A safety inspection system for the structure of supporting equipment for high-speed rail transit

Abstract

This invention discloses a safety inspection system for the structure of supporting equipment in high-speed rail transit, comprising an excitation response module, an infrared thermal source image acquisition module, a parameter monitoring module, an image preprocessing module, a thermal source feature extraction module, and a structural anomaly identification module. During train operation, the system utilizes aerodynamic disturbances of the train body to excite the response of trackside equipment structures, acquires their thermal response images through high-frame-rate infrared imaging, and performs image standardization processing based on disturbance intensity and relative distance. The system further extracts the thermal response features of the equipment structure area and compares them with normal templates to achieve abnormal structure identification and location. This invention can achieve non-contact, high-precision, and intelligent detection of the condition of trackside equipment structures without affecting train operation, and has advantages such as flexible deployment, accurate identification, and strong real-time performance, making it suitable for rail transit operation and maintenance scenarios.

Description

Technical Field

[0001] This invention relates to the field of safety inspection technology for high-speed rail transit, and in particular to a safety inspection system for the structure of supporting equipment for high-speed rail transit. Background Technology

[0002] With the large-scale construction of high-speed railways and urban rail transit systems, numerous trackside infrastructures (such as signal control boxes, communication equipment, and power cabinets) deployed along the lines have become crucial components in ensuring train operation safety. These devices typically operate outdoors for extended periods, making them susceptible to various factors such as the natural environment, equipment aging, and external impacts, leading to problems like structural loosening, shell cracking, and abnormal internal heating. If such structural defects are not detected in time, they can cause equipment failure. Due to the failure and damage to the fixed structure, the equipment may be pulled into the tracks during high-speed train operation, resulting in train malfunctions and posing a significant safety hazard.

[0003] Currently, the main technical methods used for the inspection and patrol of trackside equipment are as follows: 1. Manual visual inspection: Maintenance personnel periodically conduct offline inspections, identifying structural damage or abnormalities through visual inspection, tapping, or infrared handheld devices. This method relies on human experience, and the results are easily affected by subjective judgment, leading to problems such as missed detections, misjudgments, and low efficiency, making it difficult to meet the maintenance needs of high-density, high-speed rail lines. 2. Fixed monitoring systems: Video surveillance equipment is installed at specific locations to monitor key trackside equipment 24 hours a day. This method is suitable for monitoring anomalies in small areas and key locations, but due to limited coverage, it cannot comprehensively inspect all trackside equipment along the entire line. Furthermore, fixed cameras have a single perspective, making them unsuitable for the diverse types and deployment methods of equipment.

[0004] Therefore, current technology still lacks a technical means to monitor the structural safety issues of trackside equipment in a timely manner to ensure the safety of high-speed rail operations. Summary of the Invention

[0005] The purpose of this invention is to provide a safety detection system for the structure of supporting equipment for high-speed rail transit, which has the advantages of not requiring additional excitation devices, being able to dynamically acquire high-quality infrared thermal images, and having the ability to intelligently identify and locate structural anomalies.

[0006] The above-mentioned technical objective of the present invention is achieved through the following technical solution:

[0007] A safety inspection system for the structure of supporting equipment for high-speed rail transit includes:

[0008] The excitation response module is used to stimulate the structural response of trackside equipment in the area traversed by the train during high-speed train operation by using the aerodynamic disturbances naturally generated by the train body.

[0009] An infrared thermal source image acquisition module is installed in the side or lower area of ​​the train body to acquire the infrared thermal source response image of the trackside equipment under aerodynamic disturbance when the train passes the trackside equipment.

[0010] The parameter monitoring module includes an excitation monitoring unit and a distance measurement unit. The excitation monitoring unit is used to collect aerodynamic disturbance intensity information during train operation; the distance measurement unit is used to measure the relative spatial distance information between the train and trackside equipment.

[0011] An image preprocessing module is used to standardize the infrared heat source response image based on the aerodynamic disturbance intensity information and relative spatial distance information to obtain a heat source response spectrum.

[0012] The heat source feature extraction module is used to extract the thermal response features of each structural region of the trackside equipment from the heat source response spectrum.

[0013] The structural anomaly identification module is used to compare and analyze the thermal response characteristics with the preset normal state thermal source response template to identify whether there is a structural anomaly in the trackside equipment, and output the anomaly type and location coordinate information.

[0014] Further configuration: The infrared thermal source image acquisition module specifically includes:

[0015] The high frame rate image acquisition unit is used to continuously acquire multiple frames of infrared thermal images at a frame rate no lower than the preset frame rate as the train passes the target trackside equipment.

[0016] The image window adjustment unit is used to dynamically adjust the spatial range of the high-frequency image acquisition unit for image acquisition based on the relative spatial distance information provided by the device distance measurement module, so as to ensure the coverage integrity of the target device in the infrared thermal image.

[0017] The viewing angle offset correction unit is used to calculate the projection angle of the trackside equipment in the infrared thermal image based on the relative positional relationship between the trackside equipment and the high-frequency image acquisition unit, and to obtain the infrared heat source response image after performing geometric correction processing on the infrared thermal image.

