Method and system for monitoring galloping of power grid line in real time

Abstract

The invention discloses a power grid line galloping real-time monitoring method and system, and relates to the field of information technology monitoring, and the method comprises the steps: obtaining a GNSS satellite carrier observation value, and building a weight matrix; acquiring a current sky image at the monitoring station, and identifying the probability that each pixel in the current sky image belongs to the sky; mapping the azimuth and elevation angle of the satellite to the sky image, and calculating the probability that the satellite is shielded; adjusting the weight of a corresponding observation value according to the probability that the satellite is shielded, updating a weight matrix, and resolving a positioning coordinate; and fusing the positioning coordinate calculated by the GNSS and the positioning coordinate calculated by the INS through Kalman filtering to obtain a final positioning coordinate, and obtaining a line galloping time sequence according to the final positioning coordinate and a corresponding timestamp. According to the method, the satellite shielding probability in each sky image is analyzed, the GNSS observation value weight of each satellite is dynamically adjusted, and the positioning coordinates are solved by using the weighted least square method, so that the positioning precision is improved.

Description

Technical Field

[0001] The present invention relates to the field of information technology monitoring, and in particular to a method and system for real-time monitoring of power grid line galloping based on GNSS, INS and vision fusion. Background Art

[0002] Grid line galloping refers to the periodic or random swinging of transmission lines caused by wind, climate change and other external factors, which may pose a threat to the safety and stability of the power grid. Conventional line galloping monitoring methods mainly use a combination of GNSS positioning modules and inertial measurement units (IMUs) to monitor line galloping through data fusion algorithms. However, in urban canyons, mountainous areas or densely vegetated areas, GNSS signals may be interfered with or blocked, resulting in inaccurate or failed positioning. These environmental factors pose a challenge to the reliability of GNSS positioning systems. Summary of the invention

[0003] In view of the above-mentioned problems, the present invention is proposed.

[0004] Therefore, the problem to be solved by the present invention is that in urban canyons, mountainous areas or densely vegetated areas, GNSS signals may be interfered with or blocked, resulting in inaccurate or invalid positioning.

[0005] To solve the above technical problems, the present invention provides the following technical solutions: a real-time monitoring method for power grid line swaying based on GNSS, INS and visual fusion, which includes: obtaining GNSS satellite carrier observation values, setting the weight of each GNSS satellite carrier observation value to 1, and establishing a weight matrix; obtaining the current sky image at the monitoring station, and identifying the probability that each pixel in the current sky image belongs to the sky through a sky recognition model; mapping the azimuth and altitude angle of the satellite to the current sky image, and calculating the probability that the satellite is blocked by the position of the satellite mapped to the sky image; adjusting the corresponding satellite carrier observation value weight according to the probability that the satellite is blocked, updating the weight matrix, and solving the positioning coordinates by weighted least squares method; obtaining the final positioning coordinates by fusing the positioning coordinates solved by GNSS and the positioning coordinates solved by INS through Kalman filtering, and obtaining the line swaying time series according to the final positioning coordinates and the corresponding timestamps.

[0006] As a preferred solution of the real-time monitoring method of power line dancing based on GNSS, INS and visual fusion described in the present invention, wherein: the network structure of the sky recognition model includes a first convolution layer, a second convolution layer, a third convolution layer, a fourth convolution layer, a fifth convolution layer, a pooling layer, a first transposed convolution layer, a second transposed convolution layer, a third transposed convolution layer, a fourth transposed convolution layer and a probability calculation layer; the first convolution layer is sequentially connected with the second convolution layer, the third convolution layer, the fourth convolution layer, the fifth convolution layer, the pooling layer, the first transposed convolution layer, the second transposed convolution layer, the third transposed convolution layer, the fourth transposed convolution layer and the probability calculation layer. The first convolution layer, the second transposed convolution layer, the third transposed convolution layer, the fourth transposed convolution layer and the probability calculation layer are connected; the first convolution layer is connected to the first transposed convolution layer, the first convolution layer expands the output feature map of the first convolution layer to the same size as the output feature map of the first transposed convolution layer by nearest neighbor interpolation, and then adds the expanded output feature map of the first convolution layer to the output feature map of the first transposed convolution layer element by element to obtain a first fused transposed feature map; the second convolution layer is connected to the second transposed convolution layer, and the second convolution layer adds the output feature map of the first convolution layer to the output feature map of the first transposed convolution layer by nearest neighbor interpolation. The second convolution output feature map is expanded to the same size as the first fused transposed feature map, and then the expanded second convolution layer output feature map is added element by element to the first fused transposed feature map to obtain the second fused transposed feature map; the third convolution layer is connected to the third transposed convolution layer, and the third convolution layer expands the third convolution output feature map to the same size as the second fused transposed feature map through nearest neighbor interpolation, and then the expanded third convolution layer output feature map is added element by element to the second fused transposed feature map to obtain the third fused transposed feature map; the fourth convolution layer is connected to the fourth transposed convolution layer, and the fourth convolution layer expands the fourth convolution output feature map to the same size as the third fused transposed feature map through nearest neighbor interpolation, and then the expanded fourth convolution layer output feature map is added element by element to the third fused transposed feature map to obtain the fourth fused transposed feature map, and the size of the fourth fused transposed feature map is the same as the input image size; the probability calculation layer is a sigmoid activation function, and the score of each pixel in the fourth fused transposed feature map is converted into an occlusion probability through the sigmoid activation function.

