Power distribution network intelligent planning method, system and device and storage medium

Abstract

The invention discloses a power distribution network intelligent planning method, system and device and a storage medium, and the method comprises the steps: obtaining point cloud data, carrying out the topographic feature information extraction of the point cloud data, and obtaining a topographic data set; acquiring remote sensing data, and performing earth surface change information extraction on the remote sensing data to obtain an earth surface change data set; fusing the topographic data set and the surface change data set to obtain a topographic point cloud data set; according to the terrain point cloud data set, a debris flow scouring path of the gully edge collapse area is analyzed, and a first terrain model is constructed; acquiring unmanned aerial vehicle monitoring data, simulating a line cross-trench path in combination with the first terrain model, optimizing cross-trench points and tower footing coordinates, and obtaining a power grid planning scheme; and the stability of the power grid planning scheme is evaluated, tower footing reinforcement parameters are adjusted, a final power grid planning scheme is obtained, and technical support is provided for power grid construction in the desert area.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of information technology, and in particular to a power distribution network intelligent planning method, system, device and storage medium. BACKGROUND

[0002] Power distribution network planning is an important link to ensure the stability of power supply and regional economic development, especially in complex terrain environment, its scientificity and adaptability are directly related to the safe operation and long-term benefit of power grid. However, the existing power distribution network planning method often appears to be inadequate when facing the special terrain of the transition zone between desert gobi and loess plateau. These methods mostly rely on static terrain data and single geographic information system analysis, which is difficult to cope with the challenges brought by dynamic changes in terrain, and also cannot adjust the planning scheme in real time to adapt to sudden environmental changes. This makes the power grid line prone to stability problems when facing extreme weather and landform evolution. In the transition zone between desert gobi and loess plateau, the terrain of alternating sand dunes and gullies brings unique difficulties to power distribution network planning. The gully debris flow caused by heavy rain can change the accumulation pattern of sand dunes, leading to irregular changes in terrain. The pressure of sand dune accumulation further aggravates the collapse risk of gully edge, forming the fault phenomenon of terrain point cloud data. This fault not only destroys the continuity of the terrain model, but also changes the ground laser reflectivity due to the collapse of the wet layer of sand dunes, thereby interfering with the accuracy of unmanned aerial vehicle and satellite remote sensing data. The emergence of terrain point cloud fault makes it difficult for traditional static modeling methods to capture real-time changes in terrain, and the reflectivity anomaly reduces the reliability of remote sensing data fusion. These problems are interrelated, the decline in modeling accuracy caused by terrain point cloud fault further aggravates the difficulty of adjusting the line crossing scheme, and ultimately threatens the stability of the power grid.

[0003] Therefore, how to fuse unmanned aerial vehicle oblique photography, satellite remote sensing and geographic information system to reconstruct the terrain model in real time and dynamically adjust the line crossing scheme has become a key problem to ensure the adaptability and stability of power distribution network planning in complex terrain. SUMMARY

[0004] In view of the above existing problems, the present application is proposed.

[0005] The present application provides a power distribution network intelligent planning method, system, device and storage medium to solve the problem of how to fuse unmanned aerial vehicle oblique photography, satellite remote sensing and geographic information system to reconstruct the terrain model in real time and dynamically adjust the line crossing scheme.

[0006] To solve the above technical problems, the present application provides the following technical solutions:

[0007] In a first aspect, the present application provides a power distribution network intelligent planning method, comprising:

[0008] Acquiring point cloud data, extracting terrain feature information from the point cloud data, and obtaining a terrain dataset;

[0009] Acquiring remote sensing data, and extracting surface change information from the remote sensing data to obtain a surface change dataset;

[0010] fusing the terrain dataset and the surface change dataset to obtain a terrain point cloud dataset;

[0011] Constructing a first terrain model based on the terrain point cloud dataset and analyzing the debris flow scouring path in the gully edge collapse area;

[0012] Acquire UAV monitoring data and combine it with the first terrain model to simulate the line crossing path, optimize the crossing points and tower base coordinates, and obtain a power grid planning solution;

[0013] The stability of the power grid planning scheme is evaluated, and the tower base reinforcement parameters are adjusted to obtain the final power grid planning scheme.

[0014] As a preferred solution of the distribution network intelligent planning method described in the present invention, the terrain dataset includes:

[0015] Calculate the reflectivity value of the sampling point based on the point cloud data;

[0016] Performing cluster analysis on the reflectivity values ​​using the first clustering method to obtain abnormal reflectivity areas and extracting the point cloud spatial coordinates of the abnormal reflectivity areas;

[0017] Perform the first elevation interpolation on the point cloud spatial coordinates to generate a terrain change distribution map, extract the terrain change boundary line, and determine the terrain change area;

[0018] Extract visible light image data from areas with terrain changes, calculate texture feature parameters and color values, and determine the exposed areas of the wet layer based on the color values ​​and preset dry sand reference color values;

[0019] According to the judgment result, the first dimensionality reduction method is used to reduce the dimensionality of the texture feature parameters and color values ​​of the exposed area of ​​the wet layer to generate a feature vector;

[0020] The terrain dataset is obtained according to the spatial distribution of the feature vectors.

[0021] The beneficial effects of this preferred technical solution are to improve the accuracy of identifying terrain changes and enhance the accuracy of determining wet layer exposure areas.

[0022] As a preferred solution of the distribution network intelligent planning method described in the present invention, the surface change dataset includes:

[0023] Spectral values of remote sensing data are extracted, a land cover type distribution map is calculated, and a land dynamic change area is determined based on spectral value differences and spatial position offsets;

[0024] A first segmentation method is used to separate a bare sand distribution area from the land dynamic change area, a dune ridge line is extracted, ridge line position offsets are calculated, a dune wind erosion migration area is generated, and a dune wind erosion migration area is obtained;

[0025] Slope data and elevation change information are extracted from the dune wind erosion migration area, a potential confluence area is calculated, and the influence range of a gully mudslide is comprehensively judged;

[0026] Based on the influence range of the gully mudslide and the multi-temporal elevation data of the dune wind erosion migration area, a land change data set is generated.

[0027] The beneficial effects of the preferred technical solution are to improve the accuracy of land dynamic change recognition and enhance the ability of dune migration and mudslide risk assessment.

[0028] As a preferred scheme of the power distribution network intelligent planning method, the terrain point cloud data set includes:

[0029] Same-name registration points are extracted from the overlapping area of the terrain data set and the land change data set by a first matching method, and spatial transformation parameters are calculated;

[0030] The terrain data set and the land change data set are coordinate-converted according to the spatial transformation parameters to obtain a registered terrain data set;

[0031] The registered terrain data set and the land change data set are evaluated in terms of elevation difference value distribution in the overlapping area, and an elevation discontinuity area is marked;

[0032] A buffer area is established in the elevation discontinuity area, unmanned aerial vehicle image point clouds and satellite remote sensing point clouds are fused, and a terrain point cloud data set is generated.

[0033] The beneficial effects of the preferred technical solution are to improve the accuracy of multi-source data fusion, enhance the continuity and reliability of the terrain model.

[0034] As a preferred scheme of the power distribution network intelligent planning method, the construction of the first terrain model includes:

[0035] An irregular triangle network terrain surface is generated according to the terrain point cloud data set, slope and slope direction values are calculated, and a potential mudslide occurrence area is determined;

[0036] A digital elevation model grid is generated from the potential mudslide occurrence area, water flow direction and confluence accumulation values are determined, scour depth is calculated, path center line data and influence range polygon data are generated;

[0037] The first terrain model is constructed according to the path center line data and the influence range polygon data.

[0038] As a preferred scheme of the power distribution network intelligent planning method, the optimization of the crossing ditch point and the tower foundation coordinate comprises:

[0039] The ditch distribution data and the elevation information are extracted from the first terrain model, the initial line route is calculated, and the initial crossing ditch point coordinate is determined;

[0040] According to the sand dune change monitored by the unmanned aerial vehicle, the initial crossing ditch point coordinate is adjusted to a stable area, and the tower foundation candidate position is determined;

[0041] The line route corridor is constructed from the tower foundation candidate position, the comprehensive risk index is calculated, the optimized path is searched, and the power grid planning scheme containing the tower foundation coordinate and the line center line is generated.

[0042] The preferred technical scheme has the beneficial effects of improving the line path safety and reducing the sand dune movement and geological disaster risk.

