Low-altitude unmanned aerial vehicle route safety risk assessment method and system based on multi-source geographic information

Abstract

The invention discloses a low-altitude unmanned aerial vehicle route safety risk assessment method and system based on multi-source geographic information, relates to the field of unmanned aerial vehicle route assessment, solves the problem of insufficient intelligence of existing unmanned aerial vehicle route planning, and comprises the following steps: S1, obtaining a regional map and geographic information data, and constructing a grid map; s2, acquiring multi-dimensional data of the grid units according to the grid map, and calculating a risk cost level to obtain a risk value list; s3, carrying out weight setting on the risk dimension, and obtaining a dynamic weight matrix in combination with an adjustment coefficient; s4, calculating a comprehensive risk value according to the risk value list in combination with the dynamic weight matrix; s5, the flight route of the unmanned aerial vehicle is obtained, and a comprehensive safety score is calculated according to the flight route; performing dynamic adjustment according to the comprehensive safety score; the intelligent level of the low-altitude unmanned aerial vehicle in the course planning stage can be effectively improved, and the course safety risk can be dynamically evaluated and adjusted.

Description

Technical Field

[0001] This invention belongs to the field of unmanned aerial vehicle (UAV) flight path assessment, and more specifically relates to low-altitude UAV flight path planning, specifically a method and system for assessing the safety risks of low-altitude UAV flight paths based on multi-source geographic information. Background Technology

[0002] The existing low-altitude drones have the following specific shortcomings when conducting flight path safety risk assessments:

[0003] 1. Existing data sources for safety risk assessment of low-altitude UAV routes are limited, typically involving the analysis of physical models in space to determine the collision status of the UAV and then assessing its safety based on the collision status. This approach only considers physical collision risks and ignores social and environmental risks such as population safety, noise pollution, and ecological disturbance, resulting in poor assessment results.

[0004] 2. The existing assessment model for safety risks of low-altitude UAV routes is fixed and cannot dynamically adjust the risk assessment standards according to the type of mission (such as emergency medical transportation and regular express delivery). It is difficult to meet the needs of low-altitude UAVs for different purposes, and the assessment has low applicability.

[0005] 3. Existing safety risk assessments for low-altitude UAV routes can only passively respond to obstacles during flight, and cannot conduct forward-looking and comprehensive safety assessments during the mission planning stage; the accuracy of risk assessments is low, making it difficult to avoid risks in a timely manner.

[0006] To this end, we propose a method and system for assessing the safety risks of low-altitude unmanned aerial vehicle (UAV) routes based on multi-source geographic information. Summary of the Invention

[0007] To address the shortcomings of existing technologies, the purpose of this invention is to provide a method and system for assessing the flight safety risks of low-altitude unmanned aerial vehicles (UAVs) based on multi-source geographic information. This invention aims to improve the accuracy of low-altitude UAV flight safety risk assessment and ensure the flight safety of UAVs.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: a method for assessing the safety risks of low-altitude unmanned aerial vehicle (UAV) routes based on multi-source geographic information, the specific working process of each step of which is as follows:

[0009] Step S1: Obtain the low-altitude flight area of ​​the UAV, obtain the area map, collect geographic information data based on the area map, and perform rasterization processing on the area map based on the collected geographic information data to construct a raster map.

[0010] Step S2: Acquire multi-dimensional data of raster cells based on the raster map, statistically analyze the dimensions of the multi-dimensional data to obtain multiple risk dimensions; calculate the risk cost surface for each risk dimension to obtain the risk value; statistically analyze the risk values ​​to obtain a risk value list;

[0011] Step S3: Set weights for the risk dimensions to obtain the AHP weight matrix and the EWM weight matrix; combine the AHP weight matrix and the EWM weight matrix to obtain the basic weight matrix; obtain the preset task-specific weight library; extract the adjustment coefficients for different risk dimensions based on the task-specific weight library; calculate the dynamic weight matrix by combining the basic weight matrix.

[0012] Step S4: Based on the risk value list and combined with the dynamic weight matrix, the risk values ​​of each risk cost surface are weighted and superimposed to calculate the risk value of the comprehensive risk cost surface, thus obtaining the comprehensive risk value;

[0013] Step S5: Obtain the flight path of the UAV, extract the corresponding grid cells based on the flight path, sum the comprehensive risk values ​​of the grid cells, and calculate the comprehensive safety score; make dynamic adjustments based on the comprehensive safety score.

[0014] Furthermore, the specific steps of step S1 are as follows:

[0015] Step S11: Collect urban 3D model data, topographic data, population and land use data, and ecological environment data;

[0016] Geographic information data consists of urban 3D model data, topographic data, population and land use data, and ecological environment data.

[0017] Step S12: Analyze the regional map based on geographic information data, divide the regional map into internal regions, divide the regional map into terrain regions based on urban 3D model data and topographic data, divide the regional map into population distribution regions based on population and land use data, and divide the regional map into ecological regions based on ecological environment data; and rasterize the regional map based on the internal region division results to construct a raster map.

