A method, system, device and medium for predicting wind direction and wind speed of a transmission line

Abstract

The present invention belongs to the technical field of power grid monitoring, and relates to a method for predicting wind direction and wind speed of a transmission line. The pixel point area of ​​each wire in a video frame is detected by a target detection method, and then the pixel point area of ​​the wire between two frames is matched; the difference in depth values ​​of the two frames of images in the overlapping area is calculated by a binocular matching method, and the difference in depth values ​​is used as the longitudinal offset of the wire; the average distance between each unique area of ​​the two frames of images is calculated to obtain the lateral offset of the wire; the wind direction and wind speed on the day of video shooting are obtained by crawling historical weather data, and the two are used as supervision information, and the longitudinal and lateral offsets of the wire are used as training samples to train a random forest classification model, and the categories of the two attributes of wind direction and wind speed are obtained on the test data. The disturbance of the transmission line is calculated by the pixel matching method, which can be better applied to the prediction of wind force in strong wind weather, and an early warning is given according to the wind force level to ensure stable power supply.

Description

Technical Field

[0001] The present invention belongs to the technical field of power grid monitoring, and in particular relates to a method, system, equipment and medium for predicting wind direction and wind speed of a power transmission line. Background Art

[0002] Most existing methods for detecting wind direction and wind speed in transmission line scenarios rely on dense sensors that collect micro-weather information such as humidity, temperature, wind speed, and air pressure near power facilities. These sensors then arrange the information in chronological order and input it into a sequence modeling model based on deep learning to predict the output. The above methods require large data inputs and have many hyperparameters, making them difficult to tune. At the same time, since sensors can only be placed near power facilities, the above methods cannot estimate the disturbance of power lines well. Summary of the invention

[0003] The purpose of the present invention is to provide a method, system, device and medium for predicting wind direction and wind speed of a power transmission line, which solves the problem that the disturbance of the power line cannot be well estimated at present.

[0004] The present invention is achieved through the following technical solutions:

[0005] A method for predicting wind direction and wind speed of a power transmission line comprises the following steps:

[0006] S1, detecting the pixel area of ​​each wire in the video frame by using a target detection method;

[0007] S2, matching the pixel regions of the wires between the two frames of images, and obtaining the overlapping regions and the respective unique regions of a group of wires at the same position;

[0008] S3, calculating the difference in depth values ​​of the two frames of images in the overlapping area by a binocular matching method, and using the difference in depth values ​​as the longitudinal offset of the wire;

[0009] Calculate the average distance between the unique regions of the two frames of images, and use the average distance as the lateral offset of the wire;

[0010] S4. According to the time and location of the camera shooting, crawl the historical weather data to obtain the wind direction and wind speed at the specified time and location;

[0011] The longitudinal offset and lateral offset obtained by S3 are combined with wind direction and wind speed to form training data, which are input into the random forest classification model for training to obtain a prediction model;

[0012] S5. Acquire the actual image, and after processing in S1-S3, input the image into the prediction model to predict the wind direction and wind speed in the transmission line scene.

[0013] Furthermore, in S1, the Cascade Mask R-CNN model is used to detect the wires in the image captured by the transmission line monitoring camera, and the pixel position of each wire is output; all pixel positions are connected to form a pixel area;

[0014] The pixel position is represented by a 0-1 mask of the same size as the image resolution, where 1 represents the pixel being a wire and 0 represents the pixel being the background.

[0015] Further, S2 is specifically:

[0016] For each wire region detected in the first image, the number of intersecting "1"s with each wire region in the second image is calculated, and the region with the most intersecting "1"s in the second image is taken as a matching region of the wire region in the first image;

[0017] Set the threshold of intersection "1" to X. If the number of intersection "1" between two regions is less than X, the two regions are judged to be mismatched; if the number of intersection "1" between two regions is greater than or equal to X, the two regions are judged to be matched.

[0018] The overlapping area and unique area of ​​a set of successfully matched wires are calculated.

