A visualization method for track acoustic vibration anomaly detection

Abstract

The application discloses a kind of visualization methods for track acoustic vibration anomaly detection, belong to track transportation infrastructure state monitoring and intelligent operation and maintenance technical field;Detection date, train number, running direction, station name, section mileage, running speed, original vibration index, abnormal evaluation value after speed correction, abnormal discrimination result and section statistical result are unified reading and structured organization, construct original vibration display index, speed correction after display index and section statistical display index, and form track state contrast chart, inter-station normalized statistical chart, inter-station section color coding track state chart, map mapping chart and long-term change contrast chart respectively.Compared with prior art, the method of the application can significantly improve the readability, interpretability and engineering application value of track anomaly detection results, and is suitable for track state visualization display and operation and maintenance auxiliary decision-making in continuous detection scenario.

Description

Technical Field

[0001] This invention relates to a visualization method for detecting track acoustic and vibration anomalies; more specifically, it relates to a method for structuring, expressing, spatially mapping, and analyzing long-term status comparisons of vehicle vibration, location mileage, and related detection results obtained during the operation of in-service trains; it can be applied to scenarios such as track defect screening, key section location, and maintenance effect evaluation; and belongs to the field of rail transit infrastructure condition monitoring and intelligent operation and maintenance technology. Background Technology

[0002] Currently, my country's conventional railway network is approaching 120,000 kilometers, high-speed railway network exceeds 40,000 kilometers, and urban rail transit network exceeds 10,000 kilometers. This vast network, coupled with increasing operating mileage and service life, leads to rising demands for track maintenance and management. Track condition directly affects the safety, smoothness, and comfort of train operation. To ensure the lines are in good working order, railway operators typically conduct line inspections, track geometry checks, and repairs for track defects.

[0003] Existing track condition detection methods based on vehicle vibration can achieve continuous monitoring by utilizing the vibration response of in-service trains during operation, offering advantages such as convenient deployment, wide coverage, and relatively low cost. However, current data aggregation and analysis methods largely rely on manual analysis, resulting in scattered results that are not easily interpreted intuitively, making it difficult to meet the needs of daily high-frequency monitoring, rapid anomaly detection, and continuous tracking of condition changes.

[0004] Further analysis revealed significant shortcomings in the expression and application of existing technologies for detection results: First, the vibration amplitude of the vehicle body is significantly affected by the train's operating speed. If the original vibration amplitude or a single detection curve is used directly for display, it is easy to misjudge the increased vibration caused by speed changes as track deterioration, reducing the engineering credibility of the results. Second, existing detection results are mostly output in the form of raw data tables, single curves, or scattered statistical values, lacking the ability to intuitively display key abnormal sections, which is not conducive to maintenance personnel quickly screening and locating defects. Third, existing results expression usually lacks unified display methods such as inter-station normalized statistics, track map spatial mapping, and long-term time series comparison, making it difficult to form a closed-loop application chain of "anomaly discovery - spatial location - long-term tracking - maintenance assessment". Fourth, for high vibration sections that recur at the same location and sections that deteriorate again after maintenance, existing technologies lack continuous, comparable, and interconnected visual expression methods, making it difficult to effectively support maintenance decisions.

[0005] Therefore, there is an urgent need to propose a visualization method that can uniformly organize the detection results of track acoustic and vibration anomalies, express the velocity influence correction, display the section-level graphical representation, map the display, and perform long-term comparative analysis, so as to improve the readability, interpretability, and engineering application value of the detection results. Summary of the Invention

[0006] To address the aforementioned shortcomings in existing technologies, this invention provides a visualization method for detecting track acoustic and vibration anomalies. This method organizes and graphically represents the raw detection results, speed correction results, section anomaly results, and statistical results output by the track acoustic and vibration anomaly detection system in a structured manner, thereby enabling rapid identification, spatial positioning, local magnified analysis, long-term status tracking, and maintenance effect evaluation of key defect sections.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: A visualization method for detecting track acoustic vibration anomalies, comprising the following steps: S1 reads the detection result data output by the track acoustic and vibration anomaly detection system and organizes the detection result data in a structured manner. The detection result data includes the detection date, train number, direction of operation, station name, section start mileage, section end mileage, operating speed, original vibration index, speed-corrected anomaly evaluation value, anomaly judgment result, and section statistical results. S2 Based on the detection result data, construct display indicators, which include original vibration display indicators, velocity-corrected display indicators, and segment statistical display indicators; S3 Based on the original vibration display index and the speed-corrected display index, generate a comparison chart of the track state before and after speed correction; S4 generates a normalized statistical chart of inter-stations and a color-coded track status chart of inter-station sections based on the inter-station statistical results. S5 maps the results of abnormal events onto a route map or route diagram, generating a route-station-abnormal point distribution map. S6 performs long-term comparative analysis of historical detection results within multiple detection cycles of the same section, generating a long-term change comparison chart; S7 outputs at least one visualization result for the identification, location, tracking, and maintenance support decision-making of abnormal track sections.

[0008] Preferably, in step S1, the detection result data is organized in a structured manner, specifically including the following: The S11 section basic table is used to record the starting mileage location, the station to which each section belongs, and the inspection date; The S12 vibration index table is used to record the original vibration value, corrected vibration value, maximum absolute value, and whether the threshold is exceeded for each section. S13 Statistical Results Table is used to record the frequency of exceeding the threshold, average outlier and maximum outlier of the section in multiple detection cycles; S14 Key Section Table is used to record high-vibration-frequent sections, continuous abnormal sections, and their local magnified display range.

[0009] Preferably, in step S2, constructing the display metrics specifically includes the following: S21 uses Mahalanobis distance to identify and remove outliers in a two-dimensional feature space composed of operating speed and vehicle vibration index. S22 Based on the data after removing outliers, the least squares method was used to fit the velocity-vibration regression relationship; S23 Based on the regression relationship, predict the expected vibration level under the same velocity conditions; S24 The ratio of the measured vibration value to the expected vibration level or the standardized residual is used as the velocity-corrected anomaly evaluation value.

[0010] Preferably, in step S4, the process of forming the inter-station normalized statistical chart is as follows: S411 Count the number of abnormal events exceeding a preset threshold within the section between two adjacent stations; S412 retrieves the valid length between the corresponding stations; S413 Calculate the normalized high acceleration frequency index between stations based on the ratio of the number of abnormal events to the effective length between stations; S414 Based on the inter-station normalized high acceleration frequency index, sort and highlight different inter-station sections.

[0011] Preferably, in step S4, the process of forming the color-coded track status diagram for the inter-station section is as follows: S421 Constructs a two-dimensional data matrix of "time-mileage-segment value" according to the time dimension and the mileage dimension; S422 generates a track status heatmap with mileage as the horizontal axis, detection date or detection cycle as the vertical axis, and color coding based on the maximum absolute vibration value or anomaly evaluation value of the section. S423 statistically analyzes the average value of the maximum vibration in each section over multiple detection cycles to generate a distribution map of the average value of the maximum vibration, which shares the same mileage coordinate axis as the heat map, i.e., a color-coded track status map of the inter-station section.