[0043] As a preferred scheme of the power distribution network intelligent planning method, the final power grid planning scheme comprises:

[0044] The multi-temporal orthographic image is acquired, the displacement amount of the sand dune ridge line and the slope foot line feature point is calculated, and the sand dune movement rate distribution map is generated;

[0045] According to the sand dune movement rate distribution map and the elevation data, the sand dune morphological evolution trend is predicted, and the sand dune movement influence range is determined;

[0046] The initial crossing ditch point coordinate is adjusted according to the sand dune movement influence range, and the tower foundation candidate position is recalculated;

[0047] The geological survey data of the tower foundation candidate position is acquired, the stability evaluation value is calculated, and the tower foundation to be reinforced is marked;

[0048] The three-dimensional foundation model is established based on the soil layer distribution parameters of the tower foundation to be reinforced, the stress distribution field and the displacement field are calculated, the foundation depth increase value and the anchor rod parameter are determined;

[0049] The line layout map is generated according to the foundation depth increase value and the anchor rod parameter, the buffer zone of the debris flow high-risk area is set, and the final power grid planning scheme is generated.

[0050] In a second aspect, the present application provides a power distribution network intelligent planning system, comprising:

[0051] The terrain feature information extraction module is used for acquiring point cloud data, performing terrain feature information extraction on the point cloud data, and obtaining a terrain data set;

[0052] The ground surface change information extraction module is configured to acquire remote sensing data, perform ground surface change information extraction on the remote sensing data, and obtain a ground surface change data set.

[0053] The terrain data fusion module is configured to fuse the terrain data set and the ground surface change data set to obtain a terrain point cloud data set.

[0054] The model construction module is configured to construct a first terrain model according to the terrain point cloud data set and analyze a debris flow scouring path of a gully edge collapse area.

[0055] The path planning module is configured to acquire unmanned aerial vehicle monitoring data, simulate a line crossing gully path in combination with the first terrain model, optimize a crossing gully point and a tower foundation coordinate, and obtain a power grid planning scheme.

[0056] The stability evaluation module is configured to evaluate the stability of the power grid planning scheme, adjust a tower foundation reinforcement parameter, and obtain a final power grid planning scheme.

[0057] In a third aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and wherein the processor implements the steps of the power distribution network intelligent planning method when executing the computer program.

[0058] In a fourth aspect, the present application provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the steps of the power distribution network intelligent planning method.

[0059] Compared with the prior art, the present application has the following beneficial effects: the present application discloses a power distribution network intelligent planning method, which identifies sand dune wet layer exposure features and ground surface change features through unmanned aerial vehicle images and satellite remote sensing data, fuses to generate a unified terrain point cloud data set, establishes a three-dimensional GIS terrain model (geographic information system terrain model), combines sand dune movement analysis and debris flow risk evaluation, optimizes line crossing gully paths and tower foundation arrangement, adjusts tower foundation reinforcement parameters, and finally outputs a power grid planning scheme suitable for a desert environment. The present application realizes accurate modeling of complex desert terrain, effectively avoids geological disaster risks such as sand dune movement and debris flow, improves the scientificity and reliability of power grid planning, and provides technical support for power grid construction in desert areas. BRIEF DESCRIPTION OF DRAWINGS

[0060] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0061] Figure 1 The overall flow logic diagram of the power distribution network intelligent planning method provided for an embodiment of the present application is shown. DETAILED DESCRIPTION

[0062] In order to make the above objectives, characteristics and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should fall within the scope of protection of the present application.

[0063] Embodiment 1, refer to Figure 1 For an embodiment of the present application, a power distribution network intelligent planning method is provided, comprising:

[0064] S100: Obtain point cloud data, extract terrain feature information from the point cloud data, and obtain a terrain data set;

[0065] S200: Obtain remote sensing data, extract surface change information from the remote sensing data, and obtain a surface change data set;

[0066] S300: Fuse the terrain data set and the surface change data set to obtain a terrain point cloud data set;

[0067] S400: According to the terrain point cloud data set, analyze the debris flow erosion path of the gully edge collapse area, and construct a first terrain model;

[0068] S500: Obtain unmanned aerial vehicle monitoring data, simulate the line crossing path of the gully based on the first terrain model, optimize the crossing point and tower foundation coordinates, and obtain a power grid planning scheme;

[0069] S600: Perform stability evaluation on the power grid planning scheme, adjust the tower foundation reinforcement parameters, and obtain a final power grid planning scheme.

[0070] It should be noted that through multi-source data fusion and dynamic analysis, the accuracy and safety of power distribution network planning under complex terrain are improved, geological disaster risks such as sand dune movement and debris flow are effectively avoided, and the long-term stability and reliability of the power grid are enhanced.

[0071] In the embodiment of the present application, the above step S100 comprises the following sub-steps A1-A6;

[0072] In A1: Calculate the reflectivity value of the sampling point according to the point cloud data;

[0073] In A2: Use a first clustering method to perform clustering analysis on the reflectivity value to obtain an abnormal reflectivity area, and extract the point cloud spatial coordinates of the abnormal reflectivity area;

[0074] In A3: the first elevation interpolation is performed on the point cloud space coordinates, a terrain change amount distribution map is generated, a terrain change boundary line is extracted, and a terrain change region is determined;

[0075] In A4: visible light image data in the terrain change region is extracted, texture feature parameters and color values are calculated, and a wet layer exposure region is determined according to the color values and a preset dry sand reference color value;

[0076] In A5: according to the determination result, the texture feature parameters and the color values of the wet layer exposure region are processed by a first dimension reduction method, and a feature vector is generated;

[0077] In A6: the terrain data set is obtained according to the spatial distribution of the feature vector.

[0078] In an optional embodiment, the first clustering method can be a K-means clustering method, and the reflectivity values of each sampling point in the point cloud data are divided into K clusters, wherein some clusters can correspond to abnormal reflectivity regions, and by setting a distance threshold or combining domain knowledge, a central region far away from the main cluster is identified as an abnormal reflectivity region;

[0079] In an optional embodiment, the first clustering method can be a DBSCAN density clustering method, and the spatial coordinates and reflectivity values of the point cloud data are identified to obtain high-density regions and low-density regions, and the DBSCAN automatically marks the low-density points as noise, and these noise points are the abnormal reflectivity regions;

[0080] In the embodiment of the present application, the first clustering method includes clustering analysis of the reflectivity values by a Gaussian mixture model;

[0081] The point cloud data collected by the unmanned aerial vehicle equipped with a laser radar is obtained, and the reflectivity values of each sampling point are calculated, the clustering analysis of the reflectivity values is performed by a Gaussian mixture model, the abnormal reflectivity region is determined according to the point cloud set deviating from the center of the main distribution by more than a preset standard deviation threshold in the clustering result, the point cloud spatial coordinates and the corresponding reflectivity values in the abnormal region are extracted to form an abnormal point set, and the preset standard deviation threshold is a multiple of the standard deviation.

[0082] Specifically, when the unmanned aerial vehicle equipped with a laser radar performs terrain monitoring, the laser pulse will produce different intensity echo signals when encountering different ground materials, and the echo intensity is the reflectivity. In a sand dune environment, the reflectivity of dry sand layer is usually uniform, and when the underground wet layer is exposed due to wind erosion or other factors, the reflectivity will change significantly due to the difference in water content. Therefore, the Gaussian mixture model can effectively identify multiple distribution patterns in the data.

[0083] The Gaussian mixture model assumes that the reflectance data is composed of multiple Gaussian distributions, and the mean, variance and weight parameters of each distribution are estimated by the expectation maximization algorithm. When the reflectance value of some points deviates from the center of the main distribution by more than a preset standard deviation threshold, these points correspond to areas where the physical properties of the ground surface change, such as wet layer exposure, vegetation coverage or rock outcrop, etc.

[0084] In an optional embodiment, the first elevation interpolation can be inverse distance weighted interpolation, the point cloud data is spatially divided, the neighboring points of each point to be interpolated are determined; the weight of each neighboring point is calculated; the elevation value of the point to be interpolated is calculated by weighted average;

[0085] In an optional embodiment, the first elevation interpolation can be Kriging interpolation, the spatial correlation of the point cloud data is analyzed to determine the variogram model; the optimal weight is calculated according to the variogram and the position of the neighboring points; the elevation of the point to be interpolated is calculated using the weight and the elevation values of the neighboring points; the interpolation variance is calculated to evaluate the interpolation accuracy;

[0086] In the embodiment of the present application, the first elevation interpolation includes a time series comparison method, which is based on comparing digital elevation models of different periods to calculate the elevation difference of each grid point. The dynamic change of dune terrain mainly manifests as dune migration, wind erosion depression and accumulation increase, etc. When the change amplitude exceeds a preset change threshold, it indicates that the area has undergone significant terrain change. Such change is often closely related to the exposure of underground wet layer, because the wet sand layer has stronger anti-wind erosion capacity, and when the upper dry sand is blown away, the wet layer will form a relatively stable terrain feature.

