Contact network thermal defect detection method and system based on large-field-of-view infrared temperature measurement

Abstract

The invention discloses an overhead line system thermal defect detection method and system based on large-view-field infrared temperature measurement, relates to the technical field of railway overhead line system detection, and discloses the overhead line system thermal defect detection method and system based on large-view-field infrared temperature measurement. Through synchronous acquisition of various data, image registration fusion, component identification, temperature extraction, thermal defect analysis and associated alarm, wide-area coverage, accurate identification and real-time response are realized, and the wide-area coverage capability, accurate identification degree and real-time response of detection can be improved, so that the detection efficiency and detection reliability of the thermal defect of the contact network are improved.

Description

A Method and System for Detecting Thermal Defects in Contact Networks Based on Wide Field-of-View Infrared Thermometry Technical Field

[0001] This application relates to the field of railway catenary inspection technology, and in particular to a method and system for detecting thermal defects in catenary based on large field-of-view infrared thermometry. Background Technology

[0002] During the operation and maintenance of railway overhead contact lines, thermal defects such as overheated clamps and loose connectors have long posed a significant threat to power supply safety. Currently, the industry commonly uses manual close-range observation or traditional onboard inspection systems for detection. Manual methods suffer from inherent drawbacks such as low efficiency and limited coverage, making effective monitoring particularly difficult during train operation, resulting in early-stage thermal defects going undetected. While onboard inspection systems incorporate infrared thermography, their design limitations, particularly small field-of-view devices, necessitate precise pre-alignment with specific components, preventing comprehensive coverage of various components like catenary cables, droppers, and clamps during high-speed dynamic inspections. Furthermore, the inherent low resolution of infrared images results in a lack of textural detail found in visible light images, blurring the boundaries of overhead contact line components and hindering identification, directly impacting the accurate extraction of temperature data. More significantly, existing technologies fragment image acquisition, data processing, and defect analysis, leading to data chain breaks during transmission and preventing the establishment of a real-time closed-loop mechanism from raw image acquisition to thermal defect alarms. The aforementioned technical bottlenecks result in insufficient reliability of the test results, making it difficult to meet the urgent needs of railway power supply systems for wide-area coverage, accurate identification, and immediate response, thus restricting the timeliness and comprehensiveness of contact wire thermal defect detection.

[0003] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0004] The main purpose of this application is to provide a method and system for detecting thermal defects in contact wires based on large field-of-view infrared thermometry, aiming to improve the detection efficiency and reliability of thermal defects in contact wires.

[0005] To achieve the above objectives, this application proposes a method for detecting thermal defects in overhead contact lines based on wide-field-of-view infrared thermometry. The method includes: acquiring synchronously collected raw infrared image data, visible light image data, and vehicle positioning data of overhead contact line components; performing image registration and fusion processing on the raw infrared image data and the visible light image data to obtain fused image data; performing overhead contact line component identification processing on the fused image data to obtain overhead contact line component area location data; extracting temperature data of the corresponding area from the raw infrared image data based on the overhead contact line component area location data; performing thermal defect analysis processing on the temperature data to obtain thermal defect status data; and performing correlation processing on the thermal defect status data and the vehicle positioning data to obtain correlated alarm data and send it to the ground terminal system.

[0006] In one embodiment, the step of performing image registration and fusion processing on the original infrared image data and the visible light image data to obtain fused image data includes: performing feature point extraction processing on the original infrared image data and the visible light image data to obtain infrared image feature point data and visible light image feature point data; performing image registration processing on the infrared image feature point data and the visible light image feature point data using a video synchronization and alignment algorithm to obtain registered infrared image data and visible light image data; and performing pixel-level fusion processing on the registered infrared image data and visible light image data based on wavelet transform to obtain the fused image data.

[0007] In one embodiment, the step of performing pixel-level fusion processing on the registered infrared image data and visible light image data based on wavelet transform to obtain the fused image data includes: constructing a weight lookup table based on an exponential function to obtain pixel fusion weight data; performing weighted fusion processing on the registered infrared image data and visible light image data according to the pixel fusion weight data to obtain preliminary fused image data; and performing morphological operations on the preliminary fused image data to adaptively fuse image seams and eliminate ghosting in overlapping areas to obtain the final fused image data.

[0008] In one embodiment, the step of performing catenary component identification processing on the fused image data to obtain catenary component region location data includes: preprocessing the fused image data to obtain preprocessed fused image data; inputting the preprocessed fused image data into a trained catenary component identification model for processing; obtaining the identification result output by the catenary component identification model; determining the bounding box position of the catenary component based on the identification result; and obtaining the catenary component region location data.

[0009] In one embodiment, the step of extracting temperature data of a corresponding region from the original infrared image data based on the contact wire component region location data includes: determining the region location of the contact wire component in the image based on the contact wire component region location data; performing boundary expansion processing on the region location to obtain expanded region location data; cropping a corresponding infrared image region from the original infrared image data based on the expanded region location data, and extracting temperature data from the cropped infrared image region.

[0010] In one embodiment, the step of performing thermal defect analysis on the temperature data to obtain thermal defect status data includes: performing bright spot rate calculation on the temperature data to obtain first thermal defect level data; performing absolute temperature discrimination processing on the temperature data to obtain second thermal defect level data; performing relative temperature difference discrimination processing on the temperature data to obtain third thermal defect level data; comparing the first thermal defect level data, the second thermal defect level data, and the third thermal defect level data, and selecting the highest level as the final thermal defect status data.

[0011] In one embodiment, the step of performing absolute temperature discrimination processing on the temperature data to obtain second thermal defect level data includes: extracting the highest temperature value from the temperature data to obtain absolute temperature data; comparing the absolute temperature data with a preset temperature range threshold to determine the temperature range in which the absolute temperature data is located; and obtaining the second thermal defect level data according to the thermal defect level corresponding to the temperature range.

[0012] In one embodiment, the step of associating the thermal defect status data with the vehicle positioning data to obtain associated alarm data and sending it to the ground terminal system includes: matching the thermal defect status data with the vehicle positioning data using timestamps to obtain spatiotemporal association data; sending the spatiotemporal association data to the ground terminal system via an onboard wireless transmission module; and the ground terminal system receiving the spatiotemporal association data and performing real-time alarm and display processing.