[0007] As a preferred solution of the real-time monitoring method of power line dancing based on GNSS, INS and visual fusion described in the present invention, the loss function of the network structure of the sky recognition model is:

[0008]

[0009] Where N represents the number of training images, y i (x,y) represents the true label of the pixel with position coordinates (x,y) in the i-th image. Indicates the probability that the pixel with position coordinates (x, y) in the i-th image is occluded.

[0010] As a preferred solution of the real-time monitoring method of power line dancing based on GNSS, INS and visual fusion described in the present invention, wherein: the mapping of the satellite's azimuth and altitude angle to the current sky image includes obtaining the image length L and width W, obtaining the pitch angle θ and yaw angle ψ of the current sky image, and obtaining the azimuth angle Az of each observable satellite j and the altitude angle El j ; Calculate the local azimuth Az of each observable satellite j ' and the local altitude angle El j ' is expressed as,

[0011] Az j '=Az j -ψ

[0012] El j '=El j +θ

[0013] The distance between the position of each observable satellite mapped to the current sky image and the center point of the current sky image is calculated as:

[0014]

[0015] As a preferred solution of the real-time monitoring method of power line dancing based on GNSS, INS and visual fusion described in the present invention, wherein: the calculation of the probability of satellite being blocked includes querying the blockage probability p corresponding to all pixel points within the influence circle with radius d and the position of each observable satellite mapped to the current sky image as the center j (x,y);

[0016] Calculate the weight of each pixel k in the circle

[0017] Calculate the probability P of each observable satellite being blocked j , the calculation method is as follows:

[0018]

[0019] Among them, r k It represents the distance between the kth pixel point in the influence circle and the center point of the influence circle, s represents the set weight index, j represents the number of the observable satellite, and M represents the total number of pixels in the influence circle.

[0020] As a preferred solution of the real-time monitoring method of power line dancing based on GNSS, INS and visual fusion described in the present invention, wherein: the weight of the corresponding satellite carrier observation value is adjusted according to the probability of satellite being blocked, including obtaining the probability P of each observable satellite being blocked j; Calculate the corresponding satellite carrier observation value weight, the calculation formula is expressed as,

[0021]

[0022] Where T represents the total number of observable satellites, w j represents the weight of the jth observable satellite.

[0023] As a preferred solution of the real-time monitoring method of power line galloping based on GNSS, INS and visual fusion described in the present invention, wherein: the updating of the weight matrix includes updating the j-th element value on the diagonal of the weight matrix W to w j ; The positioning coordinates calculated by weighted least squares method are expressed as,

[0024]

[0025] in, represents the position coordinate vector to be determined, A represents the design matrix, and L represents the observation value vector.

[0026] Another object of the present invention is to provide a real-time monitoring system for power grid line swing based on GNSS, INS and vision fusion, which can obtain the time series of periodic or random swings of power transmission lines caused by wind, climate change and other external factors.

