A Dynamic Detection Method for Metro Overhead Contact System Based on Multi-Sensor Collaboration

Abstract

This application provides a multi-sensor collaborative dynamic detection method for subway catenary overhead contact lines, relating to the field of data processing. In this method, multi-source detection data corresponding to the subway catenary overhead contact lines are acquired, the sequence of changes from multiple sensors is extracted and compared with the normal sequence of changes, and a multi-sensor synchronous abnormal state is determined when changes in visible light, laser, infrared, and ultraviolet light are synchronously enhanced while changes in pantograph-catenary contact current and vehicle posture are not synchronously enhanced. Based on the duration of the abnormality, mobile abnormal states and fixed abnormal states are distinguished, detection interference and catenary defects are identified respectively, and corresponding processing strategies are generated to update the detection data and output the dynamic detection results of the subway catenary overhead contact lines. Implementing the technical solution provided in this application facilitates improved accuracy in the dynamic detection of subway catenary overhead contact lines.

Description

Technical Field

[0001] This application relates to the field of data processing technology, specifically to a dynamic detection method for subway overhead contact lines based on multi-sensor collaboration. Background Technology

[0002] With the increasing density and speed of subway lines, the subway overhead contact system is constantly exposed to high-frequency current collection, high humidity, strong vibration, and complex electromagnetic environments. Insulators, segmented insulators, and connecting components within the contact system are prone to localized insulation performance degradation during long-term operation. Existing dynamic inspection systems for subway overhead contact systems typically employ a combination of sensors, including visible light detection, infrared thermal imaging, and ultraviolet discharge detection, to jointly monitor the appearance, temperature, and discharge status of the contact system, thereby improving the ability to identify defects.

[0003] Existing multi-sensor collaborative detection methods for subway catenary systems mostly employ independent judgment by a single sensor or direct superposition of results from multiple sensors, lacking a joint analysis mechanism for the correlation between different spectral bands. When the local insulation performance of the catenary deteriorates, weak corona discharge is easily generated in high humidity environments or under catenary voltage fluctuations. This weak corona discharge usually does not form obvious structural anomalies in the visible light band, but it will form local ultraviolet bright spots in the ultraviolet band and local thermal anomalies in the infrared band. Due to the lack of obvious defects in visible light images, existing detection systems easily identify infrared thermal anomalies as environmental thermal noise, water mist interference, or equipment errors, while classifying ultraviolet bright spots as occasional discharge noise, thus leading to perception conflicts between different spectral bands. Furthermore, existing systems rely heavily on structural defect features in visible light images, lacking continuous analysis of the correlation between changes in ultraviolet bright spots and infrared temperature changes, and cannot uniformly correlate weak visible light response states, changes in ultraviolet bright spots, and infrared temperature changes, easily overlooking early insulation aging problems in the catenary. Therefore, the above methods result in low accuracy in the dynamic detection of subway catenary systems.

[0004] Therefore, there is an urgent need for a dynamic detection method for subway overhead contact lines based on multi-sensor collaboration. Summary of the Invention

[0005] This application provides a method for dynamic detection of subway catenary based on multi-sensor collaboration, which facilitates the improvement of the accuracy of dynamic detection of subway catenary.

[0006] The first aspect of this application provides a dynamic detection method for subway catenary based on multi-sensor collaboration. The method includes: acquiring visible light images, laser contour data, infrared thermal images, ultraviolet discharge images, pantograph-catenary contact current data, vehicle posture data, train running position data, and catenary structure information corresponding to the location of the subway catenary, to form detection data; extracting the multi-sensor change sequence corresponding to the location of the subway catenary based on the detection data; comparing the multi-sensor change sequence with the normal change sequence formed during historical normal detection processes, and determining the corresponding subway when visible light edge changes, laser contour changes, infrared temperature changes, and ultraviolet bright spot changes are simultaneously enhanced, while pantograph-catenary contact current changes and vehicle posture changes are not synchronously enhanced. Multiple sensors exhibit synchronous anomalies at the contact wire location. Based on the persistence of these anomalies between adjacent contact wire locations, mobile and fixed anomalies are identified. Mobile anomalies are used to identify detection equipment attitude interference, detection window contamination, and tunnel airflow interference. Fixed anomalies are used to identify insulation aging, contact wire structure loosening, and hard point anomalies. Corresponding processing strategies are generated based on these factors, and the detection data is updated accordingly. The updated detection data is then used to generate corresponding dynamic detection results for the subway contact wire.

[0007] A second aspect of this application provides a dynamic detection device for subway catenary based on multi-sensor collaboration. The device includes an acquisition module and a processing module. The acquisition module is used to acquire visible light images, laser contour data, infrared thermal images, ultraviolet discharge images, pantograph-catenary contact current data, vehicle posture data, train running position data, and catenary structure information corresponding to the location of the subway catenary, to form detection data. The processing module is used to extract the multi-sensor change sequence corresponding to the location of the subway catenary based on the detection data. The processing module is also used to compare the multi-sensor change sequence with the normal change sequence formed during historical normal detection processes, and to determine whether the visible light edge change, laser contour change, infrared temperature change, and ultraviolet bright spot change are simultaneously enhanced, while the pantograph-catenary contact current change and vehicle posture change are not synchronously enhanced. The processing module identifies multiple sensor synchronization anomalies at corresponding subway catenary locations. It further determines moving and fixed anomalies based on the duration of these anomalies between adjacent catenary locations. The module also identifies moving anomalies such as detection equipment posture interference, detection window contamination, and tunnel airflow interference, and fixed anomalies such as insulation aging, catenary structure loosening, and hard point anomalies. Finally, the module generates corresponding processing strategies based on these factors, and updates the detection data accordingly to generate corresponding dynamic detection results for the subway catenary.

[0008] A third aspect of this application provides an electronic device comprising a processor and a memory; the memory having a stored computer program, wherein the computer program, when executed by the processor, implements the multi-sensor collaborative dynamic detection method for subway overhead contact lines as described above.

[0009] In a fourth aspect of this application, a non-transitory computer-readable storage medium is provided, the non-transitory computer-readable storage medium storing instructions that, when executed, perform the multi-sensor collaborative dynamic detection method for subway catenary as described above.

[0010] In summary, one or more technical solutions provided in this application have at least the following technical effects or advantages: By acquiring visible light images, laser contour data, infrared thermal images, ultraviolet discharge images, pantograph-catenary contact current data, and vehicle attitude data, and combining this with train operation position data and catenary structure information to form unified detection data, different sensor data can be correlated around the same subway catenary location, improving the positional consistency of multi-sensor detection results. By constructing a multi-sensor change sequence and comparing it with a normal change sequence, the catenary anomaly judgment shifts from traditional static result judgment to dynamic change process judgment, improving the identification capability in complex anomaly scenarios. Simultaneously, by introducing pantograph-catenary contact current changes and vehicle attitude changes as reverse verification criteria, it can identify pseudo-consistent anomalies caused by multiple sensors being simultaneously affected by external interference. Furthermore, by distinguishing between moving and fixed anomaly states, it effectively separates detection equipment attitude interference, detection window contamination, and tunnel airflow interference from actual catenary defects, thereby improving the accuracy of anomaly identification. In addition, this technical solution can also identify early insulation aging states through ultraviolet bright spot changes and infrared temperature changes when no obvious structural defects are formed in the visible light image, improving the early defect detection capability and operational safety of the subway catenary. Therefore, it facilitates improving the accuracy of dynamic detection of the subway catenary. Attached Figure Description

[0011] Figure 1 A flowchart illustrating a dynamic detection method for subway overhead contact lines based on multi-sensor collaboration, provided in an embodiment of this application; Figure 2 A schematic diagram of a dynamic detection device for subway catenary based on multi-sensor collaboration, provided for an embodiment of this application; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0012] Explanation of reference numerals in the attached figures: 21. Acquisition module; 22. Processing module; 31. Processor; 32. Communication bus; 33. User interface; 34. Network interface; 35. Memory. Detailed Implementation

[0013] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0014] In the description of the embodiments of this application, the words "for example" or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design that is described as "for example" or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design options. Rather, the use of the words "for example" or "for instance" is intended to present the relevant concepts in a specific manner.

[0015] In the description of the embodiments of this application, the term "multiple" means two or more. For example, multiple systems means two or more systems, and multiple screen terminals means two or more screen terminals. In addition, the terms "comprising," "including," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0016] To address the aforementioned technical problems, this application provides a dynamic detection method for subway overhead contact lines based on multi-sensor collaboration, referring to... Figure 1 , Figure 1 This is a flowchart illustrating a multi-sensor collaborative dynamic detection method for subway overhead contact lines, provided as an embodiment of this application. The method is applied to a server and includes steps S110 to S160, as follows:

[0017] S110. Acquire visible light images, laser contour data, infrared thermal images, ultraviolet discharge images, pantograph-catenary contact current data, vehicle attitude data, train running position data, and catenary structure information corresponding to the location of the subway catenary to form detection data.

[0018] Specifically, the server first reads the installation location, component number, line mileage, and connection relationship of the contact wire, droppers, positioners, insulators, segment insulators, anchor joints, and supporting devices from the contact wire structure information. The contact wire structure information refers to pre-established contact wire basic ledger data, used to describe the spatial distribution and interconnection relationships of each contact wire component in the subway line; the installation location refers to the installation coordinates of the contact wire component in the detection coordinate system or line coordinate system; the component number is a code used to uniquely identify the contact wire component; the line mileage refers to the location marker of the contact wire component along the subway line direction; and the connection relationship refers to the physical connection, support connection, or adjacent relationship between the contact wire, droppers, positioners, insulators, segment insulators, anchor joints, and supporting devices. After the server reads the overhead contact system structure information, it arranges each overhead contact system component in ascending order of line mileage and establishes a corresponding subway overhead contact system location for each component. This allows the subsequently collected visible light images, laser contour data, infrared thermal images, ultraviolet discharge images, pantograph-catenary contact current data, and vehicle attitude data to be aggregated based on the same subway overhead contact system location, rather than simply stored according to the collection time.

[0019] When the server binds the overhead contact line structure information with the train's running position data, it first reads the train's running position data corresponding to the current moment. This train running position data refers to the real-time position of the detected train as it travels along the subway line, which can be obtained from onboard positioning devices, odometers, transponders, track section positioning information, or inertial positioning information. The server converts the train running position data into subway line mileage and matches it with the line mileage in the overhead contact line structure information to determine the current subway overhead contact line position of the detected train. To avoid short-term fluctuations in the train running position data, the server can perform smoothing processing by combining train running position data from adjacent moments, ensuring that the current subway line mileage of the detected train remains continuously changing. The smoothing calculation formula for the subway line mileage is:

[0020] in, This represents the smoothed mileage of the subway line at the current moment. This indicates the original subway line mileage directly calculated from the train's operating position data at the current moment. This parameter is obtained through onboard positioning devices, odometers, transponders, or inertial positioning information. This represents the subway line mileage after smoothing at the previous moment. This parameter is saved by the server at the previous detection moment. This represents the smoothing coefficient, with a value greater than 0 and less than or equal to 1. This parameter can be preset based on train speed and positioning data stability. It takes a larger value when train position data is stable and a smaller value when train position data fluctuates significantly. This formula weakens the impact of single positioning jumps on the subway catenary position matching results by weighted fusion of the current original subway line mileage and the subway line mileage at the previous moment.

[0021] After determining the mileage of the subway line currently being inspected by the inspection train based on the train's operating location data, the server performs distance matching between the subway line mileage and the line mileage of each contact wire component in the contact wire structure information. The location of the contact wire component closest to the inspection train and within a preset matching range is then identified as the corresponding subway contact wire location. The subway contact wire location is not simply a spatial coordinate, but a location object defined by the line mileage, contact wire component number, and contact wire component type, used to represent the contact wire structure area currently being inspected or about to be inspected by the inspection train. When multiple contact wire components have similar line mileages, the server further combines connection relationships and installation locations to determine the main component corresponding to the current subway contact wire location, ensuring that contact wires, droppers, positioners, insulators, segment insulators, anchor joints, and support devices are accurately assigned according to their actual structural relationships. The matching expression for the subway contact wire location is:

[0022] in, This indicates the location of the subway overhead contact line matched at the current moment; This represents the pre-established set of all subway catenary locations in the catenary structure information. This set consists of the location objects corresponding to the contact wire, droppers, positioners, insulators, segment insulators, anchor joints, and support devices. Represents the first position in the set of all subway catenary locations. Location of the subway overhead contact line; This indicates the mileage of the subway line corresponding to the detected train at the current moment; Indicates the first The line mileage corresponding to each subway catenary location is read from the catenary structure information. This expression compares the difference between the current subway line mileage of the detected train and the line mileage corresponding to each subway catenary location, and selects the subway catenary location with the smallest difference as the current matching result.

[0023] After determining the location of the subway catenary, the server controls or receives various sensors to collect visible light images, laser contour data, infrared thermal images, ultraviolet discharge images, pantograph-catenary contact current data, and vehicle attitude data. Visible light images, captured by visible light cameras, record the appearance of the catenary, including the contact wire edge, dropper shape, locator connection status, insulator surface condition, segmented insulator appearance, and support device connection status. Laser contour data, collected by laser scanning devices, records the spatial contour points or lines of the catenary, showing the spatial position and geometry of catenary components. Infrared thermal images, captured by infrared thermal imaging devices, record local temperature changes around catenary components. Ultraviolet discharge images, captured by ultraviolet imaging devices, record discharge response images, showing weak corona discharges or ultraviolet bright spots formed by partial discharges. Pantograph-catenary contact current data reflects the current changes between the pantograph and the contact wire, indicating the pantograph-catenary contact status. Vehicle attitude data detects pitch, roll, vertical vibration, and lateral vibration changes in the train body, reflecting the motion of the platform on which the detection equipment is located.

[0024] When the server binds a visible light image to the corresponding subway catenary location, it records the acquisition time, camera number, exposure status, image clarity, and the corresponding subway catenary location of the visible light image, and associates the image coordinates in the visible light image with the subway catenary location. The server can extract the contact wire edge region, dropper region, locator region, insulator region, segmented insulator region, anchor joint region, and support device region from the visible light image, and write a corresponding component number for each region. This allows for accurate identification of the subway catenary location from which the visible light edge changes originate during subsequent extraction. The purpose of visible light image binding is to transform the two-dimensional visual information in the image into detection data with subway catenary location attributes, thereby preventing multiple catenary components from being mixed together in the same frame and failing to be accurately assigned.