[0018] Furthermore, the specific steps of step S12 are as follows:

[0019] Step S121: Based on the city's 3D model data and terrain data, perform vector mapping on the regional map, divide different terrains using curves, obtain the flight altitude of the low-altitude UAV, extract buildings above the flight altitude from the city's 3D model, and represent the buildings using data points.

[0020] Based on the data points and the dividing curves, the grid size is set, and the regional map is rasterized to obtain the terrain grid size;

[0021] Based on population and land use data, the population distribution is obtained, and based on the population distribution, the grid size is set to obtain the population grid size.

[0022] Based on ecological and environmental data, ecological and environmental protection areas are extracted from the regional map, and grid settings are made according to the ecological and environmental protection areas to obtain the environmental grid size;

[0023] Step S122: Statistically analyze the terrain raster size, population raster size, and environment raster size; extract the common divisor of the terrain raster size, population raster size, and environment raster size; sort the common divisors in descending order and compile a list of common divisors; iterate through the common divisors in the list; use the common divisors as preset raster sizes to rasterize the regional map; verify the rasterized regional map using geographic information data to determine the optimal raster size; and construct a raster map based on the optimal raster size.

[0024] Furthermore, the specific steps of step S2 are as follows:

[0025] Step S21: Count the number of grid cells on the grid map to obtain the number of grid cells xs×ys; acquire the multi-dimensional data of each grid cell, extract the number of dimensions of the multi-dimensional data, denoted as is, and construct the risk dimension based on the number of dimensions; acquire the data of each grid cell and each risk dimension based on the number of grid cells and the risk dimension, denoted as the dimension acquisition data wcj(x, y, i);

[0026] Step S22: Extract data from different risk dimensions based on the collected data, calculate the risk cost surface for each risk dimension, and obtain the risk value.

[0027] Furthermore, the specific steps of step S22 are as follows:

[0028] Step S221: Acquire the height dimension data based on the dimension acquisition data to obtain the building height H of each grid cell. obs (x, y); H obs (x, y) represents the building height of the raster cell at (x, y). The highest building height within the raster map is extracted and denoted as H. max The safe flight altitude of the drone is obtained and denoted as H. safe Based on the building height H of each grid cell obs Using (x, y), the highest building height, and the safe height, the geographical collision risk cost surface is calculated to obtain the collision risk value C.geo (x, y);

[0029]

[0030] Step S222: Acquire population dimension data based on dimensional data collection to obtain the population density P of each grid cell. density (x, y) and land use risk coefficient L risk (x, y); P density (x, y) represents the population density of the raster cell at (x, y), L risk (x, y) represents the land use risk coefficient of the raster cell at (x, y); based on the population density of each raster cell, the maximum population density within the raster map is extracted to obtain P. max Based on the population density and land use risk coefficient of each grid cell, combined with the maximum population density, the population security risk cost surface is calculated to obtain the population security risk value C. pop (x, y);

[0031]

[0032] Step S223: Acquire environmental dimension data based on dimensional data collection to obtain the distance d(x, y) between each grid cell and the nearest ecological protection zone, and obtain the buffer distance of the ecological protection zone, denoted as D; calculate the ecological disturbance risk cost surface based on the distance and buffer distance between each grid cell and the nearest ecological protection zone to obtain the ecological disturbance risk value C. eco (x, y);

[0033]

[0034] Furthermore, the specific steps of step S3 are as follows:

[0035] Step S31: Evaluate each risk dimension through expert review, analyze the evaluation results using the Analytic Hierarchy Process (AHP) to obtain a risk score for each risk dimension; integrate the risk scores for each risk dimension, calculate the weight of each risk dimension, and statistically analyze the weights of each risk dimension to obtain the AHP weight matrix, denoted as W. AHP ;

[0036] Step S32: Normalize the data for each risk dimension. Based on the normalized data, calculate the dispersion of the risk dimensions using the entropy weight method. Calculate the proportions of the dispersion of different risk dimensions and set weights for each risk dimension to obtain the EWM weight matrix, denoted as W. EWM ;

[0037] Step S33: Establish the λ parameter. Based on the λ parameter, construct the combined basic weight matrix by combining the AHP weight matrix and the EWM weight matrix. Calculate the combined basic weight W by using the elements of the AHP weight matrix and the EWM weight matrix to obtain the combined basic weight. base(i) ;

[0038] W base(i) =λ×W AHP(i) +(1-λ)×W EWM(i) ;

[0039] Among them: W AHP(i) W represents the i-th element in the AHP weight matrix. EWM(i) This represents the i-th element in the EWM weight matrix;

[0040] A dynamic weight matrix of the same size is constructed based on the combined basic weight matrix. By combining the basic weights with adjustment coefficients for different risk dimensions, the elements in the dynamic weight matrix are calculated to obtain the dynamic weight W. final(i) ;

[0041] W final(i) =W base(i) ×a i ;

[0042] Where: a i This represents the adjustment coefficient for the i-th risk dimension.

[0043] Furthermore, the specific steps of step S31 are as follows:

[0044] Step S311: Arrange all risk dimensions in a matrix, compare each risk dimension pairwise according to the matrix arrangement, and quantify the comparison results by expert evaluation to obtain quantified values;

[0045] The quantified values ​​are statistically analyzed to obtain the judgment matrix pdj; feature vectors are extracted based on the judgment matrix to obtain the feature vector matrix tjz; the judgment matrix is ​​extracted row by row to obtain the judgment row matrix pdh(i); the feature vector matrix and the judgment row matrix are calculated to obtain the risk score fpf(i).

