A method, system and device for detecting defects of a catenary support post and a storage medium

Abstract

The application relates to the technical field of defect detection, and provides a catenary support defect detection method, system, device and storage medium, which solve the problems of insufficient detection sensitivity and accuracy. The application calibrates a preset temperature anomaly judgment benchmark by acquiring material defect correlation data of a catenary support and a visible light image of a catenary support to be detected, and combining a target material type and a target thermal conductivity coefficient of the catenary support to be detected. Based on this, a surface temperature field distribution of the catenary support to be detected is collected. A signal filtering device is adjusted, an infrared signal corresponding to the temperature field distribution is filtered and converted, a digital electric signal is obtained, the digital electric signal is synchronized and aligned with the visible light image, integrated feature data is obtained, temperature information and visual information are extracted, the target material type is combined, a defect detection result of the catenary support to be detected is generated, and the accuracy and anti-interference capability of multi-material catenary support defect detection are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of defect detection, in particular to a catenary support post defect detection method, system, device and storage medium. BACKGROUND

[0002] In high-speed railway and urban rail transit systems, catenary support posts are key infrastructure that support and fix catenaries, and the structural integrity thereof is directly related to power supply safety and train operation stability. With the rapid expansion of rail transit networks and the continuous improvement of train operating speeds, efficient and accurate real-time monitoring of the health status of catenary support posts has become an urgent technical requirement in operation and management. Especially in complex and variable outdoor environmental conditions, cracks, rust or material degradation may occur in the support posts due to long-term bearing of mechanical load, environmental corrosion, temperature alternation and other factors, and if not discovered in time, it may cause serious safety accidents.

[0003] Currently, research and application attempts have been made to construct a multi-modal defect recognition system based on fixed threshold judgment by fusing infrared thermal imaging and visible light visual analysis technology. This scheme uses an infrared camera to collect temperature distribution data of the support post surface, and at the same time, a high-definition camera is used to obtain the corresponding visible light image. Then, an image registration algorithm is used to realize the spatial alignment of the two types of data, and a unified temperature anomaly threshold is used for defect discrimination. However, the existing scheme has significant limitations in actual application, for example, due to the use of a unified fixed temperature judgment threshold and fixed signal filtering parameters, it fails to fully consider the essential differences in thermal conductivity characteristics of different material support posts, resulting in large fluctuations in detection sensitivity and accuracy when facing concrete, steel or composite material support posts. SUMMARY

[0004] The purpose of the present application is to provide a catenary support post defect detection method, system, device and storage medium to solve the problem of insufficient detection sensitivity and accuracy when facing concrete, steel or composite material support posts in the prior art.

[0005] To solve the above technical problems, in a first aspect, the present application provides a catenary support post defect detection method, comprising:

[0006] Obtaining material defect correlation data of a catenary support post and a visible light image of a catenary support post to be detected, the material defect correlation data including crack-related parameters and a corresponding first thermal conductivity coefficient of a concrete support post, rust-related parameters and a corresponding second thermal conductivity coefficient of a steel support post, and a temperature anomaly threshold and a corresponding third thermal conductivity coefficient of a supplementary material support post;

[0007] According to the target material type and the corresponding target thermal conductivity of the contact net support to be detected, the preset temperature anomaly judgment reference is adaptively temperature calibrated combined with the material defect correlation data, and a calibrated anomaly judgment reference is obtained.

[0008] Based on the calibrated anomaly judgment reference, the surface temperature field distribution of the contact net support to be detected is collected.

[0009] Adjust the working frequency band of the signal filtering device to obtain an adjusted signal filtering device for anti-interference filtering processing of the infrared signal corresponding to the temperature field distribution, and obtain a filtered temperature field distribution.

[0010] Convert the filtered temperature field distribution into a digitized electrical signal, and perform timestamp synchronization and alignment processing on the digitized electrical signal and the visible light image to obtain integrated feature data.

[0011] From the integrated feature data, temperature information and visual information are extracted, and the target material type of the contact net support to be detected is combined to generate a defect detection result of the contact net support to be detected.

[0012] Optionally, according to the target material type and the corresponding target thermal conductivity of the contact net support to be detected, the preset temperature anomaly judgment reference is adaptively temperature calibrated combined with the material defect correlation data, and a calibrated anomaly judgment reference is obtained, including:

[0013] According to the visible light image, the target material type of the contact net support to be detected is determined, and the target material type is a concrete material, a steel material or a supplementary material.

[0014] According to the target material type, the target thermal conductivity and the target temperature anomaly threshold corresponding to the target material type are extracted from the material defect correlation data.

[0015] The difference value between the target thermal conductivity and the reference thermal conductivity is calculated, and the boundary value of the preset general temperature anomaly judgment reference is adjusted combined with the target temperature anomaly threshold, to obtain a calibrated anomaly judgment reference matched with the target material type.

[0016] Optionally, the working frequency band of the signal filtering device is adjusted to obtain an adjusted signal filtering device for anti-interference filtering processing of the infrared signal corresponding to the temperature field distribution, and a filtered temperature field distribution is obtained, including:

[0017] extract a target filtering frequency range corresponding to the target material type from the material defect correlation data, wherein the concrete material corresponds to a first filtering frequency range, the steel material corresponds to a second filtering frequency range, and the supplementary material corresponds to a third filtering frequency range;

[0018] select a target working frequency range within the target filtering frequency range from working frequency ranges of the signal filtering device according to the target thermal conductivity of the contact net support to be detected, wherein the working frequency range is a filtering frequency range preset by the signal filtering device when the signal filtering device is manufactured and is suitable for a general material contact net support;

[0019] replace the working frequency range of the signal filtering device with the target working frequency range to obtain an adjusted signal filtering device;

[0020] filter interference signals in an infrared signal corresponding to the surface temperature field distribution through the adjusted signal filtering device to obtain a filtered temperature field distribution.

[0021] In a second aspect, the present application provides a contact net support defect detection system, comprising:

[0022] an acquisition module configured to acquire material defect correlation data of a contact net support and a visible light image of a contact net support to be detected, wherein the material defect correlation data comprises crack-related parameters and a corresponding first thermal conductivity of a concrete support, rust-related parameters and a corresponding second thermal conductivity of a steel support, and a temperature anomaly threshold and a corresponding third thermal conductivity of a supplementary material support;

[0023] a calibration module configured to perform adaptive temperature calibration processing on a preset temperature anomaly determination reference according to a target material type and a corresponding target thermal conductivity of the contact net support to be detected in combination with the material defect correlation data to obtain a calibrated anomaly determination reference;

[0024] an acquisition module configured to acquire a surface temperature field distribution of the contact net support to be detected based on the calibrated anomaly determination reference;

[0025] an adjustment module configured to adjust a working frequency range of a signal filtering device to obtain an adjusted signal filtering device for anti-interference filtering processing of an infrared signal corresponding to the temperature field distribution to obtain a filtered temperature field distribution;

[0026] a processing module configured to convert the filtered temperature field distribution into a digital electrical signal, and perform time stamp synchronization and alignment processing on the digital electrical signal and the visible light image to obtain integrated feature data;

[0027] The generating module is configured to extract temperature information and visual information from the integrated feature data, combine the target material type of the contact net support to be detected, and generate a defect detection result of the contact net support to be detected.

[0028] In a third aspect, the present application provides an electronic device, comprising:

[0029] a memory configured to store a computer program;

[0030] a processor configured to execute the computer program to implement the steps of the contact net support defect detection method according to the first aspect.

[0031] In a fourth aspect, the present application provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is executable by a processor to implement the steps of the contact net support defect detection method according to the first aspect.

[0032] The present application has the following beneficial effects:

[0033] The contact net support post defect detection method provided in the application can overcome the problem of inaccurate judgment of fixed threshold in a multi-material scene, improve the applicability and sensitivity of the temperature criterion, acquire material defect correlation data containing the heat conduction characteristics of different material support posts and defect correlation parameters, and adaptively calibrate the preset temperature anomaly judgment benchmark in combination with the target material type of the support post to be detected and the thermal conductivity thereof. The surface temperature field distribution is collected based on the calibrated anomaly judgment benchmark, and the working frequency band of the signal filtering device is dynamically adjusted to implement targeted anti-interference filtering on the infrared signal, thereby enhancing the stability and reliability of the temperature field data in a complex environment. Then, the filtered temperature field is converted into a digital electrical signal and synchronized and aligned with the visible light image in time stamp, thereby realizing consistent fusion of multi-modal data in the time and space dimensions. Finally, the temperature and visual information is extracted from the integrated feature data, and the defect detection result is generated in combination with the material type of the support post, which not only fully utilizes the complementarity of thermal response and appearance characteristics, but also improves the accuracy and robustness of defect recognition through the material perception analysis mechanism, and the overall technical path realizes high-precision, anti-interference and adaptive detection of defects of a multi-material contact net support post. Further, the temperature values of each region and the visual features such as line shape, color block distribution and color change are extracted from the fused data, and the defect-related parameters matched with the material of the support post are called, the temperature anomaly region is identified based on the adaptively calibrated judgment benchmark, and then the visual features and defect representation parameters specific to the material of the region are comprehensively analyzed to generate the final defect judgment result. The misjudgment and omission problem caused by lack of material differentiation modeling in the traditional fusion method is solved, the recognition ability for hidden defects is enhanced, and the adaptability and discrimination accuracy are higher when dealing with environmental thermal interference and thermal response differences of different materials, thereby breaking through the detection barriers caused by unified processing of multi-material structures in the prior art. BRIEF DESCRIPTION OF DRAWINGS

[0034] In order to more clearly illustrate the technical solutions of the embodiments of the application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0035] Figure 1 A flowchart of a contact net support post defect detection method provided by an embodiment of the application;

[0036] Figure 2 A specific implementation schematic diagram of a contact net support post defect detection method provided by an embodiment of the application;

[0037] Figure 3A structural schematic diagram of a catenary support defect detection system provided by an embodiment of the present application. DETAILED DESCRIPTION

[0038] To solve the problem that the existing technology uses fixed temperature threshold and unified filtering strategy to accurately identify different types of damage such as concrete cracks and steel corrosion, and is easily affected by environmental thermal interference and material thermal conductivity differences to produce misjudgment, the present application focuses on the relationship between material properties and thermal response, and by pre-establishing an associated data model of the thermal conductivity coefficient of different material supports and typical defect parameters, the material type of the support to be detected is first identified during the detection process, and the temperature anomaly judgment criterion and signal filtering frequency band are dynamically adjusted accordingly to realize personalized calibration and noise suppression of the temperature field data; then the processed infrared temperature information is time-synchronized and spatially aligned with the visible light image, the thermal distribution characteristics and visual topographic information are fused, and finally the defect judgment is completed combined with the material properties, thereby improving the stability and accuracy of the detection results in the multi-material scene.

[0039] In order to enable personnel in the technical field to better understand the present application scheme, the present application will be further described in detail below in combination with the drawings and specific embodiments. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0040] The core of the present application is to provide a catenary support defect detection method, and a flowchart of one specific embodiment of the method is shown in Figure 1 The method comprises the following steps.

[0041] Step 101: Obtain the material defect correlation data of the catenary support and the visible light image of the catenary support to be detected, wherein the material defect correlation data comprises crack-related parameters and a corresponding first thermal conductivity coefficient of a concrete support, corrosion-related parameters and a corresponding second thermal conductivity coefficient of a steel support, and a temperature anomaly threshold and a corresponding third thermal conductivity coefficient of a supplementary material support.