[0025] When the server binds the laser contour data to the corresponding subway catenary location, it records the scanning time, scanning angle, echo intensity, contour point coordinates, and the corresponding subway catenary location of the laser scanning device. It then performs coordinate correction on the laser contour data based on the current subway line mileage and vehicle attitude data of the detected train. The contour points in the laser contour data are typically initially in the sensor coordinate system and need to be transformed to the line coordinate system or catenary detection coordinate system based on the detected train's position and vehicle attitude, so that the laser contour data can be compared with the installation location in the catenary structure information. The coordinate transformation expression for the laser contour data is:

[0026] in, This represents the coordinates of the transformed contour points; This indicates the coordinates of the contour points acquired in the sensor coordinate system in the laser contour data. This parameter is directly output by the laser scanning device. This represents the attitude rotation matrix composed of vehicle attitude data; This represents the attitude angle corresponding to the pitch change, and this parameter is obtained from the vehicle attitude data. This represents the attitude angle corresponding to the roll change, and this parameter is obtained from the vehicle attitude data. This indicates the attitude angle corresponding to the change in heading; this parameter can be obtained from the train's direction of travel and inertial positioning information. This represents the position translation vector corresponding to the current subway line mileage. This parameter is obtained from the train's operating position data and the mapping relationship between the line coordinate system. This expression transforms the local contour points collected by the laser scanning device into spatial points that can correspond to the contact wire structure information through attitude rotation and translation transformation, thus providing a unified spatial basis for subsequent laser contour change extraction.

[0027] When the server binds an infrared thermal image to the corresponding subway contact network location, it records the acquisition time, infrared imaging area, temperature response area, and corresponding subway contact network location of the infrared thermal image. It also correlates high-temperature areas in the infrared thermal image with insulators, segmented insulators, contact wire connections, positioner connections, and support devices in the contact network structure information. The temperature distribution in the infrared thermal image is used not only to determine the presence of localized thermal anomalies but also to correlate with changes in ultraviolet bright spots in the ultraviolet discharge image. The server can compare the pixel temperature in the infrared thermal image with the surrounding background temperature to obtain the infrared temperature change at the corresponding subway contact network location. This allows the infrared thermal image to no longer be stored as an isolated image but as detection data that can participate in multi-sensor change sequence analysis.

[0028] When the server binds the ultraviolet (UV) discharge image to the corresponding subway contact network location, it records the image acquisition time, UV bright spot location, UV bright spot duration, number of recurrences, and the corresponding subway contact network location. It then associates the UV bright spot with insulators, segment insulators, anchor joints, or contact network connections in the contact network structure information. A UV bright spot refers to a localized response area in the UV discharge image whose brightness is significantly higher than the surrounding background; it is typically used to characterize weak corona discharge or partial discharge. After binding the UV discharge image, the server can determine in subsequent processing whether changes in the UV bright spot are fixed at the same subway contact network location along with changes in infrared temperature, thus providing a basis for identifying insulation aging conditions.

[0029] When the server binds the pantograph-catenary contact current data to the corresponding subway catenary location, it records the sampling time, current fluctuation range, current abrupt change segments, and the corresponding subway catenary location. The pantograph-catenary contact current data is typically continuous time-series data. The server segments this continuous time-series data into current segments corresponding to different subway catenary locations based on train position data, ensuring that each subway catenary location has corresponding pantograph-catenary contact current data. This data is used to determine whether a genuine pantograph-catenary contact disturbance exists when visible light edge changes, laser profile changes, infrared temperature changes, and ultraviolet bright spot changes simultaneously increase. If the pantograph-catenary contact current data does not increase synchronously, it indicates that the corresponding anomaly may not be caused by a genuine catenary structural disturbance, but rather may belong to a pseudo-consistency anomaly in a multi-sensor synchronization anomaly state.

[0030] When the server binds the vehicle attitude data to the corresponding subway catenary location, it records the pitch, roll, vertical vibration, and lateral vibration changes within the vehicle attitude data, as well as the corresponding subway catenary location. The vehicle attitude data characterizes the attitude changes of the inspection train as an inspection platform. When significant changes occur in the vehicle attitude data, the edge positions in the visible light image and the contour positions in the laser profile data may shift due to changes in the observation angle. Therefore, by binding the vehicle attitude data to the subway catenary location, the server can determine in subsequent processing whether changes in visible light edge and laser profile are caused by interference from the inspection equipment's attitude, thus avoiding misjudging measurement shifts caused by the inspection platform's movement as a loose catenary structure or an abnormal hard point.

[0031] The server uses the location of the subway catenary as a unified index to integrate visible light images, laser contour data, infrared thermal images, ultraviolet discharge images, pantograph-catenary contact current data, vehicle attitude data, train running position data, and catenary structure information to form detection data. The unified index means that the server uses the subway catenary location as the primary key when storing and retrieving detection data, enabling simultaneous association of appearance imaging, spatial contours, temperature distribution, discharge response, current response, vehicle attitude response, train running position, and catenary structure information for the same location. The detection data is not a single sensor dataset, but rather a multi-source correlated data unit centered on the subway catenary location, which can be directly used for subsequent extraction of the sequence of changes from multiple sensors. The detection data can be represented as:

[0032] in, Indicates the location of the subway overhead contact line The corresponding test data; Indicates binding to the subway overhead contact line location. The visible light image, whose parameters are acquired by a visible light camera and bound to the subway catenary location by a server; Indicates binding to the subway overhead contact line location. The laser contour data, which is acquired by the laser scanning device and obtained through coordinate transformation and position binding; Indicates binding to the subway overhead contact line location. The infrared thermal image, whose parameters are acquired by the infrared thermal imaging device and bound to the temperature response region; Indicates binding to the subway overhead contact line location. The ultraviolet discharge image, the parameter of which is acquired by the ultraviolet imaging device and obtained by binding the location of the ultraviolet bright spot; Indicates binding to the subway overhead contact line location. The pantograph-catenary contact current data is obtained by the pantograph-catenary current sampling device and segmented according to the train's running position data. Indicates binding to the subway overhead contact line location. The vehicle body attitude data, which is obtained by an inertial measurement unit or a vehicle body attitude sensor; This indicates the mileage of the subway line corresponding to the detected train at the current moment. This parameter is calculated from the train's running location data. Indicates the location of the subway overhead contact line The corresponding overhead contact line structure information is read from pre-established overhead contact line structure information. This expression integrates multiple types of sensor data and structural ledger data at the same subway overhead contact line location, enabling subsequent analysis of the sequential and corresponding relationships between changes in different sensors around the same location.

[0033] S120. Extract the sequence of changes in multiple sensors corresponding to the location of the subway catenary based on the detection data.

[0034] Specifically, after acquiring the detection data, the server first reads visible light images, laser contour data, infrared thermal images, ultraviolet discharge images, pantograph-catenary contact current data, and vehicle attitude data, using the subway catenary location as a unified index. It then processes each type of data within a detection segment corresponding to the same subway catenary location. A detection segment refers to a continuous sampling interval before the train arrives at the subway catenary location, while passing through the location, and after leaving the location, ensuring that all sensor data are extracted based on changes around the same spatial position. Visible light edge variations are used to represent changes in the edge position, edge sharpness, or edge continuity of contact wires, droppers, positioners, insulators, segmented insulators, or support devices in visible light images relative to the normal state; laser profile variations are used to represent changes in the profile position, profile height, profile continuity, or echo intensity of contact wire components in laser profile data relative to the normal state; infrared temperature variations are used to represent changes in the local temperature response in infrared thermal images relative to the background temperature or historical normal temperature; ultraviolet bright spot variations are used to represent the appearance, enhancement, persistence, or repetition of ultraviolet bright spots in ultraviolet discharge images; pantograph-catenary contact current variations are used to represent abrupt changes, amplified fluctuations, or delayed recovery in pantograph-catenary contact current data; and vehicle attitude variations are used to represent pitch, roll, vertical vibration, or lateral vibration variations in vehicle attitude data.

[0035] When the server extracts visible light edge variations from visible light images, it first determines the image region corresponding to the contact wire component in the visible light image based on the contact wire structure information. Then, within the image region, it identifies the edges of the contact wire, droppers, positioners, insulators, segmented insulators, and support devices. It then compares the edge position, edge sharpness, and edge continuity in the current detection segment with the normal edge features corresponding to the same subway contact wire position during historical normal inspections, thus obtaining the visible light edge variations. Visible light edge variations can reflect both actual structural edge offsets and imaging blurring, local occlusion, or contamination of the detection window. Therefore, they cannot be used alone as a basis for judging contact wire defects; they need to be analyzed in conjunction with laser contour changes, pantograph-catenary contact current changes, and vehicle body attitude changes. The calculation expression for visible light edge variations is:

[0036] in, Indicates time Corresponding visible light edge changes; Indicates time The edge location is extracted from the visible light image; this parameter is obtained through edge detection, structural region localization, or image contour extraction. Indicates the location of the subway overhead contact line The average value of normal edge positions formed during historical normal detection processes; this parameter is obtained from statistical analysis of historical normal visible light images. Indicates the location of the subway overhead contact line The allowable fluctuation range of the corresponding normal edge position is determined by the maximum deviation or standard fluctuation range of the historical normal edge position. Indicates time The corresponding edge sharpness, which can be obtained from the edge gradient intensity or the edge grayscale change rate; Indicates the location of the subway overhead contact line The corresponding mean normal edge sharpness, which is obtained from historical normal visible light images; This indicates the allowable fluctuation range of normal edge sharpness, and this parameter is determined by the historical fluctuations in normal edge sharpness. Indicates time The corresponding edge continuity parameter is obtained by the edge segment breakage ratio or the continuous edge length. Indicates the location of the subway overhead contact line The corresponding mean value of normal edge continuity; This indicates the allowable fluctuation range for normal edge continuity; , , These represent the weighting coefficients corresponding to edge position, edge sharpness, and edge continuity, respectively. All three can be preset according to the detection accuracy requirements and are all greater than or equal to 0. This represents an extremely small positive number to prevent the denominator from being zero. This expression normalizes and weights edge position deviation, edge sharpness deviation, and edge continuity deviation, enabling different image features to form a unified visible light edge variation result.

[0037] When the server extracts laser contour changes from the laser contour data, it first transforms the contour points in the laser contour data to a detection coordinate system consistent with the catenary structure information. Then, it filters contour points based on the type of catenary component corresponding to the location of the subway catenary and calculates the contour height offset, lateral contour offset, contour discontinuity, and echo intensity changes in the current detection segment. Laser contour changes are used to reflect whether the spatial morphology of the catenary components has undergone abnormal changes. However, under the influence of changes in train attitude or tunnel airflow interference, non-structural contour offsets may also occur. Therefore, the server simultaneously records vehicle attitude changes and train speed data when extracting laser contour changes to facilitate subsequent correlation verification of the initial change sequence. The calculation expression for laser contour changes is:

[0038] in, Indicates time The corresponding changes in the laser profile; Indicates time The corresponding contour height is calculated from the coordinates of the contact wire component contour points in the laser contour data. Indicates the location of the subway overhead contact line The corresponding average height of the normal contour; This indicates the allowable fluctuation range of the normal profile height; Indicates time The corresponding lateral contour position, this parameter is determined by the coordinates of the laser contour point in the lateral direction; Indicates the location of the subway overhead contact line The average value of the corresponding normal horizontal contour position; This indicates the permissible range of fluctuation in the normal lateral profile position. Indicates time The corresponding echo intensity or contour completeness is obtained from the echo intensity or effective contour point percentage output by the laser scanning device. Indicates the location of the subway overhead contact line The corresponding average normal echo intensity or normal contour integrity; This indicates the allowable fluctuation range of normal echo intensity or normal profile integrity. , , These represent the weighting coefficients corresponding to contour height, lateral contour position, and echo intensity or contour completeness, respectively. This represents an extremely small positive number to prevent the denominator from being zero. This expression unifies the characterization of spatial profile offset and echo quality variation, enabling laser profile changes to simultaneously reflect geometric changes in contact wire components and disturbances in the detection path.

[0039] When the server extracts infrared temperature changes from infrared thermal images, it first determines the temperature response areas of insulators, segmented insulators, contact network connections, positioner connections, and support devices in the infrared thermal image based on the contact network structure information. Then, it calculates the temperature difference, temperature range, and temperature duration between the temperature response area and the surrounding background area. Infrared temperature changes reflect localized heating responses, especially for assisting in identifying insulation aging conditions. However, infrared temperature changes can also be affected by environmental thermal noise, water mist reflection, or equipment temperature measurement deviations. Therefore, when extracting infrared temperature changes, the server simultaneously retains the corresponding subway contact network location and contact network component type, enabling subsequent determination of whether the infrared temperature change and the ultraviolet bright spot change are both fixed near the same contact network component. The calculation expression for infrared temperature change is:

[0040] in, Indicates time The corresponding infrared temperature change; Indicates time The representative temperature of the temperature response area, which can be determined by the mean, maximum or high temperature quantile of the pixel temperature within the temperature response area; Indicates time The representative temperature of the background region surrounding the temperature response region is calculated from the pixel temperatures of adjacent non-abnormal regions. Indicates the location of the subway overhead contact line The corresponding allowable fluctuation range of normal temperature difference is obtained from historical normal infrared thermal images; Indicates time The extended range of the temperature response region, which is determined by the area or contour range of a continuous pixel region that exceeds the background temperature by a certain deviation; Indicates the location of the subway overhead contact line Reference range for temperature response under normal conditions; Indicates the end time The duration of infrared temperature change is obtained by accumulating sampling times that continuously exceed the temperature change judgment condition; Indicates the location of the subway overhead contact line The duration of temperature fluctuations allowed under normal conditions; , , These represent the weighting coefficients corresponding to temperature difference, temperature range, and temperature duration, respectively. This represents a very small positive number to prevent the denominator from being zero. This expression distinguishes short-term ambient thermal noise from continuous localized heating by combining local temperature difference, range of spread, and duration.

[0041] When the server extracts ultraviolet (UV) bright spot changes from UV discharge images, it first identifies local response regions in the UV discharge image whose brightness is higher than the surrounding background, and then correlates these local response regions with the positions of insulators, segment insulators, anchor joints, or contact network connections in the contact network structure information. UV bright spot changes represent changes in the UV response caused by weak corona discharge or partial discharge, including UV bright spot intensity, UV bright spot duration, and the number of times the UV bright spot recurs. Since weak corona discharge may only form weak UV bright spots in the early stages, the server does not use the obviousness of visible light edge changes as a prerequisite for extracting UV bright spot changes; instead, it extracts UV bright spot changes independently and then correlates them with infrared temperature changes. The calculation expression for UV bright spot changes is:

[0042] in, Indicates time Corresponding changes in ultraviolet bright spots; Indicates time The brightness intensity of the ultraviolet bright spot region is obtained by statistical analysis of the pixel brightness of the local response region in the ultraviolet discharge image. Indicates time The brightness intensity of the background region surrounding the ultraviolet bright spot area, which is obtained by statistical analysis of adjacent background pixels; Indicates the location of the subway overhead contact line The corresponding allowable fluctuation range of normal ultraviolet brightness is obtained from historical normal ultraviolet discharge images; Indicates the end time The duration of continuous existence of ultraviolet bright spots is calculated from the number of sampling frames and the sampling interval of continuously detected ultraviolet bright spots; Indicates the location of the subway overhead contact line The permissible duration of ultraviolet bright spots under normal conditions; Indicates the end time The number of times the ultraviolet bright spot appears repeatedly in the current detection segment. This parameter is obtained by counting the number of events from when the ultraviolet bright spot disappears to when it reappears. Indicates the location of the subway overhead contact line The number of times a UV bright spot is allowed to reappear under normal conditions; , , These represent the weighting coefficients corresponding to the intensity, duration, and number of repetitions of the ultraviolet bright spot, respectively. This represents an extremely small positive number to prevent the denominator from being zero. The expression characterizes changes in the ultraviolet bright spot through brightness enhancement, duration, and number of repetitions, enabling the distinction between incidental ultraviolet noise and stable discharge responses.