[0046]

[0047] Among them: tjz i’ Let pdh(i) represent the i'-th element in the eigenvector matrix. i’ This represents the i'-th element in the judgment row matrix of the i-th row;

[0048] Step S312: Integrate the risk scores of each risk dimension, calculate the ratio between the integrated risk score and the risk scores of each risk dimension, and obtain the weight of each risk dimension, denoted as the analysis weight fqz(i).

[0049]

[0050] By statistically analyzing the weights fqz(i), we obtain the AHP weight matrix, denoted as W. AHP W AHP =[fqz(1), fqz(2),..., fqz(is)].

[0051] Furthermore, the specific steps of step S32 are as follows:

[0052] Step S321: Traverse the raster map, normalize each risk dimension of each raster cell to obtain standardized values, and denote the standardized values ​​as p. ij ;p ij Let represent the standardized value of the j-th grid cell in the i-th risk dimension; the discrete value lsz is obtained by calculating the dispersion of the risk dimension based on the standardized value. i ;

[0053]

[0054] Where: xs×ys is the number of grid cells, and k is a constant;

[0055] Step S322: Calculate the weight of each risk dimension based on the discrete values ​​to obtain the entropy weight sqz(i);

[0056]

[0057] By performing statistics on the entropy weights sqz(i), the EWM weight matrix is ​​obtained, denoted as W. EWM W EWM =[sqz(1),sqz(2),…,sqz(is)].

[0058] Furthermore, the specific steps of step S4 are as follows:

[0059] The risk value is retrieved from the risk value list and denoted as C. i (x, y), C i (x, y) represents the risk value of the i-th risk dimension at grid (x, y); the dynamic weights W in the dynamic weight matrix final(i) The dynamic weights are extracted and combined with the risk value to calculate the comprehensive risk value C. total (x, y);

[0060]

[0061] A low-altitude unmanned aerial vehicle (UAV) flight path safety risk assessment system based on multi-source geographic information. The assessment system includes:

[0062] Data acquisition module: acquires the low-altitude flight area of ​​the UAV, obtains a regional map, collects geographic information data based on the regional map, and performs rasterization processing on the regional map based on the collected geographic information data to construct a raster map;

[0063] Risk Calculation Module: Acquires multi-dimensional data of grid cells based on the grid map, performs statistical analysis on the dimensions of the multi-dimensional data to obtain multiple risk dimensions; calculates the risk cost surface for each risk dimension to obtain a risk value; and performs statistical analysis on the risk values ​​to obtain a list of risk values.

[0064] Weight setting module: Set weights for risk dimensions to obtain AHP weight matrix and EWM weight matrix; combine AHP weight matrix and EWM weight matrix to obtain basic weight matrix; obtain preset task-specific weight library, extract adjustment coefficients for different risk dimensions based on task-specific weight library, and calculate dynamic weight matrix by combining with basic weight matrix.

[0065] Comprehensive Analysis Module: Based on the risk value list and combined with the dynamic weight matrix, the risk values ​​of each risk cost surface are weighted and superimposed to calculate the risk value of the comprehensive risk cost surface, thus obtaining the comprehensive risk value;

[0066] Flight path assessment module: Obtain the flight path of the UAV, extract the corresponding grid cells based on the flight path, sum the comprehensive risk values ​​of the grid cells, calculate the comprehensive safety score, and make dynamic adjustments based on the comprehensive safety score.

[0067] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0068] 1. This invention conducts risk analysis on UAV flight areas using multi-dimensional data, including urban 3D model data, terrain and landform data, population and land use data, and ecological environment data; ensuring the comprehensiveness of risk analysis and improving its accuracy.

[0069] 2. Multi-dimensional data is weighted in multiple ways to improve the accuracy of weight settings. The weight settings are used to control the multi-dimensional data and meet the specificity of different risks. At the same time, a task-specific weight library is introduced to adjust the risk weights in real time under different task conditions, dynamically adjust the risk assessment standards, improve the application of UAVs, and enhance the applicability of UAVs.

[0070] 3. By processing the UAV flight area into a grid, the risk of each grid cell can be analyzed independently, improving the accuracy of risk analysis; by statistically analyzing the grid cells traversed by the UAV flight path, a safety assessment of the UAV flight path can be conducted, enabling a forward-looking and comprehensive safety assessment during the mission planning phase; thus enhancing the safety of UAV flight paths. Attached Figure Description

[0071] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0072] Figure 1 This is a schematic diagram of the method of the present invention;

[0073] Figure 2 This is a schematic diagram of the grid processing of the present invention;

[0074] Figure 3 This is a schematic diagram illustrating the weight settings of the present invention; Detailed Implementation

[0075] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0076] Example 1

[0077] Please see Figure 1 The low-altitude UAV flight path safety risk assessment method based on multi-source geographic information of the present invention belongs to the field of UAV flight path assessment, and provides the following:

[0078] Step S1: Obtain the low-altitude flight area of ​​the UAV, obtain the area map, collect geographic information data based on the area map, and perform rasterization processing on the area map based on the collected geographic information data to construct a raster map.

