Unmanned aerial vehicle base station investigation planning method and system

Abstract

The invention discloses an unmanned aerial vehicle base station investigation planning method and system, and relates to the field of intelligent route planning, and the method comprises the steps: defining an investigation demand, distinguishing stock base stations, building new base stations, planning three scenes in a region, setting corresponding demand parameters, collecting multi-source data, carrying out the fusion calculation of an elevation, and constructing a three-dimensional model with a demand mark. The system comprises an unmanned aerial vehicle platform, a pilot terminal, a task management module, a data acquisition and transmission module, a data processing module and other six modules, can solve the problems of poor adaptability, low efficiency, insufficient precision and the like of traditional investigation, meets the requirements of base station investigation for precision and intelligence, and has a wide application prospect. The method is suitable for 5G deployment, stock base station operation and maintenance and other scenes.

Description

Technical Field

[0001] This invention relates to the field of intelligent route planning, specifically to a method and system for unmanned aerial vehicle (UAV) base station survey and planning. Background Technology

[0002] The accuracy and efficiency of base station surveys directly affect the rationality of base station site selection, signal coverage quality, and subsequent operation and maintenance costs. With the large-scale deployment of 5G networks, the surge in demand for upgrading and transforming existing base stations, and the advancement of signal coverage improvement work in remote areas, traditional base station survey methods have gradually revealed problems such as poor adaptability, low efficiency, and insufficient accuracy, making it difficult to meet the needs of diverse survey scenarios.

[0003] Traditional base station surveys mainly rely on two methods: manual on-site measurement and general-purpose drone aerial photography. Manual on-site measurement requires staff to carry tools such as total stations and GPS positioning equipment to the survey area to collect data on elevation, terrain, and obstacles point by point. Although this method can obtain relatively accurate measured data, it is greatly limited by terrain and transportation conditions. When working in remote or complex areas, it suffers from problems such as high labor costs, long operation cycles, and significant safety risks, and it is difficult to achieve efficient coverage for large-scale planning surveys.

[0004] While general-purpose drone aerial photography has improved survey efficiency to some extent, it lacks customized design for base station survey scenarios. On the one hand, data processing often relies on a single data source and fails to differentiate data weights based on the needs of the survey scenario, resulting in insufficient accuracy and scenario adaptability of the fused elevation data, which cannot provide accurate support for subsequent modeling. On the other hand, flight path planning often takes "shortest path" or "largest coverage" as a single objective, without comprehensively considering the accuracy requirements, terrain adaptability, safety distance, energy consumption, and time constraints of different survey scenarios. This can easily lead to problems such as unreasonable waypoint density, high flight path safety risks, mission timeouts, or energy consumption exceeding limits.

[0005] Meanwhile, traditional surveying methods have significant shortcomings in model building and quality control: model building often uses basic terrain modeling algorithms without optimization for base station surveying needs, resulting in models that cannot provide contextual information for subsequent analysis; quality assessment relies heavily on manual post-processing verification of image clarity, lacking a real-time quantitative evaluation mechanism, which easily leads to the failure to promptly retake low-quality data, affecting subsequent base station design and maintenance decisions. Furthermore, traditional systems are functionally fragmented, with task management, data acquisition, processing, and evaluation not integrated, lacking historical data storage and abnormal scenario early warning capabilities, making it difficult to support efficient control and iterative optimization of the entire surveying process, and failing to meet the current demands for more precise, intelligent, and efficient base station surveying. Summary of the Invention

[0006] The purpose of this invention is to provide a method and system for UAV base station survey and planning. This method identifies three survey scenarios—existing base stations, new base stations, and regional planning—and sets corresponding requirement parameters. It then collects data from satellite remote sensing, ground measurements, and historical data, calculates weights to obtain fused elevations, constructs a compliant 3D model using an improved particle swarm optimization algorithm, embeds requirement markers, and generates an optimal flight path based on a multi-dimensional constraint system built from the model. Finally, the UAV performs the survey according to the flight path, and dynamic calibration and image quality assessment address the problems of poor adaptability, low efficiency, and insufficient accuracy in traditional base station surveys. This method meets the precision and intelligent requirements of three scenarios: existing base station survey, new base station survey, and regional planning survey.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] A method for surveying and planning unmanned aerial vehicle (UAV) base stations, characterized by the following steps:

