Insulator discharge fault position identification method based on image channel separation

Abstract

The invention discloses an insulator discharge fault position identification method based on image channel separation, and the method comprises the steps: obtaining an insulator original image in an outdoor working state, carrying out the image decomposition processing, carrying out the image enhancement processing according to an image decomposition result, and obtaining an insulator preprocessing image; carrying out color component feature analysis processing on the insulator preprocessing image, carrying out image channel separation processing on the insulator preprocessing image to obtain an insulator image channel with a single color component, calculating a color component weighting coefficient, and carrying out weighting processing on the corresponding insulator image channel to obtain an insulator image channel with a single color component; and carrying out image channel merging and image recombination processing on the weighted insulator image channel to obtain an insulator merged image, obtaining a target area of the insulator merged image, analyzing a discharge incidence relation between the target area and the ultraviolet light signal, and carrying out edge adjustment on the target area to obtain an insulator discharge fault position. The method has the effect of improving the identification accuracy of the insulator discharge fault.

Description

Technical Field

[0001] This application relates to the technical field of insulator fault identification, and in particular to a method for identifying the location of insulator discharge faults based on image channel separation. Background Technology

[0002] Currently, the most direct and fundamental cause of electrical equipment failures is the deterioration of equipment insulation performance. As a key electrical device in an electrical system, insulators operate under high voltage and strong field conditions for extended periods, while also being subjected to outdoor wind and rain. This inevitably leads to deterioration of insulation performance. When the deterioration reaches a certain level, discharge phenomena will occur. If insulator discharge phenomena are not detected in time, dust accumulation on the surface of the insulator during actual use can also cause excessively high local field strength, resulting in partial discharge of the insulator. Therefore, it is necessary to repair or replace faulty insulators in a timely manner. However, there is a risk of insulator breakdown and flashover. Therefore, it is necessary to promptly inspect and repair discharge faults in insulators during operation.

[0003] When an insulator discharges, it emits ultraviolet light with a wavelength range of 240-280 nm. Existing ultraviolet imaging techniques for insulator discharge detection often suffer from a significant impact on the number of discharge photons measured by the measurement distance. Furthermore, current ultraviolet imaging detection equipment lacks distance measurement capabilities, making it impossible to objectively evaluate the discharge intensity without knowing the specific distance. This leads to errors in identifying the location of discharge faults in insulators. There is room for further optimization in the aforementioned technologies to improve the accuracy of insulator discharge fault identification. Summary of the Invention

[0004] To address the problem of insufficient accuracy in identifying the location of insulator discharge faults in existing technologies, this invention provides an insulator discharge fault location identification method based on image channel separation, which can improve the accuracy of insulator discharge fault identification.

[0005] Firstly, the above-mentioned inventive objective of this application is achieved through the following technical solution: A method for identifying the location of insulator discharge faults based on image channel separation, the method comprising: Acquire the original image of the insulator under outdoor working conditions, perform image decomposition processing on the original image of the insulator, and perform image enhancement processing according to the image decomposition results to obtain the pre-processed image of the insulator. The preprocessed image of the insulator is subjected to color component feature analysis, and the image channel separation process is performed on the preprocessed image of the insulator according to each color feature component to obtain an insulator image channel with a single color component. Calculate the color component weighting coefficient for each color component feature, perform weighting processing on the corresponding insulator image channels, and perform image channel merging and image reconstruction processing on the weighted insulator image channels to obtain the merged insulator image. The target region of the merged image of the insulator is obtained, the discharge correlation between the target region and the ultraviolet light signal is analyzed, and the edge of the target region is adjusted according to the discharge correlation to obtain the location of the insulator discharge fault.

[0006] In a preferred embodiment, this application can be further configured as follows: performing color component feature analysis on the preprocessed insulator image, and performing image channel separation processing on the preprocessed insulator image according to each color feature component to obtain an insulator image channel with a single color component, specifically includes: The preprocessed insulator image is subjected to RGB color component feature analysis, and the preprocessed insulator image is decomposed into three image channels: R, G, and B, to obtain insulator image channels with single color components of R, G, and B.

[0007] In a preferred embodiment, this application can be further configured as follows: the process of calculating the color component weighting coefficients for each color component feature, weighting the corresponding insulator image channels, and performing image channel merging and image reconstruction on the weighted insulator image channels to obtain the merged insulator image, specifically includes: The image of the R channel is segmented into units, the weight coefficient of the gray value of the unit in the R channel is calculated, and the R channel image is processed by median filtering and histogram equalization to obtain the processed image of the R channel. The grayscale mean of the G channel image is calculated, and gamma correction and image sharpening are performed on the G channel image to obtain the G channel processed image; The interference threshold of the B channel image is calculated, and the B channel image is subjected to Gaussian filtering based on the interference threshold. The image interference area exceeding the interference threshold is filtered out to obtain the processed B channel image. The insulator image channels are weighted based on the processed images of the R channel, G channel, and B channel.

