Cable joint overheating diagnosis method and system based on infrared thermal imaging

Abstract

The invention discloses a cable joint overheating diagnosis method and system based on infrared thermal imaging. The method comprises the following steps: acquiring a plurality of infrared thermal images of a cable joint to be detected through a plurality of visual camera devices; fusing the infrared thermal images based on the position and direction information of the visual camera equipment; establishing an infrared temperature identification model based on a neural network, performing training according to a pre-collected infrared image sample and a temperature result label corresponding to the sample, and inputting the fused infrared image into the infrared temperature identification model to obtain a first temperature result; constructing a three-dimensional temperature field based on the fused infrared image to perform temperature analysis on the to-be-measured cable joint to obtain a second temperature result; and fusing the two temperature results through a fusion mode to obtain a fused temperature result, and performing cable joint overheating diagnosis based on the fused temperature result to obtain a diagnosis result. The problems that an existing cable connector is insufficient in overheating diagnosis accuracy and low in detection efficiency are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power equipment detection, and particularly relates to a cable joint overheating diagnosis method and system based on infrared thermal imaging. Background Art

[0002] In a power cable system, a cable joint is a key component to ensure stable power transmission. Therefore, the cable is connected to electrical equipment based on the cable connection end. However, due to long-term current load, environmental factors, and its own material aging, the cable joint is prone to overheating. Overheating not only accelerates the damage of the cable joint, but may also cause safety accidents such as fires in severe cases, threatening the safe operation of the power system. Thus, the cable connection end has become an important influencing factor for the safe operation of power cables. How to detect the above diseases in the cable connection end has become a technical problem to be solved urgently.

[0003] Currently, traditional cable joint overheating detection methods generally judge whether the temperature of the cable connection end is too high based on manual inspection. However, the manual detection method requires high experience for the inspection personnel, and the detection accuracy is low for new employees. Moreover, due to manual detection, the detection efficiency is also relatively low.

[0004] Chinese Patent with publication number CN113269748A discloses an infrared and visible light image fusion cable joint fault warning system and method. The system includes an infrared thermal imager, a network camera, an image collector, and an image processing computer. The invention first uses the infrared thermal imager and the network camera to collect real-time infrared and visible light image information of the cable joint in the cable well, and then transmits the displayed infrared and visible light image information to the image processing computer through the image collector. The image processing computer first performs filtering, defogging, and enhancement processing on the obtained infrared and visible light images, then generates a new fused image by using an image fusion method based on sparse representation for the images after image processing. Finally, the image processing computer judges the current overheating situation of the cable joint through the fused image, thereby realizing the function of fault warning for the cable joint. The invention mainly relies on detecting whether the highest temperature exceeds the threshold according to the relationship between the color and temperature of the fused image, without considering the influence of different environmental conditions (such as high humidity, strong wind) on temperature detection, resulting in insufficient temperature diagnosis accuracy and scene adaptability. Summary of the Invention

[0005] The purpose of the present invention is to provide a cable joint overheating diagnosis method and system based on infrared thermal imaging to solve the problems of insufficient accuracy and low detection efficiency in the existing cable joint overheating diagnosis.

[0006] The technical solution of the present invention is as follows: On the one hand, the present invention provides a method for diagnosing overheating of cable joints based on infrared thermal imaging, comprising the following steps: Collect a plurality of infrared thermal images of the cable joint to be measured through a plurality of visual camera devices.

[0007] Fuse the infrared thermal images based on the position information and direction information of the plurality of visual camera devices to obtain a fused infrared image.

[0008] Establish an infrared temperature recognition model based on a neural network, train the infrared temperature recognition model according to the pre-collected infrared image samples and the temperature result labels corresponding to the samples, and input the fused infrared image into the infrared temperature recognition model to obtain a first temperature result.

[0009] Construct a three-dimensional temperature field based on the fused infrared image, and perform temperature analysis on the cable joint to be measured based on the three-dimensional temperature field to obtain a second temperature result.

[0010] Fuse the first temperature result and the second temperature result through a fusion mode to obtain a fused temperature result, and perform cable joint overheating diagnosis on the cable joint to be measured based on the fused temperature result to obtain a diagnosis result.

[0011] Preferably, fusing the infrared thermal images based on the position information and direction information of the plurality of visual camera devices to obtain a fused infrared image specifically includes: Perform non-subsampled contourlet transform on the infrared thermal images to obtain a plurality of sub-images.

[0012] Based on the position information and direction information of the plurality of visual camera devices, construct a spatio-temporal correlation matrix for each sub-image, and determine the spatio-temporal correlation weight of each sub-image according to the spatio-temporal correlation matrix.

[0013] Perform weighted fusion on each sub-image and its corresponding spatio-temporal correlation weight to obtain a fused sub-image.

[0014] Perform inverse non-subsampled contourlet transform on the fused sub-image to obtain a fused infrared image.

[0015] Preferably, the calculation formula of the spatio-temporal correlation weight is specifically:[[]]

[0016]

[0017] In the formula, represents the spatio-temporal correlation weight between the th visual camera device and the th visual camera device; 、 both represent adjustment coefficients; Represents the total number of visual camera devices, ; 、 respectively represent the position vectors of the th visual camera device and the th visual camera device; 、 、 respectively represent the direction vectors of the th visual camera device; 、 represent the combined vector form of the direction vectors of the th visual camera device and the th visual camera device; represents the Euclidean distance between the position vectors of the th visual camera device and the th visual camera device; represents the distance metric between the direction vectors of the th visual camera device and the th visual camera device.

[0018] Preferably, constructing a three-dimensional temperature field based on the fused infrared image is specifically as follows: Align the feature points of the fused infrared image with the three-dimensional model of the cable joint to obtain an initial three-dimensional model, and segment the initial three-dimensional model to obtain a plurality of grid cells divided by unit grid lines.