[0042] In this step, the catenary support refers to the key infrastructure in high-speed railway and urban rail transit systems for supporting and fixing the catenary (power supply line for trains), reflecting the installation support state of the catenary, manufactured based on railway engineering design standards, and long-term subjected to mechanical load generated by train operation, corrosion effect of outdoor environment, and influence of day and night temperature variation, whose structural integrity is directly related to power supply safety and train operation stability. Material defect correlation data refers to a structured data set pre-established, correlating different material catenary supports and their defect judgment parameters, and heat conduction characteristic parameters, reflecting defect characteristics and heat conduction law of different material supports, used for subsequent temperature anomaly judgment benchmark calibration, signal filtering parameter adjustment, and defect type judgment, obtained based on a large number of defect experiments and thermal conduction test data statistical analysis of different material supports. The catenary support to be detected refers to the specific catenary support currently requiring defect detection, reflecting the object whose health state needs to be evaluated, determined based on suspected problems found by rail transit operation and maintenance plan or daily inspection, and is the direct object of all detection steps such as visible light image acquisition, temperature field distribution acquisition, signal processing, and defect judgment. Visible light image refers to the image obtained by shooting the surface of the catenary support to be detected by a high-definition industrial camera, reflecting the visual features of the support surface, used to determine the target material type of the support to be detected and subsequent spatial alignment with digital electrical signals, obtained based on visible light imaging principles, and the image resolution needs to meet the accuracy requirements of material identification and physical marker positioning. Crack-related parameters refer to a set of characteristic parameters used to judge whether a concrete support has a crack defect, reflecting the typical physical characteristics of concrete support cracks, obtained based on experimental data of crack generation and expansion of concrete materials, specifically including crack length threshold, crack direction characteristics, and crack fracture morphology, used for comparison with visual information of temperature anomaly areas to judge crack defects. The first thermal conductivity coefficient refers to a physical parameter reflecting the heat conduction capacity of a concrete material catenary support, with a unit of watts per meter kelvin (W / (m・K)), reflecting the speed of heat transfer of concrete material, obtained based on thermal conduction experiments of concrete standard test pieces, used to calculate the adjustment difference value of the temperature anomaly judgment benchmark and determine the signal filtering frequency range of concrete material supports, and is a heat conduction characteristic parameter exclusive to concrete material supports. Corrosion-related parameters refer to a set of characteristic parameters used to judge whether a steel support has a corrosion defect, reflecting the typical physical characteristics of steel support corrosion, obtained based on corrosion experimental data of steel in different environments, specifically including corrosion area threshold, corrosion edge morphology, and corrosion color gradient range, used for comparison with visual information of temperature anomaly areas to judge corrosion defects.The second thermal conductivity coefficient refers to a physical parameter reflecting the heat conduction capacity of the contact net support made of steel material, with a unit of watt per meter kelvin (W / (m K)), reflecting the speed of heat transfer of the steel material, and is obtained based on the heat conduction experiment test of the standard test piece of the steel material, and is used for calculating the adjustment difference value of the temperature anomaly judgment criterion and determining the signal filtering frequency range of the support made of steel material, and is a heat conduction characteristic parameter exclusive to the support made of steel material. The concrete material refers to one of the common materials of the contact net support, which is made by mixing and stirring cement, sand aggregate, water and additives in a certain proportion and then hardening, reflects the material composition characteristics of the support, has the characteristics of high hardness, large density, low thermal conductivity and strong corrosion resistance (relative to steel material), and is prone to crack defects due to stress concentration in long-term use, and needs to be detected through crack-related parameters and the first thermal conductivity coefficient. The steel material refers to one of the common materials of the contact net support, which is mainly composed of iron and carbon elements, and part of which contains alloy elements (such as manganese and chromium), reflects the material composition characteristics of the support, has the characteristics of high strength, good toughness and high thermal conductivity, and is prone to corrosion defects in outdoor humid and salt fog environment, and needs to be detected through corrosion-related parameters and the second thermal conductivity coefficient. The supplementary material support refers to the contact net support made of other materials (such as glass fiber composite material and carbon fiber composite material) other than the concrete material and the steel material, reflects the material composition characteristics of the support, and its thermal conductivity, mechanical properties and defect types (such as material aging and delamination) are quite different from those of the concrete and the steel material, and need to be detected using the exclusive temperature anomaly threshold value and the third thermal conductivity coefficient, and are used for covering the detection needs of the special material support that may appear in the rail transit system. The temperature anomaly threshold value refers to a numerical standard exclusive to the supplementary material support for defining whether the surface temperature of the support is abnormal, with a unit of degree Celsius (℃), reflecting the temperature boundary between the normal working state and the abnormal state of the supplementary material support, and is obtained based on the thermal characteristics test of the supplementary material support in different health states, and is used for adjusting the temperature anomaly judgment criterion of the supplementary material support, and ensures the accuracy of the temperature anomaly judgment of the supplementary material support. The third thermal conductivity coefficient refers to a physical parameter reflecting the heat conduction capacity of the supplementary material support, with a unit of watt per meter kelvin (W / (m K)), reflecting the speed of heat transfer of the supplementary material, and is obtained based on the heat conduction experiment test of the standard test piece of the supplementary material, and is used for calculating the adjustment difference value of the temperature anomaly judgment criterion and determining the signal filtering frequency range of the supplementary material support, and is a heat conduction characteristic parameter exclusive to the supplementary material support.

[0043] In the embodiments of the present application, in order to build a basic data system for detecting defects of catenary support, two types of core data are first acquired: one is material defect correlation data of catenary support, and the other is a visible light image of the catenary support to be detected. The material defect correlation data of the catenary support is obtained by calling a pre-constructed material defect thermal conductivity characteristic database of the catenary support. The data specifically includes crack-related parameters and a corresponding first thermal conductivity coefficient of a concrete support, corrosion-related parameters and a corresponding second thermal conductivity coefficient of a steel support, and a temperature anomaly threshold and a corresponding third thermal conductivity coefficient of a supplementary material support. The visible light image of the catenary support to be detected is obtained by omnidirectionally shooting the catenary support to be detected by a high-definition industrial camera. The image can directly present the surface appearance characteristics of the support, and provide a visual basis for subsequently determining the target material type of the support to be detected.

[0044] Step 102: According to the target material type and the corresponding target thermal conductivity coefficient of the catenary support to be detected, the pre-set temperature anomaly judgment reference is adaptively temperature calibrated in combination with the material defect correlation data, to obtain a calibrated anomaly judgment reference.

[0045] In this step, the target material type refers to the material category to which the contact net support to be detected actually belongs, and is only one of concrete material, steel material or supplementary material, reflecting the core material property of the support to be detected, which is determined by analyzing the visual features such as color, texture and surface texture of the support in the visible light image, and is used to extract the target thermal conductivity, target temperature anomaly threshold and target defect related parameters from the material defect related data, which is the core basis for all subsequent material adaptability operations. The target thermal conductivity refers to the thermal conductivity corresponding to the target material type of the contact net support to be detected. If the target material is concrete, it corresponds to the first thermal conductivity, if it is steel material, it corresponds to the second thermal conductivity, and if it is supplementary material, it corresponds to the third thermal conductivity, reflecting the actual heat conduction capacity of the support to be detected, and is used to calculate the adjustment difference value of the temperature anomaly judgment criterion and determine the target working frequency band of the signal filtering device. The preset temperature anomaly judgment criterion refers to the temperature anomaly judgment standard preset for the contact net support of general material (such as ordinary steel and conventional concrete), which includes the upper and lower boundary values of temperature, reflects the temperature range of the support of general material in normal working state, and is set based on the average thermal characteristic experimental data of the support of general material. It needs to be adaptively adjusted according to the target material type and target thermal conductivity of the support to be detected, and is the initial template for generating the calibrated anomaly judgment criterion. The calibrated anomaly judgment criterion refers to the exclusive judgment standard obtained by adjusting the preset temperature anomaly judgment criterion according to the target material type, target thermal conductivity and target temperature anomaly threshold of the contact net support to be detected, which includes the temperature upper and lower boundary values completely adapted to the material of the support to be detected, reflects the temperature range of the support to be detected in normal working state, is obtained by adjusting the preset temperature anomaly judgment criterion, and is used to guide the setting of temperature field distribution collection parameters and subsequent temperature anomaly region marking, solving the problem of inadaptability of general criterion to different materials.

[0046] Step 103: Collecting the surface temperature field distribution of the contact net support to be detected based on the calibrated anomaly judgment criterion.

[0047] In this step, the surface temperature field distribution refers to the spatial distribution state of temperature values of all regions on the surface of the contact net support to be detected, which is usually presented in the form of a thermal map, reflecting the temperature difference of the support surface, and is collected based on the scanning of an infrared thermal imager. The temperature range of the calibrated anomaly judgment criterion needs to be adapted, and the temperature anomaly region it contains may correspond to the defect position of the support, which is the original data for subsequent signal filtering processing.

[0048] In the embodiment of the present application, in order to obtain temperature data reflecting defects of the to-be-detected support, the collection parameters (such as temperature measurement range and sampling frequency) of the infrared thermal imager are adjusted based on the post-calibration abnormality determination reference, so as to ensure that the collection range is adapted to the temperature interval of the post-calibration abnormality determination reference, and then the surface of the to-be-detected overhead contact system support is scanned and collected region by region by using the adjusted infrared thermal imager, and the temperature values of each region of the surface of the support are recorded, so as to form a surface temperature field distribution of the to-be-detected overhead contact system support. The temperature field distribution directly presents the temperature spatial difference of the surface of the support, and is the original temperature data source for subsequent anti-interference processing.

[0049] Step 104: adjusting the working frequency band of the signal filtering device to obtain an adjusted signal filtering device, so as to perform anti-interference filtering processing on the infrared signal corresponding to the temperature field distribution, and obtain a filtered temperature field distribution.

[0050] In this step, the working frequency band of the signal filtering device refers to the frequency interval of the signal filtering device for filtering signals, which is in hertz (Hz) and reflects the filtering ability of the device for signals of different frequencies. The working frequency band is pre-set as a frequency range suitable for infrared signals of overhead contact system supports made of general materials when the device is shipped, and needs to be adjusted to a target working frequency band according to the material characteristics of the to-be-detected support, so as to filter out effective temperature signals and filter out interference signals. The adjusted signal filtering device refers to the signal filtering device after the working frequency band is adjusted to the target working frequency band suitable for the material of the to-be-detected overhead contact system support. The adjusted signal filtering device reflects the targeted filtering ability of the device for infrared signals of a specific material, is obtained based on the original signal filtering device by adjusting the working frequency band, and is used to perform anti-interference processing on the infrared signal corresponding to the temperature field distribution, so as to ensure that the output filtered temperature field distribution is accurate and reliable. The infrared signal refers to an infrared band electromagnetic signal corresponding to the surface temperature field distribution of the to-be-detected overhead contact system support. The frequency of the infrared signal is positively correlated with the temperature of the surface of the support, reflects the temperature information of the surface of the support, is synchronously obtained by the infrared thermal imager when the temperature field distribution is collected, and contains effective temperature signals and interference signals (such as environmental electromagnetic radiation and clutter generated by external light). The interference needs to be filtered by the adjusted signal filtering device. The filtered temperature field distribution refers to the surface temperature field distribution of the to-be-detected overhead contact system support after the interference signals are filtered by the adjusted signal filtering device. The filtered temperature field distribution reflects the more accurate temperature spatial difference of the surface of the support, is obtained by filtering the infrared signal corresponding to the original temperature field distribution, eliminates the influence of irrelevant interference signals, and is a high-quality temperature data source for subsequent conversion into digital electrical signals.

[0051] Step 105: converting the filtered temperature field distribution into a digital electrical signal, and performing time stamp synchronization and alignment processing on the digital electrical signal and the visible light image, to obtain integrated feature data.

[0052] In this step, the digitalized electrical signal refers to the electrical signal formed by combining the temperature field analog signals of each acquisition unit in the filtered temperature field distribution in the order of the acquisition unit arrangement, reflecting the digital form information of the surface temperature of the detected support, obtained based on analog-digital conversion and signal combination, and needs to be time-stamped synchronized and spatially aligned with the visible light image. It is the source of temperature information in the integrated feature data. The integrated feature data refers to the comprehensive data set formed by associating and storing the temperature signal data of the target digitalized electrical signal segment after position adaptation with the visual image data of the corresponding pixel range in the visible light image, reflecting the fusion result of the temperature information and visual information of the detected support, obtained based on the synchronization and alignment of the digitalized electrical signal and the visible light image, and providing complete data support for subsequent extraction of temperature information, visual information and generation of defect detection results.

