Cross-water transmission line SAR multi-scattering feature recognition and prediction method

Abstract

The invention discloses a cross-water power transmission line SAR multi-scattering feature recognition and prediction method, and relates to the field of power transmission line synthetic aperture radar monitoring, and the method comprises the steps: obtaining an SAR image, DEM, weather and power transmission tower coordinate data of a target region; establishing a point scattering tracking model based on a power transmission line catenary configuration, and accurately positioning a secondary scattering straight line which is not influenced by elevation distortion and theoretical physical positions of drifting primary and tertiary scattering; further combining mirror reflection and creeping wave theories, analyzing a physical modulation mechanism of the water surface roughness and the track included angle on multiple scattering signals, and determining a characteristic evolution boundary; and finally, extracting a space geometric included angle and a water surface backscattering coefficient as input, and quantitatively predicting and identifying geometric shape evolution characteristics of the power transmission line in the current environment. According to the method, the problems of missing detection and misjudgment pain points caused by dynamic water surface interference in a traditional method are solved, and high-reliability physical constraint and algorithm support are provided for cross-water-area power transmission line SAR intelligent inspection.

Description

Technical Field

[0001] This invention relates to the field of synthetic aperture radar (SAR) monitoring of power transmission lines, specifically to a method for identifying and predicting multiple scattering features of SAR across water power transmission lines. Background Technology

[0002] Cross-water power transmission channels are the choke points of the power grid topology. Their open, humid, and strongly convective micro-meteorological characteristics make them extremely susceptible to disasters such as icing overload and abnormal sag. Synthetic aperture radar (SAR) has become an important means of monitoring wide-area power facilities due to its all-weather and all-time advantages. However, the specular reflection and multipath effect against the background of the water surface make the SAR imaging mechanism of power transmission lines far more complex than that on land, directly affecting the reliability of the data. Existing SAR multiple scattering studies mainly focus on strong scatterers such as bridges and ships. These targets have huge radar cross-sections and stable structures, forming stable dihedral reflections. Their multiple scattering signals appear as bright spots with high signal-to-noise ratios and fixed positional shifts. However, the theoretical framework based on strong point targets cannot be applied to power transmission lines with diameters only a few centimeters. Power transmission lines across water are weak linear targets, and their SAR echoes exhibit a complex combination of primary, secondary, and tertiary scattering components. Moreover, the response thresholds of the three physical channels to environmental factors are different. Due to the combined modulation of water surface roughness and the geometric orientation of the conductor, various combinations of states may occur during observation, such as simultaneous high brightness, only primary scattering, or even overall stealth. Existing technologies lack a quantitative characterization of this multi-channel parallel-environmental differential modulation mechanism and fail to reveal how wind field and water surface wave spectrum determine the existence and combination patterns of each order of scattering components. Precisely because of the lack of understanding of the combination mechanism of scattering components, existing wire extraction algorithms are still at the primary level of "binary detection" and cannot answer the core question of "which physical path the observed signal originates from": they lack hierarchical recognition capabilities and cannot distinguish whether the signal is a real single scattering or a second / third scattering generated by multipath; they also cannot analyze the morphological evolution law of "point-line-nothingness", resulting in a large number of weak but physically meaningful multiple scattering signals being missed or misjudged. To address the aforementioned problems, this invention proposes a method for identifying and predicting multiple scattering features of SAR for cross-water power transmission lines based on the coupling of geometric optics and water surface roughness. Summary of the Invention

[0003] To address the aforementioned technical problems, a method for identifying and predicting multiple scattering features of SAR across water transmission lines is provided. This technical solution solves the problems of lacking hierarchical identification capabilities, being unable to distinguish whether a signal is a true primary scattering or a secondary / tertiary scattering generated by multipath, and being unable to analyze the morphological evolution law of "point-line-nothing", which leads to the missed detection or misjudgment of a large number of weak but physically significant multiple scattering signals.

[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for identifying and predicting SAR multiple scattering features of cross-water power transmission lines, comprising: Step S1: Data preparation; acquire high-resolution raw SAR data of the target area and convert it into single-view complex SLC data; simultaneously acquire digital elevation model (DEM), meteorological data and transmission tower coordinate data covering the target area, wherein the digital elevation model (DEM) is a digital elevation model within the coverage area of ​​the SAR image. Step S2: Data preprocessing; using the data from step S1, perform fine registration and geocoding, and convert the registered SAR image intensity information into decibel values ​​to obtain a preprocessed high-resolution SAR backscattering coefficient map. Step S3, Existence Identification: Calculate the first geometric angle between the transmission line direction and the satellite flight direction by combining the transmission tower coordinates and satellite orbit parameters. The first geometric angle is the angle between the transmission line direction and the satellite flight direction. Use the meteorological data to retrieve the water surface roughness, calculate the radar cross section of the transmission line based on the electromagnetic scattering physical model, and combine the water surface backscattering coefficient threshold analysis. With the help of numerical simulation and statistical methods, establish the existence discrimination boundaries for primary, secondary, and tertiary scattering respectively. Step S4, Location Identification: Combining the transmission tower coordinates, the catenary equation of the transmission line determined by equating the catenary configuration of the transmission line to an ideal conductive inclined parabola model, and satellite observation geometry, construct a three-dimensional spatial configuration and scattering tracking model of the transmission line, calculate the theoretical physical locations of the first, second, and third scatterings on the SAR image, and determine the precise locations of point features and line features. Step S5, Feature Recognition and Classification: Identify multipath scattering features in the theoretical physical location, and analyze the SAR scattering features of primary, secondary, and tertiary scattering under different time periods and water surface environments, classifying them as point features, line features, or non-existent states; Step S6: Prediction; Establish the functional relationship between environmental factors, geometric factors and the geometric shape of multiple scattering features, and use the Bayesian probabilistic inference model to classify and predict the appearance of power transmission lines in the target area in SAR images.

[0005] The beneficial effects of the above scheme are: (1) Combining satellite observation geometry and real-time meteorological data (water surface roughness), the critical angle and roughness threshold were established through rigorous electromagnetic numerical simulation and statistical methods; (2) For the first time, specular reflection, diffuse reflection and creeping wave theory were introduced into feature recognition, accurately revealing how dynamic water surfaces modulate multiple scattering into discrete "point features" (first / third) or continuous "line features" (second); (3) Using Bayes' maximum a posteriori probability inference, meteorological data (water surface backscattering coefficient) and satellite orbit configuration were transformed into quantitative predictions of power transmission line morphology (point, line, non-existent), which can guide the planning of the best meteorological and geometric windows for satellite inspection missions.