[0019] Perform ray tracing on each pixel point of the fused infrared image, and assign the corresponding temperature when the ray emitted from the pixel point intersects the unit grid line to the intersecting grid cell to obtain a preliminary three-dimensional temperature distribution field; Determine the heat conduction characteristics of each material part based on the material composition of the cable joint to be measured.

[0020] Iteratively optimize the preliminary three-dimensional temperature distribution field based on the heat conduction characteristics of each material part, preset boundary conditions, and preset iteration termination conditions to obtain a three-dimensional temperature field.

[0021] Preferably, the calculation formula for iteratively optimizing the preliminary three-dimensional temperature distribution field is specifically as follows:

[0022] In the formula, represents the temperature value of the grid cell with coordinates at time in the three-dimensional temperature field; represents the temperature prediction value of the grid cell with coordinates at time after one iteration; represents the time step; represents the density of the material corresponding to the grid cell with coordinates ; represents the specific heat capacity of the material corresponding to the grid cell with coordinates ; represents the set of all grid cells adjacent to the grid cell with coordinates ; represents the thermal conductivity between the adjacent grid cell and the current grid cell ; represents the temperature value of the adjacent grid cell at the moment , represents the distance between the current grid cell and the adjacent grid cell ; represents the heat source intensity inside the grid cell with coordinates at the moment .

[0023] Preferably, the fusion mode includes a first mode, and the first mode characterizes that the current environment is a high-humidity environment; the first temperature result and the second temperature result are fused through the first mode to obtain the fused temperature result, specifically: The first correction factor and the second correction factor of the humidity for the first temperature result and the second temperature result are respectively obtained.

[0024] The first temperature result and the second temperature result are corrected respectively based on the first correction factor and the second correction factor to obtain the corrected first temperature result and the corrected second temperature result.

[0025] The first fusion weight and the second fusion weight of the corrected first temperature result and the corrected second temperature result are respectively determined based on a preset humidity function.

[0026] Based on the first fusion weight and the second fusion weight, the corrected first temperature result and the corrected second temperature result are weighted and fused to obtain the fused temperature result.

[0027] Preferably, the fusion mode includes a second mode, and the second mode characterizes that the current environment is a strong-wind or large-temperature-difference environment; the first temperature result and the second temperature result are fused through the second mode to obtain the fused temperature result, specifically: Based on Newton's law of cooling, a physical model of the influence of strong wind on the heat dissipation of the cable joint to be measured is constructed.

[0028] Based on the physical model, the first compensation value of the first temperature result and the second compensation value of the second temperature result are determined.

[0029] Based on the first compensation value and the second compensation value, compensate the first temperature result and the second temperature result respectively to obtain the compensated first temperature result and the second temperature result.

[0030] Based on the preset temperature difference weight formula, determine the first dynamic weight and the second dynamic weight of the compensated first temperature result and the second temperature result respectively.

[0031] Based on the first dynamic weight and the second dynamic weight, perform weighted fusion on the compensated first temperature result and the second temperature result to obtain the fused temperature result.

[0032] Preferably, the preset temperature difference weight formula is expressed as:

[0033] In the formula, represents the temperature difference weight function; represents the temperature difference; represents the adjustment coefficient; represents the reference temperature difference; when the temperature difference is large, tends to 1, focusing on the second temperature result; when the temperature difference is small, tends to 0, focusing on the first temperature result.

[0034] Performing weighted fusion on the compensated first temperature result and the second temperature result based on the first dynamic weight and the second dynamic weight is expressed as:

[0035] In the formula, is the fused temperature result; is the second temperature result; is the first temperature result.

[0036] On the other hand, the present invention provides an overheating diagnosis system for cable joints based on infrared thermal imaging, including a collection unit, a fusion unit, a first processing unit, a second processing unit, and a diagnosis unit.

[0037] The collection unit is used to collect a plurality of infrared thermal images of the cable joint to be measured through a plurality of visual camera devices.

[0038] The fusion unit is used to fuse the infrared thermal images based on the position information and the direction information of the plurality of visual camera devices to obtain a fused infrared image.

[0039] The first processing unit is used to establish an infrared temperature recognition model based on a neural network, train the infrared temperature recognition model according to pre-collected infrared image samples and the corresponding temperature result labels of the samples, input the fused infrared image into the infrared temperature recognition model, and obtain a first temperature result.

[0040] The second processing unit is used to construct a three-dimensional temperature field based on the fused infrared image, and perform temperature analysis on the cable joint to be measured based on the three-dimensional temperature field, so as to obtain a second temperature result.

[0041] The diagnosis unit is used to fuse the first temperature result and the second temperature result through a fusion mode to obtain a fused temperature result, and perform overheating diagnosis on the cable joint to be measured based on the fused temperature result to obtain a diagnosis result.

[0042] On the other hand, the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the method for diagnosing overheating of cable joints based on infrared thermal imaging according to any embodiment of the present invention.

[0043] Compared with the prior art, the present invention has the following technical effects: 1. By fusing the position information and direction information of multiple infrared thermal images and multiple visual camera devices, the present invention obtains a fused infrared image, which can reduce the perspective deviation and information loss between images, so as to obtain a more comprehensive and accurate infrared thermal image, which is helpful to more accurately reflect the actual temperature distribution of the cable joint.

[0044] 2. Through the fusion mode, the present invention fuses the first temperature result and the second temperature result to obtain the final fused temperature result, which can further improve the accuracy of the temperature result, complement the advantages of the two types of temperature results, make the diagnosis result more reliable, and avoid misdiagnosis or missed diagnosis that may occur due to a single temperature analysis method.