[0053] Step 106: Extracting temperature information and visual information from the integrated feature data, combining the target material type of the detected overhead contact system support, and generating a defect detection result of the detected overhead contact system support.

[0054] In this step, the temperature information refers to the temperature-related data corresponding to each pixel range extracted from the integrated feature data, mainly including the specific temperature values of each pixel range, reflecting the surface temperature of the support corresponding to each pixel range, obtained based on the extraction of temperature signal information in the integrated feature data, and used for comparison with the calibrated abnormality determination reference to mark the temperature abnormal area. The visual information refers to the visual-related data corresponding to each pixel range extracted from the integrated feature data, mainly including the line shape, color block distribution and color depth change of each pixel range, reflecting the appearance features of the support corresponding to each pixel range, obtained based on the extraction of visual image information in the integrated feature data, and used for comparison with the target defect-related parameters to determine the defect type. The defect detection result refers to the final generated determination conclusion about the health status of the detected overhead contact system support, including no defect or defect + specific defect type (such as crack defect, rust defect, and supplementary material-specific defect) and defect location information, reflecting the actual health status of the detected support, and providing direct basis for defect handling for rail transit operation and maintenance personnel.

[0055] The embodiments of the present application fully adapt to the characteristics of supports of different materials, improve the sensitivity and accuracy of defect detection of overhead contact system supports in complex outdoor environments, reduce safety accidents caused by defects not being discovered in time, and effectively meet the urgent technical needs of efficient and accurate real-time monitoring of the health status of overhead contact system supports in rail transit operation and maintenance management.

[0056] The application provides one specific embodiment, step 102, according to the target material type of the contact net support to be detected and the corresponding target thermal conductivity coefficient, combining the material defect correlation data, the preset temperature anomaly judgment reference is adaptively temperature calibrated, and the calibrated anomaly judgment reference is obtained, specifically including the following steps:

[0057] Step 201: According to the visible light image, the target material type of the contact net support to be detected is determined, and the target material type is concrete material, steel material or supplementary material.

[0058] In this step, the target temperature anomaly threshold refers to the exclusive numerical standard (unit: Celsius degree ℃) corresponding to the target material type of the contact net support to be detected, which defines whether the surface temperature of the material support is abnormal, reflects the temperature boundary between the normal and abnormal states of the specific material support, is obtained based on the thermal characteristic experimental data statistics of the material support under different health states (no defect, slight defect, serious defect), is used to adjust the boundary value of the preset general temperature anomaly judgment reference in combination with the difference value of the target thermal conductivity coefficient and the reference thermal conductivity coefficient, and ensure that the subsequent temperature anomaly judgment can adapt to the material characteristics of the detected support, and avoid the judgment deviation caused by the material difference.

[0059] In the embodiment of the application, in order to determine the material attribute of the detected support to carry out subsequent adaptive operation, the visible light image of the contact net support to be detected is analyzed by image recognition technology, the texture (the concrete material usually presents rough, sand grain feeling texture, the steel material presents smooth, metal reflection texture, the supplementary material such as composite material presents uniform fiber texture) on the surface of the support in the image is detected, the color (the concrete is light gray or dark gray, the steel material is silver gray or red brown after rust, the color of the supplementary material is various but more uniform) and the surface texture (the concrete is hard and has no ductility gloss, the steel material has metal toughness gloss, the supplementary material has no metal texture and the surface is more smooth) are analyzed, the visual features are matched with the preset material visual feature comparison table (the table is established based on a large number of visible light image features of concrete, steel material and supplementary material supports) one by one, if the concrete features are matched, the target material type is determined as concrete material, if the steel material features are matched, the target material type is determined as steel material, and if the supplementary material features are matched, the target material type is determined as supplementary material, and finally the target material type of the contact net support to be detected is obtained, which will be directly used as the basis for extracting parameters from the material defect correlation data in the subsequent step.

[0060] Step 202: According to the target material type, the target thermal conductivity coefficient and the target temperature anomaly threshold corresponding to the target material type are extracted from the material defect correlation data.

[0061] In the embodiments of the present application, in order to obtain the heat conduction and temperature determination parameters matched with the material of the to-be-detected support, the corresponding parameters are extracted from the material defect association data with the target material type determined in the first step as the index: if the target material type is concrete material, the first heat conduction coefficient corresponding to the concrete support in the data (as the target heat conduction coefficient of the to-be-detected support) and the temperature anomaly threshold value exclusive to the concrete support (as the target temperature anomaly threshold value of the to-be-detected support) are extracted; if the target material type is steel material, the second heat conduction coefficient corresponding to the steel support (as the target heat conduction coefficient) and the temperature anomaly threshold value exclusive to the steel support (as the target temperature anomaly threshold value) are extracted; if the target material type is supplementary material, the third heat conduction coefficient corresponding to the supplementary material support (as the target heat conduction coefficient) and the temperature anomaly threshold value of the supplementary material support (as the target temperature anomaly threshold value) are extracted. Through the extraction process, the target heat conduction coefficient and the target temperature anomaly threshold value completely corresponding to the material of the to-be-detected overhead contact system support are obtained, and the two parameters will be the core input for calculating the difference value and adjusting the reference in the third step.

[0062] Step 203: Calculate the difference value between the target heat conduction coefficient and the reference heat conduction coefficient, adjust the boundary value of the preset universal temperature anomaly determination reference in combination with the target temperature anomaly threshold value, and obtain the calibrated anomaly determination reference matched with the target material type.

[0063] In this step, the reference thermal conductivity refers to the reference value of the thermal conductivity of the contact net support pillar of the general material (such as ordinary carbon steel and conventional concrete) corresponding to the preset general temperature anomaly judgment reference, which reflects the average heat conduction capacity of the general material pillar, is obtained based on the standard test piece heat conduction experiment test of a large number of general material pillars and taking the average value, and is used to calculate the difference value of the target thermal conductivity of the detected pillar. The difference value is the core basis for judging the adjustment range of the preset general reference boundary value, and a correlation bridge between the general reference and the material-specific reference is established through the coefficient. The difference value refers to the numerical difference between the target thermal conductivity of the detected contact net support pillar and the reference thermal conductivity (the calculation method is the target thermal conductivity minus the reference thermal conductivity), which reflects the difference degree of the heat conduction capacity of the detected pillar material and the general material. If the difference value is positive, it means that the heat conduction capacity of the detected material is stronger than that of the general material, and if the difference value is negative, it means that it is weaker than the general material. It is used to quantitatively determine the adjustment direction (increase or decrease) and adjustment range (large or small) of the preset general temperature anomaly judgment reference boundary value, and is a key quantitative index for realizing the temperature judgment reference material adaptation. The preset general temperature anomaly judgment reference refers to the initial temperature anomaly judgment standard preset for the contact net support pillar of the general material, which includes the upper boundary value (the critical value for judging high temperature anomaly) and the lower boundary value (the critical value for judging low temperature anomaly), and reflects the temperature range of the general material pillar in the normal working environment. It is set based on the heat distribution data of the general material pillar in the standard environment (temperature 25℃, humidity 50%) under the defect-free state, and is the basis template for subsequent material-specific calibration, which solves the problem of no initial judgment standard, and provides a unified reference framework for the reference adjustment of different materials. The boundary value refers to the upper and lower critical values (in Celsius) for determining whether the temperature is abnormal in the preset general temperature anomaly judgment reference and the calibrated anomaly judgment reference, including the upper boundary value (exceeding which is judged as high temperature anomaly) and the lower boundary value (being lower than which is judged as low temperature anomaly), which reflects the temperature effective range of the specific judgment reference, and is obtained based on the thermal conductivity characteristics of the material (reflected by the difference value) and the target temperature anomaly threshold adjustment. The adjusted boundary value can accurately match the material of the detected pillar, ensure the accuracy of the temperature anomaly area marking, and avoid missed judgment or misjudgment.

[0064] In the embodiments of the present application, in order to obtain the temperature anomaly determination standard suitable for the material of the detected support, first, the reference thermal conductivity coefficient (which is a reference value corresponding to the preset general temperature anomaly determination reference, obtained based on the average thermal conductivity test of the general material catenary support) is determined, and then the difference value between the target thermal conductivity coefficient extracted in the second step and the reference thermal conductivity coefficient is calculated (the calculation method is to subtract the reference thermal conductivity coefficient from the target thermal conductivity coefficient, and if the result is positive, it means that the target thermal conductivity coefficient is greater than the reference value, and if the result is negative, it means that the target thermal conductivity coefficient is less than the reference value); then, the preset general temperature anomaly determination reference (which is the initial temperature determination standard suitable for the general material support, including the upper and lower boundary values) is analyzed, and the boundary values of the reference are adjusted according to the size of the difference value and the target temperature anomaly threshold value extracted in the second step: if the difference value is positive (the target thermal conductivity coefficient is greater, and the heat transfer is faster), the upper boundary value of the general reference is appropriately increased based on the target temperature anomaly threshold value, the lower boundary value is maintained or slightly lowered (to avoid misjudgment of low temperature anomaly); if the difference value is negative (the target thermal conductivity coefficient is smaller, and the heat transfer is slower), the upper boundary value of the general reference is appropriately reduced based on the target temperature anomaly threshold value, and the lower boundary value is appropriately increased (to avoid missing the low temperature anomaly); through the above adjustment process, the calibrated anomaly determination reference accurately matched with the target material type of the detected catenary support is finally obtained, which will be used for setting the collection parameters of the surface temperature field distribution.

[0065] The embodiments of the present application solve the core defects of adapting all materials to a unified threshold, so that the temperature anomaly determination reference can accurately match the material characteristics of the detected support, providing accurate determination standards for subsequent surface temperature field distribution collection and temperature anomaly area marking, reducing the detection deviation caused by material differences, improving the consistency and accuracy of catenary support defect detection of different materials, and meeting the precise detection needs of diversified material supports in complex rail transit scenarios.

[0066] The present application provides a specific embodiment, step 104, adjusting the working frequency range of the signal filtering device to obtain an adjusted signal filtering device for anti-interference filtering processing of the infrared signal corresponding to the temperature field distribution, and obtaining the filtered temperature field distribution, which specifically includes the following steps:

[0067] Step 401: Based on the target material type and the corresponding target thermal conductivity coefficient of the detected catenary support, the target filtering frequency range corresponding to the target material type is extracted from the material defect association data, wherein the concrete material corresponds to the first filtering frequency range, the steel material corresponds to the second filtering frequency range, and the supplementary material corresponds to the third filtering frequency range.

[0068] In this step, the target filtering frequency range refers to the frequency interval (unit: Hz) corresponding to the material type of the contact net support target to be detected, which is used to filter the infrared signal. It reflects the frequency distribution range of the infrared effective signal of a specific material support, which is obtained based on the infrared signal characteristics experiment of the material support (by testing the infrared signal frequency of the material at different temperatures, and counting the frequency interval of the effective signal), and stored in the material defect association data, which is used to select the target working frequency range from the general frequency range of the signal filtering device, to ensure that only the effective infrared signal of the material is retained in the subsequent filtering, and to avoid filtering useful data or retaining interference signals. The first filtering frequency range refers to the target filtering frequency range specially adapted for the concrete material contact net support, which reflects the infrared effective signal frequency characteristics (usually relatively low frequency, such as 5-15 Hz) of the concrete material due to its low thermal conductivity coefficient (corresponding to the first thermal conductivity coefficient). It is obtained based on a large amount of infrared signal test data of concrete supports, stored in the material defect association data, and used for signal filtering frequency range selection of concrete material supports, to solve the filtering adaptation problem caused by the frequency difference between concrete material signals and other material signals. The second filtering frequency range refers to the target filtering frequency range specially adapted for the steel material contact net support, which reflects the infrared effective signal frequency characteristics (usually relatively high frequency, such as 10-25 Hz) of the steel material due to its high thermal conductivity coefficient (corresponding to the second thermal conductivity coefficient). It is obtained based on a large amount of infrared signal test data of steel supports, stored in the material defect association data, and used for signal filtering frequency range selection of steel material supports, to ensure that the high-frequency effective signal of the steel material is not filtered. The third filtering frequency range refers to the target filtering frequency range specially adapted for the supplementary material contact net support, which reflects the infrared effective signal frequency characteristics (frequency range has no overlap with the first two, such as 2-8 Hz or 22-30 Hz) of the supplementary material (such as glass fiber composite material) due to its large difference in thermal characteristics from concrete and steel (corresponding to the third thermal conductivity coefficient). It is obtained based on the exclusive infrared signal test data of the supplementary material support, stored in the material defect association data, and used for signal filtering frequency range selection of the supplementary material support, to cover the filtering needs of special materials.