[0087] Specifically, the spatial coordinates in the abnormal point set are subjected to terrain elevation interpolation operation to generate a digital elevation model, a time series comparison method is used to calculate the difference between adjacent period digital elevation models to obtain a terrain change amount distribution map, a terrain change boundary line is extracted from the area in the change amount distribution map where the change amplitude is greater than a preset change threshold, and a terrain change area surrounded by the terrain change boundary line is determined.

[0088] The UAV visible light image data corresponding to the position of the terrain change area is obtained, the RGB (red, green, and blue) color values are extracted from the visible light image data and a gray level co-occurrence matrix is calculated, the texture feature parameters including contrast, correlation, energy and homogeneity are calculated through the gray level co-occurrence matrix, the wetness degree is judged according to the Euclidean distance between the RGB color values and a preset dry sand reference color value, and if the Euclidean distance is greater than a preset distance threshold, it is determined as a wet layer exposure feature area.

[0089] It should be noted that it is difficult to comprehensively judge the wet layer exposure feature only by relying on the lidar data, and therefore it is necessary to comprehensively analyze in combination with the visible light image. The RGB color value can intuitively reflect the color feature of the ground surface, and the wet sand layer usually presents a relatively deep color, which is in sharp contrast with the light yellow color of the dry sand layer. The gray level co-occurrence matrix describes the texture feature by statistically analyzing the spatial relationship of the gray level values of adjacent pixels in the image. The contrast parameter reflects the clarity of the texture, the wet layer exposure area presents a relatively high contrast due to uneven distribution of water; the correlation parameter describes the linear correlation degree between pixels; the energy parameter represents the uniformity of the texture; and the homogeneity parameter reflects the similarity of the local texture. By calculating the Euclidean distance of the color value and the preset dry sand reference color value, the wetness degree can be quantitatively evaluated.

[0090] In an optional embodiment, the first dimension reduction method can be t-SNE (t-distribution stochastic neighbor embedding), each wet layer exposure area sample is represented as a high-dimensional feature vector, the high-dimensional feature vector is mapped to a two-dimensional or three-dimensional space by using t-SNE, and the clustering of the wet layer exposure area in the low-dimensional space is observed;

[0091] In an optional embodiment, the first dimension reduction method can be LDA (linear discriminant analysis), the mean and the covariance matrix of each category are calculated, the inter-class scatter matrix and the intra-class scatter matrix are constructed, the generalized eigenvalue problem is solved, and the optimal projection direction is found, so that the inter-class variance and the intra-class variance of the projected data in the low-dimensional space are maximum and minimum respectively.

[0092] In the embodiment of the application, the first dimension reduction method includes the principal component analysis method, which can effectively integrate the multi-dimensional feature information.

[0093] Specifically, the principal component analysis method is used to perform dimension reduction processing on the texture feature parameters, the color value Euclidean distance and the wetness degree index of the wet layer exposure feature area to obtain a feature vector, a sand dune wet layer exposure feature data set is constructed according to the spatial distribution law of the feature vector in the terrain change area, a wet layer exposure mode classification is determined by calculating the similarity matrix between each feature vector in the data set, and a terrain data set containing terrain change information and wet layer exposure features is obtained.

[0094] It should be noted that the principal component analysis method converts multiple related variables such as original texture parameters, color distance and wetness degree into a set of linearly independent principal components through linear transformation. The first principal component usually contains the largest data variation information, which may mainly reflect the humidity change; the second principal component may mainly reflect the texture complexity. The spatial distribution law of the characteristic vector in the topographic change area reveals the spatial pattern of the wet layer exposure, such as the strip-shaped distribution which may be related to the wind direction, and the patch-shaped distribution which may be related to the local topographic relief. The similarity matrix can identify the areas with similar exposure characteristics by calculating the cosine similarity or Euclidean distance between the characteristic vectors at different positions, and then determine the typical wet layer exposure mode. This method not only can accurately identify the current wet layer exposure condition, but also can provide an important basis for predicting the future topographic evolution trend, and has important application value for desertification monitoring and ecological environment assessment.

[0095] In the embodiment of the present application, the above step S200 comprises the following sub-steps B1-B2.

[0096] In B1, the spectral value of remote sensing data is extracted, the distribution map of land cover type is calculated, and the area of land dynamic change is determined based on the spectral value difference and spatial position offset.

[0097] In B2, the first segmentation method is used to separate the exposed sand distribution area from the area of land dynamic change, the sand dune ridge line is extracted, the ridge line position offset is calculated, the sand dune erosion migration area is generated, and the sand dune erosion migration area is obtained.

[0098] In B3, the slope data and elevation change information are extracted from the sand dune erosion migration area, the potential convergence area is calculated, and the influence range of gully mudflow is comprehensively judged.

[0099] In B4, the land change data set is generated based on the multi-temporal elevation data of the gully mudflow influence range and the sand dune erosion migration area.

[0100] In an optional embodiment, the first segmentation method can be a supervised classification method based on pixel classification, training samples of exposed sand and water body are selected, spectral features of each type of sample are extracted, a maximum likelihood classifier is used to classify the entire remote sensing image, and the exposed sand and water body area in the classification result is extracted to realize segmentation.

[0101] In an optional embodiment, the first segmentation method can be a region growing method based on image segmentation, seed points of exposed sand and water body are manually or automatically selected in the remote sensing image, similarity criteria are set, and from the seed points, adjacent similar pixels are gradually merged to form complete regions, and finally the exposed sand and water body distribution area is extracted.

[0102] In the embodiment of the present application, the first segmentation method comprises separating the exposed sand distribution area and the water system distribution area from the dynamic change area by using a threshold segmentation method.

[0103] Specifically, multi-temporal satellite remote sensing image data is acquired and spectral values of each band are extracted, a land cover type distribution map is calculated by using a normalized vegetation index and a normalized water index, a dynamic change area of the land surface is determined according to spectral value differences and spatial position offsets of adjacent time phase images, and an exposed sand distribution area and a water system distribution area are separated from the dynamic change area by using a threshold segmentation method. Texture feature extraction is performed on the exposed sand distribution area, a sand dune ridge line is identified by using an edge detection operator, a sand dune movement vector is calculated by using spatial offsets of ridge line positions in different time phases, a wind erosion degree parameter is determined according to a sand dune height profile and a slope change, and if the wind erosion degree parameter exceeds a preset threshold value, the area is determined as a sand dune wind erosion migration area.

[0104] Topographic slope data and elevation change information are extracted from the sand dune wind erosion migration area, a potential confluence area is calculated by using a topographic humidity index, a comprehensive judgment is performed on the slope data, the elevation change information and the confluence area distribution by using a random forest algorithm, and an influence range of a gully mudslide is determined according to an area in which a probability value in a judgment result exceeds a preset probability threshold value.

[0105] Multi-temporal elevation data of the gully mudslide influence range and the sand dune wind erosion migration area are acquired, an erosion depth spatial distribution is calculated by using an elevation difference calculation method, a topographic dynamic change feature vector is constructed according to the erosion depth spatial distribution, the sand dune movement vector and the wind erosion degree parameter, and a land surface change data set containing displacement, shape and erosion depth is generated by integrating time and space distribution information of all feature vectors.

[0106] Specifically, acquisition of multi-temporal satellite remote sensing image data involves repeated observation of the same geographical area at different time points. A normalized vegetation index reflects a land surface vegetation coverage condition by calculating a ratio of a difference value and a sum value of a near-infrared band and a red band, and a value range thereof is between-1 and 1, and a vegetation coverage area usually presents a higher positive value. A normalized water index identifies water body distribution by using a combination of a green band and a near-infrared band, and a water body area presents a positive value while other objects present a negative value. Combination use of the two indexes can effectively distinguish an exposed sand area, a vegetation coverage area and a water system distribution area.

[0107] Specifically, the determination of the dynamic change area of the ground surface is based on the difference analysis of adjacent time phase images. When the sand dune migrates by wind erosion, the area originally covered by sand may expose the underlying ground surface, while other areas are covered by new sand layers. This change is manifested as a significant difference in reflectivity in spectral characteristics. The detection of spatial position offset is realized by feature point matching or correlation coefficient calculation, which can quantify the moving distance and direction of the ground object. The threshold segmentation method automatically determines the segmentation threshold according to the statistical distribution characteristics of the spectral value, so as to accurately separate the bare sand land and water system area.