[0045] 3. The method of the present invention avoids relying on the personal experience of the inspection personnel, reduces the detection errors caused by insufficient personnel experience, can ensure the accuracy, reliability and stability of the detection results, and improves the detection efficiency at the same time. Description of the Drawings

[0046] Figure 1 is the overall flowchart of the method for diagnosing overheating of cable joints based on infrared thermal imaging according to the present invention; Figure 2 is the structural schematic diagram of the system for diagnosing overheating of cable joints based on infrared thermal imaging according to the present invention. Detailed Embodiments

[0047] To make the objectives, technical solutions, and advantages of the present invention clearer, the following will describe the technical solutions of the present invention clearly and completely in conjunction with specific embodiments of the present application and with reference to the accompanying drawings.

[0048] Embodiment 1 This embodiment provides a method for diagnosing overheating of cable joints based on infrared thermal imaging, which completes the judgment of overheating of cable joints and realizes the automatic diagnosis of overheating of cable joints. The method in this embodiment does not rely on the personal experience of inspection personnel, avoiding detection errors caused by insufficient personnel experience, making the detection results more accurate, reliable, and stable, and at the same time improving the detection efficiency. Refer to Figure 1 As shown in the figure, it includes the following steps: Collect multiple infrared thermal images of the cable joint to be measured through multiple vision camera devices.

[0049] Specifically, during the process of diagnosing overheating of the cable joint to be measured, by arranging multiple vision camera devices, the thermal information of the cable joint is comprehensively obtained: multiple perspective camera devices are distributed at different positions of the cable joint to be measured, realizing shooting of the cable joint from multiple angles. Each vision camera device can capture the infrared radiation situation of the cable joint under a corresponding specific perspective, thereby generating a corresponding infrared thermal image. In this embodiment, there are no specific restrictions on the installation position and quantity of the vision camera devices, which can be set according to actual needs. In one embodiment, for example, in a substation, there is a cable joint, and infrared camera devices (i.e., vision camera devices) are installed above, on the left, and on the right of it respectively. When the three devices work simultaneously, three infrared thermal images from different perspectives are collected respectively. That is, multi-perspective acquisition is realized to obtain more comprehensive thermal information of the cable joint, avoiding information loss caused by a single perspective.

[0050] Fuse the infrared thermal images based on the position information and direction information of multiple vision camera devices to obtain a fused infrared image.

[0051] Specifically, since each vision camera device has its specific position coordinates and direction information, using this information, through a certain image fusion algorithm, multiple infrared thermal images are fused. The fusion process mainly integrates the information about the cable joint in different images, eliminates the differences caused by different perspectives, and forms a fused infrared image containing more complete information. For example, there are two vision camera devices. The first vision camera device is closer to the cable joint, and the second device is farther away. For a certain pixel point in the fused image, the fused pixel value is calculated according to the fusion algorithm. The advantage of this image fusion is that it can improve the quality and information content of the image, enhance the characterization ability of the overall thermal state of the cable joint, and reduce temperature misjudgment caused by perspective differences.

[0052] As a preferred implementation manner of this embodiment, the infrared thermal images are fused based on the position information and orientation information of multiple visual camera devices to obtain the fused infrared image, specifically as follows: Perform a non-subsampled contourlet transform on the infrared thermal images to obtain multiple sub-images.

[0053] Specifically, the non-subsampled contourlet transform (NSCT) can effectively capture detailed information such as edges and textures of an image, and sparsely represent the image in multiple scales and multiple directions. Therefore, each infrared thermal image is decomposed into a low-frequency sub-image (i.e., low-frequency sub-band) and high-frequency sub-images (i.e., high-frequency sub-bands) through the non-subsampled pyramid (NSP) of the non-subsampled contourlet transform (NSCT). The low-frequency sub-band contains the main contours and general profile information of the image, and the high-frequency sub-bands contain the detailed information of the image. Then, the high-frequency sub-bands are decomposed in different directions using the non-subsampled directional filter bank (NSDFB) at each scale to obtain sub-images in different directions, and thus multiple sub-images with different features are obtained, which can analyze the local features of the image more meticulously.

[0054] Based on the position information and orientation information of multiple visual camera devices, construct a spatio-temporal correlation matrix for each sub-image, and determine the spatio-temporal correlation weight of each sub-image according to the spatio-temporal correlation matrix.

[0055] Specifically, the spatio-temporal correlation matrix is used to represent the spatio-temporal correlation relationship between the sub-images of multiple visual camera devices. Among them, the position information of the visual camera device is represented by a position vector, and the orientation information is represented by an orientation vector. If the total number of visual camera devices is , for the sub-image of the th visual camera device, its position vector is , and the orientation vector is represented by the Euler angles , then the spatio-temporal correlation matrix is a matrix, and the element in the matrix represents the spatio-temporal correlation weight between the th visual camera device and the th visual camera device, which is used for subsequent weighted fusion operations on the sub-images to determine the contribution degree of the sub-images of different devices in the fusion process.

[0056] As a preferred implementation manner of this embodiment, the calculation formula of the spatio-temporal correlation weight is specifically as follows:

[0057]

[0058] In the formula, represents the The spatio-temporal correlation weight between the th visual camera device and the - Both represent adjustment coefficients, which are used to balance the influence degrees of position difference and direction difference on weight calculation; represents the total number of visual camera devices, ; - respectively represent the th visual camera device and the th visual camera device's position vectors; - - respectively represent the th visual camera device's direction vectors; - represent the combined vector form of the th visual camera device's and the th visual camera device's direction vectors; represents the Euclidean distance between the th visual camera device's and the th visual camera device's position vectors, which reflects the difference in spatial positions of the devices. The closer the distance, the higher the possible correlation of the images captured by the two camera devices in space. Therefore, a higher weight should be given in weight calculation; represents the distance metric between the th visual camera device's and the th visual camera device's direction vectors. For example, by converting Euler angles into a rotation matrix and then calculating a certain distance between matrices, it is used to measure the difference degree of the shooting directions of the two visual camera devices. The closer the directions, the higher the possible correlation of the images captured by the two camera devices in terms of viewing angle, and the higher the weight should be.