[0012] Preferably, in step S5, the specific process of mapping the abnormal event results onto the route map or route diagram is as follows: S511 When the detection results have geographic coordinate information, the abnormal events, stations and line boundaries will be projected onto the electronic map or GIS base map; S512 When the detection results do not have geographic coordinate information, construct a route diagram based on the route mileage, and map the stations and abnormal events onto the route diagram according to their mileage locations. S513 Abnormal events are divided into at least two levels based on preset abnormal thresholds and displayed with different colors or different symbols.

[0013] Preferably, step S5 further includes a hotspot segment identification and local magnification display step, as follows: S521 Based on the normalized high acceleration frequency index between stations, the top few stations are selected as hotspot stations; S522 Statistical analysis of the clustered distribution of abnormal events along the mileage within the hotspot station section; S523 defines a local mileage range in which abnormal events repeatedly occur within multiple detection cycles as a hot spot sub-segment; S524 and generate a locally enlarged display image for the hot spot station section or hot spot sub-section.

[0014] Preferably, in step S6, the process of forming the long-term change comparison chart is as follows: S61 aligns historical test results for the same key section across multiple test cycles according to a unified mileage benchmark. S62 divides the target section into multiple sub-segments and establishes a two-dimensional monitoring sequence of time and sub-segments; S63 Based on the two-dimensional monitoring sequence, construct at least one of the following: long-term peak index, long-term average index, corrected long-term anomaly index, long-term fluctuation index, and trend slope index. S64 maps maintenance event times to a long-term change graph; S65 obtains the evaluation results of maintenance effect and the results of re-deterioration identification by comparing the changes of vibration index in the pre-maintenance window and the post-maintenance window.

[0015] Another objective of this invention is to provide a track acoustic vibration anomaly detection and visualization system.

[0016] The above-mentioned objective of this invention is achieved through the following technical solution: A track acoustic vibration anomaly detection and visualization system includes a processor and a memory. The memory stores program instructions that can be executed by the processor. When the program instructions are executed by the processor, the system performs a visualization method for track acoustic vibration anomaly detection.

[0017] Compared with the prior art, the present invention has the following beneficial technical effects: The method of this invention can organize the track acoustic and vibration anomaly detection results in a unified structure and form a graphical output oriented towards engineering applications. Compared with the method of only outputting raw data tables or single curves, it significantly improves the readability and interpretability of the detection results. Before visualization, a speed influence correction mechanism is introduced, which can effectively reduce the anomaly display error caused by changes in train running speed, so that the graphical results more accurately reflect the actual track deterioration location, thereby reducing the risk of misjudgment. Through station-to-station normalized statistics and color-coded track status diagrams, a hierarchical display is achieved from station-to-station screening to detailed inspection of local sections, which is conducive to maintenance personnel quickly identifying high-vibration-frequency sections and key anomalies. Location; by mapping the detection results to the route map or GIS base map, the spatial location of abnormal events and the unified display of risk ranking between stations can be realized, improving the screening efficiency of key sections; through long-term comparative visualization and maintenance event overlay display, the status changes of the same section in different detection cycles can be intuitively reflected, and it can be used for maintenance effect evaluation, disease recurrence identification and re-deterioration early warning, providing a continuous basis for operation and maintenance decision-making; by constructing an integrated display system that links data visualization, map visualization and long-term comparative visualization, a complete result display chain of "graphic overview - map positioning - long-term tracking" is formed, which has good engineering promotion value and operation and maintenance application prospects.

[0018] The present invention will be further described below with reference to the accompanying drawings and specific embodiments; however, this does not imply any limitation on the scope of protection of the present invention. Attached Figure Description

[0019] Figure 1 This is a flowchart of the data visualization method for track acoustic vibration anomaly detection results in Embodiment 1 of the present invention; Figure 2 This is a flowchart of the visualization method for mapping track acoustic vibration anomaly detection results onto a map in Embodiment 2 of the present invention; Figure 3 This is a flowchart of the long-term comparative visualization method for track acoustic vibration anomaly detection results in Embodiment 3 of the present invention; Figure 4 This is a diagram of the track state before velocity correction in the track acoustic vibration anomaly detection results of Embodiment 1 of the present invention; Figure 5 This is a velocity-corrected track state diagram based on the track acoustic vibration anomaly detection results in Embodiment 1 of the present invention; Figure 6 This is a normalized high acceleration frequency statistics chart of the track acoustic vibration anomaly detection results between stations in Embodiment 1 of the present invention; Figure 7 This is a color-coded track status diagram of the inter-station section and the corresponding average distribution diagram of the maximum value in the track acoustic and vibration anomaly detection results of Embodiment 1 of the present invention; Figure 8This is a magnified view of a key section of the track acoustic vibration anomaly detection results in Embodiment 1 of the present invention; Figure 9 This is a visualization diagram of the track acoustic vibration anomaly detection results mapped onto the track map in Embodiment 2 of the present invention; Figure 10 This is a long-term comparative visualization of the track acoustic vibration anomaly detection results in Embodiment 3 of the present invention. Detailed Implementation