[0069] In the embodiments of the present application, to determine the signal filtering frequency range suitable for the material of the to-be-detected support, first, the target material type and the corresponding target thermal conductivity of the to-be-detected overhead contact system support are determined, then the target material type + target thermal conductivity are taken as the joint search condition, and the signal filtering frequency interval corresponding to the material is selected from the material defect association data, and the interval is the target filtering frequency range. The concrete material has a relatively low thermal conductivity and a relatively low infrared signal frequency, and corresponds to the first filtering frequency range. The steel material has a relatively high thermal conductivity and a relatively high infrared signal frequency, and corresponds to the second filtering frequency range. The supplementary material has a large difference in thermal conductivity from the first two, and corresponds to the exclusive third filtering frequency range. The target filtering frequency range obtained through the extraction process will be used as the basis for selecting the working frequency range of the signal filtering device.

[0070] Step 402: According to the target thermal conductivity of the to-be-detected overhead contact system support, the target working frequency range within the target filtering frequency range is selected from the working frequency range of the signal filtering device. The working frequency range is the filtering frequency range preset by the signal filtering device when it is shipped and is suitable for the overhead contact system support of general material.

[0071] In this step, the target working frequency range refers to a specific filtering frequency sub-interval (unit: Hz) selected from the general frequency range of the signal filtering device and suitable for the target material type and the target thermal conductivity of the to-be-detected overhead contact system support. It reflects the accurate frequency range of the infrared effective signal of the to-be-detected support and is obtained based on the signal frequency characteristics of the target filtering frequency range and the target thermal conductivity. It is used to replace the preset frequency range of the signal filtering device and is the core parameter for the adjusted signal filtering device to achieve accurate anti-interference. The overhead contact system support of general material refers to the overhead contact system support widely used in the rail transit system and having relatively uniform material characteristics. It mainly includes ordinary carbon steel material support (thermal conductivity close to the second thermal conductivity) and conventional concrete material support (thermal conductivity close to the first thermal conductivity). It reflects the material type of the support in most scenarios and is based on the statistics of commonly used materials of the rail transit support. It is the design basis for the preset working frequency range of the signal filtering device and meets the basic filtering needs of most general scenarios. The filtering frequency range of the overhead contact system support of general material refers to the filtering frequency interval (unit: Hz) preset by the signal filtering device when it is shipped and suitable for the overhead contact system support of general material. It reflects the average frequency range of the infrared effective signal of the overhead contact system support of general material and is set by the manufacturer based on a large amount of infrared interference test data of the overhead contact system support of general material (such as 8-18 Hz). It can perform basic anti-interference processing on the infrared signal of the overhead contact system support of general material, but needs to be adjusted to the target working frequency range according to the specific material of the to-be-detected support. It is the initial template for subsequent frequency range adjustment.

[0072] In the embodiments of the present application, in order to select the specific working frequency band suitable for the to-be-detected support from the general filtering frequency band, first, the working frequency band preset when the signal filtering device leaves the factory is read. The frequency band is set by the manufacturer based on the infrared signal characteristics of the general material catenary support (such as ordinary carbon steel support and conventional concrete support), and is suitable for the anti-interference needs of most general materials, which is called the filtering frequency band of the general material catenary support. Then, combined with the target thermal conductivity coefficient of the to-be-detected catenary support, the infrared signal frequency characteristics corresponding to the coefficient (such as the signal frequency being close to the upper limit when the target thermal conductivity coefficient is high, and the signal frequency being close to the lower limit when the target thermal conductivity coefficient is low) are analyzed. In the target filtering frequency band range, the sub-frequency band with the highest matching degree to the target thermal conductivity coefficient is selected (for example, when the target thermal conductivity coefficient is the second thermal conductivity coefficient of steel and the target filtering frequency band range is 10-20 Hz, if the signal frequency corresponding to the second thermal conductivity coefficient is concentrated in 15-18 Hz, 15-18 Hz is selected as the sub-frequency band), which is the target working frequency band, ensuring that the subsequent filtering can accurately retain the effective infrared signal of the to-be-detected support.

[0073] Step 403: replacing the working frequency band of the signal filtering device with the target working frequency band to obtain an adjusted signal filtering device.

[0074] In the embodiments of the present application, in order to complete the parameter adjustment of the signal filtering device, first, the current working frequency band (i.e. the preset filtering frequency band of the general material catenary support) of the device is read through the parameter adjustment interface (such as a physical knob or a software control interface) of the signal filtering device. Then, the frequency band parameters of the device are modified according to the numerical range of the target working frequency band (such as modifying the original 12-18 Hz to 15-18 Hz), and it is confirmed through the frequency band verification function of the device that the modified frequency band has been stabilized and taken effect. Finally, the adjusted signal filtering device with the working frequency band completely adapted to the material of the to-be-detected catenary support is obtained, which will be used for subsequent anti-interference processing of the infrared signal.

[0075] Step 404: filtering the interference signal in the infrared signal corresponding to the surface temperature field distribution through the adjusted signal filtering device to obtain a filtered temperature field distribution.

[0076] In this step, the interference signal refers to the noise signal mixed in the infrared signal corresponding to the surface temperature field distribution of the to-be-detected catenary support, which is irrelevant to the temperature of the support, mainly comes from environmental electromagnetic radiation (such as 50 Hz interference generated by the surrounding power lines), external light reflection (such as high-frequency noise generated by strong light irradiation) and device self-noise (such as infrared thermal imager circuit noise), reflects the invalid information in the infrared signal, and will cause the temperature field distribution data to be distorted, so it needs to be filtered through the adjusted signal filtering device to ensure the accuracy of the subsequent digital signal.

[0077] In the embodiment of the present application, in order to eliminate the interference signal in the temperature field distribution and obtain accurate temperature data, first, the infrared signal corresponding to the surface temperature field distribution (including the effective signal corresponding to the surface temperature of the detected support column, and the clutter signal generated by the environmental electromagnetic radiation and the reflection of external light) is extracted, and then the infrared signal is input into the adjusted signal filtering device. The device will filter the infrared signal according to the range of the target working frequency band, and only keep the effective temperature signal within the target working frequency band, and filter out the interference signal outside the frequency band (such as the 50Hz electromagnetic interference signal in the environment and the high-frequency clutter generated by external strong light). After filtering, the filtered temperature field distribution containing only the effective temperature information of the detected catenary support column surface is output as the original data for conversion into digital electrical signals.

[0078] The embodiment of the present application extracts the exclusive target filtering frequency band range based on the material and the thermal conductivity coefficient, solves the problem of mismatching of the frequency band range, selects the accurate target working frequency band from the general frequency band in combination with the thermal conductivity coefficient, avoids the problem of too wide / narrow frequency band, filters the interference through the adjusted device, ensures that only the effective signal is kept, makes the signal filtering completely adapt to the material characteristics of the detected support column, solves the limitation of fixed filtering parameters, improves the accuracy of the filtered temperature field distribution, provides a high-quality temperature data source for converting the temperature field into digital electrical signals and realizing the fusion of temperature and visual data, and further improves the accuracy of the final defect detection, and meets the anti-interference detection requirements of catenary support columns of different materials in complex outdoor environments.

[0079] The present application provides a specific embodiment, step 105, converting the filtered temperature field distribution into a digital electrical signal, synchronizing and aligning the digital electrical signal with the visible light image, and obtaining integrated feature data, specifically including the following steps:

[0080] Step 501: According to the surface area division rule of the detected catenary support column, the surface of the detected catenary support column is divided into a plurality of acquisition units.

[0081] In this step, the surface area division rule refers to the standard and method for dividing the surface of the contact net support to be detected into multiple collection units, reflecting the structured division logic of the support surface, based on the geometric shape (such as cylindrical, rectangular) of the support, the size and the detection accuracy requirement (minimum defect recognition size), and is formulated, including the shape (such as fan-shaped, rectangular) of the divided unit, the size (such as 0.5 meters x 0.5 meters), the arrangement mode (such as according to the height and the circumferential direction), for converting the continuous support surface into discrete units that can be processed individually, providing a basic framework for the correspondence of the temperature signal and the image data. The collection unit refers to the independent small area obtained by dividing the surface of the contact net support to be detected according to the surface area division rule, reflecting the basic detection unit of the support surface, based on the surface area division rule, each collection unit corresponds to a specific local area of the support surface, for collecting the temperature field signal of the area individually and associating with the corresponding pixel range in the visible light image, realizing the fine correspondence of temperature and visual information.

[0082] In the embodiments of the present application, in order to realize the fine correspondence of the temperature and visual information of the support surface to be detected, first, the surface area division rule of the contact net support to be detected is determined, which is based on the structural characteristics (such as the cylindrical support is divided into segments of 0.5 meters in height direction and fan-shaped areas of 90° in circumferential direction) of the support and the detection accuracy requirement (minimum defect recognition size), and the size, shape and arrangement mode of the divided unit are determined; then according to the rule, the surface of the contact net support to be detected is divided into multiple independent and continuous small areas using a region division tool (such as virtual grid division based on the three-dimensional model of the support), each small area is a collection unit, and the sum of all collection units covers the entire surface of the support, which will be the basic corresponding unit of the temperature signal and the image data in the subsequent process.

[0083] Step 502: converting the temperature field signal corresponding to each collection unit in the filtered temperature field distribution into a discrete digital signal, and combining all discrete digital signals into a digitized electrical signal.

[0084] In this step, the temperature field signal refers to the analog electric signal corresponding to each acquisition unit in the filtered temperature field distribution, reflecting the temperature change of the unit, whose voltage value is positively correlated with the temperature of the acquisition unit (the higher the temperature, the greater the voltage), reflecting the real-time temperature state of the acquisition unit, obtained based on the temperature detection of the acquisition unit by the infrared thermal imager, and after anti-interference processing by the signal filtering device, needs to be converted into a discrete digital signal for subsequent processing. The discrete digital signal refers to the signal composed of a series of digital quantities obtained by analog-to-digital conversion of the temperature field signal (analog signal) of each acquisition unit, each digital quantity corresponding to the temperature value at a specific sampling time, reflecting the digital expression of the temperature of the acquisition unit, obtained based on timed sampling and quantization of the temperature field signal, and multiple discrete digital signals are combined in sequence to form a digital electric signal, which is convenient for computer storage and processing. The digital electric signal refers to the complete digital signal formed by combining the discrete digital signals of all acquisition units in the order of their arrangement on the surface of the support pillar, reflecting the digital overall distribution of the temperature of the entire surface of the detected overhead contact system support pillar, obtained based on the ordered splicing of the discrete digital signals, containing the temperature information of each acquisition unit at different times, and needs to be time-synchronized and spatially aligned with the visible light image to realize data fusion.