[0027] To solve the above technical problems, the present invention provides the following technical solutions: a system for real-time monitoring of power line dancing based on GNSS, INS and visual fusion, comprising: an observation value acquisition module, an occlusion probability calculation module, a positioning coordinate calculation module and a dancing time series module; the observation value acquisition module obtains GNSS satellite carrier observation values, sets the weight of each GNSS satellite carrier observation value to 1, and establishes a weight matrix; the occlusion probability calculation module obtains the current sky image at the monitoring station, and identifies the probability that each pixel in the current sky image belongs to the sky through a sky recognition model; the azimuth and altitude angle of the satellite are mapped to the current sky image, and the probability of the satellite being blocked is calculated by the position of the satellite mapped to the sky image; the positioning coordinate calculation module adjusts the corresponding satellite carrier observation value weight according to the probability of the satellite being blocked, updates the weight matrix, and solves the positioning coordinates by weighted least squares method; the dancing time series module obtains the final positioning coordinates by fusing the positioning coordinates solved by GNSS and the positioning coordinates solved by INS through Kalman filtering, and obtains the line dancing time series according to the final positioning coordinates and the corresponding timestamp.

[0028] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the above-mentioned method for real-time monitoring of power grid line sway based on GNSS, INS and visual fusion are implemented.

[0029] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a real-time monitoring method for power grid line sway based on GNSS, INS and visual fusion as described above.

[0030] Beneficial effects of the present invention: The present invention analyzes the satellite occlusion probability in each sky image through a neural network model, dynamically adjusts the GNSS observation value weight of each observable satellite, and uses the weighted least squares method to solve the positioning coordinates. Finally, the positioning coordinates solved by INS and GNSS are fused through Kalman filtering. In the line dance monitoring and positioning, the observation values ​​with larger errors are given lower weights, and the observation values ​​with smaller errors are given higher weights, thereby improving the positioning accuracy.

[0031] The present invention combines multiple convolutional layers and transposed convolutional layers, extracts and fuses image features layer by layer, so that the network can not only capture image features from local to global, but also accurately restore image details. Finally, through the sigmoid activation function, the model can output a refined occlusion probability map, providing a more accurate input for the weight adjustment of GNSS signals. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0033] Figure 1 This is a flow chart of a method for real-time monitoring of power grid line sway based on GNSS, INS and vision fusion in Example 1.

[0034] Figure 2 This is a structural diagram of a real-time monitoring system for power grid line sway based on GNSS, INS and vision fusion in Example 3. DETAILED DESCRIPTION

[0035] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0036] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0037] Example 1, reference Figure 1 , which is the first embodiment of the present invention, provides a real-time monitoring method for power line dancing based on GNSS, INS and visual fusion, including: Figure 1 As shown:

[0038] Step 1: Obtain GNSS satellite carrier observations, set the weight of each GNSS satellite carrier observation to 1, and establish a weight matrix.

[0039] Step 2: Obtain the current sky image at the monitoring station, and use the sky recognition model to identify the probability that each pixel in the current sky image belongs to the sky.

[0040] The network structure of the sky recognition model includes the first convolution layer, the second convolution layer, the third convolution layer, the fourth convolution layer, the fifth convolution layer, the pooling layer, the first transposed convolution layer, the second transposed convolution layer, the third transposed convolution layer, the fourth transposed convolution layer and the probability calculation layer.

[0041] The first convolution layer is sequentially connected to the second convolution layer, the third convolution layer, the fourth convolution layer, the fifth convolution layer, the pooling layer, the first transposed convolution layer, the second transposed convolution layer, the third transposed convolution layer, the fourth transposed convolution layer and the probability calculation layer.

[0042] The first convolutional layer is connected to the first transposed convolutional layer. The first convolutional layer expands the output feature map of the first convolutional layer to the same size as the output feature map of the first transposed convolutional layer through nearest neighbor interpolation, and then adds the expanded output feature map of the first convolutional layer and the output feature map of the first transposed convolutional layer element by element to obtain the first fused transposed feature map.

[0043] The second convolutional layer is connected to the second transposed convolutional layer. The second convolutional layer expands the second convolution output feature map to the same size as the first fused transposed feature map through nearest neighbor interpolation, and then adds the expanded second convolutional layer output feature map to the first fused transposed feature map element by element to obtain the second fused transposed feature map.

[0044] The third convolutional layer is connected to the third transposed convolutional layer. The third convolutional layer expands the third convolution output feature map to the same size as the second fused transposed feature map through nearest neighbor interpolation, and then adds the expanded third convolutional layer output feature map to the second fused transposed feature map element by element to obtain the third fused transposed feature map.