[0079] It should be noted that geographic information data includes topographic data (such as digital elevation models, DEM), urban 3D model data (such as lidar point cloud data), population density and land use data, and ecological environment data.

[0080] It should be noted that: regional maps refer to defining the data collection area by determining the low-altitude flight area of ​​drones; they only contain spatial range information; raster maps refer to dividing a regional map into multiple raster grids using geographic information data.

[0081] Step S11: Generate a 3D city model using oblique photogrammetry or LiDAR, and collect 3D city model data (including the spatial distribution and height information of fixed obstacles such as buildings and bridges);

[0082] The region's topographic data (including elevation, obstacle distribution, and topographic change information) is collected through digital elevation modeling, LiDAR, and urban 3D models.

[0083] By using remote sensing imagery, land use data, and nighttime light remote sensing methods, we can extract land use types (such as residential areas, commercial areas, and industrial areas) and population distribution within the region to obtain population and land use data.

[0084] Based on the ecological protection information released by the state, spatial distribution information of ecologically sensitive areas (ecological protection areas, bird migration routes, and water source protection areas) within the region is extracted to obtain ecological and environmental data;

[0085] Geographic information data consists of urban 3D model data, topographic data, population and land use data, and ecological environment data.

[0086] Please see Figure 2 Step S12: Analyze the regional map based on geographic information data, divide the regional map into internal regions, divide the regional map into terrain regions based on urban 3D model data and topographic data, divide the regional map into population distribution regions based on population and land use data, and divide the regional map into ecological regions based on ecological environment data; and rasterize the regional map based on the internal region division results to construct a raster map.

[0087] Step S121: Based on the city's 3D model data and terrain data, perform vector mapping on the regional map, divide different terrains using curves, obtain the flight altitude of the low-altitude UAV, extract buildings above the flight altitude from the city's 3D model, and represent the buildings using data points.

[0088] Based on the data points and the dividing curves, the grid size is set, and the regional map is rasterized to obtain the terrain grid size;

[0089] Based on population and land use data, the population distribution is obtained, and based on the population distribution, the grid size is set to obtain the population grid size.

[0090] Based on ecological and environmental data, ecological and environmental protection areas are extracted from the regional map, and grid settings are made according to the ecological and environmental protection areas to obtain the environmental grid size;

[0091] It should be noted that: grid setting refers to separating the corresponding positions of data on the regional map according to the distribution of data. Based on the separation, a limit value is set to divide the entire regional map. For example, if there are two population distribution clusters on the regional map, the midpoint between the two population distribution clusters is used as the separation boundary to separate the population distribution clusters and set the grid.

[0092] Step S122: Statistically analyze the terrain raster size, population raster size, and environment raster size; extract the common divisor of the terrain raster size, population raster size, and environment raster size; sort the common divisors in descending order and compile a list of common divisors; iterate through the common divisors in the list; use the common divisors as preset raster sizes to rasterize the regional map; verify the rasterized regional map using geographic information data to determine the optimal raster size; and construct a raster map based on the optimal raster size.

[0093] Step S2: Acquire multi-dimensional data of raster cells based on the raster map, statistically analyze the dimensions of the multi-dimensional data to obtain multiple risk dimensions; calculate the risk cost surface for each risk dimension to obtain the risk value; statistically analyze the risk values ​​to obtain a risk value list;

[0094] Step S21: Count the number of grid cells on the grid map to obtain the number of grid cells xs×ys; acquire the multi-dimensional data of each grid cell, extract the number of dimensions of the multi-dimensional data, denoted as is, and construct the risk dimension based on the number of dimensions; acquire the data of each risk dimension for each grid cell based on the number of grid cells and the risk dimension, denoted as the dimension acquisition data wcj(x, y, i); wcj(x, y, i) represents the acquisition data of the i-th risk dimension at grid cell (x, y), where x∈[1, xs], y∈[1, ys], and i∈[1, is].

[0095] Step S22: Extract data from different risk dimensions based on the collected data, calculate the risk cost surface for each risk dimension, and obtain the risk value;

[0096] Step S221: Acquire the height dimension data based on the dimension acquisition data to obtain the building height H of each grid cell. obs (x, y); H obs (x, y) represents the building height of the raster cell at (x, y). The highest building height within the raster map is extracted and denoted as H. max The safe flight altitude of the drone is obtained and denoted as H. safe Based on the building height H of each grid cell obsUsing (x, y), the highest building height, and the safe height, the geographical collision risk cost surface is calculated to obtain the collision risk value C. geo (x, y);

[0097]

[0098] Step S222: Acquire population dimension data based on dimensional data collection to obtain the population density P of each grid cell. density (x, y) and land use risk coefficient L risk (x, y); P density (x, y) represents the population density of the raster cell at (x, y), L risk (x, y) represents the land use risk coefficient of the raster cell at (x, y); based on the population density of each raster cell, the maximum population density within the raster map is extracted to obtain P. max Based on the population density and land use risk coefficient of each grid cell, combined with the maximum population density, the population security risk cost surface is calculated to obtain the population security risk value C. pop (x, y);

[0099]

[0100] Step S223: Acquire environmental dimension data based on dimensional data collection to obtain the distance d(x, y) between each grid cell and the nearest ecological protection zone, and obtain the buffer distance of the ecological protection zone, denoted as D; calculate the ecological disturbance risk cost surface based on the distance and buffer distance between each grid cell and the nearest ecological protection zone to obtain the ecological disturbance risk value C. eco (x, y);

[0101]

[0102] It should be noted that the calculation of the ecological disturbance risk value is based on a transformation of the Sigmoid function to quantify the potential impact or cost of distance on the ecosystem; for example, the closer to the ecological reserve, the greater the ecological disturbance risk value; conversely, the farther away, the smaller the ecological disturbance risk value.