[0009] S1: Define the requirements for survey types and quantify the model accuracy requirements for different scenarios based on differentiated requirements;

[0010] S2: Collect relevant data and calculate elevation based on data weights according to requirements, accuracy, and timeliness;

[0011] S3: Build a model using elevation as the input of the vertical axis and embed requirement markers in the model;

[0012] S4: Based on the model with demand tags, construct a constraint system and generate a survey route;

[0013] S5: Based on the planned route and the associated tasks, it performs reconnaissance operations and dynamically calibrates to ensure waypoint accuracy.

[0014] Step S1 defines the survey type requirements, including existing base station survey, new base station survey, and regional planning survey, and sets the survey type requirement parameter T. k Specifically, for existing base stations, T1 = 0.65; for newly built base stations, T2 = 0.85; and for regional planning, T3 = 0.45.

[0015] Step S2 involves collecting relevant data and calculating elevations based on requirements, accuracy, and timeliness. This specifically includes collecting satellite remote sensing terrain elevation data H. s (x,y), ground-measured base station coordinates (x,y) b,i ,y b,i ,z b,i ), historical survey equipment operating parameters data and regional signal data;

[0016] Data weights are calculated based on demand, accuracy, and timeliness:

[0017]

[0018] In the formula, w s For the weights of satellite remote sensing terrain elevation data, w h The weights are determined by historical survey equipment operating parameters and regional signal data, and satisfy w. s +w g +w h =1,T k σ is a parameter used to quantify the requirements of different exploration types. s λ represents the standard deviation of the satellite remote sensing terrain elevation data, t represents the time since the satellite remote sensing data was collected, and λ represents the error of the satellite remote sensing terrain elevation data. s T represents the cloud cover coefficient for satellite remote sensing data. k σ is a parameter used to quantify the requirements of different exploration types. g σ represents the standard deviation of the ground-based base station coordinate data. h t represents the standard deviation of the historical data. h This refers to the time elapsed since the present for historical data.

[0019] The elevation is obtained after merging:

[0020] H fusion =w s H s +w g H g +w h H h

[0021] In the formula, H fusion For the merged elevation data, w s H represents the weight of satellite remote sensing data. s For satellite remote sensing terrain elevation data, w g H represents the weight of the ground-measured data. g For ground-measured elevation data, w h H represents the weight of historical data. h This refers to elevation information from historical data.

[0022] Step S3 uses elevation as the input for the vertical axis to construct the model and embeds requirement markers into the model. Specifically, it involves using an improved particle swarm optimization algorithm to construct the model, with a dual objective function that combines minimizing height deviation with optimizing surface smoothness. The formula is as follows:

[0023]

[0024] In the formula, ΔZ is the model height deviation, n is the number of ground-based measured base stations involved in the height deviation calculation, and z model,i Let z be the elevation value of the model at the location of the i-th ground-based measured base station. b,i Let w be the actual elevation value of the i-th ground-based measured base station. gHere, S represents the weight of the measured ground data, m represents the smoothness of the model surface, k represents the number of computational units divided into the model surface, j represents the index of the k-th computational unit, and θ represents the index of the neighboring units of the k-th computational unit. k,j The angle between the k-th calculation unit and the j-th adjacent unit;

[0025] Iterate the model until ΔZ < 0.04∑z b,i Furthermore, after S>50°, demand markers are embedded to mark key areas in the model, including the equipment aging warning zone S for existing base stations. old New base station marking of geologically stable zone S stab Regional planning marker signal blind zone S blind Synchronously record the area S key Obstacle coordinates (x) o ,y o ,z o ).