[0085] In the embodiments of the present application, in order to convert the temperature field distribution into a digital signal that can be processed, the temperature field signal of each acquisition unit is converted by an analog-to-digital conversion device, the continuously changing temperature field signal is sampled at fixed time intervals, and converted into a series of discrete digital quantities (i.e. discrete digital signal, each digital quantity corresponding to the temperature value at the sampling time); finally, according to the arrangement order of the acquisition units on the surface of the support pillar (such as from top to bottom, from 0° to 360° on the circumference), the discrete digital signals of all acquisition units are spliced in sequence to form a complete digital electric signal that can reflect the temperature change of the entire support pillar surface, which will be used for time synchronization with the visible light image.

[0086] Step 503: comparing the acquisition time stamp of the visible light image with each signal acquisition time in the signal acquisition time set of the digital electric signal to screen out a target signal acquisition time with the same acquisition time stamp or the smallest time difference value of the visible light image, and determine a target digital electric signal segment corresponding to the acquisition time stamp of the visible light image.

[0087] In this step, the acquisition timestamp refers to the accurate time record (usually accurate to milliseconds) when the visible light image is taken, reflecting the acquisition time of the image data, stored in the metadata of the visible light image, used for comparison with the acquisition time of the digitized electrical signal to ensure the consistency of the temperature signal and the image data in time, which is the key reference benchmark for time synchronization. The signal acquisition time set refers to the collection of acquisition times corresponding to all discrete digital signals in the digitized electrical signal, reflecting the complete time sequence of the temperature signal, extracted based on the acquisition record of each discrete digital signal, containing multiple time points (each time point corresponds to the sampling time of a discrete digital signal), used for comparison with the acquisition timestamp of the visible light image to filter out the temperature signal segment that matches in time. The signal acquisition time refers to a single time point in the signal acquisition time set, i.e., the accurate time when a certain discrete digital signal is sampled, reflecting the time of the temperature state corresponding to the discrete digital signal, obtained based on the time record during analog-to-digital conversion, and the time difference value is calculated by comparing with the acquisition timestamp, which is the basic element for screening the target signal acquisition time. The time difference value refers to the numerical difference between the signal acquisition time and the acquisition timestamp of the visible light image, reflecting the degree of deviation in time between the temperature signal and the image data, a positive value indicating that the temperature signal is acquired later than the image, and a negative value indicating that it is earlier than the image, used for quantitative judgment of the time matching degree of the temperature signal and the image, which is the core indicator for screening the target signal acquisition time. The target signal acquisition time refers to the signal acquisition time selected from the signal acquisition time set that is the same as or has the smallest time difference value with the acquisition timestamp of the visible light image, reflecting the temperature signal sampling time that best matches the image data in time, determined based on the comparison result of the time difference value, used to intercept the segment of the digitized electrical signal that matches the image in time to ensure that both reflect the state of the support at the same time. The target digitized electrical signal segment refers to the part of the digitized electrical signal corresponding to the target signal acquisition time, reflecting the digitized distribution of the support surface temperature that matches the visible light image in time, obtained by intercepting from the digitized electrical signal based on the target signal acquisition time, containing the discrete digital signals of all acquisition units at that time, which is the basis for the subsequent spatial alignment of the temperature signal.

[0088] In the embodiments of the present application, in order to ensure the time consistency of the temperature signal and the image data, first, the collection timestamp of the visible light image (i.e. the accurate time of image shooting, accurate to milliseconds) is extracted from the metadata of the visible light image; at the same time, the collection time corresponding to each discrete digital signal is extracted from the collection record of the digitalized electric signal obtained in the second step, and is summarized to form a signal collection time set of the digitalized electric signal; then the collection timestamp of the visible light image and each signal collection time in the set are compared one by one, and the difference (i.e. the time difference, the calculation method is signal collection time minus image collection timestamp) between each signal collection time and the image collection timestamp is calculated; the signal collection time with the time difference of 0 (the time is completely the same) or the minimum absolute value (the time is closest) is selected from the set, and is determined as the target signal collection time; finally, the part of the digitalized electric signal corresponding to the target signal collection time is intercepted, and the target digitalized electric signal segment matched with the visible light image in time is obtained, so as to ensure that the two reflect the state of the support at the same time.

[0089] Step 504: identifying the surface physical marker of the contact net support to be detected, locating the pixel coordinates of the surface physical marker and the pixel range corresponding to each collection unit in the visible light image, determining the signal physical marker position of the collection unit where the surface physical marker is located in the target digitalized electric signal segment based on the target signal collection time, and the surface physical marker includes the support edge contour, inherent feature marker and interface joint.

[0090] In this step, the surface physical marker refers to the feature mark inherent to the surface of the overhead contact system support to be detected, which can be identified in the visible light image, including the support edge contour (the boundary line between the support and the background), the inherent feature mark (such as the protrusion, the notch, and the signboard in the manufacturing process), and the interface joint (the connecting joint of the support and the base, the crossbeam, and other components), reflecting the spatial positioning reference of the support surface, for establishing the spatial correspondence of the temperature signal and the image data. The pixel coordinates of the surface physical marker refer to the pixel position corresponding to the surface physical marker in the visible light image, taking the upper left corner of the image as the origin, expressed by the horizontal pixel number and the vertical pixel number (such as (x=150, y=300)), reflecting the accurate position of the physical marker in the image, determined based on the coordinate reading tool of the image coordinate system, serving as the reference point for adjusting the spatial alignment of the temperature signal and the image data. The pixel range corresponding to the acquisition unit refers to the pixel area covered by each acquisition unit in the visible light image, usually expressed by the minimum value and the maximum value of the pixel coordinates (such as 100-200 pixels horizontally and 300-400 pixels vertically), reflecting the spatial range of the acquisition unit in the image, calculated based on the corresponding relationship between the acquisition unit and the support surface and the image scale, for determining the spatial corresponding range of the temperature signal and the image data. The signal physical marker position refers to the position of the discrete digital signal corresponding to the acquisition unit where the surface physical marker is located in the target digitized electrical signal segment, reflecting the corresponding point of the physical marker in the temperature signal, determined based on the target signal acquisition time and the acquisition unit where the surface physical marker is located, for establishing the spatial correspondence with the image physical marker in the temperature signal, which is the key reference for spatial alignment. The inherent feature mark refers to the surface feature formed in the manufacturing process of the overhead contact system support to be detected, which remains unchanged for a long time, such as the production number notch, the reserved mounting hole, and the material identification protrusion, reflecting the inherent identity feature of the support, which is fixed in position and easy to identify, and is the most stable spatial positioning reference among the surface physical markers.

[0091] In the embodiments of the present application, in order to establish the spatial correspondence reference of the temperature signal and the image data, first, the visible light image is analyzed to identify the surface physical markers of the contact net support to be detected, including the support edge contour (the boundary line of the support and the background), the inherent characteristic markers (such as the protruding marks and the notches reserved during manufacturing), and the interface joint (the connecting joint of the support and the foundation or other components); then in the visible light image, the specific pixel coordinates (the horizontal and vertical pixel values with the upper left corner of the image as the origin) of the surface physical markers are determined by using a pixel coordinate positioning tool (such as the coordinate reading based on the image coordinate system); meanwhile, according to the correspondence relationship between the collection units divided in the first step and the surface of the support, the pixel range covered by each collection unit in the visible light image is marked (such as the area of 100-200 pixels horizontally and 300-400 pixels vertically in the image corresponding to a certain collection unit); then based on the target signal collection time, the position of the discrete digital signal corresponding to the collection unit where the surface physical marker is located in the target digitized electric signal segment is found, which is determined as the signal physical marker position, so as to establish the corresponding point of the image physical marker in the temperature signal.

[0092] Step 505: Taking the pixel coordinates of the surface physical markers as the reference, the arrangement order of each discrete digital signal in the target digitized electric signal segment is adjusted, so that the collection unit associated with each discrete digital signal is aligned with the corresponding pixel range, and the target digitized electric signal segment after position adaptation is obtained.

[0093] In this step, the discrete digital signal refers to the signal composed of a series of digital quantities obtained by analog-to-digital conversion of the temperature field signal (analog signal) of each collection unit, each digital quantity corresponding to the temperature value at a specific sampling time, reflecting the digital expression of the temperature of the collection unit. In this step, the arrangement order is adjusted to realize the spatial alignment with the image pixel range. The target digitized electric signal segment after position adaptation refers to the target digitized electric signal segment in which the collection unit associated with each discrete digital signal is completely aligned with the corresponding pixel range in the visible light image after the arrangement order is adjusted, reflecting the temperature signal matched with the image data in space. The arrangement order of the discrete digital signal is adjusted based on the pixel coordinates of the surface physical markers, which ensures that the temperature signal and the image data are one-to-one corresponding in space, and provides a spatially aligned temperature data source for subsequent data association storage.

[0094] In the embodiment of the present application, in order to realize the spatial alignment of the temperature signal and the image data, the pixel coordinates of the surface physical markers are taken as the spatial reference points (for example, the pixel coordinates of the top edge of the support are taken as the reference origin), the current arrangement order of each discrete digital signal in the target digitized electrical signal segment is compared with the arrangement order of the pixel range of the acquisition unit in the visible light image, if it is found that the pixel range corresponding to the acquisition unit associated with a certain discrete digital signal does not match the position of the signal in the segment (for example, the signal is arranged in the 5th position, but the corresponding pixel range is on the left side of the image, while it should actually be on the right side), the arrangement order of the signal is adjusted (for example, it is moved to the 8th position), the position of all discrete digital signals is adjusted one by one, so that the acquisition unit associated with each discrete digital signal completely coincides with the corresponding pixel range in the visible light image, and finally the position-adapted target digitized electrical signal segment is obtained, so as to ensure that the temperature signal and the image data correspond to each other in space.

[0095] Step 506: The signal data of the position-adapted target digitized electrical signal segment is associated and stored with the image data of the corresponding pixel range to form the integrated feature data.

[0096] In the embodiment of the present application, in order to form the comprehensive data set integrating the temperature and visual information, the signal data (the temperature value corresponding to each discrete digital signal and the acquisition unit information) of the position-adapted target digitized electrical signal segment is associated with the image data (for example, the color value, the gray value, the texture feature of the pixel range) of the corresponding pixel range in the visible light image, and the mapping relationship table of the acquisition unit identifier-pixel range coordinate-temperature signal data-image data is established to store the acquisition unit identifier corresponding to the same acquisition unit, the pixel range coordinate of the acquisition unit in the visible light image, the temperature signal data of the acquisition unit and the image data of the pixel range into the same record in the database; after the associated data of all acquisition units are stored, the data is summarized to finally form the integrated feature data containing the temperature information and the visual information which are time-synchronized (consistent with the time of collecting the visible light image) and spatially aligned (the temperature signal corresponding to the acquisition unit completely matches the image pixel range), which provides a complete and adaptive fusion data source for extracting the temperature information and the visual information and generating the defect detection result of the contact net support to be detected.

[0097] The embodiment of the present application solves the core defects of time asynchronization and spatial misalignment of multi-modal data, enables the temperature information and the visual information to correspond accurately, provides a high-quality integrated data source for extracting associated features and determining defect types from the fusion data, improves the positioning accuracy and determination accuracy of the contact net support defect detection, and meets the monitoring requirements of accurate positioning and clear type of defects in complex scenes.

[0098] The present application provides a specific embodiment, for example,Figure 2 As shown, step 106, temperature information and visual information are extracted from the integrated feature data, combined with the target material type of the contact net support to be detected, to generate a defect detection result of the contact net support to be detected, specifically including the following steps:

[0099] Step 601: Extract temperature information and visual information corresponding to each pixel range from the integrated feature data, wherein the temperature information includes temperature values corresponding to each pixel range, and the visual information includes line shape, color block distribution, and color depth change information.