[0018] Further configuration: The excitation monitoring unit includes:

[0019] The acceleration estimation subunit is used to acquire data on the acceleration changes of the train body in the lateral or vertical direction during train operation, and to estimate the intensity of aerodynamic disturbance around the train body based on the acceleration data.

[0020] The environmental parameter calibration subunit is used to acquire external meteorological parameters in the environment where the train is located, and to correlate the external environmental parameters with the acceleration estimation results to correct and compensate the estimated disturbance intensity to obtain gas disturbance intensity information.

[0021] Further configuration: The distance measurement unit includes:

[0022] The train positioning subunit is used to obtain the real-time position information of the train in the track coordinate system based on the inertial navigation system, odometer sensor or on-board track positioning system carried by the train.

[0023] The equipment location information retrieval subunit is used to extract the standard deployment location of the target trackside equipment in front of the train from the preset track infrastructure map database;

[0024] The relative distance calculation unit is used to calculate the relative spatial distance between the train and the trackside equipment based on the real-time location information of the train and the standard deployment location to obtain relative spatial distance information.

[0025] Further details: The image preprocessing module specifically includes:

[0026] The spatial distance correction unit adjusts the scale of each frame of infrared heat source image according to the relative spatial distance information corresponding to each frame of infrared heat source response image to obtain a spatial correction image, which is used to make the scale of the area corresponding to the trackside equipment in each frame of infrared heat source image consistent.

[0027] The normalization unit is used to correct the pixel value representing the heat source response intensity in the spatially corrected image based on the gas disturbance intensity information to obtain the heat source response spectrum.

[0028] Further steps: The step of correcting the pixel values ​​representing the heat source response intensity in the spatially corrected image based on the gas disturbance intensity information to obtain the heat source response spectrum specifically includes the following steps:

[0029] Extract thermal response image data of the corresponding area of ​​the trackside equipment from the spatially corrected image, and identify the thermal response intensity corresponding to each pixel value;

[0030] The thermal response correction parameters are calculated based on the gas disturbance intensity information and the preset disturbance-thermal response mapping model.

[0031] The standardized thermal response intensity is obtained by correcting the thermal response intensity corresponding to each pixel value based on the thermal response correction parameters.

[0032] The heat source response map is generated by updating the pixel values ​​corresponding to the standardized thermal response intensity back into the spatially corrected image.

[0033] Further configuration: The heat source feature extraction module includes:

[0034] The area division unit is used to divide the area corresponding to the trackside equipment into multiple structural areas and encode each structural area;

[0035] The temporal extraction unit is used to extract the temporal features of each structural region based on the time series of multiple frames of thermal response maps. The temporal features characterize the changes in the thermal response intensity of each structural region over time.

[0036] The distribution pattern extraction unit is used to extract the heat source pattern features of each structural region, including concentration, heat source symmetry and heat flow variability features.

[0037] An intensity feature extraction unit is used to extract the intensity features of each structural region, including the thermal response peak value and the thermal response fluctuation amplitude.

[0038] The feature output unit outputs time features, intensity features, and heat source mode features as thermal response features to the structural anomaly identification module.

[0039] Further configuration: The structural anomaly identification module includes:

[0040] The comparison and analysis unit is used to calculate the similarity between the thermal response features and the standard response features in the normal state heat source response template by using the feature vector comparison method. If any similarity is lower than the preset threshold, it is judged that there is a structural anomaly.

[0041] The type identification unit determines the anomaly type based on the types of thermal response features with similarity below a preset threshold;

[0042] The abnormal location unit obtains the corresponding code of the structural area that the comparison and analysis unit determines to have structural abnormalities as the location of the abnormal area, and generates positioning coordinate information by combining the standard layout location of the trackside equipment.

[0043] In summary, the present invention has the following beneficial effects:

[0044] The excitation process is completed based on the natural disturbance of train operation, which eliminates the need for additional active excitation devices or triggering mechanisms. This simplifies the system structure, reduces equipment costs, and the excitation method can be adapted to different lines and environmental conditions by moving with the train, thus providing good deployment flexibility and engineering feasibility.

[0045] An infrared thermal image acquisition module is used for dynamic imaging. By setting up a high frame rate acquisition unit, a window adjustment unit, and a viewing angle correction unit, thermal images can completely and stably cover the trackside equipment area, ensuring image clarity and structural integrity even under dynamic conditions such as high-speed train operation and constantly changing viewing angles. A dual monitoring mechanism of disturbance intensity and relative spatial distance is introduced. Acceleration estimation and environmental calibration improve the accuracy of aerodynamic disturbance quantification. Distance correction is performed by combining train positioning and trackside equipment map location data, achieving precise standardization of equipment thermal response and providing a solid foundation for cross-time and cross-equipment image comparison.

[0046] The image preprocessing module performs spatial and calorific value normalization, unifying infrared images under different environments, distances, and speeds into physically comparable thermal response spectra. This significantly improves the accuracy and stability of subsequent feature extraction and comparative analysis. The thermal source feature extraction module obtains multi-dimensional response features at the structural level, including time evolution features, heat flow distribution features, and thermal response intensity features, comprehensively reflecting the thermal response patterns of various structural areas of the trackside equipment under disturbance excitation. This helps identify potential local anomalies or structural degradation problems.