[0026] Step S4, based on the model with demand tags, constructs a constraint system and generates a survey route, including:

[0027] Key area waypoint density:

[0028] Non-key areas: ρ non-keu ≤0.3 units / km 2

[0029] Terrain adaptation constraint: Flight altitude h(x,y)≥H fusion (x,y)+50×(1+0.15(1-T k ))

[0030] Flight slope: θ slope ≤25°-3°×(1-T k )

[0031] safe distance d safe (n)≥10×(1+0.2δ obs )

[0032] Energy consumption per segment of a flight route:

[0033] Time window constraint: Total flight time

[0034] In the formula, ρ key For waypoint density in key areas, T k To quantify the parameters required for different exploration types, S key ρ represents the area of ​​the key region. non-keyH represents the waypoint density in non-critical areas, h(x,y) represents the UAV's flight altitude at the planar coordinates (x,y), and H... fusion (x,y) represents the merged elevation at the plane coordinates (x,y), and θ slope d represents the slope of the drone's flight path. safe (n) represents the safe distance between the UAV and the obstacle at the nth flight segment or waypoint, where n is the index of the flight segment or waypoint, and δ obs Let E(n) be the obstacle hazard coefficient, E(n) be the energy consumption of the UAV in the nth segment, m be the mass of the UAV, v be the flight speed of the UAV in this segment, g be the acceleration due to gravity, and Δz be the altitude change in this segment. wind E represents the wind speed in the flight area for this segment. max T represents the maximum total energy consumption of the drone. total Let v(n) be the total flight time of the UAV reconnaissance, v(n) be the flight speed of the UAV in the nth segment, and t be the total flight time of the UAV reconnaissance. stay (p) represents the dwell time of the UAV at the p-th waypoint, where p is the waypoint index, and T limit The maximum permissible total time for the exploration mission;

[0035] The final output is a sequence of route node coordinates {(x1,y1,z1),(x2,y2,z2),…} that satisfies all constraints.

[0036] Step S5, based on the planned flight path and the associated task, performs a reconnaissance operation and dynamically calibrates to ensure waypoint accuracy. Specifically, as the UAV flies according to the flight path sequence, it receives waypoint coordinates and task parameters in real time. After arriving at the preset waypoint, it takes images in sequence: an eight-sided view, a panoramic view, an overall view of the tower and mast, a view of the platform and antenna, and a view of the covered target. The image quality index is calculated in real time using the following formula:

[0037] Q total = 0.35D + 0.25C + 0.2E + 0.15S feat +0.05T k

[0038] In the formula, Q total The overall image quality index is defined by D, where D is the image sharpness score, C is the image contrast score, E is the image exposure score, and S is the image quality score. feat T represents the image feature recognition rate. k To quantify the parameters required for different types of exploration, Q total A reshoot is triggered when the time is less than 85 seconds.

[0039] A UAV base station survey and planning system, characterized by comprising a UAV platform, a pilot terminal, and a task management module, a data acquisition and transmission module, a data processing module, a constraint control module, a flight path optimization module, and a quality assessment module deployed on a cloud or local server; the task management module allows administrators to create or import base station survey tasks, associates survey type requirements with quantitative parameters corresponding to existing base station surveys and new base station surveys, and sends task information to the pilot terminal; the pilot terminal receives task information, sends flight control commands to the UAV platform, and controls the UAV platform to collect survey data; the survey data includes base station image data and elevation-related data, the base station image data at least includes octagonal views, coverage target maps, and overall tower and mast maps, and the elevation-related data includes satellite remote sensing terrain elevation data, ground-measured elevation data corresponding to the base station coordinates, and historical elevation data corresponding to historical survey data; the data acquisition and transmission module collects the survey data and transmits it to the data processing module. The data processing module includes an elevation fusion unit and a model building unit. The elevation fusion unit calculates the fused elevation based on the weights of satellite remote sensing data, ground measured data, and historical data, combined with the corresponding elevation data. The model building unit constructs a 3D model of the base station or survey area. When surveying a new base station, a 3D model of the proposed location is performed, and during regional planning surveys, a 3D model of the entire area is performed. The constraint control module constructs a multi-dimensional constraint system based on the model with demand markings and the fused elevation, including demand-related constraints, terrain adaptation constraints, obstacle avoidance constraints, energy consumption control constraints, and time window constraints, to verify the flight parameters of the UAV platform. The flight path optimization module optimizes the UAV flight path based on the verification results of the multi-dimensional constraint system. During regional planning surveys, the flight path is also optimized based on the results of batch survey point selection within the area. The quality assessment module evaluates the image data quality of the base station based on the total image quality index. If the total image quality index is less than 85, a re-acquisition command is triggered.