[0043] When the server extracts pantograph-catenary contact current changes from the pantograph-catenary contact current data, it first segments the continuous pantograph-catenary contact current data into current detection segments corresponding to the subway catenary positions, based on the train's operating position data. Then, it identifies current abrupt changes, fluctuation amplification, and recovery delays within these current detection segments. The pantograph-catenary contact current change is used to determine whether a real contact disturbance has occurred between the contact wire and the pantograph. When changes in visible light edges, laser contours, infrared temperature, and ultraviolet bright spots simultaneously increase, but the pantograph-catenary contact current change does not increase synchronously, the server can further determine whether the corresponding anomaly belongs to a multi-sensor synchronous anomaly state caused by multiple sensors being affected by external interference. The calculation expression for the pantograph-catenary contact current change is:

[0044] in, Indicates time The corresponding changes in pantograph-catenary contact current; Indicates time The pantograph-catenary contact current value, which is acquired by the pantograph-catenary current sampling device; Indicates the location of the subway overhead contact line The average normal pantograph-catenary contact current formed during historical normal testing processes; This indicates the allowable fluctuation range of normal pantograph-catenary contact current; Indicates the end time The degree of local current fluctuation, which is calculated from the degree of dispersion of the pantograph-catenary contact current value within the current short time window; Indicates the location of the subway overhead contact line Reference value for current fluctuation under normal conditions; This parameter indicates the degree of current recovery delay and is calculated from the sampling time required for the current to recover to the normal range after deviating from it. Indicates the location of the subway overhead contact line The allowable current recovery time under normal conditions; , , These represent the weighting coefficients corresponding to current amplitude deviation, current fluctuation degree, and current recovery delay degree, respectively. This represents an extremely small positive number to prevent the denominator from being zero. The expression reflects the pantograph-catenary contact state through the current amplitude, fluctuation level, and recovery delay, allowing for a before-and-after comparison of actual contact disturbances with image-based or contour-based anomalies.

[0045] When the server extracts vehicle attitude changes from the vehicle attitude data, it first reads the pitch, roll, vertical vibration, and lateral vibration changes of the detected train in the current detection segment, and calculates the deviation of the vehicle attitude data from the historical normal attitude state. Vehicle attitude changes are used to interpret whether visible light edge changes and laser profile changes may be caused by changes in the detection equipment's attitude. When vehicle attitude changes occur before or simultaneously with visible light edge changes and laser profile changes, the server can mark the corresponding changes as potentially affected by detection equipment attitude interference in subsequent validity marking. The calculation expression for vehicle attitude changes is:

[0046] in, Indicates time The corresponding changes in vehicle attitude; Indicates time The pitch change value, which is obtained by the vehicle attitude sensor or inertial measurement unit; Indicates the location of the subway overhead contact line Corresponding normal pitch change reference value; This indicates the permissible range of normal pitch variation; Indicates time The roll change value, which is obtained by the vehicle attitude sensor or inertial measurement unit; Indicates the location of the subway overhead contact line Corresponding normal roll change reference value; This indicates the allowable fluctuation range for normal roll changes; Indicates time The vertical vibration variation value, which is obtained by the vehicle attitude sensor, acceleration sensor or inertial measurement device; Indicates the location of the subway overhead contact line The corresponding normal vertical vibration reference value; This indicates the permissible fluctuation range of normal vertical vibration; Indicates time The lateral vibration variation value, which is obtained by the vehicle attitude sensor, acceleration sensor or inertial measurement device; Indicates the location of the subway overhead contact line The corresponding normal lateral vibration reference value; This indicates the permissible range of normal lateral vibration. , , , These represent the weighting coefficients corresponding to pitch, roll, vertical vibration, and lateral vibration changes, respectively. This indicates a very small positive number to prevent the denominator from being zero. This expression characterizes the motion state of the detection platform through changes in attitude angle and vibration, enabling subsequent determination of whether image edge shifts and laser contour shifts have attitude interference sources.

[0047] After acquiring changes in visible light edge intensity, laser contour, infrared temperature, ultraviolet bright spot intensity, pantograph-catenary contact current, and vehicle attitude, the server determines the start time, enhancement time, peak time, and recovery time for each type of change. The start time is the earliest moment when the change transitions from the normal fluctuation range to the abnormal change range; the enhancement time is the moment when the change continues to show an increasing trend after entering the abnormal change range; the peak time is the moment when the change reaches its maximum change within the current detection segment; and the recovery time is the moment when the change returns from the abnormal change range to the normal fluctuation range. The server uses the same time extraction rules for all six types of changes, ensuring that all changes can be ranked according to a unified set of rules. For any given change... For example, the expressions for determining the start time of change, the peak time of change, and the recovery time of change are:

[0048] , , in, Indicates the first The result of the class change at time The change value, It can respond to changes in visible light edge, laser profile, infrared temperature, ultraviolet bright spot, pantograph contact current, or vehicle attitude. Indicates the first The moment when the change in the result of the class change begins; Indicates the first Peak time of change in the results of class changes; Indicates the first The time to recover the changes in the results of the class change; Indicates the location of the subway overhead contact line First The anomaly detection threshold corresponding to the class of changes is determined by the parameter from the first normal detection process in history. The normal fluctuation range of the change results is determined, and it can also be adjusted according to the detection sensitivity requirements. The above expression determines the key moments of each change result through threshold overshoot, peak location, and threshold fallback, so that changes from different sensors can be converted into sortable time-series events.

[0049] When determining the moment of enhancement, the server can search for the point after the start of the change when the change result continues to rise and exceeds the enhancement criteria. The moment of enhancement differs from the start of the change; it indicates that the change result is no longer a short-term exceedance, but rather shows a continuous enhancement trend. The expression for determining the moment of enhancement is:

[0050] in, Indicates the first The moment when the change in the class of results is enhanced; Indicates the first The result of the class change at time The change value; Indicates the first The change in the class of results at the previous comparison time; This parameter represents the sampling interval between adjacent comparison times, and is determined by the sampling frequency of each sensor or a uniform resampling interval. Indicates the location of the subway overhead contact line First The enhancement threshold corresponding to the type of change result is obtained by statistically analyzing the change increments during historical normal detection processes. This expression determines the moment when the change result transitions from anomaly occurrence to anomalous enhancement by judging whether the change result continues to increase between adjacent time points.

[0051] When the server generates the initial change sequence for the corresponding subway catenary location based on the start time, enhancement time, and peak time of the change, it first sorts the results according to the start time. If two or more results have the same start time or their difference falls within the same time window, they are further sorted according to the enhancement time. If the enhancement times of two or more results are still close, they are sorted according to the peak time. The initial change sequence indicates the sequential relationship of various sensor changes at the same subway catenary location. For example, if the change in vehicle posture appears first, followed by changes in laser contour and visible light edge, it usually means that image and contour anomalies may be related to interference from the posture of the detection equipment. If the change in pantograph contact current appears first, followed by changes in laser contour, it may indicate a loose catenary structure or an abnormal hard point. The initial change sequence can be represented as:

[0052] in, Indicates the location of the subway overhead contact line The corresponding initial change sequence; Indicates the first The types of changes are identified, including visible light edge changes, laser profile changes, infrared temperature changes, ultraviolet bright spot changes, pantograph contact current changes, and vehicle attitude changes. Indicates the first The moment when the change in the result of the class change begins; Indicates the first The moment when the change in the class of results is enhanced; Indicates the first Peak time of change in the results of class changes; This expression represents a sorting process used to determine the order of changes based on the start time, the intensification time, and the peak time. By using a multi-level time-sequencing method, the initial change order is independent of a single sampling time, thus reducing the impact of sensor sampling frequency differences on the sorting results.

[0053] When the server performs correlation verification on the initial change sequence based on train speed data, it first reads the train speed data when the detected train passes the subway contact network position and determines whether the occurrence time of each change result in the initial change sequence is reasonably advanced, delayed, or its duration compressed as the train speed data changes. Train speed data refers to the speed of the detected train at the corresponding subway contact network position, which can be obtained from the train control system, odometer, or positioning system. If the laser contour change and visible light edge change are advanced or their duration shortened as the train speed data changes, and the vehicle body attitude change also occurs synchronously, the server marks the correlation between these changes as speed-related and attitude-related. If the infrared temperature change and ultraviolet bright spot change are fixed near the same insulator, segment insulator, or contact network connection and do not shift as a whole as the train speed data changes, the server marks the correlation between these changes as structurally fixed correlation. The verification expression for the duration of the train speed data is:

[0054] in, Indicates the first The results of this type of change are reflected in the location of the subway catenary. Duration deviation at the location; Indicates the first The actual duration of the change result, which is obtained by subtracting the change start time from the change recovery time; Indicates the location of the subway overhead contact line The corresponding detection influence length can be determined by the component coverage area, sensor field of view, or detection window range in the contact wire structure information. This indicates the location where the inspection train passes through the subway overhead contact line. The train's operating speed data at that time is obtained from the train control system, odometer, or positioning system. This represents an extremely small positive number to prevent the denominator from being zero. The expression compares the actual duration with the estimated passage time based on train speed data to determine if the change conforms to the timing pattern of a train passing through the subway overhead contact line location.

[0055] When the server marks the validity of each change result in the initial change sequence, it adds a corresponding mark to each change result based on the initial change sequence, train speed data, catenary structure information, and the sequential relationship between each change result. If the vehicle body attitude change precedes the laser profile change and visible light edge change, and the laser profile change and visible light edge change are synchronously enhanced with the vehicle body attitude change, then the laser profile change and visible light edge change are marked as changes that may be affected by the attitude interference of the detection equipment. If the ultraviolet bright spot change and infrared temperature change are fixed near the insulator, segment insulator, or support device, and both appear repeatedly in multiple detection segments, then the ultraviolet bright spot change and infrared temperature change are marked as changes that may be related to the insulation aging state. If the pantograph-catenary contact current change precedes the laser profile change, and the laser profile change is fixed near the positioner, dropper, anchor joint, or contact wire connection, then the pantograph-catenary contact current change and laser profile change are marked as changes that may be related to the loose state of the contact wire structure. If the pantograph-catenary contact current change, laser profile change, and visible light edge change continuously enhance within a short period of time, then the corresponding changes are marked as changes that may be related to the abnormal state of hard points.

[0056] After completing the validity marking, the server generates a multi-sensor change sequence for the corresponding subway catenary location based on the initial change sequence. This multi-sensor change sequence is not simply a time-ordered result, but rather a structured temporal relationship that includes the change type, start time, enhancement time, peak time, recovery time, validity marking, corresponding subway catenary location, and corresponding catenary components for each change. The server writes the multi-sensor change sequence into the detection data, enabling subsequent comparison with the normal change sequence formed during historical normal detection processes. When visible light edge changes, laser contour changes, infrared temperature changes, and ultraviolet bright spot changes simultaneously increase, but the pantograph-catenary contact current changes and vehicle attitude changes do not increase synchronously, a multi-sensor synchronization anomaly is determined for the corresponding subway catenary location. The expression for the multi-sensor change sequence is:

[0057] in, Indicates the location of the subway overhead contact line The corresponding sequence of changes from multiple sensors; Indicates the first The types of changes are identified, including visible light edge changes, laser profile changes, infrared temperature changes, ultraviolet bright spot changes, pantograph contact current changes, and vehicle attitude changes. Indicates the first The moment when the change in the result of the class change begins; Indicates the first The moment when the change in the class of results is enhanced; Indicates the first Peak time of change in the results of class changes; Indicates the first The time to recover the changes in the results of the class change; Indicates the first The validity flag corresponding to the class change result is obtained by correlation check and order relationship judgment; Indicates the first The parameter for the contact wire component corresponding to the type of change result is determined by the binding relationship between the contact wire structure information and the sensor data location. This indicates the current location of the subway overhead contact line. This expression, by uniformly storing six types of change results along with their timing information, validity markers, and structural attribution relationships, enables the sequence of changes from multiple sensors to directly support the generation of subsequent abnormal state identification and processing strategies.

[0058] S130. Compare the sequence of changes of multiple sensors with the sequence of normal changes formed during historical normal detection. When the changes of visible light edge, laser contour, infrared temperature and ultraviolet bright spots are enhanced simultaneously, and the changes of pantograph contact current and vehicle posture are not enhanced synchronously, it is determined that there is a multi-sensor synchronous abnormal state at the corresponding subway catenary location.

[0059] Specifically, when the server retrieves the normal change sequence corresponding to the location of the subway catenary during historical normal inspections, it first uses the subway catenary location as an index to read historical normal inspection data formed under similar operating conditions and for the same catenary component from the historical inspection database. It then extracts the normal sequence of visible light edge changes, laser profile changes, infrared temperature changes, ultraviolet bright spot changes, pantograph-catenary contact current changes, and vehicle attitude changes from this historical normal inspection data. The normal change sequence refers to the baseline temporal relationship formed around the same subway catenary location when there are no catenary defects, no interference with the inspection equipment's attitude, no contamination of the inspection window, and no interference from tunnel airflow. This sequence serves as a reference for the current multi-sensor change sequence. When retrieving the normal change sequence, the server not only reads whether each change result appears, but also reads the normal change start time, normal change enhancement time, normal change peak time, normal change recovery time, and normal duration for each change result. This ensures that subsequent item-by-item comparisons can simultaneously cover the change sequence, the degree of change concentration, and the duration of the change. The normal change sequence can be represented as:

[0060] in, Indicates the location of the subway overhead contact line The corresponding normal sequence of changes; Indicates the first The types of changes are identified, including visible light edge changes, laser profile changes, infrared temperature changes, ultraviolet bright spot changes, pantograph contact current changes, and vehicle attitude changes. Indicates the first The parameter is the start time of the normal change in the historical normal detection process, which is obtained by calculating the statistical representative value of the corresponding change start time in multiple historical normal detections. Indicates the first The parameter is obtained by calculating the statistical representative value of the corresponding enhancement time of the change result in the historical normal detection process. Indicates the first The peak time of normal change in the class of change results during the historical normal detection process is obtained by calculating the statistical representative value of the corresponding peak time of change in multiple historical normal detections. Indicates the first The normal change recovery time of the class change result in the historical normal detection process is obtained by calculating the statistical representative value of the corresponding change recovery time in multiple historical normal detections; Indicates the first The normal duration of the change result in the historical normal detection process, which is obtained by statistically analyzing the difference between the normal change recovery time and the normal change start time; Indicates the first The reference value for the intensity of normal changes in the class of changes during historical normal detection processes. This parameter is determined by the mean or median value of the corresponding changes in historical normal detection processes. Indicates the first The parameter represents the allowable range of normal changes in the results of a type of change within the historical normal detection process. This parameter is determined by the fluctuation range or dispersion of the corresponding change results in the historical normal detection. This expression constructs a benchmark time series reference for the same subway catenary location using historical normal detection data, enabling the current sequence of changes from multiple sensors to be compared with a normal state that is consistent in location and structure.