[0008] In a preferred embodiment, this application can be further configured as follows: the process of calculating the color component weighting coefficients for each color component feature, weighting the corresponding insulator image channels, and performing image channel merging and image reconstruction on the weighted insulator image channels to obtain the signal channel merging and image reconstruction process in the merged insulator image specifically includes: Based on the insulator position in the original insulator image, the weighted insulator image channels are merged. The boundary mean values ​​of the R, G, and B image channels are calculated and the boundary smoothing is performed to obtain the merged image. Obtain the boundary of the target region in the channel merged image from the B image channel, perform image reconstruction processing on the target region in the channel merged image, and obtain the merged image of the insulator with the target region highlighted.

[0009] In a preferred embodiment, this application can be further configured as follows: the process of acquiring the target region of the merged image of the insulator, analyzing the discharge correlation between the target region and the ultraviolet light signal, and adjusting the edge of the target region according to the discharge correlation to obtain the discharge fault location of the insulator, specifically includes: The target region in the merged image of the insulator at different discharge stages is obtained, and the boundary changes between the target region and the ultraviolet light signal range at each discharge stage are compared to dynamically adjust the interference threshold of the target region. Based on the interference threshold adjustment results, the boundary overlap state between the target area and the ultraviolet light signal range is dynamically adjusted, and the discharge correlation between the target area and the ultraviolet light signal is analyzed based on the boundary overlap state. Based on the discharge correlation, the boundary of the target area is adjusted, and the area range related to the ultraviolet light signal range of the target area is located to obtain the location of the insulator discharge fault.

[0010] In a preferred embodiment, this application can be further configured as follows: adjusting the target area boundary according to the discharge correlation, and performing location processing on the target area related to the ultraviolet light signal range to obtain the insulator discharge fault location, specifically includes: Based on the discharge correlation, a boundary threshold is calculated for the target area boundary, and the parameters of the target area range are adjusted based on the boundary threshold. The boundary threshold includes a confidence threshold and an intersection-exchange ratio threshold. The minimum distance between the target area and the ultraviolet light signal range is calculated and the relevant area range is located. The pixel coordinates of the relevant area range location information are transformed to obtain the location of the insulator discharge fault in the image.

[0011] In a preferred embodiment, this application can be further configured as follows: performing color component feature analysis on the preprocessed insulator image, and performing image channel separation processing on the preprocessed insulator image according to each color feature component to obtain an insulator image channel with a single color component, specifically includes: The pre-processed image of the insulator is subjected to CMYK color component feature analysis and processing, and the pre-processed image of the insulator is decomposed into four image channels of C, M, Y and K, respectively, to obtain insulator image channels of single color components of C, M, Y and K.

[0012] Secondly, the above-mentioned inventive objective of this application is achieved through the following technical solutions: An insulator discharge fault location identification system based on image channel separation, the system being applied to the aforementioned insulator discharge fault location identification method based on image channel separation, the method comprising: The data preprocessing module is used to acquire the original image of the insulator under outdoor working conditions, perform image decomposition processing on the original image of the insulator, and perform image enhancement processing according to the image decomposition results to obtain the preprocessed image of the insulator. The channel separation module is used to perform color component feature analysis on the preprocessed image of the insulator, and to perform image channel separation on the preprocessed image of the insulator according to each color feature component to obtain an insulator image channel with a single color component. The image reconstruction module is used to calculate the color component weighting coefficient of each color component feature, perform weighting processing on the corresponding insulator image channels, and perform image channel merging and image reconstruction processing on the weighted insulator image channels to obtain the merged insulator image. The location module is used to acquire the target area of ​​the merged image of the insulator, analyze the discharge correlation between the target area and the ultraviolet light signal, and adjust the edge of the target area according to the discharge correlation to obtain the location of the insulator discharge fault.

[0013] Thirdly, the above-mentioned objectives of this application are achieved through the following technical solutions: A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described insulator discharge fault location identification method based on image channel separation.

[0014] Fourthly, the above-mentioned objectives of this application are achieved through the following technical solutions: A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described insulator discharge fault location identification method based on image channel separation.

[0015] In summary, this application includes at least one of the following beneficial technical effects: 1. This application improves the visual difference between the target area and the normal area by separating the image channels of the original image of the insulator and adjusting the image channels of each color component accordingly. By combining the boundary changes between the target area boundary and the ultraviolet light signal range under different discharge states of the insulator, the boundary of the target area is finely adjusted to improve the correlation between the ultraviolet light discharge area and the blue signal target area of ​​the insulator. Thus, the location of the ultraviolet light discharge area is located by displaying the blue signal area in the image, which reduces the measurement error limited by the test distance in the insulator discharge location detection and improves the accuracy of insulator discharge fault identification.

