Electric transmission line insulator ultraviolet corona discharge image segmentation method and system

Abstract

The invention provides a power transmission line insulator ultraviolet corona discharge image segmentation method and system, and the method comprises the steps: obtaining a plurality of ultraviolet images of a target insulator under the discharge condition, and carrying out the region marking of each ultraviolet image; respectively extracting a first feature, a second feature, a third feature and a fourth feature from the discharge region and the background region, and respectively constructing a morphological difference model, a texture difference model, a spectrum difference model and a spatial difference model; constructing a target function according to the morphological difference model, the texture difference model, the spectrum difference model and the spatial difference model, optimizing the target function by taking difference maximization as a target, and obtaining a segmentation mask of the ultraviolet image according to an optimization result; and inputting the segmentation mask and all the marked ultraviolet images into the initial ultraviolet image segmentation model for training to obtain a final ultraviolet image segmentation model. According to the method, the segmentation accuracy of the ultraviolet discharge image can be greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of transfer learning, and particularly to a method and system for segmenting ultraviolet corona discharge images of transmission line insulators. Background Art

[0002] In the power system, the operating state of transmission line insulators is directly related to the safety and stability of the power grid. During long-term operation, insulators may experience corona discharge on the surface due to aging, pollution, environmental factors, etc. This kind of discharge not only causes energy loss, but may also further develop into insulation breakdown, leading to serious power accidents. Therefore, regular inspection of transmission line insulators to timely detect and handle corona discharge problems is one of the important measures to ensure the safe operation of the power grid.

[0003] As a non-contact detection method, ultraviolet imaging technology can effectively capture the corona discharge phenomenon on the surface of insulators. By analyzing the discharge characteristics in the ultraviolet image, the health status of the insulators can be evaluated. However, the contrast between the discharge area and the background area in the ultraviolet image is often low, and affected by factors such as environmental illumination and shooting angle, the accurate segmentation of the discharge area becomes a challenging task. Traditional image segmentation methods, such as threshold segmentation and edge detection, often fail to achieve ideal results when dealing with ultraviolet images with complex backgrounds, resulting in the problem of low segmentation accuracy of the discharge area. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for segmenting ultraviolet corona discharge images of transmission line insulators, aiming to solve the problem that traditional technologies are prone to low segmentation accuracy of the discharge area due to the low contrast between the discharge area and the background area when segmenting ultraviolet images.

[0005] In the first aspect, the present invention provides a method for segmenting ultraviolet corona discharge images of transmission line insulators, the method comprising:

[0006] Obtain multiple ultraviolet images of a target insulator under discharge conditions, and perform regional annotation on each of the ultraviolet images, and the annotation results include a discharge area and a background area;

[0007] Extract first features, second features, third features, and fourth features from the discharge area and the background area respectively, and construct a morphological difference model, a texture difference model, a spectral difference model, and a spatial difference model according to the first features, second features, third features, and fourth features of the discharge area and the background area;

[0008] Construct an objective function based on the morphological difference model, the texture difference model, the spectral difference model, and the spatial difference model, optimize the objective function with the goal of maximizing the difference, and obtain the segmentation mask of the ultraviolet image according to the optimization result;

[0009] Input the segmentation mask and all the labeled ultraviolet images into the initial ultraviolet image segmentation model for training to obtain the final ultraviolet image segmentation model.

[0010] Further, the steps of respectively extracting the first feature, the second feature, the third feature, and the fourth feature from the discharge area and the background area, and respectively constructing the morphological difference model, the texture difference model, the spectral difference model, and the spatial difference model according to the first feature, the second feature, the third feature, and the fourth feature of the discharge area and the background area include:

[0011] Respectively obtain the area, perimeter, and circularity of the discharge area and the background area in each ultraviolet image;

[0012] Construct a morphological difference model based on the area, perimeter, and circularity, and obtain the morphological difference value of each ultraviolet image according to the morphological difference model:

[0013]

[0014] where M i represents the morphological difference value of the i-th ultraviolet image, and α1, α2, and α3 are all weight coefficients of the morphological difference model. A 1i , A 2i respectively represent the areas of the discharge area and the background area in the i-th ultraviolet image, Y 1i , Y 2i respectively represent the circularities of the discharge area and the background area in the i-th ultraviolet image, and L 1i , L 2i respectively represent the perimeters of the discharge area and the background area in the i-th ultraviolet image;

[0015] Construct a first feature matrix based on the morphological difference value:

[0016] M = [M1, M2,..., M m ;

[0017] where M represents the first feature matrix, and M1, M2, M m respectively represent the morphological difference values of the 1st, 2nd, and m-th ultraviolet images, and m represents the total number of ultraviolet images.