[0061] When the server compares the sequence of changes from multiple sensors with the normal sequence of changes item by item, it reads the start time, enhancement time, peak time, recovery time, duration, and intensity of each change result in the current detection process, and compares them with the reference values ​​for the normal time, duration, and intensity of the corresponding change result in the normal sequence. Item-by-item comparison refers to the corresponding comparison between the same change result, rather than mixing the six types of change results into a single abnormal score; for example, comparing the current visible light edge change with the normal visible light edge change, the current laser profile change with the normal laser profile change, and the current infrared temperature change with the normal infrared temperature change. Through item-by-item comparison, the server can determine whether a certain change result in the current sequence of changes from multiple sensors appears earlier, appears more frequently, or has a longer duration than in the normal sequence of changes, providing a basis for subsequently determining candidate positions for multi-sensor synchronous enhancement. The degree of temporal deviation in item-by-item comparison can be expressed by the following expression:

[0062] in, Indicates the location of the subway overhead contact line First The degree of temporal deviation of the result of a class of changes relative to the normal order of changes; Indicates the current detection process. The parameter indicates the start time of the change in the class of changes, and is read from the current multi-sensor change sequence. Indicates the first The parameter is read in the order of normal changes at the start time of the class change result; Indicates the first The allowable fluctuation range of the normal change start time of the class change result is determined by the dispersion of the change start time in historical normal detections; Indicates the current detection process. The moment when the change in the class of results is enhanced; Indicates the first The moment when the normal change in the result of the class of changes is enhanced; Indicates the first The allowable fluctuation range for normal changes in the outcome of class-specific changes; Indicates the current detection process. Peak time of change in the results of class changes; Indicates the first Peak time of normal change in the result of class change; Indicates the first The allowable fluctuation range of the peak time of normal changes in the results of similar changes; Indicates the current detection process. The duration of the change outcome, which is obtained by subtracting the change start time from the change recovery time; Indicates the first The normal duration of the result of the type of change; Indicates the first The allowable fluctuation range for the normal duration of such changes; , , , These represent the comparison weights corresponding to the start time of the change, the time of the increase in the change, the time of the peak of the change, and the duration, respectively, and all values ​​are greater than or equal to 0. This indicates a very small positive number that prevents the denominator from being zero. The expression determines whether the change result significantly deviates from the normal timing state by comparing the current change result with the corresponding timing parameters in the normal change sequence after normalization.

[0063] When determining whether changes in visible light edge, laser contour, infrared temperature, and ultraviolet bright spots occur earlier, occur in clusters, or last longer than normal, the server reads the start time, enhancement time, and recovery time of each of the four types of changes. It then determines whether the start time of each type of change is earlier than the normal start time, whether the enhancement times fall within the same detection segment, and whether the duration exceeds the allowable range. "Early occurrence" means the current change starts earlier than the normal start time and exceeds the allowable fluctuation range; "clustered occurrence" means the enhancement times or peak times of the four types of changes fall within the same detection segment or a similar time window; "extended duration" means the time between the start and recovery times of the current change significantly exceeds the normal duration in the normal change sequence. The server can use early occurrence markers, clustered occurrence markers, and extended duration markers to jointly determine whether there is an abnormal synchronization trend among the four types of changes. The expression for determining clustered occurrence is:

[0064] in, Indicates the location of the subway overhead contact line The degree of temporal concentration of changes in visible light edge, laser profile, infrared temperature, and ultraviolet bright spot; This represents the set of changes that participated in the synchronous enhancement judgment, specifically including visible light edge changes, laser profile changes, infrared temperature changes, and ultraviolet bright spot changes; Indicates the first When the change in the class of changes is enhanced, this parameter is read from the current multi-sensor change sequence; This indicates the latest point in time when the change intensifies among the four types of changes. This represents the earliest enhancement time among the four types of changes. This expression determines whether the four types of changes are concentrated within the same detection segment by calculating the difference between the latest and earliest enhancement times. When the result is less than or equal to the preset synchronization time window, it can be considered that the four types of change results have a concentrated occurrence characteristic.

[0065] When determining whether the changes in pantograph-catenary contact current and vehicle attitude are within the range corresponding to the normal change sequence, the server compares the current changes in pantograph-catenary contact current and vehicle attitude with the reference values ​​for normal change intensity and the allowable range of normal change in the normal change sequence. The change range refers to the normal fluctuation range allowed for changes in pantograph-catenary contact current and vehicle attitude during historical normal detection processes. Exceeding the change range usually indicates the presence of actual pantograph-catenary contact disturbances or detection platform motion disturbances. If the changes in pantograph-catenary contact current are still within the normal change range, it means that the current image, contour, infrared, and ultraviolet changes are not supported by actual pantograph-catenary contact disturbances; if the changes in vehicle attitude are still within the normal change range, it means that the current image and contour changes cannot be directly explained by obvious changes in detection platform attitude. The expression for determining the change range is:

[0066] in, Indicates the location of the subway overhead contact line First Whether the changes in the class are within the normal range; Indicates the current detection process. The intensity of the change in the type of change at the location of the subway catenary is calculated from the corresponding change in the current detection data. Indicates the first The reference value of the intensity of normal change corresponding to the class of change results in the historical normal detection process. This parameter is read from the normal change sequence. Indicates the first The permissible range of normal changes corresponding to the type of change result is obtained from historical normal detection data statistics; when... When, it indicates the first The results of the class change are within the normal range; when When, it indicates the first The result of this type of change exceeds the normal range. This expression determines whether changes in pantograph-catenary contact current and vehicle attitude can be considered as accompanying evidence of a real disturbance by comparing the current change intensity with the historical normal range.

[0067] When visible light edge changes, laser contour changes, infrared temperature changes, and ultraviolet bright spot changes simultaneously enhance within the same detection segment, the server marks the corresponding subway catenary location as a candidate location for multi-sensor synchronous enhancement. The same detection segment refers to a continuous sampling interval surrounding the same subway catenary location, which can include adjacent sampling data before the train arrives at the subway catenary location, when the train passes the subway catenary location, and after the train leaves the subway catenary location. Simultaneous enhancement does not require the four types of changes to reach their peak values ​​at exactly the same sampling time; rather, it requires that the enhancement times of the four types of changes fall within the same preset synchronization time window, and that the intensity of the changes in all four types of changes exceeds the corresponding enhancement judgment threshold. The candidate location for multi-sensor synchronous enhancement is the location object for subsequent reverse verification of pantograph-catenary contact current changes and vehicle attitude changes. Its meaning is not to directly confirm a catenary defect, but to confirm that multiple sensors have jointly enhanced their responses at that subway catenary location. The judgment expression for the candidate location for multi-sensor synchronous enhancement is:

[0068] in, Indicates the location of the subway overhead contact line Whether it is marked as a candidate location for multi-sensor synchronization enhancement; This represents a conditional function, which takes the value 1 when the condition inside the parentheses is true and takes the value 0 when the condition inside the parentheses is false. This indicates the degree of temporal concentration of the four types of change outcomes; Indicates the location of the subway overhead contact line The corresponding preset synchronization time window, this parameter can be determined based on the sampling frequency of each sensor, train speed data and detection segment length; This represents the set of changes resulting from the synchronization enhancement judgment. Indicates the first The results of this type of change are reflected in the location of the subway catenary. The intensity of the change at that location; Indicates the first The enhancement threshold corresponding to the type of change result is determined by the normal change range and detection sensitivity requirements in historical normal detection data. This expression filters subway catenary locations with common enhancement phenomena as candidate locations for multi-sensor synchronous enhancement by simultaneously satisfying two conditions: concentrated enhancement time and change intensity exceeding the threshold.

[0069] After the server reads the changes in pantograph-catenary contact current and vehicle attitude corresponding to the candidate locations for multi-sensor synchronous enhancement, it performs a conditional judgment on the accompanying disturbances of the actual catenary structure at these candidate locations. Actual catenary structure disturbances refer to physical disturbances caused by defects in the catenary itself, loose catenary components, hard spots in the contact wire, or abnormal pantograph-catenary contact. If such disturbances truly exist, they should typically be accompanied by abrupt changes, amplified fluctuations, or delayed recovery in the pantograph-catenary contact current changes, or by pitch, roll, vertical vibration, or lateral vibration changes in the vehicle attitude. If the pantograph-catenary contact current changes do not exhibit corresponding abrupt changes, amplified fluctuations, or delayed recovery, and the vehicle attitude changes also do not exhibit corresponding pitch, roll, vertical vibration, or lateral vibration changes, it indicates that although the candidate locations for multi-sensor synchronous enhancement exhibit the phenomenon of multiple sensors enhancing the same location, they lack the accompanying responses that should be present in actual catenary structure disturbances. The expression for determining the accompanying conditions is:

[0070] in, Indicates the location of the subway overhead contact line The degree to which the accompanying conditions for actual catenary structural disturbances are satisfied; This parameter represents the intensity of the pantograph-catenary contact current change at the location of the subway catenary. It is calculated from the current abrupt change, fluctuation amplification, and recovery delay in the pantograph-catenary contact current data. This indicates the threshold for determining the accompanying changes in pantograph-catenary contact current. This parameter is determined by the normal range of pantograph-catenary contact current changes in historical normal testing. This parameter represents the intensity of the vehicle's attitude change corresponding to the location of the subway catenary. It is calculated from the pitch, roll, vertical vibration, and lateral vibration changes in the vehicle's attitude data. This indicates the threshold for determining the accompanying changes in vehicle attitude. This parameter is determined by the normal range of vehicle attitude changes in historical normal detection. This represents a conditional judgment function. This expression determines whether the candidate location for multi-sensor synchronous augmentation has a real disturbance-accompanied response by judging whether the changes in pantograph-catenary contact current and vehicle attitude exceed the corresponding associated judgment thresholds; when... When this occurs, it indicates that the candidate locations for multi-sensor synchronous enhancement do not meet the accompanying conditions of the actual catenary structure disturbance.

[0071] After the server determines whether the candidate locations for multi-sensor synchronous enhancement meet the accompanying conditions of actual catenary structural disturbance, it continues to judge whether changes in visible light edge, laser profile, infrared temperature, and ultraviolet bright spots have the same enhancement direction, similar enhancement time, and similar recovery time. Same enhancement direction means that all four types of changes show enhancement or deviation from the normal change sequence, rather than partial enhancement and partial weakening; similar enhancement time means that the enhancement times of the four types of changes fall within the same preset synchronization time window; similar recovery time means that the difference between the recovery times of the four types of changes is less than the preset recovery time window. If all four types of changes simultaneously meet the above characteristics, but the changes in pantograph contact current and vehicle attitude are not synchronously enhanced, it indicates that the anomaly is more consistent with the pseudo-uniformity characteristics formed after multiple sensors are jointly affected by the same external interference, rather than typical catenary defect characteristics. The expression for similar recovery time is:

[0072] in, Indicates the location of the subway overhead contact line The degree of similarity in recovery time for the four types of changes; Indicates the first The time for recovering the change in the class of changes is determined by reading the current sequence of changes from multiple sensors. This represents the set of changes resulting from the synchronization enhancement judgment. Indicates the latest change recovery time among the four types of changes; This indicates the earliest recovery time among the four types of changes. This expression determines whether the four types of changes have recovered to their normal range within a similar timeframe by calculating the difference in recovery time among the four types of changes; when... When the recovery time is less than or equal to the preset recovery time window, the four types of change results can be considered to have similar recovery times.

[0073] When the server determines that a multi-sensor synchronous anomaly exists at a corresponding subway catenary location, several conditions must be met simultaneously: visible light edge changes, laser profile changes, infrared temperature changes, and ultraviolet bright spot changes must all enhance simultaneously within the same detection segment; these four types of changes must have the same enhancement direction, similar enhancement time, and similar recovery time; and the pantograph-catenary contact current changes and vehicle attitude changes must not enhance synchronously. The multi-sensor synchronous anomaly indicates a surface-consistent abnormal enhancement among multiple sensor detection results. However, this abnormal enhancement lacks accompanying support from pantograph-catenary contact current changes and vehicle attitude changes. Therefore, in subsequent processing, it is necessary to further determine whether the anomaly is mobile or fixed based on its persistence between adjacent catenary locations. The comprehensive judgment expression for the multi-sensor synchronous anomaly is:

[0074] in, Indicates the location of the subway overhead contact line Is there an abnormal state in the multi-sensor synchronization? Indicate whether the location of the subway overhead contact line is a candidate location for multi-sensor synchronous enhancement; This indicates the degree to which the accompanying conditions for disturbances in the actual overhead contact system structure are satisfied; This indicates the degree of temporal concentration of the four types of change outcomes; Indicates the preset synchronization time window; This indicates the degree to which the recovery times of the four types of changes are similar; This indicates the preset recovery time window, which can be determined based on the sampling frequency of each sensor, the length of the detection segment, and the historical normal recovery time fluctuation range. This indicates whether the four types of change results have the same direction of enhancement. The value is 1 when all four types of change results show enhancement or deviation from the normal change order, and 0 otherwise. This represents a conditional judgment function. This expression determines the synchronous abnormal state of multiple sensors through joint constraints including synchronous enhancement candidate judgment, elimination of accompanying conditions, proximity of enhancement time, proximity of recovery time, and consistency of enhancement direction. This avoids directly misjudging consistent surface anomalies of multiple sensors as actual catenary defects.

[0075] S140. Based on the continuity of the multi-sensor synchronization anomaly between adjacent contact network positions, determine the moving anomaly and the fixed anomaly.

[0076] Specifically, after determining that a multi-sensor synchronization anomaly exists at a certain subway catenary location, the server first reads the adjacent catenary locations ahead and behind the current location along the train's direction of travel from the catenary structure information. These locations are then arranged according to the line mileage (from smallest to largest) or the actual train direction. Adjacent catenary locations refer to those spatially adjacent to the current location and belonging to the same detection section. These locations can correspond to the positions of contact wires, droppers, positioners, insulators, segment insulators, anchor joints, or support devices. The catenary location arrangement refers to the server arranging the anomaly's starting position and multiple adjacent catenary locations into a continuous sequence according to the train's direction of travel. This sequence is used to determine whether the multi-sensor synchronization anomaly propagates along the train's direction of travel. The server then reads the visible light edge changes, laser profile changes, infrared temperature changes, and ultraviolet bright spot changes corresponding to each subway catenary location in this continuous sequence. It then determines whether these four types of changes occur continuously among multiple adjacent catenary locations, thus avoiding directly determining the source of the anomaly based solely on the anomaly at a single subway catenary location.

