Transmission Line Sag Detection Method Based on Synthetic Aperture Radar Imagery

Abstract

The present invention relates to the field of sag detection technology and discloses a transmission line sag detection method based on synthetic aperture radar images. The method comprises the following steps: Step S101, determining whether the current wind speed is less than a threshold, and calculating the ice-free, wind-free load ratio and the ice-covered, wind-free load ratio; otherwise, proceeding to Step S104; Step S102, obtaining the conductor stress value using a stress detection model; Step S103, calculating the ice-free, wind-free sag and the ice-covered, wind-free sag, and proceeding to Step S105; Step S104, obtaining the ice-free, wind-free sag and the ice-covered, wind-free sag using a sag detection model; and Step S105, determining whether there is a risk of ground contact. The present invention uses a stress detection model and a target detection model combined with an empirical formula to obtain the ice-free, wind-free sag and the ice-covered, wind-free sag. The sag detection model also dynamically tracks the changing trend of conductor sag over time in synthetic aperture radar images, thereby achieving all-weather, real-time transmission line sag detection.

Description

Technical Field

[0001] The present invention relates to the technical field of sag detection, and more particularly to a transmission line sag detection method based on synthetic aperture radar images. Background Art

[0002] As an important part of the power system, transmission lines undertake the critical task of transmitting electricity over long distances. Under the influence of their own weight and environmental loads, the conductors will form a suspension curve between the towers. The sag refers to the vertical distance between the lowest point of the conductor and the connection point of the tower. Excessive sag indicates that there is a risk of contact between the conductor and the ground or vegetation, which may cause accidents such as conductor grounding and short circuit. Too little sag indicates that the conductor tension is too high, which may cause the conductor to break or damage related equipment.

[0003] Traditionally, transmission line sag detection relies primarily on manual inspections, using simple tools like hanging line distance meters or sag rulers combined with empirical formulas to calculate sag. However, manual inspections are inefficient, especially in complex terrain, where inspections are extremely difficult. Furthermore, empirical formulas are based on an approximate parabola assumption and are only suitable for situations with smaller suspension curves, resulting in low sag detection accuracy.

[0004] With the development of drone technology and high-performance cameras, existing methods use drones to fly along transmission lines to capture high-definition images. These images are then extracted using image processing techniques (such as edge detection algorithms and Hough line detection algorithms). Catenary or parabola fitting is then used to calculate sag and the lowest point. While these approaches improve detection efficiency and accuracy to a certain extent, drone aerial photography is significantly affected by weather, making it difficult to achieve all-weather, real-time transmission line sag detection. Summary of the Invention

[0005] The present invention provides a transmission line sag detection method based on synthetic aperture radar images, which solves the technical problems in the above-mentioned background technology.

[0006] The present invention provides a method for detecting sag of a power transmission line based on synthetic aperture radar images, comprising the following steps:

[0007] Step S101: If the current wind speed is less than a preset wind speed threshold, then the ice-free wind load ratio and the ice-covered wind load ratio are calculated based on the first conductor parameter. Otherwise, the process proceeds to step S104.

[0008] The first conductor parameters include: conductor mass per unit length, cross-sectional area and diameter;

[0009] Step S102: Acquire meteorological data and second conductor parameters, perform normalization processing on them, and input them into the trained stress detection model to obtain the conductor stress value;

[0010] Meteorological data include: temperature, rainfall, ice thickness, wind speed, and the angle between wind direction and the line connecting the tower;

[0011] The second conductor parameters include: the first conductor parameters, the conductor expansion coefficient and the elastic coefficient;

[0012] Step S103: Output the synthetic aperture radar image to the target detection model to obtain the basic parameters of the tower, and calculate the ice-free and wind-free sag and ice-covered and wind-free sag based on the conductor stress value, the ice-free and wind-free load, and the ice-covered and wind-free load, respectively, and then proceed to step S105;

[0013] The basic parameters of the tower include: span, height difference angle and distance from the detection point to the trumpet side;

[0014] Step S104: Collect synthetic aperture radar images within a preset time period and construct an image sequence to input into the trained sag detection model to obtain wind sag without ice and wind sag with ice covering.