[0047] The structural anomaly identification module possesses template comparison and anomaly classification and localization capabilities, enabling quantitative judgment of equipment structural status. Once a deviation of thermal response characteristics from the normal template is detected, the anomaly type is automatically determined and precise coordinates are output, providing structured and visualized inspection data for the backend maintenance system. The entire system can complete the closed-loop data acquisition, processing, and identification of equipment during normal train operation, without requiring shutdown or manual intervention. It boasts high-frequency, wide-coverage, and low-cost inspection capabilities, contributing to the construction of a full lifecycle equipment health management platform for future intelligent railways. The system architecture of this invention is modular and highly scalable, capable of being integrated into different types of trains such as high-speed rail and urban rail, and applicable to both new and existing line renovation projects, demonstrating significant engineering promotion value and industry application prospects. Attached Figure Description

[0048] Figure 1 This is an overall structural block diagram of the embodiment. Detailed Implementation

[0049] The present invention will be further described in detail below with reference to the accompanying drawings.

[0050] Example:

[0051] like Figure 1As shown, the system proposed in this invention is suitable for real-time detection tasks during high-speed railway train operation. It can perform long-distance, non-contact structural defect identification of trackside infrastructure such as communication equipment, signal boxes, and power control cabinets deployed along the track without affecting the normal operation of the train. The system is integrated and installed on the train body. Relying on the aerodynamic disturbances generated by the train's own movement, it guides the equipment to generate thermal response signals. Combined with infrared thermal imaging, spatial positioning, and thermal image analysis, a complete process for structural anomaly identification is constructed.

[0052] By standardizing and extracting features from infrared thermal source response images, this invention can accurately identify different types of structural defects, including but not limited to: poor enclosure sealing, loose structure, and internal overheating. It can also achieve anomaly type judgment and spatial positioning output, offering advantages such as high detection accuracy, fast response speed, convenient deployment, and suitability for high-speed operating environments. It is applicable to scenarios such as intelligent track inspection and railway infrastructure safety monitoring. The specific solution is as follows:

[0053] A safety inspection system for the structure of supporting equipment for high-speed rail transit includes:

[0054] The excitation response module is used to stimulate the structural response of trackside equipment in the area traversed by the train during high-speed train operation by using the aerodynamic disturbances naturally generated by the train body.

[0055] An infrared thermal source image acquisition module is installed in the side or lower area of ​​the train body to acquire the infrared thermal source response image of the trackside equipment under aerodynamic disturbance when the train passes the trackside equipment.

[0056] The parameter monitoring module includes an excitation monitoring unit and a distance measurement unit. The excitation monitoring unit is used to collect aerodynamic disturbance intensity information during train operation; the distance measurement unit is used to measure the relative spatial distance information between the train and trackside equipment.

[0057] An image preprocessing module is used to standardize the infrared heat source response image based on the aerodynamic disturbance intensity information and relative spatial distance information to obtain a heat source response spectrum.

[0058] The heat source feature extraction module is used to extract the thermal response features of each structural region of the trackside equipment from the heat source response spectrum.

[0059] The structural anomaly identification module is used to compare and analyze the thermal response characteristics with the preset normal state thermal source response template to identify whether there is a structural anomaly in the trackside equipment, and output the anomaly type and location coordinate information.

[0060] In an embodiment of the present invention, the excitation response module does not contain a specific physical structure. Instead, it utilizes the aerodynamic disturbances naturally generated by the train during operation as an excitation source to perform non-contact structural excitation on the trackside equipment in the area it passes through, thereby guiding it to generate a response.

[0061] When a train operates at high speeds (e.g., 250 km / h and above), a high-intensity aerodynamic disturbance field is generated around the train due to air compression, flow around the train, and boundary layer disturbances at the leading edge of the train head and the sides of the train body. These disturbances include turbulent wakes, pressure pulsations, and lateral / vertical vortex structures. Especially when the train passes near equipment areas, these disturbances are transmitted to the equipment surface in the form of shock waves or vibration waves, thereby triggering the equipment structure's response behavior. This response mainly manifests in the following ways: microstructural vibrations, such as slight deformation of the equipment shell or support under instantaneous pressure; low-frequency mechanical resonance of internal connecting components (bolts, connecting plates, etc.); and temporary thermal anomalies in surface temperature due to minute deformations or stress concentration areas. Since these structural response processes are usually accompanied by subtle changes in heat conduction or radiation, combined with high-sensitivity infrared thermal imaging, the changes in the equipment's structural state can be perceived without contacting the equipment. In this invention, the excitation response module is a purely passive structure that does not rely on active excitation sources outside the vehicle body (such as mechanical impact, electromagnetic pulse, etc.). It uses natural aerodynamic disturbances to replace traditional physical excitation devices, which not only reduces system complexity and maintenance costs, but also improves safety and detection efficiency during train operation.

