Power transmission line icing monitoring method, equipment and medium

Abstract

The invention relates to a power transmission line icing monitoring method, equipment and a medium. The method comprises the following steps: acquiring original image information about a power transmission line and target meteorological characteristic data of the position of a tower; performing image segmentation on the original image information through an image segmentation algorithm; conducting wire skeleton extraction on the image segmentation information through a skeleton refinement algorithm; determining actual maximum icing thickness information of the power transmission line according to the image segmentation information, the skeleton image information and the pixel mapping distance; inputting the target meteorological characteristic data into a meteorological prediction model, and outputting an icing index of the power transmission line through the meteorological prediction model; and determining the icing risk index of the power transmission line according to the actual maximum icing thickness information, the icing index, the maximum icing thickness information that the electric wire can bear, the current temperature information of the power transmission line, the freezing point temperature information of water, the image weight, the meteorological weight and the temperature smoothing factor. And the real-time monitoring accuracy of the icing risk of the power transmission line is greatly improved.

Description

Technical Field

[0001] This invention relates to the field of line monitoring technology, and in particular to a method, equipment and medium for monitoring icing on transmission lines. Background Technology

[0002] When transmission lines are covered by ice or wet snow, it increases the overall weight and internal stress of the conductors, thereby increasing the vertical load on the conductors and their surface area exposed to wind loads, which can lead to tower collapse. Severe icing can not only cause serious damage but also endanger lives.

[0003] One highly effective method for addressing the hazards of icing is real-time monitoring of icing thickness. In overhead transmission lines, various methods can be employed to obtain information on line icing, depending on the observation methods, measuring instruments, and equipment used. However, existing monitoring methods have several limitations. Traditional monitoring methods generally suffer from a lack of data dimension, relying solely on manual inspections or data from a single sensor, making it difficult to comprehensively reflect the spatiotemporal evolution of conductor icing. Summary of the Invention

[0004] The purpose of this invention is to provide a method, equipment, and medium for monitoring icing on power transmission lines.

[0005] According to one aspect of this application, a method for monitoring icing on transmission lines is provided, the method comprising: Obtain raw image information about the transmission line and target meteorological feature data of the location of the towers of the transmission line; The original image information is segmented using an image segmentation algorithm to obtain image segmentation information, which includes the target region and the pixel-level segmentation mask of the target region. The skeleton image information is obtained by extracting the conductor skeleton from the image segmentation information using a skeleton thinning algorithm. The skeleton image information includes the conductor skeleton of the transmission line. The actual maximum icing thickness of the transmission line is determined based on image segmentation information, skeleton image information, and pixel mapping distance. Input the target meteorological characteristic data into the meteorological prediction model, and output the icing index of the transmission line through the meteorological prediction model; The icing risk index of the transmission line is determined based on the actual maximum icing thickness information, icing index, maximum icing thickness that the power line can withstand, current temperature information of the transmission line, freezing point temperature of water, image weight, meteorological weight, and temperature smoothing factor.

[0006] According to another aspect of this application, a computer device is provided, including a memory and a processor, wherein a computer program capable of being loaded by the processor and executing the methods described above is stored in the memory.

[0007] According to another aspect of this application, a computer-readable storage medium is provided, storing a computer program that can be loaded by a processor and executed as described above.

[0008] Compared with existing technologies, this application determines the icing risk index based on the original image information of the transmission line and the target meteorological feature data of the tower location, so as to monitor the icing risk of the transmission line in real time. For the original image information, the original image information is segmented by an image segmentation algorithm to obtain image segmentation information; the skeleton image information is extracted from the image segmentation information by a skeleton thinning algorithm to obtain skeleton image information; the actual maximum icing thickness information of the transmission line is determined based on the image segmentation information, skeleton image information, and pixel mapping distance. This accurately determines the actual maximum icing thickness information. For the target meteorological feature data of the tower location, the target meteorological feature data is input into a meteorological prediction model to output the icing index of the transmission line. This combines meteorological information to further improve the accuracy of icing risk judgment. Finally, the icing risk index of the transmission line is determined based on the actual maximum icing thickness information determined by the image, the icing index determined by the target meteorological feature data, the maximum icing thickness that the wire can withstand, the current temperature information of the transmission line, the freezing point temperature of water, image weights, meteorological weights, and temperature smoothing factors. This scheme precisely processes the algorithms for both images and meteorological data. Ultimately, it determines the icing risk index of the transmission line by using the actual maximum icing thickness information, the icing index determined based on the target meteorological feature data, the maximum icing thickness that the power line can withstand, the current temperature information of the transmission line, the freezing point temperature of water, image weights, meteorological weights, and temperature smoothing factors. This greatly improves the accuracy of real-time monitoring of icing risk. Attached Figure Description

[0009] Figure 1 A flowchart of a method for monitoring icing on a transmission line according to an embodiment of this application is shown; Figure 2 A schematic diagram of the structure of a transmission line icing monitoring device according to an embodiment of this application is shown; Figure 3 A flowchart of another real-time transmission line icing monitoring method according to this application is shown; Figure 4 Exemplary systems that can be used to implement the various embodiments described in this application are shown. Detailed Implementation

[0010] The present application will now be described in further detail with reference to the accompanying drawings.

[0011] In a typical configuration of this application, the terminal, the device of the service network, and the trusted party all include one or more processors (e.g., a central processing unit (CPU)), input / output interfaces, network interfaces, and memory.

[0012] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash memory. Memory is an example of computer-readable media.