[0020] The visualization method for detecting track acoustic and vibration anomalies of the present invention reads, organizes, analyzes, and displays the detection result data output by the track acoustic and vibration anomaly detection system. The detection result data includes the detection date, train number, direction of travel, station name, starting mileage of the section, ending mileage of the section, section length, operating speed, original vibration index, speed-corrected anomaly evaluation value, anomaly judgment result, and section statistical results. The detection results are organized using fixed sections as the basic analysis unit, and a section basic table, a vibration index table, a statistical result table, and a key section table are established. The method includes the following three functional modules or processing procedures: First, the data visualization process for track acoustic and vibration anomaly detection results includes data reading and structured organization, display index construction, generation of track status comparison charts before and after speed correction, generation of inter-station normalized statistical charts, generation of inter-station section color-coded track status charts, local magnification display, and interactive graphic linkage. In the process of constructing the display indicators, original vibration display indicators, speed-corrected display indicators, and section statistical display indicators are established. Among them, the speed-corrected display indicators are obtained by modeling the relationship between "operating speed and vehicle vibration". Mahalanobis distance is preferably used to identify and remove outliers, and then the least squares method is used to fit the speed-vibration regression relationship to obtain the expected vibration level under the same speed conditions. The ratio of the measured value to the predicted value or the standardized residual is used as the corrected anomaly evaluation quantity. Based on the indicators, a comparison map of track status before and after speed correction, a heat map of inter-station sections, a corresponding average distribution map of large values, and a local enlarged map of key sections are generated. Second, the visualization process of mapping track acoustic and vibration anomaly detection results onto a map includes data reading and spatial benchmark establishment, anomaly event extraction, map layer generation, hotspot segment identification, map zoom-in display, and interactive linkage. When the detection results have geographic coordinate information, the abnormal events, stations, and line boundaries are directly mapped onto the electronic map or GIS base map; when the detection results only have mileage information, a line diagram is created, and the line, stations, and abnormal events are mapped and displayed according to the mileage relationship. Abnormal events are classified into at least two levels according to preset thresholds and displayed with different colors or different symbols; at the same time, the number of occurrences of high acceleration events between stations is normalized to form normalized statistical results between stations, so as to achieve rapid screening of hot inter-station sections. Third, the process of visualizing the long-term comparison of track acoustic and vibration anomaly detection results includes long-term data reading and unification, time-series segmentation organization, construction of long-term evaluation indicators, alignment of maintenance events, generation of long-term change diagrams, comparative analysis of maintenance effects, identification of re-deterioration trends, and output of results. Historical test results of the same key section over multiple testing cycles are organized into a time series to construct long-term peak indicators, long-term average indicators, corrected long-term abnormal indicators, fluctuation indicators, and trend indicators. Maintenance operation time markers are overlaid on the long-term change map. By comparing the test results before and after maintenance, maintenance improvement indicators are formed, and sections that deteriorate again after maintenance are further identified. The three processing steps mentioned above share a unified data organization structure and visualization interface, and form a multi-map collaborative display mechanism through mileage linkage, time linkage, threshold linkage and map linkage. This ensures that the same abnormal section maintains a consistent index relationship in the track status map, inter-station statistics map, map layer and long-term change map, thereby achieving continuous visualization of "overview of detection results - identification of key sections - spatial positioning - long-term tracking". The output results include: a comparison chart of track status before and after speed correction, a normalized statistical chart between stations, a color-coded track status chart for inter-station sections, a distribution chart of the average distribution of the largest values ​​in the sections, a magnified view of key sections, a distribution chart of the line-station-abnormal points, a local map of hotspot sections, a comparison chart of long-term changes, a comparison chart of status before and after maintenance, a chart of re-deterioration trend analysis, as well as a corresponding list of abnormal sections, a list of spatial distributions, and a long-term monitoring and analysis report; The output results can take the form of images, vector graphics, PDF reports, web page graphic components, GIS layers, or dynamic chart objects from the operation and maintenance platform.

[0021] The present invention will be further described below with reference to specific embodiments. It should be understood that the following text is only used to describe one or more specific embodiments of the present invention and does not strictly limit the scope of protection specifically claimed by the present invention. In the absence of conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0022] Unless otherwise specified, the components involved in the following embodiments are all commercially available conventional components, and the methods used are all commonly used methods in this technical field.