[0006] Preferably, step S1 specifically includes: Acquire high-resolution SAR data; Acquire wind speed, wind direction, temperature, and rainfall data corresponding to the SAR image imaging time; Obtain the coordinates of the transmission tower; Obtain the digital elevation model (DEM) within the coverage area of ​​SAR imagery.

[0007] The beneficial effect of the above-mentioned further solutions is that the above technical solutions prepare the data that needs to be processed later, which facilitates the identification and prediction of SAR signals of transmission lines.

[0008] Preferably, step S2 specifically includes: Step S2-1: Acquire all images of the same study area and select one of them as the reference master image; Step S2-2: Convert the original SAR satellite imagery into SLC data; Step S2-3: Obtain a lookup table using the DEM of the area covered by satellite imagery and satellite orbit parameters to facilitate subsequent geocoding; Step S2-4: Based on the DEM, register the remaining SLCs to the reference master image; Step S2-5: Set the multiview ratio to 1:1 for range and azimuth of the SLC data to acquire intensity images; Step S2-6: Convert the intensity to dB. The conversion formula is: in, For pixel grayscale values, It is the absolute scaling constant; Step S2-7: Use the DEM and lookup table to transfer the intensity image in SAR coordinates to the geographic coordinate system; Step S2-8: Crop the meteorological data to fit within the coverage area of ​​the intensity image.

[0009] The beneficial effects of the above-mentioned further solutions are: through the above technical solutions, the data preprocessing is realized, the original SAR image is converted into intensity data, which facilitates the subsequent identification and prediction of multiple scattering characteristics of transmission lines by combining with external meteorological data.

[0010] Preferably, the specific steps for determining the existence boundary of the primary scattering in step S3 are as follows: Establishment of the catenary geometry of the transmission line: Considering computational complexity, the catenary configuration of the transmission line is equivalent to an ideal conductive inclined parabolic model; using the height difference, span, and maximum sag of the transmission towers, the three-dimensional spatial coordinates of any point on the transmission line are determined. Its geometric trajectory equation is expressed as: In the formula, The span between two transmission towers, This represents the maximum sag of the transmission line. To match the slope angle of the terrain The relevant horizontal offset coefficient; Geometric discretization: Calculate the total length of the ideal conductive inclined parabolic model. The specific formula is as follows: According to the radar incident wavelength The total length is The power transmission line is divided into axial sections. A number of tiny cylindrical elements are used to meet the mesh density requirements for high-frequency computing. Construction and expansion of surface current equations: Under plane wave illumination, the electric field integral equation of the target surface is established based on Maxwell's equations, using the Rao-Wilton-Glisson basis functions. For unknown surface current Discretization expansion: Matrix transformation and MLFMA solution; Galerkin method is used for testing, transforming the integral equation into a discrete impedance matrix equation, specific formula: The impedance matrix is ​​solved quickly using a multilayer fast multipole algorithm to obtain the surface current coefficient. ; Far-field scattering and RCS inversion: Calculating the total scattered electric field in the far-field region based on the solved surface current. Then, the radar cross-section of the transmission line is calculated. The specific formula is as follows: Based on the solved RCS, analyze its curve of change with the incident wave direction angle to determine the second geometric angle that produces the maximum backscattering response. The second geometric angle is the geometric angle corresponding to the primary scattering discrimination boundary. The second geometric angle is preferred as the primary scattering existence discrimination boundary. The secondary scattering existence discrimination boundary in step S3 includes: statistically analyzing the water surface backscattering coefficient under the scenarios of the presence and absence of secondary scattering signal, and calculating the water surface backscattering coefficient threshold range.

[0011] Preferably, the existence discrimination boundary of triple scattering in step S3 includes: statistically analyzing the backscattering coefficient of the water surface in scenarios where triple scattering signals exist and do not exist, and calculating the threshold range of the backscattering coefficient of the water surface; wherein, the triple scattering is highly dependent on the strong specular reflection effect of the calm water surface, and when the backscattering coefficient of the water surface is lower than the secondary scattering threshold and the specular reflection condition is met, the existence of triple scattering is determined.

[0012] The beneficial effect of the above-mentioned further solution is that the existence of multiple scattering characteristics of transmission lines can be identified through the above technical solution.

[0013] Preferably, step S4 specifically includes the following steps: Step S4-1: Identification of primary and tertiary scattering locations; the primary scattering in the SAR image appears as radar echoes directly from the power transmission line. Since the power transmission line is catenary-shaped, its primary scattering appears as a continuously curved trajectory of bright pixels in the SAR image; firstly, bright discrete pixels spanning the water area are extracted from the SAR image as candidate points for primary scattering; then, any point on the power transmission line... The first scattering at the range-up position in the SAR image The reference point is used as the reference point; the third scattering is similar to the first scattering in that it also retains the catenary geometry of the transmission line, which appears as a continuously curved trajectory in the SAR image; due to the added additional propagation path delay of the microwave signal between the transmission line and the water surface, the third scattering ( The position is located at the primary scattering position, shifted backward along the range. The formula for calculating the offset at each pixel is: in, Indicates the target point of the transmission line The height difference between its vertical projection point on the water surface; This refers to the local incident angle of the radar. This represents the range resolution of the SAR image.