[0013] Computer-readable media, including both permanent and non-permanent, removable and non-removable media, can store information using any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PCM), programmable random access memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0014] The devices referred to in this application include, but are not limited to, terminals, network devices, or devices formed by integrating terminals and network devices through a network. The terminals include, but are not limited to, any mobile electronic product capable of human-computer interaction (e.g., via a touchpad), such as smartphones and tablets. These mobile electronic products can use any operating system, such as Android or iOS. The network devices include electronic devices capable of automatically performing numerical calculations and information processing according to pre-set or stored instructions. Their hardware includes, but is not limited to, microprocessors, application-specific integrated circuits (ASICs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), and embedded devices. The network devices include, but are not limited to, computers, network hosts, single network servers, multiple network server clusters, or clouds composed of multiple servers. Here, a cloud consists of a large number of computers or network servers based on cloud computing, where cloud computing is a type of distributed computing, consisting of a virtual supercomputer composed of a group of loosely coupled computer clusters. The network includes, but is not limited to, the Internet, wide area network, metropolitan area network, local area network, VPN network, wireless ad hoc network, etc. Preferably, the device can also be a program running on the terminal, network device, or a device formed by integrating the terminal and network device, network device, touch terminal, or network device and touch terminal through a network.

[0015] Of course, those skilled in the art should understand that the above-described devices are merely examples, and other existing or future devices that are applicable to this application should also be included within the scope of protection of this application, and are hereby incorporated by reference.

[0016] In the description of this application, "multiple" means two or more, unless otherwise expressly and specifically defined.

[0017] refer to Figure 1This invention provides a method for monitoring icing on power transmission lines, comprising steps S11, S12, S13, S14, S15, and S16. In step S11, the original image information of the transmission line and the target meteorological feature data of the tower location of the transmission line are obtained; in step S12, the original image information is segmented using an image segmentation algorithm to obtain image segmentation information, wherein the image segmentation information includes the target region and the pixel-level segmentation mask of the target region; in step S13, the conductor skeleton is extracted from the image segmentation information using a skeleton thinning algorithm to obtain skeleton image information, wherein the skeleton image information includes the conductor skeleton of the transmission line; in step S14, the actual maximum icing thickness information of the transmission line is determined based on the image segmentation information, the skeleton image information, and the pixel mapping distance; in step S15, the target meteorological information is input into the meteorological prediction model, and the predicted icing thickness information of the transmission line is output by the meteorological prediction model; in step S16, the icing risk index of the transmission line is determined based on the actual maximum icing thickness information, the predicted icing thickness information, the maximum icing thickness that the wire can withstand, the current temperature information of the transmission line, the freezing point temperature information, the image weight, the meteorological weight, and the temperature smoothing factor.

[0018] Specifically, in step S11, raw image information of the transmission line and target meteorological feature data of the location of the transmission line towers are acquired. In some embodiments, raw image information of the transmission line is acquired by a camera device installed on the tower. In some embodiments, the camera device includes an industrial-grade high-definition camera that captures JPEG format images at 10-minute intervals and simultaneously records EXIF ​​metadata. In some embodiments, the EXIF ​​metadata includes, but is not limited to, timestamps and GPS coordinates. In some embodiments, the target meteorological feature data includes, but is not limited to, temperature, humidity, wind speed, precipitation probability, weather field codes, and historical time-series meteorological information. In some embodiments, the target meteorological feature data includes target meteorological feature data of the tower location, for example, meteorological feature data corresponding to the latitude and longitude coordinates of the tower location. In some embodiments, the target meteorological feature data of the tower location is determined by a bilinear interpolation algorithm, the target meteorological grid data, and the tower location information. For a detailed explanation of this part, please refer to the corresponding embodiments below, which will not be repeated here.

[0019] In step S12, the original image information is segmented using an image segmentation algorithm to obtain image segmentation information, which includes a target region and a pixel-level segmentation mask for that target region. In some embodiments, the original image information is input into an image segmentation model, and the image segmentation model outputs image segmentation information. In some embodiments, the original image information is preprocessed before image segmentation to obtain an enhanced image. The enhanced image is then input into the image segmentation model to output image segmentation information. For specific descriptions of the preprocessing and image segmentation model, please refer to the corresponding embodiments below, which will not be repeated here. In some embodiments, the target region includes a conductor region (e.g., when there is no ice on the conductor). In other embodiments, the target region includes both a conductor region and an icy region (e.g., when there is ice on the conductor). In some embodiments, the pixel-level segmentation mask includes, but is not limited to, a mask with a pixel value equal to 1. For example, the target region is marked by a mask with a pixel value of 1. In other words, the image segmentation image includes the target region, the pixel value of the target region is 1, and the pixel value of the region outside the target region (e.g., the background region) is 0.

[0020] In step S13, the image segmentation information is processed by a skeleton thinning algorithm to extract the conductor skeleton, resulting in skeleton image information, which includes the conductor skeleton of the transmission line. In some embodiments, the skeleton thinning algorithm includes, but is not limited to, extracting the conductor skeleton from the target region by deleting pixels that meet the deletion criteria in the target region. For a detailed description of the skeleton thinning algorithm, please refer to the corresponding embodiments below; it will not be repeated here.

[0021] In step S14, the actual maximum icing thickness of the transmission line is determined based on image segmentation information, skeleton image information, and pixel mapping distance. In some embodiments, the pixel mapping distance includes, but is not limited to, the actual distance represented by a single pixel. In some embodiments, the actual icing thickness is determined by calculating the shortest Euclidean distance between each pixel on the region contour of the target area and the pixels on the conductor skeleton, and then using the shortest Euclidean distance and the pixel mapping distance. For a detailed explanation of this part, please refer to the corresponding embodiments below; it will not be repeated here.