[0040] The task management module also includes a task data preview unit, which displays basic information, survey information and base station image data. The basic information is the task data information that comes with the issued task, the survey information is the base station data output based on the UAV survey data, and the base station image data may also include at least one of the following: panoramic view, platform and antenna view, engineering parameter view, abnormal scene view, and proposed location view.

[0041] The system includes a storage module that stores historical data for all exploration tasks, including existing site exploration records from previous years and regional planning iteration models, and supports filtering and viewing by time range and exploration type.

[0042] The system supports marking and photographing abnormal scenes. If situations such as tower tilting or equipment damage are detected, abnormal scene images can be manually photographed and text annotations can be added. After being uploaded to the system, the system will automatically mark the abnormal location in the 3D model and trigger a maintenance warning.

[0043] The planning and surveying of UAV base stations is based on the core logic of "demand-driven, data fusion, model support, constraint control, and dynamic execution" to achieve precise and intelligent surveying operations. The specific mechanism is as follows:

[0044] First, the survey scenario types are clearly defined, distinguishing between three core scenarios: existing base stations, newly built base stations, and regional planning. Specific survey type requirement parameters are then set for each scenario. These parameters directly relate to subsequent data weighting, model accuracy requirements, and constraint thresholds. For example, the survey of newly built base stations has higher accuracy requirements, so the corresponding parameter values ​​are higher than those for existing base stations and regional planning. This approach anchors the survey priority and technical standards for different scenarios from the outset, ensuring that subsequent steps accurately match actual needs.

[0045] Three core data types were collected: satellite remote sensing topographic elevation, ground-based measured base station coordinates, and historical survey equipment parameters and regional signals. The weights of each data type were calculated from three dimensions: "demand adaptability, data accuracy, and timeliness." The weight of satellite remote sensing data was determined by considering survey demand parameters, data error standard deviation, time elapsed, and cloud cover; higher priority, better data accuracy, and stronger timeliness resulted in a higher weight. The weight of ground-based measured data was negatively correlated with survey demand parameters, emphasizing the high accuracy advantage of measured data in low-demand scenarios. The weight of historical data was set based on its own error and timeliness to ensure basic data support. The weights of the three data types satisfied a normalization condition that summed to 1. A weighted calculation was then performed to obtain fused elevation data, which combined the advantages of various data sources and became the core vertical axis input for subsequent model construction.

[0046] An improved particle swarm optimization algorithm was used to construct a 3D model of the survey area. The model was optimized with the dual objectives of "minimizing model height deviation" and "maximizing model surface smoothness." Height deviation was calculated by comparing the difference between the model elevation and the measured elevation of the ground base station, and by combining the weights of the measured ground data, to ensure the vertical accuracy of the model. Surface smoothness was determined by statistically analyzing the average angle between each calculated unit on the model surface and its adjacent units, and it had to meet a minimum smoothness threshold requirement. After the model was iterated until both height deviation and smoothness met the standards, requirement markers were embedded in the model to mark key areas for different survey scenarios. At the same time, key information such as the area of ​​key areas and the coordinates of obstacles were recorded to provide scenario-based support for subsequent flight path planning.