[0077] When the server constructs the catenary location arrangement, it can use the catenary location exhibiting a multi-sensor synchronization anomaly as the anomaly initiation location, and the catenary location exhibiting a subsequent multi-sensor synchronization anomaly along the train's direction of travel as the anomaly continuation location. The server then determines whether these locations are consecutive based on the difference in track mileage between them. The anomaly initiation location refers to the catenary location where a multi-sensor synchronization anomaly first occurs during the current detection process; the anomaly continuation location refers to the adjacent catenary location where a multi-sensor synchronization anomaly continues after the anomaly initiation location. If the difference in track mileage between two adjacent catenary locations exceeds a preset adjacent distance, the server will not consider them consecutive locations to avoid misjudging independent anomalies that are spatially distant as continuations of the same anomaly. The continuity of catenary locations can be expressed as follows:

[0078] in, Indicates the first The location of the subway overhead contact line is related to the first Whether there is a continuous abnormal relationship between the locations of the subway overhead contact system; This represents a conditional function, which takes the value 1 when the condition inside the parentheses is true and takes the value 0 when the condition inside the parentheses is false. Indicates the first The line mileage corresponding to each subway catenary location is obtained from the catenary structure information; Indicates the first The line mileage corresponding to each subway catenary location is also read from the catenary structure information; This indicates the preset adjacent distance, which is preset based on the spacing of subway catenary components, detection resolution, and allowable continuous tracking range. Indicates the first Whether there is a multi-sensor synchronization anomaly at a certain location of the subway catenary is determined by the aforementioned multi-sensor synchronization anomaly determination process. Indicates the first Does a subway overhead contact line location exhibit a multi-sensor synchronous anomaly? This expression determines whether a continuous chain of anomalies can be formed between adjacent overhead contact line locations by considering both the continuity of the line mileage and the simultaneous existence of the anomaly.

[0079] When the server tracks whether visible light edge changes, laser contour changes, infrared temperature changes, and ultraviolet bright spot changes occur continuously in the catenary position arrangement, it not only determines whether all four types of changes exceed their corresponding anomaly thresholds, but also whether these four types of changes maintain similar change patterns across multiple adjacent catenary positions. Change pattern refers to the change characteristics formed by the combination of enhancement direction, enhancement amplitude, enhancement duration, peak position, and recovery process in the detected segment; similar change patterns mean that the four types of changes corresponding to the anomaly continuation position are similar in the above characteristics to the four types of changes corresponding to the anomaly initiation position. The server can construct change pattern vectors corresponding to the anomaly initiation position and the anomaly continuation position respectively, and determine whether the anomaly continues with the train's running direction by using change pattern similarity. The expression for calculating change pattern similarity is:

[0080] in, Indicates the starting position of the exception With the position of abnormal continuation The similarity of the morphological changes between them; This represents the set of changes involved in the motion feature judgment, specifically including visible light edge changes, laser contour changes, infrared temperature changes, and ultraviolet bright spot changes; Indicates the first The similarity weights corresponding to the different types of change results can be preset according to the importance of each type of change result to the judgment of movement features; Indicates the starting position of the exception First The change pattern vector of the change result, which is composed of the enhancement direction, enhancement magnitude, duration, peak time and recovery time of the change result; Indicates the position of abnormal continuation First The change pattern vector of the class change result, which is calculated from the detection data corresponding to the abnormal continuation position; This function represents the similarity calculation between two vectors with different forms, and can be obtained based on the similarity of the vector angle, normalized distance similarity, or correlation similarity. This indicates an extremely small positive number to prevent the denominator from being zero. The expression determines whether anomalies in multiple adjacent contact wire locations belong to the continuation of the same anomaly chain by comparing the similarity in the morphology of four types of changes at the anomaly initiation and continuation positions.

[0081] When the location of the anomaly continuation corresponding to the multi-sensor synchronization anomaly appears sequentially after the anomaly initiation location along the train's direction of travel, and the visible light edge changes, laser profile changes, infrared temperature changes, and ultraviolet bright spot changes corresponding to the anomaly continuation location have similar change patterns to those corresponding to the anomaly initiation location, the server determines that the multi-sensor synchronization anomaly has a moving characteristic. A moving characteristic means that the anomaly is not fixed at a single subway contact network location or a single contact network component, but rather appears sequentially between multiple adjacent contact network locations as the detection train travels along the line. This characteristic typically corresponds to anomalies caused by detection window contamination, tunnel airflow interference, or changes in the detection equipment's status with the train, because the source of such anomalies moves with the detection train and is not permanently fixed at a particular contact network component. The expression for determining the moving characteristic is:

[0082] in, Indicates the starting position of the exception Does the corresponding multi-sensor synchronization anomaly exhibit movement characteristics? Indicates the position after the start of the exception. The continuous anomaly relationship between adjacent contact wire positions is calculated from the contact wire position continuity expression. This indicates the number of adjacent catenary locations involved in the tracking; this parameter is determined by the number of catenary locations within the preset tracking range. This parameter represents the minimum number of consecutive anomalies required to determine a movement feature; it is preset based on detection sensitivity and the spacing between line components. This represents the set of locations where anomalies continue, consisting of adjacent contact wire locations where multiple sensors exhibit synchronous anomalies sequentially along the train's direction of travel. Indicates the starting position of the exception With the position of abnormal continuation The similarity of the morphological changes between them; This parameter represents the threshold for morphological similarity, determined based on the degree of morphological similarity corresponding to the same external disturbance in historical detection data. The expression uses both the number of consecutive anomalies and the morphological similarity to determine whether an anomaly exhibits characteristics of movement along the train's direction of travel.

[0083] After determining that the multi-sensor synchronization anomaly has movement characteristics, the server further reads the changes in pantograph-catenary contact current and vehicle attitude corresponding to the anomaly's initiation and continuation positions, and determines whether both are not synchronously enhanced. Synchronous enhancement refers to the pantograph-catenary contact current change or vehicle attitude change exceeding the corresponding judgment threshold within a detection segment that is the same as or similar to visible light edge changes, laser contour changes, infrared temperature changes, and ultraviolet bright spot changes. If the pantograph-catenary contact current changes at the anomaly's initiation and continuation positions do not show abrupt changes, amplified fluctuations, or delayed recovery, and the vehicle attitude changes do not show corresponding pitch changes, roll changes, vertical vibration changes, or lateral vibration changes, it indicates that the movement characteristic lacks the accompanying support of real pantograph-catenary contact disturbances and strong attitude disturbances of the detection platform, and is more consistent with the movement anomaly state caused by external interference accompanying vehicle movement. The judgment expression for the non-synchronous enhancement of pantograph-catenary contact current changes and vehicle attitude changes is:

[0084] in, Indicates the starting position of the exception Whether the changes in pantograph contact current and vehicle attitude did not increase synchronously at the locations where the abnormality continued; This represents the set of locations consisting of the starting position and the continuation position of the anomaly; Indicates the location of the subway overhead contact line The corresponding pantograph-catenary contact current variation intensity, this parameter is calculated from the pantograph-catenary contact current data; Indicates the location of the subway overhead contact line The corresponding threshold for judging changes in pantograph-catenary contact current is determined by historical normal detection data. Indicates the location of the subway overhead contact line The corresponding intensity of vehicle attitude change, this parameter is calculated from vehicle attitude data; Indicates the location of the subway overhead contact line The corresponding threshold for judging vehicle attitude change is determined by historical normal detection data. This represents a conditional judgment function. The expression requires that the changes in pantograph-catenary contact current and vehicle attitude at both the initial and all subsequent abnormal positions not exceed a threshold, thus eliminating the influence of actual pantograph-catenary contact disturbances and obvious detection platform attitude disturbances on the judgment of movement anomalies.

[0085] When a multi-sensor synchronous anomaly exhibits mobile characteristics, and the changes in pantograph-catenary contact current and vehicle attitude at the anomaly's inception and continuation points do not increase synchronously, the server classifies the multi-sensor synchronous anomaly as a mobile anomaly. A mobile anomaly refers to a multi-sensor synchronous anomaly occurring continuously between multiple adjacent catenary locations and shifting with the train's direction of travel, but without exhibiting repetitiveness fixed to a single catenary component. The technical implication of this state is that the source of the anomaly is more likely to move with the detected train, such as contamination of the detection window, tunnel airflow interference, or localized changes in the detection equipment's condition, rather than a defect in the catenary itself. The comprehensive judgment expression for a mobile anomaly is:

[0086] in, Indicates the starting position of the exception Whether the corresponding multi-sensor synchronization anomaly is determined to be a movement anomaly; Indicates whether the abnormal state of multi-sensor synchronization has movement characteristics; This indicates whether there is no synchronous enhancement of changes in pantograph-catenary contact current and vehicle attitude at either the initial or continuation location of the anomaly. This expression determines the abnormal movement state by combining movement characteristics and accompanying response exclusion conditions, enabling the distinction between external disturbances moving with the vehicle and actual defects fixed to the catenary structure.

[0087] When determining whether a multi-sensor synchronization anomaly has a fixed characteristic, the server first checks whether the anomaly does not sequentially shift along the train's direction of travel, and then checks whether the anomaly occurs a preset number of times within a preset range at the same subway catenary location or within the same catenary component. A fixed characteristic refers to an anomaly that does not move with the detected train but repeatedly occurs around the same line mileage, the same subway catenary location, or the same catenary component. The preset range for the same catenary component refers to a spatial neighborhood defined around the installation location of the catenary component, used to absorb sensor positioning errors, line mileage matching errors, or image recognition deviations. The preset number of occurrences refers to the minimum number of times the anomaly must recur within a continuous segment of a single detection or during multiple detection processes. By using preset ranges and preset number of occurrences, the server can avoid misjudging fixed characteristics due to occasional noise. The expression for determining a fixed characteristic is:

[0088] in, Indicates the location of the subway overhead contact line Do the corresponding multi-sensor synchronization anomalies have fixed characteristics? Indicates whether the abnormal state of multi-sensor synchronization has movement characteristics; This indicates the number of detections or the number of detection segments involved in the fixed feature judgment. This parameter is determined by the current detection task and historical retest data. Indicates the first The line mileage corresponding to the location of the anomaly in this test is obtained by matching the train's running location data and the catenary structure information. Indicates the location of the subway overhead contact line The corresponding line mileage is obtained from the overhead contact line structure information; This parameter represents the allowable deviation of the line mileage corresponding to the preset range of the same contact wire component. This parameter is set according to the positioning error, the sensor field of view, and the spatial dimensions of the contact wire component. Indicates the first The location of the subway overhead contact line during the second inspection Is there an abnormal state in the multi-sensor synchronization? This parameter represents the preset number of times a fixed feature is required. This parameter can be preset according to the reliability requirements of the retest. The expression determines the fixed feature by excluding moving features and requiring the anomaly to repeat a preset number of times within the same subway catenary location or the same catenary component.

[0089] After determining that the multi-sensor synchronization anomaly has fixed characteristics, the server further determines whether the subway catenary location exhibits the same or similar visible light edge changes, laser contour changes, infrared temperature changes, and ultraviolet bright spot changes during multiple detections. The same or similar changes refer to the consistent enhancement direction, similar intensity, similar peak location, similar duration, or similar recovery process of the four types of changes across multiple detections. The server can construct fixed anomaly change morphology vectors for each of the four types of changes from multiple detections and calculate the similarity of change morphology between each detection. If the similarity reaches a preset fixed similarity threshold, it indicates that the anomaly is not an occasional interference but is stably fixed at the same subway catenary location or near the same catenary component. The expression for calculating the similarity of change morphology across multiple detections is:

[0090] in, Indicates the location of the subway overhead contact line The similarity of fixed changes in morphology during multiple detection processes; Indicates the number of tests included in the comparison; Indicates the first The location of the subway overhead contact line during the second inspection The corresponding change pattern data consists of visible light edge changes, laser profile changes, infrared temperature changes, and ultraviolet bright spot changes; Indicates the first The location of the subway overhead contact line during the second inspection Corresponding change pattern data; Indicates the first The second test and the first The similarity of change patterns between multiple detections is calculated using the expression for change pattern similarity. This expression determines whether anomalies at the same subway catenary location exhibit stable and recurring change patterns by averaging the pairwise similarities between multiple detections.

[0091] When the subway catenary exhibits similar or identical visible light edge changes, laser profile changes, infrared temperature changes, and ultraviolet bright spot changes during multiple detections, and the multi-sensor synchronous anomaly state exhibits fixed characteristics, the server classifies the multi-sensor synchronous anomaly state as a fixed anomaly state. A fixed anomaly state refers to an anomaly repeatedly occurring around the same subway catenary location or the same catenary component, exhibiting similar changes in multiple detections. This typically corresponds more likely to insulation aging, catenary structural loosening, or hard spot anomalies. The comprehensive judgment expression for a fixed anomaly state is:

[0092] in, Indicates the location of the subway overhead contact line Whether the corresponding multi-sensor synchronization anomaly is determined to be a fixed anomaly; Indicates whether the abnormal state of multi-sensor synchronization has fixed characteristics; This indicates the similarity of the fixed changes in the location of the subway catenary during multiple inspections; This indicates a preset fixed similarity threshold, which is pre-determined based on historical defect samples, retest consistency requirements, and sensor stability. This expression uses fixed features and the consistency of changes in multiple detections to jointly determine fixed abnormal states, enabling further differentiation between incidental interference, positional errors, and genuine fixed defects.

[0093] S150. Based on the abnormal state of movement, identify the attitude interference of the detection equipment, the contamination of the detection window and the airflow interference in the tunnel, and based on the abnormal state of fixation, identify the state of insulation aging, the state of loose contact wire structure and the state of hard point abnormality.

[0094] Specifically, when identifying attitude interference from the detection equipment, the server first reads the abnormal start and continuation positions corresponding to the abnormal movement state, and extracts laser profile changes, visible light edge changes, vehicle body attitude changes, and pantograph-catenary contact current changes from the detection data. Attitude interference from the detection equipment refers to pitch, roll, vertical vibration, or lateral vibration changes that occur during the operation of the detection train, causing changes in the observation angles of the visible light and laser acquisition devices installed on the train. This leads to a synchronous shift in visible light edge and laser profile changes, but this shift is not caused by defects in the catenary itself. When judging attitude interference from the detection equipment, the server not only determines whether the vehicle body attitude change is enhanced, but also whether there is a synchronous change relationship between the vehicle body attitude change and the laser profile and visible light edge changes, and whether the pantograph-catenary contact current change is not synchronously enhanced. If the pantograph-catenary contact current change does not show abrupt changes, amplified fluctuations, or delayed recovery, it indicates that there is no obvious actual contact disturbance between the pantograph and the contact wire, thus further attributing the abnormal movement state to attitude interference from the detection equipment. The matching expression for attitude interference from the detection equipment is:

[0095] in, Indicates the location of the subway overhead contact line The degree of attitude interference matching of the corresponding detection equipment; Indicates the location of the subway overhead contact line The corresponding laser profile change is calculated from the profile height offset, lateral profile offset, profile discontinuity, and echo intensity change in the laser profile data. Indicates the location of the subway overhead contact line The corresponding visible light edge variation is calculated from the edge position, edge sharpness, and edge continuity in the visible light image. Indicates the location of the subway overhead contact line The corresponding change in vehicle attitude is calculated from the pitch change, roll change, vertical vibration change and lateral vibration change in the vehicle attitude data; Indicates the location of the subway overhead contact line The corresponding change in pantograph-catenary contact current is calculated from the current abrupt change, fluctuation amplification, and recovery delay in the pantograph-catenary contact current data. Indicates the location of the subway overhead contact line The corresponding threshold for judging changes in pantograph-catenary contact current is obtained from historical normal detection data. This parameter indicates the degree of synchronous correlation between two change results within the same detection segment. It can be determined by the proximity of the change times and the consistency of the change directions of the two change results. , , The weighting coefficients for the laser profile change related terms, visible light edge change related terms, and pantograph contact current change suppression terms can all be preset based on historical test samples and equipment installation stability. This expression represents an extremely small positive number to prevent the denominator from being zero. It enhances the interpretative contribution of vehicle attitude changes to laser profile and visible light edge variations, and suppresses the influence of actual contact disturbances through changes in the pantograph-catenary contact current, enabling the server to distinguish between attitude interference from the detection equipment and defects in the catenary itself.