[0059] Furthermore, the denominator part in the above formula normalizes the correlation weights of all devices, making ; ensuring the reasonable distribution of information during the fusion process. And ensuring that each sub-image of each camera device can contribute corresponding information according to its association degree with other devices during the fusion process, so as to avoid the situation that the information of a certain device is overemphasized or underemphasized.

[0060] Perform weighted fusion on each sub-image and its corresponding spatio-temporal correlation weight to obtain the fused sub-image.

[0061] Specifically, for the sub-images of each scale and direction, the fused sub-image is expressed as:

[0062] Wherein, represents the fused sub-image, which is the result obtained by fusing the corresponding sub-images captured by multiple visual camera devices, represents the scale index, represents the direction index. The fused sub-image synthesizes the information of different visual camera devices and has better characteristics at a specific scale and direction , such as clearer edges, more accurate temperature information, etc.; also represents the total number of camera devices; represents the th weight corresponding to the visual camera device, which is obtained through the spatio-temporal correlation weight calculation method and reflects the importance of the th visual camera device in the current fusion process. The size of the weight depends on the spatio-temporal correlation degree of the device with other devices, including factors such as position and direction; the larger the weight, the greater the proportion of the sub-band image captured by the device in the fused sub-band image and the greater the impact on the fusion result; represents the sub-image of the image captured by the th visual camera device at the current scale and direction . Each visual camera device has its corresponding sub-image.

[0063] Perform an inverse non-subsampled contourlet transform on the fused sub-image to obtain the fused infrared image.

[0064] Specifically, the NSCT inverse transform is the inverse process of the NSCT transform. By processing the fused sub-image in the reverse order of decomposition, first merge the sub-band information in different directions through the inverse transform of the directional filter bank, and then merge the information at different scales through the inverse transform of the non-subsampled pyramid, and finally obtain the fused infrared image. The reconstruction process can restore the multi-scale and multi-direction sub-band information after fusion processing into a complete and high-quality infrared image. The fused infrared image synthesizes the information of multiple visual camera devices and has richer details and a more accurate representation of temperature distribution.

[0065] Establish an infrared temperature recognition model based on a neural network, train the infrared temperature recognition model according to the pre-collected infrared image samples and the temperature result labels corresponding to the samples, and input the fused infrared image into the infrared temperature recognition model to obtain the first temperature result.

[0066] Specifically, during the training process of the infrared temperature recognition model, first, a neural network model based on deep learning is constructed. The choice of the neural network model is not restricted here and can be selected according to actual accuracy requirements, computing resources, or real-time requirements. For example, a convolutional neural network (CNN) can be selected. The model is trained with a large number of infrared image samples and the corresponding temperature result labels until the model converges or reaches the preset goal and then stops. After the training is completed, the fused infrared image is input into the trained model. The model learns and analyzes the image features and outputs the corresponding temperature result, denoted as the first temperature result. In this embodiment, the deep learning model is used for temperature recognition, which can automatically learn the complex features in the image, improve the accuracy and efficiency of temperature recognition, and does not require manual detection.

[0067] Construct a three-dimensional temperature field based on the fused infrared image, and perform temperature analysis on the cable joint to be measured based on the three-dimensional temperature field to obtain the second temperature result.

[0068] Specifically, this step first constructs a three-dimensional temperature field according to the temperature information at different positions in the fused infrared image and combines it with the geometric structure information of the cable joint. Then, the three-dimensional temperature field is used to find the position with the highest temperature through finite element analysis, and this temperature is the second temperature result. By constructing a three-dimensional temperature field, the internal temperature distribution of the cable joint can be more intuitively and accurately reflected, and potential overheating hazards inside the cable joint can be discovered.

[0069] As a preferred implementation manner of this embodiment, constructing a three-dimensional temperature field based on the fused infrared image is specifically as follows: Align the feature points of the fused infrared image with the three-dimensional model of the cable joint to obtain an initial three-dimensional model, and segment the initial three-dimensional model to obtain multiple grid units divided by unit grid lines. Specifically: First, establish a three-dimensional model of the cable joint, such as obtaining it through CAD modeling. The three-dimensional model of the cable joint contains information such as the geometric shape and size of the cable joint. Then, extract feature points on the fused infrared image and the three-dimensional model of the cable joint respectively. The feature points can be significant features such as corner points and edge points in the image. Then, find the corresponding relationship between the two sets of feature points through a preset algorithm to achieve their alignment, thereby obtaining the initially matched initial three-dimensional model. And for the convenience of subsequent numerical calculations, the initial three-dimensional model is divided into grid units and segmented into multiple grid units, including tetrahedral units, hexahedral units, etc. in common grid unit types.

[0070] Further, the preset algorithm used in this embodiment is, for example, the Scale-Invariant Feature Transform (SIFT) algorithm combined with the Random Sample Consensus (RANSAC) algorithm. In the SIFT algorithm, the scale-space extreme value detection of key points is achieved by constructing a Difference of Gaussian (DOG) scale space, and the DOG scale space is defined as:

[0071] In the formula, represents the pixel coordinates in the image; represents the standard deviation of the Gaussian kernel function, which is used to control the degree of Gaussian blur; represents the Gaussian kernel function, represents the image at position the pixel value at; represents the scale factor between adjacent scales.

[0072] In the RANSAC algorithm, it is assumed that there are b data points in the data set. Randomly select a points as the inlier set (assumed model parameters), calculate the error between the model and other points. If the error is less than the threshold, then this point is an inlier, and continuously iterate and update the inlier set until a certain number of iterations or the inlier ratio requirement is met. Thus, by aligning the feature points, the infrared image information is combined with the three-dimensional model, and the accurate geometric information of the three-dimensional model can be utilized to provide an accurate spatial position basis for the subsequent construction of the temperature field.