[0016] 2. This application improves the particle swarm optimization algorithm to output the optimal weights for each image channel, thereby enhancing the accuracy of image enhancement and background interference suppression for each image channel. It can obtain a blue halo region with obvious distinction. By enhancing the features of the blue spectrum, it provides a reliable basis for identifying the ultraviolet range of insulator discharge. Attached Figure Description

[0017] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0018] Figure 1 This is a flowchart illustrating the implementation of the insulator discharge fault location identification method based on image channel separation in this embodiment.

[0019] Figure 2 This is a flowchart illustrating the implementation of step S30 of the insulator discharge fault location identification method in this embodiment.

[0020] Figure 3 This is a flowchart illustrating the image channel processing implementation of the insulator discharge fault location identification method in this embodiment.

[0021] Figure 4 This is a schematic diagram of the blue halo in the merged image of the insulator in this embodiment.

[0022] Figure 5 This is a flowchart illustrating the implementation of step S40 in the insulator discharge fault location identification method of this embodiment.

[0023] Figure 6 This is a comparison table showing the relationship between the blue halo and the ultraviolet light signal at different discharge stages in this embodiment.

[0024] Figure 7 This is a structural block diagram of the insulator discharge fault location identification system based on image channel separation in this embodiment.

[0025] Figure 8 This is a schematic diagram of the internal structure of a computer device used to implement a method for identifying the location of insulator discharge faults. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0028] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0029] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0030] In one embodiment, such as Figure 1 As shown, this application discloses a method for identifying the location of insulator discharge faults based on image channel separation, which specifically includes the following steps: S10: Acquire the original image of the insulator under outdoor working conditions, perform image decomposition processing on the original image of the insulator, and perform image enhancement processing according to the image decomposition results to obtain the preprocessed image of the insulator.

[0031] Specifically, the original images of insulators in outdoor working conditions are acquired using a high-definition industrial camera. The original images of insulators are then decomposed according to features such as brightness, chroma, and hue. Based on the decomposition results, each decomposed image is subjected to image enhancement processing such as noise reduction, color space conversion, and color component decomposition to obtain a pre-processed image of the insulator.

[0032] S20: Perform color component feature analysis on the preprocessed image of the insulator, and perform image channel separation processing on the preprocessed image of the insulator according to each color feature component to obtain an insulator image channel with a single color component.

[0033] Specifically, the RGB color component feature analysis is performed on the pre-processed image of the insulator, and the pre-processed image of the insulator is decomposed into three image channels: R, G, and B, to obtain the insulator image channels with single color components of R, G, and B.

[0034] In this embodiment, the pre-processed insulator image can also be subjected to CMYK color component feature analysis, decomposing the pre-processed insulator image into four image channels: C, M, Y, and K, to obtain insulator image channels with single color components of C, M, Y, and K. The method can be selected according to actual needs and is not limited to one method in this embodiment. The processing method for the insulator image channels is the same; this embodiment uses RGB color components to illustrate image channel separation and merging.

[0035] S30: Calculate the color component weighting coefficient for each color component feature, perform weighting processing on the corresponding insulator image channels, and perform image channel merging and image reconstruction processing on the weighted insulator image channels to obtain the merged insulator image.

[0036] Specifically, such as Figure 2 As shown, step S30 includes: S301: Perform cell segmentation on the R channel image, calculate the cell grayscale weight coefficient of the R channel, and perform median filtering and histogram equalization on the R channel image to obtain the processed R channel image.

[0037] Specifically, step S301 includes: S3011: Divide the image into sub-regions in the R channel. Select the sub-region size according to the image size; options include 8×8, 16×16, and 32×32. Each sub-region is treated as an independent processing unit. For each sub-region, count its gray-level frequency. The gray-level value of the discharge region in the R channel is significantly higher than that of the normal region. Enhance the proportion of high gray-level values ​​in the histogram to obtain the histogram. The expression for calculating the weighting coefficient of the gray-level value is as follows: (1) (2) in, For grayscale value equal to The corresponding weight function, The grayscale value of a pixel. To enhance the coefficient, For the set enhancement threshold, for Weighted frequency For grayscale value equal to The number of pixels.