[0019] Furthermore, before matching, the two regions of the two frames of images are respectively subjected to the morphological dilation method, and then the intersection is taken.

[0020] Further, in S3, the specific process of calculating the difference in depth values ​​of the two frames of images in the overlapping area by the binocular matching method is as follows:

[0021] The relevant parameters of the camera are obtained through binocular camera calibration;

[0022] Through relevant parameters, the right image is fused with the left image, and the depth map of the 3D reconstructed image is obtained;

[0023] The depth maps of the two frames are subtracted and combined with the overlapping area obtained by S2 to obtain the difference in depth values.

[0024] Furthermore, the relevant parameters of the camera include focal length, imaging origin, radial distortion parameter, tangential distortion parameter and the relative position between the binocular cameras.

[0025] Further, in S3, the specific process of calculating the average distance between the unique regions of the two frames of images is as follows:

[0026] Considering the different numbers of pixels in the two regions, the shortest Euclidean distance between each pixel in the unique region with fewer pixels and any pixel in the other unique region is calculated, and finally the shortest Euclidean distances are added and divided by the total number of pixels to obtain the average distance between the unique regions of each group of wires;

[0027] At the same time, the relative position relationship between the two areas is considered to determine the direction of the lateral offset of the wire, and the offset direction of each pair of pixel points matched by the shortest Euclidean distance is counted;

[0028] Finally, the average distances of each set of unique regions are added and averaged to obtain the final average distance.

[0029] The present invention also discloses a system for predicting wind direction and wind speed in a transmission line scenario, comprising:

[0030] A data acquisition module is used to detect the pixel area of ​​each wire in the video frame by using a target detection method based on the data collected by the binocular camera;

[0031] A pixel matching module is used to match the pixel area of ​​the wires between two frames of images to obtain the overlapping area and the unique area of ​​a group of wires at the same position;

[0032] A longitudinal offset calculation module is used to calculate the difference in depth values ​​of two frames of images in the overlapping area by a binocular matching method, and use the difference in depth values ​​as the longitudinal offset of the wire;

[0033] A lateral offset calculation module is used to calculate the average distance between the unique areas of the two frames of images, and use the average distance as the lateral offset of the wire;

[0034] The historical data acquisition module is used to crawl historical weather data according to the time and location of camera shooting, and obtain the wind direction and wind speed at the specified time and location;

[0035] The training module is used to combine the obtained longitudinal offset and lateral offset with wind direction and wind speed to form training data, input them into the random forest classification model for classification, and obtain a prediction model.

[0036] The present invention also discloses a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method for predicting the wind direction and wind speed of a transmission line when executing the computer program.

[0037] The present invention also discloses a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for predicting the wind direction and wind speed of a power transmission line are implemented.

[0038] Compared with the prior art, the present invention has the following beneficial technical effects:

[0039] The invention discloses a method for predicting the wind direction and wind speed of a transmission line. The pixel point area of ​​each wire in a video frame is detected by a target detection method, and then the pixel point area of ​​the wire between two frames is matched; thereafter, the difference in depth values ​​of the two frames of images in the overlapping area is calculated by a binocular matching method, and the difference in depth values ​​is used as the longitudinal offset of the wire; the lateral offset of the wire is obtained by calculating the average distance between each unique area of ​​the two frames of images; finally, the wind direction and wind speed on the day of video shooting are obtained by crawling historical weather data, and the two are used as supervision information, and the longitudinal and lateral offsets of the wire are used as training samples to train a random forest classification model, and the categories of the two attributes of wind direction and wind speed are obtained on the test data. The disturbance of the transmission line is calculated by the pixel matching method, which has the characteristics of real-time and rapidity, can be better applied to the prediction of wind force in strong wind weather, and an early warning is given according to the wind force level to ensure stable power supply, which is of great significance to the maintenance of the transmission line. The present invention can directly observe the disturbance of the transmission line in a windy environment by means of target detection, and simultaneously consider the lateral and longitudinal offsets of the power line, which is more comprehensive and accurate than the previous method of only considering the lateral offset; in calculating the longitudinal offset, the current main method is a monocular depth estimation method based on deep learning. The main problem with the above method is that it is only applicable to specific low-resolution data sets, and has low accuracy in high-resolution transmission line scenarios. The binocular matching method can calculate the three-dimensional coordinates pixel by pixel, and thus has higher accuracy; the present invention adopts a simple pixel matching method, comprehensively considers the lateral and longitudinal offsets of the power line, and obtains a more accurate overall offset, thereby predicting the wind force and direction based on the overall offset, which has the advantages of simplicity, real-time and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is the overall flow chart of the present invention.