[0014] The precise location of the point feature is determined by the following method: By setting the azimuth direction of the SAR image as the X-axis and the range direction as the Y-axis, the curved geometric trajectories of the first and third scatterings are fitted using a quadratic polynomial: in, and These are the trajectory functions for first-order and third-order scattering, respectively; , , These are the polynomial coefficients; The center of the high-brightness feature precisely falls at the lowest point of the ideal conductive inclined parabolic model in the range direction projection, and the tangent is parallel to the stationary point in the azimuth direction, thus satisfying the condition that the first derivative is zero: If there is no point within the span between two transmission towers that satisfies the condition that the first derivative is zero, then it is determined that there are no visible first or third scattering point features within the span. Step S4-2: Secondary scattering location identification; wherein secondary scattering ( The position is located at the primary scattering position, shifted backward along the range. At each pixel, and unaffected by the catenary geometry of the transmission line, it consistently appears as a straight line. Based on the point scattering tracking model, the identification of the secondary scattering location of cross-water transmission lines relies on the slant range path compensation principle and the geometric orthogonal projection relationship; In the geometry of oblique-look SAR observation, because the power transmission line is suspended high above the water surface, the microwave signal undergoes a secondary scattering propagation path of "radar-power transmission line-water surface-radar". Compared to the single scattering path where the radar wave returns directly via the power transmission line, this adds an extra propagation distance in the range direction. This additional path precisely compensates for the slant range reduction effect caused by the elevation of the transmission line; in the SAR image coordinate system, taking any point on the transmission line... The position of the first scattering As a reference point, its secondary scattering position Drift backward along the distance; the number of pixels in the drift distance. It conforms to the following equation: in, For the target point of the transmission line Distance from its vertical foot on the water surface The elevation difference; This refers to the local incident angle of the radar. This represents the range resolution of the SAR image. Based on the above geometric derivation, the secondary scattering point The location is mathematically strictly equivalent to the orthographic projection point of the transmission line on the water surface in a single reflection. SAR coordinates ; Preferably, step S5 specifically includes the following process: Step S5-1: Identification and determination of point features of primary and secondary scattering; the physical mechanism of primary and secondary scattering is constrained by the specular reflection law, that is, only when the local lateral incident angle of the radar beam on the suspension geometry of the transmission line is zero, that is, at the stagnation point where the tangent is parallel to the azimuth direction, will the extreme response of backscattering be generated, which will be manifested as discrete high-point features on the predicted trajectory; furthermore, the secondary scattering is the radar-water surface-transmission line-radar feature dynamically modulated by the roughness of the water surface. Step S5-2: Line feature identification and determination of secondary scattering; The physical mechanism of secondary scattering, which is radar-transmission line-water surface-radar, originates from the bistatic SAR or creeping wave propagation effect formed between the transmission line and the water surface by radar waves; Since the propagation path of the secondary scattering eliminates the elevation phase sensitivity of the transmission line, the secondary scattering signal is mapped on the line connecting the single reflection points on the water surface in the SAR image, and is not affected by the overhang curvature. Step S5-3: Feature evolution determination under angle constraints; combined with the geometric angle analysis of the satellite and the transmission line, under the conventional oblique angle, the primary scattering exhibits the point feature described in S5-1; only when the transmission line and the radar line of sight satisfy the strict orthogonal condition, the entire transmission line satisfies the specular reflection condition, and the primary scattering will evolve from the point feature to an extremely strong line feature.

[0015] The beneficial effect of the above-mentioned further solution is that, through the above technical solution, the geometric morphological features of multiple scattering characteristics of transmission lines can be identified and classified.

[0016] Preferably, step S6 specifically includes the following steps: Step S6-1: Construct the predicted input feature vector; extract known observation conditions as model input features, let the joint observation feature vector be... ,in, The spatial geometric angle characteristics are calculated by combining satellite geometric orbit information and transmission line coordinate information; To extract the radar backscattering coefficient from the water area near the power transmission line, which is used to characterize real-time water surface roughness information; Step S6-2: Bayesian inference and maximum a posteriori probability classification; Let the formula for the power transmission line morphology category space that the system needs to determine be: in Representing point features, Indicates line characteristics, This indicates that it does not exist; A joint conditional probability model of the scattering pattern based on the geometric angle and water surface roughness is established. According to Bayes' theorem, given the observed eigenvectors... Under these conditions, the transmission line exhibits the first Posterior probability of a morphological category The calculation formula is as follows: Because of the denominator The constant is assumed, and the geometric configuration is assumed to be ( ) and environmental factors ( The two sides are physically independent, therefore the likelihood function is decoupled as follows: .

[0017] This method uses the maximum a posteriori probability (MAP) criterion for state determination, and its decision function... Defined as: in, This is a priori probability based on historical statistics. Characterize the probability of a certain scattering feature being generated under the first geometric angle; The system characterizes the probability of forming a certain scattering feature under different water surface roughness conditions; the system will use the measured data... Substituting into the decision function, the output is one of the three states with the highest probability: When the inference result points to the point feature state ( When the geometric angle does not satisfy the orthogonality condition for linear primary reflection, and the water surface backscattering coefficient is extremely low ( This indicates that the water surface is in an extremely calm state, at which time secondary scattering is suppressed, and the morphology consists only of local stagnation points of primary scattering or strong tertiary scattering points. When the inference result points to the linear characteristic state ( When the water surface backscattering coefficient is high, it indicates that the backscattering coefficient is relatively high at this time. The water surface exhibits moderate microwave diffuse reflection, supporting the formation of continuous line features by secondary and tertiary scattering, or when the first geometric angle satisfies the orthogonal condition, the line features are formed by primary scattering. When the inference result points to a non-existent state ( When this occurs, it indicates that the feature is at an extreme boundary, such as a drastically increased backscattering coefficient of the water surface ( ). The environmental degradation causes the signal-to-noise ratio of the multiple scattering signal to drop sharply and be completely submerged in the background noise of the water surface.

[0018] The beneficial effect of the above-mentioned further solution is that, through the above technical solution, the geometric morphological features of the multiple scattering characteristics of the transmission line are predicted by utilizing SAR satellite imaging geometry, transmission line orientation and water surface roughness data, and based on the Bayesian inference model.

[0019] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention introduces water surface roughness as a key state variable to construct a physical constraint model where geometric optics determines theoretical position and environmental coupling determines the form of existence. This enables accurate identification and hierarchical interpretation of primary, secondary, and tertiary scattering signals, improving detection accuracy to 86.5% and reducing the false negative rate to 8%. A quantitative characterization mechanism for primary-secondary-tertiary scattering components is proposed, filling the technical gap in multi-channel parallel-environmentally differentiated modulation mechanisms. Critical angles and roughness thresholds are established through MLFMA electromagnetic numerical simulation and statistical methods. Theories of specular reflection, diffuse reflection, and dihedral reflection are introduced into feature recognition, achieving a technological leap from "binary detection" to "physical semantic-level classification." Attached Figure Description

[0020] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a system architecture diagram of the present invention. Detailed Implementation