[0018] Further, the method further includes:

[0019] Convert the ultraviolet image into a grayscale histogram, and respectively obtain the KL divergence, the energy of the gray-level co-occurrence matrix, and the entropy of the gray-level co-occurrence matrix in each grayscale histogram;

[0020] Construct a texture difference model according to the following formula:

[0021]

[0022] where T i represents the texture difference value in the i-th grayscale histogram, represents the KL divergence of the probability distribution discharge area relative to the probability distribution background area in the i-th grayscale histogram, and β1, β2, and β3 are all weight coefficients of the texture difference model. P 1i , P 2i respectively represent the discharge area and the background area in the i-th image. E 1i , E 2i respectively represent the energy of the gray-level co-occurrence matrix of the discharge area and the background area in the i-th grayscale histogram. S 1i , S 2i respectively represent the entropy of the gray-level co-occurrence matrix of the discharge area and the background area in the i-th grayscale histogram;

[0023] Construct a second feature matrix according to the texture difference value:

[0024] T = [T1, T2,..., T m ;

[0025] where T represents the second feature matrix, and T1, T2, T m respectively represent the texture feature difference values of the 1st, 2nd, and m-th grayscale histograms.

[0026] Furthermore, the method further includes:

[0027] Respectively obtain the spectral intensity and spectral bandwidth of the discharge area and the background area in each ultraviolet image at a preset wavelength;

[0028] Construct a spectral difference model according to the following formula:

[0029]

[0030] where G i represents the spectral difference value of the i-th ultraviolet image, and γ1, γ2 are both weight coefficients of the spectral difference model. I 1i , I 2i respectively represent the spectral intensity of the discharge area and the background area in the i-th ultraviolet image. B 1i , B 2i respectively represent the spectral bandwidth of the discharge area and the background area in the i-th ultraviolet image;

[0031] Construct a third feature matrix according to the spectral difference value:

[0032] G = [G1, G2, …, G m ;

[0033] where G represents the third feature matrix, and G1, G2, G m respectively represent the spectral difference values of the 1st, 2nd, and mth ultraviolet images.

[0034] Furthermore, the method further includes:

[0035] Calculate the centroid distance between the discharge area and the background area in each ultraviolet image, and respectively obtain the regional density of the discharge area and the background area in each ultraviolet image;

[0036] Construct a spatial difference model according to the following formula:

[0037]

[0038] where K i represents the spatial difference value of the ith ultraviolet image, O i represents the centroid distance of the ith ultraviolet image, L0 represents the diagonal length of the ultraviolet image, and Q 1i , Q 2i respectively represent the regional density of the discharge area and the background area in the ith image;

[0039] Construct a fourth feature matrix according to the spatial difference value:

[0040] K = [K1, K2, …, K m ;

[0041] where K represents the fourth feature matrix, and K1, K2, K m respectively represent the spatial difference values of the 1st, 2nd, and mth ultraviolet images.

[0042] Furthermore, the step of constructing an objective function according to the morphological difference model, the texture difference model, the spectral difference model, and the spatial difference model, optimizing the objective function with the goal of maximizing the difference, and obtaining the segmentation mask of the ultraviolet image according to the optimization result includes:

[0043] Construct an objective function according to the following formula:

[0044] maxF = A·M + B·T + C·G + D·K;

[0045] where A, B, C, and D are all parameter matrices, a1, am respectively represent the 1st and m-th weight parameters related to the first feature matrix, b1, b m respectively represent the 1st and m-th weight parameters related to the second feature matrix, c1, c m respectively represent the 1st and m-th weight parameters related to the third feature matrix, d1, d m respectively represent the 1st and m-th weight parameters related to the fourth feature matrix, and maxF represents the sum of maximized feature differences;

[0046] Iteratively solve the objective function to obtain each parameter value in each parameter matrix.

