A ring network line ice melting method, system, device and medium based on switch control

Abstract

The application discloses a kind of ring network line ice-melting method, system, equipment and medium based on switch control, belong to power grid deicing technical field, including the target image of each icing line is gray scale;Based on fuzzy region division, the gray-scale target image is binarized;Based on vertical projection method, the target ice thickness of each icing line is obtained by analyzing the ice thickness of binarized target image;Determine the power supply importance coefficient of icing line, based on power supply importance coefficient and target ice thickness, analyze the priority of each icing line;Based on the length and wire diameter of each icing line, determine the corresponding ice-melting voltage;Determine target ice-melting mode, determine switch action information based on priority and target ice-melting mode, based on switch action information, combine ice-melting voltage to control multifunctional ice-melting distribution transformer to icing line for ice-melting treatment.The application is based on the multi-power characteristics of ring network, while ensuring the line ice-melting effect, improve the ice-melting efficiency of line.

Description

Technical Field

[0001] This invention relates to the field of power grid de-icing technology, specifically to a method, system, equipment, and medium for de-icing ring network lines based on switch control. Background Technology

[0002] Ice accumulation on distribution network lines is a major hidden danger to the safe operation of power systems. Ice accumulation increases the weight of conductors, causing them to sag, and may even lead to serious accidents such as line breaks and collapses. Therefore, timely de-icing of distribution network lines is of paramount importance.

[0003] However, traditional line de-icing technologies have several limitations. First, traditional line de-icing solutions only support unidirectional de-icing, and the effective de-icing distance is limited, resulting in low de-icing efficiency. Second, in multi-segment de-icing scenarios of ring network lines, traditional unidirectional de-icing solutions require a large number of de-icing transformers, which not only incurs high investment costs but also fails to fully utilize the multi-directional power supply characteristics of ring network lines, causing serious resource waste. Furthermore, traditional de-icing solutions lack flexibility in parameter settings, failing to meet the differentiated de-icing needs of multi-branch lines, further limiting their application scope and effectiveness.

[0004] Meanwhile, the current ice thickness of the line is mainly analyzed by traditional edge detection operators. However, this method is sensitive to image noise, resulting in insufficient accuracy of the analyzed ice thickness and affecting the reliability of the de-icing parameter analysis, thus causing the line de-icing to fail to achieve the expected results. Summary of the Invention

[0005] In view of the above-mentioned problems, the present invention provides a method, system, equipment and medium for de-icing ring network lines based on switch control.

[0006] Therefore, the purpose of this invention is to overcome the shortcomings of the prior art and effectively improve the de-icing efficiency of the line while ensuring the de-icing effect.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for de-icing ring network lines based on switch control, comprising, Target images of each icing line are acquired, and icing thickness is analyzed to obtain the target icing thickness of each icing line. The power supply importance coefficient of each icing line is determined, and the priority of each icing line is analyzed based on the power supply importance coefficient and the target icing thickness. The corresponding de-icing voltage is determined based on the length and conductor diameter of each icing line. The target de-icing mode is determined, and the switch action information is determined based on the priority and the target de-icing mode. The icing process is performed on the icing line based on the switch action information and the de-icing voltage.

[0008] As a preferred embodiment of the switch-controlled ring network line de-icing method of the present invention, wherein: the ice thickness analysis of the target image includes: Collect target images of each icing line and calculate the information entropy of the color channels of the target images; The sensitivity coefficient of each color channel is calculated based on the information entropy. The target image is converted to grayscale based on the sensitivity coefficient, and the initial grayscale value is calculated to obtain the grayscale target image.

[0009] As a preferred embodiment of the ring network line de-icing method based on switch control described in this invention, the step of analyzing the ice thickness of the target image further includes: The grayscale target image is divided into different image blocks, and the edge gradient of the image blocks is calculated based on the Sobel operator to obtain the target edge gradient. The image patch is enhanced based on the target edge gradient to obtain an enhanced image patch, and the enhanced image patch is fused to generate an enhanced grayscale target image. Obtain the grayscale range of the enhanced grayscale target image, and divide the enhanced grayscale target image into different blurred regions based on the grayscale range; Based on information entropy, grayscale threshold analysis is performed on the fuzzy region to obtain the target grayscale threshold. The blurred region is binarized based on the target grayscale threshold to obtain a binarized blurred region, and the binarized target image is determined based on the binarized blurred region.

[0010] As a preferred embodiment of the ring network line de-icing method based on switch control described in this invention, the enhancement process includes: Extract the maximum and minimum gray values ​​of each image block, and calculate the weighting factor based on the target edge gradient to determine the weighted comparison coefficient of the image block; The image patch is enhanced based on the weighted comparison coefficient to obtain the enhanced image patch; The enhanced image blocks are stitched together to obtain the enhanced grayscale target image.

[0011] As a preferred embodiment of the switch-controlled ring network line de-icing method described in this invention, the ice thickness analysis includes: The vertical projection curve of the binarized target image is analyzed to obtain the target vertical projection curve; The pixel width of the icing line is determined based on the target vertical projection curve; Based on the pixel width, the first thickness estimate of the icing line is obtained using the vertical projection method; The second thickness estimate of the icing line is obtained based on the edge gradient method; The first thickness estimate and the second thickness estimate are fused to determine the final target icing thickness.

[0012] As a preferred embodiment of the switch-controlled ring network line de-icing method described in this invention, the step of analyzing the priority of each iced line includes: Obtain the load level of the icing line, and determine the power supply importance coefficient of the icing line based on the load level; The vulnerability index of icing lines is obtained by weighted summation of the length of the icing line and the conductor diameter of the icing line. The icing lines are weighted and fused based on the vulnerability index, power supply importance coefficient, and target icing thickness to obtain the priority of each icing line.

[0013] As a preferred embodiment of the de-icing method for ring network lines based on switch control described in this invention, the de-icing process includes: Determine the basic de-icing voltage requirement based on the line length and conductor diameter, and calculate the target de-icing voltage. Determine the target de-icing voltage, and determine the switch switching direction and switch closing state based on the priority; Based on the switching direction of the disconnector and the closing state of the switch, switch action information is generated to determine the dynamic de-icing sequence; The ice-covered lines are de-iced based on switch action information and the dynamic de-icing sequence determined.

[0014] This invention provides a ring network line de-icing system based on switch control.

[0015] To solve the above technical problems, the present invention provides the following technical solution: a ring network line de-icing system based on switch control, comprising: an image grayscale module acquiring target images of each iced ring network line, and converting the target images to grayscale to obtain grayscale target images; The image binarization module performs binarization processing on the grayscale target image based on the fuzzy region division to obtain a binarized target image; The ice thickness analysis module analyzes the ice thickness of the binary target image based on the vertical projection method to obtain the target ice thickness of each ice-covered line. The priority analysis module determines the power supply importance coefficient of each icing line and analyzes the priority of each icing line based on the power supply importance coefficient and the target icing thickness. The de-icing voltage analysis module determines the corresponding de-icing voltage based on the length of each iced line and the diameter of the conductor. The line melting module determines the target de-icing mode, and determines the switch action information based on the priority and the target de-icing mode. Based on the switch action information and the de-icing voltage, it controls the multi-functional de-icing distribution transformer to perform de-icing treatment on the iced line.