[0103] It should be noted that by analyzing risk data from multiple different dimensions, the accuracy of drone flight risk assessment can be improved, thus ensuring drone flight safety.

[0104] Step S3: Set weights for the risk dimensions using the Analytic Hierarchy Process (AHP) to obtain the AHP weight matrix; calculate the dispersion of the risk dimensions using the entropy weight method; set weights based on the dispersion of the risk dimensions to obtain the EWM weight matrix; combine the AHP weight matrix and the EWM weight matrix to obtain the basic weight matrix; obtain a preset task-specific weight library; extract adjustment coefficients for different risk dimensions based on the task-specific weight library; and calculate the dynamic weight matrix by combining the dynamic weight matrix with the basic weight matrix.

[0105] Please see Figure 3 Step S31: Evaluate each risk dimension through expert assessment, analyze the assessment results using the Analytic Hierarchy Process (AHP) to obtain a risk score for each risk dimension; integrate the risk scores for each risk dimension, calculate the weight of each risk dimension, and statistically analyze the weights of each risk dimension to obtain the AHP weight matrix, denoted as W. AHP ;

[0106] Step S311: Arrange all risk dimensions in a matrix, compare each risk dimension pairwise according to the matrix arrangement, and quantify the comparison results by expert evaluation to obtain quantified values;

[0107] It should be noted that the quantification mentioned refers to assigning values ​​to the comparison results based on a scaling table in the Analytic Hierarchy Process (AHP) and expert judgment. The scaling table is as follows:

[0108]

[0109]

[0110] The quantified values ​​are statistically analyzed to obtain the judgment matrix pdj; feature vectors are extracted based on the judgment matrix to obtain the feature vector matrix tjz; the judgment matrix is ​​extracted row by row to obtain the judgment row matrix pdh(i); the feature vector matrix and the judgment row matrix are calculated to obtain the risk score fpf(i).

[0111]

[0112] Among them: tjz i’ Let pdh(i) represent the i'-th element in the eigenvector matrix. i’ This represents the i'th element in the judgment row matrix of the i-th row.

[0113] Step S312: Integrate the risk scores of each risk dimension, calculate the ratio between the integrated risk score and the risk scores of each risk dimension, and obtain the weight of each risk dimension, denoted as the analysis weight fqz(i).

[0114]

[0115] By statistically analyzing the weights fqz(i), we obtain the AHP weight matrix, denoted as W. AHP W AHP =[fqz(1), fqz(2),..., fqz(is)].

[0116] It should be noted that the analytic hierarchy process (AHP) quantifies subjective judgments, reduces the error of subjective factors, and improves the accuracy of weighted analysis.

[0117] Step S32: Normalize the data for each risk dimension. Based on the normalized data, calculate the dispersion of the risk dimensions using the entropy weight method. Calculate the proportions of the dispersion of different risk dimensions and set weights for each risk dimension to obtain the EWM weight matrix, denoted as W. EWM ;

[0118] Step S321: Traverse the raster map, normalize each risk dimension of each raster cell to obtain standardized values, and denote the standardized values ​​as p. ij ;p ij Let represent the standardized value of the j-th grid cell in the i-th risk dimension; the discrete value lsz is obtained by calculating the dispersion of the risk dimension based on the standardized value. i ;

[0119]

[0120] Where: xs×ys is the number of grid cells, and k is a constant.

[0121] Step S322: Calculate the weight of each risk dimension based on the discrete values ​​to obtain the entropy weight sqz(i);

[0122]

[0123] By performing statistics on the entropy weights sqz(i), the EWM weight matrix is ​​obtained, denoted as W. EWM W EWM =[sqz(1),sqz(2),…,sqz(is)].

[0124] Step S33: Establish the λ parameter. Based on the λ parameter, construct the combined basic weight matrix by combining the AHP weight matrix and the EWM weight matrix. Calculate the combined basic weight W by using the elements of the AHP weight matrix and the EWM weight matrix to obtain the combined basic weight. base(i) ;

[0125] W base(i) =λ×W AHP(i) +(1-λ)×WEWM(i) ;

[0126] Among them: W AHP(i) W represents the i-th element in the AHP weight matrix. EWM(i) This represents the i-th element in the EWM weight matrix;

[0127] It should be noted that the λ parameter is set to an initial value based on the task type. For example, for emergency medical tasks, the λ value can be set to 0.7, favoring the AHP method; for routine logistics tasks, the λ value can be set to 0.5, maintaining a balance between AHP and EWM.