[0108] It should be noted that the sand dune ridge line as an important feature of the sand dune morphology, its position change directly reflects the moving state of the sand dune. The edge detection operator such as Sobel operator or Canny operator can identify the gradient change in the image, and the sand dune ridge line usually corresponds to the maximum value position of the elevation gradient. By tracking the ridge line positions of different time phases, the moving vector of the sand dune can be calculated, including the moving speed and the moving direction. The extraction of the sand dune height profile needs to sample along the direction perpendicular to the ridge line to obtain the cross-sectional morphology of the sand dune. The slope change reflects the intensity of the wind erosion, and the steepening of the windward slope and the slowing down of the leeward slope are the typical characteristics of the wind erosion.

[0109] Specifically, the calculation of the terrain humidity index comprehensively considers the influence of the upstream catchment area and the local slope. In desert areas, although the precipitation is rare, the low-lying areas may still form temporary confluence after extreme precipitation events. These potential confluence areas are high-risk areas for debris flow. As an integrated learning method, the random forest algorithm can effectively handle the nonlinear relationship of multidimensional data such as terrain factors and confluence characteristics by constructing multiple decision trees and integrating their prediction results. The algorithm outputs the probability value of debris flow occurrence for each area, and when the probability exceeds the preset threshold, it is determined as the debris flow affected range.

[0110] It should be noted that the elevation difference calculation is realized by comparing the digital elevation models of different periods. The spatial distribution of erosion depth presents obvious spatial heterogeneity, and the erosion depth in gully areas is usually larger, while the erosion in flat areas is relatively smaller. The construction of the terrain dynamic change feature vector integrates multiple dimensions of information: the displacement feature reflects the horizontal movement of the ground surface, the morphological feature describes the change of the terrain profile, and the erosion depth quantifies the loss of material in the vertical direction. The comprehensive expression of such multi-dimensional features provides comprehensive data support for subsequent surface change pattern recognition and prediction. Through the integration of spatial and temporal distribution information, the formed surface change data set not only records the current state of the ground surface, but also retains the historical evolution information, providing important basic data for geological disaster warning and ecological environment assessment.

[0111] In the embodiment of the present application, the above-mentioned step S300 includes the following sub-steps C1-C4;

[0112] In C1: extract homonymous registration point pairs from the overlapping area of the terrain dataset and the ground change dataset by the first matching method, and calculate the spatial transformation parameters;

[0113] In C2: coordinate transform the terrain dataset and the ground change dataset according to the spatial transformation parameters, to obtain a registered terrain dataset;

[0114] In C3: evaluate the elevation difference value distribution of the registered terrain dataset and the ground change dataset in the overlapping area, and mark the elevation discontinuous area;

[0115] In C4: establish a buffer zone in the elevation discontinuous area, fuse the UAV image point cloud and the satellite remote sensing point cloud, and generate a terrain point cloud dataset.

[0116] In an optional embodiment, the first matching method can be an edge detection segmentation method, the remote sensing image is subjected to gray processing, a Canny or Sobel operator is used to detect the image edge, the boundary of the bare sand and the water body is divided according to the edge information, the morphological operation is combined to remove the noise, and the complete area is extracted;

[0117] In an optional embodiment, the first matching method can be a region growing method, the seed points of the bare sand and the water body are manually or automatically selected in the remote sensing image, the similarity threshold is set according to the specific condition, the similar pixels are gradually merged from the seed points, the complete area is formed, and finally the distribution area of the bare sand and the water body is extracted;

[0118] In the embodiment of the application, the first matching method includes using a feature point matching method to extract homonymous registration point pairs from the overlapping area of the two datasets;

[0119] Specifically, the coordinate system parameters of the terrain dataset and the land surface change dataset are obtained, the spatial reference systems of the two datasets are unified through a coordinate conversion matrix, the same name registration point pairs are extracted from the overlapping area of the two datasets by using a feature point matching method, and the registration transformation matrix containing translation, rotation and scaling parameters is calculated according to the registration point pairs. The registration transformation matrix is used to perform coordinate transformation on the terrain dataset to obtain a registered terrain dataset, and the accuracy is evaluated by calculating the elevation difference distribution of the registered terrain dataset and the land surface change dataset in the overlapping area. If the elevation difference exceeds a preset threshold, the area is marked as an elevation discontinuous area, and the boundary line coordinate set of the elevation discontinuous area is extracted. A buffer range is established according to the boundary line coordinate set, the point density of the unmanned aerial vehicle image point cloud and the satellite remote sensing point cloud in the buffer range is calculated respectively, the elevation values of the two kinds of point clouds are fused and calculated by using a distance inverse weighting method, the optimal weight coefficient is determined by minimizing the fusion error, and a buffer fusion point cloud is generated. The buffer fusion point cloud and all point cloud data in the registered terrain dataset and the land surface change dataset outside the buffer are obtained, the merged point cloud is grid processed by using a three-dimensional Delaunay triangulation method, the spatial interpolation calculation is performed according to the neighborhood relationship of the grid nodes, and the spatial continuous unified terrain point cloud dataset is generated by point cloud density uniformization processing.

[0120] The fusion registration of the terrain dataset and the land surface change dataset is a key link to realize the unified expression of multi-source data. The coordinate system parameters include projection type, central meridian, ellipsoid parameters and other elements. Different data sources may use different coordinate systems, such as local coordinate system for unmanned aerial vehicle data and WGS84 coordinate system for satellite data. The coordinate conversion matrix realizes the conversion between different coordinate systems through seven-parameter or four-parameter conversion method, ensuring the consistency of spatial position. The feature point matching method extracts feature points with rotation and scale invariance from two datasets based on SIFT or SURF feature descriptors, and finds corresponding same name point pairs through similarity measurement of feature vectors.

[0121] The registration transformation matrix contains three translation parameters, three rotation parameters and one scaling parameter. The least square method is used to solve these parameters to minimize the sum of squares of distances between same name point pairs. The accuracy evaluation after registration is crucial because the terrain dataset is usually obtained from high-precision unmanned aerial vehicle lidar, and the land surface change dataset is obtained from satellite remote sensing with relatively low resolution. Systematic deviations may exist in the overlapping area of the two. The grid method is used to calculate the average elevation difference in each grid when the elevation difference of a certain area exceeds the preset threshold, which often means that there is a terrain change caused by the time difference of data acquisition, or the system error of the two data sources is large.

[0122] It should be noted that the boundary line extraction of the elevation discontinuous area adopts the gradient analysis method, and the mutation position is identified by calculating the spatial gradient of the elevation difference. The buffer zone is established to provide a transition area during data fusion, so as to avoid discontinuity caused by hard splicing. The difference in point density reflects the spatial resolution characteristics of different data sources. The point cloud density of the unmanned aerial vehicle is usually hundreds of points per square meter, and the point cloud density derived from satellite remote sensing is relatively low. The core idea of the distance inverse weighting method is that the closer the points, the greater the contribution to the fusion result, and the weight is inversely proportional to the distance. This method can achieve smooth transition and avoid mutation of the fusion boundary.

[0123] Specifically, the determination of the optimal weight coefficient is realized by iterative optimization. The initial weight can be set according to the point density ratio, and then the deviation of the fused elevation value from the original data is calculated, and the weight is adjusted to minimize the deviation. Three-dimensional Delaunay triangulation is a spatial data organization method, which constructs discrete point cloud data into non-overlapping tetrahedral grids, and each tetrahedron does not contain other points in its circumscribed sphere. This data structure is beneficial to subsequent spatial interpolation and query operations. After gridding, there may be data gaps in some areas, especially in places where the original data coverage is incomplete. Spatial interpolation estimates the elevation of the gap position by the elevation values of the adjacent points. Common methods include natural neighborhood interpolation or spline interpolation.

[0124] It should be noted that the point cloud density uniformization processing solves the problem of density difference of different data sources. This process divides the space into regular cubic units through the voxelization method, and retains one representative point in each voxel, thereby realizing the uniformity of the density. This processing not only reduces data redundancy, but also improves the efficiency of subsequent processing. The generated uniform terrain point cloud dataset combines the advantages of high-precision local data of unmanned aerial vehicles and wide coverage data of satellites, realizes the organic combination of multi-scale terrain information, and provides complete, continuous and high-quality basic data for terrain analysis and change monitoring.

[0125] In the embodiment of the present application, the above step S400 includes the following sub-steps D1-D3;

[0126] In D1: generating an irregular triangular mesh terrain surface according to the terrain point cloud dataset, calculating the slope and slope direction values, and determining the potential debris flow occurrence area;

[0127] In D2: generating a digital elevation model grid for the potential debris flow occurrence area, determining the water flow direction and flow accumulation value, calculating the scour depth, and generating path center line data and impact range polygon data;

[0128] In D3: constructing a first terrain model according to the path center line data and the impact range polygon data.