[0016] The present invention provides a computer device, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the described method for de-icing a ring network line based on switch control.

[0017] The present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the aforementioned method for de-icing a ring network line based on switch control.

[0018] The beneficial effects of this invention are as follows: In the embodiments of this invention, the target image is converted to grayscale, and the grayscale target image is binarized based on the fuzzy region division, thereby improving the binarization effect of the image. Ice thickness analysis is performed on the binarized target image using the vertical projection method, making the obtained target ice thickness more accurate. The power supply importance coefficient of each icing line is determined, and the priority of each icing line is analyzed based on the power supply importance coefficient and the target ice thickness, improving the reliability of priority analysis by analyzing the priority of icing lines from multiple aspects. The corresponding de-icing voltage is determined based on the length and conductor diameter of each icing line; the target de-icing mode is determined, and the switch action information is determined based on the priority and the target de-icing mode. Based on the switch action information and the de-icing voltage, a multi-functional de-icing distribution transformer is controlled to perform de-icing treatment on the icing lines, enabling orderly de-icing control of ring network lines, effectively improving the de-icing efficiency of the lines while ensuring the de-icing effect. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 The present invention provides an overall flowchart of a ring network line de-icing method based on switch control, which is an embodiment of the present invention.

[0021] Figure 2This is a schematic diagram of the structural composition of a switch-controlled ring network line de-icing device, which is provided as an embodiment of the present invention for a switch-controlled ring network line de-icing method.

[0022] Figure 3 This is a schematic diagram of the layout of a multifunctional de-icing distribution transformer and switch for a ring network line de-icing method based on switch control, provided in one embodiment of the present invention.

[0023] Figure 4 This is a circuit diagram of the forward de-icing mode of a ring network line de-icing method based on switch control, provided as an embodiment of the present invention.

[0024] Figure 5 This is a circuit diagram illustrating the reverse de-icing mode of a ring network line de-icing method based on switch control, provided as an embodiment of the present invention. Detailed Implementation

[0025] To make the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0026] Example 1, referring to Figures 1-5 This is one embodiment of the present invention, which provides a method for de-icing ring network lines based on switch control, comprising: S100. Collect target images of each icing line, and convert the target images to grayscale to obtain grayscale target images; S200. The grayscale target image is binarized to obtain a binarized target image; S300. Perform ice thickness analysis on the binarized target image to obtain the target ice thickness of each icing line. S400. Determine the power supply importance coefficient of each icing line, and analyze the priority of each icing line based on the power supply importance coefficient and the target icing thickness. S500. Determine the corresponding de-icing voltage based on the length of each icing line and the diameter of the conductor. S600. Determine the target de-icing mode, and determine the switch action information based on the priority and the target de-icing mode, and perform de-icing treatment on the iced line based on the switch action information and the de-icing voltage.

[0027] Therefore, to address the aforementioned problems, steps S100-S600 effectively solve the issues of limited unidirectional melting distance, low melting efficiency, and significant resource waste caused by the underutilization of the multidirectional power supply characteristics of ring network lines in traditional line de-icing technologies. Simultaneously, it resolves the shortcomings of traditional edge detection operators, which are susceptible to image noise interference, leading to insufficient accuracy in ice thickness analysis. Furthermore, through differentiated de-icing parameter configuration and priority scheduling, it meets the personalized de-icing needs of multi-branch lines, thereby effectively improving the de-icing efficiency of ring network lines and reducing de-icing costs while ensuring the de-icing effect. It should be noted that all thresholds in this invention can be obtained through expert experience and experiments.

[0028] Example 2, an embodiment of the present invention, provides a method for de-icing ring network lines based on switch control, based on the previous embodiment, comprising: In this embodiment of the invention, step S100 involves acquiring target images of each icing line, converting the target images to grayscale to obtain grayscale target images, and includes the following steps S101-S102: S101. Acquire target images of each icing line, calculate the information entropy of the color channels of the target images, and calculate the sensitivity coefficient of each color channel based on the information entropy. Specifically: Target images of each icing line are acquired using an image acquisition device, and the sensitivity coefficients of each color channel of the target image are obtained. These sensitivity coefficients are adaptively determined based on the information entropy of each color channel and the separability of icing features. Specifically: Calculate the information entropy of each color channel (R, G, B, red, green, blue) of the target image; perform inter-class separability analysis between the icing region and the background region for each color channel to obtain the icing feature separability index for each color channel. The sensitivity coefficients of each color channel are calculated based on the information entropy and the separability index of icing features. It should be noted that the entropy of each color channel is: The separability index for icing characteristics is: Sensitivity coefficient for: in, For color channels Information entropy For color channels medium grayscale value The probability of occurrence For color channels The separability index of icing characteristics , In the color channels of the icy area and the background area respectively The mean gray level in the image. , These represent the variances of the icy region and the background region in color channel c, respectively. For color channels The sensitivity coefficient, This indicates that the summation range includes the R, G, and B color channels. The number of gray levels (usually 256); i is just a variable index and is a positive integer.

[0029] S102. Based on the sensitivity coefficient, the target image is converted to grayscale, and an initial grayscale value is calculated to obtain the grayscale target image. Specifically: S1021. Calculate the initial grayscale value based on the sensitivity coefficient, specifically: Initial gray value based on sensitivity coefficient for: ,in, These are the sensitivity coefficients of the R, G, and B channels, as determined in the experiment. , , These represent the original pixel grayscale values ​​of the red, green, and blue channels at pixel coordinates (x, y) in the original color image.

[0030] S1022. Perform local contrast analysis on the initial grayscale image, specifically: Local contrast calculation in pixel coordinates Centered on, take Local window, calculate local mean : Local contrast: in, To prevent small amounts from being divided by zero, For local contrast; in this invention, i and j are merely variable indices and are positive integers; This refers to the size of a local window. For pixel coordinates The initial grayscale value at that location.

[0031] S1023. Adaptive Gamma correction based on local contrast, specifically: Adaptive Gamma correction determines the Gamma value based on local contrast: ,in, As the baseline Gamma value, , These are manually defined adjustment parameters. Used to control the magnitude of the overall offset. Used to control the rate of decay of the exponent. For adaptive Gamma values, e is the base of the natural logarithm; Areas with low contrast (which may be areas with uniform icing) are enhanced with a small Gamma value lower than the baseline Gamma value, while areas with high contrast remain unchanged.

[0032] S1024. Output the enhanced grayscale target image: in, To output a grayscale target image The gray value at that location is 255, which is the maximum gray value of an 8-bit image.