[0128] A dynamic weight matrix of the same size is constructed based on the combined basic weight matrix. By combining the basic weights with adjustment coefficients for different risk dimensions, the elements in the dynamic weight matrix are calculated to obtain the dynamic weight W. final(i) ;

[0129] W final(i) =W base(i) ×a i ;

[0130] Where: a i This represents the adjustment coefficient for the i-th risk dimension;

[0131] It should be noted that by adjusting the coefficients to perform specialized processing on drone flight missions, the application breadth of drone flights can be improved, avoiding the limitation of drone applications caused by fixed data.

[0132] Step S4: Based on the risk value list and combined with the dynamic weight matrix, the risk values ​​of each risk cost surface are weighted and superimposed to calculate the risk value of the comprehensive risk cost surface, thus obtaining the comprehensive risk value;

[0133] Step S41: Obtain the risk value from the risk value list, denoted as C. i (x, y), C i (x, y) represents the risk value of the i-th risk dimension at grid (x, y); the dynamic weights W in the dynamic weight matrix final(i) The dynamic weights are extracted and combined with the risk value to calculate the comprehensive risk value C. total (x, y);

[0134]

[0135] Step S5: Obtain the flight path of the UAV, extract the corresponding grid cells based on the flight path, sum the comprehensive risk values ​​of the grid cells, and calculate the comprehensive safety score; make dynamic adjustments based on the comprehensive safety score.

[0136] Step S51: Extract the grid cells corresponding to the flight path, count the number of grid cells corresponding to the flight path to obtain the number of cells ds; record the position of each upper cell as (x, y) according to the number of grid cells. d ; indicates the position of the d-th raster cell on the raster map;

[0137] The overall safety score S is obtained by summing the comprehensive risk values ​​of the grid cells.

[0138]

[0139] Step S52: Obtain the threshold for the comprehensive safety score according to industry standards, denoted as S. thresh If the overall safety score is lower than S thresh This indicates that the flight path is a high-risk path; the dimensional data of the corresponding grid cells of the path are collected, and the main sources of risk in the high-risk section are analyzed (e.g., whether the risk is too high due to densely populated areas, ecological protection areas or geographical obstacles); the flight path is adjusted through manual intervention.

[0140] Example 2

[0141] The low-altitude UAV flight route safety risk assessment system based on multi-source geographic information includes: a data acquisition module, a risk calculation module, a weight setting module, a comprehensive analysis module, and a flight route assessment module.

[0142] Data acquisition module: acquires the low-altitude flight area of ​​the UAV, obtains a regional map, collects geographic information data based on the regional map, and performs rasterization processing on the regional map based on the collected geographic information data to construct a raster map;

[0143] Risk Calculation Module: Acquires multi-dimensional data of grid cells based on the grid map, performs statistical analysis on the dimensions of the multi-dimensional data to obtain multiple risk dimensions; calculates the risk cost surface for each risk dimension to obtain a risk value; and performs statistical analysis on the risk values ​​to obtain a list of risk values.

[0144] Weight setting module: Set weights for risk dimensions to obtain AHP weight matrix and EWM weight matrix; combine AHP weight matrix and EWM weight matrix to obtain basic weight matrix; obtain preset task-specific weight library, extract adjustment coefficients for different risk dimensions based on task-specific weight library, and calculate dynamic weight matrix by combining with basic weight matrix.

[0145] Comprehensive Analysis Module: Based on the risk value list and combined with the dynamic weight matrix, the risk values ​​of each risk cost surface are weighted and superimposed to calculate the risk value of the comprehensive risk cost surface, thus obtaining the comprehensive risk value;

[0146] Flight path assessment module: Obtain the flight path of the UAV, extract the corresponding grid cells based on the flight path, sum the comprehensive risk values ​​of the grid cells, calculate the comprehensive safety score, and make dynamic adjustments based on the comprehensive safety score.

[0147] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for assessing the safety risks of low-altitude unmanned aerial vehicle (UAV) routes based on multi-source geographic information, characterized in that, include: Step S1: Obtain the low-altitude flight area of ​​the UAV, obtain the area map, collect geographic information data based on the area map, and perform rasterization processing on the area map based on the collected geographic information data to construct a raster map. Step S2: Obtain multi-dimensional data of the grid cells based on the grid map, and statistically analyze the dimensions of the multi-dimensional data to obtain multiple risk dimensions; The risk value is obtained by calculating the risk cost surface for each risk dimension. The risk values ​​are statistically analyzed to obtain a list of risk values; Step S3: Set weights for the risk dimensions to obtain the AHP weight matrix and the EWM weight matrix; combine the AHP weight matrix and the EWM weight matrix to obtain the basic weight matrix; Obtain a preset task-specific weight library, extract adjustment coefficients for different risk dimensions based on the task-specific weight library, and calculate the dynamic weight matrix by combining it with the basic weight matrix. Step S4: Based on the risk value list and combined with the dynamic weight matrix, the risk values ​​of each risk cost surface are weighted and superimposed to calculate the risk value of the comprehensive risk cost surface, thus obtaining the comprehensive risk value; Step S5: Obtain the flight path of the UAV, extract the corresponding grid cells based on the flight path, sum the comprehensive risk values ​​of the grid cells, and calculate the comprehensive safety score; make dynamic adjustments based on the comprehensive safety score.