[0022] In step S15, the target meteorological characteristic data is input into the meteorological prediction model, and the icing index of the transmission line is output through the meteorological prediction model. For example, the target meteorological characteristic data of the tower location is input into the meteorological prediction model, and the icing index is directly output through the meteorological prediction model. In some embodiments, a higher icing index indicates a greater probability of icing on the transmission line. For a detailed explanation of this part, please refer to the corresponding embodiments below, which will not be repeated here.

[0023] In step S16, the icing risk index of the transmission line is determined based on the actual maximum icing thickness information, the predicted icing thickness information, the maximum icing thickness that the power line can withstand, the current temperature information of the transmission line, the freezing point temperature information, image weights, meteorological weights, and temperature smoothing factors. For example, ， Here, R includes the icing risk index. Including image weights, Including meteorological weights, H includes actual maximum icing thickness information. This includes information on the maximum icing thickness that the power line can withstand. Including the icing index, Including information on the freezing point of water, This includes a temperature smoothing factor. In some embodiments, the image weight is preferably 0.7, and the meteorological weight is preferably 0.3. In some embodiments, =0, =0.1 can prevent the denominator from being zero. In some embodiments, 0.6≤R<0.8 indicates that icing has entered a rapid accumulation period, 0.8≤R<1.0 indicates that the line is approaching its theoretical load threshold, 1.0≤R<1.2 indicates that the line exceeds its design load, and R≥1.2 indicates an extreme dangerous situation.

[0024] In some embodiments, obtaining target meteorological feature data of the location of a transmission line tower includes: obtaining target meteorological grid data from a meteorological center based on the tower's location information, wherein the target meteorological network data covers the tower's location information and includes multiple reference meteorological feature data; and determining the target meteorological feature data using a bilinear interpolation algorithm and the target meteorological grid data. For example, the meteorological center includes meteorological data for multiple locations. The meteorological grid data includes multiple location coordinates and meteorological information corresponding to each location coordinate. In this embodiment, to accurately determine the meteorological information of the tower's location, the target meteorological feature data of the tower's location is determined using a bilinear interpolation algorithm. For example, the tower's location information includes (x, y), and target meteorological grid data including this location information (x, y) is obtained from the meteorological center based on the location information (x, y). For example, the target meteorological grid data includes (x, y). ), ( ), ( ), ( ),as well as( The corresponding reference meteorological characteristic data, ( The corresponding reference meteorological characteristic data, ( The corresponding reference meteorological characteristic data, ( The corresponding reference meteorological characteristic data. ), ( ), ( ), ( ) covers (x,y), in other words, (x,y) lies in ( ), ( ), ( ), ( Between ) . In some embodiments, ( The corresponding reference meteorological characteristic data includes, but is not limited to, location ( The corresponding meteorological information includes temperature, humidity, and wind speed. In some embodiments, the target meteorological characteristic data of the tower location is determined using a bilinear interpolation algorithm and target meteorological grid data. For example, Here, This includes target meteorological characteristic data (e.g., temperature) at the location of the pole. include( The corresponding reference meteorological characteristic data (e.g., temperature). include( The corresponding reference meteorological characteristic data (e.g., temperature). include( The corresponding reference meteorological characteristic data (e.g., temperature). include( The reference meteorological feature data (e.g., temperature) corresponding to the tower location information (x, y) is included in some embodiments, including but not limited to latitude and longitude information, such as x indicating latitude and y representing latitude. Similarly, the meteorological grid data includes ( This includes, but is not limited to, latitude and longitude information.

[0025] In some embodiments, the method further includes step S17 (not shown) before step S12. In step S17, the original image information is corrected using a distortion model algorithm to obtain a corrected image. The distortion model algorithm includes... ; ； Here, (x, y) includes the pixel coordinates from the original image information, , This includes the corrected pixel coordinates. , Including radial distortion coefficient, , This includes tangential distortion coefficients; enhancing the corrected image using a multi-scale retinal cortex theory algorithm to obtain an enhanced image; step S12 includes: inputting the enhanced image into an image segmentation model, and outputting image segmentation information through the image segmentation model. In some embodiments, a coordinate system is established in the original image information, where each pixel has its own pixel coordinates. The original image information is corrected using a distortion model algorithm to correct the original image and make subsequent analysis more accurate. In some embodiments, radial distortion coefficients and tangential distortion coefficients can be calibrated experimentally. For example, the distortion coefficients corresponding to the camera device can be calibrated using a checkerboard calibration method. In some embodiments, a multi-scale retinal cortex theory algorithm (multi-scale Retinex algorithm) is used to enhance the corrected image. Specifically, this includes: enhancing the corrected image using three different scales (e.g., Gaussian kernels with values ​​of 15, 80, and 250 are used to estimate the illumination components. Reflectance component enhancement is performed in the Log domain, and dynamic range compression and nonlinear tone mapping are applied. The formula is as follows: log{R}_{MSR}(x,y)=\sum ^{n}_{k=1} {{w}_{k}}(logI(x,y)-log[{F}_{k}(x,y)\times I(x,y)]) Here, This includes the grayscale value of the pixel coordinates (x, y) in the corrected image. Including the Gaussian filter kernel at the k-th scale, The weights include the weights for the k-th scale, and N represents the number of scales. In some embodiments, the image segmentation algorithm includes, but is not limited to, an image segmentation model. For example, by inputting the preprocessed enhanced image into the image segmentation model, the enhanced image is segmented, and image segmentation information is output.