[0033] In summary, the grayscale conversion process not only considers the contribution of color channels, but also adaptively enhances the icy areas based on local contrast.

[0034] In this embodiment of the invention, step S200 involves binarizing the grayscale target image to obtain a binarized target image, including the following steps S201-S205: S201. Divide the grayscale target image into different image blocks, and calculate the edge gradient of the image blocks based on the Sobel operator to obtain the target edge gradient. Specifically: S2011. In the specific implementation of this invention, the image block size is adaptively determined based on the local information complexity of the grayscale target image, and the grayscale target image is divided into several image blocks, specifically: Calculate the local variance or local entropy of each pixel in the grayscale target image as an indicator of local information complexity. Based on the local information complexity, the image region is divided into high-complexity regions and low-complexity regions according to the rule of determining the image patch size: small image patches (such as 16×16 pixels) are used for high-complexity regions to retain more details; large image patches (such as 32×32 pixels) are used for low-complexity regions to reduce the amount of computation. The image block division is completed based on the above image block size.

[0035] It should be noted that the local variance is: in, To grayscale the target image in pixel coordinates grayscale value at that location This represents the average grayscale value within a local window. This refers to the size of a local window. For local variance, pixel coordinates in the image patch The pixel values ​​at the specified locations, i and j, are merely variable indices and are positive integers.

[0036] It should also be noted that the rule for determining the image patch size is as follows: in, pixel coordinates The size of the image patch, For small image patch sizes (e.g., 16×16). For large image block sizes (e.g., 32×32). The complexity threshold set for experts.

[0037] In summary, this image block division can adaptively adjust according to the local features of the icy line image, ensuring that areas with rich edge details are processed in detail while reducing the overall computational cost.

[0038] S2012. Based on the Sobel operator, the edge gradient of each image block is calculated. The edge gradient of the corresponding pixel in each image block is calculated by the convolution kernel of the Sobel operator to obtain the target edge gradient. It should be noted that the edge gradient calculation for each image patch based on the Sobel operator is as follows: The Sobel operator includes a horizontal convolution kernel and a vertical convolution kernel. The horizontal and vertical convolution kernels are convolved with each image block to obtain the horizontal and vertical edge gradient components of each pixel. The edge gradient magnitude of each pixel is calculated based on the horizontal and vertical edge gradient components to obtain the target edge gradient.

[0039] Wherein, the horizontal convolution kernel for: The vertical convolution kernel for: The horizontal edge gradient component and vertical edge gradient components The expression is: in, For horizontal edge gradient components, For vertical edge gradient components, The kernel is a horizontal convolution kernel. The kernel is a vertical convolution kernel. pixel coordinates in the image patch The pixel value at the location; i and j are only variable indices and are positive integers.

[0040] The expression for the magnitude of the edge gradient is: in, This represents the edge gradient magnitude, i.e., the target edge gradient.

[0041] S202. Enhance the image patch based on the target edge gradient to obtain an enhanced image patch, and fuse the enhanced image patch to generate an enhanced grayscale target image. Specifically: S2021. Extract the maximum and minimum gray values ​​of each image block, and determine the weighted comparison coefficient of the image block based on the target edge gradient and a preset weighting factor. The weighted comparison coefficient of each image block is determined based on the target edge gradient and a weighting factor, which is preset by relevant personnel. It should be noted that the weighting factor is dynamically calculated based on the grayscale dynamic range and edge gradient features of each image patch. Specifically: Calculate the normalized value of the gray-level dynamic range of the image patch; calculate the normalized value of the edge gradient intensity of the image patch; and adaptively calculate the weighting factor based on the above two normalized values ​​to reasonably enhance regions with large gray-level dynamic range and strong edge gradient.

[0042] Among them, the normalized value of grayscale dynamic range: in, This is the normalized value for the grayscale dynamic range. This represents the maximum grayscale value of the image patch. 255 represents the minimum grayscale value of the image block, and 255 represents the maximum grayscale value of an 8-bit image.

[0043] Normalized edge gradient intensity: in, pixel coordinates in the image patch Edge gradient magnitude at the location, The total number of pixels in the image block. To be the maximum possible edge gradient magnitude, This is the normalized value of the edge gradient intensity.

[0044] Adaptive weighting factor : in, This is the scaling factor. This is the offset coefficient. To prevent small amounts from being divided by zero, It is the hyperbolic tangent function.

[0045] When the grayscale dynamic range of an image patch is relatively large compared to the edge gradient... As the value approaches 1, the enhancement processing focuses more on the grayscale dynamic range of image patches, making it suitable for image areas with low overall contrast but relatively clear edge information. When the edge gradient is relatively large When the value approaches 0, the stronger the processing, the more it focuses on edge gradient information, which is suitable for image regions with blurred edges but rich gray levels; The weighting factor can be adaptively adjusted according to the characteristics of the image patch itself, avoiding the problems of insufficient or excessive enhancement caused by fixed values.

[0046] Obtain the maximum and minimum gray values ​​of each image block, and calculate the weighted comparison coefficient of each pixel in the image block based on the target edge gradient, maximum and minimum gray values, and weighting factors. The expression for the weighted control coefficient is: Where F is the weighted control coefficient. As a weighting factor, This represents the maximum grayscale value of the image patch. mag is the minimum gray level of the image patch, and mag is the gradient of the target edge.

[0047] S2022. Based on the weighted comparison coefficient, perform the above enhancement processing on each image block, that is, use the weighted comparison coefficient as the pixel value of the corresponding pixel point of the image block to replace the original pixel value, and obtain the enhanced image block.

[0048] S2023 generates an enhanced grayscale target image based on each enhanced image block, and then stitches and assembles the enhanced image blocks to obtain the enhanced grayscale target image. It should be noted that the stitching assembly is performed according to the original spatial position relationship of each image block, placing the enhanced image blocks one by one into the corresponding positions of the grayscale target image; for the overlapping boundary region of adjacent image blocks, a linear weighted fusion method is used for stitching processing, that is, for each pixel in the overlapping region of two adjacent image blocks, the corresponding fusion weight is determined according to the distance between the pixel and the boundary of the two adjacent image blocks. For the overlapping boundary region of adjacent image blocks, a non-linear fusion method based on structural similarity and gradient consistency is used for stitching processing.

[0049] Specifically: within the overlapping region, the structural similarity index of corresponding pixels of two adjacent image blocks is calculated; the gradient direction consistency of two adjacent image blocks in the overlapping region is calculated; the fusion weight is determined based on the structural similarity and gradient direction consistency; and nonlinear weighted fusion is performed according to the fusion weight to obtain the fused pixel value.

[0050] Among them, the above-mentioned structural similarity indicators for: in, , At pixel coordinates The average gray value of two adjacent image blocks A and B within the corresponding window of the overlapping region. , For pixel coordinates The variance of two adjacent image patches A and B at a given location. For pixel coordinates The covariance of two adjacent image patches, , It is the stability constant.