[0096] When changes in the laser profile and visible light edge change synchronously with changes in vehicle attitude, and the changes in the pantograph-catenary contact current do not increase synchronously, the server identifies the abnormal movement state as attitude interference of the detection equipment and marks the changes in the laser profile and visible light edge as attitude-related changes in the detection data. Synchronous change means that the start time, enhancement time, or peak time of the vehicle attitude change is close to the corresponding time of the changes in the laser profile and visible light edge, and the directions of these three changes can be mutually explained. For example, after a pitch change in the vehicle, the laser profile height shifts overall, and the visible light edge position shifts synchronously in the image direction. Non-synchronous enhancement means that the changes in the pantograph-catenary contact current do not exceed the corresponding judgment threshold, or although the changes in the pantograph-catenary contact current exhibit normal fluctuations, they do not form abrupt changes, amplified fluctuations, or delayed recovery corresponding to changes in the laser profile and visible light edge. After the server completes the identification, it does not directly write this abnormal movement state into the catenary defect statistics object, but saves it as an anomaly of the detection platform, enabling subsequent processing strategies to trigger attitude compensation, equipment installation verification, or processing to reduce the reliability of the corresponding detection segment.

[0097] When identifying detection window contamination, the server first reads the abnormal continuation positions of the moving anomaly across multiple adjacent contact wire locations and compares whether the visible light edge changes and laser contour changes at each abnormal continuation position maintain similar change patterns. Detection window contamination refers to the impact of water mist, dust, oil, carbon powder, or condensation on the detection window in front of the visible light or laser acquisition equipment, causing continuous edge blurring, partial occlusion, brightness attenuation, echo weakening, discontinuous contours, or contour diffusion at multiple adjacent contact wire locations. Because detection window contamination moves with the inspection train, its effects maintain similar change patterns across multiple adjacent contact wire locations, rather than being fixed near a single insulator, segment insulator, or contact wire connection. The server also needs to determine whether the infrared temperature changes and ultraviolet bright spot changes are not fixed within the range of the insulator, segment insulator, or contact wire connection; if the infrared temperature changes and ultraviolet bright spot changes are not stably bound within the range of these components, it indicates that the moving anomaly is more consistent with detection window contamination than with insulation aging on fixed components. The matching expression for detection window contamination is:

[0098] in, Indicates the starting position of the exception The corresponding degree of contamination matching in the detection window; This represents the set of locations where abnormalities continue, consisting of adjacent contact wire locations where abnormal movement occurs continuously along the train's direction of travel; This parameter represents the similarity of visible light edge changes and laser profile changes in the set of anomalous continuation locations. It is calculated from the similarity between the degree of edge blurring, edge position offset, profile discontinuity, and echo changes in each anomalous continuation location. This parameter represents the degree of blurring or occlusion in a set of abnormally continuous locations in a visible light image. It is calculated from the decrease in image sharpness, the proportion of local occlusion, and the brightness attenuation. This parameter represents the degree of echo attenuation of laser profile data in the set of abnormal continuity locations. It is calculated from the decrease in echo intensity, the reduction in effective profile points, and the proportion of profile discontinuity. This parameter indicates the degree to which infrared temperature changes and ultraviolet bright spot changes are fixed within the range of insulators, segmented insulators, or contact wire connections. It is calculated from the number of times the infrared temperature changes and ultraviolet bright spot changes coincide with the component range in the contact wire structure information, the stability of the coincidence position, and the recurrence rate. , , , These represent the weighting coefficients for the similarity of change morphology, image contamination, echo attenuation, and fixed discharge heating suppression terms, respectively. This expression enables the server to identify contamination in the detection window by emphasizing similar image anomalies and similar contour anomalies in multiple adjacent contact wire locations and weakening infrared temperature changes and ultraviolet bright spot changes fixed within the insulation component range.

[0099] When visible light edge variations and laser contour variations maintain similar patterns across multiple adjacent contact network locations, and infrared temperature variations and ultraviolet bright spot variations are not fixed within the range of insulators, segmented insulators, or contact network connections, the server identifies the abnormal movement as detection window contamination and marks the corresponding continuation location of the anomaly as the location affected by window contamination. The range of insulators, segmented insulators, or contact network connections refers to the spatial neighborhood of the component determined based on its installation location, component size, and positioning tolerance in the contact network structure information. This range is used to determine whether infrared temperature variations or ultraviolet bright spot variations truly fall on the corresponding component. If infrared temperature variations and ultraviolet bright spot variations do not recur stably within this range, it indicates that the abnormal movement lacks fixed component anomaly characteristics. After completing the identification, the server reduces the reliability of defect judgments regarding visible light edge variations and laser contour variations within the window contamination-affected location, while retaining infrared temperature variations and ultraviolet bright spot variations as a verification reference. This avoids misjudging continuous image anomalies and continuous contour anomalies caused by detection window contamination as simultaneous anomalies in multiple contact network components.

[0100] When identifying tunnel airflow interference, the server reads the train speed data corresponding to the abnormal movement state and determines whether the changes in laser contour and visible light edge appear earlier, later, or have a different duration as the train speed data changes. Tunnel airflow interference refers to the pressure waves, wake disturbances, or water mist disturbances generated by the detected train during its operation in the tunnel, causing short-term disturbances to the visible light acquisition path and laser scanning path, thus leading to temporal changes in visible light edge and laser contour changes with the train speed data. "Early appearance" means that the anomaly appears earlier in adjacent contact wire locations when the train speed data increases; "later appearance" means that the anomaly's appearance time is delayed relative to the normal passage time after the train speed data changes; "changed duration" means that the time length between the start of the anomaly and the recovery time changes regularly with the train speed data. The server also determines whether the changes in vehicle posture are insufficient to explain the changes in laser contour and visible light edge. If the changes in vehicle posture are weak or their direction is inconsistent with the direction of image and contour changes, the anomaly is more likely to originate from tunnel airflow interference. The matching expression for tunnel airflow interference is:

[0101] in, Indicates the location of the subway overhead contact line The corresponding degree of tunnel airflow interference matching; This indicates the location where the inspection train passes through the subway overhead contact line. The train's operating speed data at that time is obtained from the train control system, odometer, or positioning system. This parameter represents the combined duration of visible light edge changes and laser profile changes. It is determined by combining the start and recovery times of the visible light edge changes and laser profile changes. This parameter represents the start time of the combined change of visible light edge and laser profile. It is determined by the earlier of the two start times or the statistically representative time of both. Indicates changes in vehicle attitude; This parameter represents the combined intensity of visible light edge variation and laser profile variation; it is obtained by weighting the intensity of both variations. It indicates the degree of correlation between two data sequences. This parameter is calculated from corresponding data from multiple adjacent catenary locations or multiple detection processes. , , These represent the weighting coefficients for the duration-based velocity-related term, the start-time velocity-related term, and the vehicle attitude interpretation suppression term, respectively. This expression enhances the interpretation of anomalous temporal changes in images and contours by train speed data, while weakening the interpretative contribution of vehicle attitude changes, enabling the server to distinguish tunnel airflow interference from detection equipment attitude interference.

[0102] When changes in laser profile and visible light edge appear earlier, later, or have altered duration depending on train speed data, and the vehicle attitude change is insufficient to explain these changes, the server identifies the abnormal movement as tunnel airflow interference. Insufficient vehicle attitude change means the change does not exceed the attitude judgment threshold, or although the change exists, its direction, start time, and duration cannot correspond to the laser profile and visible light edge changes. After identifying tunnel airflow interference, the server writes the corresponding anomaly start and continuation positions into the detection data, and retains the train speed data, the combined change start time, and the combined duration. This allows for interference verification when the same anomaly reappears in the same section under similar train speed data, rather than directly including the anomaly in the catenary defect statistics.

[0103] When identifying insulation aging conditions, the server first reads the location and components of the subway contact network corresponding to the fixed abnormal state, and then determines whether the fixed abnormal state is fixed within the range of insulators, segment insulators, or supporting devices based on the contact network structure information. Insulation aging refers to early abnormal states near insulators, segment insulators, or supporting devices caused by a decline in insulation performance, resulting in weak corona discharge, localized heating, or surface contamination accumulation. Since early insulation aging may only manifest as slight dirt, slight discoloration, or blurred edges in visible light images, the server does not use obvious visible light edge changes as an exclusion criterion. Instead, it focuses on whether changes in ultraviolet bright spots and infrared temperature recur in multiple detection processes and whether they stably fall within the range of insulators, segment insulators, or supporting devices. The matching expression for insulation aging conditions is:

[0104] in, Indicates the location of the subway overhead contact line The degree of matching of the corresponding insulation aging state; This parameter indicates whether the fixed abnormal state is fixed within the range of insulators, segmented insulators, or support devices. It is determined by the overlap between the location of the ultraviolet bright spot change and the infrared temperature change and the component range in the contact network structure information. This indicates the number of times the change in ultraviolet bright spots reappears at the location of the subway catenary during multiple tests. This parameter is obtained by statistical analysis of historical and current test data. This indicates the number of times the infrared temperature change occurred repeatedly at the location of the subway contact network during multiple detections. This parameter is obtained by statistical analysis of the detection records corresponding to the infrared thermal images. This parameter represents the degree of positional and temporal similarity between changes in ultraviolet bright spots and changes in infrared temperature. It is calculated from the degree of overlap between their corresponding positions, the number of repetitions, and the proximity of their change times. It indicates the change in pantograph-catenary contact current, which is used to suppress the interference of significant mechanical contact disturbances on the assessment of insulation aging; , , , , These represent the weighting coefficients for the component location term, ultraviolet (UV) repetition term, infrared (IR) repetition term, UV-IR correlation term, and pantograph-catenary contact current suppression term, respectively. This expression characterizes the early insulation aging state through fixed component range, UV repetition, IR repetition, and UV-IR correlation, and avoids misjudging mechanical shock anomalies as insulation aging states.

[0105] When a fixed abnormal state is located within the insulator, segmented insulator, or support device, and the changes in ultraviolet bright spots and infrared temperature repeatedly occur during multiple detection processes, the server identifies the fixed abnormal state as an insulation aging state. Multiple detection processes can include multiple pass records from the current detection task, or retest records of the same subway catenary location in the historical detection database. Repetition means that the changes in ultraviolet bright spots and infrared temperature occur more than a preset number of times within a preset range at the same subway catenary location or the same catenary component, and the change patterns are similar. After the server completes the identification, it marks the subway catenary location as a location of concern for insulation aging and writes the changes in ultraviolet bright spots, infrared temperature changes, corresponding catenary components, and retest requirements into the detection data, enabling subsequent dynamic detection results of the subway catenary to provide early indications of insulation aging.

[0106] When identifying a loose contact wire structure, the server first determines whether the abnormal fixation is within the range of the positioner, dropper, anchor joint, or contact wire connection, and then reads the sequence of changes in pantograph-catenary contact current and laser profile. A loose contact wire structure refers to a localized mechanical response in the positioner, dropper, anchor joint, or contact wire connection when the pantograph passes by, causing a change in pantograph-catenary contact current to appear first, followed by short-term offset, short-term rebound, or profile jitter in the laser profile data. When determining this state, the server focuses on whether the change in pantograph-catenary contact current precedes the change in laser profile, excluding the explanation of changes in vehicle attitude for the laser profile change. If changes in vehicle attitude are insufficient to explain the laser profile change, and the change in pantograph-catenary contact current precedes the change in laser profile, it indicates that the contact wire components may have experienced a localized loosening response due to the pantograph-catenary action. The matching expression for a loose contact wire structure is:

[0107] in, Indicates the location of the subway overhead contact line The degree of matching between the corresponding loose state of the overhead contact line structure; This parameter indicates whether the abnormal status is fixed within the range of the positioner, dropper, anchor joint, or contact wire connection. This parameter is determined by whether the abnormal location overlaps with the corresponding component range in the contact wire structure information. This indicates the moment when the change in pantograph-catenary contact current begins; this parameter is determined from the pantograph-catenary contact current data. This parameter indicates the start time of the change in the laser profile and is determined by the laser profile data. This represents a conditional function; Indicates changes in the laser profile; This indicates the change in pantograph-catenary contact current; Indicates changes in vehicle attitude; This indicates the degree to which changes in vehicle attitude synchronously interpret changes in laser profile; , , , , These represent the weighting coefficients for the mechanical component position term, the sequence relationship term, the laser profile change term, the pantograph-catenary contact current change term, and the vehicle attitude interpretation and suppression term, respectively. This expression uses the range of mechanical components, the precedence of pantograph-catenary contact current changes, the response to laser profile changes, and the vehicle attitude exclusion to jointly determine the loosening state of the catenary structure.

[0108] When a fixed abnormality is located within the area of ​​the locator, dropper, anchor joint, or contact wire connection, and the change in pantograph-catenary contact current precedes the change in laser profile, the server identifies the fixed abnormality as a loose contact wire structure. The area of ​​the locator, dropper, anchor joint, or contact wire connection is determined by the installation location, component dimensions, line mileage, and connection relationships in the contact wire structure information. The change in pantograph-catenary contact current preceding the change in laser profile indicates that the contact state between the pantograph and the contact wire is disturbed first, followed by a spatial profile response from the corresponding mechanical structure. After the server completes the identification, it marks the corresponding subway contact wire location as a location of concern for structural loosening and retains the order of the change in pantograph-catenary contact current preceding the change in laser profile, the corresponding contact wire components, and the retest requirements in the detection data. This allows subsequent maintenance personnel to focus on reviewing the locator, dropper, anchor joint, or contact wire connection.