[0026] In some embodiments, the image segmentation model is trained using the following method: a first base model is trained using multiple training images and labeled masks for each training image to obtain the image segmentation model. The labeled masks include the target region and a pixel-level segmentation mask for the target region. The model architecture of the first base model includes a ResNeXt-101+FPN backbone network, a region proposal network, and a joint loss function. For example, an image segmentation model is constructed based on Mask R-CNN, using ResNeXt-101 as the backbone network and fusing an FPN feature pyramid. Anchor boxes with sizes covering 8×8 to 512×512 pixels are generated using an RPN (e.g., a region proposal network). During training, a joint loss function is used in conjunction with a cosine annealing strategy, and training is performed on a self-built wireline dataset. The joint loss function is as follows: Here, classification loss Focal Loss (α=0.25, γ=2.0) is used for bounding box loss. For GIoU Loss, The loss is the mask prediction loss, and λ is the balancing hyperparameter, set to 1.0. In some embodiments, the training images include, but are not limited to, photographs of power transmission lines. For example, a large number of photographs including power transmission lines are taken. The labeled mask for each training image includes, but is not limited to, the conductor region or the conductor + icing region in the training image, and the target region is labeled using a pixel-level segmentation mask. In some embodiments, the labeled mask also includes the pixel coordinates of the boundary of the target region. For example, a first base model is trained using a large number of training images and the labeled masks of the training images, with the training images as input and the labeled masks as output, so that the final image segmentation model outputs image segmentation information after inputting an image, wherein the image segmentation information includes the target region and the pixel-level segmentation mask of the target region.

[0027] In some embodiments, step S13 includes: iteratively scanning the pixels in the target region, deleting pixels that satisfy the first deletion condition and the second deletion condition in each iteration; if, after the iterative scan, there are no pixels in the target region that can be deleted, the remaining pixels in the target region are used as the wire skeleton, wherein the first deletion condition includes: 2≤N( ) ≤6; S( ) = 1; ; The second step of deletion includes: 2 ≤ N ( ) ≤6; S( ) = 1; ; Here, N( (Including pixels) Among the 8 neighboring pixels, the pixel is The number of pixels with the same pattern number (or pixel value), S ( (Including pixels) The number of pixel groups with pixel changes among the neighboring pixels, where two pixels in the same pixel group are adjacent. , , , Including the pixels The neighboring pixels. In some embodiments, pixels of the same type include, but are not limited to, pixels with the same type as the pixel. Pixels with the same pattern count. For example, pixels in the target region have a pattern count of 1 and are displayed as white in the image segmentation information; pixels outside the target region have a pattern count of 0 and are displayed as black in the image segmentation information. For example, the Zhang-Suen skeleton thinning algorithm is used to perform morphological processing on the segmented wire binary image (e.g., image segmentation information), and boundary pixels that meet the conditions (e.g., the first step deletion condition, the second step deletion condition) are deleted through 8-neighborhood iterative scanning. After 8 iterations, a single-pixel wide skeleton is obtained. For example, using pixels... For example: pixel , , , , , , , Yes, 8 neighboring pixels, and, , , , , , , , Adjacent in clockwise order. Statistics. ~ The number of white dots, for example, if the number is 4, then 2 ≤ N ( ≤6. Traverse clockwise. ~ This counts the number of times the pattern number changes from 0 to 1. For example, if... , , , , , , , Then S( ) = 3, which does not satisfy condition S ( =1. The conditions are met. .satisfy However, because condition S is not satisfied... ) = 3, The first step's deletion condition is not met. The specific judgment process for the second step's deletion condition is the same as or similar to that of the first step, and will not be elaborated here. For example, in the first iteration: scan all foreground pixels (e.g., pixels in the target region). For each pixel, determine whether it meets the four sub-conditions of the first step's deletion condition. If it does, mark it as to be deleted. In the second iteration: for the remaining foreground pixels, determine whether they meet the four sub-conditions of the second step's deletion condition. If they do, mark them as to be deleted. Delete the pixels marked for deletion in the target region. If no pixels are deleted in any of the chain wheel iterations, the algorithm ends; otherwise, repeat the iteration steps. Finally, the pixels in the target region that have not been deleted are used as the wire skeleton.

[0028] In some embodiments, step S14 includes: for each pixel within the target area, calculating the shortest Euclidean distance between the pixel and pixels on the conductor skeleton based on a 3×3 neighborhood chamfer distance algorithm to obtain multiple shortest Euclidean distances; determining the maximum Euclidean distance from the multiple shortest Euclidean distances; and determining the actual maximum icing thickness information based on the maximum Euclidean distance and the pixel mapping distance, wherein... Here, T includes the actual maximum icing thickness information, and D includes the maximum Euclidean distance. This includes pixel mapping distance. In some embodiments, Here, The target region includes pixels (x, y), and (x', y') includes pixels on the guide frame. In some embodiments, the 3×3 neighborhood chamfer distance algorithm includes: dividing all pixels within the target region into multiple 3×3 regions, each 3×3 region including 3 rows of pixels and 3 columns of pixels, for a total of 9 pixels. The Euclidean distance between the middle pixel within the 3×3 region and all pixels on the guide frame is calculated, and the shortest Euclidean distance (e.g., D1) is taken. The shortest Euclidean distance D2 = D1 + A for pixels directly above, below, to the left, and to the right of the middle pixel, and the shortest Euclidean distance D3 = D1 + B for pixels diagonally to the middle pixel, where A and B include the cost distance from the center pixel to each neighbor within the 3×3 region. For example, A = 1, B = 1.4. The 3×3 neighborhood chamfer distance algorithm can significantly improve computation speed. The maximum Euclidean distance is selected from all the shortest Euclidean distances obtained within the target area, and the actual maximum icing thickness information is determined based on the pixel mapping distance. In some embodiments, the pixel mapping distance... Here, This includes the actual length of the calibration plate, and Lpixel includes the pixel length of the calibration plate in the image. Pixel distance in the image is converted to actual distance using pixel mapping distance. This yields the actual maximum icing thickness information.