[0129] Specifically, a unified terrain point cloud dataset is obtained and the density value of each point is calculated. The number of points in a unit volume is counted through an octree spatial index structure. If the point density value is lower than the preset density threshold, it is marked as a sparse region. The data missing boundary is determined according to the spatial connectivity analysis of the sparse region. The local geometric features of the point cloud in the data missing boundary are calculated. The normal vector and curvature value of each point are obtained through principal component analysis. The abnormal points of the dune accumulation form are identified according to the mutation degree of the curvature value. If the difference between the curvature value of a point and the average curvature of the neighborhood exceeds the preset threshold, the point is determined to be an abnormal point and is removed. The preliminary optimized point cloud is obtained. The gully edge points are identified from the preliminary optimized point cloud through elevation gradient calculation. The collapse affected area is determined according to the spatial distribution characteristics and elevation change rate of the edge points. The point cloud coordinates in the collapse affected area are iteratively adjusted using the Laplace smoothing algorithm to obtain the edge smoothed point cloud data. The sparse regions in the edge smoothed point cloud data are encrypted using the Kriging interpolation method. The quality is evaluated according to the spatial distribution consistency of the interpolated points and the original points. The surface of the local area whose interpolation quality does not meet the preset standard is reconstructed by the Poisson reconstruction method to generate an optimized terrain point cloud dataset with uniform point density and complete terrain features.

[0130] It should be noted that the octree spatial index structure is a high-efficiency three-dimensional spatial data organization method, which recursively divides a three-dimensional space into eight subspaces, forming a tree-like hierarchical structure. In point cloud density analysis, each node of the octree records the number of points it contains. By traversing the tree structure, the point density in any spatial region can be quickly counted. When the number of points in a voxel is lower than the preset threshold, the region is marked as a sparse region. Spatial connectivity analysis distinguishes isolated sparse points from large missing areas by checking the sparse state of adjacent voxels, thereby accurately identifying the true data missing boundary.

[0131] Principal component analysis plays an important role in extracting local geometric features of point cloud. For the neighborhood point set of each point, a covariance matrix is constructed and its eigenvalues and eigenvectors are calculated. The eigenvector corresponding to the smallest eigenvalue is the normal vector of the point, and the relative size relationship of the three eigenvalues reflects the geometric properties of the local surface. The curvature value is calculated by the ratio of the eigenvalues. The curvature of a flat area is close to zero, while the curvature value of a dune ridge or accumulation edge is larger. The irregular accumulation of dunes under the action of wind will produce local curvature anomalies. If these abnormal points are not removed, they will affect the accuracy of subsequent terrain analysis.

[0132] The calculation of the elevation gradient uses the finite difference method, which obtains the ratio of the elevation difference between adjacent points to the horizontal distance. The edge of the gully is usually characterized by a sharp change in the elevation gradient, especially in the scarp area formed by erosion. The determination of the collapse-affected area not only considers the size of the gradient but also analyzes the consistency of the gradient direction. The Laplace smoothing algorithm achieves smoothing by adjusting the position of each point to the weighted average position of its neighborhood points. In the iterative process of this algorithm, the movement of edge points is constrained, which not only maintains the overall topographic features but also eliminates local irregularities.

[0133] The Kriging interpolation method is based on the principles of geostatistics and takes into account the correlation and variability of spatial data. This method establishes a variogram model through the spatial distribution and attribute values of known points to predict the attribute values of unknown locations. In the point cloud densification process, Kriging interpolation not only fills in data gaps but also maintains the continuity and smoothness of the terrain. The quality assessment of the interpolated points is achieved through cross-validation, which temporarily removes some known points, uses the remaining points for interpolation prediction, and then compares the difference between the predicted value and the actual value.

[0134] It should also be noted that the Poisson reconstruction method converts the point cloud surface reconstruction problem into the solution of the Poisson equation. This method uses the normal vector information of the point cloud to construct the gradient field of the indicator function, solves the Poisson equation to obtain an implicit surface function, and finally extracts the isosurface as the reconstructed surface. This method is particularly suitable for handling noisy and incomplete point cloud data and can generate smooth and continuous surfaces. For areas where the interpolation quality does not meet the requirements, Poisson reconstruction can generate reasonable surface morphology based on the distribution trend of surrounding points. The optimized terrain point cloud dataset not only has uniform density, eliminating the influence of abnormal points and data missing, but also maintains the main features of the original terrain, providing a high-quality data foundation for subsequent terrain analysis and application.

[0135] Specifically, the optimized terrain point cloud dataset is acquired and a Delaunay triangulation algorithm is used to generate an irregular triangular mesh terrain surface. The slope and aspect values are obtained by calculating the normal vector of each triangular facet. The potential debris flow occurrence area is determined according to the continuous region with slope values greater than a preset threshold, and the boundary coordinates are output. The irregular triangular mesh within the boundary coordinates of the potential debris flow occurrence area is converted to a digital elevation model grid through rasterization. The D8 flow direction algorithm is used to determine the water flow direction of each grid cell. The upstream catchment area of each grid is accumulated to obtain the flow accumulation value. The debris flow starting point position is determined according to the grid whose product of the flow accumulation value and the slope value exceeds a preset threshold. The downstream path is tracked from the debris flow starting point position along the direction of the maximum slope. The local slope sequence is obtained by calculating the ratio of the elevation difference and the horizontal distance between adjacent points on the path. The scour depth at each position is calculated according to the changes of the local slope sequence and the flow accumulation value. The debris flow impact boundary is determined by the lateral extension range of the scour depth, and the path centerline data and the impact range polygon data are generated. The path centerline data and the impact range polygon data are converted into vector element format. The element dataset is created by establishing the spatial topological relationship and attribute association between elements. The first terrain model containing three-dimensional visualization and spatial query function, i.e., the GIS terrain model, is constructed according to the original irregular triangular mesh terrain surface, the vector element dataset and the scour depth attribute table.

[0136] It should be noted that the Delaunay triangulation algorithm has unique advantages in terrain modeling. The algorithm ensures that the generated triangles are as close to equilateral triangles as possible, avoiding excessively narrow triangles, thereby improving the expression quality of the terrain surface. The algorithm is implemented by the empty circle criterion, that is, no other points are contained in the circumscribed circle of any triangle, which guarantees the uniqueness and stability of the triangular mesh. When processing terrain point clouds, each triangular facet represents a local plane. The normal vector of the facet can be obtained by calculating the plane equation formed by the three vertices. The slope value is the angle between the normal vector and the vertical direction, and the aspect is the direction angle of the normal vector projection on the horizontal plane.

[0137] Specifically, the identification of the potential debris flow occurrence area is based on the spatial distribution characteristics of the terrain slope. When the slope of a plurality of continuous triangular facets exceeds the critical value, these regions have the terrain conditions for debris flow occurrence. The extraction of the boundary coordinates is realized by tracking the contour line of the slope threshold, forming a closed polygon region. This vector-based expression method provides an accurate spatial range for subsequent rasterization conversion.

[0138] It should be noted that the conversion from the irregular triangulated network to the raster data involves a resampling process. When rasterizing, a proper resolution needs to be determined to maintain the terrain details and control the data volume. The elevation value of each raster cell is calculated by interpolating the central point on the triangulated network. The D8 flow direction algorithm is a classic method for hydrological analysis, which assumes that water flow can only flow to the cell with the lowest elevation among the eight adjacent cells. This simplification, although may not be accurate in some cases, is computically efficient and suitable for large-scale flow direction analysis.

[0139] Specifically, the calculation process of the flow accumulation is recursive. Starting from the highest point of the terrain, the upstream contributing area of each grid is accumulated step by step. The flow accumulation value reflects the size of the water that can be collected at that location, and the product of the slope value comprehensively considers the water volume and dynamic conditions. The starting point of the debris flow usually appears at the location where the product value suddenly increases, indicating that there is enough water and enough slope to provide power.

[0140] The debris flow path tracking adopts the principle of maximum slope direction, that is, starting from the starting point, the direction with the steepest slope is selected at each step. The scouring depth on the path is related to multiple factors, including local slope, flow volume, and surface material. The larger the local slope, the faster the water flow speed and the stronger the scouring ability. The lateral expansion of the scouring depth presents a central deep and edge shallow feature, and this distribution pattern determines the impact range of the debris flow. By setting a scouring depth threshold, the impact boundary can be determined to form a complete impact area polygon. The conversion of the vector element format enables the path and impact range to be managed and analyzed in the GIS environment. The establishment of spatial topological relationships includes the inclusion relationship between the path and the impact area, the intersection relationship between different paths, etc. Attribute association binds information such as scouring depth, flow rate, and occurrence probability to spatial elements.