[0013] In one embodiment, the steps of the ground terminal system receiving the spatiotemporal correlation data and performing real-time alarm and display processing include: the ground terminal system establishing an encrypted communication tunnel with the vehicle-mounted wireless transmission module through a VPN server to receive the spatiotemporal correlation data; storing the spatiotemporal correlation data in a disk array to obtain persistent storage data; displaying the location, level, and temperature information of thermal defects graphically on a display terminal based on the persistent storage data; and performing statistical analysis on the thermal defect status data to generate a thermal defect detection report.

[0014] Furthermore, to achieve the above objectives, this application also proposes a contact wire thermal defect detection system based on wide field-of-view infrared thermometry. The contact wire thermal defect detection system based on wide field-of-view infrared thermometry includes: a memory, a processor, and a contact wire thermal defect detection program based on wide field-of-view infrared thermometry stored in the memory and executable on the processor. The contact wire thermal defect detection program based on wide field-of-view infrared thermometry is configured to implement the steps of the contact wire thermal defect detection method based on wide field-of-view infrared thermometry.

[0015] The contact wire thermal defect detection method and system proposed in this application, based on wide field-of-view infrared thermometry, achieves wide-area coverage, accurate identification, and real-time response by simultaneously acquiring multiple data, image registration and fusion, component identification, temperature extraction, thermal defect analysis, and associated alarms. This improves the wide-area coverage, accurate identification, and real-time responsiveness of the detection, thereby enhancing the detection efficiency and reliability of contact wire thermal defects. Attached Figure Description

[0016] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 is a flowchart illustrating an embodiment of the contact wire thermal defect detection method based on large field-of-view infrared thermometry according to this application; Figure 2 is a structural diagram illustrating an embodiment of the contact wire thermal defect detection system based on large field-of-view infrared thermometry according to this application.

[0019] The diagram shows the following symbols: 10, memory; 20, processor.

[0020] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] The technical solutions of this application 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. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of this application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0022] It should be understood that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0023] Existing technologies for detecting thermal defects in railway overhead contact lines generally suffer from limitations such as small detection range, low positioning accuracy, difficulty in component identification, and non-real-time closed-loop data processing. For example, small field-of-view infrared equipment cannot fully cover all types of overhead contact line equipment, infrared images have low resolution and lack visible light texture information, which limits the accuracy of component identification and temperature extraction. Furthermore, data acquisition and defect analysis are independent processes, making it impossible to achieve a real-time closed loop from image acquisition to defect alarm.

[0024] Based on this, this application provides a method for detecting thermal defects in overhead contact lines based on wide field-of-view infrared thermometry. Referring to Figure 1, the method includes steps S100 to S600, wherein: Step S100, acquiring synchronously collected original infrared image data, visible light image data, and vehicle positioning data of the overhead contact line components; Step S200, performing image registration and fusion processing on the original infrared image data and the visible light image data to obtain fused image data; Step S300, performing overhead contact line component identification processing on the fused image data to obtain overhead contact line component area location data; Step S400, extracting temperature data of the corresponding area from the original infrared image data based on the overhead contact line component area location data; Step S500, performing thermal defect analysis processing on the temperature data to obtain thermal defect status data; Step S600, performing association processing on the thermal defect status data and the vehicle positioning data to obtain associated alarm data and sending it to the ground terminal system.

[0025] In this embodiment, wide-field-of-view infrared thermometry refers to using an infrared thermal imager with a large field of view to measure temperature, thereby achieving wide-area coverage and efficient detection of the railway catenary and its components. Image registration refers to spatially aligning image data collected by different sensors (such as infrared thermal imagers and visible light cameras) so that the same physical point corresponds to the same pixel position in different images. Image fusion refers to combining registered different image data (such as infrared image data and visible light image data) to generate a single image data containing more information and easier to analyze. Thermal defect analysis and processing refers to performing a series of calculations and judgments on the temperature data extracted from the catenary component area to assess whether there is abnormal heating in the catenary component and determine its thermal defect level. The ground terminal system refers to a remote monitoring and management platform that receives thermal defect alarm data sent by onboard equipment, and stores, displays, analyzes, and generates reports.

[0026] In this embodiment, raw infrared image data, visible light image data, and vehicle positioning data of the overhead contact system components are acquired synchronously. Specifically, the raw infrared image data can be acquired in real time using a wide-field-of-view infrared thermal imager mounted on the roof of the electric locomotive, which continuously records the infrared radiation information of the overhead contact system components. Visible light image data can be acquired synchronously using a visible light camera mounted on the roof of the electric locomotive, which records the visible light image information of the overhead contact system components. Vehicle positioning data can be acquired using an onboard positioning module, for example, a Global Positioning System (GPS) receiver that outputs the vehicle's latitude and longitude information, or calculates kilometer markers and pole numbers using an odometer and preset route information.

[0027] In this embodiment, the original infrared image data and the visible light image data are then registered and fused to obtain fused image data. Image registration can be achieved by manually selecting corresponding points in the images and performing geometric transformations, or by using pre-calibrated camera parameters for rough spatial alignment. Image fusion can employ a simple pixel averaging method, which involves arithmetically averaging the corresponding pixel values ​​of the registered infrared and visible light images to generate preliminary fused image data. Next, the fused image data is processed for contact wire component identification to obtain contact wire component location data. This identification process can be implemented using preset image processing rules, such as simple target detection based on color, brightness, or shape features to identify potential contact wire components in the image and determine their approximate location. Alternatively, the operator can manually select the contact wire component area on the fused image.

[0028] Furthermore, temperature data for the corresponding area is extracted from the original infrared image data based on the location data of the contact wire component area. This extraction process can directly use the bounding box of the identified contact wire component area to extract the corresponding image block from the original infrared image, and calculate the average temperature value of all pixels within the image block as the temperature data of the component.