[0029] In some embodiments, the meteorological prediction model is trained using the following method: a second base model is trained using multiple sets of meteorological feature data and the icing index corresponding to each set of meteorological feature data to obtain the meteorological prediction model. The model architecture of the second base model includes the XGBoost algorithm and an objective function. In some embodiments, each set of meteorological feature data includes, but is not limited to, temperature, humidity, wind speed, snowfall, precipitation type, and historical icing thickness. Different sets of meteorological feature data correspond to different icing indices. For example, the lower the temperature, the higher the humidity, the greater the wind speed, and the greater the snowfall, the larger the corresponding icing index. In some embodiments, a large number of actual icing records are used, for example, recording the actual icing thickness corresponding to multiple sets of actual meteorological feature data. The icing index corresponding to the set of meteorological feature data is determined based on the actual icing thickness; the greater the actual icing thickness, the larger the corresponding icing index. The second base model is trained using a large number of meteorological feature data and the icing index corresponding to each set of meteorological feature data to obtain the meteorological prediction model. When the target meteorological feature data of the transmission line is input into the meteorological prediction model, the icing index is output by the meteorological prediction model. Specifically, XGBoost is used as the second base model, and mean squared error is used as the loss function, as shown in the following formula: ,in, To represent the actual icing thickness, This is the icing index predicted by the model. A regularization term is used to prevent overfitting; the formula is as follows: Where T is the number of leaf nodes. The weights of the leaf nodes are... and Here, represents the regularization coefficient. Combining the loss function and the regularization term, we obtain the objective function of XGBoost: Where K is the total number of trees. The objective function is minimized by iteratively adding decision trees to obtain the optimal model parameters. During model training, the model parameters are optimized using cross-validation and grid search.

[0030] Figure 2The diagram illustrates the structure of a transmission line icing monitoring device according to an embodiment of this application. The device includes modules one, two, three, four, five, and six. Module one is used to acquire raw image information about the transmission line and target meteorological feature data of the location of the transmission line towers. Module two is used to segment the raw image information using an image segmentation algorithm to obtain image segmentation information, wherein the image segmentation information includes a target region and a pixel-level segmentation mask of the target region. Module three is used to extract the conductor skeleton from the image segmentation information using a skeleton thinning algorithm. The system obtains skeleton image information, which includes the conductor skeleton of the transmission line; Module 14 is used to determine the actual maximum icing thickness of the transmission line based on image segmentation information, skeleton image information, and pixel mapping distance; Module 15 is used to input target meteorological feature data into the meteorological prediction model and output the icing index of the transmission line through the meteorological prediction model; Module 16 is used to determine the icing risk index of the transmission line based on the actual maximum icing thickness information, icing index, maximum icing thickness that the wire can withstand, current temperature information of the transmission line, freezing point temperature of water, image weight, meteorological weight, and temperature smoothing factor.

[0031] Here, the specific implementation methods corresponding to modules one, two, three, four, five, and six are the same as or similar to the specific embodiments of steps S11, S12, S13, S14, S15, and S16 above, and therefore will not be repeated here, but are included by reference.

[0032] Figure 3 A schematic diagram of a transmission line icing monitoring method according to another embodiment of this application is shown. It includes: Step 1, deploying industrial-grade high-definition cameras at key towers of the target line, acquiring JPEG format images (e.g., raw image information) at 10-minute intervals and simultaneously recording EXIF ​​metadata (including timestamps and GPS coordinates). Gridded weather forecast data (e.g., meteorological grid data) within the target line area is obtained from a meteorological center every 5 minutes. Data fields include temperature, humidity, wind speed, precipitation probability, and weather phenomenon codes. All devices use the NTP protocol for clock synchronization to ensure strict correspondence between image acquisition time and meteorological data release time.

[0033] Step 2: Using a bilinear interpolation algorithm, based on grid data provided by the meteorological center, an interpolation window is constructed centered on the GPS coordinates of the transmission line towers. By calculating the distance weights from the target point (e.g., the location of the tower) to four surrounding grid points, meteorological elements such as temperature and humidity are dynamically mapped from the geographic coordinate system to the physical coordinate system of the transmission line. The formula is as follows: ,in, and These are the coordinates of the four vertices of the weather grid.

[0034] Step 3: Construct a correction mapping relationship based on the Brown-Conrady distortion model. Using pre-calibrated radial and tangential distortion coefficients, a polynomial fitting algorithm is used to perform a nonlinear transformation on the image pixel coordinates. The formula is as follows: ; Where (x,y) represents the normalized image coordinates, () indicates the position of the undistorted projection point. , The radial distortion coefficient is... , For example, this represents the tangential distortion coefficient. This yields the corrected image.

[0035] Step 4: The transmission line image (e.g., the corrected image) is enhanced using the multi-scale Retinex algorithm. Illumination components are estimated using Gaussian kernels at three different scales (σ=15, 80, 250). Reflectance component enhancement is performed in the Log domain, and dynamic range compression and nonlinear tone mapping are applied. The formula is as follows: log{R}_{MSR}(x,y)=\sum ^{n}_{k=1} {{w}_{k}}(logI(x,y)-log[{F}_{k}(x,y)\times I(x,y)]) ,in, It is the original input image. This represents the Gaussian filter kernel at the k-th scale. This represents the weight of the k-th scale, and N represents the number of scales. For example, this yields an enhanced image.