[0047] In a second aspect, the present invention provides a system for segmenting ultraviolet corona discharge images of transmission line insulators, and the system includes:

[0048] An image annotation module, configured to obtain multiple ultraviolet images of a target insulator under a discharge condition, and perform regional annotation on each ultraviolet image, and the annotation result includes a discharge area and a background area;

[0049] A difference model construction module, configured to respectively extract the first feature, the second feature, the third feature, and the fourth feature from the discharge area and the background area, and respectively construct a morphological difference model, a texture difference model, a spectral difference model, and a spatial difference model according to the first feature, the second feature, the third feature, and the fourth feature of the discharge area and the background area;

[0050] An objective function construction module, configured to construct an objective function according to the morphological difference model, the texture difference model, the spectral difference model, and the spatial difference model, optimize the objective function with the goal of maximizing the difference, and obtain a segmentation mask of the ultraviolet image according to the optimization result;

[0051] A segmentation model construction module, configured to input the segmentation mask and all the annotated ultraviolet images into an initial ultraviolet image segmentation model for training to obtain a final ultraviolet image segmentation model.

[0052] In a third aspect, the present invention provides a storage medium that stores one or more programs, and when the program is executed by a processor, the above-mentioned method for segmenting ultraviolet corona discharge images of transmission line insulators is implemented.

[0053] In a fourth aspect, the present invention provides an electronic device, and the electronic device includes a memory and a processor, wherein:

[0054] The memory is used to store a computer program;

[0055] When the processor is used to execute the computer program stored in the memory, the above-mentioned method for segmenting the ultraviolet corona discharge image of the transmission line insulator is realized.

[0056] Compared with the prior art, the present invention has the following advantages:

[0057] 1. By fusing four types of features, namely morphology, texture, spectrum, and space, of the discharge area and the background area, and constructing a morphology difference model, a texture difference model, a spectrum difference model, and a space difference model, a multi-dimensional accurate characterization of the ultraviolet corona discharge image of the transmission line insulator is realized. This method effectively overcomes the limitations of traditional single-feature segmentation methods, can distinguish the discharge area and the background area more comprehensively and meticulously, significantly improves the accuracy and reliability of image segmentation, and provides a solid technical support for subsequent insulator health status assessment and fault warning.

[0058] 2. By optimizing the specific parameter values of each parameter matrix, it can adapt to different image conditions, improve the stability and generalization ability of segmentation, and at the same time make the features that have a greater impact on the segmentation result obtain higher weights, giving full play to the advantages of each feature and further improving the segmentation accuracy.

[0059] 3. By constructing four parameter matrices for the four feature matrices and jointly optimizing the objective function, a uniquely determined segmentation mask value is obtained, and this segmentation mask is used as the label of all the images of the labeled areas and input into the model for training. This process ensures that all features work together under a unified optimization framework, avoiding inconsistent segmentation results caused by differences in parameter settings. Description of the Drawings

[0060] Figure 1 It is a flowchart of the method for segmenting the ultraviolet corona discharge image of the transmission line insulator proposed in an embodiment of the present invention;

[0061] Figure 2 It is a schematic structural diagram of the system for segmenting the ultraviolet corona discharge image of the transmission line insulator proposed in an embodiment of the present invention.

[0062] The following specific embodiments will further illustrate the present invention in conjunction with the above-mentioned drawings. Specific Embodiments

[0063] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention. Unless otherwise defined, the technical terms or scientific terms used herein shall have the ordinary meanings as understood by those of ordinary skill in the art to which the present invention pertains. The words such as "including" used herein mean that the elements or items appearing before this word cover the elements or items listed after this word and their equivalents, without excluding other elements or items.

[0064] As Figure 1 shown, an embodiment of the present invention provides a method for segmenting ultraviolet corona discharge images of transmission line insulators. The method includes steps S101 to S104, where:

[0065] Step S101: Obtain multiple ultraviolet images of a target insulator under the condition of discharge, and perform region annotation on each of the ultraviolet images. The annotation results include a discharge region and a background region;

[0066] It should be noted that in this step, first, multiple ultraviolet images of it under the condition of discharge need to be obtained through a professional ultraviolet imaging device. These images contain rich discharge feature information, but for the convenience of subsequent analysis and processing, region annotation needs to be performed on each ultraviolet image.