[0029] In this embodiment, the temperature data undergoes thermal defect analysis to obtain thermal defect status data. This thermal defect analysis can be based on a single temperature threshold. For example, if the extracted temperature data exceeds a preset alarm temperature threshold, the component is determined to have a thermal defect and its status is marked as "abnormal"; otherwise, it is marked as "normal." Finally, the thermal defect status data is correlated with the vehicle location data to obtain correlated alarm data, which is then sent to the ground terminal system. This correlation can be achieved by performing a simple time match between the thermal defect status data and the most recently collected vehicle location data. The alarm data can be sent to the ground terminal system via a wired connection or a basic wireless communication module. After receiving the alarm data, the ground terminal system can display the alarm information in text format.

[0030] Thermal defects are detected by simultaneously acquiring data from overhead contact line components using a wide-field-of-view infrared thermal imager and a visible light camera, combined with image registration and fusion technology. This improves the identification capability of overhead contact line components and the accuracy of temperature extraction. Consequently, wide-area coverage and effective detection of thermal defects in the overhead contact line are achieved. Furthermore, by correlating thermal defect status data with vehicle positioning data and transmitting it in real time to the ground terminal system, a closed-loop process from data acquisition to defect alarm is constructed, solving the problem of data link breaks in traditional methods and improving the automation level and response efficiency of railway overhead contact line operation and maintenance.

[0031] In one feasible implementation, the step of performing image registration and fusion processing on the original infrared image data and the visible light image data to obtain fused image data includes: performing feature point extraction processing on the original infrared image data and the visible light image data to obtain infrared image feature point data and visible light image feature point data; performing image registration processing on the infrared image feature point data and the visible light image feature point data using a video synchronization and alignment algorithm to obtain registered infrared image data and visible light image data; and performing pixel-level fusion processing on the registered infrared image data and visible light image data based on wavelet transform to obtain the fused image data.

[0032] In this embodiment, feature point extraction is first performed on the original infrared and visible light image data to obtain infrared and visible light image feature point data. Feature point extraction is a crucial step in image processing, aiming to identify unique and repeatable local feature points from the image. These feature points typically exhibit robustness to image rotation, scale changes, and illumination variations. For example, algorithms such as Scale Invariant Feature Transform (SIFT), Speed-Up Robust Feature Transform (SURF), or Oriented Fast and Rotationally Descriptive (ORB) can be used to detect and describe these feature points. By extracting these stable feature points, a reliable correspondence basis is provided for subsequent image registration.

[0033] In this embodiment, based on the acquired infrared and visible light image feature point data, image registration is performed using a video synchronization and alignment algorithm to obtain registered infrared and visible light image data. The purpose of image registration is to establish the geometric transformation relationship between two or more images, ensuring their spatial alignment. The video synchronization and alignment algorithm uses extracted feature point pairs and matching algorithms (such as nearest neighbor matching and cross-validation matching) to find corresponding point pairs. These corresponding point pairs are then used to estimate the geometric transformation model between the images, such as affine transformation and perspective transformation. Commonly used robust estimation methods, such as the Random Sample Consensus (RANSAC) algorithm, can effectively eliminate mismatched points, thereby improving the accuracy of the transformation model and ensuring precise pixel-level correspondence between the infrared and visible light images.

[0034] In this embodiment, the registered infrared and visible light image data are finally fused at the pixel level using wavelet transform to obtain fused image data. Wavelet transform is a multi-resolution analysis tool that can decompose an image into different frequency sub-bands, including low-frequency components (representing approximate information and main structure of the image) and high-frequency components (representing detailed information such as edges and textures). In pixel-level fusion, the registered infrared and visible light images can be decomposed using wavelet transform respectively, and then their corresponding sub-band coefficients are combined according to specific fusion rules (e.g., averaging or weighted averaging of low-frequency components, selecting the higher energy of high-frequency components, or using a more complex fusion strategy). Finally, the fused coefficients are reconstructed into a single fused image using inverse wavelet transform. This method can effectively integrate the temperature information of the infrared image and the rich texture details of the visible light image, while reducing artifacts and information loss that may occur during the fusion process.

[0035] In this embodiment, feature points are extracted from the original infrared and visible light image data using the aforementioned technical solution. Based on these feature points, image registration is performed using a video synchronization and alignment algorithm. This effectively solves the spatial misalignment problem of images under different imaging principles and viewpoints, ensuring precise alignment of the infrared and visible light images. Furthermore, pixel-level fusion processing is performed using wavelet transform. This fully utilizes the multi-scale decomposition characteristics of wavelet transform to effectively integrate the temperature information of the infrared image with the texture details of the visible light image, avoiding detail loss or artifacts that may occur with traditional fusion methods. The resulting fused image data not only contains rich thermal information but also possesses clear visual context, greatly improving the accuracy of subsequent contact network component identification and the reliability of thermal defect analysis, thereby enhancing the accuracy and robustness of the entire detection method.

[0036] In one feasible implementation, the step of performing pixel-level fusion processing on the registered infrared image data and visible light image data based on wavelet transform to obtain the fused image data includes: constructing a weight lookup table based on an exponential function to obtain pixel fusion weight data; performing weighted fusion processing on the registered infrared image data and visible light image data according to the pixel fusion weight data to obtain preliminary fused image data; and performing morphological operations on the preliminary fused image data to adaptively fuse image seams and eliminate ghosting in overlapping areas to obtain the final fused image data.

[0037] In this embodiment, pixel fusion weight data is a key parameter determining the respective contributions of the infrared and visible light images during the fusion process. This application employs a weight lookup table based on an exponential function, aiming to dynamically allocate weights according to pixel characteristics (such as brightness, gradient, and texture). The exponential function provides non-linear weight allocation, allowing higher weights to be assigned to certain regions (e.g., regions with high information content) and lower weights to others, thereby better preserving image details and suppressing noise. Constructing a weight lookup table allows for pre-calculation of weights corresponding to different pixel features, which can be quickly retrieved during actual fusion by looking up the table, improving processing efficiency.