[0038] S3012: For each sub-region, pixels are redistributed, and the portion of pixels with values ​​greater than the cropping threshold is extracted. The cropping threshold is determined based on the average gray level within the sub-region. The total number of extracted pixels is counted, and they are redistributed evenly across the gray levels of the histogram. The expression for the cropping threshold is as follows: (3) in, To set a cropping threshold, limit the maximum frequency of a certain gray level in the histogram to prevent that gray level from being over-enhanced. for The highest frequency in and All are threshold proportional coefficients. The average gray level, The upper limit of the grayscale value is determined based on sample statistics. For example, the R value of a normal porcelain area is usually no greater than 100, so it is set as follows: ,when At this time, it can be determined as a potential discharge zone. This allows for higher frequency retention, avoiding the removal of high grayscale features from the discharge region. At that time, it can be determined as the background area. Strictly limit high-frequency noise, such as tiny bright spot noise, to reduce noise amplification.

[0039] S3013: Perform histogram equalization on the cropped region, and use the pixel value of the center of the equalized sub-block as the reference point. Calculate and output the median value of each pixel in the image using bilinear interpolation to perform median filtering on the R-channel image. Perform histogram equalization based on the median filtering result to obtain the processed R-channel image.

[0040] S302: Calculate the grayscale mean of the G channel image, perform gamma correction and image sharpening on the G channel image, and obtain the G channel processed image.

[0041] Specifically, the image pixel values ​​of the G channel are normalized to [0, 1] to eliminate the influence of absolute gray value differences. The gray value mean of the normalized pixel values ​​is calculated, and then gamma correction is performed based on the calculated gray value mean. The corrected image pixel values ​​of the G channel are mapped from [0, 1] back to [0, 255]. The image sharpening process uses a second-order Laplacian operator convolution kernel for edge enhancement. The mapped image is convolved to obtain the sharpened image. The weighted coefficients calculated from multiple sets of experimental data are used to weight and fuse the gamma-corrected image and the convolutional image to obtain the G channel processed image. This process retains the overall brightness improvement after gamma correction, while also superimposing edge details and amplifying the difference between the discharge area and the normal area.

[0042] S303: Calculate the interference threshold for the B channel image, perform Gaussian filtering on the B channel image based on the interference threshold, and filter out image interference areas that exceed the interference threshold to obtain the processed B channel image.

[0043] Specifically, the interference threshold is determined by the mean and standard deviation of the grayscale values ​​of the B-channel image. Gaussian filtering is then applied to the B-channel image based on this threshold. Areas that are too bright or too dark exceeding the threshold are marked as interference regions. These areas are then filtered out to obtain the processed B-channel image. Specifically, the grayscale values ​​of excessively dark areas (such as shadow areas) are significantly lower than those of effective areas (such as discharge areas), while the grayscale values ​​of excessively bright areas (such as metallic reflective areas) are significantly higher than those of effective areas. The specific formula is as follows: (4) (5) in, and These are the lower threshold and the upper threshold for truncation, respectively. The average grayscale value of the B channel image. and All are adjustment coefficients. The higher the value, the more shadow information is retained; it needs to be adjusted according to the shadow intensity. The higher the value, the more shadow information is retained; it needs to be adjusted according to the shadow intensity. is the standard deviation of the grayscale values ​​of the B channel image.

[0044] S304: Weight the insulator image channels based on the processed images of the R channel, G channel, and B channel.

[0045] Specifically, the corresponding RGB channels are subjected to targeted enhancement and suppression processing according to the R channel processed image, G channel processed image, and B channel processed image to obtain a weighted image of the insulator image channels.

[0046] Specifically, in step S30, the weighted insulator image channels are merged and reconstructed to obtain a merged insulator image, such as... Figure 3 As shown, it specifically includes: S305: Based on the insulator position in the original insulator image, perform image channel merging processing on the weighted insulator image channels, calculate the boundary mean of the R, G, and B image channels, perform boundary smoothing processing, and obtain the channel merged image.

[0047] Specifically, the position of the insulator in the original image is determined based on the pixel differences between the bright and dark areas and the pre-set insulator shape. Using the insulator position as the location, the weighted RGB insulator image channels are merged. The pixel mean of the insulator image boundary of the RGB image channels is calculated. The insulator boundary pixels of the merged image are then subjected to mean filtering to smooth the noise in the merged insulator boundary pixels, resulting in a merged channel image.

[0048] S306: Obtain the boundary of the target region in the channel merged image of the B image channel, perform image reconstruction processing on the target region in the channel merged image, and obtain the insulator merged image with the target region highlighted.

[0049] Specifically, step S306 includes: S3061: Select several samples from the original images of insulators at different discharge stages. The proportion of normal samples and fault samples in the samples is the same. Manually label the discharge fault area, with the fault area as 1 and the non-fault area as 0. The samples are divided into training set and validation set.

[0050] S3062: Initialize the number of particles, spatial dimension, position of each particle, maximum inertia weight, minimum inertia weight, individual learning factor, group learning factor, and maximum number of iterations. The spatial dimension is three, and the position vector of each particle represents the weight coefficients of the R, G, and B channels, respectively.