[0041] Figure 2 It is a schematic diagram of the unique area and overlapping area of ​​the wires.

[0042] Figure 3 is a schematic diagram of calculating the average depth difference.

[0043] Figure 4 It is a schematic diagram for calculating the average distance of unique areas. DETAILED DESCRIPTION

[0044] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] like Figure 1 As shown, the present invention discloses a method for predicting wind direction and wind speed of a power transmission line, comprising the following steps:

[0046] S1. Detect the pixel area of ​​each wire in the video frame through the target detection method:

[0047] Train a Cascade Mask R-CNN model to detect wires in images taken by a power line surveillance camera and output the pixel location of each wire. The pixel location is represented by a 0-1 mask of the image resolution, where 1 represents the pixel as a wire and 0 represents the pixel as background.

[0048] S2. Match the pixel areas of the wires between the two frames of images to obtain the overlapping areas and unique areas of a group of wires at the same position:

[0049] For each wire area detected in one image, the number of intersecting "1"s with each wire area in the second image is calculated, and the area with the most intersecting "1"s in the second image is taken as a match for the wire area in the first image. In order to prevent the wires in the two images from not being detected by the detection model in the first step, resulting in unsuccessful matching, the threshold of intersecting "1" is set to 80. If the number of intersecting "1"s in two areas is less than 80, the two areas are considered to be unmatched. In order to prevent the disturbance of the wires from causing excessive displacement of the pixel position, the morphological dilation method is used for the two areas respectively, and then the intersection is taken. Finally, the overlapping area and unique area of ​​a group of successfully matched wires are calculated.

[0050] S3. Calculate the difference in depth values ​​of the two frames of images in the overlapping area by binocular matching method, and use the difference in depth values ​​as the longitudinal offset of the wire:

[0051] Specifically, since the image is collected by a binocular camera, each frame of the image has a left-eye image and a right-eye image. At the same time, the camera calibration can be used to obtain the focal length, imaging origin, radial distortion parameters, tangential distortion parameters, relative position between binocular cameras and other parameters of the camera. Through the above parameters, the right-eye image can be fused to the left-eye image, and the depth map of the three-dimensional reconstructed image can be obtained. Finally, the depth maps of the two frames are subtracted, and the difference in depth values ​​can be obtained by combining the overlapping area obtained in the above steps. Unless otherwise specified, "frame image" refers to the image captured by the left camera in the binocular camera.

[0052] Calculate the average distance between the unique regions of the two frames and use the average distance as the lateral offset of the wire:

[0053] Considering the different number of pixels in the two regions, the shortest Euclidean distance from each pixel in the unique region with fewer pixels to any pixel in the other unique region is calculated, and finally the distances are added and divided by the total number of pixels to obtain the average distance between each group of unique regions of the wires. At the same time, the relative position relationship of the two regions needs to be considered to determine the direction of the lateral offset of the wires. Therefore, the offset direction of each pair of pixels matched by the shortest Euclidean distance is counted and voted. If the overall offset is to the upper left, the average distance is positive, and if the overall offset is to the lower right, it is negative. Finally, the average distances of each group of unique regions are added and averaged to obtain the final average distance.