[0023] Example 1: A data visualization method for track acoustic and vibration anomaly detection results This embodiment provides a data visualization method for track acoustic and vibration anomaly detection results. It presents the original detection results, speed correction results, section anomaly results, and statistical results output by the track acoustic and vibration anomaly detection system in a unified visual format, thereby enabling rapid identification, local magnified analysis, and maintenance-assisted interpretation of key defect sections. This method is applicable to result display modules deployed in industrial control computers, vehicle-mounted terminals, ground analysis platforms, or visualization management platforms, and is particularly suitable for the graphical display of a large number of section-level results in continuous detection scenarios of operating trains. Based on existing research, track condition diagrams can be presented through a comparison of "before speed correction—after speed correction," a color-coded diagram of inter-station sections, and a local magnified diagram of key sections to highlight the actual track deterioration location and suppress misjudgments caused by speed factors. In this embodiment, the data visualization method includes at least eight steps: data reading and structured organization, display indicator construction, speed correction of track status comparison chart, generation of inter-station normalized statistical chart, generation of inter-station section color-coded track status chart, local zoom-in and display of key defects, graphical interaction and linkage, and result output, as detailed below: S1 Data Reading and Structured Organization S11 reads the detection result dataset generated by the track acoustic vibration anomaly detection system. The detection result dataset includes: detection date, train number, running direction, station name, section start mileage, section end mileage, section length, running speed, original car body vertical acceleration value, speed-corrected acceleration evaluation value, anomaly judgment result, and section statistical results. S12 divides the track line into basic display units in 50m sections and establishes a section-level data table with one detection record corresponding to each 10m interval. During the corresponding detection, a new 50m long unit is added every 10m step, and the average effective value of the vehicle's vertical acceleration is used as diagnostic data. To facilitate graphical representation, S13 organizes the original detection results into the following data structure: (1) Section Basic Table: Records the starting mileage position, station to which each section belongs, and inspection date; (2) Vibration index table: Record the original vertical acceleration, corrected acceleration, maximum absolute value and whether the threshold is exceeded for each section; (3) Statistical results table: Record the frequency of exceeding the threshold, average outlier and maximum outlier of the segment in multiple detections; (4) Key Section Table: Records the high-frequency vibration sections, continuous abnormal sections and their local magnified display ranges selected through statistical screening; The above data tables can be stored using relational database tables, CSV tables, in-memory matrices, or key-value mapping objects; S2 Display Metrics Construction In this embodiment, to achieve the visualization effect, at least three types of display indicators are first constructed; The first category is the original vibration indicator, denoted as (A) i ), representing the measured vertical acceleration value of the vehicle body or its average effective value in the (i)th segment, used to draw the track state diagram before speed correction; The second category is the display index after speed correction, denoted as (B). i ), representing the corrected evaluation value of the (i)th segment after considering the speed effect; First, in the two-dimensional space of "running speed - vehicle vertical acceleration", Mahalanobis distance is used to identify and remove outliers. Then, the least squares method is used to fit the speed-acceleration regression relationship. Based on the regression relationship, the expected vehicle vibration level at the same speed is predicted from the speed. The ratio of the measured value to the predicted value or the corresponding standardized residual is used as the corrected anomaly evaluation quantity, thereby realizing the identification of abnormal large value sections. The third category is the segment statistical display indicators, including: (1) the segment maximum absolute acceleration (M i (2) Average maximum acceleration over multiple periods in the segment (E) i (3) Frequency of exceeding the threshold in a section (N) i (4) Section anomaly color level (C) i ); where, (C i Preferred criteria (M) i ), (E) i ) or (B i The abnormalities are classified and coded as follows: normal is blue, mild abnormality is yellow, moderate abnormality is orange, and severe abnormality is red. By "visualizing the maximum absolute acceleration measured on different dates by segment", a track status diagram of the inter-station segment is constructed. S3 performs velocity correction on the orbital state comparison diagram. In this embodiment, the first core graphic is a comparison diagram of the track state before and after speed correction. The specific steps are as follows: Step S3-1: Select the target segment to be displayed. Preferably, it can be a single inter-station segment, a key abnormal mileage segment, or any mileage range selected by the user. The system reads the (A) of all segments within this range. i ) and (B i ); Step S3-2: Establish a shared horizontal axis for mileage, where the horizontal axis represents actual or relative mileage values, and the vertical axis corresponds to the original vibration display indicators (A). i ) and speed correction indicator (B) i ); Step S3-3: Draw the track state diagram before correction: Use a line graph, bar chart, strip chart, or heat strip chart to represent the original vibration level of each section, and overlay highlighted marks at locations exceeding a preset threshold. This view is used to show the vibration distribution of the entire section before the velocity effect is eliminated, such as... Figure 4 The image shown is a diagram of the track state before velocity correction in the track acoustic vibration anomaly detection results of Embodiment 1 of the present invention. Step S3-4: Draw the corrected track status diagram: Using the same horizontal axis as in step S3-3, arrange a second image or a second layer below, and draw the corrected display indicators (B). i ); when (B i When an error exceeds a preset threshold, it will be highlighted in red or dark orange; for example... Figure 5 The image shown is a velocity-corrected track state diagram based on the track acoustic vibration anomaly detection results in Embodiment 1 of the present invention. Since the velocity factor is significantly weakened after velocity correction, this figure is used to reflect anomalous locations that are closer to the actual trajectory degradation. Without correction, large vertical accelerations are detected at most locations, which are mostly due to velocity effects. Step S3-5: Construct a linked comparison display: When the user clicks on any abnormal point, Figure 4 and Figure 5 The system synchronously displays the original value, correction value, speed, predicted acceleration, and anomaly description for the mileage segment, allowing users to directly compare the "state changes of the same mileage before and after correction"; By following the steps above, you can generate a [database name] with [database name]. Figure 4 and Figure 5 An equivalent comparison chart of the track state before and after velocity correction is used to visually display the effect of velocity correction. Generation of S4 Inter-station