[0045] The fourth convolutional layer is connected to the fourth transposed convolutional layer. The fourth convolutional layer expands the fourth convolution output feature map to the same size as the third fused transposed feature map through nearest neighbor interpolation, and then adds the expanded fourth convolutional layer output feature map and the third fused transposed feature map element by element to obtain a fourth fused transposed feature map. The size of the fourth fused transposed feature map is the same as the input image size.

[0046] The probability calculation layer is a sigmoid activation function, which converts the score of each pixel in the fourth fused transposed feature map into an occlusion probability.

[0047] The loss function of the network structure of the sky recognition model is,

[0048]

[0049] Where N represents the number of training images, y i (x,y) represents the true label of the pixel with position coordinates (x,y) in the i-th image. Indicates the probability that the pixel with position coordinates (x, y) in the i-th image is occluded.

[0050] Step 3: Map the satellite's azimuth and altitude angle onto the current sky image, and calculate the probability that the satellite is blocked by the position of the satellite mapped onto the sky image.

[0051] Mapping the satellite's azimuth and altitude angles to the current sky image includes obtaining the image length L and width W, obtaining the pitch angle θ and yaw angle ψ of the current sky image, and obtaining the azimuth angle Az of each observable satellite j and the altitude angle El j .

[0052] Calculate the local azimuth Az of each observable satellite j ' and the local altitude angle El j ' is expressed as,

[0053] Az j '=Az j -ψ

[0054] El j '=El j +θ

[0055] The distance between the position of each observable satellite mapped to the current sky image and the center point of the current sky image is calculated as:

[0056]

[0057]

[0058] Query the occlusion probability p of all pixels within the influence circle with radius d, centered on the position of each observable satellite mapped to the current sky image. j (x,y).

[0059] Calculate the weight of each pixel k in the circle

[0060] Calculate the probability P of each observable satellite being blocked j , the calculation method is as follows:

[0061]

[0062] Among them, r k It represents the distance between the kth pixel point in the influence circle and the center point of the influence circle, s represents the set weight index, j represents the number of the observable satellite, and M represents the total number of pixels in the influence circle.

[0063] Step 4: Adjust the corresponding satellite carrier observation weights according to the probability of satellite obstruction, update the weight matrix, and solve the positioning coordinates using the weighted least squares method.

[0064] Get the probability P of each observable satellite being blocked j ; Calculate the corresponding satellite carrier observation value weight, the calculation formula is expressed as,

[0065]

[0066] Where T represents the total number of observable satellites, w j Represents the weight of the jth observable satellite

[0067] Update the j-th element value on the diagonal of the weight matrix W to w j ;

[0068] The positioning coordinates are calculated by weighted least squares method and expressed as:

[0069]

[0070] in, represents the position coordinate vector to be determined, A represents the design matrix, and L represents the observation value vector

[0071] Step 5: The final positioning coordinates are obtained by fusing the positioning coordinates solved by GNSS and the positioning coordinates solved by INS through Kalman filtering, and the line dancing time series is obtained according to the final positioning coordinates and the corresponding timestamps.

[0072] Embodiment 2 is the second embodiment of the present invention, which is different from the first embodiment in that: a real-time monitoring method for power grid line dancing based on GNSS, INS and visual fusion also includes: in order to verify and illustrate the technical effects adopted in this method, this embodiment adopts traditional technical solutions and the method of the present invention for comparative testing, and compares the test results by means of scientific demonstration to verify the real effect of this method.

[0073] Experimental design:

[0074] Environmental settings: Typical urban canyons, mountainous areas, and dense vegetation areas were selected as the test environments.

[0075] Equipment configuration: The monitoring system based on GNSS, INS and vision fusion mentioned in this invention is deployed and compared with the traditional system based on GNSS and INS.

[0076] Data Collection: Monitoring data for both systems were collected over the same time period.

[0077] Data analysis method: Calculate the positioning accuracy of the two systems separately.

[0078] Compare and analyze the performance differences of the two systems in different environments.

[0079] Evaluation Metrics:

[0080] Positioning accuracy: The positioning error caused by the measurement system.

[0081] System stability: Analyze system performance fluctuations under various environmental conditions.

[0082] Signal Quality: Assess the quality of GNSS signal reception under obstruction conditions.