[0021] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0022] Reference Figure 1 , Figure 2 As shown, a method for identifying and predicting SAR multiple scattering features of cross-water power transmission lines includes: Step S1: Data preparation; acquire high-resolution raw SAR data of the target area and convert it into single-view complex SLC data; simultaneously acquire digital elevation model (DEM), meteorological data and transmission tower coordinate data covering the target area, wherein the digital elevation model (DEM) is a digital elevation model within the coverage area of ​​the SAR image. Step S2: Data preprocessing; using the data from step S1, perform fine registration and geocoding, and convert the registered SAR image intensity information into decibel values ​​to obtain a preprocessed high-resolution SAR backscattering coefficient map. Step S3, Existence Identification: Calculate the first geometric angle between the transmission line direction and the satellite flight direction by combining the transmission tower coordinates and satellite orbit parameters. The first geometric angle is the angle between the transmission line direction and the satellite flight direction. Use the meteorological data to retrieve the water surface roughness, calculate the radar cross section of the transmission line based on the electromagnetic scattering physical model, and combine the water surface backscattering coefficient threshold analysis. With the help of numerical simulation and statistical methods, establish the existence discrimination boundaries for primary, secondary, and tertiary scattering respectively. Step S4, Location Identification: Combining the transmission tower coordinates, the catenary equation of the transmission line determined by equating the catenary configuration of the transmission line to an ideal conductive inclined parabola model, and satellite observation geometry, construct a three-dimensional spatial configuration and scattering tracking model of the transmission line, calculate the theoretical physical locations of the first, second, and third scatterings on the SAR image, and determine the precise locations of point features and line features. Step S5, Feature Recognition and Classification: Identify multipath scattering features in the theoretical physical location, and analyze the SAR scattering features of primary, secondary, and tertiary scattering under different time periods and water surface environments, classifying them as point features, line features, or non-existent states; Step S6: Prediction; Establish the functional relationship between environmental factors, geometric factors and the geometric shape of multiple scattering features, and use the Bayesian probabilistic inference model to classify and predict the appearance of power transmission lines in the target area in SAR images.

[0023] In addition, data preparation in S1 includes the following steps: Acquire high-resolution SAR data; Acquire wind speed, wind direction, temperature, and rainfall data corresponding to the SAR image imaging time; Obtain the coordinates of the transmission tower; Obtain the digital elevation model (DEM) within the coverage area of ​​SAR imagery; This embodiment uses high-resolution SAR data from the TerraSAR-X satellite's X-band, with a range resolution of 1.0m and an azimuth resolution of 1.0m. The polarization mode is HH polarization, and the incident angle range is 25° to 45°. The DEM uses ALOS PALSAR RTCDEM data, with a spatial resolution of 12.5m and a vertical accuracy better than 10m. The meteorological data comes from real-time measurements by automatic weather stations within 30 minutes before and after the imaging time, including wind speed, wind direction, temperature, and rainfall, with a time synchronization accuracy better than 15 minutes. The coordinates of the transmission towers are obtained through RTK-GNSS static measurement mode, with an observation time of no less than 30 minutes, a horizontal accuracy better than 0.05m, an elevation accuracy better than 0.1m, and the coordinate system used is CGCS2000.

[0024] Data preprocessing in S2 includes the following steps: Step S2-1: Acquire all images of the same study area and select one of them as the reference master image; Step S2-2: Convert the raw SAR satellite imagery into SLC data Step S2-3: Obtain a lookup table using the DEM of the area covered by satellite imagery and satellite orbit parameters to facilitate subsequent geocoding; Step S2-4: Using the DEM, register the remaining SLCs to the reference master image. Step S2-5: Set the multiview ratio to 1:1 for range and azimuth of the SLC data to obtain intensity imagery. Step S2-6: Convert the intensity to dB according to the following formula. This step is designed to enhance the contrast of the scattered signal of the transmission line in a cross-water environment.

[0025] in For pixel grayscale values, absolute scaling constant Step S2-7: Use the DEM and lookup table to transfer the intensity image in SAR coordinates to the geographic coordinate system; Step S2-8: Crop the meteorological data to fit within the coverage area of ​​the intensity image; GAMMA software was used for precise registration based on DEM, with a registration accuracy better than 0.1 pixels; geocoding adopted UTM projection and WGS84 ellipsoid; the calibration constant K in dB conversion was obtained from the satellite payload parameter file or determined by field calibration of uniformly distributed corner reflectors; the multi-look ratio of 1:1 corresponds to 1 look in the range direction and 1 look in the azimuth direction, with an equivalent look number of approximately 1 look, maintaining the original resolution.

[0026] Identification of the presence of multiple scattering features of transmission lines in S3 includes the following steps: Step S3-1: The method for identifying the existence of primary scattering specifically includes the following steps: (a) Establishment of the catenary geometry of the transmission line: Considering the computational complexity, the catenary configuration of the transmission line is equivalent to an ideal conductive inclined parabolic model. Using the height difference, span, and maximum sag of the transmission towers, the three-dimensional spatial coordinates of any point on the transmission line are determined. Its geometric trajectory equation is expressed as: (in, The span between two transmission towers; This represents the maximum sag of the transmission line; To match the slope angle of the terrain (Related horizontal offset coefficient).

[0027] (b) Geometric discretization (mesh generation): Calculate the total length of the ideal conductive inclined parabolic model. : According to the radar incident wavelength The total length is The power transmission line is divided into axial sections. A number of tiny cylindrical elements are used to meet the mesh density requirements for high-frequency computing. (c) Construction and expansion of surface current equations: Under plane wave illumination, the electric field integral equations of the target surface are established based on Maxwell's equations. Rao-Wilton-Glisson basis functions are used. For unknown surface current Discretization expansion: (d) Matrix transformation and MLFMA solution: The Galerkin method is used to transform the integral equation into a discrete impedance matrix equation. The impedance matrix is ​​solved quickly using the Multilevel Fast Multipole Algorithm (MLFMA) to obtain the surface current coefficient. .

[0028] (e) Far-field scattering and RCS inversion: Based on the solved surface current, the total scattered electric field in the far-field region is calculated. Then, the radar cross section (RCS) of the transmission line can be calculated. : Based on the solved RCS, we analyze its curve as a function of the incident wave direction angle to determine the second geometric angle that produces the maximum backscattering response. The second geometric angle is the geometric angle corresponding to the primary scattering discrimination boundary, and the second geometric angle is used as the existence discrimination boundary for primary scattering.

[0029] In the MLFMA calculation, the number of layers L satisfies kD / 2L≈1 (k is the wave number, D is the electrically large target size), and in this embodiment, L=4 is taken; the smallest layer size of the box is 0.5λ, and the far-field aggregation threshold is 10-3; the RWG basis function discrete density is 10 triangle sides per wavelength; the incident wave frequency is 5.405GHz (C-band), the elevation angle range is 20°-60°, and the azimuth angle range is 0°-360°; it is calculated that when the angle between the radar line of sight and the tangent of the transmission line is less than 5°, the RCS peak reaches -15dBsm, which meets the presence identification threshold (SNR>10dB).

[0030] Step S3-2: Identification of the existence of secondary and tertiary scattering, specifically including statistical analysis of the backscattering coefficient of the water surface in scenarios where secondary and tertiary scattering signals exist and do not exist, and calculation of the threshold range of the backscattering coefficient of the water surface.