2. The method for assessing the safety risks of low-altitude unmanned aerial vehicle (UAV) routes based on multi-source geographic information according to claim 1, characterized in that, The specific steps of step S1 are as follows: Step S11: Collect urban 3D model data, topographic data, population and land use data, and ecological environment data; Geographic information data consists of urban 3D model data, topographic data, population and land use data, and ecological environment data. Step S12: Analyze the regional map based on geographic information data, divide the regional map into internal regions, divide the regional map into terrain regions based on urban 3D model data and topographic data, divide the regional map into population distribution regions based on population and land use data, and divide the regional map into ecological regions based on ecological environment data; and rasterize the regional map based on the internal region division results to construct a raster map.

3. The method for assessing the safety risks of low-altitude unmanned aerial vehicle (UAV) routes based on multi-source geographic information according to claim 2, characterized in that, The specific steps of step S12 are as follows: Step S121: Based on the city's 3D model data and terrain data, perform vector mapping on the regional map, divide different terrains using curves, obtain the flight altitude of the low-altitude UAV, extract buildings above the flight altitude from the city's 3D model, and represent the buildings using data points. Based on the data points and the dividing curves, the grid size is set, and the regional map is rasterized to obtain the terrain grid size; Based on population and land use data, the population distribution is obtained, and based on the population distribution, the grid size is set to obtain the population grid size. Based on ecological and environmental data, ecological and environmental protection areas are extracted from the regional map, and grid settings are made according to the ecological and environmental protection areas to obtain the environmental grid size; Step S122: Statistically analyze the terrain raster size, population raster size, and environment raster size; extract the common divisor of the terrain raster size, population raster size, and environment raster size; sort the common divisors in descending order and compile a list of common divisors; iterate through the common divisors in the list; use the common divisors as preset raster sizes to rasterize the regional map; verify the rasterized regional map using geographic information data to determine the optimal raster size; and construct a raster map based on the optimal raster size.

4. The method for assessing the safety risks of low-altitude unmanned aerial vehicle (UAV) routes based on multi-source geographic information according to claim 1, characterized in that, The specific steps of step S2 are as follows: Step S21: Count the number of grid cells on the grid map to obtain the number of grid cells xs×ys; acquire the multi-dimensional data of each grid cell, extract the number of dimensions of the multi-dimensional data, denoted as is, and construct the risk dimension based on the number of dimensions; Based on the number of grid cells and the risk dimension, data for each grid cell and each risk dimension is acquired and denoted as dimension acquisition data wcj(x, y, i); Step S22: Extract data from different risk dimensions based on the collected data, calculate the risk cost surface for each risk dimension, and obtain the risk value.

5. The method for assessing the safety risks of low-altitude unmanned aerial vehicle (UAV) routes based on multi-source geographic information according to claim 4, characterized in that, The specific steps of step S22 are as follows: Step S221: Acquire the height dimension data based on the dimension acquisition data to obtain the building height H of each grid cell. obs (x, y); H obs (x, y) represents the building height of the raster cell at (x, y). The highest building height within the raster map is extracted and denoted as H. max The safe flight altitude of the drone is obtained and denoted as H. safe Based on the building height H of each grid cell obs Using (x, y), the highest building height, and the safe height, the geographical collision risk cost surface is calculated to obtain the collision risk value C. geo (x, y); Step S222: Acquire population dimension data based on dimensional data collection to obtain the population density P of each grid cell. density (x, y) and land use risk coefficient L risk (x, y); P density (x, y) represents the population density of the raster cell at (x, y), L risk (x, y) represents the land use risk coefficient of the raster cell at (x, y); based on the population density of each raster cell, the maximum population density within the raster map is extracted to obtain P. max Based on the population density and land use risk coefficient of each grid cell, combined with the maximum population density, the population security risk cost surface is calculated to obtain the population security risk value C. pop (x, y); Step S223: Acquire environmental dimension data based on dimensional data collection to obtain the distance d(x, y) between each grid cell and the nearest ecological protection zone, and obtain the buffer distance of the ecological protection zone, denoted as D; calculate the ecological disturbance risk cost surface based on the distance and buffer distance between each grid cell and the nearest ecological protection zone to obtain the ecological disturbance risk value C. eco (x, y); 6. The method for assessing the safety risks of low-altitude unmanned aerial vehicle (UAV) routes based on multi-source geographic information according to claim 1, characterized in that, The specific steps of step S3 are as follows: Step S31: Evaluate each risk dimension, analyze the evaluation results using the Analytic Hierarchy Process (AHP), and obtain the risk score for each risk dimension; integrate the risk dimensions based on the risk scores, calculate the weight of each risk dimension, and statistically analyze the weights of each risk dimension to obtain the AHP weight matrix, denoted as W. AHP ; Step S32: Normalize the data for each risk dimension, and calculate the dispersion of the risk dimension based on the normalized data using the entropy weight method. The proportions are calculated based on the dispersion of different risk dimensions, and weights are set for each risk dimension to obtain the EWM weight matrix, denoted as W. EWM ; Step S33: Set the λ parameter, and construct the combined basic weight matrix by combining the AHP weight matrix and the EWM weight matrix; calculate the combined basic weight W by using the elements of the AHP weight matrix and the EWM weight matrix to calculate the elements of the combined basic weight matrix. base(i) ; W base(i) =λ×W AHP(i) +(1-λ)×W EWM(i) ; Among them: W AHP(i) W represents the i-th element in the AHP weight matrix. EWM(i) This represents the i-th element in the EWM weight matrix; A dynamic weight matrix of the same size is constructed based on the combined basic weight matrix. By combining the basic weights with adjustment coefficients for different risk dimensions, the elements in the dynamic weight matrix are calculated to obtain the dynamic weight W. final(i) ; IN final(i) =In base(i) ×a i ; Where: a i This represents the adjustment coefficient for the i-th risk dimension.