[0073] Perform ray tracing on each pixel point of the fused infrared image, and assign the corresponding temperature when the ray emitted from the pixel point intersects the unit grid line to the intersecting grid cell, obtaining a preliminary three-dimensional temperature distribution field.

[0074] Specifically, ray tracing is a method for simulating the propagation of light. For each pixel point of the fused infrared image, a ray is emitted from this pixel point, and this ray propagates in three-dimensional space along a specific direction (determined according to the position and direction information of the imaging device). When the ray intersects the unit grid line obtained by the previous segmentation, the temperature value corresponding to this pixel point is assigned to the intersecting grid cell. By performing ray tracing on all pixel points, the temperature value corresponding to each grid cell can be obtained, thereby constructing a preliminary three-dimensional temperature distribution field.

[0075] Further, if the pixel point the coordinates in the image, according to the parameters of the visual imaging device such as the position and the direction vector , the parametric equation of the ray is:

[0076] In the formula, is a parameter, where Denote the coordinates of points on the light ray. When the light ray intersects with the unit grid lines, solve the intersection points of the light ray equation and the boundary equations of the unit grid lines, and assign the temperature corresponding to the pixel points to the grid cell where the intersection points are located. By means of ray tracing, the temperature information on the two-dimensional image can be intuitively mapped onto the grid cells in the three-dimensional space, establishing the temperature connection between the image and the three-dimensional model, and providing the basic data for accurately calculating the three-dimensional temperature field subsequently. Based on the material composition of the cable joint to be measured, determine the heat conduction characteristics of each material part.

[0077] Specifically, a cable joint is usually composed of various materials, such as metal conductors, insulating materials, etc. Different materials have different heat conduction characteristics, including parameters such as thermal conductivity and specific heat capacity. These heat conduction characteristic parameters of the materials can be obtained by referring to material manuals, experimental measurements, etc. For example, for a common copper conductor, its thermal conductivity has a definite value at a certain temperature. Thus, clarify the heat conduction characteristics of each material part. Considering the heat conduction characteristics of the materials can more truly reflect the heat transfer law inside the cable joint to be measured, making the constructed three-dimensional temperature field more in line with the actual physical process and improving the accuracy of temperature analysis.

[0078] Based on the heat conduction characteristics of each material part, the preset boundary conditions, and the preset iteration termination conditions, perform iterative optimization on the preliminary three-dimensional temperature distribution field to obtain the three-dimensional temperature field.

[0079] Specifically, according to the basic principle of heat conduction, in this embodiment, numerical calculation methods such as the finite element method are used to perform iterative optimization on the preliminary three-dimensional temperature distribution field. The preset boundary conditions include the heat transfer conditions between the surface of the cable joint and the surrounding environment (such as the convective heat transfer coefficient, ambient temperature, etc.) and the possible heat source conditions, etc. During the iteration process, according to the heat conduction characteristics of each material part, continuously adjust the temperature values of each grid cell, so that the entire three-dimensional temperature field gradually converges to a stable state that conforms to the actual physical laws. When the preset iteration termination conditions are met (such as the temperature change between two adjacent iterations is less than the preset threshold, or the set maximum number of iterations is reached), stop the iteration to obtain the final three-dimensional temperature field.

[0080] As a preferred implementation manner of this embodiment, the calculation formula for performing iterative optimization on the preliminary three-dimensional temperature distribution field is specifically:

[0081] In the formula,

[0082] respectively represent the indices of the grid cell in the three coordinate axis directions in the three-dimensional space; represents the temperature value of the grid cell with coordinates in the three-dimensional temperature field at time; ​Denotes the predicted temperature value of the grid cell with coordinates at the moment after one iteration; ; Denotes the time step, which determines the time interval corresponding to each iteration and is used to control the rate of change of temperature over time; Denotes the density of the material corresponding to the grid cell with coordinates ; Denotes the specific heat capacity of the material corresponding to the grid cell with coordinates ; Denotes the set of all grid cells adjacent to the grid cell with coordinates ; Denotes the thermal conductivity between the adjacent grid cell and the current grid cell ; Denotes the temperature value of the adjacent grid cell at the moment , Denotes the distance between the current grid cell and the adjacent grid cell ; Denotes the heat source intensity inside the grid cell with coordinates at the moment .

[0083] Furthermore, by considering the heat conduction between the grid cell and the adjacent grid cells through the thermal conductivity and the distance , the heat source inside the grid cell, as well as the thermophysical properties of the material, density and specific heat capacity , the heat conduction process of the three-dimensional temperature field of the cable joint can be more accurately simulated.

[0084] The first temperature result and the second temperature result are fused through the fusion mode to obtain the fused temperature result, and the cable joint overheating diagnosis is performed on the cable joint to be measured based on the fused temperature result to obtain the diagnosis result.

[0085] Specifically, the fusion mode includes a first mode and a second mode, wherein the first mode indicates that the current environment is in a high humidity environment, and the second mode indicates that the current environment is in a strong wind or a large temperature difference environment, so that different values ​​of the temperature results after fusion are calculated to be different in different environments, thereby minimizing the interference of different temperature environments on temperature detection and improving the calculation accuracy of the fused temperature results. After obtaining the fused temperature result, it is compared with the normal temperature range of the cable joint. If it exceeds the normal temperature range, it is judged that the cable joint to be tested is overheated, and according to different data exceeding the normal temperature range, the overheating is divided into different stages such as initial overheating, intermediate overheating and severe overheating, etc., so as to obtain the final diagnosis result. The advantage of fusing the temperature results obtained by two different methods is that their advantages can be comprehensively utilized to improve the accuracy and reliability of diagnosis.

[0086] As a preferred implementation of this embodiment, the fusion mode includes a first mode, and the first mode indicates that the current environment is in a high humidity environment; the first temperature result and the second temperature result are fused by the first mode, and the fused temperature result is specifically: A first correction factor and a second correction factor of humidity to the first temperature result and the second temperature result are obtained respectively.