[0036] Step 5: Construct a transmission line instance segmentation model (e.g., an image segmentation model) based on Mask R-CNN, using ResNeXt-101 as the backbone network and fusing it with the FPN feature pyramid. Anchor boxes are generated via RPN, covering sizes from 8×8 to 512×512 pixels. During training, a joint loss function is used in conjunction with a cosine annealing strategy on a self-built conductor dataset (e.g., multiple training images and labeled masks for each training image). The joint loss function is as follows: Among them, classification loss Focal Loss (α=0.25, γ=2.0) is used for bounding box loss. For GIoU Loss, Let λ be the mask prediction loss, and let λ be the balancing hyperparameter, set to 1.0.

[0037] Step 6: Ice Thickness Calculation. The Zhang-Suen skeleton thinning algorithm is used to perform morphological processing on the segmented binary image of the conductor. Boundary pixels satisfying conditions (such as 2≤N(p1)≤6, S(p1)=1, p2×p4×p6=0, etc.) are deleted through 8-neighborhood iterative scanning. A single-pixel wide skeleton is obtained after 8 iterations. The axial width distribution of the skeleton is calculated based on calibration parameters, and the maximum radial expansion of the icing area (e.g., the target area) is taken as the thickness value, as shown in the following formula: First step deletion condition: 2≤N( ) ≤6; S( ) = 1; ; ; Second step deletion condition: 2≤N( ) ≤6; S( ) = 1; ; .

[0038] in, Represents pixels The number of attractions within the 8 neighboring areas, S( ) represents a pixel The number of 0-1 modes, , , , Represents pixels The 8 neighboring pixels.

[0039] Step 7: Based on the skeleton thinning results, perform Euclidean distance transformation, and use the 3×3 neighborhood chamfer distance algorithm to calculate the maximum inscribed circle radius from each pixel (e.g., each pixel within the target area) to the background. Combine this with calibration parameters to generate a thickness mapping map. The formula is as follows:

[0040] in This represents the Euclidean distance from pixel (x,y) to the wire skeleton, and (x',y') represents the pixel coordinates on the wire skeleton.

[0041] Step 8, Pixel Calibration: Perform pixel calibration on the camera using stereo vision reconstruction or a traditional calibration board method, establishing the transformation relationship between pixel coordinates and actual physical coordinates. The transformation coefficient (Scale) is determined using the formula... The calculations are as follows (Lreal is the actual length of the calibration board, and Lpixel is the pixel length of the calibration board in the image); subsequently, images of the icy conductor are acquired, preprocessed, and then the single-pixel width skeleton of the conductor is extracted using a skeleton thinning algorithm; next, a distance transform is performed on the skeleton image to generate a distance map, where each pixel value represents the distance from that point to the nearest conductor skeleton pixel; finally, the distance is calculated using the formula... Calculate the actual thickness of the icing (T is the thickness of the icing). (This is the maximum distance value in the distance map).

[0042] Step 9: Meteorological Data Feature Engineering. Select key features such as temperature, humidity, wind speed, precipitation, precipitation type, and historical icing thickness as inputs to the model (e.g., a meteorological forecasting model). Considering the cumulative effect of icing growth and meteorological conditions, add time-series features, such as the average temperature and humidity change rate over the past 24 hours.

[0043] Step 10: Meteorological data feature scaling. The Min-Max normalization method was used to scale the feature values ​​to the [0, 1] interval. This step is achieved using the following formula: ; Where x is the original feature value, and These are the minimum and maximum values ​​of the feature, respectively. These are the normalized eigenvalues.

[0044] Step 11: Use XGBoost as the weather forecasting model. Use the mean squared error as the loss function, as shown in the following formula: ; in, To represent the actual icing thickness, This is the icing index predicted by the model.

[0045] Regularization is used to prevent overfitting of the model, and the formula is as follows: Where T is the number of leaf nodes. The weights of the leaf nodes are... and is the regularization coefficient.

[0046] Combining the loss function and the regularization term, we obtain the objective function of XGBoost: , Here, K represents the total number of trees. The objective function is minimized by iteratively adding decision trees to obtain the optimal model parameters. During model training, the model parameters are optimized using cross-validation and grid search.

[0047] Step 12: Using an improved risk assessment function, combined with the relative value of icing thickness, the icing index correction term, and the temperature sensitivity factor, the weight allocation of each data source is determined through the analytic hierarchy process (AHP). The specific formula is as follows: ; In image segmentation contribution ( In this context, H represents the real-time icing thickness (e.g., actual maximum icing thickness information), calculated by step 2.2; Maximum ice-resistant thickness of the conductor design; Image data weights (typical value 0.7); In the contribution of meteorological forecasting ( )middle, The icing index is represented by the meteorological forecasting model; T represents the ambient temperature (°C). =0℃ indicates the freezing point temperature of water; Indicates the temperature smoothing factor (to prevent the denominator from being zero); This indicates the weight of the meteorological data (typical value 0.3).

[0048] Furthermore, based on the risk assessment value output in step 11, the degree of icing warning is divided, and the warning guidelines are as follows: 0.6≤R<0.8 indicates that icing has entered a rapid accumulation period, 0.8≤R<1.0 indicates that it is close to the theoretical load threshold of the transmission line, 1.0≤R<1.2 indicates that it exceeds the line design load, and R≥1.2 indicates an extreme dangerous situation.

[0049] In addition to the methods and devices described in the above embodiments, this application also provides a computer-readable storage medium storing computer code that, when executed, performs the method described in any of the preceding embodiments.

[0050] This application also provides a computer program product that, when executed by a computer device, performs the method described in any of the preceding claims.