[0109] When identifying hard-point anomalies, the server first determines whether the fixed anomaly is confined to the contact wire, locator, or dropper, and then further determines whether the changes in pantograph-catenary contact current, laser profile, and visible light edge are continuously enhanced. A hard-point anomaly refers to a momentary mechanical impact caused when the pantograph passes over a local unevenness, a sudden change in local stiffness, or a local abnormal position of the contact wire, resulting in a sequential or synchronous enhancement of the changes in pantograph-catenary contact current, laser profile, and visible light edge within a short period. Continuous enhancement means that the enhancement of the changes in pantograph-catenary contact current, laser profile, and visible light edge occurs within the same short time window, and the intensity of all three exceeds the corresponding enhancement threshold; this continuous enhancement reflects the continuous anomaly relationship between the current response, profile response, and visual edge response. The matching expression for hard-point anomalies is:

[0110] in, Indicates the location of the subway overhead contact line The degree of matching of the corresponding hardpoint abnormal state; This parameter indicates whether the abnormal condition is fixed within the range of the contact wire, positioner, or dropper. This parameter is determined by whether the abnormal location overlaps with the corresponding component range in the contact wire structure information. This parameter represents the degree of temporal concentration of the changes in pantograph contact current, laser profile, and visible light edge. It is determined by the difference between the latest and earliest times of the three changes. This indicates the continuous enhancement time window, which is preset based on the sensor sampling frequency, train speed data, and the duration characteristics of hard point impact. This indicates the change in pantograph-catenary contact current; Indicates changes in the laser profile; Indicates changes in the visible light edge; This parameter indicates whether the changes in infrared temperature and ultraviolet bright spots exhibit fixed and repetitive characteristics of the insulating component. It is determined by whether the changes in infrared temperature and ultraviolet bright spots are fixed within the range of the insulator, segmented insulator, or support device and occur repeatedly. , , , These represent the weighting coefficients for the hard spot location term, continuous enhancement term, three types of intensity variation term, and insulation anomaly suppression term, respectively. This represents an extremely small positive number to prevent the denominator from being zero. The expression uses a combination of factors, including the range of contact wire-related components, visually enhanced current profile, and exclusion of insulation anomalies, to determine the abnormal state of hard spots.

[0111] When a fixed anomaly is confined to the area of ​​the contact wire, positioner, or dropper, and the changes in pantograph-catenary contact current, laser profile, and visible light edge reflect a continuous enhancement, the server identifies the fixed anomaly as a hard point anomaly. The area of ​​the contact wire, positioner, or dropper is determined by the installation location, component dimensions, line mileage, and connection relationships in the catenary structure information. Continuous enhancement indicates that the electrical contact disturbance reflected by the pantograph-catenary contact current change, the spatial profile response reflected by the laser profile change, and the appearance edge response reflected by the visible light edge change form a correlation within a short period of time. After the server completes the identification, it marks the corresponding subway catenary location as a hard point anomaly of interest and saves the pantograph-catenary contact current change, laser profile change, visible light edge change, continuous enhancement time window, and corresponding catenary component in the detection data, so that subsequent dynamic detection results of the subway catenary can indicate the risk of a hard point anomaly at that location.

[0112] When a server simultaneously meets the conditions of multiple anomaly source types under the same mobile or fixed anomaly state, it determines the anomaly source based on its explanatory power. For mobile anomalies, if the vehicle attitude change can explain the laser profile change and visible light edge change, it is preferentially identified as detection equipment attitude interference; if the vehicle attitude change is insufficient to explain it, and there are similar edge blurring, partial occlusion, brightness attenuation, echo weakening, profile discontinuity, or profile diffusion among multiple adjacent contact wire positions, it is identified as detection window contamination; if neither of the above is sufficient to explain it, and the laser profile change and visible light edge change occur earlier, later, or with a change in duration depending on the train speed data, it is identified as tunnel airflow interference. For fixed abnormal conditions, if the abnormality is fixed within the range of insulators, segmented insulators, or support devices, and the changes in ultraviolet bright spots and infrared temperature changes occur repeatedly, it is preferentially identified as an insulation aging state; if the abnormality is fixed within the range of positioners, droppers, anchor joints, or contact wire connections, and the changes in pantograph-catenary contact current precede the changes in laser profile, it is identified as a loose contact wire structure; if the abnormality is fixed within the range of contact wires, positioners, or droppers, and the changes in pantograph-catenary contact current, laser profile, and visible light edge changes form a continuous enhancement, it is identified as a hard spot abnormal condition.

[0113] The server can compare the matching degree of each exception source type to obtain the final exception source type. The selection expression for the final exception source type is: in, Indicates the location of the subway overhead contact line The corresponding final exception source type; This represents the set of candidate anomaly source types, specifically including interference with the attitude of the detection equipment, contamination of the detection window, interference with tunnel airflow, insulation aging, loose contact wire structure, and hard point anomaly. Represents the first in the set of candidate anomaly source types. One type of anomaly source; Indicates the first The types of anomalies are located in the subway overhead contact system. The degree of matching at each location is calculated from the matching expressions for the attitude interference of the detection equipment, the pollution of the detection window, the airflow interference of the tunnel, the insulation aging state, the loose state of the contact wire structure, or the abnormal state of the hard point. Indicates the first The pre-state constraints corresponding to each anomaly source type are as follows: when the moving anomaly state corresponds to the attitude interference of the detection equipment, the contamination of the detection window, or the airflow interference in the tunnel, the value is 1; otherwise, the value is 0. When the fixed anomaly state corresponds to the insulation aging state, the loose state of the contact wire structure, or the hard point anomaly state, the value is 1; otherwise, the value is 0. This expression selects the anomaly source type that maximizes the value within the parentheses. It determines the final anomaly source type by combining the degree of matching with pre-existing state constraints, avoiding misidentification of mobile anomalies as fixed defects or fixed anomalies as vehicle-mounted interference.

[0114] S160. Based on the interference of the detection equipment posture, contamination of the detection window, interference of tunnel airflow, insulation aging status, loose status of the contact network structure, and abnormal status of hard points, corresponding processing strategies are generated, and the detection data is updated according to the processing strategies, so as to generate corresponding dynamic detection results of the subway contact network based on the updated detection data.

[0115] Specifically, after obtaining the anomaly source type, the server first binds the anomaly source type with the corresponding subway catenary location, catenary component, multi-sensor change sequence, moving anomaly state, or fixed anomaly state. The anomaly source type refers to the previously identified issues such as detection equipment posture interference, detection window contamination, tunnel airflow interference, insulation aging state, catenary structure loosening state, or hard point anomaly state; the catenary defect statistics object refers to the data object that ultimately needs to be included in catenary defect statistics, retesting, maintenance prompts, or risk assessment; the processing strategy refers to the differentiated data update rules generated for different anomaly source types. After the server completes the binding, it ensures that each anomaly source type corresponds to a specific subway catenary location, catenary component, and sensor change result, thereby avoiding generalized processing of the entire line.

[0116] When the anomaly source type is detection equipment attitude interference, the server generates a detection equipment attitude correction strategy. Detection equipment attitude interference refers to non-contact network defect anomalies caused by changes in the train's attitude leading to changes in the observation angles of the visible light acquisition equipment and the laser acquisition equipment, resulting in synchronous shifts in visible light edge changes and laser profile changes. The detection equipment attitude correction strategy includes retaining the train's attitude changes in the corresponding detection segment, reducing the reliability of visible light edge changes and laser profile changes in defect judgment, marking visible light images and laser profile data as attitude-affected data, and excluding the corresponding subway contact network location from the contact network defect statistics. Through this processing, the server will not incorrectly output the contact network structure loosening state or hard point anomaly state due to the observation shift caused by the train's own attitude changes.

[0117] When the anomaly source type is "detection window contamination," the server generates a detection window contamination handling strategy. Detection window contamination refers to the impact of water mist, dust, oil, toner, or condensation on the transparent window in front of the visible light or laser acquisition equipment, causing similar edge blurring, partial occlusion, brightness attenuation, reduced echo, discontinuous outline, or outline diffusion at multiple adjacent contact wire locations. The detection window contamination handling strategy includes marking the corresponding multiple subway contact wire locations as window contamination-affected locations, reducing the reliability of defect judgments based on visible light edge changes and laser outline changes at window contamination-affected locations, and excluding visible light edge changes and laser outline changes from the contact wire defect statistics. Simultaneously, the server retains infrared temperature changes and ultraviolet bright spot changes as verification references to subsequently determine whether there are any fixed component anomalies unrelated to detection window contamination.

[0118] When the anomaly source type is tunnel airflow interference, the server generates a tunnel airflow interference handling strategy. Tunnel airflow interference refers to the pressure waves, wake disturbances, or water mist disturbances generated by the detected train during its operation within the tunnel, causing short-term disturbances to the visible light acquisition path and laser scanning path, resulting in contour shift, edge jitter, or short-term imaging blur. The tunnel airflow interference handling strategy includes recording the train speed data, anomaly start location, anomaly continuation location, contour shift, edge jitter, and short-term imaging blur at the corresponding subway catenary location, and marking the contour shift, edge jitter, or short-term imaging blur related to the train speed data as airflow interference impact results. The server excludes the airflow interference impact results from the catenary defect statistics object, but retains the correlation record between them and the train speed data, so that interference verification can be performed when similar anomalies occur again in the same section at similar train speeds.

[0119] When the anomaly source type is insulation aging, the server generates an insulation aging processing strategy. Insulation aging refers to early contact network defects near insulators, segment insulators, or support devices, characterized by changes in ultraviolet bright spots and infrared temperature due to deterioration in insulation performance. The insulation aging processing strategy includes retaining the ultraviolet bright spot and infrared temperature changes at the corresponding subway contact network location, recording the recurrence of these changes in multiple detection processes, and marking the corresponding subway contact network location as an insulation aging concern. Even if the visible light edge change only manifests as slight contamination, slight discoloration, or blurred edges, the server will not use this weak visible light response as a basis for excluding insulation aging, but rather as an auxiliary description of early insulation aging in the detection data.

[0120] When the anomaly source type is "loose contact wire structure," the server generates a loose contact wire structure handling strategy. A loose contact wire structure refers to a localized mechanical response occurring in the positioner, dropper, anchor joint, or contact wire connection after the pantograph passes, causing a change in the pantograph-catenary contact current first, followed by a change in the laser profile near the same contact wire component, manifesting as short-term offset, short-term rebound, or profile jitter. The loose contact wire structure handling strategy includes preserving the order in which the pantograph-catenary contact current change precedes the laser profile change, recording the short-term offset, short-term rebound, or profile jitter manifestations of the laser profile change, and marking the corresponding subway contact wire location as a location of concern regarding structural loosening. The server also retains the component number and connection relationship of the corresponding contact wire component, enabling subsequent inspection and repair checks to pinpoint the specific positioner, dropper, anchor joint, or contact wire connection.

[0121] When the anomaly source type is a hard spot anomaly, the server generates a hard spot anomaly handling strategy. A hard spot anomaly refers to localized unevenness, sudden changes in local stiffness, or pantograph-catenary impact points near the contact wire, positioner, or dropper, causing a continuous enhancement of changes in pantograph-catenary contact current, laser profile, and visible light edge variations within a short period. The hard spot anomaly handling strategy includes preserving the continuous enhancement relationship between changes in pantograph-catenary contact current, laser profile, and visible light edge variations; recording current abrupt changes, fluctuation amplification, recovery delay, short-term profile enhancement, and short-term edge enhancement at the corresponding subway contact wire location; and marking the corresponding subway contact wire location as a hard spot anomaly of interest. The server can also record the fact that infrared temperature changes and ultraviolet bright spot changes do not form fixed, repetitive features of the insulation components to avoid confusing hard spot anomalies with insulation aging conditions.

[0122] When updating detection data according to various processing strategies, the server uniformly marks anomalies belonging to detection equipment attitude interference, detection window contamination, and tunnel airflow interference as non-contact network defect anomalies. Non-contact network defect anomalies refer to abnormal results caused by the detection platform, detection window, or tunnel environment, and do not inherently indicate a real defect in the contact network structure. For non-contact network defect anomalies, the server retains the anomaly source type, affected sensor data, anomaly start location, anomaly continuation location, and cause of elimination when updating detection data, but does not write the corresponding anomalies into the contact network defect statistics object, thereby reducing the false alarm rate and providing a basis for detection equipment maintenance or environmental interference verification.

[0123] When updating detection data according to various processing strategies, the server uniformly marks anomalies belonging to insulation aging, loose contact network structure, and hard point abnormalities as contact network defect anomalies. Contact network defect anomalies refer to abnormal results fixed within the contact network component area and possessing structural, electrical, thermal, or discharge-related characteristics, which need to enter subsequent defect statistics, retesting, or maintenance prompt processes. For contact network defect anomalies, the server records the corresponding subway contact network location, contact network component, anomaly source type, multi-sensor change sequence, key change results, retesting requirements, and attention location markers when updating detection data, updating the detection data from the original sensor data set to data results containing anomaly source explanations and maintenance directions.

[0124] When the server generates dynamic inspection results for the subway catenary based on updated inspection data, it categorizes the output into non-catenary defect anomalies, catenary defect anomalies, and anomalies requiring retesting. For non-catenary defect anomalies, the output includes the corresponding subway catenary location, anomaly source type, affected sensor data, and reason for elimination. For catenary defect anomalies, the output includes the corresponding subway catenary location, catenary components, anomaly source type, key changes, areas of concern, and retesting requirements. For anomalies that cannot yet be stably categorized, the output includes the anomalies requiring retesting and the types of inspection data that need to be re-collected. Therefore, the dynamic inspection results for the subway catenary not only indicate the existence of anomalies but also whether the anomalies are genuine catenary defects, the source of the anomalies, which sensor data were retained or excluded, and whether subsequent equipment maintenance, environmental verification, or catenary repair and retesting should be performed.

[0125] This application also provides a dynamic detection device for subway overhead contact lines based on multi-sensor collaboration, referring to... Figure 2 , Figure 2This application provides a schematic diagram of a multi-sensor collaborative dynamic detection device for subway catenary, which is a server. The server includes an acquisition module 21 and a processing module 22. The acquisition module 21 is used to acquire visible light images, laser contour data, infrared thermal images, ultraviolet discharge images, pantograph-catenary contact current data, vehicle posture data, train running position data, and catenary structure information corresponding to the location of the subway catenary, to form detection data. The processing module 22 is used to extract the sequence of changes of the multiple sensors corresponding to the location of the subway catenary based on the detection data. The processing module 22 is also used to compare the sequence of changes of the multiple sensors with the normal sequence of changes formed in the historical normal detection process, and to simultaneously enhance the changes in visible light edge, laser contour, infrared temperature, and ultraviolet bright spots, as well as the changes in pantograph-catenary contact current. When the changes in the vehicle body attitude are not synchronized, it is determined that there is a multi-sensor synchronization anomaly at the corresponding subway catenary location. The processing module 22 is also used to determine the moving anomaly and the fixed anomaly based on the persistence of the multi-sensor synchronization anomaly between adjacent catenary locations. The processing module 22 is also used to identify the attitude interference of the detection equipment, the contamination of the detection window, and the airflow interference in the tunnel based on the moving anomaly, and to identify the insulation aging state, the loosening state of the catenary structure, and the hard point anomaly based on the fixed anomaly. The processing module 22 is also used to generate corresponding processing strategies based on the attitude interference of the detection equipment, the contamination of the detection window, the airflow interference in the tunnel, the insulation aging state, the loosening state of the catenary structure, and the hard point anomaly, and to update the detection data according to the processing strategies, so as to generate the corresponding dynamic detection results of the subway catenary based on the updated detection data.

[0126] This application also provides an electronic device, with reference to... Figure 3 , Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include: at least one processor 31, at least one network interface 34, a user interface 33, a memory 35, and at least one communication bus 32.