[0015] Step S105 , calculating the height above the ground based on the ice-free and wind-free sag, ice-covered and wind-free sag, ice-covered and wind-free sag, or ice-covered and wind-free sag combined with the ground elevation, and determining whether there is a risk of touching the ground based on a preset safety distance threshold.

[0016] Furthermore, the wind speed threshold is preset The calculation formula is as follows: ,in Indicates the preset reference wind speed, Indicates the height of the wire hanging point on the trumpet side. Indicates the preset reference altitude, Represents the terrain factor; the length of the image sequence is equal to the ratio of the preset time period to the preset acquisition time interval; the preset reference wind speed, preset reference height, terrain factor, preset time period, preset acquisition time interval, and preset safety distance threshold are all custom parameters.

[0017] Furthermore, no ice and no wind load The calculation formula is as follows: ,in Indicates the mass per unit length of the conductor, represents the cross-sectional area of the conductor, Represents the acceleration due to gravity.

[0018] Furthermore, ice-covered no-wind load The calculation formula is as follows: ,in Indicates the mass per unit length of the conductor, represents the cross-sectional area of the conductor, represents the acceleration due to gravity, Indicates the wire diameter, Indicates the ice thickness, Indicates ice density.

[0019] Furthermore, the stress detection model consists of two branches;

[0020] The first branch is used to obtain a first stress value by performing matrix operation on the normalized meteorological data and the second conductor parameter;

[0021] First stress value The calculation formula is as follows:

[0022] ,in , , represents the combined vector formed by the normalized meteorological data and the second conductor parameters, Represents the weight parameter between the i-th dimension value and the j-th dimension value of the combination vector, represents the first weight parameter, represents the second weight parameter, represents the exponential function with the natural constant e as the base, Represents a transpose operation;

[0023] The second branch is used to obtain a second stress value by performing a weighted sum operation on the normalized meteorological data and the second conductor parameter;

[0024] Performing a weighted sum operation on the first stress value and the second stress value to obtain a stress value of the conductor;

[0025] The weight coefficients corresponding to the first stress value and the second stress value are custom parameters whose total value is 1, and the sample labels of the training samples of the stress detection model are acquired through the collection of the fiber Bragg grating stress sensor.

[0026] Furthermore, the target detection model is built based on the YOLOV8 model, and the sample labels of the training samples of the target detection model are obtained through manual annotation.

[0027] Furthermore, the calculation formulas for ice-free and wind-free sag and ice-covered wind-free sag are the same, where ice-free and wind-free sag is The calculation formula is as follows:

[0028] ,in Indicates no ice and no wind load. represents the conductor stress value, Indicates gear spacing, Indicates the distance from the detection point to the trumpet side, Indicates the elevation angle.

[0029] Furthermore, the sag detection model consists of N hidden layers, where N is the same as the length of the image sequence, and each hidden layer consists of a first unit and a second unit;

[0030] The first unit of the nth hidden layer inputs the synthetic aperture radar image corresponding to the nth sequence unit of the image sequence and outputs the first eigenvector, 1≤n≤N;

[0031] The second unit of the nth hidden layer inputs the first eigenvector output by the first unit of the nth hidden layer and outputs the second eigenvector;

[0032] The second feature vector output by the second unit of the Nth hidden layer is input to the first classifier and the second classifier, and the category spaces of the two classifiers represent ice-free wind sag and ice-covered wind sag respectively;

[0033] The sample labels of the training samples of the sag detection model are obtained through a handheld laser rangefinder.

[0034] Furthermore, the first unit is constructed based on a sliding window self-attention transformer, and the second unit is constructed based on a long short-term memory recurrent network unit.

[0035] Furthermore, the height above ground The calculation formula is as follows:

[0036] ,in and Respectively represent the wire hanging point heights on the small side and the large side, Indicates gear spacing, Distance from the detection point to the trumpet side, Indicates the ground elevation below the detection point, Indicates one of no ice and no wind sag, ice-covered and no wind sag, no ice and wind sag, or ice-covered and wind sag.