Normalized Statistical Chart like Figure 6 The image shows a normalized high acceleration frequency statistical chart (inter-station normalized statistical chart) of the track acoustic vibration anomaly detection results in Embodiment 1 of the present invention. The specific generation steps are as follows: Step S4-1: Using the station number or the combination of the first and last station names as the x-axis, and the normalized high acceleration frequency index (Q) as the y-axis. i Use the vertical axis as the ordinate to create a bar chart, line chart, or column chart. Figure 6 In the diagram, the horizontal axis uses letters to represent the first and last stations. Step S4-2: Highlight the station sections that exceed the preset screening threshold with color, preferably red or dark orange; other station sections can be displayed in gray or blue to create visual contrast. Step S4-3: When a user clicks on a column or label between stations, the corresponding station range is automatically highlighted; Generation of color-coded track status diagram for S5 station section like Figure 7 The image shows the color-coded track status map of the inter-station section and the corresponding average distribution map of the maximum values ​​(color-coded track status map of the inter-station section) of the track acoustic vibration anomaly detection results in Embodiment 1 of the present invention; the specific generation steps are as follows: Step S5-1: Determine the target inter-station section: Based on the inter-station statistics, select key inter-station sections with a high frequency of high acceleration as the target inter-station section, such as... Figure 7 First, the frequency of high acceleration is counted at the station-to-station scale. Based on this, it can be determined that the high vibration frequency is highest between stations designated as N and O. Then, a more detailed diagnosis can be made for the station-to-station section. Step S5-2: Organize the detection results by time dimension: For each segment within the target station area, extract the maximum absolute acceleration (M / s) of the segment from multiple detection dates or multiple calendar months. i (or corrected abnormal evaluation values, forming a two-dimensional data matrix of "time-mileage-segment value"); Step S5-3: Generate orbital state heat map ( Figure 7 (Upper section): Using mileage as the horizontal axis and the detection date as the vertical axis, colors represent the maximum absolute acceleration value or abnormal index of a section. If a section is red or orange on multiple dates, it indicates that there is a persistent high risk of vibration at that location. Figure 7 The upper part is the track status diagram of the NO station section, which uses color coding to indicate the status level of different mileage positions at different times; Step S5-4: Generate the average distribution map of the maximum values ​​at the corresponding locations ( Figure 7 Bottom section: For each segment, calculate the average of its maximum acceleration over multiple detection cycles (E). i ), and display them in the form of bar charts, area charts or line charts, forming Figure 7 The lower part shows the average value distribution; this figure is compared with the thermogram of the generated orbital state ( Figure 7 The upper part shares a mileage axis, which facilitates the identification of key abnormal locations from two dimensions: "frequent occurrence in multiple periods" and "average intensity"; Step S5-5: Set color coding rules: Set the color level to 4 or 5 levels: Blue indicates normal range; Yellow indicates slight abnormality; Orange indicates obvious abnormality; Red indicates significant abnormality; Dark red indicates key abnormality that recurs for multiple consecutive periods; For segments with high-level colors for multiple consecutive periods, the system can automatically add a "key focus" mark. By following the steps above, you can generate something like... Figure 7 The color-coded track status diagram and corresponding average distribution diagram of large values ​​shown in the station section are used to identify local high vibration locations that recur over a long period of time. S6 Local Magnification and Key Disease Display like Figure 8 The image shown is a magnified view of a key section of the track acoustic vibration anomaly detection results in Embodiment 1 of the present invention. To enhance visual diagnostic capabilities, a local magnification display method has been further implemented. Specifically, when a certain mileage section is within... Figure 7 The upper part of the thermal map showing the formation orbit state shows a continuous occurrence of high-level red or orange colors, or their average anomaly intensity (E). i When the threshold is exceeded, the system automatically adds the section to the key section table and generates a magnified view of the key section of the track vibration anomaly detection results. Figure 8 ); Figure 8 and Figure 7 The analytical logic is consistent; The enlarged view of the key section of the track acoustic vibration anomaly detection results shows the following: (1) Detailed heat map of several consecutive sections near this mileage ( Figure 8 (above); (2) Histogram of the frequency of high acceleration at this location ( Figure 8 (Lower part); (3) Multi-period maximum acceleration line graph at this location; (4) On-site verification instructions or defect notes at this location; Enlarged view of key sections of track acoustic vibration anomaly detection results ( Figure 8 It can achieve layered display from an overall overview of stations to details of key locations; S7 Graphical Interaction and Linkage The data visualization method in this embodiment is preferably an interactive visualization method, rather than being limited to static graphic output; therefore, the following linkage mechanism is further set up: S7-1 Mileage Linkage: In Figure 6 and Figure 7 When the same mileage location is clicked, the corresponding segment is highlighted simultaneously on all maps; S7-2 Time Linkage: In the Heatmap ( Figure 7 The upper part shows the heat map of the generated orbital state and Figure 8 In the detailed heat map at the top, after selecting a detection date, the system synchronously displays the original waveform summary, velocity information and anomaly location of the corresponding section on that day; S7-3 Threshold Linkage: Users can dynamically adjust the high vibration threshold or color grading threshold, and the image will be refreshed in real time to meet the display needs of different routes, different vehicle types and different management standards. S7-4 Map Linkage: The key section results generated in this embodiment can be linked with the subsequent map visualization module. When the user clicks on a certain section, the map module will automatically jump to the spatial location of that section. S8 Result Output After the graphic generation is completed, the system outputs at least one visualization result. Preferably, the output results include: (1) a comparison diagram of track status before and after speed correction; (2) a color-coded track status diagram of inter-station sections; (3) a diagram of the average distribution of the maximum value of the section; (4) a magnified view of the key section; (5) a list of the corresponding abnormal sections. The output format can be PNG, SVG, PDF, web graphic components or dynamic chart objects that can be embedded in the operation and maintenance platform. Furthermore, the system can generate an automatic report by combining graphics and text descriptions. The report preferably includes: detection date, line section, main abnormal mileage, abnormality level, graphic number, and suggested review instructions. In this way, the originally scattered detection results can be transformed into visual results that can be directly used for technical analysis and operation and maintenance reporting.