[0062] In an embodiment of the present invention, the infrared thermal source image acquisition module, as the core data acquisition component of the entire system, is installed on the side or bottom of the high-speed train body. It is used to acquire the infrared thermal source response image sequence generated by the trackside equipment under aerodynamic disturbance when the train passes the equipment. This module has a compact structure, strong vibration resistance, and high response speed, and mainly includes the following three functional units:

[0063] 1. High frame rate image acquisition unit

[0064] The high frame rate image acquisition unit employs an uncooled infrared focal plane detector (such as a microbolometer), featuring high temporal resolution. It supports the continuous acquisition of multiple frames of infrared thermal images during high-speed train operation at a frame rate no lower than a preset limit (e.g., 200 frames / second). The unit's optical field of view is spatially matched to the equipment deployment area, ensuring that the thermal imaging equipment can complete high-density image acquisition in a short time when the train passes key locations of the trackside equipment. This high frame rate acquisition capability ensures: acquiring information on changes in thermal response before, during, and after disturbances; effectively suppressing image blurring, ghosting, or missing frames; and supporting subsequent dynamic analysis and trend modeling of thermal features. Thermal image acquisition data is transmitted to the back-end image processing module via a high-speed interface (such as Camera Link or GigE Vision), maintaining the overall real-time performance of the system.

[0065] 2. Image window adjustment unit

[0066] This unit is used to dynamically adjust the spatial window for image acquisition based on the real-time relative spatial distance between the trackside equipment and the infrared acquisition equipment during train operation.

[0067] The image window adjustment unit calculates the distance between the device and the train in real time by calling the position information provided by the device distance measurement module, and performs the following operations accordingly: automatically adjusting the focal length or field of view of the infrared lens to keep the device centered or fully visible in the image; controlling the start and end points of the acquisition frame window to avoid premature / delayed acquisition; and dynamically adjusting the region of interest (ROI) range to improve image resolution utilization. For example, when the train approaches the device at a distance less than a set threshold, the system automatically shrinks the image window to avoid near-field imaging distortion; when the distance is greater, the window is enlarged to ensure that the device is not cropped. This function allows the system to adapt to device deployment density, train speed fluctuations, and differences in device structure, ensuring the availability and integrity of the image.

[0068] 3. Viewpoint shift correction unit

[0069] Due to the spatial offset and projection angle between the infrared acquisition unit installed on the train and the trackside equipment, directly acquired infrared images often suffer from angular distortion, scale distortion, and contour tilt, which is detrimental to subsequent heat source feature comparison and template comparison. Therefore, the perspective offset correction unit uses a perspective transformation algorithm to perform geometric correction processing on the infrared images based on the relative position coordinates between the train and the equipment, the installation height, and the tilt angle. Its basic process includes:

[0070] The projection angle of the computing device onto the image;

[0071] Image rotation, scaling, and stretching are performed using affine transformations or four-point mapping methods.

[0072] Output the corrected infrared thermal image, also known as the "heat source response image".

[0073] This correction process ensures that the device presents a standardized geometric structure in the image under different shooting angles and distances, guaranteeing the consistency of the position and distribution of heat source features in the spectrum and enhancing the accuracy of subsequent thermal image comparison. In summary, the infrared heat source image acquisition module, with its integrated architecture of "high-speed imaging + distance adjustment + projection correction," ensures that the system still possesses the ability to acquire high-quality thermal response images even under high-speed operation, providing stable and reliable raw data support for structural defect identification.

[0074] In this embodiment of the invention, the parameter monitoring module, as a key subsystem supporting the standardized processing of infrared images, is mainly used to acquire two types of key information during train operation: the intensity of aerodynamic disturbance and the spatial relative distance between the train and trackside equipment. This module includes two subsystems: an excitation monitoring unit and a distance measurement unit.

[0075] The excitation monitoring unit is used to evaluate the intensity of aerodynamic disturbances generated during train operation. This index will be used for normalization and correction of thermal response intensity in subsequent images. This unit includes the following two sub-units:

[0076] The acceleration estimation subunit is deployed at multiple locations along the lateral and vertical directions of the train body (such as the front, middle, and rear sections) to collect acceleration variation information of the train body structure under high-speed operation. Specifically, a three-axis accelerometer array records the acceleration fluctuation data of the train body in the X (lateral) and Z (vertical) directions in real time. Sampling is performed at a frequency of 500 times per second to form an acceleration time series.

[0077] a x (t), a z (t)

[0078] Subsequently, the system estimates the disturbance intensity based on the following empirical model:

[0079]

[0080] Here, A represents the estimated intensity of the current external disturbance to the train body. This disturbance intensity is highly correlated with the degree of airflow turbulence generated outside the train. This data will be used in subsequent image processing to estimate the excitation background conditions for the thermal response of trackside equipment.