[0051] Among them, the above gradient direction consistency for: in, , The pixel coordinates of two adjacent image blocks A and B are respectively The gradient vector of the corresponding pixel.

[0052] Among them, the aforementioned fusion weights for: Among them, the above-mentioned fused pixel values for: ,in, , The pixel coordinates of two adjacent image blocks A and B are respectively The overlapping area corresponds to the pixel value of the pixel. , The pixel coordinates of image blocks A and B are respectively. The fusion weights and the fused pixel values.

[0053] The stitching and fusion process considers not only pixel distance but also local structural similarity and edge continuity, which can reduce boundary blurring and artifacts caused by linear fusion.

[0054] S203. Obtain the grayscale range of the enhanced grayscale target image, and divide the enhanced grayscale target image into different blurred regions based on the grayscale range. Specifically: S2031. In the specific implementation of this invention, the grayscale range is determined based on the grayscale values ​​of the enhanced grayscale target image. It should be noted that all pixels in the enhanced grayscale target image are traversed to obtain the grayscale value of each pixel. Based on the grayscale histogram distribution, an adaptive truncation method is used to determine the grayscale range, specifically: Construct a grayscale histogram and calculate the cumulative distribution function; Based on a preset truncation ratio (such as the low tail truncation ratio) and high tail cutoff ratio Determine the truncated grayscale value; The truncated low and high gray values ​​are used as the lower and upper limits of the gray range.

[0055] It should be noted that the cumulative distribution function for: in, grayscale value The probability of occurrence Let be the cumulative distribution function. is the total number of grayscale values, representing the proportion of all pixels with grayscale values ​​from 0 to g; i is only a variable index and is a positive integer.

[0056] The adaptive grayscale range is: in, To adaptively truncate the lower limit of the grayscale range, To adaptively truncate the upper limit of the grayscale range, The low tail truncation ratio can be determined based on the average of a large number of experiments, and the value is 0.01. The high tail truncation ratio can be determined based on the average of a large number of experiments, and the value is 0.01. These are the operators for taking the minimum and maximum values ​​of the corresponding parameters, respectively.

[0057] Adaptive truncation of grayscale range is .

[0058] In summary, this defined grayscale range can eliminate interference from a small number of overexposed or underexposed noise pixels, making subsequent blurry region segmentation and binarization threshold analysis more stable.

[0059] S2032. Randomly select several gray values ​​within the gray range, and divide the enhanced grayscale target image into several blurred regions based on the selected gray values.

[0060] S204. Perform grayscale threshold analysis on the fuzzy region based on information entropy to obtain the target grayscale threshold, specifically: S2041. Obtain the probability of each gray value appearing in the blurred area, calculate the information entropy based on the probability of gray values ​​appearing, and calculate the probability of gray values ​​in the blurred area in the background and foreground. It should be noted that the probability of each gray value occurring is specifically calculated by traversing all pixels in the blurred region and counting the frequency of each gray value; the probability of each gray value occurring is then calculated based on the frequency of each gray value. The expression for the probability of each grayscale value appearing is: in, This represents the probability of a grayscale value appearing. This represents the frequency of grayscale values ​​appearing in the blurred region. This represents the total number of pixels in the blurred region.

[0061] It should also be noted that the calculation of information entropy specifically involves: The overall information entropy of the blurred region is calculated based on the probability of each gray value appearing. The expression for the information entropy is: in, The overall information entropy of the fuzzy region, This represents the number of gray levels (usually 256). This represents the probability of grayscale values ​​appearing.

[0062] Furthermore, the probability of grayscale values ​​in the blurred region within the background and foreground is calculated, specifically: Within the grayscale range of the blurred region, a set of candidate grayscale thresholds is determined based on histogram peak-valley analysis and the maximum inter-class variance method. Specifically: Perform gray-level histogram statistics on the fuzzy region; use histogram peak-valley analysis to determine the trough positions in the histogram as initial candidate thresholds; use the Otsu method to calculate the global optimal threshold for the fuzzy region; merge the trough thresholds and the Otsu thresholds to construct a candidate gray-level threshold set; select T sequentially from the candidate set to calculate information entropy, and finally determine the target gray-level threshold that maximizes the total entropy, where T is the candidate gray-level threshold.

[0063] It should be noted that the maximum inter-class variance is: ,in, Candidate threshold The corresponding maximum inter-class variance, These represent the probabilities of the background and foreground regions, respectively. These are the average gray levels of the background and foreground regions, respectively, where: Otsu's optimal threshold for: Candidate threshold set for: ,in, For the histogram of the th The gray values ​​corresponding to each trough position.

[0064] In summary, the selection of candidate grayscale thresholds T is based on both histogram statistics and the classic Otsu method, moving beyond blind iteration; it prioritizes grayscale values ​​smaller than... The pixels are divided into background regions, and the gray values ​​greater than or equal to are... The pixels are divided into foreground regions; the grayscale probability distributions of the background and foreground regions are calculated separately. The expressions for the grayscale probability distributions of the background and foreground regions are as follows: in, The grayscale value in the background area The conditional probability at that time, The grayscale value in the foreground area The conditional probability at that time, The probability of being the background region. The probability of the foreground region.

[0065] The expressions for the probabilities of the background region and the foreground region are: S2042. Calculate the background information entropy and foreground information entropy based on the information entropy and the probability of grayscale values ​​in the background and foreground. Determine the target grayscale threshold based on the background information entropy and foreground information entropy. Specifically: The expressions for the background information entropy and the foreground information entropy are: in, For background information entropy, Foreground information entropy.

[0066] Furthermore, the target grayscale threshold is determined based on the background information entropy and foreground information entropy. Specifically: Calculate the sum of background entropy and foreground entropy to obtain the candidate grayscale threshold. The corresponding total entropy; traverse all candidate grayscale thresholds within the grayscale range of the blurred region, and select the candidate grayscale threshold that maximizes the total entropy as the target grayscale threshold.

[0067] The total entropy is The target grayscale threshold is ,in, Candidate grayscale threshold The corresponding total entropy, The target grayscale threshold.

[0068] S205. Binarize the blurred region based on the target grayscale threshold to obtain a binarized blurred region, and determine the binarized target image based on the binarized blurred region. Specifically: When the gray value of the blurred area is greater than or equal to the target gray value threshold, the area is set as the foreground pixel, and the foreground pixel is set to black. If the gray value of the blurred region is less than the target gray value threshold, then the region is set as the background pixel, and the background pixel is set to white, thus realizing the binarization of the blurred region and obtaining the binarized blurred region. The binarized blurred regions are stitched together to obtain the binarized target image. It should be noted that the binarized blurred regions are stitched together according to their original spatial position relationship in the enhanced grayscale target image to obtain the complete binarized target image.