[0087] Furthermore, humidity can interfere with the accuracy of infrared thermal imaging, causing a deviation between the measured temperature and the actual temperature. Therefore, the first correction factor can be obtained through experimental results and data analysis. and the second correction factor In a high humidity environment, the traditional temperature measurement method may produce large errors due to humidity interference. By introducing a correction factor, the possibility of such errors can be effectively reduced, making subsequent overheating diagnosis more reliable.

[0088] The first temperature result and the second temperature result are corrected based on the first correction factor and the second correction factor respectively to obtain corrected first temperature result and second temperature result.

[0089] Specifically, after obtaining the first correction factor and the second correction factor Then, directly compare it with the corresponding first temperature result and the second temperature result Perform calculations to obtain the corrected first temperature results. , and the corrected second temperature result , in order to correct the temperature results, remove the interference of humidity on temperature measurement, and provide a more accurate data basis for subsequent fusion and diagnosis.

[0090] Determine the first fusion weight and the second fusion weight of the corrected first temperature result and the second temperature result respectively based on a preset humidity function.

[0091] Specifically, establish a preset humidity function according to the relationship between humidity and the reliability of the two temperature results, which is expressed as:

[0092] In the formula, represents the humidity function; represents a constant determined according to experiments; represents the current ambient humidity; Take the calculated value of the preset humidity function as the first fusion weight, and calculate the second fusion weight through the first fusion weight, which is expressed as:

[0093] In the formula, is the second fusion weight.

[0094] The purpose of introducing the preset humidity function is to dynamically adjust the weights of the two temperature results according to the magnitude of humidity. When the humidity is low is close to 1, indicating that the weight of the first temperature result is larger. When the humidity is high decreases, and the weight of the second temperature result relatively increases. In this way, the influence of humidity on the two temperature measurement methods is comprehensively considered, making the fusion result more accurate.

[0095] Based on the first fusion weight and the second fusion weight, perform weighted fusion on the corrected first temperature result and the second temperature result to obtain the fused temperature result.

[0096] Specifically, when obtaining the first fusion weight , the second fusion weight and the corrected first temperature result , the corrected second temperature result , calculate the two according to the weighted fusion method: . Weighted fusion can comprehensively consider the information of the two temperature results and perform reasonable weight allocation according to their reliability.

[0097] As a preferred implementation manner of this embodiment, the fusion mode includes a second mode, and the second mode represents that the current environment is in a strong wind or a large temperature difference environment; fuse the first temperature result and the second temperature result through the second mode, and the specific fused temperature result is: Based on Newton's law of cooling, construct a physical model of the influence of strong wind on the heat dissipation of the cable joint to be measured.

[0098] Specifically, Newton's law of cooling describes that during the convective heat dissipation process of an object, the heat dissipation rate is related to the temperature difference between the object and the surrounding environment and the convective heat transfer coefficient. In a strong wind or a large temperature difference environment, the heat dissipation of the cable joint will change significantly. Therefore, it is first necessary to determine the surface heat transfer coefficient of the cable joint, which is related to factors such as wind speed, the shape of the cable joint, and surface roughness. For example, the surface heat transfer coefficient can be calculated through experiments or empirical formulas. Then, based on Newton's law of cooling, a mathematical model between the heat dissipation and temperature of the cable joint is established.

[0099] Furthermore, the formula of Newton's law of cooling is:

[0100] In the formula, represents the heat dissipation per unit time; represents the surface heat transfer coefficient; represents the surface area of the cable joint; represents the temperature of the cable joint; represents the ambient temperature.

[0101] Based on Newton's law of cooling, a differential equation for the change of the cable joint temperature with time is established:

[0102] In the formula, represents the mass of the cable joint, represents the specific heat capacity of the cable joint material.

[0103] By solving the differential equation (under certain initial conditions), the relationship between the heat dissipation and temperature change of the cable joint in a strong wind or large temperature difference environment is obtained, thus constructing a physical model.

[0104] Based on the physical model, the first compensation value for the first temperature result and the second compensation value for the second temperature result are determined.

[0105] Specifically, after obtaining the physical model, the first temperature result and the second temperature result are respectively substituted into the physical model. According to the calculated heat dissipation and temperature change relationship in the physical model, the temperature deviation caused by strong wind or temperature difference is determined, and this temperature deviation is the value that needs to be compensated. For example, through the physical model calculation, it is found that under the current environmental conditions, the actual temperature of the cable joint should be higher than the measured temperature by , then is the first compensation value for the first temperature result. Similarly, the second compensation value for the second temperature result can be obtained By determining the compensation value through a physical model, it is possible to specifically correct the temperature measurement results for environments with strong winds or large temperature differences. This enables the temperature results to more accurately reflect the true temperature of the cable joint in the actual environment, improving the accuracy and reliability of the temperature data.

[0106] Based on the first compensation value and the second compensation value, the first temperature result and the second temperature result are respectively compensated to obtain the compensated first temperature result and the compensated second temperature result.

[0107] Specifically, the determined first compensation value and the second compensation value are respectively operated with the first temperature result and the second temperature result to obtain the compensated temperature result. The compensation method can be an addition operation, that is, the compensated first temperature result , the compensated second temperature result + . Compensating the temperature result can eliminate the influence of strong winds or large temperature differences on temperature measurement, providing more accurate data for subsequent fusion and diagnosis. The compensated temperature result is more in line with the true temperature state of the cable joint in the actual environment.

[0108] Based on the preset temperature difference weight formula, the first dynamic weight and the second dynamic weight of the compensated first temperature result and the compensated second temperature result are respectively determined.