[0100] In this step, the temperature information refers to temperature-related data corresponding to each pixel range extracted from the integrated feature data, including specific temperature values corresponding to each pixel range, reflecting the temperature state of the surface area of the contact net support to be detected corresponding to each pixel range, for subsequent comparison with the calibrated abnormality judgment reference to mark the temperature abnormal area, based on the extraction of the temperature signal-related storage content in the integrated feature data. The visual information refers to visual-related data corresponding to each pixel range extracted from the integrated feature data, including line shape, color block distribution, and color depth change information, reflecting the appearance features of the surface area of the contact net support to be detected corresponding to each pixel range, for comparison with the target defect-related parameters to determine the defect type, based on the visual attribute analysis of the image data in the integrated feature data. The line shape refers to the shape features related to lines in the visual information, reflecting the contour state of the linear defects (such as cracks) possibly existing on the surface of the contact net support to be detected, including the continuity or discontinuity, bending or straightness, etc. of the lines, for crack defect judgment of the concrete material support, based on the identification of the line contour in the visual information. The color block distribution refers to the distribution of different color regions in the pixel range in the visual information, reflecting the position and range of the planar defects (such as rust) possibly existing on the surface of the contact net support to be detected, including the distribution coordinates, coverage area, etc. of the color blocks, for rust defect judgment of the steel material support, based on the division and statistics of the color regions in the visual information. The color depth change information refers to the change of the depth gradient in the same color region in the visual information, reflecting the severity difference of the defects (such as different depths of rust) on the surface of the contact net support to be detected, including the transition range, gradient level, etc. of the color from light to dark, for detailed judgment of defects of various material supports, based on the analysis of the color gray value in the visual information.

[0101] Step 602: Extract the target defect-related parameters of the target material type of the contact net support to be detected from the material defect association data.

[0102] In this step, the target defect related parameter refers to the exclusive defect judgment parameter corresponding to the target material type of the contact net support to be detected, which is extracted from the material defect correlation data and reflects the judgment standard of common defects of the material support, such as crack related parameters of concrete material and corrosion related parameters of steel material, which is used to compare with the visual information of the temperature abnormal area to generate the defect judgment basis.

[0103] Step 603: Compare the temperature value of each pixel range in the temperature information with the calibrated abnormal judgment reference, and mark the pixel range with the temperature value exceeding the calibrated abnormal judgment reference as a temperature abnormal area.

[0104] In this step, the temperature abnormal area refers to the pixel range in the temperature information with the temperature value exceeding the calibrated abnormal judgment reference, which reflects the area of the surface temperature of the contact net support to be detected not conforming to the normal range, which may correspond to the potential defect position, and is used to focus the visual information of the area for defect judgment.

[0105] Step 604: Generate the defect detection result of the contact net support to be detected based on the visual information corresponding to the temperature abnormal area and the target defect related parameter.

[0106] Optionally, in step 604, the defect detection result of the contact net support to be detected is generated based on the visual information corresponding to the temperature abnormal area and the target defect related parameter, which specifically includes the following steps:

[0107] Step 611: If the target material type is concrete material and the target defect related parameter is crack related parameter, compare the line feature in the visual information corresponding to the temperature abnormal area with the crack length threshold, crack direction feature and crack fracture form in the crack related parameter to obtain a first comparison result, and the line feature includes line length, line direction, line fracture and connection relationship between lines.

[0108] In this step, the line feature refers to the specific attributes related to the line extracted from the visual information corresponding to the temperature anomaly area, including the actual length of the line (line length), the direction of the line extension (line direction), whether the line is disconnected and the number of disconnections (line breakage), and whether there is intersection or connection between different lines (line connection relationship), reflecting the morphological characteristics of possible crack defects in the temperature anomaly area, which are used for comparison with crack-related parameters of concrete materials and are extracted based on the detailed analysis of the line morphology in the visual information. The crack length threshold refers to the length standard in the crack-related parameters for determining whether the line constitutes a crack, reflecting the minimum length requirement of the concrete material pillar crack defect, which is used for comparison with the line length in the visual information of the temperature anomaly area, and is obtained based on the statistical data of concrete pillar crack failure experiments, which is part of the target defect-related parameters. The crack direction feature refers to the extension direction of the common crack of the concrete material pillar (such as horizontal, vertical, and diagonal) in the crack-related parameters, reflecting the typical morphological law of the crack, which is used for comparison with the line direction in the visual information of the temperature anomaly area, and is obtained based on a large number of concrete pillar crack cases, which is part of the target defect-related parameters. The crack breakage morphology refers to the common breakage state (such as continuous breakage and intermittent breakage) of the concrete material pillar crack in the crack-related parameters, reflecting the structural characteristics of the crack, which is used for comparison with the line breakage in the visual information of the temperature anomaly area, and is obtained based on the analysis of the concrete crack formation mechanism, which is part of the target defect-related parameters. The first comparison result refers to the matching degree result obtained by comparing the line features of the temperature anomaly area with the crack-related parameters when the target material type is concrete material, reflecting the conformity of the line features with the crack standards, which is used for subsequent statistical feature proportion to generate the crack determination result, and is obtained based on the item-by-item comparison of the line features with the crack-related parameters.

[0109] Step 612: Alternatively, if the target material type is steel material and the target defect-related parameters are corrosion-related parameters, the color block and color features in the visual information corresponding to the temperature anomaly area are compared with the corrosion area threshold, corrosion edge morphology, and corrosion color gradient range in the corrosion-related parameters, to obtain a second comparison result. The color block and color features include color block area, color block edge blur degree, and color depth gradient change characteristics.

[0110] In this step, the color block and color feature refers to the specific attributes related to color block and color extracted from the visual information corresponding to the temperature anomaly area, including the actual area of the color block (color block area), the clarity or blur degree of the color block edge (color block edge blur degree), and the gradient change of color depth within the same color block (color depth gradient change feature), reflecting the morphological features of possible corrosion defects in the temperature anomaly area, which are used for comparison with corrosion-related parameters of steel material quality and are extracted based on detailed analysis of color block distribution and color depth change information in visual information. The corrosion area threshold refers to the area standard in the corrosion-related parameters for determining whether the color block constitutes corrosion, reflecting the minimum area requirement of the steel material quality support corrosion defect, and is used for comparison with the color block area in the visual information of the temperature anomaly area. The corrosion area threshold is obtained based on the statistical data of steel corrosion impact strength experiments and belongs to part of the target defect-related parameters. The corrosion edge morphology refers to the edge state of the common corrosion of the steel material quality support (such as fuzzy, jagged), reflecting the typical appearance features of corrosion, and is used for comparison with the color block edge blur degree in the visual information of the temperature anomaly area. The corrosion edge morphology is obtained based on a large number of steel corrosion cases and belongs to part of the target defect-related parameters. The corrosion color gradient range refers to the color gradient interval (such as from light yellow to brown red) of the steel material quality support corrosion from light to deep, reflecting the color change corresponding to the severity of corrosion, and is used for comparison with the color depth gradient change feature in the visual information of the temperature anomaly area. The corrosion color gradient range is obtained based on the steel corrosion degree and color correlation experiment and belongs to part of the target defect-related parameters. The second comparison result refers to the matching degree result obtained by comparing the color block and color features of the temperature anomaly area with the corrosion-related parameters when the target material type is steel material quality, reflecting the conformity of the color block and color features with the corrosion standard, and is used for subsequent statistical feature proportion to generate the corrosion determination result.

[0111] Step 613: Alternatively, if the target material type is a supplementary material and the target defect-related parameters are supplementary material defect parameters, the supplementary material defect features in the visual information corresponding to the temperature anomaly area are compared with the defect shape standard, distribution density threshold, and attachment state feature in the supplementary material defect parameters, to obtain a third comparison result. The supplementary material defect features include defect feature shape, distribution density, and surface attachment state.

[0112] In this step, the supplementary material defect feature refers to specific attributes related to the supplementary material pillar defect extracted from the visual information corresponding to the temperature anomaly area, including the typical shape of the defect (defect feature shape), the number of defects per unit area (distribution density), and the attachment tightness of the defect on the pillar surface (surface attachment state), reflecting the morphological characteristics of the possible supplementary material exclusive defects in the temperature anomaly area, which are used for comparison with the supplementary material defect parameters. The supplementary material defect parameters refer to the target defect-related parameters corresponding to the supplementary material pillar, including the typical shape standard of the defect (defect shape standard), the critical value of the number of defects per unit area (distribution density threshold), and the normal attachment state requirement of the defect (attachment state feature), reflecting the judgment standard of the supplementary material pillar defect, which is used to compare the supplementary material defect features of the temperature anomaly area and is based on the experimental data of the supplementary material pillar defect. The third comparison result refers to the matching degree result obtained by comparing the supplementary material defect features of the temperature anomaly area with the supplementary material defect parameters when the target material type is the supplementary material, reflecting the conformity of the supplementary material defect features with the supplementary material defect standard, which is used for subsequent statistical feature proportion to generate the supplementary material defect judgment result, and is based on the item-by-item comparison of the supplementary material defect features and the supplementary material defect parameters.

[0113] Step 614: generating a defect detection result of the contact net pillar to be detected based on the first comparison result, the second comparison result, or the third comparison result.

[0114] Optionally, in step 614, the defect detection result of the contact net pillar to be detected is generated based on the first comparison result, the second comparison result, or the third comparison result, specifically including the following steps:

[0115] Step 621: if the target material type is a concrete material, then according to the first comparison result, the number of first features satisfying the line and crack matching conditions is counted, and a crack judgment result is generated by combining a preset crack matching proportion threshold.

[0116] In this step, the line and crack matching condition refers to a condition for determining whether the line feature of the temperature anomaly area meets the crack standard, including that the line length reaches the crack length threshold, the line direction is consistent with the crack direction feature, the line fracture morphology matches the crack fracture morphology, etc., which is used to count the number of features meeting the crack standard in the first comparison result, and is set based on the correspondence between the crack related parameters and the line feature. The first feature quantity ratio refers to, when the target material type is the concrete material, the proportion of the number of features meeting the line and crack matching condition in the total number of line features to be compared in the first comparison result, reflecting the overall matching degree of the line feature and the crack standard, which is used to generate the crack determination result in combination with the preset crack matching proportion threshold, and is calculated based on the division of the number of features meeting the condition and the total number of features. The preset crack matching proportion threshold refers to a proportion standard for determining whether the concrete material support has a crack defect, reflecting the minimum requirement for matching the line feature and the crack standard, which is used to compare the first feature quantity ratio to determine whether the crack exists, and is set based on the reliability requirement of the concrete support crack determination. The crack determination result refers to, when the target material type is the concrete material, a defect determination conclusion obtained based on the comparison between the first feature quantity ratio and the preset crack matching proportion threshold, including the existence of crack defects or no crack defects, and the position, severity, etc. of the crack, which is used as the defect detection result of the concrete material support.

[0117] Step 622: Alternatively, if the target material type is the steel material, the second feature quantity ratio meeting the color block and rust matching condition is counted according to the second comparison result, and the rust determination result is generated in combination with the preset rust matching proportion threshold.

[0118] In this step, the color block and rust matching condition refers to the condition for determining whether the color block and color feature of the temperature abnormal area meet the rust standard, including that the color block area reaches the rust area threshold, the color block edge blur degree is consistent with the rust edge shape, the color depth gradient change feature is within the rust color gradient range, etc., which is used to count the number of features meeting the rust standard in the second comparison result, and is set based on the correspondence between the rust related parameters and the color block and color feature. The second feature quantity ratio refers to the proportion of the number of features meeting the color block and rust matching condition in the second comparison result to the total number of color block and color feature to be compared when the target material type is steel material, which reflects the overall matching degree of the color block and color feature with the rust standard, and is used to generate a rust determination result in combination with a preset rust matching proportion threshold, which is calculated based on the number of features meeting the condition and the total number of features. The preset rust matching proportion threshold refers to the proportion standard for determining whether the steel material support has rust defects, which reflects the minimum requirement for matching the color block and color feature with the rust standard, and is used to compare the second feature quantity ratio to determine whether rust exists, which is set based on the reliability requirement of the steel support rust determination. The rust determination result refers to the defect determination conclusion obtained based on the comparison between the second feature quantity ratio and the preset rust matching proportion threshold when the target material type is steel material, including the existence of rust defects or no rust defects, and the position, severity, etc. of the rust, which is used as the defect detection result of the steel material support, and is generated based on the comparison between the second feature quantity ratio and the preset threshold.