[0069] It should also be noted that the purpose of the stitching combination here is the same as that of the stitching assembly in S2023 above, which is to integrate the images after regional processing into a complete image; however, the specific operation methods are different: the stitching assembly in S2023 above involves the overlapping boundary areas of adjacent image blocks, and linear weighted fusion is required to eliminate the stitching gaps; while in this step, each blurred area is a non-overlapping area divided based on the grayscale range, so it can be directly combined according to the original spatial position relationship without weighted fusion processing.

[0070] Furthermore, considering the unique characteristics of iced railway images under natural operating conditions, such as transparent and semi-transparent ice layers, complex lighting, and strong background interference, to further improve the accuracy and robustness of icing region extraction, a semantic segmentation of the icing region based on physical perception constraints is introduced on the basis of the binarized target image: Specifically, a priori library of icing physical characteristics is constructed, which includes the spectral reflectance characteristics of ice layer, the cylindrical geometric constraints of icing conductors, and the spatial continuity constraints of icing regions. The enhanced grayscale target image and the original target image are input into a lightweight semantic segmentation network. A physical perception loss function is designed based on the prior library of icing physical features. The semantic segmentation network is then trained to obtain a probability map of the icing region. The probability map of the icing area is binarized based on a preset probability threshold to obtain a refined icing target mask, and the binarized target image is updated based on the refined icing target mask. Ice thickness analysis is performed based on the updated binarized target image, and the results of the ice thickness analysis are fed back to the ice melting voltage adjustment module and the ice melting mode selection module, forming a closed-loop feedback between image processing and ice melting control.

[0071] It should be noted that, in constructing the prior library of physical characteristics of icing, the ice layer appears translucent or milky white under visible light, and there are spectral differences between it and the conductor, sky, and vegetation background; the iced conductor is cylindrically swollen with continuous and smooth edges; under different lighting conditions, the surface of the ice layer exhibits a mixture of specular and diffuse reflection characteristics.

[0072] It should be noted that the physical perception loss function introduces physical consistency constraints into the deep learning segmentation network, including: spatial continuity constraint, the icing region should be a connected region; shape constraint, the icing contour should be close to a cylinder; and spectral constraint, the pixels in the icing region should satisfy the semi-transparent reflection model of the ice layer.

[0073] It should be noted that the lightweight semantic segmentation network used here employs U-Net (a convolutional neural network for image segmentation) or MobileNet-DeepLab (a lightweight semantic segmentation model) as the backbone network; the input includes the original RGB image and the grayscale image after the aforementioned enhancement; the output is an ice-covered region probability map, and a fine binary mask is obtained through thresholding.

[0074] It should be noted that the closed-loop feedback is driven by the image processing results to dynamically adjust the ice melting parameters. Based on the aforementioned analysis of ice thickness, and combined with line parameters, the required current density and duration for ice melting are calculated; and a mapping table of ice thickness, ice melting voltage, and ice melting duration is established.

[0075] When the image segmentation confidence level is lower than the threshold or abnormal lighting conditions are detected, the image acquisition parameters are adjusted (such as supplemental lighting and exposure time); when the rate of change of ice thickness exceeds the threshold, the sampling frequency and monitoring points are adjusted.

[0076] Images are continuously collected during the ice melting process. The ice melting effect is evaluated by comparing the changes in ice thickness before and after melting. If the ice melting effect does not meet expectations, feedback is provided to adjust the ice melting voltage or ice melting mode.

[0077] In an embodiment of the present invention, step S300 involves analyzing the ice thickness of the binarized target image to obtain the target ice thickness of each icing line, including the following steps S301-S303: S301. Perform vertical projection curve analysis on the binarized target image to obtain the target vertical projection curve, specifically: Morphological closing operations are performed on the binarized target image to connect broken foreground pixels and remove isolated noise points; morphological closing operations for: ,in, To binarize the target image, As a structural element, For expansion operations, For erosion calculation.

[0078] Based on the morphologically processed image, pixel values ​​are statistically analyzed column by column along the vertical direction, and row position weights are introduced to make pixels in the central region of the image contribute more; the weighted vertical projection is: in, For the first The vertical projection value of the column. For the number of rows in the image, belong , For the first Row position weights (e.g., Gaussian weights, which give higher weights to rows in the middle of the image): in, denoted as the standard deviation of the Gaussian weighted distribution.

[0079] Gaussian smoothing is applied to the statistical results to eliminate local fluctuations. Gaussian smoothing is as follows: in, For the smoothed first Column projection values, For smoothing coefficients, The width of the Gaussian smoothed window is half; k is only the variable index. For the first The vertical projection value of the column; It is an abbreviation for Exponential. Pi is the mathematical constant of a circle.

[0080] The target vertical projection curve is obtained after smoothing, which has better noise suppression capability and peak positioning accuracy.

[0081] S302. Determine the pixel width of the icing line based on the target vertical projection curve, specifically: Peak feature analysis is performed based on the target vertical projection curve to determine the pixel width of the icing line. Specifically: peak detection is performed on the smoothed target vertical projection curve to determine the peak position; half-peak points are determined on both sides of the peak position; the pixel width of the icing line is calculated based on the half-width; if there is a bimodal or multimodal structure, a Gaussian mixture model is used for fitting to determine the width of each sub-peak and sum them.

[0082] It should be noted that the peak detection is as follows: ,in, is the column coordinate of the peak value; j is the variable index.

[0083] It should be noted that the half-peak height is: in, Half-peak height, For the smoothed first Column projection values.

[0084] It should be noted that the above half-height and width are: ,in, and Satisfying the conditions on the left and right sides of the peak, respectively. Location, It is half the height and width.

[0085] It should be noted that the pixel width of the aforementioned icing line is: ,in, This is a correction factor used to convert the half-width to half-height to the equivalent pixel width, typically set to 1.0~1.2; This represents the pixel width of the icing line.

[0086] If the projection curve has a multi-peak structure, a Gaussian mixture model is used: in, , , The first The amplitude, center location, and standard deviation of each Gaussian component. H is the baseline; H is the total number of Gaussian components; and h is the variable index.

[0087] It should be noted that the above pixel width: This allows for more precise determination of the pixel width of the icing lines, especially in cases of uneven icing or blurred edges.

[0088] S303. Based on the pixel width, obtain a first thickness estimate of the icing line using the vertical projection method; obtain a second thickness estimate of the icing line using the edge gradient method; fuse the first thickness estimate and the second thickness estimate to determine the final target icing thickness. Specifically: The first thickness estimate of the icing line is obtained based on the vertical projection method; the second thickness estimate of the icing line is obtained based on the edge gradient method; the icing density correction coefficient is calculated based on the ambient temperature and humidity; and the final target icing thickness is determined based on multi-source information fusion.

[0089] It should be noted that the above-mentioned first thickness estimate is: in, This is the first thickness estimate based on the vertical projection method. This refers to the overall outer diameter of the conductor after it has become icy. The diameter of the conductor itself (excluding ice).