[0047] Based on a 3D model with requirement markers, a multi-dimensional constraint system is constructed to screen the optimal reconnaissance route: Regarding waypoint density, the waypoint density in key areas is calculated based on reconnaissance requirement parameters and area, and must not be lower than a minimum threshold to ensure reconnaissance accuracy; the waypoint density in non-key areas should not exceed a maximum threshold to control resource costs. Regarding terrain adaptability, the flight altitude must be higher than the sum of the fused elevation and safe altitude coefficients, and the route gradient must not exceed the basic maximum gradient to ensure flight stability. Regarding safety and efficiency, the safe distance between the UAV and obstacles must be dynamically adjusted according to the obstacle's hazard level; single-segment energy consumption should not exceed a fixed percentage of total energy consumption, and total flight time should not exceed a fixed percentage of the maximum allowable mission time to avoid exceeding energy limits or timeouts. Finally, a sequence of route node coordinates that satisfies all constraints is output, forming a directly executable flight path.

[0048] The drone flies along a planned route, and upon reaching the preset waypoints, it takes key images such as octagonal views, panoramic views, and overall tower and mast views according to standard procedures, while calculating the image quality index in real time. This index comprehensively considers image sharpness, contrast, exposure, feature recognition rate, and survey requirements. If the index falls below the acceptable standard, a re-shoot is triggered to ensure image data quality. Simultaneously, a dynamic calibration mechanism corrects waypoint position deviations during flight, combining real-time received waypoint coordinates with mission parameters to ensure the actual flight trajectory closely matches the planned route, ultimately completing the survey data collection and quality verification.

[0049] Compared with the prior art, the beneficial effects of the present invention are:

[0050] By setting specific requirement parameters for three types of survey scenarios—existing base stations, newly built base stations, and regional planning—subsequent steps such as data weight allocation, model accuracy requirements, and constraint thresholds are deeply tied to actual needs. This avoids the "one-size-fits-all" survey model of existing technologies, enables precise matching of survey resources in different scenarios, and improves the targeting of surveys.

[0051] By integrating satellite remote sensing, ground measurement, and historical data to generate high-precision fused elevation data, we can leverage the advantages of each type of data while avoiding the problems of insufficient accuracy from a single data source or disordered superposition of multiple data sources in existing technologies, thereby improving data utilization efficiency and elevation data accuracy.

[0052] An improved particle swarm optimization algorithm is used to construct a 3D model with the dual objectives of "minimizing height deviation and optimizing surface smoothness". After the model meets the requirements, requirement markers are embedded and key information such as obstacle coordinates are recorded simultaneously. Compared with the existing technology that only builds a basic terrain model, it can provide richer scenario-based information for route planning and reduce the blindness of subsequent route planning.

[0053] A multi-dimensional constraint system of "waypoint density, terrain adaptation, safe distance, energy consumption, and time" is constructed to control waypoint density according to regional differences. Flight altitude, slope, and other indicators are dynamically adjusted in combination with demand parameters to ensure that the route meets the requirements of accuracy, safety, and efficiency. This avoids the insufficient survey accuracy or safety risks caused by existing route planning technologies that only consider a single factor, thereby improving the practicality of the route and flight safety.

[0054] During the exploration and execution phase, data quality is assessed in real time through an image quality index. When the index falls below 85, a re-shoot is triggered. At the same time, dynamic calibration is used to ensure waypoint accuracy, forming a quality closed loop of "acquisition-assessment-calibration-re-shoot". Compared with existing technologies that lack real-time quality verification, this can effectively prevent poor-quality data from affecting subsequent analysis and improve the reliability of exploration data.

[0055] The supporting exploration and planning system integrates modules such as task management, data acquisition and transmission, processing, constraint control, and quality assessment. It supports task data preview, historical data storage and filtering, and abnormal scene marking and early warning. Compared with existing systems with scattered functions, it realizes integrated management of the entire exploration process, improving work efficiency and problem response speed. Attached Figure Description

[0056] Figure 1 This is a flowchart of a UAV base station survey and planning method according to the present invention;

[0057] Figure 2 This is a task management interface diagram of an unmanned aerial vehicle (UAV) base station survey and planning system according to the present invention. Detailed Implementation

[0058] The technical solutions of the present invention will now be described in detail with reference to the accompanying drawings.