[0067] Step S102: Extract the first feature, second feature, third feature, and fourth feature from the discharge region and the background region respectively, and construct a morphological difference model, a texture difference model, a spectral difference model, and a spatial difference model according to the first feature, second feature, third feature, and fourth feature of the discharge region and the background region;

[0068] It should be noted that the first feature mainly includes area, perimeter, and circularity. In some embodiments, first, the area, perimeter, and circularity of the discharge region and the background region in each ultraviolet image are obtained respectively;

[0069] Construct a morphological difference model according to the area, perimeter, and circularity, and obtain the morphological difference value of each ultraviolet image according to the morphological difference model:

[0070]

[0071] where, M i represents the morphological difference value of the i-th ultraviolet image, and α1, α2, and α3 are all weight coefficients of the morphological difference model, A 1i 、A2i respectively represent the areas of the discharge region and the background region in the i-th ultraviolet image, Y 1i and Y 2i respectively represent the circularity of the discharge region and the background region in the i-th ultraviolet image, L 1i and L 2i respectively represent the perimeters of the discharge region and the background region in the i-th ultraviolet image;

[0072] Construct a first feature matrix according to the morphological difference value:

[0073] M = [M1, M2,..., M m ;

[0074] where M represents the first feature matrix, and M1, M2, M m respectively represent the morphological difference values of the 1st, 2nd,..., m-th ultraviolet images, and m represents the total number of ultraviolet images.

[0075] In some embodiments, the second feature includes KL divergence, the energy of the gray-level co-occurrence matrix, and the entropy of the gray-level co-occurrence matrix. Specifically, convert the ultraviolet image into a gray-level histogram, and respectively obtain the KL divergence, the energy of the gray-level co-occurrence matrix, and the entropy of the gray-level co-occurrence matrix in each gray-level histogram;

[0076] Construct a texture difference model according to the following formula:

[0077]

[0078] where T i represents the texture difference value in the i-th gray-level histogram, represents the KL divergence of the probability distribution of the discharge region relative to the probability distribution of the background region in the i-th gray-level histogram, β1, β2, and β3 are all weight coefficients of the texture difference model, and P 1i and P 2i respectively represent the discharge region and the background region in the i-th image, and E 1i and E 2i respectively represent the energy of the gray-level co-occurrence matrix of the discharge region and the background region in the i-th gray-level histogram, and S 1i and S 2i respectively represent the entropy of the gray-level co-occurrence matrix of the discharge region and the background region in the i-th gray-level histogram;

[0079] Construct a second feature matrix according to the texture difference value:

[0080] T = [T1, T2,..., T m ;

[0081] where T represents the second feature matrix, and T1, T2, T mrespectively represent the texture feature difference values of the first, second, and m-th grayscale histograms.

[0082] In some embodiments, the third feature includes spectral intensity and spectral bandwidth. Specifically, the spectral intensity and spectral bandwidth of the discharge region and the background region in each ultraviolet image are respectively obtained at a preset wavelength;

[0083] Construct a spectral difference model according to the following formula:

[0084]

[0085] where G i represents the spectral difference value of the i-th ultraviolet image, and γ1 and γ2 are both weight coefficients of the spectral difference model. I 1i and I 2i respectively represent the spectral intensities of the discharge region and the background region in the i-th ultraviolet image, and B 1i and B 2i respectively represent the spectral bandwidths of the discharge region and the background region in the i-th ultraviolet image;

[0086] Construct a third feature matrix according to the spectral difference value:

[0087] G = [G1, G2,..., G m ;

[0088] where G represents the third feature matrix, and G1, G2, and G m respectively represent the spectral difference values of the first, second, and m-th ultraviolet images.

[0089] In some embodiments, the fourth feature includes centroid distance and regional density. Specifically, first calculate the centroid distance between the discharge region and the background region in each ultraviolet image, and respectively obtain the regional densities of the discharge region and the background region in each ultraviolet image;

[0090] Construct a spatial difference model according to the following formula:

[0091]

[0092] where K i represents the spatial difference value of the i-th ultraviolet image, O i represents the centroid distance of the i-th ultraviolet image, L0 represents the diagonal length of the ultraviolet image, and Q 1i and Q 2i respectively represent the regional densities of the discharge region and the background region in the i-th image;

[0093] Construct a fourth feature matrix according to the spatial difference value:

[0094] K = [K1, K2, …, K m ;

[0095] where K represents the fourth feature matrix, and K1, K2, K m respectively represent the spatial difference values of the 1st, 2nd, …, mth ultraviolet images.