[0024] Technical effects of this embodiment Compared with the method of only outputting numerical tables, this embodiment has the following technical advantages: First, converting existing detection results into a comparison chart before and after speed correction can intuitively reflect the difference in abnormal positioning before and after speed influence removal, thereby reducing misjudgment of track status caused by speed fluctuations. Second, converting the results of inter-station sections into color-coded heat maps and average value distribution maps can intuitively reveal high-vibration-frequency sections and local repetitive anomaly locations, making it easier to discover hidden defects; Third, through local zoom-in and graphic linkage mechanisms, it is possible to achieve hierarchical display from inter-station overview to single-point detailed investigation, thereby improving data interpretation efficiency and engineering usability; Fourth, it can provide a standardized intermediate visualization result interface for subsequent map mapping and long-term comparative display, thus forming a complete visualization chain of "graphic overview - map positioning - long-term tracking".

[0025] Example 2: A method for visualizing track acoustic and vibration anomaly detection results onto a map This embodiment provides a visualization method for mapping track acoustic and vibration anomaly detection results onto a map. It is used to convert the section anomaly results, anomaly event results, and inter-station statistical results output by the track acoustic and vibration anomaly detection system into spatial map layers and inter-station statistical layers for comprehensive display. It is applicable to industrial control computer analysis platforms, ground visualization management platforms, line operation and maintenance terminals, or GIS display terminals. It can quickly screen high-vibration-frequent sections in the line, spatially locate anomaly points, and provide graphical entry points for detailed local inspections of key sections. like Figure 9The image shown is a visualization diagram of the track acoustic vibration anomaly detection results mapped onto the line map in Embodiment 2 of the present invention. The visualization method of mapping the track acoustic vibration anomaly detection results onto the line map in Embodiment 2 of the present invention is to spatially present high acceleration events by combining "inter-station normalized statistical chart + line station distribution map". The line is represented by blue lines, the stations are represented by light blue squares, and the larger acceleration points are represented by orange and red dots. In this embodiment, the visualization method mapped to the map includes six processes: data reading and spatial benchmark establishment, abnormal event extraction, map layer generation, hotspot segment identification and map zoom-in display, map interaction and linked display methods, and result output, as detailed below: S1 Data Reading and Spatial Benchmark Establishment S1-1 reads the detection result dataset output by the track acoustic and vibration anomaly detection system, including: detection date, train number, direction of travel, station name, section start mileage, section end mileage, vehicle speed, maximum absolute value of vertical acceleration of the vehicle body, anomaly evaluation value after speed correction, threshold exceedance marker, and optional latitude and longitude coordinates. If the detection system itself only outputs mileage coordinates and not latitude and longitude coordinates, it is preferable to read the basic track data at the same time. The basic track data includes: track centerline, station location, station boundary, effective station length, turnout section marking, small radius curve section marking, and mileage-coordinate conversion relationship. S1-2 Establish the spatial reference required for map display: One of the following two methods is adopted: The first method is geographic coordinate mapping: When the detection results contain latitude and longitude information or can be converted from mileage and route coordinate tables, the abnormal events, stations and inter-station boundaries are directly projected onto the electronic map or GIS (Geographic Information System or Geo-Information system, sometimes also called "geoscience information system") base map to form a real geographic space display; The second method is the route schematic mapping method: when the system does not provide latitude and longitude information, the route mileage is used as a one-dimensional basic coordinate, the route centerline is abstracted as a schematic broken line, each station is mapped as a node in mileage order, and abnormal events are mapped on the route schematic diagram according to their mileage location. Although this method does not reflect the precise geographical location, it can maintain the station order, the length ratio between stations, and the distribution relationship of abnormalities. In this embodiment, when GIS conditions are available, the first method is preferred as the basic implementation method. S2 Abnormal Event Extraction In this embodiment, map visualization does not directly display all section data, but preferably extracts abnormal event points from the section-level detection results first, so as to reduce drawing redundancy and enhance the expression of key information; The specific steps are as follows: Step S2-1: read the detection results of each section or other preset section length units, and extract the maximum absolute value of car body vertical acceleration, the speed-corrected abnormality evaluation value and the information of the inter-station section to which they belong; Step S2-2: screen abnormal events according to preset thresholds, set a first-level abnormal threshold (T1) and a second-level abnormal threshold (T2), wherein T1> T2; when the maximum absolute acceleration of the section or the corrected abnormality evaluation value is greater than or equal to T2, the section is identified as an abnormal candidate event; when the value is greater than or equal to T1, the section is identified as a high-level abnormal event; Step S2-3: eliminate special sections not included in inter-station statistics, preferably remove switch sections, small radius curve sections, construction interference sections and missing data sections; Step S2-4: generate spatial attributes for each abnormal event: the spatial attributes include: event number, affiliated line, affiliated inter-station section, event mileage, abnormality level, event time, display color and display symbol; when the system has longitude and latitude data, the longitude and latitude coordinate fields can also be included; S3 Map Layer Generation In this embodiment, the map display is preferably composed of three parts: a basic line layer, a station layer and an abnormal event layer; S3-1 Basic line layer generation: generate a basic line layer according to the line centerline or line schematic data; represent the line track with a blue broken line or a continuous blue line; S3-2 Station layer generation: read the position of station nodes, and overlay and display the station layer on the basic line layer; stations are represented by light blue squares, rectangles or station icons; If it is a schematic map, the stations are arranged uniformly or proportionally on the line according to the mileage sequence; if it is a GIS map, the stations are arranged according to real coordinates and marked with light blue squares; S3-3 Abnormal event layer generation: overlay the abnormal events extracted in step S2 onto the line layer according to their spatial positions; the following color grading method is adopted: when the abnormal event satisfies (T1≤V i <T2), it is represented by an orange dot; when the abnormal event satisfies (V i ≥T2), it is represented by a red dot; wherein (V i ) represents the maximum absolute acceleration value of the section or the abnormality evaluation value of the i-th abnormal event; S3-4 sets the point size, transparency, and display level: the higher the degree of anomaly, the larger the displayed point or the darker the color; when multiple events occur in spatially close locations, aggregated display, transparent overlay, or slight offset display can be used to avoid occlusion problems caused by point overlap; S4 Hotspot Area Identification and Map Zoom-in Display To enhance the practicality of the project, the functions of automatic hotspot section identification and map zoom-in display have been further added. The specific steps are as follows: Step S4-1: Based on the inter-station normalized high acceleration frequency index (Q) i Sort all stations and select the top few stations as hotspot stations. Step S4-2: Within the hot spot station section, statistically analyze the clustering distribution of abnormal event points along the mileage. If orange or red abnormal points repeatedly appear within a certain local mileage range in multiple detection cycles, then that local range is defined as a hot spot sub-segment. Step S4-3: Generate a local magnified map for hotspot stations or hotspot sub-segments: The local magnified map preferably includes mileage scales, event point density display, anomaly level legends, and a jump entry to the segment heatmap in Example 1. This allows users to access the map directly from... Figure 8 The spatial distribution of the lines shown will jump directly to... Figure 7 The detailed distribution map of the sections enables the linkage analysis of "spatial positioning - detailed section inspection"; S5 Map Interaction and Linked Display This embodiment preferably uses an interactive map display method, rather than being limited to static image output. To this end, the following interactive functions can be set: Step S5-1: Threshold filtering function: Users can adjust (T1) and (T2) according to different management needs, and the orange and red dots on the map are updated in real time; Step S5-2: Time Filtering Function: Users can filter and display abnormal events by time range such as single day, single month, quarter, and year, so as to observe the spatial distribution under different time scales; Step S5-3: Direction filtering function: Displays uplink, downlink, or bidirectional data separately to avoid confusion between spatial events under different operating directions; Step S5-4: Attribute pop-up function: After the user clicks on any anomaly point, the detailed attributes of the event are displayed, including the detection date, mileage, maximum absolute acceleration, speed correction result, station to which it belongs, direction of operation and remarks; Step S5-5: Multi-map linkage function: The hot spot station sections in the map are linked with the track status map in Example 1 and the long-term change map in Example 3; that is, after the user clicks on the key section in the map, the color-coded track status map and long-term tracking detection map of that section can be directly brought up. S6 Result Output After the map generation is completed, the system outputs at least one mapping result; preferably, the output results include: a normalized high acceleration frequency statistics map between stations, a distribution map of line-station-anomaly points, a highlighted map of hot inter-station sections, a magnified map of key sections, and a list of spatial distribution of abnormal events; The output format can be PNG images, PDF report diagrams, web page GIS layers, desktop software interactive interfaces, or dynamic chart objects embedded in the operation and maintenance platform; furthermore, the system can also automatically generate text descriptions, such as "The normalized frequency of high acceleration events between XX stations is the highest, and it is recommended to carry out detailed section diagnosis" to assist operation and maintenance decisions.