[0081] An environmental parameter calibration subunit is introduced to improve estimation accuracy, as aerodynamic disturbances around the vehicle body may be affected by environmental factors (such as natural wind speed and direction). This subunit acquires environmental data in two ways: real-time wind field data is collected by a roof-mounted meteorological sensor array (anemometer, wind vane); or the track section meteorological data interface provided by the Train Management System (TMS) is connected. The acquired parameters include: real-time wind speed V. wind Wind direction angle θ wind The system corrects this data against the estimated disturbance intensity A using the following calibration model (which can be linear / empirical weighted):

[0082] A′=A-k1·V wind ·cos(θ wind -θ train )

[0083] Where: A′ is the corrected disturbance intensity index; θ train is the direction of train travel; k1 is the empirical correction coefficient.

[0084] The distance measurement unit is used to obtain spatial relative distance information between the train and the trackside equipment that is about to pass. It consists of the following three sub-units:

[0085] The train positioning subunit integrates multiple position sensing components, including: a high-precision GNSS receiver for acquiring the train's latitude and longitude position information in the track coordinate system; an inertial navigation system (INS) for providing continuous position compensation over short periods, improving stability in tunnels and obstructed areas; and wheel pulse meter or odometer sensors for providing cumulative distance correction information. The system obtains the train's real-time position P using a multi-source fusion positioning algorithm (such as extended Kalman filtering). train (t), its accuracy can reach the sub-meter level.

[0086] The device location information retrieval subunit of this invention pre-defines a track infrastructure map database, which records the standard deployment locations of all target devices, including: device code; track segment number; center coordinates (projected coordinates or latitude and longitude) P. dev The system retrieves the location of the target equipment that the train is about to pass from the database based on the train's current location and direction of travel.

[0087] According to the train's current position P train With the target device location P dev The system calculates the spatial distance D between the two in real time, that is:

[0088] D = ||P train -P dev ||

[0089] This result will serve as input parameters for image window adjustment and scale normalization, ensuring that the target device can be accurately captured and standardized during image acquisition. In summary, the parameter monitoring module, through dual-channel information acquisition and real-time calculation, provides two key fundamental variables for system operation (perturbation intensity + relative distance), forming the input starting point of the image processing chain and ensuring that the entire detection system possesses environmental adaptability and positioning accuracy.

[0090] In an embodiment of the present invention, an image preprocessing module is used to standardize the scale and thermal intensity of the original infrared heat source image to eliminate image inconsistencies caused by changes in spatial distance and disturbance intensity during train operation, providing a unified heat source response spectrum for subsequent heat source feature extraction and anomaly identification. This module mainly includes a spatial distance correction unit and a normalization unit.

[0091] The spatial distance correction unit addresses the issue that the relative distance between the train and trackside equipment continuously changes during operation, causing variations in the display ratio of the same equipment in the thermal image across different frames, thus affecting subsequent image comparison and thermal feature consistency. Therefore, this unit performs spatial scale correction on the image based on the relative spatial distance information of the equipment in each frame (provided by the distance measurement module). The correction method is as follows:

[0092] For each frame of the infrared thermal response image, let the original image be I(x, y) and the corresponding distance be D. The system calculates the image scaling factor s based on the preset reference distance D0:

[0093]

[0094] Then perform a proportional scaling transformation on the image:

[0095] I s (x′,y′)=I(s·x,s·y)

[0096] Among them, I s (x′, y′) represents the spatially scaled image, indicating the actual size of the device area displayed in the image at a uniform reference distance. After processing by this unit, the projected area size of the same type of device is consistent in all images, facilitating subsequent spatial alignment and atlas synthesis.

[0097] The normalization unit addresses the issue that thermal response intensity is directly affected by disturbance excitation intensity, leading to potential shifts in the thermal response values ​​of the same equipment under different train disturbance conditions. Therefore, this unit introduces disturbance intensity information A′ to normalize the thermal value of each pixel in the image, resulting in a stable and comparable thermal source response spectrum.

[0098] The normalization process is as follows:

[0099] Step 1: Extraction of thermal response region

[0100] Image I corrected from spatial scale s In (x, y), extract the image data corresponding to the preset device region, denoted as:

[0101]

[0102] Where R(x, y) represents the set of pixels within the device area, and each pixel value T(x, y) represents the thermal response intensity.

[0103] Step 2: Invoking the perturbation-thermal response mapping model

[0104] The system is based on a perturbation-thermal response mapping model constructed through experimental fitting or simulation:

[0105] T std =f(T, A′)

[0106] Where: T: original thermal response value; A′: disturbance intensity; T std : The corrected standard thermal response value; f(.): The mapping function, which can be a linear function, a piecewise function, or an empirical formula. The mapping model considers the nonlinear effect of perturbation intensity on the thermal response, such as response delay under low perturbation and response enhancement under high perturbation.

[0107] Step 3: Pixel value correction

[0108] For each pixel value T(x, y) ∈ R, perform mapping correction to obtain the normalized pixel value:

[0109] T std (x, y) = f(T(x, y), A′)

[0110] Step 4: Generating the heat source response spectrum

[0111] The above-corrected standard calorific value T std Replace (x, y) back with the device region to form a new heat source response map:

[0112]

[0113] Final output image I std This is a heat source response spectrum, which has standardized characteristics of consistent scale and intensity, and meets the requirements for comparative analysis across devices, frames, and environments.