[0096] Step S103: Construct an objective function according to the morphological difference model, the texture difference model, the spectral difference model, and the spatial difference model, optimize the objective function with the goal of maximizing the difference, and obtain the segmentation mask of the ultraviolet image according to the optimization result;

[0097] In this step, specifically construct the objective function according to the following formula:

[0098] maxF = A·M + B·T + C·G + D·K;

[0099] where A, B, C, and D are all parameter matrices, a1, a m respectively represent the 1st and mth weight parameters related to the first feature matrix, b1, b m respectively represent the 1st and mth weight parameters related to the second feature matrix, c1, c m respectively represent the 1st and mth weight parameters related to the third feature matrix, d1, d m respectively represent the 1st and mth weight parameters related to the fourth feature matrix, and maxF represents the sum of the maximized feature differences;

[0100] Iteratively solve the objective function to obtain each parameter value in each parameter matrix, and then determine each parameter matrix.

[0101] Step S104: Input the segmentation mask and all the labeled ultraviolet images into the initial ultraviolet image segmentation model for training to obtain the final ultraviolet image segmentation model.

[0102] It should be noted that the segmentation mask is the objective value that maximizes the objective function. After obtaining the unique segmentation mask, for the sake of achieving consistent segmentation, it is also necessary to use this segmentation mask as a label and input the segmentation mask and all the labeled images under this segmentation mask into the model for training.

[0103] In summary, according to the above-mentioned method for segmenting ultraviolet corona discharge images of transmission line insulators, the present invention has the following advantages:

[0104] 1. By fusing the morphological, texture, spectral, and spatial features of the discharge region and the background region, and constructing a morphological difference model, a texture difference model, a spectral difference model, and a spatial difference model, a multi-dimensional and accurate characterization of the ultraviolet corona discharge image of the transmission line insulator is achieved. This method effectively overcomes the limitations of traditional single-feature segmentation methods, can distinguish the discharge region and the background region more comprehensively and meticulously, significantly improves the accuracy and reliability of image segmentation, and provides a solid technical support for subsequent insulator health status assessment and fault warning.

[0105] 2. By optimizing the specific parameter values of each parameter matrix, it can adapt to different image conditions, improve the stability and generalization ability of segmentation, and at the same time make the features that have a greater impact on the segmentation results obtain higher weights, giving full play to the advantages of each feature and further improving the segmentation accuracy.

[0106] 3. By constructing four parameter matrices for the four feature matrices and jointly optimizing the objective function, a uniquely determined segmentation mask value is obtained, and this segmentation mask is used as the label of all the labeled region images and input into the model for training. This process ensures that all features work together under a unified optimization framework, avoiding inconsistent segmentation results caused by differences in parameter settings.

[0107] As Figure 2 shown, an embodiment of the present invention provides a system for segmenting ultraviolet corona discharge images of transmission line insulators, and the system includes:

[0108] An image annotation module 10, configured to obtain multiple ultraviolet images of a target insulator under discharge conditions, and perform region annotation on each of the ultraviolet images, and the annotation results include a discharge region and a background region;

[0109] A difference model construction module 20, configured to extract first features, second features, third features, and fourth features from the discharge region and the background region respectively, and construct a morphological difference model, a texture difference model, a spectral difference model, and a spatial difference model according to the first features, second features, third features, and fourth features of the discharge region and the background region;

[0110] An objective function construction module 30, configured to construct an objective function according to the morphological difference model, the texture difference model, the spectral difference model, and the spatial difference model, optimize the objective function with the maximization of differences as the goal, and obtain a segmentation mask of the ultraviolet image according to the optimization result;

[0111] A segmentation model construction module 40, configured to input the segmentation mask and all the labeled ultraviolet images into an initial ultraviolet image segmentation model for training to obtain a final ultraviolet image segmentation model.

[0112] On the other hand, the present invention also provides a storage medium storing one or more programs, which, when executed by a processor, implement the above-mentioned method for segmenting ultraviolet corona discharge images of transmission line insulators.