[0051] S3063: Perform image channel separation and weighted merging processing on the samples in each training set, resynthesize the image according to the weight coefficients corresponding to the positions of each particle, output the reconstructed image, and use the overlap rate between the fault region in the reconstructed image and the fault region in the original image as the objective function.

[0052] S3064: Iteratively updates the particle's velocity and position, introducing adaptive inertia weights and a sine / cosine search algorithm. The adaptive inertia weights decrease as the number of iterations increases. After updating the particle position, a sine / cosine search algorithm is added, with the corresponding formula as follows: (6) (7) (8) (9) (10) in, For the number of iterations, The maximum number of iterations, For the first The first particle Vi in the Speed ​​at the next iteration For the first Inertia weights in the next iteration For the first The first particle Vi in the Speed ​​at the next iteration For individual learning factors, As a group learning factor, and All are random numbers within the interval [0,1]. For the first The particle in the first The optimal position of an individual in dimension. The globally optimal position is at the th Dimension value, For the first The first particle Vi in the Position at the next iteration For the first The first particle Vi in the Position at the next iteration For adaptive parameters, It decreases as the number of iterations increases. [0, 2] Random angles on ] and All are random numbers uniformly distributed on [0,1]. To set a constant. S3065: Determine the termination condition; stop when the maximum number of iterations is reached, thus determining the region boundary of the target area of ​​the B image channel in the channel-merged image. Reassemble the adjusted RGB channel images according to the region boundary of the target area to obtain a color image, and output the reassembled insulator merged image, such as... Figure 4 As shown, the color difference between the blue halo area and the normal area is significantly enhanced at this time.

[0053] S40: Obtain the target area of ​​the merged image of the insulator, analyze the discharge correlation between the target area and the ultraviolet light signal, adjust the edge of the target area according to the discharge correlation, and obtain the location of the insulator discharge fault.

[0054] Specifically, such as Figure 5 As shown, step S40 includes: S401: Obtain the target region in the merged image of the insulator at different discharge stages, compare the boundary changes of the target region and the ultraviolet light signal range at each discharge stage, and dynamically adjust the interference threshold of the target region.

[0055] Specifically, based on the merged images of insulators at different discharge stages, including the no-discharge stage, the corona initiation stage, the blue halo stage, and the flashover stage, the comparison of the relationship between the blue halo and the ultraviolet light signal at different discharge stages is as follows: Figure 6 As shown, the area containing the blue halo is taken as the target area. By comparing the boundary changes between the target area and the ultraviolet light signal range at each discharge stage, the interference threshold of the target area is analyzed, and the interference area is filtered out by adjusting the interference threshold. In this embodiment, the target area is determined by improving the YOLOv8 model, as follows: S4011: Using the CSPDarknet backbone network of the YOLOv8 model, the C2f module is used to extract feature maps at five scales (80×80×256, 40×40×512, 20×20×1024, 10×10×2048, 5×5×4096).

[0056] S4012: Construct an improved PAN-FPN structure. The top-down path uses 1×1 convolution to reduce the dimensionality of high-level features (5×5×4096) and then fuses them with mid-level features (10×10×2048). The bottom-up path uses 3×3 convolution to increase the dimensionality of low-level features (80×80×256) and then fuse them with mid-level features (40×40×512).

[0057] S4013: Introduces a spatial attention module, adding a spatial attention mechanism after the fused feature maps of three detection scales (80×80×256, 40×40×512, 20×20×1024) to enhance the feature response of the discharge region.

[0058] S4014: Output feature map, which is then compressed by convolution to obtain the detection scale feature map and determine the target region. In this embodiment, the target region is the blue halo region.

[0059] S402: Based on the interference threshold adjustment results, dynamically adjust the boundary overlap state between the target area boundary and the ultraviolet light signal range, and analyze the discharge correlation between the target area and the ultraviolet light signal based on the boundary overlap state.

[0060] Specifically, signal correlation prediction is performed using a decoupled detection head based on the YOLOv8 model, and discharge correlation analysis is conducted by combining a dynamic target allocation mechanism with an improved loss function optimization model. This includes: S4021: A decoupled detection head using the YOLOv8 model, divided into a classification branch and a regression branch. The classification branch outputs the overlap probability between the target area and the ultraviolet light signal, while the regression branch outputs the offset of the overlapping bounding box and the target confidence level.

[0061] S4022: Dynamic target allocation adopts a dynamic allocation mechanism for blue halo detection of target regions in the single discharge stage of the task alignment, and matches the best predicted box for each ground truth box according to the product of the classification score and the cross-union ratio.