[0038] In this embodiment, after obtaining the pixel fusion weight data, these weights are applied to the registered infrared and visible light image data. Weighted fusion processing refers to linearly or non-linearly combining the corresponding pixel values ​​of the two images according to their respective weights to generate preliminary fused image data. For example, for each pixel (x, y), the fused pixel value F(x, y) can be expressed as "F(x, y) = W_IR(x, y) × IR(x, y) + W_VL(x, y) × VL(x, y)", where "IR" and "VL" are the pixel values ​​of the infrared and visible light images, respectively, and "W_IR" and "W_VL" are the corresponding weights, satisfying "W_IR + W_VL = 1". Through the above weighting method, the balance between information preservation and noise suppression can be flexibly controlled according to the characteristics of different images and the fusion objective, thereby obtaining a preliminary fusion result that is better than simple averaging or fixed-weight fusion.

[0039] In this embodiment, the preliminary fused image data may still have some unsatisfactory fusion effects, such as unnatural image seams and ghosting in overlapping areas. Morphological operations are shape-based image processing techniques commonly used in image preprocessing, segmentation, and feature extraction. Common morphological operations include erosion, dilation, opening, and closing operations. By performing appropriate morphological operations on the preliminary fused image data, for example, closing operations can be used to fill small holes or broken areas in the image, thereby smoothing the image seams; opening operations can be used to eliminate small bright spots or noise, thereby reducing ghosting in overlapping areas. These operations can adaptively adjust parameters according to the local features of the image to optimize the fusion effect, improve image quality, and ultimately generate high-quality fused image data.

[0040] In this embodiment, by constructing a weight lookup table based on an exponential function using the aforementioned technical solution, fusion weights can be dynamically and non-linearly allocated according to pixel characteristics, making the fusion process more refined and effectively preserving the temperature information of the infrared image and the texture details of the visible light image. Furthermore, morphological operations are performed on the preliminary fused image data to adaptively smooth image seams and eliminate ghosting that may occur in overlapping areas, thereby significantly improving the overall quality and visual effect of the fused image. This high-quality fused image data provides a more reliable input for the accurate identification of subsequent contact wire components, thus improving the accuracy and reliability of thermal defect detection.

[0041] In one feasible implementation, the step of performing catenary component identification processing on the fused image data to obtain catenary component region location data includes: preprocessing the fused image data to obtain preprocessed fused image data; inputting the preprocessed fused image data into a trained catenary component identification model for processing; obtaining the identification result output by the catenary component identification model; determining the bounding box position of the catenary component based on the identification result; and obtaining the catenary component region location data.

[0042] In this embodiment, the fused image data is first preprocessed to obtain preprocessed fused image data. This preprocessing aims to optimize image quality, eliminate noise, enhance features, or adjust the image data to a format suitable for subsequent recognition model processing. Specifically, preprocessing may include, but is not limited to: image denoising (e.g., using smoothing algorithms such as Gaussian filtering and median filtering to reduce random noise in the image), contrast enhancement (e.g., improving the visual effect and feature recognizability of the image through histogram equalization or adaptive histogram equalization), image normalization (e.g., scaling pixel values ​​to a specific range, such as 0-1 or 0-255, to adapt to the model's input requirements), and image resizing (e.g., scaling the image to a fixed input size required by the recognition model). Through these preprocessing operations, the image clarity and feature saliency can be effectively improved, laying the foundation for subsequent accurate recognition.

[0043] In this embodiment, the preprocessed fused image data is then input into a trained overhead contact line component recognition model for processing. This model is an intelligent model trained on a large amount of labeled data, capable of automatically learning and recognizing overhead contact line components in the fused image data. This model can be implemented using various advanced computer vision technologies, such as deep learning-based convolutional neural network (CNN) models, including target detection algorithms like YOLO (You Only Look Once), Faster R-CNN, or Mask R-CNN. These models extract and learn image features through multi-layer neural networks, thereby achieving accurate classification and localization of overhead contact line components (such as contact wires, catenary wires, insulators, and droppers). During the training phase, the model learns the visual features of different components, and during the inference phase, it detects components in the input image based on these features.

[0044] Further, the recognition results output by the contact wire component recognition model are obtained, and the bounding box positions of the contact wire components are determined based on the recognition results, thereby obtaining the contact wire component region location data. The recognition model typically outputs the category label, confidence score, and bounding box coordinates of each detected object in the image. The bounding box position is usually represented by the coordinates of the top-left and bottom-right corners of a rectangle, or the coordinates of the center point and its width and height. After obtaining these raw recognition results, post-processing may be required, such as non-maximum suppression (NMS) to eliminate overlapping redundant bounding boxes, and reliable detection results are filtered based on a confidence threshold. Ultimately, these precise bounding box coordinates constitute the contact wire component region location data, providing a spatial basis for accurately extracting temperature data of the corresponding region from the raw infrared image data.

[0045] In this embodiment, preprocessing the fused image data significantly improves image quality and enhances the features of the overhead contact line components, thus providing a higher-quality input for the recognition model. Based on this, automated processing using the trained overhead contact line component recognition model enables rapid and accurate identification of these components and precise determination of their bounding box positions. This model-based recognition method not only overcomes the limitations of traditional manual recognition or simple image processing methods, which are susceptible to environmental factors, inefficient, and inaccurate, but also adapts to the needs of overhead contact line component recognition in different scenarios and complex backgrounds. Ultimately, the obtained precise location data of the overhead contact line component regions provides reliable spatial positioning for the subsequent accurate extraction of temperature information from the original infrared image data, thereby ensuring the accuracy and reliability of thermal defect analysis and significantly improving the automation level and overall performance of overhead contact line thermal defect detection.

[0046] In one feasible implementation, the step of extracting temperature data of a corresponding region from the original infrared image data based on the contact wire component region location data includes: determining the region location of the contact wire component in the image based on the contact wire component region location data; performing boundary expansion processing on the region location to obtain expanded region location data; cropping a corresponding infrared image region from the original infrared image data based on the expanded region location data, and extracting temperature data from the cropped infrared image region.

[0047] In this embodiment, the step of determining the location of the contact wire component in the image based on the aforementioned contact wire component location data aims to accurately locate the specific spatial range of the contact wire component in the image using the contact wire component location data obtained in the preceding processing. The location data is typically represented in the form of bounding boxes, pixel masks, or polygonal regions, clearly defining the starting point, size, or contour information of the contact wire component in the image coordinate system. Through this step, the system can clearly identify the target area for subsequent temperature data extraction.