[0051] This application also provides a computer device, the computer device comprising: One or more processors; Memory, used to store one or more computer programs; When the one or more computer programs are executed by the one or more processors, the one or more processors cause the one or more processors to perform the method as described in any of the preceding methods.

[0052] Figure 4 Exemplary systems that can be used to implement the various embodiments described in this application are shown; like Figure 4 As shown in some embodiments, system 300 can function as any of the devices described in each of the embodiments. In some embodiments, system 300 may include one or more computer-readable media having instructions (e.g., system memory or NVM / storage device 320) and one or more processors (e.g., one or more processors 305) coupled to the one or more computer-readable media and configured to execute the instructions to implement the module and thus perform the actions described in this application.

[0053] In one embodiment, the system control module 310 may include any suitable interface controller to provide any suitable interface to at least one of the processors 305 and / or any suitable device or component communicating with the system control module 310.

[0054] The system control module 310 may include a memory controller module 330 to provide an interface to the system memory 315. The memory controller module 330 may be a hardware module, a software module, and / or a firmware module.

[0055] System memory 315 can be used, for example, to load and store data and / or instructions for system 300. In one embodiment, system memory 315 may include any suitable volatile memory, such as suitable DRAM. In some embodiments, system memory 315 may include double data rate type quad synchronous dynamic random access memory (DDR4 SDRAM).

[0056] In one embodiment, the system control module 310 may include one or more input / output (I / O) controllers to provide interfaces to the NVM / storage device 320 and (one or more) communication interfaces 325.

[0057] For example, NVM / storage device 320 may be used to store data and / or instructions. NVM / storage device 320 may include any suitable non-volatile memory (e.g., flash memory) and / or may include any suitable (one or more) non-volatile storage devices (e.g., one or more hard disk drives (HDDs), one or more optical disc drives (CDs), and / or one or more digital universal optical disc (DVD) drives).

[0058] NVM / storage device 320 may include storage resources that are physically part of a device on which system 300 is mounted, or that can be accessed by the device without necessarily being part of it. For example, NVM / storage device 320 may be accessed via a network through one or more communication interfaces 325.

[0059] One or more communication interfaces 325 may provide the system 300 with an interface to communicate over one or more networks and / or with any other suitable device. The system 300 may wirelessly communicate with one or more components of a wireless network in accordance with any of one or more wireless network standards and / or protocols.

[0060] In one embodiment, at least one of the processors 305 may be logically packaged with one or more controllers of the system control module 310 (e.g., memory controller module 330). In one embodiment, at least one of the processors 305 may be logically packaged with one or more controllers of the system control module 310 to form a system-in-package (SiP). In one embodiment, at least one of the processors 305 may be integrated with the logic of one or more controllers of the system control module 310 on the same die. In one embodiment, at least one of the processors 305 may be integrated with the logic of one or more controllers of the system control module 310 on the same die to form a system-on-a-chip (SoC).

[0061] In various embodiments, system 300 may be, but is not limited to, a server, workstation, desktop computing device, or mobile computing device (e.g., laptop computing device, handheld computing device, tablet computer, netbook, etc.). In various embodiments, system 300 may have more or fewer components and / or different architectures. For example, in some embodiments, system 300 includes one or more cameras, a keyboard, a liquid crystal display (LCD) screen (including a touchscreen display), a non-volatile memory port, multiple antennas, a graphics chip, an application-specific integrated circuit (ASIC), and a speaker.

[0062] It should be noted that this application can be implemented in software and / or a combination of software and hardware, for example, using an application-specific integrated circuit (ASIC), a general-purpose computer, or any other similar hardware device. In one embodiment, the software program of this application can be executed by a processor to implement the steps or functions described above. Similarly, the software program of this application (including related data structures) can be stored in a computer-readable recording medium, such as RAM memory, a magnetic or optical drive, a floppy disk, or similar devices. Furthermore, some steps or functions of this application can be implemented in hardware, for example, as circuitry that cooperates with a processor to perform the various steps or functions.

[0063] Furthermore, a portion of this application can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to this application through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0064] Communication media include media through which communication signals containing, for example, computer-readable instructions, data structures, program modules, or other data are transmitted from one system to another. Communication media can include guided transmission media (such as cables and wires (e.g., optical fibers, coaxial cables, etc.)) and wireless (unguided transmission) media capable of propagating energy waves, such as sound, electromagnetic, RF, microwave, and infrared. Computer-readable instructions, data structures, program modules, or other data can be embodied as modulated data signals in, for example, wireless media (such as carrier waves or similar mechanisms embodied as part of spread spectrum technology). The term "modulated data signal" refers to a signal whose one or more characteristics are altered or set in a manner that encodes information in the signal. Modulation can be analog, digital, or a hybrid modulation technique.

[0065] By way of example and not limitation, computer-readable storage media may include volatile and non-volatile, removable and non-removable media implemented by any method or technique for storing information such as computer-readable instructions, data structures, program modules or other data. For example, computer-readable storage media include, but are not limited to, volatile memories such as random access memory (RAM, DRAM, SRAM); and non-volatile memories such as flash memory, various read-only memories (ROM, PROM, EPROM, EEPROM), magnetic and ferromagnetic / ferroelectric memories (MRAM, FeRAM); and magnetic and optical storage devices (hard disks, magnetic tapes, CDs, DVDs); or other media now known or hereafter developed capable of storing computer-readable information / data for use by a computer system.

[0066] Herein, one embodiment of this application includes an apparatus comprising a memory for storing computer program instructions and a processor for executing the program instructions, wherein when the computer program instructions are executed by the processor, the apparatus is triggered to run a method and / or technical solution based on the foregoing embodiments of this application.