[0062] S4023: In the loss function, the classification loss uses the Focal Loss function, with weights added to the blue halo range. Regression loss compensation is applied to the overlapping frame between the target region and the ultraviolet signal region based on the loss function, thereby constructing the discharge correlation between the target region and the ultraviolet signal. The regression loss uses the complete intersection-union loss, with the corresponding formula as follows: (11) in, This is the total loss value of the model, used to measure the difference between the predicted results and the true labels, and to guide the optimization of model parameters. The classification loss weighting coefficient adjusts the proportion of classification loss in the total loss. Focal Loss is used to address fault class imbalance issues, such as a shortage of arc discharge samples and an abundance of corona discharge samples, thereby enhancing the focus on hard-to-classify samples. The regression loss weighting coefficient adjusts the proportion of the bounding box regression loss in the total loss. Because the discharge location accuracy requirement is high, its weight is larger. The full intersection-union ratio loss is used to optimize the accuracy of bounding box coordinate prediction, taking into account bounding box overlap, center distance, and aspect ratio.

[0063] S403: Adjust the boundary of the target area according to the discharge correlation, perform positioning processing on the area range of the target area related to the ultraviolet light signal range, and obtain the location of the insulator discharge fault.

[0064] Specifically, step S403 includes: S4031: Calculate the boundary threshold of the target area based on the discharge correlation, and adjust the parameters of the target area range based on the boundary threshold. The boundary threshold includes the confidence threshold and the cross-parallel ratio threshold. Specifically, based on the discharge correlation, the boundary threshold of the target area boundary, i.e. the blue halo range boundary value, is calculated through regression compensation to fit the blue halo range boundary with the ultraviolet light signal range of the corresponding discharge stage. The confidence threshold and cross-parallel ratio (CPAR) threshold are adjusted using a non-maximum suppression algorithm. The confidence threshold is 0.6 (arc), 0.5 (spark), and 0.4 (corona), and the CPAR threshold is 0.3.

[0065] S4032: Calculate the minimum distance between the target area and the ultraviolet light signal range and perform relevant area range positioning. Perform pixel coordinate transformation on the relevant area range positioning information to obtain the insulator discharge fault location in the image.

[0066] Specifically, the minimum distance between the boundary of the ultraviolet light signal range and the boundary of the blue halo range is calculated through spatial constraints. Data with a minimum distance greater than the edge of the insulator image are removed, and the insulator range within the blue halo range is located. The boundary normalized coordinates in the range location information are converted into pixel coordinates, and the location, type, and confidence level of the insulator discharge fault in the image are output. At the same time, a labeled visualization image is generated.

[0067] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0068] In one embodiment, an insulator discharge fault location identification system based on image channel separation is provided. This system corresponds one-to-one with the insulator discharge fault location identification method based on image channel separation described in the above embodiments. Figure 7 As shown, the insulator discharge fault location identification system based on image channel separation includes a data preprocessing module, a channel separation module, an image reconstruction module, and a location positioning module. Detailed descriptions of each functional module are as follows: The data preprocessing module is used to acquire the original image of the insulator under outdoor working conditions, perform image decomposition processing on the original image of the insulator, and perform image enhancement processing according to the image decomposition results to obtain the preprocessed image of the insulator.

[0069] The channel separation module is used to perform color component feature analysis on the preprocessed image of the insulator. It separates the image channels of the preprocessed image of the insulator according to each color feature component to obtain an insulator image channel with a single color component.

[0070] The image reconstruction module is used to calculate the color component weighting coefficient of each color component feature, perform weighting processing on the corresponding insulator image channels, and perform image channel merging and image reconstruction processing on the weighted insulator image channels to obtain the merged insulator image.

[0071] The location module is used to acquire the target area of ​​the merged image of the insulator, analyze the discharge correlation between the target area and the ultraviolet light signal, and adjust the edge of the target area according to the discharge correlation to obtain the location of the insulator discharge fault.

[0072] Preferably, the channel separation module specifically includes: The preprocessed insulator image is subjected to RGB color component feature analysis, decomposing it into three image channels: R, G, and B, to obtain insulator image channels with single R, G, and B color components; or, The pre-processed image of the insulator is subjected to CMYK color component feature analysis and processing. The pre-processed image of the insulator is decomposed into four image channels: C, M, Y, and K, respectively, to obtain the insulator image channels with single color components of C, M, Y, and K.

[0073] Preferably, the weighted processing of the insulator image channels in the image reconstruction module specifically includes: The R-channel image processing submodule is used to perform cell segmentation on the R-channel image, calculate the cell grayscale weight coefficients of the R-channel, and perform median filtering and histogram equalization on the R-channel image to obtain the processed R-channel image.

[0074] The G-channel image processing submodule is used to calculate the grayscale mean of the G-channel image, perform gamma correction and image sharpening on the G-channel image, and obtain the G-channel processed image.