[0054] S4. According to the time and location of the camera shooting, crawl the historical weather data to obtain the wind direction and wind speed at the specified time and location;

[0055] The longitudinal offset and lateral offset obtained by S3 are combined with wind direction and wind speed to form training data in CSV format, which are input into the random forest classification model for classification to obtain a prediction model;

[0056] S5. Obtain the actual image, and after S1-S3, input it into the prediction model to predict the wind direction and wind speed in the transmission line scene.

[0057] The present invention builds a Cascade Mask R-CNN target detection model based on the pytorch and mmdetection frameworks. The data set used in the present invention contains 8177 training images and 1500 test images, and the data set is annotated with polygon annotations in the COCO format. The model is trained using the Adam gradient descent method with a learning rate of 1e-5. In the test phase, the pixel position of the wire area is obtained using the pycocotool tool.

[0058] Perform chessboard calibration on the binocular camera. Select a chessboard of appropriate size (for example, a 7*7 square chessboard with alternating black and white and a side length of 25mm), and use the binocular camera to take a batch of chessboard images at different angles and distances. The images need to contain the complete chessboard. After that, use Matlab's Stereo Calibration Toolbox to calibrate the image. During calibration, you need to input the side length of the chessboard square and the path of the left and right eye images. After calibration, both the left and right eye cameras will get the following parameters: focal length, origin, offset (Skew), radial distortion and tangential distortion. At the same time, the relative position relationship between the left and right eye cameras will also be obtained, expressed as translation and rotation. After that, the focal length and origin are respectively composed of the left eye matrix and the right eye matrix, and the radial distortion and tangential distortion are composed of the left eye distortion matrix and the right eye distortion matrix. At this time, introduce the python-opencv library in the python environment. The rotation relationship vector is transformed through cv2.Rodrigues(). Use cv2.stereoRectify() to input parameters such as focal length and origin for stereo correction. Use cv2.initUndistortRectifyMap() to initialize the undistorted maps of the left and right images, and reconstruct them through cv2.remap(), and finally convert them into grayscale images through cv2.cvtColor(). Generate a visual difference map through cv2.StereoBM_create(), and calculate the difference between the two grayscale undistorted maps through the compute() method. Finally, project the above difference into three-dimensional space through cv2.reprojectImageTo3D() to generate a three-dimensional tensor. The value of the tensor in the z direction is the depth value corresponding to each pixel of the image.

[0059] Specifically, wire area matching is performed based on the Python environment: the unique area and overlapping area of ​​the wires are as follows: Figure 2 As shown in the figure, the sum() function of numpy is used to count the number of "1", and the dilate() function of python-opencv is used to dilate the wire area.

[0060] Specifically, the average depth difference is calculated based on the Python environment: use numpy's zeros() function to create an empty mask, then add the masks of each overlapping area to it to get the total mask, then multiply it with the two depth maps and subtract them, and finally divide it by the total number of pixels in the overlapping area to get the average depth difference. The calculation diagram is as follows Figure 3 shown.

[0061] Specifically, the average distance of the unique region is calculated based on the Python environment: the coordinate of "1" is obtained using the where function of numpy, and then for each pixel in a smaller unique region, a matching pixel in another unique region is found so that the Euclidean distance between the two pixels is the shortest. The average distance of a group of unique regions can be obtained by adding and averaging several distances, and the overall average distance can be obtained by analogy. At the same time, a variable is set to count the offset of the pixels in the two regions.

[0062] In order to improve the matching efficiency, the divide-and-conquer method is used instead of the double loop method when searching for the pixel with the shortest Euclidean distance. Figure 4 As shown in the figure, two unique regions are shown on a mask image, distinguished by a bold italic "1" and an underlined "1".

[0063] Specifically, the wind direction and wind speed are classified based on the python environment and the sklearn framework. The wind direction is divided into eight categories: north, south, east, west, northeast, northwest, southeast, and southwest. The wind speed is divided into seven categories: 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, and 4.0. Use sklearn's train_test_split() to divide the training set and the test set, of which the training set accounts for 70%, and the training set has a total of 800 sets of data. Use MultiOutputClassifier(RandomForestClassifier(n_estimators)) as the multi-label training model, where n_estimators is set to 10 after testing. Use classification_report() to print out the test accuracy, which is above 90%.