[0037] The beneficial effects of the present invention are as follows: the present invention divides the detection environment into four conditions, namely, ice-free and windless conditions, ice-covered and windless conditions, ice-free and windy conditions, and ice-covered and windy conditions. The present invention first obtains the conductor stress value through a stress detection model based on meteorological data and conductor parameters, then identifies the basic parameters of the tower pole in the synthetic aperture radar image through a target detection model, and finally obtains the ice-free and windless sag and ice-covered and windless sag in combination with an empirical formula. In addition, the present invention performs feature extraction and time series analysis on the image sequence of the synthetic aperture radar image through the sag detection model, dynamically tracks the changing trend of the sag over time, and thus realizes all-weather, real-time transmission line sag detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 The present invention is a flow chart of a method for detecting sag of a power transmission line based on synthetic aperture radar images. DETAILED DESCRIPTION

[0039] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed solely to enable those skilled in the art to better understand and implement the subject matter described herein, and that the functions and arrangements of the elements discussed may be varied without departing from the scope of this specification. Various examples may omit, substitute, or add various processes or components as needed. In addition, features described with respect to some examples may also be combined in other examples.

[0040] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in one or more embodiments of the present invention should have the usual meanings understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in one or more embodiments of the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprising" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the object being described changes, the relative positional relationship may also change accordingly.

[0041] like Figure 1 As shown, the transmission line sag detection method based on synthetic aperture radar images includes the following steps:

[0042] It should be noted that the calculation formula of the empirical formula mentioned in the background technology is as follows: ,in Indicates sag (unit: m), Indicates the deadweight of the conductor per unit length (unit: N / m), which can be obtained by referring to the engineering design file based on the conductor model. Indicates the span, that is, the horizontal projection distance between the tower connection points (unit: m), which can be obtained by referring to the engineering design file based on the tower number. It represents the conductor tension (unit: N). It can be obtained by referring to the standard value of conductor tension under normal temperature and conventional load specified in the engineering design archive or by empirical estimation. In addition to being approximately based on the parabola assumption, the empirical formula tends to ignore the influence of environmental factors such as temperature, icing, and wind load, resulting in low accuracy in sag calculations using the empirical formula.

[0043] Step S101: If the current wind speed is less than a preset wind speed threshold, then the ice-free wind load ratio and the ice-covered wind load ratio are calculated based on the first conductor parameter. Otherwise, the process proceeds to step S104.

[0044] The first conductor parameters include: conductor mass per unit length, cross-sectional area and diameter;

[0045] Step S102: Acquire meteorological data and second conductor parameters, perform normalization processing on them, and input them into the trained stress detection model to obtain the conductor stress value;

[0046] Meteorological data include: temperature, rainfall, ice thickness, wind speed, and the angle between wind direction and the line connecting the tower;

[0047] The second conductor parameters include: the first conductor parameters, the conductor expansion coefficient and the elastic coefficient;

[0048] Step S103: Output the synthetic aperture radar image to the target detection model to obtain the basic parameters of the tower, and calculate the ice-free and wind-free sag and ice-covered and wind-free sag based on the conductor stress value, the ice-free and wind-free load, and the ice-covered and wind-free load, respectively, and then proceed to step S105;

[0049] The basic parameters of the tower include: span, height difference angle and distance from the detection point to the trumpet side;

[0050] Step S104: Collect synthetic aperture radar images within a preset time period and construct an image sequence to input into the trained sag detection model to obtain wind sag without ice and wind sag with ice covering.

[0051] Step S105 , calculating the height above the ground based on the ice-free and wind-free sag, ice-covered and wind-free sag, ice-covered and wind-free sag, or ice-covered and wind-free sag combined with the ground elevation, and determining whether there is a risk of touching the ground based on a preset safety distance threshold.

[0052] In one embodiment of the present invention, the wind speed threshold is preset The calculation formula is as follows: ,in Indicates the preset reference wind speed, Indicates the height of the wire hanging point on the trumpet side. Indicates the preset reference altitude, represents the terrain factor; the length of the image sequence is equal to the ratio of the preset time period to the preset acquisition time interval; wherein the preset reference wind speed, preset reference height, terrain factor, preset time period, preset acquisition time interval and preset safety distance threshold are all custom parameters; preferably, the preset reference wind speed is set to 5m / s, the preset reference height is set to 10m, the terrain factor is set according to the terrain, for example, it is set to 0.16 in plain areas, 0.22 in hilly areas, and 0.30 in urban areas, the preset time period is set to 30min, the preset acquisition time interval is set to 1min, then the length of the image sequence is equal to 30, and the preset safety distance threshold is set to 6m.

