vector into a decoder-based compensation module to obtain a decoded value of the vibration compensation parameter.

9. A contact network-track coordinated comprehensive inspection system, using the contact network-track coordinated comprehensive inspection method according to claim 1, characterized in that: include: A calibration module, used for calibrating the catenary detection system and the laser profiler system to convert the catenary line scanning camera coordinate system of the catenary detection system and the left and right laser profiler coordinate systems of the laser profiler system into a unified measurement coordinate system, so that the measurement coordinate system of the vehicle when it is static coincides with the track center coordinate system; A data acquisition module, used for driving the vehicle to move, acquiring the measured values ​​of the contact network geometric parameters, collecting the time queue of the vehicle body vibration amplitude value through a vibration sensor, and calculating the vibration compensation parameters based on the time queue of the vehicle body vibration amplitude value; A correction module, used for correcting the measured value of the contact network geometric parameter by using the vibration compensation parameter to obtain the corrected measured value of the contact network geometric parameter; The measurement value visualization module is used to display the measured values ​​of the corrected contact network geometric parameters in real time on the vehicle control screen.

Images