[0064] The present invention also discloses a system for predicting wind direction and wind speed in a transmission line scene based on binocular vision, comprising:

[0065] A data acquisition module is used to detect the pixel area of ​​each wire in the video frame by using a target detection method based on the data collected by the binocular camera;

[0066] A pixel matching module is used to match the pixel area of ​​the wires between two frames of images to obtain the overlapping area and the unique area of ​​a group of wires at the same position;

[0067] A longitudinal offset calculation module is used to calculate the difference in depth values ​​of two frames of images in the overlapping area by a binocular matching method, and use the difference in depth values ​​as the longitudinal offset of the wire;

[0068] A lateral offset calculation module is used to calculate the average distance between the unique areas of the two frames of images, and use the average distance as the lateral offset of the wire;

[0069] The historical data acquisition module is used to crawl historical weather data according to the time and location of camera shooting, and obtain the wind direction and wind speed at the specified time and location;

[0070] The training module is used to combine the obtained longitudinal offset and lateral offset with wind direction and wind speed to form training data, input them into the random forest classification model for classification, and obtain a prediction model.

[0071] The method for predicting wind direction and wind speed of a power transmission line of the present invention can be implemented in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention can be implemented in the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0072] If the method for predicting the wind direction and wind speed of a transmission line of the present invention is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. Computer-readable storage media include permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules or other data. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals. Among them, the computer storage medium can be any available medium or data storage device that can be accessed by the computer, including but not limited to magnetic storage (such as floppy disks, hard disks, magnetic tapes, magneto-optical disks (MO), etc.), optical storage (such as CD, DVD, BD, HVD, etc.), and semiconductor storage (such as ROM, EPROM, EEPROM, non-volatile memory (NANDFLASH), solid-state drive (SSD)), etc.

[0073] In an exemplary embodiment, a computer device is also provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the steps of the method for predicting the wind direction and wind speed of a power transmission line when executing the computer program. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc.

[0074] The above is a detailed introduction to the method proposed by the present invention for predicting the wind direction and wind speed of the day by detecting power line disturbances. This article applies the current mainstream deep learning and machine learning framework for implementation. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for those of ordinary skill in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation method and scope of application. In summary, the content of this specification should not be understood as a limitation of the present invention.

Claims

1. A method for predicting wind direction and wind speed of a power transmission line, characterized in that: The following steps are involved: S1, detecting the pixel area of ​​each wire in the video frame by using a target detection method; S2, matching the pixel regions of the wires between the two frames of images, and obtaining the overlapping regions and the respective unique regions of a group of wires at the same position; S3, calculating the difference in depth values ​​of the two frames of images in the overlapping area by a binocular matching method, and using the difference in depth values ​​as the longitudinal offset of the wire; Calculate the average distance between the unique regions of the two frames of images, and use the average distance as the lateral offset of the wire; S4. According to the time and location of the camera shooting, crawl the historical weather data to obtain the wind direction and wind speed at the specified time and location; The longitudinal offset and lateral offset obtained by S3 are combined with wind direction and wind speed to form training data, which are input into the random forest classification model for training to obtain a prediction model; S5, obtaining an actual image, and after processing in S1-S3, inputting the image into the prediction model to predict the wind direction and wind speed in the transmission line scene; In S3, the specific process of calculating the average distance between the unique regions of the two frames of images is as follows: Considering the different numbers of pixels in the two regions, the shortest Euclidean distance between each pixel in the unique region with fewer pixels and any pixel in the other unique region is calculated, and finally the shortest Euclidean distances are added and divided by the total number of pixels to obtain the average distance between the unique regions of each group of wires; At the same time, the relative position relationship between the two areas is considered to determine the direction of the lateral offset of the wire, and the offset direction of each pair of pixel points matched by the shortest Euclidean distance is counted; Finally, the average distances of each set of unique regions are added and averaged to obtain the final average distance.