[0048] In this embodiment, the aforementioned area location is then expanded by a predetermined ratio to obtain expanded area location data. This step involves appropriately magnifying the determined contact wire component area location. The "predetermined ratio" can be a fixed pixel value, such as expanding outward by 5 pixels; it can also be a relative ratio, such as increasing the width and height of the bounding box by 5% or 10% respectively; or it can be an adaptively adjusted ratio based on the type, size, or historical data of the contact wire component. For example, a larger expansion ratio can be used for smaller components to ensure coverage; a smaller expansion ratio can be used for larger components. This expansion processing aims to compensate for minor deviations that may exist during image registration and takes into account the physical phenomenon that heat may diffuse to surrounding areas during actual operation of the contact wire component, ensuring that potential hotspots of the component and its neighboring areas are fully covered when extracting temperature data.

[0049] In this embodiment, finally, based on the expanded regional location data, a corresponding infrared image region is extracted from the original infrared image data, and temperature data is extracted from the extracted infrared image region. This step utilizes the expanded regional location data to accurately crop an infrared image segment containing the contact wire component and its surrounding area from the original infrared image data. Each pixel value in the original infrared image data typically corresponds directly to or can be converted into a temperature value. Therefore, extracting temperature data from the extracted infrared image region involves reading and collecting the temperature values ​​of all pixels within that region. This may include obtaining statistical information such as the region's highest temperature, average temperature, and temperature distribution, providing basic data for subsequent thermal defect analysis.

[0050] In this embodiment, by employing the above-described technical solution, when extracting temperature data of the contact wire components from the original infrared image data, a predetermined proportion of boundary expansion processing is first applied to the identified contact wire component area. This expansion processing effectively compensates for minor errors that may exist in image registration and fully considers the physical characteristics of heat diffusion from the contact wire components to the surrounding areas. Therefore, when subsequently extracting infrared images and temperature data from the expanded area, it ensures that all thermal information of the component and its neighboring areas is captured, avoiding the omission of key hot spots due to overly precise boundary identification or thermal diffusion effects. This significantly improves the completeness and accuracy of temperature data extraction, providing a more reliable and comprehensive data foundation for subsequent thermal defect analysis, thereby enhancing the accuracy and reliability of contact wire thermal defect detection.

[0051] In one feasible implementation, the step of performing thermal defect analysis on the temperature data to obtain thermal defect status data includes: performing bright spot rate calculation on the temperature data to obtain first thermal defect level data; performing absolute temperature discrimination processing on the temperature data to obtain second thermal defect level data; performing relative temperature difference discrimination processing on the temperature data to obtain third thermal defect level data; comparing the first thermal defect level data, the second thermal defect level data, and the third thermal defect level data, and selecting the highest level as the final thermal defect status data.

[0052] In this embodiment, the bright spot rate calculation process involves setting one or more preset temperature thresholds, counting the number of pixels in the corresponding area of ​​the original infrared image data whose temperature exceeds these thresholds, and calculating the percentage of these pixels in the total number of pixels in that area. Based on the calculated bright spot rate, it can be compared with a preset bright spot rate range to determine the first thermal defect level. For example, a higher bright spot rate indicates a more significant local overheating phenomenon, and the corresponding thermal defect level is also higher.

[0053] In this embodiment, the absolute temperature discrimination processing involves directly analyzing the highest temperature value in the extracted temperature data and comparing it with a series of preset absolute temperature thresholds. Based on the temperature range within which the highest temperature value falls, the corresponding thermal defect level can be determined as the second thermal defect level data. For example, when the highest temperature value exceeds a certain critical threshold, even if other indicators are normal, it may be judged as a higher-level thermal defect. Furthermore, the relative temperature difference discrimination processing involves comparing the extracted temperature data with the ambient temperature or the reference temperature of a similar normal component to calculate the temperature difference value. By comparing this temperature difference value with a preset temperature difference threshold, the third thermal defect level data can be determined. This method effectively eliminates the influence of ambient temperature fluctuations on the detection results and more accurately reflects the abnormal heating of the component itself.

[0054] In this embodiment, to ensure the comprehensiveness and security of the detection, this application comprehensively compares the first thermal defect level data, the second thermal defect level data, and the third thermal defect level data obtained through the three different discrimination methods described above. The final thermal defect status data will be the highest level among these three levels. This "highest level selected" strategy can maximize the identification of potential thermal defects, avoid missed detections due to the limitations of a single indicator, and thus improve the reliability of the detection.

[0055] In this embodiment, the present application overcomes the limitations of a single thermal defect discrimination standard through the above technical solution. By comprehensively considering three different discrimination dimensions—bright spot rate, absolute temperature, and relative temperature difference—the thermal defect status of the contact wire components can be comprehensively evaluated from multiple perspectives, such as local overheating range, component temperature limits, and environmental adaptability. This multi-dimensional fusion discrimination mechanism significantly improves the accuracy and robustness of thermal defect detection, effectively avoiding misjudgments or omissions that may be caused by a single indicator. Simultaneously, the strategy of selecting the highest level as the final thermal defect status data ensures the highest priority response to any potential hazards, thereby greatly improving the reliability and safety of contact wire thermal defect detection.

[0056] In one feasible implementation, the step of performing absolute temperature discrimination processing on the temperature data to obtain second thermal defect level data includes: extracting the highest temperature value from the temperature data to obtain absolute temperature data; comparing the absolute temperature data with a preset temperature range threshold to determine the temperature range in which the absolute temperature data is located; and obtaining the second thermal defect level data according to the thermal defect level corresponding to the temperature range.

[0057] In this embodiment, the highest temperature value is extracted from the temperature data to obtain absolute temperature data. The aim is to identify the most representative high-temperature point, i.e., the highest temperature value, from the temperature data of the contact wire component area. This highest temperature value is a key indicator for assessing whether the component is overheating and its severity. After obtaining the temperature data of the contact wire component area, the temperature values ​​of all pixels within that area can be traversed, and the highest value can be selected. For example, a maximum value filtering algorithm from image processing can be used, or the temperature matrix can be directly scanned to determine the highest temperature point and its corresponding temperature value within that area. This highest temperature value is the absolute temperature data used for subsequent absolute temperature determination.