[0026] Compared with methods that only output inter-station tables or lists of outliers, this embodiment has the following technical advantages: First, by mapping the normalized statistical results between stations and the spatial line distribution results into the same visualization system, information at both the "risk order between stations" and "specific location of anomalies" levels can be reflected simultaneously. Secondly, by displaying "blue lines on the line + light blue squares at stations + orange / red anomalies", the distribution density, grade differences and clustering locations of anomalies on the line can be intuitively revealed, improving the identification efficiency of key sections. Third, by normalizing the inter-station data, the impact of differences in the length between different stations on the statistical results of high acceleration events was eliminated, making the risks between stations comparable. Fourth, by magnifying hotspot sections and linking them with section heat maps and long-term trend maps, a continuous and visualized chain of "inter-station screening - spatial positioning - detailed section inspection - long-term tracking" can be formed, providing a more intuitive basis for subsequent maintenance decisions.

[0027] Example 3: A Long-Term Comparative Visualization Method for Track Acoustic and Vibration Anomaly Detection Results This embodiment provides a long-term comparative visualization method for track acoustic and vibration anomaly detection results. It organizes, compares, and graphically presents the section detection results generated by the track acoustic and vibration anomaly detection system over multiple detection cycles, thereby displaying the vibration state change pattern of the same section during long-term operation. It is applicable to long-term monitoring data in continuous detection scenarios for in-service trains, and also to comparative analysis scenarios before and after maintenance operations. The system can continuously record vehicle vibration and position information, and uses the vehicle's vertical acceleration as a state evaluation quantity. Through long-term statistics, it reveals the trend of track condition deterioration and quantifies the maintenance effect. In this embodiment, the long-term comparative visualization method includes at least eight processes: long-term data reading and unification, time-series segmentation organization, long-term evaluation index construction, maintenance event alignment, long-term change graph generation, maintenance effect comparative analysis, re-deterioration trend identification, and result output, as detailed below: S1 Long-Term Data Reading and Unification S1-1 reads historical detection results data generated by the track acoustic and vibration anomaly detection system over multiple detection dates, multiple train trips, or multiple calendar months, including: detection date, train number, direction of travel, line name, station name, section start mileage, section end mileage, section maximum vertical acceleration, section average vertical acceleration, speed-corrected anomaly evaluation value, anomaly level marker, and the key section number to which the section belongs; it also reads maintenance record data associated with key sections, including: maintenance date, maintenance type, maintenance section range, and remarks. S1-2 performs unified processing on long-term historical data: data from different testing cycles are aligned according to a unified mileage benchmark; when there are slight deviations in the start and end mileage of segments in different batches of data, they are mapped to preset standard segments; when there are missing records, abnormal interruption records, or special working condition records in the data, it is preferable to mark, remove, or separately classify and display null values; this ensures that the same segment under different dates is directly comparable. S2 temporal segmentation organization In this embodiment, in order to achieve Figure 10 The long-term comparison results shown are achieved by first segmenting key sections in both the time and spatial dimensions. The specific steps are as follows: Step S2-1: Determine the target long-term tracking section. Based on the section heat map, inter-station statistical chart, or list of sections with frequent anomalies, select the key mileage sections that need to be tracked for a long time. Step S2-2: Re-divide the long-term tracking segment of the target by dividing the target segment into multiple sub-segments of fixed length, with each sub-segment serving as an independent long-term comparison unit; Step S2-3: Establish a two-dimensional data structure of "time-sub-segment". For each sub-segment, record the corresponding maximum vertical acceleration value, average acceleration value and corrected anomaly evaluation value under each detection date, thereby forming a long-term monitoring sequence sorted by time. In a preferred embodiment, if multiple trains pass through the same section on the same day, the maximum, average, or median value of that day can be taken as the representative value for that day; if there is no valid detection on that day, that moment is recorded as a null value, and the graph is left blank or connected with dashed lines. S3 Long-Term Evaluation Indicator Construction To enable visualization of long-term comparisons, displaying not only changes in original values ​​but also supporting subsequent evaluation of maintenance effectiveness and re-deterioration analysis, this embodiment preferably constructs at least the following types of long-term evaluation indicators: The first category is the long-term peak indicator, denoted as (P m,t ), representing the maximum vertical acceleration value of the (m)th sub-segment on the (t)th detection date; The second category is long-term average indicators, denoted as (E m,t ), representing the average maximum acceleration or average outlier of the (m)th sub-segment on the (t)th detection date, used to reduce the impact of single extreme value fluctuations; The third category is the corrected long-term abnormal indicator, denoted as (R m,t The value represents the long-term anomaly evaluation value after considering the speed effect correction, which is used to more accurately reflect the changes in the actual track condition. Since the vibration of the vehicle body is significantly affected by the operating speed, Mahalanobis distance is introduced to identify outliers, and the speed-acceleration regression relationship is used to achieve explicit correction of the speed effect. The fourth category is long-term volatility indicators, denoted as (V m This is used to describe the degree of fluctuation of the same sub-segment within an observation period; it is preferred to use standard deviation, range, or sliding window fluctuation amplitude to represent it; the greater the fluctuation, the more unstable the state of the segment; The fifth category is the trend slope indicator, denoted as (K). m ), used to represent the changing trend of the same sub-segment over a period of time; preferably by using (P) m,t ) or (R m,t Linear fitting is performed to obtain the result; when (K) m When K > 0), it indicates that the overall level of abnormality is rising and there is a trend of further deterioration; when (K m When <0), it indicates that the overall situation has improved; S4 Maintenance Event Alignment In long-term comparative visualization, maintenance events are key explanatory information; Figure 9 The red dashed line indicates the time for track maintenance operations, and the maintenance effect is observed by the changes in acceleration before and after maintenance. Therefore, this embodiment preferably includes a maintenance event alignment step: Step S4-1: Read the maintenance record table and extract the time, section range, and operation type of each maintenance operation; Step S4-2: Map maintenance events onto the long-term monitoring timeline; for maintenance events falling within the target long-term tracking range, generate maintenance markers at the corresponding time positions; preferably, they are represented in the long-term change graph by red dashed lines, red vertical lines, red bands, or maintenance icons; Step S4-3: If multiple maintenance operations occur in the same section, they are all displayed in chronological order, and each maintenance operation is assigned a unique number to facilitate subsequent comparison of the effects of different maintenance measures. In a preferred embodiment, the maintenance marker not only displays the time and location, but can also include maintenance instructions, such as "tamping, grinding, retesting, temporary repairs," etc. Users can click to view detailed work records. S5 Long-Term Change Chart Generation In this embodiment, the core graphic is a long-term change comparison chart, and its graphic organization method is preferably in accordance with... Figure 9 The specific steps are as follows: Step S5-1: Plot time on the x-axis and the long-term peak index (P) on the y-axis. m,t ) or the corrected long-term abnormality index (R m,t Using the vertical axis as the ordinate, plot a long-term line graph for each sub-segment; preferably, use lines to connect data points under each detection date, with detection points represented by solid circles, squares, or triangles; Step S5-2: Overlay maintenance event markers on the sub-graph; preferably, use a red dashed line to run through the entire height of the sub-graph, and add text related to the maintenance event type at the top or bottom; Step S5-3: Set a uniform ordinate range: In order to facilitate horizontal comparison between different sub-segments, it is preferable to use a uniform ordinate scale for all sub-plots within the same key segment; when the differences between different sub-segments are too large, a hybrid approach of consistent principal axis and local adaptive scaling can also be used. Step S5-4: Set the baseline reference line; add a health reference line, alarm threshold line, and severity threshold line to each subgraph to display the range of abnormality at each time point; By following the steps above, you can generate a [database name - likely a string of characters]. Figure 9 Similar long-term change comparison charts show the state evolution of key sections across multiple detection cycles; Comparative Analysis of S6 Maintenance Effects To ensure that the long-term change map not only has a display function but also has engineering evaluation capabilities, this embodiment further quantifies and compares the state before and after maintenance. The specific steps are as follows: Step S6-1: Set up a pre-maintenance window and a post-maintenance window for each maintenance event; the pre-maintenance window contains a number of valid checks before the maintenance date, and the post-maintenance window contains a number of valid checks after the maintenance date; Step S6-2: Calculate the average peak value before maintenance (P) m pre ) and the average peak value after maintenance (P) m post ), and construct maintenance improvement indicators:

[0028] Among them, (I) m The larger the value, the more obvious the improvement effect of maintenance; Step S6-3: If (I m If the value of (I) is greater than a preset threshold, then the condition of the sub-section is determined to have significantly improved after maintenance; if (I) m If the value is close to zero, the maintenance effect is considered insignificant; if (I) mIf <0, it indicates that higher vibrations occur in the short term after maintenance, which should be marked as an abnormal maintenance effect event. Step S6-4: Since the vertical acceleration of some sections decreased significantly after maintenance, but the acceleration of some sections still increased significantly in the short term after maintenance, this embodiment further sets up a re-deterioration identification process to determine whether the section has deteriorated again after maintenance. Step S6-4-1: Based on the post-maintenance window, analyze the long-term peak index (P) after maintenance. m,t ) or the corrected long-term abnormality index (R m,t Trend fitting was performed to obtain the trend slope (P) after maintenance. m post ); Step S6-4-2: If (P) m post If the value is greater than 0 and continues to exceed the preset time window, then the sub-segment is determined to have a re-deterioration trend; if it eventually approaches or reaches the pre-maintenance average level (P), then the sub-segment is considered to have a re-deterioration trend. m pre If the condition is as described above, it is considered a relapsed segment that has "returned to the pre-maintenance level". Step S6-4-3: For sections with multiple maintenance records, compare the rate of re-deterioration, recurrence time, and maximum rebound amplitude after different maintenance events to determine whether the section is "prone to recurrence, repetitive, and persistent." The original study gave such conclusions for some locations; In a preferred embodiment, the re-deterioration time index (T) can also be calculated. m re This indicates the time it takes for the abnormal level to exceed the warning threshold again after maintenance; the shorter the value, the faster the recurrence in that segment. S7 Interactive Long-Term Comparison Display To improve the efficiency of graphic explanation, this embodiment preferably adopts an interactive long-term visualization method, rather than being limited to static image output; specifically including: Step S7-1: Time range switching: Supports switching the display by the last 1 month, last 3 months, last 6 months, last 1 year or the entire period; Step S7-2: Sub-segment switching: Users can select a single sub-segment to view long-term changes, or display multiple adjacent sub-segments simultaneously for horizontal comparison; Step S7-3: View Maintenance Events: After clicking on any red dotted line, the user can view the corresponding maintenance type, implementation time, and maintenance instructions. Step S7-4: Peak Point Details Pop-up: When a user clicks on any point in the long-term variation graph, the date, train number, maximum vertical acceleration, speed correction value, and anomaly level of the detection will be displayed. Step S7-5: Link with other graphics: The long-term change map is linked with the track status heat map in Example 1 and the map hotspot section in Example 2. When the user selects a key location in the heat map or map, the system automatically opens the corresponding long-term comparison map. S8 Result Output After completing the generation and comparative analysis of long-term comparison charts, the system outputs at least one long-term visualization result; preferably, the output results include: an overview of long-term changes in key sections, a long-term comparison sub-chart group divided by sub-sections, a comparison chart of the state before and after maintenance, a re-deterioration trend analysis chart, and a long-term monitoring and analysis report; The long-term monitoring and analysis report includes: key section name, sub-section number, peak value change trend, maintenance effect judgment, whether there is further deterioration, recommended review time and recommended maintenance level. This method can not only be used for hazard location, but also for maintenance effect assessment and further deterioration early warning, providing a quantitative basis for formulating more refined maintenance schedules and resource allocation.