2. A method for predicting wind direction and wind speed of a power transmission line according to claim 1, characterized in that: In S1, the Cascade Mask R-CNN model is used to detect the wires in the image captured by the power line monitoring camera and output the pixel position of each wire; all pixel positions are connected to form a pixel area; The pixel position is represented by a 0-1 mask of the same size as the image resolution, where 1 represents the pixel being a wire and 0 represents the pixel being the background.

3. A method for predicting wind direction and wind speed of a power transmission line according to claim 2, characterized in that: S2 is specifically: For each wire area detected in the first image, the number of intersecting "1"s with each wire area in the second image is calculated, and the area with the most intersecting "1"s in the second image is taken as a matching area of ​​the wire area in the first image; Set the intersection "1" threshold to X. If the number of intersection "1"s between two regions is less than X, the two regions are considered to be mismatched. If the number of intersection "1"s between two regions is greater than or equal to X, the two regions are considered to be matched. The overlapping area and unique area of ​​a set of successfully matched wires are calculated.

4. A method for predicting wind direction and wind speed of a power transmission line according to claim 3, characterized in that: Before matching, the morphological dilation method is used on the two regions of the two frames of images respectively, and then the intersection is taken.

5. A method for predicting wind direction and wind speed of a power transmission line according to claim 1, characterized in that: In S3, the specific process of calculating the difference in depth values ​​of two frames of images in the overlapping area by the binocular matching method is as follows: The relevant parameters of the camera are obtained through binocular camera calibration; Through relevant parameters, the right image is fused with the left image, and the depth map of the 3D reconstructed image is obtained; The depth maps of the two frames are subtracted and combined with the overlapping area obtained by S2 to obtain the difference in depth values.

6. A method for predicting wind direction and wind speed of a power transmission line according to claim 5, characterized in that: The relevant parameters of the camera include focal length, imaging origin, radial distortion parameters, tangential distortion parameters and the relative position between the binocular cameras.

7. A system for predicting wind direction and wind speed in a transmission line scenario, characterized in that: include: A data acquisition module is used to detect the pixel area of ​​each wire in the video frame by using a target detection method based on the data collected by the binocular camera; A pixel matching module is used to match the pixel area of ​​the wires between two frames of images to obtain the overlapping area and the unique area of ​​a group of wires at the same position; A longitudinal offset calculation module is used to calculate the difference in depth values ​​of two frames of images in the overlapping area by a binocular matching method, and use the difference in depth values ​​as the longitudinal offset of the wire; A lateral offset calculation module is used to calculate the average distance between the unique areas of the two frames of images, and use the average distance as the lateral offset of the wire; The historical data acquisition module is used to crawl historical weather data according to the time and location of camera shooting, and obtain the wind direction and wind speed at the specified time and location; A training module is used to combine the obtained longitudinal offset and lateral offset with wind direction and wind speed to form training data, input them into a random forest classification model for classification, and obtain a prediction model; The specific process of calculating the average distance between the unique regions of the two frames of images is as follows: Considering the different numbers of pixels in the two regions, the shortest Euclidean distance between each pixel in the unique region with fewer pixels and any pixel in the other unique region is calculated, and finally the shortest Euclidean distances are added and divided by the total number of pixels to obtain the average distance between the unique regions of each group of wires; At the same time, the relative position relationship between the two areas is considered to determine the direction of the lateral offset of the wire, and the offset direction of each pair of pixel points matched by the shortest Euclidean distance is counted; Finally, the average distances of each set of unique regions are added and averaged to obtain the final average distance.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method for predicting wind direction and wind speed of a power transmission line as claimed in any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method for predicting wind direction and wind speed of a power transmission line as claimed in any one of claims 1 to 6 are implemented.

Images