[0053] For example, if the preset reference wind speed is set to 5m / s, the preset reference height is set to 10m, the terrain factor is set to 0.30, and the height of the wire hanging point on the small side is 8m, then the preset wind speed threshold is approximately equal to 4.68m / s.

[0054] In one embodiment of the present invention, no ice no wind load The calculation formula is as follows: ,in Indicates the mass per unit length of the conductor (unit: kg / km), Indicates the cross-sectional area of the conductor (unit: mm²), It represents the acceleration due to gravity (unit: m / s², assigned value: 9.8m / s²). The unit of ice-free and wind-free load is Pa / m.

[0055] In one embodiment of the present invention, ice-covered no-wind load The calculation formula is as follows: ,in Indicates the mass per unit length of the conductor (unit: kg / km), Indicates the cross-sectional area of the conductor (unit: mm²), represents the acceleration due to gravity (unit: m / s², assigned value is 9.8m / s²), Indicates the wire diameter (unit: mm), Indicates ice thickness (unit: mm), It represents the ice density (unit: kg / m³, assigned as 900kg / m³). The unit of ice no-wind load is Pa / m.

[0056] In one embodiment of the present invention, the stress detection model consists of two branches;

[0057] The first branch is used to obtain a first stress value by performing matrix operation on the normalized meteorological data and the second conductor parameter;

[0058] First stress value The calculation formula is as follows:

[0059] ,in , , Represents a combined vector (8 dimensions) formed by concatenating the normalized meteorological data and the second conductor parameters. Represents the weight parameter between the i-th dimension value and the j-th dimension value of the combination vector, represents the first weight parameter, represents the second weight parameter, represents the exponential function with the natural constant e as the base, Represents a transpose operation;

[0060] Then according to the above calculation formula, the transpose of the combined vector and the multiplication of the combined vector are obtained to form an 8×8 matrix. Then the first weight parameter needs to be designed as a 1×8 matrix, and the second weight parameter needs to be designed as an 8×1 matrix.

[0061] The second branch is used to obtain a second stress value by performing a weighted sum operation on the normalized meteorological data and the second conductor parameter;

[0062] Performing a weighted sum operation on the first stress value and the second stress value to obtain a stress value of the conductor;

[0063] The weight coefficients corresponding to the first stress value and the second stress value are custom parameters whose sum is 1. Preferably, the weight coefficient corresponding to the first stress value is set to 0.8, and the weight coefficient corresponding to the second stress value is set to 0.2.

[0064] In one embodiment of the present invention, sample labels of training samples of the stress detection model are acquired through collection of a fiber Bragg grating stress sensor.

[0065] It should be noted that the parameters in the stress detection model (including the weight parameters and bias parameters of the first branch, and the weight coefficients of the second branch) are all learnable hyperparameters. Random initial values (such as normal distribution initialization) are given at the beginning of training. During the training process, the mean square error between the predicted value output by the stress detection model and the true value corresponding to the sample label of the training sample is used as the loss function. Backpropagation is performed through a gradient optimizer (such as SGD, Adam, etc.) to minimize the loss value obtained by forward calculation of the loss function. The training of the neural network model is a conventional technical means and will not be elaborated here.

[0066] In one embodiment of the present invention, the target detection model is constructed based on the YOLOV8 model, and the sample labels of the training samples of the target detection model are obtained through manual annotation.