[0031] Identification of the location of multiple scattering features of transmission lines in S4 includes the following steps: Step S4-1: Identification of primary and tertiary scattering positions Primary and tertiary scattering location identification; the primary scattering in SAR images appears as radar echoes directly from power transmission lines. Since power transmission lines are catenary-shaped, their primary scattering in SAR images appears as continuously curved, bright pixel trajectories. First, bright discrete pixels spanning water areas are extracted from the SAR image as candidate points for primary scattering; then, any point on the power transmission line... The first scattering at the range-up position in the SAR image The reference point is used as the reference point; the third scattering is similar to the first scattering in that it also retains the catenary geometry of the transmission line, which appears as a continuously curved trajectory in the SAR image; due to the added additional propagation path delay of the microwave signal between the transmission line and the water surface, the third scattering ( The position is located at the primary scattering position, shifted backward along the range. The formula for calculating the offset at each pixel is: in, Indicates the target point of the transmission line The height difference between its vertical projection point on the water surface; This refers to the local incident angle of the radar. This represents the range resolution of the SAR image.

[0032] The precise location of the point feature is determined by the following method: By setting the azimuth direction of the SAR image as the X-axis and the range direction as the Y-axis, the curved geometric trajectories of the first and third scatterings are fitted using a quadratic polynomial: in, and These are the trajectory functions for first-order and third-order scattering, respectively; , , These are the polynomial coefficients; The center of the high-brightness feature precisely falls at the lowest point of the ideal conductive inclined parabolic model in the range direction projection, and the tangent is parallel to the stationary point in the azimuth direction, thus satisfying the condition that the first derivative is zero: If there is no point within the span between two transmission towers that satisfies the condition that the first derivative is zero, then it is determined that there are no visible first or third scattering point features within the span.

[0033] Step S4-2: Secondary scattering location identification Secondary scattering ( The position is located at the primary scattering position, shifted backward along the range. At each pixel, and unaffected by the catenary geometry of the transmission line, the secondary scattering location is consistently represented as a straight line. Based on the point scattering tracking model, the identification of the secondary scattering location of the cross-water transmission line relies on the slant range path compensation principle and the geometric orthogonal projection relationship.

[0034] In the geometry of oblique-look SAR observation, because the power transmission line is suspended high above the water surface, the microwave signal undergoes a secondary scattering propagation path of "radar-power transmission line-water surface-radar". Compared to the single scattering path where the radar wave returns directly via the power transmission line, this adds an extra propagation distance in the range direction. This extra path precisely compensates for the slant range reduction effect caused by the power line elevation. In the SAR image coordinate system, taking any point on the power line... The position of the first scattering As a reference point, its secondary scattering position Drift backward along the distance. The number of pixels drifted along the distance. It conforms to the following equation: in, For the target point of the transmission line Distance from its vertical foot on the water surface The elevation difference; This refers to the local incident angle of the radar. This represents the range resolution of the SAR image.

[0035] Based on geometric derivation, the secondary scattering point The location is mathematically strictly equivalent to the orthographic projection point of the transmission line on the water surface in a single reflection. SAR coordinates (i.e. ).

[0036] Although the transmission line presents a complex drooping catenary shape in three-dimensional space, its secondary scattering completely eliminates geometric distortions such as perspective contraction caused by elevation changes, and in SAR images it remains an indefinite and constant straight line connecting the bases of the two transmission towers.

[0037] The identification and classification of multiple scattering features of transmission lines in S5 includes the following steps: Step S5-1: Identification and determination of point features for primary and secondary scattering. The physical mechanism of primary and secondary scattering is constrained by the specular reflection law, that is, only when the local lateral incident angle of the radar beam on the transmission line suspension geometry is zero (i.e., at the stagnation point where the tangent is parallel to the azimuth direction) will an extreme response of backscattering occur, which will appear as discrete bright "point features" on the predicted trajectory. Furthermore, the features of secondary scattering (radar-water surface-transmission line-radar) are dynamically modulated by the roughness of the water surface (diffuse reflection degree).

[0038] Step S5-2: Line feature identification and determination of secondary scattering. The physical mechanism of the secondary scattering (radar-transmission line-water surface-radar) originates from the bistatic SAR or creeping wave propagation effect formed by radar waves between the transmission line and the water surface. Since the propagation path of the secondary scattering eliminates the elevation phase sensitivity of the transmission line, the secondary scattering signal is mapped in the SAR image onto the line connecting the single reflection points on the water surface, and is not affected by the overhang curvature.

[0039] Step S5-3: Characteristic Evolution Determination under Angle Constraints. Combining the geometric angle analysis of the satellite and transmission line directions, under the conventional oblique angle, primary scattering exhibits the "point feature" described in S5-1; only when the transmission line direction and the radar line of sight satisfy the strict orthogonal condition (or within the critical angle range), the entire transmission line satisfies the specular reflection condition, and primary scattering evolves from the "point feature" to an extremely strong "line feature".

[0040] Point feature determination: On the predicted trajectory, if the intensity of a pixel is more than 3 times higher than the neighborhood mean and satisfies the condition that the first derivative is zero, it is determined to be a point feature; Line feature determination: If the intensity of more than 50 consecutive pixels is higher than the threshold and the linear fit R... 2 If the intensity is greater than 0.95, it is determined to be a line feature; if the maximum intensity on the predicted trajectory is lower than the background mean plus 2 standard deviations, it is determined to be non-existent. This embodiment uses adaptive threshold segmentation combined with morphological filtering for feature extraction.

[0041] Prediction of the geometric shape of multiple scattering characteristics of transmission lines in S6 includes the following steps: Step S6-1: Construct the predicted input feature vector.

[0042] Extract known observation conditions as model input features, and let the input joint observation feature vector be... ,in, The spatial geometric angle characteristics are calculated by combining satellite geometric orbit information and transmission line coordinate information; To extract the radar backscattering coefficient from the water area near the power transmission line, which is used to characterize real-time water surface roughness information; Step S6-2: Bayesian inference and maximum a posteriori probability classification.

[0043] Let the space of power transmission line types that the system needs to determine be as follows: ,in Indicates "point feature", Indicates "line features". It means "does not exist".

[0044] A joint conditional probability model of the scattering pattern based on the geometric angle and water surface roughness is established. According to Bayes' theorem, given the observed eigenvectors... Under these conditions, the transmission line exhibits the first Posterior probability of a morphological category The calculation formula is as follows: Because of the denominator The constant is assumed, and the geometric configuration is assumed to be ( ) and environmental factors ( They are physically independent, which decouples the likelihood function. .