[0109] As a preferred implementation manner of this embodiment, the preset temperature difference weight formula is expressed as:

[0110] In the formula, represents the temperature difference weight function; represents the temperature difference; represents the adjustment coefficient; represents the reference temperature difference; when the temperature difference is large, approaches 1, focusing on the second temperature result; when the temperature difference is small, approaches 0, focusing on the first temperature result.

[0111] Similarly, the preset temperature difference weight formula is established based on the relationship between the temperature difference and the reliability of the two temperature results. Dynamically adjusting the weight according to the temperature difference situation can more reasonably integrate the two temperature results. In different temperature difference environments, the accuracy and reliability of the two temperature measurement methods will change. Through dynamic weight allocation, the more reliable temperature result can play a greater role in the fusion, improving the accuracy of the fusion result.

[0112] Based on the first dynamic weight and the second dynamic weight, the compensated first temperature result and the second temperature result are weighted and fused to obtain the fused temperature result.

[0113] Specifically, after obtaining the first dynamic weight, the second dynamic weight, the compensated first temperature result, and the compensated second temperature result, they are calculated in a weighted fusion manner, expressed as:

[0114] In the formula, is the fused temperature result; is the second temperature result; is the first temperature result.

[0115] Weighted fusion can comprehensively consider the information of the two temperature results and reasonably allocate weights according to their reliability in different temperature difference environments. It fully utilizes the advantages of the two temperature measurement methods and takes into account the influence of strong wind or large temperature difference environments on the reliability of temperature measurement. This makes the fused temperature result more accurate and reliable, providing more powerful data support for the overheat diagnosis of cable joints.

[0116] Embodiment 2 Correspondingly, as shown in Figure 2 this embodiment provides an infrared thermal imaging-based cable joint overheat diagnosis system, which is used to implement the infrared thermal imaging-based cable joint overheat diagnosis method as described in Embodiment 1 of the present invention, including an acquisition unit, a fusion unit, a first processing unit, a second processing unit, and a diagnosis unit.

[0117] The acquisition unit is used to collect multiple infrared thermal images of the cable joint to be measured through multiple vision cameras.

[0118] The fusion unit is used to fuse the infrared thermal images based on the position information and direction information of the multiple vision cameras to obtain the fused infrared image.

[0119] The first processing unit is used to establish an infrared temperature recognition model based on a neural network, train the infrared temperature recognition model according to the pre-collected infrared image samples and the temperature result labels corresponding to the samples, and input the fused infrared image into the infrared temperature recognition model to obtain the first temperature result.

[0120] The second processing unit is used to construct a three-dimensional temperature field based on the fused infrared image and perform temperature analysis on the cable joint to be measured based on the three-dimensional temperature field to obtain the second temperature result.

[0121] A diagnostic unit is configured to fuse a first temperature result and a second temperature result through a fusion mode to obtain a fused temperature result, and perform overheat diagnosis on a cable joint to be measured based on the fused temperature result to obtain a diagnosis result.

[0122] Embodiment III This embodiment provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the method for diagnosing overheating of a cable joint based on infrared thermal imaging as described in Embodiment I of the present invention.

[0123] In the embodiments of the present application, "at least one" means one or more, and "a plurality" means two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent the situation where A exists alone, A and B exist simultaneously, or B exists alone. Where A and B may be singular or plural. The character " / " generally represents an "or" relationship between the associated objects before and after. "At least one of the following" and its similar expressions refer to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, and c may represent: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, and c may be single or multiple.

[0124] Those of ordinary skill in the art can realize that the units and algorithm steps described in the embodiments disclosed herein can be implemented by a combination of electronic hardware, computer software, and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0125] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments and will not be described in detail herein.

[0126] In several embodiments provided by the present application, if any function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art or a part of this technical solution can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM for short), random access memories (RAM for short), magnetic disks, or optical discs that can store program codes.

[0127] The above are only the embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall be similarly included in the patent protection scope of the present invention.

Claims

1. A method for diagnosing overheating of cable joints based on infrared thermal imaging, characterized in that, Including the following steps: Collecting a plurality of infrared thermal images of the cable joint to be measured through a plurality of visual camera devices; Fusing the infrared thermal images based on the position information and direction information of the plurality of visual camera devices to obtain a fused infrared image; Establishing an infrared temperature recognition model based on a neural network, training the infrared temperature recognition model according to the pre-collected infrared image samples and the corresponding temperature result labels of the samples, and inputting the fused infrared image into the infrared temperature recognition model to obtain a first temperature result; Constructing a three-dimensional temperature field based on the fused infrared image, and performing temperature analysis on the cable joint to be measured based on the three-dimensional temperature field to obtain a second temperature result; Fusing the first temperature result and the second temperature result through a fusion mode to obtain a fused temperature result, and performing cable joint overheat diagnosis on the cable joint to be measured based on the fused temperature result to obtain a diagnosis result.

2. The method for diagnosing overheating of cable joints based on infrared thermal imaging according to claim 1, wherein, Fusing the infrared thermal images based on the position information and direction information of the plurality of visual camera devices to obtain a fused infrared image specifically as follows: Performing a non-subsampled contourlet transform on the infrared thermal images to obtain a plurality of sub-images; Based on the position information and direction information of the plurality of visual camera devices, constructing a spatio-temporal correlation matrix for each sub-image, and determining the spatio-temporal correlation weight of each sub-image according to the spatio-temporal correlation matrix; Performing weighted fusion on each sub-image and its corresponding spatio-temporal correlation weight to obtain a fused sub-image; Performing an inverse non-subsampled contourlet transform on the fused sub-image to obtain a fused infrared image.