[0114] This image preprocessing module effectively solves the problem of comparability of thermal images caused by changes in equipment images and inconsistent excitation during train operation by standardizing both spatial scale and thermal intensity, thus providing a unified input for the subsequent structural anomaly identification module.

[0115] In one embodiment of the present invention, a heat source feature extraction module is used to perform structured analysis on the standardized infrared heat source response spectrum, extracting the thermal response behavior features of trackside equipment in different structural regions, providing basic data support for subsequent structural anomaly identification. This module mainly includes the following five functional units: a region division unit, a time-series extraction unit, a distribution pattern extraction unit, an intensity feature extraction unit, and a feature output unit, with each unit processing data sequentially according to the data flow order.

[0116] The region division unit is used to identify the main structural components corresponding to the trackside equipment in the heat source response spectrum and divide them into several functional structural regions.

[0117] The partitioning method can be based on one of the following two approaches: Template matching: using a predefined structural template (such as a distribution box partition diagram) to slice regions from a standardized equipment map; Image segmentation: using image segmentation algorithms based on gray-level gradients, edge detection, or heat flow features to automatically identify structural boundaries in the map. The partitioning result is a set of structural regions {R1, R2, ..., R...} n Each region is assigned a unique coded ID. i This is used for subsequent feature binding and location association.

[0118] The time-series extraction unit utilizes the multi-frame standardized thermal response maps acquired previously to extract the time-varying thermal response behavior of each structural region. Specifically, for each structural region RiR_iRi, the average thermal response intensity at the corresponding location in the consecutive frame images is extracted to construct a time series.

[0119]

[0120] Wherein: T i (t): The average thermal response value of the i-th region in frame t; Normalized spectral pixel values ​​in frame t; |R i |: Number of pixels in the region.

[0121] Based on this sequence, the system can extract time-related features such as: rise / fall rate; peak occurrence time; and continuous heating or cooling cycles. These time features can reveal potential defect behaviors in the structural region, such as delayed response or abnormal heat dissipation.

[0122] The distribution pattern extraction unit analyzes the spatial distribution characteristics of the heat source within each structural region and extracts its thermal field pattern features, including the following three dimensions:

[0123] 1. Heat source concentration

[0124] Defined as the percentage of pixels within a region whose heat value is higher than the average, reflecting the degree of heat source focusing:

[0125]

[0126] Among them, C i : The heat source concentration in the i-th structural region; Region R i Internal average heat value; numerator: number of pixels with temperatures above the average temperature; denominator: total number of pixels in the region.

[0127] 2. Symmetry of the heat source

[0128] The symmetry of the thermal field distribution is measured by comparing the differences in pixel values ​​on either side of the left-right (or top-bottom) axis of symmetry within the region.

[0129]

[0130] Where (x′, y′) are pixels symmetric to (x, y) about the center of the region, S i : Symmetry score of the i-th structural region, the closer the value is to 1, the more symmetrical; (x′, y′): The symmetrical point of pixel (x, y) about the central axis of the region; the absolute value represents the difference in thermal intensity between the left and right pixels.

[0131] 3. Heat flow variability

[0132] The degree of thermal non-uniformity is calculated using the gradient change index within the region:

[0133]

[0134] F i : The intensity of the gradient change in the distribution of the heat source in the i-th region; Lateral and longitudinal temperature gradients (which can be estimated using the Sobel operator).

[0135] The above three parameters together constitute the feature vector of the regional heat source distribution pattern.

[0136] The intensity feature extraction unit is used to extract the thermal response intensity features of the structural region from the time series, including the thermal response peak value and the thermal response fluctuation amplitude. This information can determine whether there are phenomena such as short-term intense heating or intermittent thermal failure in the region.

[0137] Finally, the feature output unit integrates the features from the above four dimensions into a unified thermal response feature set:

[0138] F i =[T i (t), C i S i F i T peak，i ΔT i ]

[0139] F i : The complete thermal response feature vector of the i-th region, used by the anomaly identification module.

[0140] Each structural region forms an independent feature vector, which is bound through structural encoding and transmitted to the structural anomaly identification module for subsequent template comparison or classification. This module constructs a clear thermal response profiling mechanism through region decomposition, multi-dimensional modeling, and unified feature structure, providing a high-resolution data foundation for intelligent identification of the structural status of trackside equipment.

[0141] In one embodiment of the present invention, the structural anomaly identification module is used to compare and analyze the thermal response feature vector output by the heat source feature extraction module to identify whether there is a structural anomaly in the trackside equipment, and further output the anomaly type and location. This module includes three core functional units: a comparison and analysis unit, a type identification unit, and an anomaly location unit.

[0142] The comparison and analysis unit is used to determine whether the thermal response of a device's structural region deviates from the normal state. It employs a feature vector similarity matching method to compare the real-time extracted thermal response feature vector with a preset normal template. The normal state heat source template vector corresponding to the thermal response feature vector extracted from the i-th structural region in the current frame is:

[0143]

[0144] The system uses the following similarity function for vector comparison:

[0145]

[0146] Where ||.||2 represents the Euclidean norm (L2 distance). The closer the similarity value is to 1, the more normal the state; the closer it is to 0, the greater the deviation.