[0045] This method uses the maximum a posteriori probability (MAP) criterion for state determination, and its decision function... Defined as: in, This is a priori probability based on historical statistics. Characterize the probability of a certain scattering feature being generated under the first geometric angle; The system characterizes the probability of forming a certain scattering feature under different water surface roughness conditions. The system will use measured data... Substituting into the decision function, the output is one of the three states with the highest probability: When the inference result points to the point feature state ( When the geometric angle does not satisfy the orthogonality condition for linear primary reflection, and the water surface backscattering coefficient is extremely low ( This indicates that the water surface is in an extremely calm state, at which time secondary scattering is suppressed, and the morphology consists only of local stagnation points of primary scattering or strong tertiary scattering points. When the inference result points to the linear characteristic state ( When the water surface backscattering coefficient is high, it indicates that the backscattering coefficient is relatively high at this time. The water surface exhibits moderate microwave diffuse reflection, supporting the formation of continuous line features by secondary and tertiary scattering, or when the first geometric angle satisfies the orthogonal condition, the line features are formed by primary scattering. When the inference result points to a non-existent state ( When this occurs, it indicates that the feature is at an extreme boundary, such as a drastically increased backscattering coefficient of the water surface ( ). The signal-to-noise ratio of the multiple scattering signal drops sharply and is completely submerged in the background noise of the water surface due to environmental degradation (such as heavy rain or strong winds).

[0046] This embodiment employs a Bayesian probabilistic inference framework to establish a quantitative mapping relationship between geometric factors, environmental factors, and transmission line scattering patterns. The core idea is to transform satellite imaging geometric parameters and real-time water surface conditions into quantifiable feature vectors. Through probabilistic calculations, the posterior probabilities of three mutually exclusive morphological categories (point features, line features, and non-existence) are output, with the maximum posterior probability used as the final determination criterion. The advantage of this framework lies in its ability to explicitly integrate prior knowledge (historical observational statistical patterns) with real-time observational evidence (current imaging conditions) and to quantify the uncertainty of the prediction results.

[0047] The input feature vector contains two core dimensions. The first dimension is the spatial geometric angle feature, calculated by analyzing satellite orbit ephemeris data and transmission tower coordinates. Specifically, it extracts the horizontal angle between the satellite flight direction vector and the transmission line direction vector, and simultaneously calculates the three-dimensional angle between the radar line of sight and the local tangent plane of the transmission line. This angle determines the conditions for single specular reflection: when the angle is close to zero degrees, the entire transmission line satisfies the specular reflection law and easily forms a line feature; when the angle is large, only local stagnation points satisfy the condition, exhibiting discrete point features. The second dimension is the water surface roughness feature, characterized by statistically analyzing the mean and variance of the backscattering coefficient in the waters near the transmission line. This parameter indirectly reflects the sea surface wind speed state: extremely low backscattering coefficients (e.g., below -25 dB) correspond to calm water surfaces, supporting triple specular reflection; moderate values ​​(e.g., -22 to -18 dB) correspond to slightly rough water surfaces, where secondary scattering is dominant; high values ​​(e.g., above -15 dB) correspond to rough sea surfaces, where multiple scattering signals are submerged in background noise.

[0048] Prior probabilities are determined based on frequency statistics of historical observation data. Satellite transit records for the past three years in this study area were collected, and the frequency of occurrence of the three morphological categories was statistically analyzed. Typically, point features account for approximately 60% (conventional oblique imaging), line features account for approximately 30% (specific geometric conditions or moderate roughness), and non-existent states account for approximately 10% (extreme weather). The likelihood function is constructed using kernel density estimation to avoid making strong assumptions about the data distribution. For the geometric angle dimension, a Gaussian kernel with a bandwidth of 10 degrees is constructed to estimate the conditional probability density of each category within different angle intervals; for the water surface roughness dimension, a log-normal kernel with a bandwidth of 3 dB is constructed to accommodate the non-negativity of the backscattering coefficient. The two dimensions are assumed to be physically independent, and the joint likelihood is the product of the marginal likelihoods.

[0049] For each new observation, the real-time geometric angle and water surface roughness features are first extracted, and their likelihood values ​​under each category are calculated. Combining the prior probability, the posterior probability distribution is calculated using Bayes' theorem. The decision-making process employs the maximum a posteriori probability criterion: if the point feature has the highest posterior probability, the point feature prediction is output, along with a probability value as a confidence level; if the line feature has the highest probability, its cause (geometric line feature or environmental line feature) is further distinguished; if the absence of a state has the highest probability, an anomaly warning mechanism is triggered. When the probabilities of two classes are close (difference less than 5%), it is marked as an uncertain state, and it is recommended to combine multi-period imagery for comprehensive judgment.

[0050] Cross-validation was used to evaluate model performance. The historical dataset was divided into training, validation, and test sets to ensure a balanced distribution of data from different years, seasons, and satellites. On the test set, the model achieved an overall accuracy of 85%, with a recall rate of 88% for point features, 79% for line features, and 91% for non-existent states. The main sources of error included: geometric angle calculation errors (caused by satellite orbit determination errors), insufficient representativeness of water surface roughness (limited by the spatiotemporal resolution of meteorological data), and blurred class boundaries (misclassification in transitional states). To address the boundary ambiguity issue, cost-sensitive learning was introduced to reduce the cost weight of missed detections for line features, improving its applicability to continuous monitoring tasks.

[0051] Taking a cross-river power transmission channel as an example, the satellite passed over at 10:00 AM local time, the imaging geometry was ascending orbit with a right-looking view, and the extracted geometric angle was 28 degrees (not meeting the orthogonality condition). The backscattering coefficient of the adjacent water area was -19 dB (slightly rough). Model calculations showed that the posterior probability of point features was 62%, line features 33%, and there was no 5%. It was determined to be a point feature with moderate confidence. In the actual SAR image, this segment appeared as two bright discrete points, corresponding to the stationary points of the two power lines, consistent with the prediction results. In another case, the geometric angle was 2 degrees (close to orthogonality), the water surface roughness was -21 dB, and the line feature probability reached 71%. The actual image showed a continuous bright line, verifying the effectiveness of the model.

[0052] Considering the seasonal variations in water surface conditions and the long-term drift of satellite orbits, an online model update mechanism is established. Each quarter, the kernel density estimation parameters are incrementally updated using newly acquired imagery and measured data, ensuring that prior probabilities reflect the latest environmental statistical characteristics. Simultaneously, for the first observation after a major meteorological event (such as typhoons or freezing rain), the likelihood bandwidth is temporarily adjusted to broaden the uncertainty range and avoid overconfident and erroneous predictions. The update process employs a sliding window strategy, retaining data from the past two years and discarding outdated samples to ensure the model's adaptability to climate change.