[0127] The communication bus 32 is used to enable communication between these components.

[0128] The user interface 33 may include a display screen and a camera. Optionally, the user interface 33 may also include a standard wired interface and a wireless interface.

[0129] The network interface 34 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0130] The processor 31 may include one or more processing cores. The processor 31 connects to various parts of the server via various interfaces and lines, executing instructions, programs, code sets, or instruction sets stored in the memory 35, and calling data stored in the memory 35 to perform various server functions and process data. Optionally, the processor 31 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 31 may integrate one or a combination of several of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content to be displayed on the screen; and the modem handles wireless communication. It is understood that the modem may also not be integrated into the processor 31 and may be implemented as a separate chip.

[0131] The memory 35 may include random access memory (RAM) or read-only memory. Optionally, the memory 35 may include a non-transitory computer-readable storage medium. The memory 35 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 35 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. Optionally, the memory 35 may also be at least one storage device located remotely from the aforementioned processor 31. Figure 3 As shown, the memory 35, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and an application program for a dynamic detection method of subway catenary based on multi-sensor collaboration.

[0132] exist Figure 3In the electronic device shown, the user interface 33 is mainly used to provide an input interface for the user and to obtain the user input data; while the processor 31 can be used to call an application program stored in the memory 35 that is a dynamic detection method for subway catenary based on multi-sensor collaboration. When executed by one or more processors, the electronic device executes one or more methods as described in the above embodiments.

[0133] This application also provides a non-transitory computer-readable storage medium storing instructions. When executed by one or more processors, these instructions cause an electronic device to perform one or more of the methods described in the above embodiments.

[0134] The foregoing description is merely an exemplary embodiment of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure. Those skilled in the art will readily conceive of other embodiments of this disclosure upon considering the specification and the disclosure of practical truth. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described in this disclosure. The specification and embodiments are considered exemplary only, and the scope and spirit of this disclosure are defined by the claims.

Claims

1. A dynamic detection method for subway overhead contact lines based on multi-sensor collaboration, characterized in that, The method includes: The system acquires visible light images, laser contour data, infrared thermal images, ultraviolet discharge images, pantograph-catenary contact current data, vehicle attitude data, train running position data, and catenary structure information corresponding to the location of the subway catenary to form detection data. Based on the detection data, extract the sequence of changes from multiple sensors corresponding to the location of the subway overhead contact line; The sequence of changes in the multi-sensor is compared with the sequence of changes formed during the historical normal detection process. When the changes in visible light edge, laser contour, infrared temperature and ultraviolet bright spots are enhanced simultaneously, and the changes in pantograph contact current and vehicle posture are not enhanced synchronously, it is determined that there is a multi-sensor synchronous abnormal state at the corresponding subway catenary location. Based on the persistence of the multi-sensor synchronization anomaly between adjacent contact wire positions, the moving anomaly and the fixed anomaly are determined. Based on the described abnormal movement status, the device's posture interference, detection window contamination, and tunnel airflow interference are identified; and based on the described abnormal fixed status, the insulation aging status, contact wire structure loosening status, and hard point abnormal status are identified. Based on the attitude interference of the detection equipment, the contamination of the detection window, the airflow interference in the tunnel, the aging state of the insulation, the loose state of the contact network structure, and the abnormal state of the hard points, a corresponding processing strategy is generated, and the detection data is updated according to the processing strategy, so as to generate the corresponding dynamic detection results of the subway contact network based on the updated detection data.

2. The method for dynamic detection of subway overhead contact lines based on multi-sensor collaboration according to claim 1, characterized in that, The acquisition of visible light images, laser contour data, infrared thermal images, ultraviolet discharge images, pantograph-catenary contact current data, vehicle attitude data, train running position data, and catenary structure information corresponding to the location of the subway catenary, to form detection data, specifically includes: Read the installation position, component number, line mileage, and connection relationship of the contact wire, dropper, positioner, insulator, segment insulator, anchor joint, and support device from the contact network structure information, and bind the contact network structure information with the train running position data; Based on the train's operating location data, the mileage of the subway line that the detection train is currently passing through is determined, and the corresponding subway catenary location is matched according to the subway line mileage; Collect visible light images, laser contour data, infrared thermal images, ultraviolet discharge images, pantograph-catenary contact current data, and vehicle posture data for the corresponding subway catenary locations, and bind the visible light images, laser contour data, infrared thermal images, ultraviolet discharge images, pantograph-catenary contact current data, and vehicle posture data to the corresponding subway catenary locations respectively. Using the location of the subway catenary as a unified index, visible light images, laser contour data, infrared thermal images, ultraviolet discharge images, pantograph-catenary contact current data, vehicle attitude data, train running position data, and catenary structure information are integrated to form detection data.

3. The method for dynamic detection of subway overhead contact lines based on multi-sensor collaboration according to claim 1, characterized in that, The step of extracting the multi-sensor change sequence corresponding to the location of the subway catenary based on the detection data specifically includes: Visible light edge changes are extracted from the visible light image, laser contour changes are extracted from the laser contour data, infrared temperature changes are extracted from the infrared thermal image, ultraviolet bright spot changes are extracted from the ultraviolet discharge image, pantograph-catenary contact current changes are extracted from the pantograph-catenary contact current data, and vehicle posture changes are extracted from the vehicle posture data. The start time, enhancement time, peak time, and recovery time of the visible light edge change, laser profile change, infrared temperature change, ultraviolet bright spot change, pantograph contact current change, and vehicle body attitude change are determined respectively. The initial change sequence of the corresponding subway catenary position is formed based on the start time, enhancement time, and peak time of the change. The initial change sequence is correlated and verified based on the train speed data, and the validity of each change result in the initial change sequence is marked. The sequence of changes in the corresponding subway catenary location is formed based on the initial change sequence after the validity marking is completed.

4. The method for dynamic detection of subway overhead contact lines based on multi-sensor collaboration according to claim 1, characterized in that, The process of comparing the sequence of changes from the multiple sensors with the normal sequence of changes formed during historical normal detection, and determining that there is a multi-sensor synchronization anomaly at the corresponding subway catenary location when the changes in visible light edge, laser contour, infrared temperature, and ultraviolet bright spots are simultaneously enhanced, while the changes in pantograph contact current and vehicle attitude are not synchronously enhanced, specifically includes: The normal change sequence corresponding to the position of the subway catenary during the historical normal detection process is retrieved, and the change sequence of the multi-sensor is compared with the normal change sequence item by item; Determine whether the visible light edge change, the laser contour change, the infrared temperature change, and the ultraviolet bright spot change occur earlier, occur more frequently, or last longer than the normal change sequence, and determine whether the pantograph contact current change and the vehicle body attitude change are within the change range corresponding to the normal change sequence; When the visible light edge change, the laser profile change, the infrared temperature change, and the ultraviolet bright spot change are simultaneously enhanced within the same detection segment, the corresponding subway contact network location is marked as a candidate location for multi-sensor synchronous enhancement. The changes in pantograph contact current and vehicle attitude corresponding to the candidate positions of the multi-sensor synchronous enhancement are read. If the changes in pantograph contact current do not show corresponding sudden changes, fluctuation expansions, or recovery delays, and the changes in vehicle attitude do not show corresponding pitch changes, roll changes, vertical vibration changes, or lateral vibration changes, it is determined that the candidate positions of the multi-sensor synchronous enhancement do not meet the accompanying conditions of the actual catenary structure disturbance. When the visible light edge change, the laser profile change, the infrared temperature change, and the ultraviolet bright spot change have the same enhancement direction, similar enhancement time, and similar recovery time, and the pantograph contact current change and the vehicle body attitude change are not synchronously enhanced, it is determined that there is a multi-sensor synchronous abnormality at the corresponding subway catenary location.

5. The method for dynamic detection of subway overhead contact lines based on multi-sensor collaboration according to claim 1, characterized in that, The determination of mobile and fixed abnormal states based on the persistence of the multi-sensor synchronization abnormal state between adjacent contact wire positions specifically includes: The locations of the subway contact wire and the adjacent contact wire locations are arranged according to the direction of train operation to form a contact wire location arrangement relationship, and the visible light edge changes, laser contour changes, infrared temperature changes and ultraviolet bright spot changes corresponding to the multi-sensor synchronization abnormal state are tracked to see whether they occur continuously between multiple adjacent contact wire locations; When the abnormal continuation position corresponding to the multi-sensor synchronization abnormal state appears sequentially after the abnormal starting position along the train running direction, and the visible light edge change, laser profile change, infrared temperature change, and ultraviolet bright spot change corresponding to the abnormal continuation position have similar change patterns to the visible light edge change, laser profile change, infrared temperature change, and ultraviolet bright spot change corresponding to the abnormal starting position, it is determined that the multi-sensor synchronization abnormal state has a moving characteristic. When the changes in pantograph contact current and vehicle attitude at the corresponding abnormal start position and abnormal continuation position are not synchronously enhanced, the multi-sensor synchronous abnormal state is determined as a moving abnormal state. When the multi-sensor synchronization anomaly does not transfer sequentially along the train's running direction, and occurs a preset number of times within a preset range at the same subway contact network location or the same contact network component, the multi-sensor synchronization anomaly is determined to have fixed characteristics. When the same or similar visible light edge changes, laser profile changes, infrared temperature changes, and ultraviolet bright spot changes occur at the location of the subway contact network during multiple detections, the multi-sensor synchronous abnormal state is determined as a fixed abnormal state.

6. The method for dynamic detection of subway overhead contact lines based on multi-sensor collaboration according to claim 1, characterized in that, The process of identifying and detecting device posture interference, detection window contamination, and tunnel airflow interference based on the mobile abnormal state, and identifying insulation aging, contact network structure loosening, and hard point abnormal states based on the fixed abnormal state, specifically includes: When the changes in the laser profile and the visible light edge change synchronously with the changes in the vehicle body posture, and the changes in the pantograph contact current do not increase synchronously, the abnormal movement state is identified as a posture interference of the detection equipment. When the visible light edge change and the laser profile change maintain similar change patterns between multiple adjacent contact wire positions, and the infrared temperature change and the ultraviolet bright spot change are not fixed within the range of the insulator, segmented insulator, or contact wire connection, the abnormal movement state is identified as detection window contamination. When the changes in the laser profile and the visible light edge appear earlier, later, or have a different duration as the train speed data changes, and the changes in the vehicle body attitude are insufficient to explain the changes in the laser profile and the visible light edge, the abnormal movement state is identified as tunnel airflow interference. When the fixed abnormal state is fixed within the range of the insulator, segmented insulator, or support device, and the changes in ultraviolet bright spots and infrared temperature occur repeatedly during multiple detection processes, the fixed abnormal state is identified as an insulation aging state. When the abnormal fixation state is fixed within the range of the positioner, dropper, anchor joint, or contact wire connection, and the change in the pantograph-catenary contact current occurs before the change in the laser profile, the abnormal fixation state is identified as a loose contact wire structure. When the fixed abnormal state is fixed within the range of the contact line, positioner, or dropper, and the changes in the pantograph contact current, the laser profile, and the visible light edge continuously enhance, the fixed abnormal state is identified as a hard spot abnormal state.

7. The method for dynamic detection of subway overhead contact lines based on multi-sensor collaboration according to claim 1, characterized in that, The process involves generating corresponding processing strategies based on the attitude interference of the detection equipment, the contamination of the detection window, the airflow interference in the tunnel, the aging state of the insulation, the loosening state of the contact network structure, and the abnormal state of the hard points. The detection data is then updated according to these processing strategies to generate corresponding dynamic detection results for the subway contact network based on the updated detection data. Specifically, this includes: When the anomaly source type is detection equipment posture interference, a detection equipment posture correction strategy is generated, and the corresponding subway catenary location is excluded from the catenary defect statistics object. When the anomaly source type is detection window contamination, a detection window contamination handling strategy is generated, and visible light edge changes and laser contour changes in the corresponding subway catenary location are excluded from the catenary defect statistics object. When the anomaly source type is tunnel airflow interference, a tunnel airflow interference processing strategy is generated, and contour offset, edge jitter, or short-term imaging blur related to train speed data are marked as airflow interference impact results; When the anomaly source type is insulation aging state, an insulation aging treatment strategy is generated, and the corresponding subway contact network location is marked as an insulation aging concern location; When the anomaly source type is a loose catenary structure, a catenary structure loosening handling strategy is generated, and the corresponding subway catenary location is marked as a location of concern regarding structural loosening; When the anomaly source type is a hard point anomaly state, a hard point anomaly handling strategy is generated, and the corresponding subway catenary location is marked as a hard point anomaly concern location. The detection data is updated according to each processing strategy, and anomalies belonging to the detection equipment posture interference, detection window contamination and tunnel airflow interference are marked as non-contact network defect anomalies, and anomalies belonging to the insulation aging state, contact network structure loose state and hard point anomaly state are marked as contact network defect anomalies. The updated detection data will be used to generate the corresponding dynamic detection results for the subway overhead contact system.

8. A dynamic detection device for subway overhead contact lines based on multi-sensor collaboration, characterized in that, The device is used to execute the multi-sensor collaborative dynamic detection method for subway overhead contact lines as described in any one of claims 1 to 7. The device includes an acquisition module and a processing module, wherein... The acquisition module is used to acquire visible light images, laser contour data, infrared thermal images, ultraviolet discharge images, pantograph-catenary contact current data, vehicle posture data, train running position data, and catenary structure information corresponding to the location of the subway catenary, so as to form detection data. The processing module is used to extract the sequence of changes of multiple sensors corresponding to the location of the subway catenary based on the detection data; The processing module is also used to compare the sequence of changes of the multiple sensors with the normal sequence of changes formed during the historical normal detection process, and when the changes of visible light edge, laser contour, infrared temperature and ultraviolet bright spot are enhanced at the same time, and the changes of pantograph contact current and vehicle body posture are not enhanced synchronously, it is determined that there is a multi-sensor synchronous abnormal state at the corresponding subway catenary location. The processing module is also used to determine the moving abnormal state and the fixed abnormal state based on the persistence of the multi-sensor synchronization abnormal state between adjacent contact wire positions. The processing module is also used to identify the attitude interference of the detection equipment, the contamination of the detection window and the airflow interference in the tunnel according to the abnormal movement state, and to identify the insulation aging state, the loosening state of the contact wire structure and the abnormal hard point state according to the abnormal fixed state. The processing module is also used to generate corresponding processing strategies based on the attitude interference of the detection equipment, the contamination of the detection window, the airflow interference in the tunnel, the aging state of the insulation, the loose state of the contact network structure, and the abnormal state of the hard points, and update the detection data according to the processing strategies, so as to generate corresponding dynamic detection results of the subway contact network based on the updated detection data.

9. An electronic device, characterized in that, The electronic device includes a processor and a memory; the memory stores a computer program, wherein the computer program, when executed by the processor, implements the dynamic detection method for subway catenary based on multi-sensor collaboration as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium stores instructions that, when executed, perform the dynamic detection method for subway catenary based on multi-sensor collaboration as described in any one of claims 1 to 7.

Images