7. The method for assessing the safety risks of low-altitude unmanned aerial vehicle (UAV) routes based on multi-source geographic information according to claim 6, characterized in that, The specific steps of step S31 are as follows: Step S311: Arrange all risk dimensions into a matrix, compare each risk dimension pairwise according to the matrix arrangement, quantify the comparison results, and obtain quantified values; The quantized values ​​are statistically analyzed to obtain the judgment matrix pdj; based on the judgment matrix, feature vectors are extracted to obtain the feature vector matrix tjz; The judgment matrix is ​​extracted row by row to obtain the judgment row matrix pdh(i); The risk score fpf(i) is obtained by calculating the feature vector matrix and the judgment row matrix. Among them: tjz i’ Let pdh(i) represent the i'-th element in the eigenvector matrix. i’ This represents the i'-th element in the judgment row matrix of the i-th row; Step S312: Integrate the risk scores of each risk dimension, calculate the ratio between the integrated risk score and the risk scores of each risk dimension, and obtain the weight of each risk dimension, denoted as the analysis weight fqz(i). By statistically analyzing the weights fqz(i), we obtain the AHP weight matrix, denoted as W. AHP W AHP =[fqz(1), fqz(2),..., fqz(is)].

8. The method for assessing the safety risks of low-altitude unmanned aerial vehicle (UAV) routes based on multi-source geographic information according to claim 6, characterized in that, The specific steps of step S32 are as follows: Step S321: Traverse the raster map, normalize each risk dimension of each raster cell to obtain standardized values, and denote the standardized values ​​as p. ij ;p ij Let represent the standardized value of the j-th grid cell in the i-th risk dimension; the discrete value lsz is obtained by calculating the dispersion of the risk dimension based on the standardized value. i ; Where: xs×ys is the number of grid cells, and k is a constant; Step S322: Calculate the weight of each risk dimension based on the discrete values ​​to obtain the entropy weight sqz(i); By performing statistics on the entropy weights sqz(i), the EWM weight matrix is ​​obtained, denoted as W. EWM W EWM =[sqz(1),sqz(2),…,sqz(is)].

9. The method for assessing the safety risks of low-altitude unmanned aerial vehicle (UAV) routes based on multi-source geographic information according to claim 1, characterized in that, The specific steps of step S4 are as follows: The risk value is retrieved from the risk value list and denoted as C. i (x, y), C i (x, y) represents the risk value of the i-th risk dimension at grid (x, y); For the dynamic weights W in the dynamic weight matrix final(i) The dynamic weights are extracted and combined with the risk value to calculate the comprehensive risk value C. total (x, y); 10. A low-altitude unmanned aerial vehicle (UAV) flight path safety risk assessment system based on multi-source geographic information, applicable to the low-altitude UAV flight path safety risk assessment method based on multi-source geographic information as described in any one of claims 1-9, characterized in that, The evaluation system includes: Data acquisition module: acquires the low-altitude flight area of ​​the UAV, obtains a regional map, collects geographic information data based on the regional map, and performs rasterization processing on the regional map based on the collected geographic information data to construct a raster map; Risk Calculation Module: Acquires multi-dimensional data of grid cells based on the grid map, performs statistical analysis on the dimensions of the multi-dimensional data to obtain multiple risk dimensions; calculates the risk cost surface for each risk dimension to obtain a risk value; and performs statistical analysis on the risk values ​​to obtain a list of risk values. Weight setting module: Set weights for risk dimensions to obtain AHP weight matrix and EWM weight matrix; combine AHP weight matrix and EWM weight matrix to obtain basic weight matrix; obtain preset task-specific weight library, extract adjustment coefficients for different risk dimensions based on task-specific weight library, and calculate dynamic weight matrix by combining with basic weight matrix. Comprehensive Analysis Module: Based on the risk value list and combined with the dynamic weight matrix, the risk values ​​of each risk cost surface are weighted and superimposed to calculate the risk value of the comprehensive risk cost surface, thus obtaining the comprehensive risk value; Flight path assessment module: Obtain the flight path of the UAV, extract the corresponding grid cells based on the flight path, sum the comprehensive risk values ​​of the grid cells, calculate the comprehensive safety score, and make dynamic adjustments based on the comprehensive safety score.

Images