[0141] It should be noted that the construction of the GIS terrain model integrates three-dimensional visualization function and spatial analysis function, users can not only visually view the three-dimensional form of the terrain and the debris flow path, but also perform spatial operations such as buffer analysis and overlay analysis. This integrated model provides a powerful decision support tool for debris flow disaster assessment and prevention planning.

[0142] In the embodiment of the present application, the above step S500 includes the following sub-steps E1-E3;

[0143] In E1: extract gully distribution data and elevation information from the first terrain model, calculate the initial line direction, and determine the initial gully-crossing point coordinates;

[0144] In E2: adjust the initial gully-crossing point coordinates to the stable area according to the sand dune changes monitored by the unmanned aerial vehicle, and determine the candidate position of the tower foundation;

[0145] In E3: Construct line corridors from candidate tower base locations, calculate comprehensive risk indices, search for optimized paths, and generate power grid planning solutions that include tower base geographic coordinates and line centerlines.

[0146] Specifically, the method obtains gully distribution data and elevation information from the GIS terrain model. A shortest path algorithm is used to calculate the initial route alignment and identify the intersections between the route and the gully. Initial gully crossing point coordinates are determined based on the terrain elevation and gully width on both sides of the intersection. Line of sight analysis is used to verify inter-gully visibility between the crossing points, resulting in an initial gully crossing plan that includes the crossing point coordinates and the crossing method. For each crossing point coordinate in the initial gully crossing plan, the dune movement vector and amplitude data from drone imagery are extracted. The stability influence coefficient of the crossing point location is calculated based on the dune movement vector. If the influence coefficient exceeds a preset threshold, the crossing point coordinate is adjusted to a stable area away from the dune movement path. This results in a set of stable crossing point coordinates and candidate tower base locations determined based on the spacing between the crossing points. A route corridor is constructed by connecting the stable crossing point coordinates and the tower base candidate locations. Slope values, wind erosion rates, and debris flow risk values ​​for each grid within the corridor are extracted from the GIS terrain model. A weighted overlay method is used to calculate a comprehensive risk index for each grid. A feasible path search space is determined based on the set of grids whose comprehensive risk index falls below a preset safety threshold. A dynamic programming algorithm is used to search for the optimal path with the minimum sum of construction cost and risk cost in the feasible path search space. The span parameters and rotation angle parameters are calculated based on the distance and rotation angle between adjacent tower base candidate positions on the optimized path. The tower base coordinates are adjusted by satisfying the span upper limit constraint and the rotation angle limit constraint, and a power grid planning scheme is generated that includes the tower base geographic coordinates, line centerline, span table and risk assessment data.

[0147] The application of shortest path algorithms in power grid line planning primarily considers terrain undulations and obstacle distribution. The algorithm discretizes the terrain surface into a grid of nodes, assigning a cost to each connection, including a distance cost and a terrain difficulty cost. When identifying where a line intersects a gully, the algorithm automatically searches for the optimal location to cross the gully, typically selecting an area with a narrow gully and relatively flat terrain on either side. Line-of-sight analysis methods construct terrain profiles to verify the presence of obstructions between two gully crossing points, which is crucial for ensuring the clearance height of transmission lines.

[0148] Dune movement vectors are monitored through comparative analysis of multi-temporal drone imagery. Characteristic lines, such as dune ridges and toe lines, are extracted to calculate displacement and movement direction over time. The stability impact coefficient comprehensively considers the relative relationship between dune movement speed and direction and the location of the gully crossing point. If the dune movement path threatens the stability of the tower foundation, the gully crossing point needs to be relocated to a stable area away from active dunes. Candidate tower foundation locations must not only meet gully crossing requirements but also consider accessibility and foundation bearing capacity.

[0149] Specifically, the construction of the line corridor belt adopts a buffer zone analysis method, and a certain width is expanded to both sides based on the determined crossing ditch point and tower foundation candidate position to form a search space. The slope value reflects the steepness of the terrain, and a too steep slope will increase the construction difficulty and operation and maintenance risk. The wind erosion rate is quantified by analyzing the change of surface roughness and the degradation of vegetation coverage, and a high wind erosion rate area means that the foundation may face the risk of erosion. The debris flow risk value is based on the comprehensive evaluation of factors such as convergence accumulation, slope and loose material distribution.

[0150] The weighted superposition method needs to determine the weight of each factor when calculating the comprehensive risk index. The slope factor directly affects the construction cost and line stability, and is usually given a high weight; the wind erosion rate affects the long-term operation and maintenance safety, and the weight is second; although the debris flow risk has a relatively low probability of occurrence, once it occurs, the consequences are serious, so it also needs to be given appropriate weight. The calculation of the comprehensive risk index adds the three factors after standardization according to the weight, and the risk level of each grid is obtained.

[0151] In the embodiment of the application, the application of the dynamic programming algorithm in path optimization embodies the idea of global optimization. The algorithm divides the path planning problem into multiple stages, and each stage represents the selection of a tower foundation position. The state transition equation considers the construction cost and cumulative risk cost from the previous tower foundation to the current candidate position. The construction cost includes tower foundation construction cost, line length related cost, etc.; the risk cost converts the comprehensive risk index into potential economic loss. The span constraint ensures that the distance between adjacent tower foundations does not exceed the maximum allowable span of the conductor, which is related to the conductor material, design tension and meteorological conditions. The corner restriction constraint is to avoid excessive unbalanced tension at the tower site. The generated power grid planning scheme is a comprehensive data package, which includes the precise geographic coordinates of each tower foundation, elevation information, foundation type suggestion, etc. The line center line data records the spatial orientation of the conductor, which is convenient for subsequent three-dimensional visualization display. The span table lists the length, height difference and representative span of each span in detail, which provides a basis for conductor selection and stress calculation. The risk assessment data retains the risk indicators of each tower site and line section, which provides a reference for operation and maintenance management, so that the planning scheme not only meets the current construction needs, but also fully considers the safety and economy of future operation.

[0152] By collecting sand dune accumulation shape and displacement data through regular field investigation and unmanned aerial vehicle aerial photography, analyzing sand dune shape, moving speed and accumulation height under different wind directions and wind speeds, combining historical sand dune migration records, evaluating sand dune morphological evolution trend in a preset period, identifying potential threat areas of the evolution trend to the safety of the crossing ditch path, obtaining the sand dune movement influence range, adjusting the crossing ditch point position and checking the tower foundation coordinates to outside the sand dune movement influence range.

[0153] The sand dune morphology data obtained by field survey and the multi-temporal orthophoto generated by unmanned aerial vehicle are used to identify the feature points of sand dune ridge line and slope foot line through image matching and calculate the displacement of feature points in different periods. A corresponding relationship table is established between the wind direction and wind speed values in the same period meteorological monitoring data and the displacement. The correlation analysis method is used to obtain the sand dune movement rate distribution graph under different wind directions and wind speeds. According to the sand dune movement rate distribution graph and the latest sand dune elevation data, the time series data set is constructed by integrating the sand dune elevation change sequence over the years. The autoregressive moving average model is used to fit and predict the sand dune position and height change in the preset period, and the evolution trend data containing the sand dune morphology parameters at each future time node are obtained. For the sand dune predicted position sequence in the evolution trend data, the threat degree is judged by calculating the vertical distance between each predicted position and the center line of the current crossing ditch path. If the vertical distance is less than the preset safety distance threshold, it is determined as a potential threat point. The sand dune movement influence range is determined according to the outer envelope line of all potential threat points plus the safety buffer distance. The sand dune movement influence range is used as the forbidden area constraint, and a new crossing ditch point position that meets the visibility condition and terrain stability requirement is searched outside the influence range. The tower foundation coordinates are recalculated according to the new crossing ditch point position and the line direction constraint. After verifying that the distance between the tower foundation coordinates and the boundary of the sand dune movement influence range is greater than the preset safety threshold, the adjusted crossing ditch point and tower foundation coordinates are obtained.

[0154] The feature point identification in sand dune morphology monitoring is the basis for understanding the dynamic changes of sand dune. The sand dune ridge line represents the highest point connection line of sand dune, usually showing a winding shape, while the slope foot line identifies the junction between sand dune and surrounding ground. In multi-temporal orthophoto, these feature lines will change their positions with the movement of sand dune. Image matching technology calculates the spatial displacement of the same feature points in different periods by identifying them. Meteorological monitoring data provide the time series of wind direction and wind speed, and the corresponding relationship between these data and sand dune displacement reveals the driving mechanism of wind force on sand dune movement.