[0067] In one embodiment of the present invention, a pre-processing layer is added to the target detection model, that is, the input synthetic aperture radar image is filtered and denoised, and the filtering and denoising algorithm can be any one of the Lee filtering algorithm, the Frost filtering algorithm, and the BM3D denoising algorithm; the Neck part of the YOLOV8 model is modified, and a P2 feature layer is added, where Neck represents the intermediate module connecting the backbone network (Backbone) and the detection head (Head), which is mainly responsible for fusing feature maps of different scales, and the P2 feature layer represents the extracted shallower and finer-grained feature maps. The preset scale of the Anchor (anchor box) can be adjusted in the configuration file of the YOLOV8 model, such as 8×12, 12×16, 16×20, etc.; a high-resolution feature extractor is also added to the target detection model, and the original CSPDarknet (cross-stage partial convolutional network) is replaced with a Swin Transformer (sliding window self-attention transformer) as the Backbone (backbone network) to improve the Backbone's ability to extract high-detail information; in addition, the loss function of the target detection model can be modified to EIoU Loss, and add a small target penalty item, which will not be described in detail here.

[0068] It should be noted that the detection points refer to characteristic sampling points selected at preset intervals along the direction of the conductor, which are used to representatively detect the conductor sag and height above the ground, such as intervals of 30m, 50m, etc., because it is impossible to perform sag detection on every point of the conductor during the actual inspection process. In addition, the span can be obtained by referring to the engineering design files based on the tower number, the height difference angle can be obtained using DEM (digital elevation model) data, and the distance from the detection point to the small side can be obtained using the GIS database or high-definition map, which will not be elaborated here.

[0069] In one embodiment of the present invention, the calculation formulas for ice-free no-wind sag and ice-covered no-wind sag are the same, wherein ice-free no-wind sag The calculation formula is as follows: ,in Indicates no ice and no wind load. represents the conductor stress value, Indicates gear spacing, Indicates the distance from the detection point to the trumpet side, Indicates the elevation angle.

[0070] It should be noted that the unit of the conductor stress value is Pa, the height difference angle represents the angle between the line between the tower connection points and the horizontal direction, and the distance from the detection point to the small side represents the horizontal projection distance between the detection point and the conductor hanging point with a lower height.

[0071] In one embodiment of the present invention, the sag detection model consists of N hidden layers, where N is the same as the length of the image sequence, and each hidden layer consists of a first unit and a second unit;

[0072] The first unit of the nth hidden layer inputs the synthetic aperture radar image corresponding to the nth sequence unit of the image sequence and outputs the first eigenvector, 1≤n≤N;

[0073] The second unit of the nth hidden layer inputs the first eigenvector output by the first unit of the nth hidden layer and outputs the second eigenvector;

[0074] The second feature vector output by the second unit of the Nth hidden layer is input to the first classifier and the second classifier, and the category spaces of the two classifiers represent ice-free wind sag and ice-covered wind sag, respectively.

[0075] In one embodiment of the present invention, the sample labels of the training samples of the sag detection model are acquired by collecting with a handheld laser rangefinder or a laser scanner.

[0076] In one embodiment of the present invention, the first unit is constructed based on a Swin Transformer (sliding window self-attention transformer), and the second unit is constructed based on an LSTM (long short-term memory recurrent network) unit, and can also be constructed based on a GRU (gated recurrent network) unit, which will not be described in detail here.

[0077] For example, the size of the synthetic aperture radar image input to the first unit is 512×512×C (C represents the number of channels, for example, C is set to 1, that is, a grayscale image), and then it is divided into non-overlapping small blocks (Patches), each of which is 4×4 in size, for a total of 128×128 Patches, each of which corresponds to a vector of size 4×4×1=16. Then, each Patch is encoded into a higher-dimensional embedding vector (for example, 96) through MLP (Multi-Layer Perceptron), so the feature map size at this time is 128×128×96. Repeat the above operation to convert the size of the feature map to 16×16×512, and finally convert it into the first eigenvector through global average pooling, that is, the number of dimensions of the first eigenvector is 512.

[0078] In one embodiment of the present invention, the height above the ground The calculation formula is as follows:

[0079] ,in and Respectively represent the wire hanging point heights on the small side and the large side, Indicates gear spacing, Distance from the detection point to the trumpet side, Indicates the ground elevation below the detection point, Indicates one of no ice and no wind sag, ice-covered and no wind sag, no ice and wind sag, or ice-covered and wind sag.