[0058] In this embodiment, the absolute temperature data is then compared with a preset temperature range threshold to determine the temperature range in which the absolute temperature data falls. This step involves comparing the highest temperature value actually measured with a series of preset temperature ranges, thereby discretizing the continuous temperature value into specific temperature ranges. Different temperature ranges typically correspond to different degrees of thermal defect severity. The preset temperature range threshold can be determined based on industry standards, equipment operating experience, historical fault data, or expert knowledge. For example, multiple threshold points can be set, such as T1, T2, T3, etc., forming multiple temperature ranges: [0, T1), [T1, T2), [T2, T3), [T3, ∞). After receiving the absolute temperature data, the system compares it with these thresholds one by one to determine which range it falls into. For example, if the absolute temperature data is greater than or equal to T1 but less than T2, it is determined to be in the [T1, T2) range.

[0059] In this embodiment, based on the temperature range, the second thermal defect level data is obtained according to the thermal defect level corresponding to the temperature range. This step directly maps the temperature range determined in the previous step to a predefined thermal defect level. This is a key step in transforming quantified temperature information into an operable, graded defect status, providing a basis for subsequent alarms and maintenance. Typically, a mapping table or rule set is established to associate each temperature range with a specific thermal defect level. For example, [0, T1) might correspond to "normal," [T1, T2) to "minor defect," [T2, T3) to "moderate defect," and [T3, ∞) to "severe defect." After determining the temperature range where the absolute temperature data falls, the system queries this mapping table to obtain the corresponding thermal defect level, which serves as the second thermal defect level data. These levels can be digital (e.g., levels 1, 2, 3, 4) or descriptive (e.g., normal, attention, warning, alarm).

[0060] In this embodiment, through the above technical solution, this application can perform refined classification and discrimination of the absolute temperature of catenary components. First, by accurately extracting the highest temperature value from the temperature data, the most severe potential thermal defects are captured, avoiding the problem of average temperature or local low temperature masking the true high temperature risk. Subsequently, the highest temperature value is compared with a preset temperature range threshold, effectively discretizing the continuous temperature data into temperature ranges with clear physical meaning, thus providing a preliminary quantitative basis for the severity of thermal defects. Finally, the temperature range is directly mapped to the corresponding thermal defect level, making the thermal defect discrimination process standardized, automated, and operable. This classification and discrimination mechanism enables the system to identify thermal defects of different severity more accurately and promptly, avoiding fuzzy judgments and improving the accuracy and reliability of thermal defect analysis. When combined with bright spot rate calculation and relative temperature difference discrimination, a multi-dimensional and more comprehensive thermal defect assessment system can be formed, significantly improving the intelligence level and early warning capability of catenary thermal defect detection, and providing a more solid guarantee for railway operation safety.

[0061] In one feasible implementation, the step of associating the thermal defect status data with the vehicle positioning data, obtaining the associated alarm data, and sending it to the ground terminal system includes: matching the thermal defect status data with the vehicle positioning data using timestamps to obtain spatiotemporal correlation data; sending the spatiotemporal correlation data to the ground terminal system via an onboard wireless transmission module; and the ground terminal system receiving the spatiotemporal correlation data and performing real-time alarm and display processing.

[0062] In this embodiment, to accurately correlate thermal defect status data with vehicle positioning data, this application proposes to perform timestamp matching between the thermal defect status data and vehicle positioning data to obtain spatiotemporally correlated data. Specifically, when generating thermal defect status data, the system synchronously records its generation timestamp; simultaneously, the vehicle positioning data also includes a precise timestamp during collection. By comparing and aligning these timestamps, for example using sliding time window matching or nearest neighbor matching algorithms, thermal defect status data at a specific time point can be bound to vehicle latitude and longitude, kilometer markers, and pole number information at the same or similar time points. In this way, each piece of thermal defect information can be assigned an accurate geographical location and time attribute, forming correlated data with spatiotemporal attributes.

[0063] In this embodiment, to ensure that the aforementioned spatiotemporal correlation data can be transmitted to the ground monitoring center in a timely manner, this application further proposes to send the spatiotemporal correlation data to the ground terminal system via an onboard wireless transmission module. The onboard wireless transmission module can be a 4G / 5G cellular communication module, a satellite communication module, or a railway-specific GSM-R communication module, etc. Its main function is to encapsulate the correlated spatiotemporal data and transmit it using a wireless network. During transmission, the data can be encrypted as needed to ensure the security of data transmission. Simultaneously, to improve the reliability and efficiency of transmission, the module can also integrate data compression, error checking, and retransmission mechanisms.

[0064] In this embodiment, upon receiving the spatiotemporal correlation data, the ground terminal system will immediately perform real-time alarm and display processing. Specifically, the ground terminal system receives spatiotemporal correlation data from the vehicle-mounted system through its wireless receiving module or network interface. Once it receives spatiotemporal correlation data containing thermal defect information, the system will trigger corresponding alarm mechanisms based on the level and severity of the thermal defect, such as audible and visual alarms, SMS notifications, or email alerts, to quickly notify relevant maintenance personnel. Simultaneously, the system will visualize the location, level, and temperature of the thermal defect on a Geographic Information System (GIS) map, route map, or a customized monitoring interface. For example, it can mark thermal defect points with different colors or icons and provide detailed attribute query functions, enabling maintenance personnel to intuitively understand the distribution and specific situation of the thermal defects.

[0065] In this embodiment, the detected thermal defect status data is timestamped with precise vehicle positioning data using the aforementioned technical solution, thereby generating spatiotemporal correlation data containing thermal defect location and severity information. This effectively solves the problem of the disconnect between thermal defect information and geographical location information, ensuring that each thermal defect can be accurately located. Subsequently, this crucial spatiotemporal correlation data is transmitted in real time to the ground terminal system via the onboard wireless transmission module, ensuring the timeliness and efficiency of information transmission. Upon receiving the data, the ground terminal system can immediately perform real-time alarm and display processing, enabling maintenance personnel to grasp the specific location, severity, and related temperature information of the thermal defect at the first opportunity, thus allowing for rapid response and decision-making. This significantly improves the real-time performance, accuracy, and maintenance efficiency of overhead contact line thermal defect detection, effectively reducing operational risks caused by thermal defects and ensuring the safety of railway transportation.