3. The cable joint overheating diagnosis method based on infrared thermal imaging according to claim 2, wherein, The specific calculation formula of the spatio-temporal correlation weight is as follows: In the formula, represents the spatio-temporal correlation weight between the th visual camera device and the , both represent adjustment coefficients; represents the total number of visual camera devices, ; , respectively represent the position vectors of the th visual camera device and the th visual camera device; , , respectively represent the direction vectors of the th visual camera device; , represent the combined vector form of the direction vectors of the th visual camera device and the th visual camera device; represents the Euclidean distance between the position vectors of the th visual camera device and the th visual camera device; represents the distance metric between the direction vectors of the th visual camera device and the th visual camera device.

4. The cable joint overheating diagnosis method based on infrared thermal imaging according to claim 1, characterized in that, Constructing a three-dimensional temperature field based on the fused infrared image specifically as follows: Aligning the feature points of the fused infrared image with the three-dimensional model of the cable joint to obtain an initial three-dimensional model, and segmenting the initial three-dimensional model to obtain a plurality of grid cells segmented by unit grid lines; Performing ray tracing on each pixel point of the fused infrared image, and assigning the corresponding temperature when the ray emitted from the pixel point intersects the unit grid line to the intersecting grid cell to obtain a preliminary three-dimensional temperature distribution field; Based on the material composition of the cable joint to be measured, determining the heat conduction characteristics of each material part; Performing iterative optimization on the preliminary three-dimensional temperature distribution field based on the heat conduction characteristics of each material part, preset boundary conditions, and preset iterative termination conditions to obtain a three-dimensional temperature field.

5. The method for diagnosing overheating of cable joints based on infrared thermal imaging according to claim 4, characterized in that, The specific calculation formula for performing iterative optimization on the preliminary three-dimensional temperature distribution field is as follows: In the formula, represents at the moment, the temperature value of the grid cell with coordinates in the three-dimensional temperature field; represents the temperature prediction value of the grid cell with coordinates at the moment after one iteration; represents the time step; represents the density of the material corresponding to the grid cell with coordinates ; represents the specific heat capacity of the material corresponding to the grid cell with coordinates ; represents the set of all grid cells adjacent to the grid cell with coordinates ; represents the thermal conductivity between the adjacent grid cell and the current grid cell ; represents at the moment the temperature value of the adjacent grid cell , represents the distance between the current grid cell and the adjacent grid cell ; represents at the moment the heat source intensity inside the grid cell with coordinates .

6. The cable joint overheating diagnosis method based on infrared thermal imaging according to claim 1, characterized in that, The fusion mode includes a first mode, and the first mode represents that the current environment is a high-humidity environment; fusing the first temperature result and the second temperature result through the first mode to obtain a fused temperature result specifically as follows: Respectively obtaining a first correction factor and a second correction factor of humidity for the first temperature result and the second temperature result; Respectively correcting the first temperature result and the second temperature result based on the first correction factor and the second correction factor to obtain a corrected first temperature result and a corrected second temperature result; Respectively determining a first fusion weight and a second fusion weight of the corrected first temperature result and the corrected second temperature result based on a preset humidity function; Based on the first fusion weight and the second fusion weight, performing weighted fusion on the corrected first temperature result and the corrected second temperature result to obtain a fused temperature result.

7. The method for diagnosing overheating of cable joints based on infrared thermal imaging according to claim 1, characterized in that The fusion mode includes a second mode, which characterizes that the current environment is in a strong wind or a large temperature difference environment; the first temperature result and the second temperature result are fused through the second mode to obtain the fused temperature result specifically as follows: Based on Newton's law of cooling, a physical model of the influence of strong wind on the heat dissipation of the cable joint to be measured is constructed; Based on the physical model, the first compensation value of the first temperature result and the second compensation value of the second temperature result are determined; Based on the first compensation value and the second compensation value, the first temperature result and the second temperature result are compensated respectively to obtain the compensated first temperature result and the second temperature result; Based on the preset temperature difference weight formula, the first dynamic weight and the second dynamic weight of the compensated first temperature result and the second temperature result are determined respectively; Based on the first dynamic weight and the second dynamic weight, the compensated first temperature result and the second temperature result are weighted and fused to obtain the fused temperature result.

8. The method for diagnosing overheating of cable joints based on infrared thermal imaging according to claim 7, wherein The preset temperature difference weight formula is expressed as: In the formula, represents the temperature difference weight function; represents the temperature difference; represents the adjustment coefficient; represents the reference temperature difference; when the temperature difference is large, approaches 1, focusing on the second temperature result; when the temperature difference is small, approaches 0, focusing on the first temperature result; Based on the first dynamic weight and the second dynamic weight, the weighted fusion of the compensated first temperature result and the second temperature result is expressed as: In the formula, is the temperature result after fusion; is the second temperature result; is the first temperature result.

9. An overheating diagnosis system for cable joints based on infrared thermal imaging, characterized in that, The system is used to implement the infrared thermal imaging-based cable joint overheating diagnosis method described in any one of claims 1-8, and includes an acquisition unit, a fusion unit, a first processing unit, a second processing unit, and a diagnosis unit; The acquisition unit is used to collect a plurality of infrared thermal images of the cable joint to be measured through a plurality of visual camera devices; The fusion unit is used to fuse the infrared thermal images based on the position information and direction information of the plurality of visual camera devices to obtain a fused infrared image; The first processing unit is used to establish an infrared temperature recognition model based on a neural network, train the infrared temperature recognition model according to the pre-collected infrared image samples and the temperature result labels corresponding to the samples, and input the fused infrared image into the infrared temperature recognition model to obtain the first temperature result; The second processing unit is used to construct a three-dimensional temperature field based on the fused infrared image, and perform temperature analysis on the cable joint to be measured based on the three-dimensional temperature field to obtain the second temperature result; The diagnosis unit is used to fuse the first temperature result and the second temperature result through the fusion mode to obtain the fused temperature result, and perform cable joint overheating diagnosis on the cable joint to be measured based on the fused temperature result to obtain the diagnosis result.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the infrared thermal imaging-based cable joint overheating diagnosis method described in any one of claims 1 to 8.

Images