[0083] Table 1: Experimental data comparison table

[0084]

[0085] It can be seen from the above data that in urban canyons and dense vegetation areas, the system of the present invention shows significant superiority, with low average positioning error, high system stability, and high-quality signal reception. This is due to the use of the neural network model in the system of the present invention, which can effectively predict and adjust the GNSS signal weakening caused by satellite occlusion.

[0086] In mountainous areas, although the gap between the system of the present invention and the traditional system has been reduced, the system of the present invention still exhibits lower positioning error and higher system stability.

[0087] Example 3, reference Figure 2, which is the third embodiment of the present invention, and is different from the previous two embodiments in that: a system for real-time monitoring method of power line dancing based on GNSS, INS and visual fusion, characterized in that: it includes an observation value acquisition module, an occlusion probability calculation module, a positioning coordinate calculation module and a dancing time series module; the observation value acquisition module obtains GNSS satellite carrier observation values, sets the weight of each GNSS satellite carrier observation value to 1, and establishes a weight matrix; the occlusion probability calculation module obtains the current sky image at the monitoring station, and identifies the probability that each pixel in the current sky image belongs to the sky through a sky recognition model; the azimuth and altitude angle of the satellite are mapped to the current sky image, and the probability of the satellite being blocked is calculated by the position of the satellite mapped to the sky image; the positioning coordinate calculation module adjusts the corresponding satellite carrier observation value weight according to the probability of the satellite being blocked, updates the weight matrix, and solves the positioning coordinates by weighted least squares method; the dancing time series module obtains the final positioning coordinates by fusing the positioning coordinates solved by GNSS and the positioning coordinates solved by INS through Kalman filtering, and obtains the line dancing time series according to the final positioning coordinates and the corresponding timestamp.

[0088] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0089] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0090] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0091] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0092] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A real-time monitoring method for power line galloping based on GNSS, INS and vision fusion, characterized by: include, Obtain GNSS satellite carrier observation values, set the weight of each GNSS satellite carrier observation value to 1, and establish a weight matrix; Obtain the current sky image at the monitoring station, and identify the probability that each pixel in the current sky image belongs to the sky through the sky recognition model; Map the satellite's azimuth and altitude angle onto the current sky image, and calculate the probability of the satellite being blocked by the position of the satellite mapped onto the sky image; The corresponding satellite carrier observation weights are adjusted according to the probability of satellite being blocked, the weight matrix is ​​updated, and the positioning coordinates are solved by weighted least squares method; The final positioning coordinates are obtained by fusing the positioning coordinates solved by GNSS and INS through Kalman filtering, and the line dancing time series is obtained according to the final positioning coordinates and the corresponding timestamps.

2. A method for real-time monitoring of power line galloping based on GNSS, INS and visual fusion as claimed in claim 1, characterized in that: The network structure of the sky recognition model includes a first convolutional layer, a second convolutional layer, a third convolutional layer, a fourth convolutional layer, a fifth convolutional layer, a pooling layer, a first transposed convolutional layer, a second transposed convolutional layer, a third transposed convolutional layer, a fourth transposed convolutional layer and a probability calculation layer; The first convolution layer is sequentially connected to the second convolution layer, the third convolution layer, the fourth convolution layer, the fifth convolution layer, the pooling layer, the first transposed convolution layer, the second transposed convolution layer, the third transposed convolution layer, the fourth transposed convolution layer and the probability calculation layer; The first convolution layer is connected to the first transposed convolution layer, the first convolution layer expands the output feature map of the first convolution layer to the same size as the output feature map of the first transposed convolution layer through nearest neighbor interpolation, and then adds the expanded output feature map of the first convolution layer to the output feature map of the first transposed convolution layer element by element to obtain a first fused transposed feature map; The second convolution layer is connected to the second transposed convolution layer, the second convolution layer expands the second convolution output feature map to the same size as the first fused transposed feature map through nearest neighbor interpolation, and then adds the expanded second convolution layer output feature map to the first fused transposed feature map element by element to obtain the second fused transposed feature map; The third convolution layer is connected to the third transposed convolution layer, and the third convolution layer expands the third convolution output feature map to the same size as the second fused transposed feature map through nearest neighbor interpolation, and then adds the expanded third convolution layer output feature map to the second fused transposed feature map element by element to obtain the third fused transposed feature map. The fourth convolution layer is connected to the fourth transposed convolution layer, the fourth convolution layer expands the fourth convolution output feature map to the same size as the third fused transposed feature map through nearest neighbor interpolation, and then adds the expanded fourth convolution layer output feature map to the third fused transposed feature map element by element to obtain a fourth fused transposed feature map, and the size of the fourth fused transposed feature map is the same as the input image size; The probability calculation layer is a sigmoid activation function, which converts the score of each pixel in the fourth fused transposed feature map into an occlusion probability.