[0075] The B-channel image processing submodule calculates the interference threshold of the B-channel image, performs Gaussian filtering on the B-channel image based on the interference threshold, and filters out image interference areas that exceed the interference threshold to obtain the processed B-channel image.

[0076] The image weighting processing submodule is used to perform weighted processing on the insulator image channels based on the processed images of the R channel, G channel, and B channel.

[0077] Preferably, the signal channel merging and image reconstruction process in the image reconstruction module specifically includes: The channel merging submodule is used to perform image channel merging processing on the weighted insulator image channels based on the insulator position in the original insulator image, calculate the boundary mean of the R, G, and B image channels of the insulator image and perform boundary smoothing processing to obtain the channel merged image.

[0078] The image reconstruction submodule is used to obtain the region boundary of the target region in the B image channel in the channel-merged image, perform image reconstruction processing on the target region in the channel-merged image, and obtain the insulator merged image with the target region highlighted.

[0079] Preferably, the analysis process of the discharge correlation between the target area and the ultraviolet light signal in the location positioning module specifically includes: The region determination submodule is used to acquire the target region in the merged image of the insulator at different discharge stages, compare the boundary changes of the target region with the ultraviolet light signal range at each discharge stage, and dynamically adjust the interference threshold of the target region.

[0080] The correlation analysis submodule is used to dynamically adjust the boundary overlap state between the target area boundary and the ultraviolet light signal range based on the interference threshold adjustment results, and analyze the discharge correlation between the target area and the ultraviolet light signal based on the boundary overlap state.

[0081] The fault location submodule is used to adjust the boundary of the target area according to the discharge correlation, and to perform location processing on the area range of the target area related to the ultraviolet light signal range to obtain the location of the insulator discharge fault.

[0082] Preferably, the fault location submodule specifically includes: The parameter adjustment unit is used to calculate the boundary threshold of the target area based on the discharge correlation, and adjust the parameters of the target area range based on the boundary threshold. The boundary threshold includes a confidence threshold and an intersection-to-exchange ratio threshold.

[0083] The location unit is used to calculate the minimum distance between the target area and the ultraviolet light signal range and to perform relevant area range positioning. It performs pixel coordinate transformation on the relevant area range positioning information to obtain the location of the insulator discharge fault in the image.

[0084] Specific limitations regarding the insulator discharge fault location identification system based on image channel separation can be found in the limitations of the insulator discharge fault location identification method based on image channel separation mentioned above, and will not be repeated here. Each module in the aforementioned insulator discharge fault location identification system based on image channel separation can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in hardware or independently of the processor in a computer device, or stored in software in the memory of a computer device, so that the processor can call and execute the corresponding operations of each module.

[0085] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 8As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computational and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and the database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores image processing data for insulator discharge fault location identification. The network interface communicates with external terminals via a network connection. When executed by the processor, the computer program implements an insulator discharge fault location identification method based on image channel separation.

[0086] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of a method for identifying the location of an insulator discharge fault based on image channel separation.

[0087] Those skilled in the art will recognize that the units of the various examples described in connection with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of the invention.

[0088] In the embodiments provided by the present invention, it should be understood that the division of units is only a logical functional division. In actual implementation, there may be other division methods, such as multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored.

[0089] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0090] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the 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 to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0091] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A method for identifying the location of insulator discharge faults based on image channel separation, characterized in that, The method includes: Acquire the original image of the insulator under outdoor working conditions, perform image decomposition processing on the original image of the insulator, and perform image enhancement processing according to the image decomposition results to obtain the pre-processed image of the insulator. The preprocessed image of the insulator is subjected to color component feature analysis, and the image channel separation process is performed on the preprocessed image of the insulator according to each color feature component to obtain an insulator image channel with a single color component. Calculate the color component weighting coefficient for each color component feature, perform weighting processing on the corresponding insulator image channels, and perform image channel merging and image reconstruction processing on the weighted insulator image channels to obtain the merged insulator image. The target region of the merged image of the insulator is obtained, the discharge correlation between the target region and the ultraviolet light signal is analyzed, and the edge of the target region is adjusted according to the discharge correlation to obtain the location of the insulator discharge fault.

2. The insulator discharge fault location identification method based on image channel separation according to claim 1, characterized in that, The step of performing color component feature analysis on the preprocessed insulator image, and separating the image channel according to each color feature component to obtain an insulator image channel with a single color component, specifically includes: The preprocessed insulator image is subjected to RGB color component feature analysis, and the preprocessed insulator image is decomposed into three image channels: R, G, and B, to obtain insulator image channels with single color components of R, G, and B.