[0080] It should be noted that a training sample consists of sample data and sample labels. The sample data of the sag detection model is a synthetic aperture radar image sequence, and the sample labels of the sag detection model are ice-free wind sag and ice-covered wind sag (which can be obtained by collecting with a handheld laser rangefinder or laser scanner). During the training of the sag detection model, the mean square error is used as the loss function to calculate the deviation between the sag value predicted by the sag detection model at each iteration and the measured value (sample label). The model parameters are updated by backpropagation through the gradient descent optimization algorithm (such as Adam, SGD optimizer, etc.), and the deviation calculated by the loss function is gradually reduced until the maximum number of iterations is reached or the deviation reaches the preset minimum loss threshold. The training of the sag detection model is then completed.

[0081] It should be noted that before inputting the sag detection model, the synthetic aperture radar image can also be normalized, denoised, etc., and before training, it can also be pre-trained, with meteorological data used separately as the pre-training sample labels for the first unit and the second unit. After pre-training, parameter fine-tuning is performed for the sag detection task, which can accelerate model convergence and improve feature extraction capabilities. To a certain extent, it can reduce the number of training samples for the actual training of the sag detection model, thereby reducing the difficulty of model training.

[0082] It should be noted that the present invention divides the detection environment into four conditions, namely, ice-free and windless conditions, ice-covered and windless conditions, ice-free and windy conditions, and ice-covered and windy conditions. The present invention first obtains the conductor stress value through a stress detection model based on meteorological data and conductor parameters, and then identifies the basic parameters of the tower in the synthetic aperture radar image through a target detection model. Finally, the ice-free and wind-free sag and ice-covered and wind-free sag are obtained by combining empirical formulas. However, under windy conditions, the conductor is affected by wind force, and the above empirical formula is no longer applicable. Therefore, the present invention uses a sag detection model to perform feature extraction and time series analysis on the image sequence of the synthetic aperture radar image, and dynamically tracks the changing trend of sag over time, thereby realizing all-weather, real-time sag detection of transmission lines.

[0083] It should be noted that the intervals and thresholds are set for ease of comparison. The threshold size depends on the amount of sample data and the cardinality set by those skilled in the art for each set of sample data, as long as it does not affect the proportional relationship between the parameter and the quantized value. Furthermore, the above formulas are all dimensionless numerical calculations. These formulas are derived from software simulations of the most recent real-world conditions using large amounts of data. The preset parameters in these formulas are set by those skilled in the art based on actual conditions.

[0084] The above describes the embodiments of this embodiment, but this embodiment is not limited to the above specific implementation methods. The above specific implementation methods are merely illustrative and not restrictive. Ordinary technicians in this field can also make many forms based on the inspiration of this embodiment, all of which are protected by this embodiment.

Claims

1. A transmission line sag detection method based on synthetic aperture radar images, characterized in that: The following steps are involved: Step S101: If the current wind speed is less than a preset wind speed threshold, then the ice-free wind load ratio and the ice-covered wind load ratio are calculated based on the first conductor parameter. Otherwise, the process proceeds to step S104. The first conductor parameters include: conductor mass per unit length, cross-sectional area and diameter; Step S102: Acquire meteorological data and second conductor parameters, perform normalization processing on them, and input them into the trained stress detection model to obtain the conductor stress value; Meteorological data include: temperature, rainfall, ice thickness, wind speed, and the angle between wind direction and the line connecting the tower; The second conductor parameters include: the first conductor parameters, the conductor expansion coefficient and the elastic coefficient; Step S103: Output the synthetic aperture radar image to the target detection model to obtain the basic parameters of the tower, and calculate the ice-free and wind-free sag and ice-covered and wind-free sag based on the conductor stress value, the ice-free and wind-free load, and the ice-covered and wind-free load, respectively, and then proceed to step S105; The basic parameters of the tower include: span, height difference angle and distance from the detection point to the trumpet side; Step S104: Collect synthetic aperture radar images within a preset time period and construct an image sequence to input into the trained sag detection model to obtain wind sag without ice and wind sag with ice covering. Step S105: Calculate the height above the ground based on the ice-free and wind-free sag, ice-covered and wind-free sag, ice-covered and wind-free sag, or ice-covered and wind-free sag in combination with the ground elevation, and determine whether there is a risk of ground contact based on a preset safety distance threshold. Preset wind speed threshold The calculation formula is as follows: ,in Indicates the preset reference wind speed, Indicates the height of the wire hanging point on the trumpet side. Indicates the preset reference altitude, Represents the terrain factor; the length of the image sequence is equal to the ratio of the preset time period to the preset acquisition time interval; the preset reference wind speed, preset reference height, terrain factor, preset time period, preset acquisition time interval, and preset safety distance threshold are all custom parameters.