[0155] Correlation analysis method is used to quantify the relationship between wind condition parameters and sand dune movement rate. When the dominant wind direction is perpendicular to the sand dune direction, the lateral movement of sand dune is most obvious; when the wind direction is parallel to the sand dune direction, it mainly shows the change of sand dune morphology. Different wind speed intervals correspond to different transport capacities, low wind speed can only transport small particles, and high wind speed can move larger sand particle blocks. By establishing the quantitative relationship between wind speed and movement rate, the movement trend of sand dune can be calculated according to the future weather forecast.

[0156] The autoregressive moving average model has good adaptability in time series prediction. The model considers that the future position of the dune depends not only on the trend of the historical position, but also on the impact of random disturbances. The autoregressive part of the model captures the inertial characteristics of the dune movement, that is, the movement pattern of the previous period will continue to some extent; the moving average part considers the impact of sudden events such as storms on the position of the dune. Through the training of historical data for many years, the model can identify the seasonal regularity and long-term trend of the dune movement.

[0157] It should be noted that the generation of evolution trend data needs to consider multiple time scales. Short-term prediction focuses on changes in the next few months, mainly for risk assessment during construction period; medium-term prediction covers the next few years, for safety planning during operation period; long-term prediction focuses on the whole life cycle of the transmission line. The dune shape parameters at each time node include position coordinates, height distribution and slope change, which together depict the three-dimensional shape evolution process of the dune.

[0158] Specifically, the identification of potential threat points uses a spatial distance analysis method. The center line of the cross-ditch path represents the design direction of the transmission line, and the predicted sequence of dune positions shows the area that the dune may reach. The calculation of vertical distance considers the overall shape of the dune, not only the ridge line position, but also the extension range of the slope surface. When this distance is less than the safety threshold, it means that the dune may invade the line corridor, threatening the stability of the tower foundation or affecting the line clearance. The determination of the outer envelope line uses the convex hull algorithm to include all potential threat points, and then expands a safety buffer distance outward to form a conservative but reliable impact range. The selection of new cross-ditch points needs to meet multiple constraint conditions. The visibility condition ensures that there is no terrain obstruction between adjacent tower sites, which is verified by constructing a terrain profile. The terrain stability requires that the new position be away from unstable areas such as steep slopes and gullies, which is evaluated by analyzing the terrain slope and geological conditions. Under the premise of meeting these basic conditions, construction accessibility and economy are also considered.

[0159] It should be noted that the adjusted scheme not only avoids the threat of dune movement, but also maintains the rationality of the overall direction of the line, achieving a balance between safety and economy.

[0160] In the embodiment of the present application, the above step S600 includes the following sub-steps F1-F6;

[0161] In F1: acquire multi-temporal orthoimages, calculate the displacement of the dune ridge line and slope foot line feature points, and generate a dune movement rate distribution map;

[0162] In F2: according to the dune movement rate distribution map and elevation data, predict the evolution trend of the dune shape, and determine the influence range of the dune movement;

[0163] In F3: adjust the initial crossing point coordinates according to the sand dune moving influence range, recalculate the tower foundation candidate position;

[0164] In F4: obtain the geological survey data of the tower foundation candidate position, calculate the stability evaluation value, and mark the tower foundation to be reinforced;

[0165] In F5: establish a three-dimensional foundation model based on the soil layer distribution parameters of the tower foundation to be reinforced, calculate the stress distribution field and displacement field, determine the foundation depth increase value and anchor rod parameters;

[0166] In F6: generate a line layout map according to the foundation depth increase value and anchor rod parameters, set a buffer zone in the high-risk area of mudslides, and generate a final power grid planning scheme.

[0167] Specifically, the geological survey data and topographic parameters corresponding to the position of all tower foundations are obtained, the stability evaluation value of each tower foundation is obtained by calculating the foundation bearing capacity and anti-overturning stability coefficient, and if the stability evaluation value is lower than the preset stability threshold, the tower foundation is marked as a tower foundation to be reinforced, and the soil layer distribution parameters and underground water level elevation data of the tower foundation to be reinforced are extracted. A three-dimensional foundation model containing the physical and mechanical properties of the soil layer is established based on the soil layer distribution parameters and underground water level elevation data of the tower foundation to be reinforced, and the stress distribution field and displacement field under the current foundation condition are calculated using the finite element analysis method. The required foundation depth increase value and the corresponding concrete strength grade requirement are determined according to the difference between the maximum stress value and the maximum settlement value and the allowable value. The required uplift anchoring force of the tower foundation is calculated according to the foundation depth increase value and the concrete strength grade requirement, and the number, length and arrangement spacing parameters of the anchor rod are determined according to the anchoring force size and stratum conditions. After the increased foundation depth and anchoring parameters are used to re-perform finite element analysis to verify that the stability meets the requirements, a line layout map is generated that marks all tower foundation positions, line directions and reinforced tower foundation numbers. The line segment coordinates of the line that passes through the mudslide impact area are extracted from the line layout map, and the buffer zone range is delineated around the impact area according to the mudslide impact force decay law and safety protection requirements. A final power grid planning scheme containing spatial position information and engineering design information is generated by integrating the line layout map, the tower foundation reinforcement design parameter table, the buffer zone coordinate data and the line technical parameters.

[0168] Geological survey data is the basis for tower foundation stability evaluation, including the physical properties, mechanical parameters and underground water distribution of the soil layer. The foundation bearing capacity reflects the maximum compressive stress that the soil can withstand, which is obtained through standard penetration tests or static cone penetration tests, combined with the internal friction angle and cohesion of the soil layer. The anti-overturning stability coefficient considers the stability of the tower foundation under the action of horizontal forces such as wind load and conductor tension, and is obtained by calculating the ratio of resisting moment to overturning moment. When these indicators are lower than the safety threshold required by the design specification, it indicates that the tower foundation position is at risk of instability.

[0169] The establishment of a three-dimensional foundation model requires accurate reflection of the spatial distribution characteristics of the strata. Soil layer distribution parameters include the thickness, burial depth, and spatial variation of each layer of soil, which are obtained through the interpolation of drilling data. Groundwater level data reflect the influence of groundwater on soil strength, and the effective stress of saturated soil is lower than that of unsaturated soil, resulting in a corresponding reduction in bearing capacity. Finite element analysis discretizes the continuous foundation into a finite number of elements, each element is assigned corresponding material properties, and the stress and displacement distribution is obtained by solving the balance equation.

[0170] Specifically, the stress distribution field reveals the stress state inside the foundation. Under the action of tower load, stress spreads downward and outward from the foundation bottom, forming a stress bubble. The maximum stress usually occurs at the edge or a certain depth below the foundation, and if it exceeds the bearing capacity of the soil, it will lead to foundation failure. The displacement field reflects the deformation characteristics of the foundation, including vertical settlement and horizontal displacement. The maximum settlement is related to the size of the foundation, the size of the load, and the compressibility of the soil layer. By comparing the difference between the calculated value and the allowable value, the additional foundation depth required can be determined, and a deeper foundation can transfer the load to deeper soil with higher bearing capacity.

[0171] It should be noted that the selection of concrete strength grade not only meets the structural bearing requirements, but also considers the influence of environmental conditions. In corrosive soil or groundwater environments, high-strength concrete with better corrosion resistance is required. The calculation of anchoring force is based on the limit equilibrium principle to ensure that the tower foundation does not fail under the most unfavorable working conditions. The design parameters of the anchor rod include diameter, length, inclination angle, and spacing, which are interrelated. Longer anchor rods can penetrate stable strata to provide greater uplift resistance; a reasonable inclination angle can fully utilize the shear strength of the soil; and appropriate spacing avoids the reduction of bearing capacity caused by group anchor effect.

[0172] It should also be noted that the preset threshold values in this invention are set according to specific conditions. Stability verification uses updated model parameters to perform finite element analysis again. The reinforced foundation not only increases the burial depth, but also forms a composite foundation through the anchor rod system, significantly improving the overall stability. The generation of the line layout map integrates all the spatial information and attribute information of the tower foundations, and distinguishes between regular tower foundations and reinforced tower foundations through different symbols, providing intuitive guidance for construction management. The setting of the debris flow buffer zone is based on risk assessment and protection needs. The impact force attenuation law shows that the destructive power of debris flow decreases rapidly with distance, and setting a buffer zone of sufficient width on both sides of the main flow path can effectively protect the line safety. The final power grid planning scheme is a comprehensive document that not only includes technical design content, but also considers construction organization, operation and maintenance management, and other factors, achieving the unity of engineering safety, economy, and implementability.