[0053] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A method for identifying and predicting SAR multiple scattering features of cross-water power transmission lines, characterized in that, include: Step S1: Data preparation; acquire high-resolution raw SAR data of the target area and convert it into single-look complex SLC data; Simultaneously acquire digital elevation model (DEM) data, meteorological data, and transmission tower coordinate data covering the target area. The DEM is a digital elevation model within the coverage area of ​​the SAR image. Step S2: Data preprocessing; using the data from step S1, perform fine registration and geocoding, and convert the registered SAR image intensity information into decibel values ​​to obtain a preprocessed high-resolution SAR backscattering coefficient map. Step S3, Existence Identification: Calculate the first geometric angle between the transmission line direction and the satellite flight direction by combining the transmission tower coordinates and satellite orbit parameters. The first geometric angle is the angle between the transmission line direction and the satellite flight direction. Use the meteorological data to retrieve the water surface roughness, calculate the radar cross section of the transmission line based on the electromagnetic scattering physical model, and combine the water surface backscattering coefficient threshold analysis. With the help of numerical simulation and statistical methods, establish the existence discrimination boundaries for primary, secondary, and tertiary scattering respectively. Step S4, Location Identification: Combining the transmission tower coordinates, the catenary equation of the transmission line determined by equating the catenary configuration of the transmission line to an ideal conductive inclined parabola model, and satellite observation geometry, construct a three-dimensional spatial configuration and scattering tracking model of the transmission line, calculate the theoretical physical locations of the first, second, and third scatterings on the SAR image, and determine the precise locations of point features and line features. Step S5, Feature Recognition and Classification: Identify multipath scattering features in the theoretical physical location, and analyze the SAR scattering features of primary, secondary, and tertiary scattering under different time periods and water surface environments, classifying them as point features, line features, or non-existent states; Step S6: Prediction; Establish the functional relationship between environmental factors, geometric factors and the geometric shape of multiple scattering features, and use the Bayesian probabilistic inference model to classify and predict the appearance of power transmission lines in the target area in SAR images.

2. The method for identifying and predicting SAR multiple scattering features of cross-water transmission lines according to claim 1, characterized in that, Step S1 specifically includes: Acquire high-resolution SAR data; Acquire wind speed, wind direction, temperature, and rainfall data corresponding to the SAR image imaging time; Obtain the coordinates of the transmission tower; Obtain the digital elevation model (DEM) within the coverage area of ​​SAR imagery.

3. The method for identifying and predicting SAR multiple scattering features of cross-water transmission lines according to claim 1, characterized in that, Step S2 specifically includes: Step S2-1: Acquire all images of the same study area and select one of them as the reference master image; Step S2-2: Convert the original SAR satellite imagery into SLC data; Step S2-3: Obtain a lookup table using the DEM of the area covered by satellite imagery and satellite orbit parameters to facilitate subsequent geocoding; Step S2-4: Based on the DEM, register the remaining SLCs to the reference master image; Step S2-5: Set the multiview ratio to 1:1 for range and azimuth of the SLC data to acquire intensity images; Step S2-6: Convert the intensity to dB. The conversion formula is: in, For pixel grayscale values, It is the absolute scaling constant; Step S2-7: Use the DEM and lookup table to transfer the intensity image in SAR coordinates to the geographic coordinate system; Step S2-8: Crop the meteorological data to fit within the coverage area of ​​the intensity image.

4. The method for identifying and predicting SAR multiple scattering features of cross-water transmission lines according to claim 1, characterized in that, The specific steps for determining the existence boundary of the first scattering in step S3 are as follows: Establishment of the catenary geometry of the transmission line: Considering computational complexity, the catenary configuration of the transmission line is equivalent to an ideal conductive inclined parabolic model; using the height difference, span, and maximum sag of the transmission towers, the three-dimensional spatial coordinates of any point on the transmission line are determined. Its geometric trajectory equation is expressed as: In the formula, The span between two transmission towers, This represents the maximum sag of the transmission line. To match the slope angle of the terrain The relevant horizontal offset coefficient; Geometric discretization: Calculate the total length of the ideal conductive inclined parabolic model. The specific formula is as follows: According to the radar incident wavelength The total length is The power transmission line is divided into axial sections. A number of tiny cylindrical elements are used to meet the mesh density requirements for high-frequency computing. Construction and expansion of surface current equations: Under plane wave illumination, the electric field integral equation of the target surface is established based on Maxwell's equations, using the Rao-Wilton-Glisson basis functions. For unknown surface current Discretization expansion: Matrix Transformation and MLFMA Solution: The Galerkin method is used to transform the integral equation into a discrete impedance matrix equation. Specific formula: The impedance matrix is ​​solved quickly using a multilayer fast multipole algorithm to obtain the surface current coefficient. ; Far-field scattering and RCS inversion: Calculating the total scattered electric field in the far-field region based on the solved surface current. Then, the radar cross-section of the transmission line is calculated. The specific formula is as follows: Based on the solved RCS, we analyze its curve as a function of the incident wave direction angle to determine the second geometric angle that produces the maximum backscattering response. The second geometric angle is the geometric angle corresponding to the primary scattering discrimination boundary, and the second geometric angle is used as the existence discrimination boundary for primary scattering.

5. The method for identifying and predicting SAR multiple scattering features of cross-water transmission lines according to claim 1, characterized in that, The secondary scattering existence discrimination boundary mentioned in step S3 includes: statistically analyzing the water surface backscattering coefficient under scenarios where the secondary scattering signal exists and does not exist, and calculating the threshold range of the water surface backscattering coefficient.

6. The method for identifying and predicting SAR multiple scattering features of cross-water transmission lines according to claim 1, characterized in that, The existence discrimination boundary of triple scattering in step S3 includes: statistically analyzing the backscattering coefficient of the water surface in scenarios where triple scattering signals exist and do not exist, and calculating the threshold range of the backscattering coefficient of the water surface; wherein, the triple scattering is highly dependent on the strong specular reflection effect of the calm water surface, and when the backscattering coefficient of the water surface is lower than the secondary scattering threshold and the specular reflection condition is met, the existence of triple scattering is determined.