2. The method for detecting sag of a power transmission line based on synthetic aperture radar images according to claim 1, wherein: No ice and no wind load The calculation formula is as follows: ,in Indicates the mass per unit length of the conductor, represents the cross-sectional area of the conductor, Represents the acceleration due to gravity.

3. The method for detecting sag of a power transmission line based on synthetic aperture radar images according to claim 1, wherein: Ice-covered and windless load The calculation formula is as follows: ,in Indicates the mass per unit length of the conductor, represents the cross-sectional area of the conductor, represents the acceleration due to gravity, Indicates the wire diameter, Indicates the ice thickness, Indicates ice density.

4. The method for detecting sag of a power transmission line based on synthetic aperture radar images according to claim 1, wherein: The stress detection model consists of two branches; The first branch is used to obtain a first stress value by performing matrix operation on the normalized meteorological data and the second conductor parameter; First stress value The calculation formula is as follows: ,in , , represents the combined vector formed by the normalized meteorological data and the second conductor parameters, Represents the weight parameter between the i-th dimension value and the j-th dimension value of the combination vector, represents the first weight parameter, represents the second weight parameter, represents the exponential function with the natural constant e as the base, Represents a transpose operation; The second branch is used to obtain a second stress value by performing a weighted sum operation on the normalized meteorological data and the second conductor parameter; Performing a weighted sum operation on the first stress value and the second stress value to obtain a stress value of the conductor; The weight coefficients corresponding to the first stress value and the second stress value are custom parameters whose total value is 1, and the sample labels of the training samples of the stress detection model are acquired through the collection of the fiber Bragg grating stress sensor.

5. The method for detecting sag of a power transmission line based on synthetic aperture radar images according to claim 1, wherein: The target detection model is built based on the YOLOV8 model, and the sample labels of the training samples of the target detection model are obtained through manual annotation.

6. The method for detecting sag of a power transmission line based on synthetic aperture radar images according to claim 1, wherein: The calculation formulas for ice-free and wind-free sag and ice-covered wind-free sag are the same, among which ice-free and wind-free sag is The calculation formula is as follows: ,in Indicates no ice and no wind load. represents the conductor stress value, Indicates gear spacing, Indicates the distance from the detection point to the trumpet side, Indicates the elevation angle.

7. The method for detecting sag of a power transmission line based on synthetic aperture radar images according to claim 1, wherein: The sag detection model consists of N hidden layers, where N is the same as the length of the image sequence, and each hidden layer consists of the first unit and the second unit; The first unit of the nth hidden layer inputs the synthetic aperture radar image corresponding to the nth sequence unit of the image sequence and outputs the first eigenvector, 1≤n≤N; The second unit of the nth hidden layer inputs the first eigenvector output by the first unit of the nth hidden layer and outputs the second eigenvector; The second feature vector output by the second unit of the Nth hidden layer is input to the first classifier and the second classifier, and the category spaces of the two classifiers represent ice-free wind sag and ice-covered wind sag respectively; The sample labels of the training samples of the sag detection model are obtained through a handheld laser rangefinder.

8. The method for detecting sag of a power transmission line based on synthetic aperture radar images according to claim 7, characterized in that: The first unit is built based on a sliding window self-attention transformer, and the second unit is built based on a long short-term memory recurrent network unit.

9. The method for detecting sag of a power transmission line based on synthetic aperture radar images according to claim 1, wherein: Height from the ground The calculation formula is as follows: ,in and Respectively represent the wire hanging point heights on the small side and the large side, Indicates gear spacing, Distance from the detection point to the trumpet side, Indicates the ground elevation below the detection point, Indicates one of no ice and no wind sag, ice-covered and no wind sag, no ice and wind sag, or ice-covered and wind sag.

Images