[0029] Technical effects of this embodiment Compared with methods that only output a single detection image or a single numerical table, this embodiment has the following technical advantages: First, it can continuously display the vibration state changes of the same section in a time series manner, revealing the long-term deterioration trend of the line condition; related studies have clearly pointed out that long-term time evolution analysis can reveal the laws of line condition changes. Second, it can directly overlay maintenance event times on the graph, making the changes in state before and after maintenance visually comparable, thereby achieving a quantitative assessment of maintenance effectiveness; Third, it can identify sections that gradually increase in size after maintenance and eventually return to the pre-maintenance level, thereby providing early warning of disease recurrence and further deterioration; Fourth, it can be combined with section heat maps and map hotspot maps to form a complete visualization chain of "overview-location-tracking", improving the operation and maintenance department's ability to continuously track key diseased sections.

[0030] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A visualization method for detecting track acoustic and vibration anomalies, comprising the following steps: S1 reads the detection result data output by the track acoustic and vibration anomaly detection system and organizes the detection result data in a structured manner. The detection result data includes the detection date, train number, direction of operation, station name, section start mileage, section end mileage, operating speed, original vibration index, speed-corrected anomaly evaluation value, anomaly judgment result, and section statistical results. S2 Based on the detection result data, construct display indicators, which include original vibration display indicators, velocity-corrected display indicators, and segment statistical display indicators; S3 Based on the original vibration display index and the speed-corrected display index, generate a comparison chart of the track state before and after speed correction; S4 generates a normalized statistical chart of inter-stations and a color-coded track status chart of inter-station sections based on the inter-station statistical results. S5 maps the results of abnormal events onto a route map or route diagram, generating a route-station-abnormal point distribution map. S6 performs long-term comparative analysis of historical detection results within multiple detection cycles of the same section, generating a long-term change comparison chart; S7 outputs at least one visualization result for the identification, location, tracking, and maintenance support decision-making of abnormal track sections.

2. The visualization method for detecting track acoustic vibration anomalies according to claim 1, characterized in that, In step S1, the detection result data is organized in a structured manner, specifically including the following: The S11 section basic table is used to record the starting mileage location, the station to which each section belongs, and the inspection date; The S12 vibration index table is used to record the original vibration value, corrected vibration value, maximum absolute value, and whether the threshold is exceeded for each section. S13 Statistical Results Table is used to record the frequency of exceeding the threshold, average outlier and maximum outlier of the section in multiple detection cycles; S14 Key Section Table is used to record high-vibration-frequent sections, continuous abnormal sections, and their local magnified display range.

3. The visualization method for detecting track acoustic vibration anomalies according to claim 1, characterized in that, In step S2, the construction of display metrics specifically includes the following: S21 uses Mahalanobis distance to identify and remove outliers in a two-dimensional feature space composed of operating speed and vehicle vibration index. S22 Based on the data after removing outliers, the least squares method was used to fit the velocity-vibration regression relationship; S23 Based on the regression relationship, predict the expected vibration level under the same velocity conditions; S24 The ratio of the measured vibration value to the expected vibration level or the standardized residual is used as the velocity-corrected anomaly evaluation value.

4. The visualization method for detecting track acoustic vibration anomalies according to claim 1, characterized in that, In step S4, the process of forming the inter-station normalized statistical chart is as follows: S411 Count the number of abnormal events exceeding a preset threshold within the section between two adjacent stations; S412 retrieves the valid length between the corresponding stations; S413 Calculate the normalized high acceleration frequency index between stations based on the ratio of the number of abnormal events to the effective length between stations; S414 Based on the inter-station normalized high acceleration frequency index, sort and highlight different inter-station sections.

5. The visualization method for detecting track acoustic vibration anomalies according to claim 1, characterized in that, In step S4, the process of forming the color-coded track status diagram for the inter-station section is as follows: S421 Constructs a two-dimensional data matrix of "time-mileage-segment value" according to the time dimension and the mileage dimension; S422 generates a track status heatmap with mileage as the horizontal axis, detection date or detection cycle as the vertical axis, and color coding based on the maximum absolute vibration value or anomaly evaluation value of the section. S423 statistically analyzes the average value of the maximum vibration in each section over multiple detection cycles to generate a distribution map of the average value of the maximum vibration, which shares the same mileage coordinate axis as the heat map, i.e., a color-coded track status map of the inter-station section.

6. The visualization method for detecting track acoustic vibration anomalies according to claim 1, characterized in that, In step S5, the specific process of mapping the abnormal event results onto the route map or route diagram is as follows: S511 When the detection results have geographic coordinate information, the abnormal events, stations and line boundaries will be projected onto the electronic map or GIS base map; S512 When the detection results do not have geographic coordinate information, construct a route diagram based on the route mileage, and map the stations and abnormal events onto the route diagram according to their mileage locations. S513 Abnormal events are divided into at least two levels based on preset abnormal thresholds and displayed with different colors or different symbols.

7. The visualization method for detecting track acoustic vibration anomalies according to claim 1, characterized in that, Step S5 also includes hotspot segment identification and local magnification display steps, as detailed below: S521 Based on the normalized high acceleration frequency index between stations, the top few stations are selected as hotspot stations; S522 Statistical analysis of the clustered distribution of abnormal events along the mileage within the hotspot station section; S523 defines a local mileage range in which abnormal events repeatedly occur within multiple detection cycles as a hot spot sub-segment; S524 and generate a locally enlarged display image for the hot spot station section or hot spot sub-section.

8. The visualization method for detecting track acoustic vibration anomalies according to claim 1, characterized in that, In step S6, the process of forming the long-term change comparison chart is as follows: S61 aligns historical test results for the same key section across multiple test cycles according to a unified mileage benchmark. S62 divides the target section into multiple sub-segments and establishes a two-dimensional monitoring sequence of time and sub-segments; S63 Based on the two-dimensional monitoring sequence, construct at least one of the following: long-term peak index, long-term average index, corrected long-term anomaly index, long-term fluctuation index, and trend slope index. S64 maps maintenance event times to a long-term change graph; S65 obtains the evaluation results of maintenance effect and the results of re-deterioration identification by comparing the changes of vibration index in the pre-maintenance window and the post-maintenance window.

9. A track acoustic vibration anomaly detection and visualization system, characterized in that, The system includes a processor and a memory, the memory storing program instructions executable by the processor, which, when executed by the processor, cause the system to perform the visualization method for detecting track acoustic vibration anomalies as described in any one of claims 1 to 8.

Images