[0113] On the other hand, the present invention also provides an electronic device including a memory and a processor, where the memory is used to store a computer program, and the processor is used to execute the computer program stored on the memory to implement the above-mentioned method for segmenting ultraviolet corona discharge images of transmission line insulators.

[0114] Those skilled in the art can understand that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or in combination with these instruction execution systems, apparatuses, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

[0115] More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection part (electronic device) having one or more wirings, a portable computer disk cartridge (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpretation, or other suitable processing as necessary, and then stored in a computer memory.

[0116] It should be understood that various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following well-known techniques in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application-specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0117] Although the embodiments of the present invention have been described in detail above, it will be apparent to those skilled in the art that various modifications and variations can be made to these embodiments. However, it should be understood that such modifications and variations are all within the scope and spirit of the present invention as described in the claims. Moreover, the present invention as described herein may have other embodiments and can be implemented or realized in various ways.

Claims

1. A method for segmenting ultraviolet corona discharge images of transmission line insulators, characterized in that, The method includes: Obtaining multiple ultraviolet images of a target insulator under a discharge condition, and performing region annotation on each of the ultraviolet images, where the annotation result includes a discharge region and a background region; Extracting first features, second features, third features, and fourth features from the discharge region and the background region respectively, and constructing a morphological difference model, a texture difference model, a spectral difference model, and a spatial difference model according to the first features, second features, third features, and fourth features of the discharge region and the background region; Constructing an objective function according to the morphological difference model, the texture difference model, the spectral difference model, and the spatial difference model, optimizing the objective function with the maximization of the difference as the goal, and obtaining a segmentation mask of the ultraviolet image according to the optimization result; Inputting the segmentation mask and all the annotated ultraviolet images into an initial ultraviolet image segmentation model for training to obtain a final ultraviolet image segmentation model.

2. The method for segmenting ultraviolet corona discharge images of transmission line insulators according to claim 1, wherein The step of extracting first features, second features, third features, and fourth features from the discharge region and the background region respectively, and constructing a morphological difference model, a texture difference model, a spectral difference model, and a spatial difference model according to the first features, second features, third features, and fourth features of the discharge region and the background region includes: Respectively obtaining the area, perimeter, and circularity of the discharge region and the background region in each ultraviolet image; Constructing a morphological difference model according to the area, perimeter, and circularity, and obtaining the morphological difference value of each ultraviolet image according to the morphological difference model; Among them, M i represents the morphological difference value of the i-th ultraviolet image. α1, α2, and α3 are all weight coefficients of the morphological difference model. A 1i , A 2i respectively represent the areas of the discharge region and the background region in the i-th ultraviolet image. Y 1i , Y 2i respectively represent the circularity of the discharge region and the background region in the i-th ultraviolet image. L 1i , L 2i respectively represent the perimeters of the discharge region and the background region in the i-th ultraviolet image; Constructing a first feature matrix according to the morphological difference value; M = [M1, M2, …, M m ; Among them, M represents the first feature matrix, and M1, M2, and M m respectively represent the morphological difference values of the 1st, 2nd, and mth ultraviolet images, where m represents the total number of ultraviolet images.

3. The method for segmenting the ultraviolet corona discharge image of a transmission line insulator according to claim 2, wherein The method further includes: Converting the ultraviolet image into a grayscale histogram, and respectively obtaining the KL divergence, the energy of the gray-level co-occurrence matrix, and the entropy of the gray-level co-occurrence matrix in each grayscale histogram; Constructing a texture difference model according to the following formula: Among them, T i represents the texture difference value in the i-th gray-level histogram, represents the KL divergence of the probability distribution discharge region relative to the probability distribution background region in the i-th gray-level histogram. β1, β2, and β3 are all weight coefficients of the texture difference model. P 1i , P 2i respectively represent the discharge region and the background region in the i-th image. E 1i , E 2i respectively represent the energies of the gray-level co-occurrence matrices of the discharge region and the background region in the i-th gray-level histogram. S 1i , S 2i respectively represent the entropies of the gray-level co-occurrence matrices of the discharge region and the background region in the i-th gray-level histogram; Constructing a second feature matrix according to the texture difference value; T = [T1, T2, …, T m ; Among them, T represents the second feature matrix, and T1, T2, and T m respectively represent the texture feature difference values of the first, second, and mth grayscale histograms.