7. The method for identifying and predicting SAR multiple scattering features of cross-water transmission lines according to claim 1, characterized in that: Step S4 specifically includes the following steps: Step S4-1: Identification of primary and tertiary scattering locations; the primary scattering in the SAR image appears as radar echoes directly from the power transmission line. Since the power transmission line is catenary-shaped, its primary scattering appears as a continuously curved trajectory of bright pixels in the SAR image; firstly, bright discrete pixels spanning the water area are extracted from the SAR image as candidate points for primary scattering; then, any point on the power transmission line... The first scattering at the range-up position in the SAR image The reference point is used as the reference point; the third scattering is similar to the first scattering in that it also retains the catenary geometry of the transmission line, which appears as a continuously curved trajectory in the SAR image; due to the added additional propagation path delay of the microwave signal between the transmission line and the water surface, the third scattering ( The position is located at the primary scattering position, shifted backward along the range. The formula for calculating the offset at each pixel is: in, Indicates the target point of the transmission line The height difference between its vertical projection point on the water surface; This refers to the local incident angle of the radar. This represents the range resolution of the SAR image. The precise location of the point feature is determined by the following method: By setting the azimuth direction of the SAR image as the X-axis and the range direction as the Y-axis, the curved geometric trajectories of the first and third scatterings are fitted using a quadratic polynomial: in, and These are the trajectory functions for first-order and third-order scattering, respectively; , , These are the polynomial coefficients; The center of the high-brightness feature precisely falls at the lowest point of the ideal conductive inclined parabolic model in the range direction projection, and the tangent is parallel to the stationary point in the azimuth direction, thus satisfying the condition that the first derivative is zero: If there is no point within the span between two transmission towers that satisfies the condition that the first derivative is zero, then it is determined that there are no visible first or third scattering point features within the span. Step S4-2: Secondary scattering location identification; wherein secondary scattering ( The position is located at the primary scattering position, shifted backward along the range. At each pixel, and unaffected by the catenary geometry of the transmission line, it consistently appears as a straight line; based on the point scattering tracking model, the identification of the secondary scattering location of the cross-water transmission line is based on the slant range path compensation principle and the geometric orthogonal projection relationship. In the geometry of oblique-look SAR observation, because the power transmission line is suspended high above the water surface, the microwave signal undergoes a secondary scattering propagation path of "radar-power transmission line-water surface-radar". Compared to the single scattering path where the radar wave returns directly via the power transmission line, this adds an extra propagation distance in the range direction. This additional path precisely compensates for the slant range reduction effect caused by the elevation of the transmission line; in the SAR image coordinate system, taking any point on the transmission line... The position of the first scattering As a reference point, its secondary scattering position Drift backward along the distance; the number of pixels in the drift distance. It conforms to the following equation: in, For the target point of the transmission line Distance from its vertical foot on the water surface The elevation difference; This refers to the local incident angle of the radar. This represents the range resolution of the SAR image. Based on geometric derivation, the secondary scattering point The location is mathematically strictly equivalent to the orthographic projection point of the transmission line on the water surface in a single reflection. SAR coordinates .

8. The method for identifying and predicting SAR multiple scattering features of cross-water transmission lines according to claim 1, characterized in that, The specific process of step S5 includes: Step S5-1: Identification and determination of point features of primary and secondary scattering; the physical mechanism of primary and secondary scattering is constrained by the specular reflection law, that is, only when the local lateral incident angle of the radar beam on the suspension geometry of the transmission line is zero, that is, at the stagnation point where the tangent is parallel to the azimuth direction, will the extreme response of backscattering be generated, which will be manifested as discrete high-point features on the predicted trajectory; furthermore, the secondary scattering is the radar-water surface-transmission line-radar feature dynamically modulated by the roughness of the water surface. Step S5-2: Line feature identification and determination of secondary scattering; The physical mechanism of secondary scattering, which is radar-transmission line-water surface-radar, originates from the bistatic SAR or creeping wave propagation effect formed between the transmission line and the water surface by radar waves; Since the secondary scattering propagation path eliminates the elevation phase sensitivity of the transmission line, the secondary scattering signal is mapped on the line connecting the single reflection points on the water surface in the SAR image, and is not affected by the overhang curvature. Step S5-3: Feature evolution determination under angle constraints; combined with the geometric angle analysis of the satellite and the transmission line, under the conventional oblique angle, the primary scattering exhibits the point feature described in S5-1; only when the transmission line and the radar line of sight satisfy the strict orthogonal condition, the entire transmission line satisfies the specular reflection condition, and the primary scattering will evolve from the point feature to an extremely strong line feature.

9. The method for identifying and predicting SAR multiple scattering features of cross-water transmission lines according to claim 1, characterized in that, The specific steps of step S6 are as follows: Step S6-1: Construct the predicted input feature vector; extract known observation conditions as model input features, let the joint observation feature vector be... ,in, The spatial geometric angle characteristics are calculated by combining satellite geometric orbit information and transmission line coordinate information; To extract the radar backscattering coefficient from the water area near the power transmission line, which is used to characterize real-time water surface roughness information; Step S6-2: Bayesian inference and maximum a posteriori probability classification; Let the formula for the power transmission line morphology category space that the system needs to determine be: in Representing point features, Indicates line characteristics, This indicates that it does not exist; A joint conditional probability model of the scattering pattern based on the geometric angle and water surface roughness is established. According to Bayes' theorem, given the observed eigenvectors... Under these conditions, the transmission line exhibits the first Posterior probability of a morphological category The calculation formula is as follows: Because of the denominator The constant is assumed, and the geometric configuration is assumed to be ( ) and environmental factors ( The two sides are physically independent, therefore the likelihood function is decoupled as follows: ; This method uses the maximum a posteriori probability (MAP) criterion for state determination, and its decision function... Defined as: in, This is a priori probability based on historical statistics. Characterize the probability of a certain scattering feature being generated under the first geometric angle; The system characterizes the probability of forming a certain scattering feature under different water surface roughness conditions; the system will use the measured data... Substituting into the decision function, the output is one of the three states with the highest probability: When the inference result points to the point feature state ( When the geometric angle does not satisfy the orthogonality condition for linear primary reflection, and the water surface backscattering coefficient is extremely low ( This indicates that the water surface is in an extremely calm state, at which time secondary scattering is suppressed, and the morphology consists only of local stagnation points of primary scattering or strong tertiary scattering points. When the inference result points to the linear characteristic state ( When ), it indicates a higher backscattering coefficient of the water surface at this time ( The water surface exhibits moderate microwave diffuse reflection, supporting the formation of continuous line features by secondary and tertiary scattering, or when the first geometric angle satisfies the orthogonal condition, the line features are formed by primary scattering. When the inference result points to a non-existent state ( When this occurs, it indicates that the feature is at an extreme boundary, such as a drastically increased backscattering coefficient of the water surface ( ). The environmental degradation causes the signal-to-noise ratio of the multiple scattering signal to drop sharply and be completely submerged in the background noise of the water surface.

Images