[0067] It will be apparent to those skilled in the art that this application is not limited to the details of the exemplary embodiments described above, and that this application can be implemented in other specific forms without departing from the spirit or essential characteristics of this application. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of this application is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within this application. No reference numerals in the claims should be construed as limiting the scope of the claims. Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices recited in the apparatus claims may also be implemented by a single unit or device in software or hardware. The terms "first," "second," etc., are used to indicate names and do not indicate any particular order.

Claims

1. A method for monitoring icing on transmission lines, characterized in that, The method includes: Obtain raw image information about the transmission line and target meteorological feature data of the location of the towers of the transmission line; The original image information is segmented using an image segmentation algorithm to obtain image segmentation information, wherein the image segmentation information includes a target region and a pixel-level segmentation mask for the target region; The image segmentation information is processed by a skeleton thinning algorithm to extract the conductor skeleton, thereby obtaining skeleton image information, wherein the skeleton image information includes the conductor skeleton of the transmission line. The actual maximum icing thickness of the transmission line is determined based on the image segmentation information, the skeleton image information, and the pixel mapping distance. The target meteorological characteristic data is input into the meteorological prediction model, and the icing index of the transmission line is output through the meteorological prediction model. The icing risk index of the transmission line is determined based on the actual maximum icing thickness information, the icing index, the maximum icing thickness that the power line can withstand, the current temperature information of the transmission line, the freezing point temperature of water, image weight, meteorological weight, and temperature smoothing factor.

2. The method according to claim 1, characterized in that, The acquisition of target meteorological characteristic data at the location of the transmission line towers includes: Target meteorological grid data is obtained from the meteorological center based on the location information of the poles, wherein the target meteorological network data covers the location information of the poles and includes multiple reference meteorological feature data; The target meteorological feature data is determined using a bilinear interpolation algorithm and the target meteorological grid data.

3. The method according to claim 1, characterized in that, The method calculates by image segmentation. The method performs image segmentation on the original image information to obtain image segmentation information, and the preceding steps include: The original image information is corrected using a distortion model algorithm to obtain a corrected image, wherein the distortion model algorithm includes... ; ； Here, (x, y) includes the pixel coordinates in the original image information, the Including the corrected pixel coordinates, the , Including the radial distortion coefficient, the , Including tangential distortion coefficient; The corrected image is enhanced using a multi-scale retinal cortex theory algorithm to obtain an enhanced image. The step of segmenting the original image information using an image segmentation algorithm to obtain image segmentation information includes: The enhanced image is input into the image segmentation model, and the image segmentation model outputs the image segmentation information.

4. The method according to claim 3, characterized in that, The image segmentation model was trained using the following method: The image segmentation model is obtained by training a first base model using multiple training images and the labeled mask of each training image. The labeled mask includes the target region and the pixel-level segmentation mask of the target region. The model architecture of the first base model includes a ResNeXt-101+FPN backbone network, a region proposal network, and a joint loss function.

5. The method according to claim 1, characterized in that, The step of extracting the wire skeleton from the image segmentation information using a skeleton thinning algorithm to obtain skeleton image information includes: The target region is iteratively scanned for pixels. In each iteration, pixels that meet the first deletion condition and the second deletion condition are deleted. If, after iterative scanning, there are no pixels in the target region that can be deleted, the remaining pixels in the target region at this time are used as the wire skeleton. The first step deletion condition includes: 2 ≤ N ( ) ≤6; S( ) = 1; ; The second step of deletion includes: 2≤N( ) ≤6; S( ) = 1; ; Here, N( (Including pixels) Among the 8 neighboring pixels, the pixel is The number of pixels with the same pattern number, the S ( (Including pixels) The number of pixel groups with pixel changes in the neighboring pixels, wherein two pixels in the same pixel group are adjacent. , , , Including the pixels The neighboring pixels.

6. The method according to claim 1, characterized in that, The step of determining the actual maximum icing thickness information of the transmission line based on the image segmentation information, the skeleton image information, and the pixel mapping distance includes: For each pixel within the target area, the shortest Euclidean distance between the pixel and the pixels on the wire skeleton is calculated based on the 3×3 neighborhood chamfer distance algorithm to obtain multiple shortest Euclidean distances; Determine the maximum Euclidean distance from the plurality of shortest Euclidean distances; The actual maximum icing thickness information is determined based on the maximum Euclidean distance and the pixel mapping distance, wherein... Here, T includes the actual maximum icing thickness information, D includes the maximum Euclidean distance, and the... This includes pixel mapping distance.

7. The method according to claim 1, characterized in that, The meteorological prediction model was trained using the following method: The second basic model is trained by multiple sets of meteorological feature data and the actual icing thickness corresponding to each set of meteorological feature data to obtain the meteorological prediction model. The model architecture of the second basic model includes the XGBoost algorithm and the objective function.

8. The method according to claim 1, characterized in that, The determination of the icing risk index of the transmission line based on the actual maximum icing thickness information, the icing index, the maximum icing thickness that the power line can withstand, the current temperature information of the transmission line, the freezing point temperature of water, image weights, meteorological weights, and temperature smoothing factors includes: ， Here, R includes the icing risk index, the Including the image weights, the Including the meteorological weights, H includes the actual maximum icing thickness information, the This includes information on the maximum icing thickness that the wire can withstand, where T includes the ambient temperature. Including the icing index, the Including the freezing point temperature information of the water, Includes temperature smoothing factor.

9. A computer device, characterized in that, It includes a memory and a processor, wherein the memory stores a transmission line icing monitoring method that can be loaded by the processor and executed as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, The system stores a transmission line icing monitoring method that can be loaded by a processor and executed as described in any one of claims 1 to 8.

Images