[0090] It should be noted that the second thickness estimate mentioned above is obtained by performing edge gradient analysis on the grayscale target image, extracting the edge points on both sides of the icy guide wire, and calculating the distance between the two edges: in, and The first The x-coordinates of the left and right edges of the ice-covered conductor. These represent the icing diameter using the edge gradient method and the number of sampling rows at the edge points of the icing guide wire, respectively. is the second thickness estimate based on the edge gradient method; i is the variable index.

[0091] It should be noted that the above-mentioned icing density correction factor is based on ambient temperature. and relative humidity Sure: in, Standard icing density, , These are environmental correction factors assigned to ambient temperature and relative humidity, respectively. This is the correction factor for ice density.

[0092] It should be noted that the aforementioned final target icing thickness... for: in, , , They are respectively , , The fusion weight coefficient satisfies .

[0093] Ice thickness analysis is a fusion of multi-source information, which improves the accuracy and robustness in complex environments.

[0094] In this embodiment of the invention, step S400 determines the power supply importance coefficient of each icing line and analyzes the priority of each icing line based on the power supply importance coefficient and the target icing thickness, including the following steps S401-S403: S401. Obtain the load level of the icing line, and determine the power supply importance coefficient of the icing line based on the load level. Specifically: Obtain the load level of each icing line, and determine the power supply importance coefficient of each icing line based on the load level. Specifically: Obtain the user type, number of users, and power outage impact range of the users supplied by the icing line; determine the economic loss per unit time for each type of user based on the user type; calculate the expected economic loss of the icing line based on the economic loss per unit time, the number of users, and the expected power outage duration; and classify the load level of the icing line into Level 1 load, Level 2 load, Level 3 load, and Level 4 load based on the expected economic loss.

[0095] It should be noted that the above-mentioned economic losses per unit time are: in, Economic loss per unit time For the number of user types, For the first Number of users of each class For the first Economic losses per unit of time for users of this type.

[0096] It should be noted that the above-mentioned estimated total economic losses ,in, To estimate the total economic loss, To estimate the duration of the power outage, This is a correction factor for the severity of icing.

[0097] It should be noted that the above-mentioned icing severity correction coefficient ,in, For correction factor, Target ice thickness.

[0098] It should be noted that the above-mentioned quantitative load level classification is shown in Table 1: Table 1 Classification of Quantitative Load Levels

[0099] in, These are the economic loss thresholds for Level 1, Level 2, and Level 3 loads, which are set manually.

[0100] In summary, determining the load level based on economic loss calculations enhances objectivity and creativity.

[0101] The first-level loads are those whose power outages would cause personal injury, major equipment damage, major environmental pollution, major political impact, or serious disruption to public order, including power supply lines to users such as hospitals, mines, and important transportation hubs.

[0102] The secondary loads are those whose power outages would cause significant economic losses, significant political impact, or disruption to public order, including power supply lines to key industrial enterprises, large commercial facilities, and high-rise residential buildings.

[0103] The three-level loads are loads that would cause some economic loss or impact if the power supply were interrupted, but would not cause major safety consequences, including power supply lines of general industrial and commercial users, ordinary public buildings, etc.

[0104] The fourth level of load refers to other ordinary loads besides the first, second, and third level loads mentioned above, including power supply lines for rural lighting, ordinary residential electricity, and other users.

[0105] S402. A weighted sum is calculated based on the length of the iced line and the conductor diameter of the iced line to obtain the vulnerability index of the iced line. Specifically: The expression for the vulnerability index of the icing-covered line is: in, As an indicator of the vulnerability of icing-covered lines, This refers to the length of the icing line. The length of the longest line in the region. The diameter of the conductor in the icing-affected line. Minimum allowable wire diameter, This is the length weighting coefficient. This is the diameter weighting coefficient.

[0106] S403. Based on the vulnerability index of the iced lines, the power supply importance coefficient, and the target icing thickness, a weighted fusion is performed on each iced line to obtain the priority of the iced lines. Specifically: The vulnerability index of iced lines is obtained, and the priority of each iced line is analyzed based on the vulnerability index, power supply importance coefficient, and target icing thickness. The priority is... ,in, Priority The weighting coefficient for the target icing thickness. For the target icing thickness, The weighting coefficients for the vulnerability index of iced railway lines. As an indicator of the vulnerability of icing-covered lines, The weighting factor for the importance coefficient of power supply. This is the power supply importance coefficient.

[0107] In this embodiment of the invention, in step S500, the corresponding de-icing voltage is determined based on the length of each icing line and the diameter of the conductor. The specific steps are as follows: Based on the line length and conductor diameter, the basic de-icing voltage requirement is determined, and the target de-icing voltage is calculated. Specifically: The intelligent de-icing voltage determination model based on thermal balance and multi-parameter fusion specifically determines the basic de-icing voltage requirement based on line length and conductor diameter; calculates the de-icing heat loss correction coefficient based on ice thickness, ambient temperature, and wind speed; determines the multi-branch de-icing voltage distribution strategy based on the ring network topology; and calculates the target de-icing voltage by combining the above factors.

[0108] It should be noted that the above-mentioned basic de-icing voltage for: ,in, , This is a reference coefficient determined based on conductor materials and line standards.

[0109] It should be noted that the above heat loss correction factor is: ,in, For the target icing thickness, For ambient temperature, For wind speed, , , Assigned to respectively , , The heat loss influence coefficient, This is the heat loss correction factor.

[0110] It should be noted that the voltage distribution coefficients mentioned above for multi-branch voltage distribution are for ring networks where ice melting occurs simultaneously. Branches, based on the priority of each branch. The voltage distribution coefficient is determined as follows: in, The first Voltage distribution coefficient of each line, and the total number of branches that can be melted simultaneously; Let be the priority value of the i-th line.

[0111] It should be noted that the target de-icing voltage mentioned above is: ,in, For the first The target de-icing voltage for an icing-covered power line. For the first The basic de-icing voltage of this line. For voltage distribution coefficient, This is the heat loss correction factor.

[0112] The de-icing voltage is determined by comprehensively considering the icing state, environmental conditions, and coordination of multiple branches of the ring network, forming an intelligent method for determining the de-icing voltage.

[0113] In an embodiment of the present invention, in step S600, a target de-icing mode is determined, and switch action information is determined based on the priority and the target de-icing mode. Then, based on the switch action information and the de-icing voltage, de-icing treatment is performed on the iced line, including the following steps S601-S603: It should also be noted that the priority, de-icing mode, and switching action in this invention are not simply sequentially executed, but rather based on a condition-response disturbance rejection control strategy of real-time reconfiguration of the ring network topology. Before each switching operation, the host performs power flow calculations based on the current ring network topology to analyze the impact of the target de-icing line switching on the load distribution of the remaining lines. If the load rate of a remaining line exceeds the preset safety threshold, the de-icing sequence or de-icing mode will be adjusted so that the line is deloaded before de-icing or a reverse de-icing mode is used to avoid the heavily loaded path. If a remaining line is at risk of power outage due to switching operation, the backup power supply will be automatically activated or the switching status of adjacent lines will be adjusted to ensure continuous power supply to non-ice-melting loads. Based on the above automatic reload analysis results, the ice melting sequence, ice melting mode and switch action information are updated in real time to form a closed-loop disturbance rejection control.