[0147] The system sets a preset threshold θ∈(0,1), if: Sim(F i F i ref If θ < θ, then the i-th structural region is determined to have a structural anomaly.

[0148] The type recognition unit classifies and judges anomalies based on features with similarity below a threshold, identifying different types of potential defects. The recognition process is as follows:

[0149] If the main deviation occurs in ΔT i T peak，i This could be due to "intermittent overheating";

[0150] If S iSignificant decrease (symmetry disruption): This may indicate "structural loosening or deformation";

[0151] If F i A significant increase (drastic change in heat flow) may indicate "uneven internal distribution or insulation damage".

[0152] The system classifies data according to the following set of logical rules R:

[0153]

[0154] Anomalies are assigned to pre-trained expert rules or machine learning classification models (such as SVM, decision trees):

[0155] Category A: Abnormal heat source concentration; Category B: Structural geometric anomalies (such as warping); Category C: Poor heat dissipation or disordered internal heat sources.

[0156] The anomaly location unit, once it confirms an anomaly in a structural area, will locate the anomaly based on the area's unique coded ID. i It locates abnormal positions in the equipment and, combined with the standard equipment layout coordinates provided by the distance measurement module, outputs positioning results that can be used in the ground inspection system.

[0157] Abnormal coordinate expression:

[0158] Loc i =Pos device +δ i

[0159] P osdevice δ: The standard center position of the target equipment in the orbital coordinate system; i : Structural region R i Offset vector relative to the center of the device (can be obtained from the map template).

[0160] The final system output will be: anomaly type (e.g., structural loosening), anomaly area code (e.g., R3), and location coordinate information.

[0161] The present invention has the following beneficial effects:

[0162] The excitation process is completed based on the natural disturbance of train operation, which eliminates the need for additional active excitation devices or triggering mechanisms. This simplifies the system structure, reduces equipment costs, and the excitation method can be adapted to different lines and environmental conditions by moving with the train, thus providing good deployment flexibility and engineering feasibility.

[0163] An infrared thermal image acquisition module is used for dynamic imaging. By setting up a high frame rate acquisition unit, a window adjustment unit, and a viewing angle correction unit, thermal images can completely and stably cover the trackside equipment area, ensuring image clarity and structural integrity even under dynamic conditions such as high-speed train operation and constantly changing viewing angles. A dual monitoring mechanism of disturbance intensity and relative spatial distance is introduced. Acceleration estimation and environmental calibration improve the accuracy of aerodynamic disturbance quantification. Distance correction is performed by combining train positioning and trackside equipment map location data, achieving precise standardization of equipment thermal response and providing a solid foundation for cross-time and cross-equipment image comparison.

[0164] The image preprocessing module performs spatial and calorific value normalization, unifying infrared images under different environments, distances, and speeds into physically comparable thermal response spectra. This significantly improves the accuracy and stability of subsequent feature extraction and comparative analysis. The thermal source feature extraction module obtains multi-dimensional response features at the structural level, including time evolution features, heat flow distribution features, and thermal response intensity features, comprehensively reflecting the thermal response patterns of various structural areas of the trackside equipment under disturbance excitation. This helps identify potential local anomalies or structural degradation problems.

[0165] The structural anomaly identification module possesses template comparison and anomaly classification and localization capabilities, enabling quantitative judgment of equipment structural status. Once a deviation of thermal response characteristics from the normal template is detected, the anomaly type is automatically determined and precise coordinates are output, providing structured and visualized inspection data for the backend maintenance system. The entire system can complete the closed-loop data acquisition, processing, and identification of equipment during normal train operation, without requiring shutdown or manual intervention. It boasts high-frequency, wide-coverage, and low-cost inspection capabilities, contributing to the construction of a full lifecycle equipment health management platform for future intelligent railways. The system architecture of this invention is modular and highly scalable, capable of being integrated into different types of trains such as high-speed rail and urban rail, and applicable to both new and existing line renovation projects, demonstrating significant engineering promotion value and industry application prospects.

[0166] The embodiments described above do not constitute a limitation on the scope of protection of this technical solution. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the above embodiments should be included within the scope of protection of this technical solution.

Claims

1. A safety inspection system for the structure of supporting equipment in high-speed rail transit, characterized in that, include: The excitation response module is used to stimulate the structural response of trackside equipment in the area traversed by the train during high-speed train operation by using the aerodynamic disturbances naturally generated by the train body. An infrared thermal source image acquisition module is installed in the side or lower area of ​​the train body to acquire the infrared thermal source response image of the trackside equipment under aerodynamic disturbance when the train passes the trackside equipment. The parameter monitoring module includes an excitation monitoring unit and a distance measurement unit. The excitation monitoring unit is used to collect aerodynamic disturbance intensity information during train operation; the distance measurement unit is used to measure the relative spatial distance information between the train and trackside equipment. An image preprocessing module is used to standardize the infrared heat source response image based on the aerodynamic disturbance intensity information and relative spatial distance information to obtain a heat source response spectrum. The heat source feature extraction module is used to extract the thermal response features of each structural region of the trackside equipment from the heat source response spectrum. The structural anomaly identification module is used to compare and analyze the thermal response characteristics with the preset normal state thermal source response template to identify whether there is a structural anomaly in the trackside equipment, and output the anomaly type and location coordinate information.