[0119] Step 623: Alternatively, if the target material type is a supplementary material, a third feature quantity ratio meeting the supplementary material defect feature matching condition is counted according to the third comparison result, and a supplementary material defect determination result is generated in combination with a preset defect matching proportion threshold.

[0120] In this step, the supplementary material defect feature matching condition refers to a condition for determining whether the temperature anomaly area supplementary material defect feature meets the supplementary material defect standard, including defect feature shape consistent with defect shape standard, distribution density reaching distribution density threshold, surface attachment state matching with attachment state feature, etc., and is used to count the number of features meeting the supplementary material defect standard in the third comparison result, and is set based on the correspondence between the supplementary material defect parameter and the supplementary material defect feature. The third feature quantity proportion refers to the proportion of the number of features meeting the supplementary material defect feature matching condition in the third comparison result to the total number of supplementary material defect features that need to be compared when the target material type is the supplementary material, reflecting the overall matching degree of the supplementary material defect feature and the supplementary material defect standard, and is used to generate the supplementary material defect determination result in combination with the preset defect matching proportion threshold, and is calculated based on the division of the number of features meeting the condition and the total number of features. The preset defect matching proportion threshold refers to the proportion standard for determining whether the supplementary material support exists defects, reflecting the minimum requirement for matching the supplementary material defect feature and the supplementary material defect standard, and is used to compare the third feature quantity proportion to determine whether defects exist, and is set based on the reliability requirement of the supplementary material support defect determination. The supplementary material defect determination result refers to the defect determination conclusion obtained based on the comparison between the third feature quantity proportion and the preset defect matching proportion threshold when the target material type is the supplementary material, including the existence of the supplementary material defect or the absence of the supplementary material defect, and the position, severity, etc. of the defect, and is used as the defect detection result of the supplementary material support, and is generated based on the comparison analysis between the third feature quantity proportion and the preset threshold.

[0121] Step 624: taking the crack determination result, the corrosion determination result or the supplementary material defect determination result as the defect detection result of the overhead contact system support to be detected.

[0122] In the embodiments of the present application, first, the temperature information and visual information corresponding to each pixel range are extracted from the integrated feature data: the mapping relationship table of the acquisition unit identifier-pixel range coordinate-temperature signal data-image data is found in the integrated feature data, the corresponding acquisition unit identifier is matched in the table according to the coordinates of each pixel range (such as 100-200 pixels horizontally and 300-400 pixels vertically), and then the temperature signal data associated with the pixel range is located through the acquisition unit identifier, the specific temperature value (such as 28°C and 35°C) of the pixel range is read from the temperature signal data, all the temperature values of the pixel ranges are sorted according to the pixel range number to form the temperature information containing the specific temperature values of the pixel ranges, and for the image data of each pixel range, whether there is a line in the image is viewed, whether the line is continuous without breakpoints or there are multiple breaks (i.e. line pattern), different color regions in the image are identified, the positions of the color regions in the pixel range (such as the upper left region and the lower right region) and the pixel numbers (converted into actual area) covered by the color regions (i.e. color block distribution) are marked, the color change in the same color region is observed, and whether the transition from light to dark is obvious and whether there is a gradient level (i.e. color lightness change information) is recorded. The line pattern, color block distribution and color lightness change information are integrated according to the pixel range number to form the visual information. Second, the target defect related parameters corresponding to the target material type of the contact network support to be detected are screened: the target material type of the contact network support to be detected is first determined (such as concrete material, steel material or supplementary material), the storage file of the material defect correlation data is used as a reference: the file is stored in the classification mode of material type-defect parameter (such as recording crack related parameters under the concrete material item and recording corrosion related parameters under the steel material item), the target material type is used as a keyword for searching, the corresponding material item is found in the storage file, and all parameters related to defect determination are extracted from the item (such as crack length threshold, crack direction feature and crack breakage pattern when the target material is concrete; corrosion area threshold, corrosion edge pattern and corrosion color gradient range when the target material is steel). These extracted parameters are the target defect related parameters exclusive to the target material type. Subsequently, the temperature abnormal area is marked: first, the numbers of all pixel ranges in the temperature information and the corresponding temperature values are listed, and then the calibrated abnormal determination reference corresponding to the target material type of the contact network support to be detected is determined (including the upper and lower boundary values, such as 25°C-32°C for the concrete material reference). The temperature value of each pixel range is compared with the upper and lower boundary values of the calibrated abnormal determination reference in order according to the pixel range number: if the temperature value is greater than the upper boundary value or less than the lower boundary value, the number of the pixel range is marked with a special symbol (such as an asterisk) in the temperature information table, and all the pixel ranges marked are the temperature abnormal area.After that, the corresponding defect comparison mode is selected according to the target material type to obtain the corresponding comparison result: if the target material type is concrete material, the line feature and crack-related parameters are compared one by one for the visual information of each temperature abnormal area: the actual length of the line in the area is measured by an image measurement tool, compared with the crack length threshold (to see if it reaches or exceeds the threshold), the line extension direction (such as longitudinal, transverse, and diagonal) is observed, compared with the crack direction feature (common concrete cracks are mostly longitudinal), the number of line breaks and breakpoint distribution are recorded, compared with the crack breaking pattern (most concrete cracks are intermittent), whether there is intersection or connection between different lines is checked, compared with the standard of line connection relationship in crack-related parameters, the matching or non-matching result of each comparison is recorded to form the first comparison result; if the target material type is steel material, the color block and color feature and the rust-related parameters are compared one by one for the visual information of each temperature abnormal area: the pixel number covered by the color block in the area is counted and converted into the actual area, compared with the rust area threshold (to see if it reaches or exceeds the threshold), the pixel gray scale change of the color block edge (if the edge gray scale transitions from the color block color to the background color smoothly, it is fuzzy), compared with the rust edge shape (the rust edge of steel is mostly fuzzy), the color gradient change from light to dark in the color block is observed, compared with the rust color gradient range (such as light yellow-brown), the matching or non-matching result of each comparison is recorded to form the second comparison result; if the target material type is supplementary material, the defect features of the supplementary material and the supplementary material defect parameters are compared one by one for the visual information of each temperature abnormal area: the shape of the defect in the area (such as circular, strip-shaped) is observed, compared with the defect shape standard (to see if it is consistent), the number of defects in a unit area (i.e. distribution density) is counted, compared with the distribution density threshold (to see if it reaches or exceeds the threshold), whether the defect has signs of falling off on the surface of the support (i.e. surface adhesion state) is observed, compared with the adhesion state feature (most supplementary material defects are closely attached), the matching or non-matching result of each comparison is recorded to form the third comparison result; finally, the defect detection result of the contact net support to be detected is generated: if the target material type is concrete material, the total number of line features (such as length, direction, breaking condition, and connection relationship) that need to be compared in the first comparison result is counted, the number of matching feature items is counted, the first feature quantity ratio is obtained by dividing the number of matching feature items by the total number of features, the first feature quantity ratio is compared with the preset crack matching proportion threshold (such as 70%): if the first feature quantity ratio reaches or exceeds the preset crack matching proportion threshold, it is determined that the temperature abnormal area has a crack defect, the determination results of all temperature abnormal areas are integrated to generate a crack determination result containing the crack defect and the defect location (pixel range number);If the proportion does not reach the threshold, it is determined that there is no crack defect, and a crack defect determination result without crack defect is generated. If the target material type is steel material, the total number of color blocks and color feature items (such as area, edge blur degree, color gradient change, a total of 3 items) that need to be compared in the second comparison result is counted first, and then the number of matched feature items is counted. The number of matched feature items is divided by the total number of features to obtain the second feature quantity proportion that meets the color block and rust matching condition. The second feature quantity proportion is compared with the preset rust matching proportion threshold (such as 60%): if the second feature quantity proportion reaches or exceeds the preset rust matching proportion threshold, it is determined that the temperature abnormal area has a rust defect, and the determination results of all temperature abnormal areas are integrated to generate a rust determination result containing the rust defect and the defect location (pixel range number). If the proportion does not reach the threshold, it is determined that there is no rust defect, and a rust determination result without rust defect is generated. If the target material type is a supplementary material, the total number of supplementary material defect features (such as shape, distribution density, and attachment state, a total of 3 items) that need to be compared in the third comparison result is counted first, and then the number of matched feature items is counted. The number of matched feature items is divided by the total number of features to obtain the third feature quantity proportion that meets the supplementary material defect feature matching condition. The third feature quantity proportion is compared with the preset defect matching proportion threshold (such as 65%): if the proportion reaches or exceeds the threshold, it is determined that the temperature abnormal area has a supplementary material defect, and the determination results of all temperature abnormal areas are integrated to generate a supplementary material defect determination result containing the existence of the supplementary material defect and the defect location (pixel range number). If the proportion does not reach the threshold, it is determined that there is no supplementary material defect, and a supplementary material defect determination result without supplementary material defect is generated. Finally, the determination result (crack determination result, rust determination result, or supplementary material defect determination result) corresponding to the target material type is taken as the defect detection result of the contact net support to be detected.

[0123] The embodiments of the present application realize the material adaptability and determination accuracy of defect detection, avoid detection errors caused by parameter mismatch and single determination logic, improve the accuracy of contact net support defect type identification and the reliability of defect existence determination, can provide more actual demand defect information for rail transit operation and maintenance, help timely and accurately process support defects, and ensure the safe and stable operation of the contact net system.

[0124] Figure 3 A specific implementation structure diagram of a contact net support defect detection system provided by the embodiments of the present application is shown in Figure 3 The system can include:

[0125] The acquisition module 21 is configured to acquire material defect correlation data of the overhead line support and a visible light image of the overhead line support to be detected, and the material defect correlation data comprises crack-related parameters and a corresponding first thermal conductivity coefficient of a concrete support, corrosion-related parameters and a corresponding second thermal conductivity coefficient of a steel support, and a temperature anomaly threshold and a corresponding third thermal conductivity coefficient of a supplementary material support.

[0126] The calibration module 22 is configured to perform adaptive temperature calibration processing on a preset temperature anomaly determination reference according to a target material type and a corresponding target thermal conductivity coefficient of the overhead line support to be detected, in combination with the material defect correlation data, to obtain a calibrated anomaly determination reference.

[0127] The acquisition module 23 is configured to acquire a surface temperature field distribution of the overhead line support to be detected based on the calibrated anomaly determination reference.

[0128] The adjustment module 24 is configured to adjust a working frequency band of a signal filtering device to obtain an adjusted signal filtering device, so as to perform anti-interference filtering processing on an infrared signal corresponding to the temperature field distribution to obtain a filtered temperature field distribution.

[0129] The processing module 25 is configured to convert the filtered temperature field distribution into a digitalized electrical signal, and perform time stamp synchronization and alignment processing on the digitalized electrical signal and the visible light image to obtain integrated feature data.

[0130] The generation module 26 is configured to extract temperature information and visual information from the integrated feature data, and generate a defect detection result of the overhead line support to be detected in combination with the target material type of the overhead line support to be detected.

[0131] The overhead line support defect detection system according to the embodiment of the present application is used to implement the foregoing overhead line support defect detection method, and the specific embodiments of the overhead line support defect detection system can be found in the foregoing embodiment part of the overhead line support defect detection method, and the specific embodiments can be referred to the description of the corresponding embodiment part, which will not be described herein.

[0132] The present application further provides an electronic device, comprising a memory configured to store a computer program, and a processor configured to execute the computer program to implement the steps of any one of the overhead line support defect detection methods.

[0133] The present application further provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of any one of the overhead line support defect detection methods.

[0134] In an example embodiment, the computer readable storage medium described above can include, but is not limited to, a U disk, a read-only memory, a random access memory, a mobile hard disk, a magnetic disk or an optical disk, and various media that can store computer programs.