[0114] The condition-response rules are shown in Table 2: Table 2 Condition-Response Rules

[0115] Heavy load analysis calculation ,in, For the first Load rate after switching of the remaining lines Before switching load rate, For the reason The increase in load transfer caused by the ice melting and switching of the line.

[0116] in, For the first The load carried by the ice-melting line For the first The load of line 1 to the first The transfer coefficient of each line is determined by the ring network topology and switch states. Let be the rated transmission power of the j-th line.

[0117] S601. Determine the target de-icing voltage, and based on the priority, determine the switch switching direction and switch closing state, specifically: The switching direction of the boundary switch, the closing state of the short-circuit switch, and the switching direction of the ice-melting switch are determined based on the priority and the target ice-melting mode. In one embodiment, the de-icing sequence of iced lines is determined based on priority, i.e., dynamic priority assessment, which comprehensively considers the icing status, grid operating status, and the impact of switch operations. Specifically: The three static factors of icing thickness, line vulnerability, and power supply importance are retained. Increase real-time power grid operating status parameters, including real-time power flow load rate and system stability margin; Add parameters affecting switching operations, including the impact index of switching operations on power supply reliability. Dynamic priority is obtained by dynamically weighting and fusing the above multi-source parameters.

[0118] It should be noted that the above real-time power flow load rate is: in, For the first Real-time active power flow of an icing-covered power line. The rated transmission power of the line, For the first Real-time power flow load rate of each line.

[0119] It should be noted that the stability margin of the above system is: in, For the first The critical stable transmission power of the line, For the first The system stability margin of each line.

[0120] It should be noted that the above-mentioned impact index of switch operation is calculated based on the load transfer situation of non-ice-melting lines after the switch operation: in, For the first The set of adjacent lines affected by the de-icing of a given line. For the first The load increment of adjacent lines, The total system load, For the first The impact index of switch operation during ice melting on a line; j is the variable index.

[0121] It should be noted that the above dynamic priority is as follows: in, For the first Dynamic priority of icing lines , , , , , These are the weighting coefficients for each factor. superscript Indicates the first An ice-covered railway line.

[0122] Furthermore, the dynamic de-icing sequence is determined: the ice is de-iced first according to the dynamic priority; when the power grid operating status changes (such as a new line trip or a sudden increase in load), the dynamic priority is recalculated and the de-icing sequence is adjusted.

[0123] This prioritizes dynamic control commands, enabling them to adapt to real-time changes in the power grid's operating status and improve the intelligence and safety of ice melting control.

[0124] In another embodiment, a target de-icing mode can be determined based on the priority of the icing line. The target de-icing mode includes a forward de-icing mode, a reverse de-icing mode, a forward-then-reverse mode, or a reverse-then-forward mode, etc. The switching of the forward de-icing mode is as follows: Figure 4 As shown, the reverse ice melting mode is switched on and off as follows: Figure 5 As shown.

[0125] Based on the priority and target de-icing mode, the switching direction of the boundary switch, the closing state of the short-circuit switch, and the switching direction of the de-icing switch are determined. That is, the switching action of the ice-covered line with higher priority is determined according to the target de-icing mode and the de-icing sequence of the line. It should be noted that the de-icing sequence of each ice-covered line is determined according to the priority ranking result, and the ice-covered line with higher priority performs the de-icing operation first. Based on the relative connection position of the line to be melted with the multi-functional de-icing distribution transformer in the ring network topology, the target de-icing mode for the line is determined: if the line is connected to the forward power supply side of the transformer, the forward de-icing mode is adopted, and the boundary switch is controlled to switch to the forward contact and the de-icing switch is controlled to switch to the corresponding forward contact; if the line is connected to the reverse power supply side of the transformer, the reverse de-icing mode is adopted, and the boundary switch is controlled to switch to the reverse contact and the de-icing switch is controlled to switch to the corresponding reverse contact. The short-circuit switch closing state is determined according to the target de-icing mode, so that the current line to be de-iced forms a closed de-icing current loop.

[0126] like Figure 3 As shown, if the priority of line C > the priority of line A > the priority of line B, line C can be merged first in the forward direction, then line A in the reverse direction, and finally line B in the forward direction. like Figure 3 As shown, if the priority of line C is > the priority of line A is > the priority of line B, based on the connection position of each line with the multi-functional ice-melting distribution transformer in the ring network topology, line C and line B are connected to the forward power supply side of the transformer, and line A is connected to the reverse power supply side of the transformer. Therefore, according to the priority order, line C can be melted first, then line A can be melted in reverse, and finally line B can be melted in forward. S602. Based on the switching direction of the disconnector and the closing state of the switch, generate switch action information and determine the dynamic de-icing sequence, specifically: The switching action information is generated based on the switching direction of the boundary switch, the closing state of the short-circuit switch, and the switching direction of the de-icing switch. The boundary switch and the de-icing switch are double-throw, double-throw disconnect switches, which simplify the design and reduce costs.

[0127] It should be noted that the specific steps for generating the switch action are as follows: The switching direction of the sectionalizing switch, the closing state of the short-circuit switch, and the switching direction of the de-icing switch are encoded to generate sectionalizing switch action commands, short-circuit switch action commands, and de-icing switch action commands. The boundary switch action command, short circuit switch action command, and ice-melting switch action command are combined in the ice-melting sequence to generate switch action information; the switch action information includes switch identifier, target contact position or closed state, and action execution sequence. Based on the switch action information, the corresponding switches are controlled to perform actions in sequence. The dividing switch and the ice-melting switch are double-throw type isolating switches.

[0128] S603. Based on switch action information and a dynamically determined de-icing sequence, de-icing is performed on the iced lines. Specifically: Based on the switch action information and the de-icing voltage control, the multi-functional de-icing distribution transformer is used to de-ic the iced lines. The layout of the switch and the multi-functional de-icing distribution transformer is as follows: Figure 3 As shown; The boundary switch changes the direction of the knife switch, causing the power supply to the current ice-melting line to be cut off and the short-circuit switch of the current ice-melting line to be closed. The de-icing switch changes the direction of the disconnector, creating a loop in the current de-icing line. This controls the multi-functional de-icing distribution transformer to output de-icing voltage to melt the ice on the corresponding iced line. Taking line C as an example, the control logic of the bidirectional switch is as follows: The sectionalizing switch changes the direction of the knife switch to contact ①; the short-circuit switch 3 closes; the de-icing switch changes the direction of the knife switch to contact ②.