[0066] In one feasible implementation, the steps of real-time alarm and display processing after the ground terminal system receives the spatiotemporal correlation data include: the ground terminal system establishing an encrypted communication tunnel with the vehicle-mounted wireless transmission module through a VPN server to receive the spatiotemporal correlation data; storing the spatiotemporal correlation data in a disk array to obtain persistent storage data; displaying the location, level, and temperature information of thermal defects graphically on a display terminal based on the persistent storage data; and performing statistical analysis on the thermal defect status data to generate a thermal defect detection report.

[0067] In this embodiment, the ground terminal system establishes an encrypted communication tunnel with the vehicle-mounted wireless transmission module through a VPN server to receive the spatiotemporal correlation data, aiming to ensure the security, integrity, and confidentiality of data transmission. The VPN server can be deployed on the ground terminal system side, establishing an encrypted tunnel with the vehicle-mounted wireless transmission module via the Internet or other public networks. Before sending data, the vehicle-mounted wireless transmission module establishes a connection with the VPN server through a VPN client, and all data is transmitted encrypted within the tunnel. Common VPN protocols include IPsec, OpenVPN, L2TP / IPsec, etc. These protocols can effectively prevent data from being eavesdropped on, tampered with, or forged during transmission, especially in the complex wireless communication environment along railway lines, ensuring the secure transmission of sensitive thermal defect data.

[0068] In this embodiment, the spatiotemporal correlated data is then stored in a disk array to obtain persistent storage data. A redundant array of independent disks (RAID) is a technology that combines multiple independent physical hard drives to form a single logical hard drive. It provides data redundancy (such as RAID1, RAID 5, RAID 6, etc.), effectively preventing data loss due to the failure of a single hard drive, while also improving data read and write performance. The spatiotemporal correlated data, including key information such as thermal defect status, location, and timestamps, is written to the disk array to form persistent storage data. This storage method not only ensures long-term data preservation but also provides a stable and reliable data source for subsequent data querying, analysis, and report generation.

[0069] Based on this, and using the persistently stored data, the location, level, and temperature information of thermal defects are graphically displayed on a display terminal. The display terminal can be a computer monitor, a large-screen display wall, or a mobile device screen. The graphical display may include: marking the specific location of the thermal defect on a Geographic Information System (GIS) map (combining kilometer markers and pole number information), and using different colors or icons to represent the level of the thermal defect (e.g., normal, caution, warning, danger); displaying infrared or fused images of the thermal defect area, overlaid with a temperature distribution map; showing the temperature change trend over time in the form of a line graph or bar chart; and providing detailed tabular information listing the highest temperature, average temperature, level, and occurrence time of each thermal defect. This visualization method allows maintenance personnel to intuitively and clearly grasp the thermal defect status of the overhead contact line, facilitating rapid understanding and decision-making.

[0070] In this embodiment, the system also performs statistical analysis on the thermal defect status data to generate a thermal defect detection report. Statistical analysis may include: statistically analyzing the quantity and severity distribution of thermal defects by time period (e.g., day, week, month, year); statistically analyzing the density and type of thermal defects by geographical region (e.g., line segment, pole number interval); analyzing the frequency and severity of thermal defects in specific components (e.g., clamps, insulators, connectors); and performing trend analysis on historical data to predict potential thermal defect development. Based on these analysis results, the system can automatically generate a standardized thermal defect detection report, which includes statistical charts, key data summaries, anomaly lists, and recommended maintenance measures. This report is an important basis for the operation and maintenance management of the overhead contact system, helping to deeply explore the patterns behind the thermal defect data and providing decision support for the maintenance, repair, and fault prediction of the overhead contact system.

[0071] In this embodiment, through the above technical solution, after receiving spatiotemporal correlation data sent by the vehicle-mounted system, the ground terminal system can effectively ensure the security and integrity of data transmission by establishing an encrypted communication tunnel, preventing the leakage or tampering of sensitive information. Simultaneously, storing the received data in a disk array ensures persistent storage and high reliability, providing a solid foundation for subsequent data querying, tracing, and analysis. Based on this, the system can display key information such as the location, level, and temperature of thermal defects on the display terminal in an intuitive graphical manner, greatly improving the efficiency of maintenance personnel's perception and decision-making speed regarding thermal defect conditions. Furthermore, statistical analysis of thermal defect status data and generation of detection reports not only allows for in-depth exploration of the occurrence patterns and development trends of thermal defects, providing a scientific basis for preventative maintenance and precise repair of the overhead contact system, but also significantly improves the intelligence level of overhead contact system thermal defect detection and management, thereby effectively reducing operational risks and ensuring the safety of railway transportation.

[0072] In the embodiments of this application, the contact wire thermal defect detection method based on wide field-of-view infrared thermometry achieves wide-area coverage, accurate identification, and real-time response by simultaneously acquiring multiple data, image registration and fusion, component identification, temperature extraction, thermal defect analysis, and associated alarms. This improves the wide-area coverage, accurate identification, and real-time responsiveness of the detection, thereby enhancing the detection efficiency and reliability of contact wire thermal defects.

[0073] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the contact wire thermal defect detection method based on large field-of-view infrared thermometry. Any simple modifications based on this technical concept are within the protection scope of this application.

[0074] This application also provides a contact wire thermal defect detection system based on wide field-of-view infrared thermometry. Referring to Figure 2, the contact wire thermal defect detection system based on wide field-of-view infrared thermometry includes: a memory 10, a processor 20, and a contact wire thermal defect detection program based on wide field-of-view infrared thermometry stored in the memory 10 and executable on the processor 20. The contact wire thermal defect detection program based on wide field-of-view infrared thermometry is configured to implement the steps of the contact wire thermal defect detection method based on wide field-of-view infrared thermometry.