3. The insulator discharge fault location identification method based on image channel separation according to claim 2, characterized in that, The process of calculating the weighting coefficients for each color component feature, weighting the corresponding insulator image channels, and then merging and reconstructing the weighted insulator image channels to obtain the merged insulator image specifically includes: The image of the R channel is segmented into units, the weight coefficient of the unit gray value of the R channel is calculated, and the R channel image is processed by median filtering and histogram equalization to obtain the processed R channel image. The grayscale mean of the G channel image is calculated, and gamma correction and image sharpening are performed on the G channel image to obtain the G channel processed image; The interference threshold of the B channel image is calculated, and the B channel image is subjected to Gaussian filtering based on the interference threshold. The interference regions of the image exceeding the interference threshold are filtered out to obtain the processed B channel image. The insulator image channels are weighted based on the processed images of the R channel, G channel, and B channel.

4. The insulator discharge fault location identification method based on image channel separation according to claim 3, characterized in that, The process of calculating the color component weighting coefficients for each color component feature, weighting the corresponding insulator image channels, and then performing image channel merging and image reconstruction on the weighted insulator image channels to obtain the signal channel merging and image reconstruction process in the merged insulator image specifically includes: Based on the insulator position in the original insulator image, the weighted insulator image channels are merged. The boundary mean values ​​of the R, G, and B image channels are calculated and the boundary smoothing is performed to obtain the merged image. Obtain the boundary of the target region in the channel merged image from the B image channel, perform image reconstruction processing on the target region in the channel merged image, and obtain the merged image of the insulator with the target region highlighted.

5. The insulator discharge fault location identification method based on image channel separation according to claim 1, characterized in that, The process of acquiring the target region of the merged image of the insulator, analyzing the discharge correlation between the target region and the ultraviolet light signal, and adjusting the edges of the target region based on the discharge correlation to obtain the discharge fault location of the insulator, specifically includes: The target region in the merged image of the insulator at different discharge stages is obtained, and the boundary changes between the target region and the ultraviolet light signal range at each discharge stage are compared to dynamically adjust the interference threshold of the target region. Based on the interference threshold adjustment results, the boundary overlap state between the target area and the ultraviolet light signal range is dynamically adjusted, and the discharge correlation between the target area and the ultraviolet light signal is analyzed based on the boundary overlap state. Based on the discharge correlation, the boundary of the target area is adjusted, and the area range related to the ultraviolet light signal range of the target area is located to obtain the location of the insulator discharge fault.

6. The insulator discharge fault location identification method based on image channel separation according to claim 5, characterized in that, The step of adjusting the target area boundary based on the discharge correlation, and locating the area related to the ultraviolet light signal range within the target area to obtain the insulator discharge fault location specifically includes: Based on the discharge correlation, a boundary threshold is calculated for the target area boundary, and the parameters of the target area range are adjusted based on the boundary threshold. The boundary threshold includes a confidence threshold and an intersection-exchange ratio threshold. The minimum distance between the target area and the ultraviolet light signal range is calculated and the relevant area range is located. The pixel coordinates of the relevant area range location information are transformed to obtain the location of the insulator discharge fault in the image.

7. The insulator discharge fault location identification method based on image channel separation according to claim 1, characterized in that, The step of performing color component feature analysis on the preprocessed insulator image, and separating the image channel according to each color feature component to obtain an insulator image channel with a single color component, specifically includes: The pre-processed image of the insulator is subjected to CMYK color component feature analysis and processing, and the pre-processed image of the insulator is decomposed into four image channels of C, M, Y and K, respectively, to obtain insulator image channels of single color components of C, M, Y and K.

8. An insulator discharge fault location identification system based on image channel separation, characterized in that, The system is applied to the insulator discharge fault location identification method based on image channel separation according to any one of claims 1-7, the method comprising: The data preprocessing module is used to acquire the original image of the insulator under outdoor working conditions, perform image decomposition processing on the original image of the insulator, and perform image enhancement processing according to the image decomposition results to obtain the preprocessed image of the insulator. The channel separation module is used to perform color component feature analysis on the preprocessed image of the insulator, and to perform image channel separation on the preprocessed image of the insulator according to each color feature component to obtain an insulator image channel with a single color component. The image reconstruction module is used to calculate the color component weighting coefficient of each color component feature, perform weighting processing on the corresponding insulator image channels, and perform image channel merging and image reconstruction processing on the weighted insulator image channels to obtain the merged insulator image. The location module is used to acquire the target area of ​​the merged image of the insulator, analyze the discharge correlation between the target area and the ultraviolet light signal, and adjust the edge of the target area according to the discharge correlation to obtain the location of the insulator discharge fault.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the insulator discharge fault location identification method based on image channel separation as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the insulator discharge fault location identification method based on image channel separation as described in any one of claims 1 to 7.

Images