4. The method for segmenting the ultraviolet corona discharge image of a transmission line insulator according to claim 3, wherein The method further includes: Respectively obtaining the spectral intensity and spectral bandwidth of the discharge region and the background region in each ultraviolet image at a preset wavelength; Constructing a spectral difference model according to the following formula: Among them, G i represents the spectral difference value of the i-th ultraviolet image, and γ1 and γ2 are both weight coefficients of the spectral difference model. I 1i and I 2i respectively represent the spectral intensities of the discharge area and the background area in the i-th ultraviolet image, and B 1i and B 2i respectively represent the spectral bandwidths of the discharge area and the background area in the i-th ultraviolet image; Constructing a third feature matrix according to the spectral difference value; G = [G1, G2, …, G m ; Among them, G represents the third feature matrix, and G1, G2, and G m respectively represent the spectral difference values of the first, second, and m-th ultraviolet images.

5. The method for segmenting the ultraviolet corona discharge image of a transmission line insulator according to claim 4, characterized in that The method further includes: Calculating the centroid distance between the discharge region and the background region in each ultraviolet image, and respectively obtaining the regional density of the discharge region and the background region in each ultraviolet image; Constructing a spatial difference model according to the following formula: Among them, K i represents the spatial difference value of the i-th ultraviolet image, and O i represents the centroid distance of the i-th ultraviolet image, L0 represents the diagonal length of the ultraviolet image, and Q 1i , Q 2i respectively represent the regional density of the discharge area and the background area in the i-th image; Constructing a fourth feature matrix according to the spatial difference value; K = [K1, K2, …, K m ; Among them, K represents the fourth feature matrix, and K1, K2, and K m respectively represent the spatial difference values of the first, second, and m-th ultraviolet images.

6. The method for segmenting the ultraviolet corona discharge image of the transmission line insulator according to claim 5, characterized in that, The step of constructing an objective function according to the morphological difference model, the texture difference model, the spectral difference model, and the spatial difference model, optimizing the objective function with the maximization of the difference as the goal, and obtaining a segmentation mask of the ultraviolet image according to the optimization result includes: Constructing an objective function according to the following formula: maxF = A·M + B·T + C·G + D·K; Among them, A, B, C, and D are all parameter matrices, a1, a m respectively represent the 1st and m-th weight parameters related to the first feature matrix, b1, b m respectively represent the 1st and m-th weight parameters related to the second feature matrix, c1, c m respectively represent the 1st and m-th weight parameters related to the third feature matrix, d1, d m respectively represent the 1st and m-th weight parameters related to the fourth feature matrix, and maxF represents the sum of maximized feature differences; Performing iterative solution on the objective function to obtain each parameter value in each parameter matrix.

7. A system for segmenting ultraviolet corona discharge images of transmission line insulators, characterized in that, The system includes: An image annotation module, configured to obtain multiple ultraviolet images of a target insulator under a discharge condition, and perform region annotation on each of the ultraviolet images, where the annotation result includes a discharge region and a background region; The difference model construction module is used to extract the first feature, the second feature, the third feature, and the fourth feature for the discharge area and the background area respectively, and construct a morphological difference model, a texture difference model, a spectral difference model, and a spatial difference model according to the first feature, the second feature, the third feature, and the fourth feature of the discharge area and the background area; The objective function construction module is used to construct an objective function according to the morphological difference model, the texture difference model, the spectral difference model, and the spatial difference model, optimize the objective function with the goal of maximizing the difference, and obtain the segmentation mask of the ultraviolet image according to the optimization result; The segmentation model construction module is used to input the segmentation mask and all the labeled ultraviolet images into the initial ultraviolet image segmentation model for training to obtain the final ultraviolet image segmentation model.

8. A storage medium, characterized in that, The storage medium stores one or more programs, which when executed by a processor, implement the transmission line insulator ultraviolet corona discharge image segmentation method according to any one of claims 1-6.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor, wherein: The memory is used to store a computer program; When the processor is used to execute the computer program stored on the memory, it implements the transmission line insulator ultraviolet corona discharge image segmentation method according to any one of claims 1-6.

Images