2. The safety detection system for the structure of supporting equipment for high-speed rail transit according to claim 1, characterized in that, The infrared thermal source image acquisition module specifically includes: The high frame rate image acquisition unit is used to continuously acquire multiple frames of infrared thermal images at a frame rate no lower than the preset frame rate as the train passes the target trackside equipment. The image window adjustment unit is used to dynamically adjust the spatial range of the high-frequency image acquisition unit for image acquisition based on the relative spatial distance information provided by the device distance measurement module, so as to ensure the coverage integrity of the target device in the infrared thermal image. The viewing angle offset correction unit is used to calculate the projection angle of the trackside equipment in the infrared thermal image based on the relative positional relationship between the trackside equipment and the high-frequency image acquisition unit, and to obtain the infrared heat source response image after performing geometric correction processing on the infrared thermal image.

3. A safety inspection system for the structure of supporting equipment for high-speed rail transit according to claim 2, characterized in that, The stimulus monitoring unit includes: The acceleration estimation subunit is used to acquire data on the acceleration changes of the train body in the lateral or vertical direction during train operation, and to estimate the intensity of aerodynamic disturbance around the train body based on the acceleration data. The environmental parameter calibration subunit is used to acquire external meteorological parameters in the environment where the train is located, and to correlate the external environmental parameters with the acceleration estimation results to correct and compensate the estimated disturbance intensity to obtain gas disturbance intensity information.

4. A safety inspection system for the structure of supporting equipment for high-speed rail transit according to claim 3, characterized in that, The distance measurement unit includes: The train positioning subunit is used to obtain the real-time position information of the train in the track coordinate system based on the inertial navigation system, odometer sensor or on-board track positioning system carried by the train. The equipment location information retrieval subunit is used to extract the standard deployment location of the target trackside equipment in front of the train from the preset track infrastructure map database; The relative distance calculation unit is used to calculate the relative spatial distance between the train and the trackside equipment based on the real-time location information of the train and the standard deployment location to obtain relative spatial distance information.

5. A safety inspection system for the structure of supporting equipment for high-speed rail transit according to claim 4, characterized in that, The image preprocessing module specifically includes: The spatial distance correction unit adjusts the scale of each frame of infrared heat source image according to the relative spatial distance information corresponding to each frame of infrared heat source response image to obtain a spatial correction image, which is used to make the scale of the area corresponding to the trackside equipment in each frame of infrared heat source image consistent. The normalization unit is used to correct the pixel value representing the heat source response intensity in the spatially corrected image based on the gas disturbance intensity information to obtain the heat source response spectrum.

6. A safety inspection system for the structure of supporting equipment for high-speed rail transit according to claim 5, characterized in that, The process of correcting the pixel values ​​representing the heat source response intensity in the spatially corrected image based on the gas disturbance intensity information to obtain the heat source response spectrum specifically includes the following steps: Extract thermal response image data of the corresponding area of ​​the trackside equipment from the spatially corrected image, and identify the thermal response intensity corresponding to each pixel value; The thermal response correction parameters are calculated based on the gas disturbance intensity information and the preset disturbance-thermal response mapping model. The standardized thermal response intensity is obtained by correcting the thermal response intensity corresponding to each pixel value based on the thermal response correction parameters. The heat source response map is generated by updating the pixel values ​​corresponding to the standardized thermal response intensity back into the spatially corrected image.

7. A safety inspection system for the structure of supporting equipment for high-speed rail transit according to claim 6, characterized in that, The heat source feature extraction module includes: The area division unit is used to divide the area corresponding to the trackside equipment into multiple structural areas and encode each structural area; The temporal extraction unit is used to extract the temporal features of each structural region based on the time series of multiple frames of thermal response maps. The temporal features characterize the changes in the thermal response intensity of each structural region over time. The distribution pattern extraction unit is used to extract the heat source pattern features of each structural region, including concentration, heat source symmetry and heat flow variability features. An intensity feature extraction unit is used to extract the intensity features of each structural region, the intensity features including thermal response peak value and thermal response fluctuation amplitude; The feature output unit outputs time features, intensity features, and heat source mode features as thermal response features to the structural anomaly identification module.

8. A safety inspection system for the structure of supporting equipment for high-speed rail transit according to claim 7, characterized in that, The structural anomaly identification module includes: The comparison and analysis unit is used to calculate the similarity between the thermal response features and the standard response features in the normal state heat source response template by using the feature vector comparison method. If any similarity is lower than the preset threshold, it is judged that there is a structural anomaly. The type identification unit determines the anomaly type based on the types of thermal response features with similarity below a preset threshold; The abnormal location unit obtains the corresponding code of the structural area that the comparison and analysis unit determines to have structural abnormalities as the location of the abnormal area, and generates positioning coordinate information by combining the standard layout location of the trackside equipment.

Images