[0129] The multi-functional de-icing distribution transformer can be used as a dual-purpose transformer for both users and de-icing. In the de-icing scenario of ring network lines, the transformer supports de-icing of multiple de-icing branches in both forward and reverse directions, and can support de-icing of up to 14 branches. When the line is de-icing, the user side of the transformer that provides the de-icing power and the user side of the adjacent transformer are powered normally. The transformer body adopts a single winding multi-tap design and has two sets of low-voltage output terminals, one for power supply and one for de-icing. The power supply side provides stable power to users for a long time. On the de-icing side, de-icing current is output after the de-icing circuit is formed, and de-icing is achieved by using the principle of current heating effect. Under normal circumstances (when the de-icing circuit is not formed), no de-icing current is output. A transformer can achieve de-icing of the forward and reverse lines through the action of the boundary switch and the de-icing switch, and the de-icing distance is doubled, that is, a transformer can de-ic the distribution network line up to 10km.

[0130] Example 3, referring to Figure 2 As one embodiment of the present invention, this embodiment provides a ring network line de-icing system based on switch control, comprising: The image grayscale module is used to acquire target images of each icing line and convert the target images to grayscale to obtain grayscale target images. The image binarization module is used to perform binarization processing on the grayscale target image based on the division of blurred regions to obtain a binarized target image; The ice thickness analysis module is used to analyze the ice thickness of the binary target image based on the vertical projection method to obtain the target ice thickness of each ice-covered line. The priority analysis module is used to determine the power supply importance coefficient of each icing line and analyze the priority of each icing line based on the power supply importance coefficient and the target icing thickness. The de-icing voltage analysis module is used to determine the corresponding de-icing voltage based on the length of each iced line and the diameter of the conductor. The line melting module is used to determine the target de-icing mode, determine the switch action information based on the priority and the target de-icing mode, and control the multi-functional de-icing distribution transformer to perform de-icing treatment on the iced line based on the switch action information and the de-icing voltage.

[0131] In the specific implementation of this invention, the specific implementation of the device item can be referred to the implementation of the method item above, and will not be repeated here.

[0132] This embodiment also provides a computer device applicable to a switch-controlled ring network line de-icing method, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the switch-controlled ring network line de-icing method proposed in the above embodiment.

[0133] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements a ring network line de-icing method based on switch control as proposed in the above embodiment.

[0134] The storage medium proposed in this embodiment belongs to the same inventive concept as the method for de-icing a ring network line based on switch control proposed in the above embodiment. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0135] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0136] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for de-icing ring network lines based on switch control, characterized in that: include, Target images of each icing line are acquired, and icing thickness is analyzed in the target images to obtain the target icing thickness of each icing line. Determine the power supply importance coefficient of each icing line, and analyze the priority of each icing line based on the power supply importance coefficient and the target icing thickness; The corresponding de-icing voltage is determined based on the length of each icing line and the diameter of the conductor. The target de-icing mode is determined, and the switch action information is determined based on the priority and the target de-icing mode. The de-icing process is then performed on the iced line based on the switch action information and the de-icing voltage.

2. The de-icing method for ring network lines based on switch control as described in claim 1, characterized in that, The ice thickness analysis of the target image includes: Collect target images of each icing line and calculate the information entropy of the color channels of the target images; The sensitivity coefficient of each color channel is calculated based on the information entropy. The target image is converted to grayscale based on the sensitivity coefficient, and the initial grayscale value is calculated to obtain the grayscale target image.

3. The de-icing method for ring network lines based on switch control as described in claim 2, characterized in that, The ice thickness analysis of the target image further includes: The grayscale target image is divided into different image blocks, and the edge gradient of the image blocks is calculated based on the Sobel operator to obtain the target edge gradient. The image patch is enhanced based on the target edge gradient to obtain an enhanced image patch, and the enhanced image patch is fused to generate an enhanced grayscale target image. Obtain the grayscale range of the enhanced grayscale target image, and divide the enhanced grayscale target image into different blurred regions based on the grayscale range; Based on information entropy, grayscale threshold analysis is performed on the fuzzy region to obtain the target grayscale threshold. The blurred region is binarized based on the target grayscale threshold to obtain a binarized blurred region, and the binarized target image is determined based on the binarized blurred region.

4. The de-icing method for ring network lines based on switch control as described in claim 3, characterized in that, The enhancement process includes: Extract the maximum and minimum gray values ​​of each image block, and calculate the weighting factor based on the target edge gradient to determine the weighted comparison coefficient of the image block; The image patch is enhanced based on the weighted comparison coefficient to obtain the enhanced image patch; The enhanced image blocks are stitched together to obtain the enhanced grayscale target image.

5. The de-icing method for ring network lines based on switch control as described in claim 4, characterized in that, The ice thickness analysis includes: The vertical projection curve of the binarized target image is analyzed to obtain the target vertical projection curve; The pixel width of the icing line is determined based on the target vertical projection curve; Based on the pixel width, the first thickness estimate of the icing line is obtained using the vertical projection method; The second thickness estimate of the icing line is obtained based on the edge gradient method; The first thickness estimate and the second thickness estimate are fused to determine the final target icing thickness.

6. The de-icing method for ring network lines based on switch control as described in claim 5, characterized in that, The analysis of the priority of each icing line includes: Obtain the load level of the icing line, and determine the power supply importance coefficient of the icing line based on the load level; The vulnerability index of icing lines is obtained by weighted summation of the length of the icing line and the conductor diameter of the icing line. The icing lines are weighted and fused based on the vulnerability index, power supply importance coefficient, and target icing thickness to obtain the priority of each icing line.

7. The de-icing method for ring network lines based on switch control as described in claim 6, characterized in that, The ice-melting process includes: Determine the basic de-icing voltage requirement based on the line length and conductor diameter, and calculate the target de-icing voltage. Determine the target de-icing voltage, and determine the switch switching direction and switch closing state based on the priority; Based on the switching direction of the disconnector and the closing state of the switch, switch action information is generated to determine the dynamic de-icing sequence; The ice-covered lines are de-iced based on switch action information and the determined dynamic de-icing sequence.

8. A ring network line de-icing system based on switch control, employing the ring network line de-icing method based on switch control as described in any one of claims 1 to 7, characterized in that, include: The image grayscale module acquires target images of each icing ring network line and performs grayscale conversion on the target images to obtain grayscale target images; The image binarization module performs binarization processing on the grayscale target image based on the fuzzy region division to obtain a binarized target image; The ice thickness analysis module analyzes the ice thickness of the binary target image based on the vertical projection method to obtain the target ice thickness of each ice-covered line. The priority analysis module determines the power supply importance coefficient of each icing line and analyzes the priority of each icing line based on the power supply importance coefficient and the target icing thickness. The de-icing voltage analysis module determines the corresponding de-icing voltage based on the length of each iced line and the diameter of the conductor. The line melting module determines the target de-icing mode, and determines the switch action information based on the priority and the target de-icing mode. Based on the switch action information and the de-icing voltage, it controls the multi-functional de-icing distribution transformer to perform de-icing treatment on the iced line.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the ring network line de-icing method based on switch control as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the ring network line de-icing method based on switch control as described in any one of claims 1 to 7.

Images