3. A method for real-time monitoring of power line galloping based on GNSS, INS and visual fusion as claimed in claim 2, characterized in that: The loss function of the network structure of the sky recognition model is: Where N is the number of training images, yi(x, y) is the true label of the pixel with position coordinates (x, y) in the i-th image, Indicates the probability that the pixel with position coordinates (x, y) in the i-th image is occluded.

4. A method for real-time monitoring of power line galloping based on GNSS, INS and vision fusion as claimed in claim 3, characterized in that: The mapping of the satellite's azimuth and altitude angle to the current sky image includes obtaining the image length L and width W, obtaining the pitch angle θ and yaw angle ψ of the current sky image, and obtaining the azimuth angle Az of each observable satellite. j and the altitude angle El j ; Calculate the local azimuth Az of each observable satellite j ' and the local altitude angle El j ' is expressed as, Az j '=Az j -ψ He j '=The j +θ The distance between the position of each observable satellite mapped to the current sky image and the center point of the current sky image is calculated as:

5. A method for real-time monitoring of power line galloping based on GNSS, INS and vision fusion as claimed in claim 4, characterized in that: The calculation of the probability of satellite being blocked includes querying the blocked probability p corresponding to all pixel points within the influence circle with radius d and the position of each observable satellite mapped to the current sky image as the center j (x, y); Calculate the weight of each pixel k in the circle Calculate the probability Pj of each observable satellite being blocked. The calculation method is as follows: Among them, rk represents the distance between the kth pixel point in the influence circle and the center point of the influence circle, s represents the set weight index, j represents the number of observable satellites, and M represents the total number of pixels in the influence circle.

6. A method for real-time monitoring of power line galloping based on GNSS, INS and vision fusion as claimed in claim 5, characterized in that: The method of adjusting the corresponding satellite carrier observation value weight according to the probability of satellite being blocked includes obtaining the probability P of each observable satellite being blocked. j ; Calculate the corresponding satellite carrier observation weights, and the calculation formula is expressed as: Among them, T represents the total number of observable satellites, and wj represents the weight of the jth observable satellite.

7. A method for real-time monitoring of power line galloping based on GNSS, INS and vision fusion as claimed in claim 6, characterized in that: The updating of the weight matrix includes updating the j-th element value on the diagonal of the weight matrix W to wj; The positioning coordinates calculated by weighted least squares method are expressed as: in, represents the position coordinate vector to be determined, A represents the design matrix, and L represents the observation value vector.

8. A system using a real-time monitoring method for power line galloping based on GNSS, INS and vision fusion as claimed in any one of claims 1 to 7, characterized in that: It includes observation value acquisition module, occlusion probability calculation module, positioning coordinate calculation module and dancing time series module; The observation value acquisition module obtains GNSS satellite carrier observation values, sets the weight of each GNSS satellite carrier observation value to 1, and establishes a weight matrix; The occlusion probability calculation module obtains the current sky image at the monitoring station, identifies the probability that each pixel in the current sky image belongs to the sky through the sky recognition model; maps the azimuth and altitude angle of the satellite to the current sky image, and calculates the probability that the satellite is occluded through the position of the satellite mapped to the sky image; The positioning coordinate calculation module adjusts the corresponding satellite carrier observation value weight according to the probability of the satellite being blocked, updates the weight matrix, and solves the positioning coordinates by weighted least squares method; The dancing time series module obtains the final positioning coordinates by fusing the positioning coordinates solved by GNSS and the positioning coordinates solved by INS through Kalman filtering, and obtains the line dancing time series according to the final positioning coordinates and the corresponding timestamps.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of a real-time monitoring method for power grid line dancing based on GNSS, INS and visual fusion according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a real-time monitoring method for power grid line galloping based on GNSS, INS and visual fusion according to any one of claims 1 to 7 are implemented.

Images