[0135] Embodiments of the present application also provide a computer program product, which comprises a computer program, and the computer program is executed by a processor to implement the steps in any of the above contact net support defect detection method embodiments.

[0136] The skilled person can further realize that the units and algorithm steps of the examples described in connection with the embodiments disclosed herein can be realized in electronic hardware, computer software or a combination of both. In order to clearly illustrate the interchangeability of hardware and software, the components and steps of the examples have been described in general terms in the above description. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. The skilled person can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0137] The above describes in detail a contact net support defect detection method and system provided by the present application. The principles and implementation manners of the present application are described herein by applying specific examples, and the above example descriptions are only used to help understand the method of the present application and its core idea. It should be pointed out that, for the ordinary skilled person in the technical field, some improvements and modifications can be made to the present application without departing from the principles of the present application, and these improvements and modifications also fall within the protection scope of the present application.

Claims

1. A method for detecting defects in a catenary support post, characterized in that, The method comprises the following steps: Obtain material defect correlation data of the overhead line support and a visible light image of the overhead line support to be detected, the material defect correlation data including crack-related parameters and a corresponding first thermal conductivity coefficient of a concrete support, corrosion-related parameters and a corresponding second thermal conductivity coefficient of a steel support, and a temperature anomaly threshold value and a corresponding third thermal conductivity coefficient of a supplementary material support; According to the target material type and the corresponding target thermal conductivity coefficient of the overhead line support to be detected, and in combination with the material defect correlation data, perform adaptive temperature calibration processing on a preset temperature anomaly judgment reference to obtain a calibrated anomaly judgment reference; Based on the calibrated anomaly judgment reference, collect the surface temperature field distribution of the overhead line support to be detected; Adjust the working frequency band of the signal filtering device to obtain an adjusted signal filtering device for anti-interference filtering processing of the infrared signal corresponding to the temperature field distribution, and obtain a filtered temperature field distribution; Convert the filtered temperature field distribution into a digital electrical signal, and perform timestamp synchronization and alignment processing on the digital electrical signal and the visible light image to obtain integrated feature data; Extract temperature information and visual information from the integrated feature data, and generate a defect detection result of the overhead line support to be detected in combination with the target material type of the overhead line support to be detected; Convert the filtered temperature field distribution into a digital electrical signal, and perform timestamp synchronization and alignment processing on the digital electrical signal and the visible light image to obtain integrated feature data, comprising: According to the surface area division rule of the overhead line support to be detected, divide the surface of the overhead line support to be detected into a plurality of collection units; Convert the temperature field signal corresponding to each collection unit in the filtered temperature field distribution into a discrete digital signal, and combine all the discrete digital signals into a digital electrical signal; Compare the collection timestamp of the visible light image with each signal collection time in the signal collection time set of the digital electrical signal to filter out a target signal collection time with the same collection timestamp or the smallest time difference value of the visible light image, and determine a target digital electrical signal segment corresponding to the collection timestamp of the visible light image; Recognize the surface physical marker of the overhead line support to be detected, locate the pixel coordinates of the surface physical marker and the pixel range corresponding to each collection unit in the visible light image, and based on the target signal collection time, determine the signal physical marker position of the collection unit where the surface physical marker is located in the target digital electrical signal segment, the surface physical marker including a support edge contour, an inherent feature marker and an interface joint; Take the pixel coordinates of the surface physical marker as a reference to adjust the arrangement order of each discrete digital signal in the target digital electrical signal segment, so that the collection unit associated with each discrete digital signal is aligned with the corresponding pixel range, and a position-adapted target digital electrical signal segment is obtained; Correlate and store the signal data of the position-adapted target digital electrical signal segment and the image data of the corresponding pixel range to form integrated feature data.

2. The method for detecting defects in a catenary post according to claim 1, characterized in that, According to the target material type and the corresponding target thermal conductivity coefficient of the contact net support to be detected, the preset temperature anomaly judgment reference is adaptively temperature calibrated by combining the material defect correlation data, and a calibrated anomaly judgment reference is obtained, including: According to the visible light image, the target material type of the contact net support to be detected is determined, and the target material type is a concrete material, a steel material or a supplementary material; According to the target material type, the target thermal conductivity coefficient and the target temperature anomaly threshold corresponding to the target material type are extracted from the material defect correlation data; The difference value of the target thermal conductivity coefficient and the reference thermal conductivity coefficient is calculated, the boundary value of the preset general temperature anomaly judgment reference is adjusted by combining the target temperature anomaly threshold, and the calibrated anomaly judgment reference matched with the target material type is obtained.

3. The method for detecting defects in a catenary support pillar according to claim 1, characterized in that, Adjust the working frequency band of the signal filtering device to obtain an adjusted signal filtering device for anti-interference filtering processing of the infrared signal corresponding to the temperature field distribution, and obtain a filtered temperature field distribution, including: Based on the target material type and the corresponding target thermal conductivity coefficient of the contact net support to be detected, the target filtering frequency band range corresponding to the target material type is extracted from the material defect correlation data, wherein the first filtering frequency band range corresponds to the concrete material, the second filtering frequency band range corresponds to the steel material, and the third filtering frequency band range corresponds to the supplementary material; According to the target thermal conductivity coefficient of the contact net support to be detected, the target working frequency band in the target filtering frequency band range is selected from the working frequency band of the signal filtering device, and the working frequency band is the filtering frequency band of the signal filtering device preset when the signal filtering device is shipped and suitable for general material contact net support; The working frequency band of the signal filtering device is replaced by the target working frequency band to obtain an adjusted signal filtering device; The interference signal in the infrared signal corresponding to the surface temperature field distribution is filtered through the adjusted signal filtering device to obtain a filtered temperature field distribution.

4. The method of claim 1, wherein From the integrated feature data, temperature information and visual information are extracted, and a defect detection result of the contact net support to be detected is generated by combining the target material type of the contact net support to be detected, including: From the integrated feature data, the temperature information and the visual information corresponding to each pixel range are extracted, the temperature information includes the temperature value corresponding to each pixel range, and the visual information includes line shape, color block distribution and color depth change information; From the material defect correlation data, the target defect related parameters of the target material type of the contact net support to be detected are extracted; The temperature value of each pixel range in the temperature information is compared with the calibrated anomaly judgment reference, and the pixel range with a temperature value exceeding the calibrated anomaly judgment reference is marked as a temperature anomaly area; Based on the visual information corresponding to the temperature anomaly area and the target defect related parameters, a defect detection result of the contact net support to be detected is generated.

5. The method for detecting defects in a catenary post according to claim 4, characterized in that, Based on the visual information corresponding to the temperature anomaly area and the target defect related parameters, a defect detection result of the contact net support to be detected is generated, including: If the target material type is a concrete material and the target defect-related parameter is a crack-related parameter, line features in visual information corresponding to the temperature anomaly region are compared with a crack length threshold, a crack direction feature, and a crack fracture morphology in the crack-related parameter, to obtain a first comparison result. The line features include line length, line direction, line fracture, and connection between lines. Alternatively, if the target material type is a steel material and the target defect-related parameter is a corrosion-related parameter, color blocks and color features in the visual information corresponding to the temperature anomaly region are compared with a corrosion area threshold, a corrosion edge morphology, and a corrosion color gradient range in the corrosion-related parameter, to obtain a second comparison result. The color blocks and color features include color block area, color block edge blur degree, and color depth gradient change feature. Alternatively, if the target material type is a supplementary material and the target defect-related parameter is a supplementary material defect parameter, supplementary material defect features in the visual information corresponding to the temperature anomaly region are compared with a defect shape standard, a distribution density threshold, and an attachment state feature in the supplementary material defect parameter, to obtain a third comparison result. The supplementary material defect features include defect feature shape, distribution density, and surface attachment state. Based on the first comparison result, the second comparison result, or the third comparison result, a defect detection result of the contact net support to be detected is generated.

6. The method for detecting defects in a catenary support pillar according to claim 5, characterized in that, Based on the first comparison result, the second comparison result, or the third comparison result, a defect detection result of the contact net support to be detected is generated, including: If the target material type is a concrete material, a first feature quantity ratio that satisfies line and crack matching conditions is counted according to the first comparison result, and a crack determination result is generated in combination with a preset crack matching proportion threshold. Alternatively, if the target material type is a steel material, a second feature quantity ratio that satisfies color block and corrosion matching conditions is counted according to the second comparison result, and a corrosion determination result is generated in combination with a preset corrosion matching proportion threshold. Alternatively, if the target material type is a supplementary material, a third feature quantity ratio that satisfies supplementary material defect feature matching conditions is counted according to the third comparison result, and a supplementary material defect determination result is generated in combination with a preset defect matching proportion threshold. The crack determination result, the corrosion determination result, or the supplementary material defect determination result is taken as the defect detection result of the contact net support to be detected.

7. A catenary post defect detection system, characterized in that, including: The acquisition module is configured to acquire material defect association data of a contact net support and a visible light image of a contact net support to be detected. The material defect association data includes crack-related parameters corresponding to a concrete support and a corresponding first thermal conductivity, corrosion-related parameters corresponding to a steel support and a corresponding second thermal conductivity, and a temperature anomaly threshold and a corresponding third thermal conductivity of a supplementary material support. The calibration module is configured to perform adaptive temperature calibration processing on a preset temperature anomaly judgment reference according to a target material type of the contact net support to be detected and a corresponding target thermal conductivity, and in combination with the material defect correlation data, to obtain a calibrated anomaly judgment reference; The acquisition module is configured to acquire a surface temperature field distribution of the contact net support to be detected based on the calibrated anomaly judgment reference; The adjustment module is configured to adjust a working frequency band of the signal filtering device to obtain an adjusted signal filtering device, to perform anti-interference filtering processing on an infrared signal corresponding to the temperature field distribution, and to obtain a filtered temperature field distribution; The processing module is configured to convert the filtered temperature field distribution into a digitized electrical signal, to perform timestamp synchronization and alignment processing on the digitized electrical signal and the visible light image, and to obtain integrated feature data; The generation module is configured to extract temperature information and visual information from the integrated feature data, to generate a defect detection result of the contact net support to be detected in combination with the target material type of the contact net support to be detected. The filtered temperature field distribution is converted into a digitized electrical signal, and timestamp synchronization and alignment processing is performed on the digitized electrical signal and the visible light image to obtain integrated feature data, including: According to a surface area division rule of the contact net support to be detected, the surface of the contact net support to be detected is divided into a plurality of acquisition units; Each acquisition unit corresponding temperature field signal in the filtered temperature field distribution is converted into a discrete digital signal, and all discrete digital signals are combined into a digitized electrical signal; The acquisition timestamp of the visible light image is compared with each signal acquisition time in a signal acquisition time set of the digitized electrical signal to filter out a target signal acquisition time that is the same as or has a minimum time difference with the acquisition timestamp of the visible light image, and to determine a target digitized electrical signal segment corresponding to the acquisition timestamp of the visible light image; A surface physical marker of the contact net support to be detected is recognized, pixel coordinates and a pixel range corresponding to each acquisition unit of the surface physical marker are located in the visible light image, based on the target signal acquisition time, a signal physical marker position of an acquisition unit where the surface physical marker is located is determined in the target digitized electrical signal segment, and the surface physical marker includes a support edge contour, an inherent feature marker, and an interface joint; The pixel coordinates of the surface physical marker are taken as a reference to adjust an arrangement order of each discrete digital signal in the target digitized electrical signal segment, so that each acquisition unit associated with the discrete digital signal is aligned with the corresponding pixel range, to obtain a position-adapted target digitized electrical signal segment; Signal data of the position-adapted target digitized electrical signal segment and image data of the corresponding pixel range are associated and stored to form integrated feature data.

8. An electronic device, comprising: including: a memory configured to store a computer program; a processor configured to execute the computer program to implement the steps of the contact net support defect detection method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the method for detecting the defect of the catenary support column according to any one of claims 1 to 6.

Images