[0075] The contact wire thermal defect detection system based on wide field-of-view infrared thermometry provided in this application adopts the contact wire thermal defect detection method based on wide field-of-view infrared thermometry in the above embodiments, which can improve the detection efficiency and reliability of contact wire thermal defects. Compared with the prior art, the beneficial effects of the contact wire thermal defect detection system based on wide field-of-view infrared thermometry provided in this application are the same as those of the contact wire thermal defect detection method based on wide field-of-view infrared thermometry provided in the above embodiments, and other technical features of the contact wire thermal defect detection system based on wide field-of-view infrared thermometry are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0076] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0077] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. All equivalent structural transformations made under the technical concept of this application using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included within the scope of patent protection of this application.

Claims

1. A method for detecting thermal defects in contact wires based on large field-of-view infrared thermometry, characterized in that, The method includes: acquiring synchronously collected raw infrared image data, visible light image data, and vehicle positioning data of the overhead contact system components; performing image registration and fusion processing on the raw infrared image data and the visible light image data to obtain fused image data; performing overhead contact system component identification processing on the fused image data to obtain overhead contact system component area location data; extracting temperature data of the corresponding area from the raw infrared image data based on the overhead contact system component area location data; performing thermal defect analysis processing on the temperature data to obtain thermal defect status data; and performing correlation processing between the thermal defect status data and the vehicle positioning data to obtain correlated alarm data and sending it to the ground terminal system.

2. The contact wire thermal defect detection method based on large field-of-view infrared thermometry as described in claim 1, characterized in that, The steps of performing image registration and fusion processing on the original infrared image data and the visible light image data to obtain fused image data include: performing feature point extraction processing on the original infrared image data and the visible light image data to obtain infrared image feature point data and visible light image feature point data; performing image registration processing on the infrared image feature point data and the visible light image feature point data using a video synchronization and alignment algorithm to obtain registered infrared image data and visible light image data; and performing pixel-level fusion processing on the registered infrared image data and visible light image data based on wavelet transform to obtain fused image data.

3. The contact wire thermal defect detection method based on large field-of-view infrared thermometry as described in claim 2, characterized in that, The steps for performing pixel-level fusion processing on the registered infrared and visible light image data based on wavelet transform to obtain the fused image data include: constructing a weight lookup table based on an exponential function to obtain pixel fusion weight data; performing weighted fusion processing on the registered infrared and visible light image data according to the pixel fusion weight data to obtain preliminary fused image data; and performing morphological operations on the preliminary fused image data to adaptively fuse image seams and eliminate ghosting in overlapping areas to obtain the final fused image data.

4. The contact wire thermal defect detection method based on large field-of-view infrared thermometry as described in claim 1, characterized in that, The steps of performing catenary component identification processing on the fused image data to obtain catenary component region location data include: preprocessing the fused image data to obtain preprocessed fused image data; inputting the preprocessed fused image data into a trained catenary component identification model for processing; obtaining the identification result output by the catenary component identification model; determining the bounding box position of the catenary component based on the identification result; and obtaining the catenary component region location data.

5. The contact wire thermal defect detection method based on large field-of-view infrared thermometry as described in claim 1, characterized in that, The step of extracting temperature data of a corresponding region from the original infrared image data based on the location data of the contact wire component region includes: determining the region location of the contact wire component in the image based on the location data of the contact wire component region; performing boundary expansion processing on the region location by a predetermined ratio to obtain expanded region location data; and extracting the corresponding infrared image region from the original infrared image data based on the expanded region location data, and extracting temperature data from the extracted infrared image region.

6. The contact wire thermal defect detection method based on large field-of-view infrared thermometry as described in claim 1, characterized in that, The steps of performing thermal defect analysis on the temperature data to obtain thermal defect status data include: calculating the bright spot rate of the temperature data to obtain first thermal defect level data; performing absolute temperature discrimination processing on the temperature data to obtain second thermal defect level data; performing relative temperature difference discrimination processing on the temperature data to obtain third thermal defect level data; comparing the first thermal defect level data, the second thermal defect level data, and the third thermal defect level data, and selecting the highest level as the final thermal defect status data.

7. The contact wire thermal defect detection method based on large field-of-view infrared thermometry as described in claim 6, characterized in that, The steps of performing absolute temperature discrimination processing on the temperature data to obtain the second thermal defect level data include: extracting the highest temperature value from the temperature data to obtain absolute temperature data; comparing the absolute temperature data with a preset temperature range threshold to determine the temperature range in which the absolute temperature data is located; and obtaining the second thermal defect level data according to the thermal defect level corresponding to the temperature range.

8. The contact wire thermal defect detection method based on large field-of-view infrared thermometry as described in claim 1, characterized in that, The steps of associating the thermal defect status data with the vehicle positioning data, obtaining the associated alarm data, and sending it to the ground terminal system include: matching the thermal defect status data with the vehicle positioning data using timestamps to obtain spatiotemporal correlation data; sending the spatiotemporal correlation data to the ground terminal system via an onboard wireless transmission module; and the ground terminal system receiving the spatiotemporal correlation data and performing real-time alarm and display processing.

9. The contact wire thermal defect detection method based on large field-of-view infrared thermometry as described in claim 8, characterized in that, After receiving the spatiotemporal correlation data, the ground terminal system performs real-time alarm and display processing as follows: the ground terminal system establishes an encrypted communication tunnel with the vehicle-mounted wireless transmission module through a VPN server to receive the spatiotemporal correlation data; stores the spatiotemporal correlation data in a disk array to obtain persistent storage data; based on the persistent storage data, displays the location, level, and temperature information of thermal defects graphically on a display terminal; and performs statistical analysis on the thermal defect status data to generate a thermal defect detection report.

10. A contact wire thermal defect detection system based on large field-of-view infrared thermometry, characterized in that, The contact wire thermal defect detection system based on wide field-of-view infrared thermometry includes: a memory, a processor, and a contact wire thermal defect detection program based on wide field-of-view infrared thermometry stored in the memory and executable on the processor. The contact wire thermal defect detection program based on wide field-of-view infrared thermometry is configured to implement the steps of the contact wire thermal defect detection method based on wide field-of-view infrared thermometry as described in any one of claims 1 to 9.

Images