[0173] The above is a schematic scheme of the power distribution network intelligent planning method of the embodiment. It should be noted that the technical scheme of the power distribution network intelligent planning system belongs to the same concept as the technical scheme of the power distribution network intelligent planning method described above. The technical details of the power distribution network intelligent planning system in the embodiment are not described in detail, and can be referred to the description of the technical scheme of the power distribution network intelligent planning method described above.

[0174] The power distribution network intelligent planning system in the embodiment comprises:

[0175] The terrain feature information extraction module is configured to obtain point cloud data, perform terrain feature information extraction on the point cloud data, and obtain a terrain data set.

[0176] The ground surface change information extraction module is configured to obtain remote sensing data, perform ground surface change information extraction on the remote sensing data, and obtain a ground surface change data set.

[0177] The terrain data fusion module is configured to fuse the terrain data set and the ground surface change data set, and obtain a terrain point cloud data set.

[0178] The model construction module is configured to construct a first terrain model according to the terrain point cloud data set and analyze a debris flow erosion path of a gully edge collapse area.

[0179] The path planning module is configured to obtain unmanned aerial vehicle monitoring data, simulate a line crossing gully path in combination with the first terrain model, optimize a crossing gully point and a tower foundation coordinate, and obtain a power grid planning scheme.

[0180] The stability evaluation module is configured to evaluate the stability of the power grid planning scheme, adjust a tower foundation reinforcement parameter, and obtain a final power grid planning scheme.

[0181] The embodiment also provides a computer device suitable for power distribution network intelligent planning, comprising:

[0182] The memory is configured to store computer executable instructions, and the processor is configured to execute the computer executable instructions to implement the power distribution network intelligent planning method proposed in the above embodiment.

[0183] The embodiment also provides a storage medium having a computer program stored thereon, and the program is executed by a processor to implement the power distribution network intelligent planning method proposed in the above embodiment.

[0184] The storage medium proposed in the embodiment and the power distribution network intelligent planning method proposed in the above embodiment belong to the same inventive concept. The technical details not described in detail in the embodiment can be referred to the above embodiment, and the embodiment has the same beneficial effects as the above embodiment.

[0185] Through the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented with the help of software and necessary general-purpose hardware, and of course can also be implemented by hardware. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as a computer's floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for enabling a computing device (which can be a personal computer, server, or network device, etc.) to execute the methods of various embodiments of the present invention.

[0186] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A distribution network intelligent planning method, characterized in that: include: Acquiring point cloud data, extracting terrain feature information from the point cloud data, and obtaining a terrain dataset; Acquiring remote sensing data, and extracting surface change information from the remote sensing data to obtain a surface change dataset; fusing the terrain dataset and the surface change dataset to obtain a terrain point cloud dataset; Constructing a first terrain model based on the terrain point cloud dataset and analyzing the debris flow scouring path in the gully edge collapse area; Acquire UAV monitoring data and combine it with the first terrain model to simulate the line crossing path, optimize the crossing points and tower base coordinates, and obtain a power grid planning solution; The stability of the power grid planning scheme is evaluated, and the tower base reinforcement parameters are adjusted to obtain the final power grid planning scheme.

2. A distribution network intelligent planning method according to claim 1, characterized in that: The terrain dataset includes: Calculate the reflectivity value of the sampling point based on the point cloud data; Performing cluster analysis on the reflectivity values ​​using the first clustering method to obtain abnormal reflectivity areas and extracting the point cloud spatial coordinates of the abnormal reflectivity areas; Perform the first elevation interpolation on the spatial coordinates of the point cloud to generate a terrain change distribution map, extract the terrain change boundary line, and determine the terrain change area; Extract visible light image data from areas with terrain changes, calculate texture feature parameters and color values, and determine the exposed areas of the wet layer based on the color values ​​and preset dry sand reference color values; According to the judgment result, the first dimensionality reduction method is used to reduce the dimensionality of the texture feature parameters and color values ​​of the exposed area of ​​the wet layer to generate a feature vector; The terrain dataset is obtained according to the spatial distribution of the feature vectors.

3. A distribution network intelligent planning method according to claim 2, characterized in that: Land surface change datasets include: Extract spectral values ​​from remote sensing data, calculate the distribution map of land cover types, and determine the dynamic change areas of the surface based on spectral value differences and spatial position offsets; The first segmentation method is used to separate the exposed sand distribution area from the surface dynamic change area, extract the dune ridge line, calculate the ridge line position offset, generate the dune movement vector, and obtain the dune wind erosion migration area; Extract slope data and elevation change information from the dune wind erosion migration area, calculate the potential confluence area, and comprehensively determine the impact range of gully debris flows; A surface change dataset was generated based on multi-temporal elevation data of the gully debris flow impact area and the dune wind erosion migration area.

4. A distribution network intelligent planning method according to claim 2 or 3, characterized in that: The terrain point cloud dataset includes: Extracting the same-name registration point pairs from the overlapping area of ​​the terrain dataset and the surface change dataset by the first matching method, and calculating the spatial transformation parameters; The terrain dataset and the surface change dataset are converted into coordinates according to the spatial transformation parameters to obtain the registered terrain dataset; Evaluate the distribution of elevation differences between the registered terrain dataset and the surface change dataset in the overlapping area, and mark the elevation discontinuity areas; A buffer zone is established in the elevation discontinuous area, and the drone image point cloud and satellite remote sensing point cloud are fused to generate a terrain point cloud dataset.

5. A distribution network intelligent planning method according to claim 4, characterized in that: Constructing the first terrain model includes: Generate irregular triangulated terrain surface based on terrain point cloud dataset, calculate slope and aspect values, and identify potential debris flow occurrence areas; Generate a digital elevation model grid for the potential debris flow area, determine the flow direction and cumulative runoff, calculate the scour depth, and generate path centerline data and impact range polygon data; A first terrain model is constructed according to the path centerline data and the influence range polygon data.

6. A distribution network intelligent planning method according to claim 5, characterized in that: Optimizing the cross-ditch point and tower base coordinates includes: Extract gully distribution data and elevation information from the first terrain model, calculate the initial route direction, and determine the coordinates of the initial gully crossing point; Based on the changes in the sand dunes monitored by drones, the coordinates of the initial cross-ditch point were adjusted to the stable area, and the candidate location of the tower base was determined; Construct line corridors from candidate tower base locations, calculate comprehensive risk indexes, search for optimized paths, and generate power grid planning solutions that include tower base geographic coordinates and line centerlines.

7. A distribution network intelligent planning method according to claim 6, characterized in that: The final grid planning solution includes: Obtain multi-temporal orthophotos, calculate the displacement of characteristic points of dune ridges and toe lines, and generate a dune movement rate distribution map; Based on the dune movement rate distribution map and elevation data, the dune morphological evolution trend is predicted and the impact range of dune movement is determined; Adjust the initial cross-ditch point coordinates according to the impact range of sand dune movement and recalculate the candidate tower base location; Obtain geological survey data for candidate tower foundation locations, calculate stability assessment values, and mark tower foundations to be reinforced; A three-dimensional foundation model is established based on the soil distribution parameters of the tower foundation to be reinforced, the stress distribution field and displacement field are calculated, and the foundation depth increase and anchor parameters are determined; Generate a line layout diagram based on the foundation depth increase and anchor parameters, set a buffer zone for high-risk debris flow areas, and generate the final power grid planning scheme.

8. A distribution network intelligent planning system, applying a distribution network intelligent planning method according to any one of claims 1 to 7, characterized in that: include: A terrain feature information extraction module is used to obtain point cloud data, extract terrain feature information from the point cloud data, and obtain a terrain data set; A surface change information extraction module is used to obtain remote sensing data, extract surface change information from the remote sensing data, and obtain a surface change data set; A terrain data fusion module, configured to fuse the terrain dataset and the surface change dataset to obtain a terrain point cloud dataset; A model building module is used to build a first terrain model based on the terrain point cloud data set and by analyzing the debris flow scouring path in the gully edge collapse area; A path planning module is used to obtain UAV monitoring data and combine it with the first terrain model to simulate the line crossing trench path, optimize the trench crossing points and tower base coordinates, and obtain a power grid planning solution; The stability evaluation module is used to evaluate the stability of the power grid planning scheme, adjust the tower base reinforcement parameters, and obtain the final power grid planning scheme.

9. A computer device, characterized in that: include: A memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of a distribution network intelligent planning method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that A computer program is stored thereon, and when the computer program is executed by a processor, the steps of a distribution network intelligent planning method according to any one of claims 1 to 7 are implemented.

Images