[0059] like Figure 1 As shown, a method for surveying and planning unmanned aerial vehicle (UAV) base stations is characterized by the following steps:

[0060] S1: Define survey type requirements, quantify the model accuracy requirements for different scenarios based on differentiated needs; define survey type requirements including existing base station survey, new base station survey, and regional planning survey, and set their survey type requirement parameter T. k Specifically, for existing base stations, T1 = 0.65; for newly built base stations, T2 = 0.85; and for regional planning, T3 = 0.45.

[0061] S2: Collect relevant data and calculate elevation based on data weights according to requirements, accuracy, and timeliness; specifically, this includes collecting satellite remote sensing terrain elevation data H. s (x,y), ground-measured base station coordinates (x,y) b,i ,y b,i ,z b,i), historical survey equipment operating parameters data and regional signal data;

[0062] Data weights are calculated based on demand, accuracy, and timeliness:

[0063]

[0064] In the formula, w s For the weights of satellite remote sensing terrain elevation data, w h The weights are determined by historical survey equipment operating parameters and regional signal data, and satisfy w. s +w g +w h =1,T k σ is a parameter used to quantify the requirements of different exploration types. s λ represents the standard deviation of the satellite remote sensing terrain elevation data, t represents the time since the satellite remote sensing data was collected, and λ represents the error of the satellite remote sensing terrain elevation data. s T represents the cloud cover coefficient for satellite remote sensing data. k σ is a parameter used to quantify the requirements of different exploration types. g σ represents the standard deviation of the ground-based base station coordinate data. h t represents the standard deviation of the historical data. h This refers to the time elapsed since the present for historical data.

[0065] The elevation is obtained after merging:

[0066] H fusion =w s H s +w g H g +w h H h

[0067] In the formula, H fusion For the merged elevation data, w s H represents the weight of satellite remote sensing data. s For satellite remote sensing terrain elevation data, w g H represents the weight of the ground-measured data. g For ground-measured elevation data, w h H represents the weight of historical data. h This refers to elevation information from historical data.

[0068] S3: Construct a model using elevation as the input of the vertical axis and embed requirement markers into the model; specifically, this includes using an improved particle swarm optimization algorithm to construct the model, with a dual objective function of minimizing height deviation combined with optimizing surface smoothness, the formula of which is:

[0069]

[0070] In the formula, ΔZ is the model height deviation, n is the number of ground-based measured base stations involved in the height deviation calculation, and z model,i Let z be the elevation value of the model at the location of the i-th ground-based measured base station. b,i Let w be the actual elevation value of the i-th ground-based measured base station. g Here, S represents the weight of the measured ground data, m represents the smoothness of the model surface, k represents the number of computational units divided into the model surface, j represents the index of the k-th computational unit, and θ represents the index of the neighboring units of the k-th computational unit. k,j The angle between the k-th calculation unit and the j-th adjacent unit;

[0071] Iterate the model until ΔZ < 0.04∑z b,i Furthermore, after S>50°, demand markers are embedded to mark key areas in the model, including the equipment aging warning zone S for existing base stations. old New base station marking of geologically stable zone S stab Regional planning marker signal blind zone S blind Synchronously record the area S key Obstacle coordinates (x) o ,y o ,z o ).

[0072] S4: Based on the model with demand tags, construct a constraint system and generate survey routes; including:

[0073] Key area waypoint density:

[0074] Non-key areas: ρ non-key ≤0.3 units / km 2

[0075] Terrain adaptation constraint: Flight altitude h(x,y)≥H fusion (x,y)+50×(1+0.15(1-T k ))

[0076] Flight slope: θ slope ≤25°-3°×(1-T k )

[0077] safe distance d safe (n)≥10×(1+0.2δ obs )

[0078] Energy consumption per segment of a flight route:

[0079] Time window constraint: Total flight time

[0080] In the formula, ρ key For waypoint density in key areas, T k To quantify the parameters required for different exploration types, S key ρ represents the area of ​​the key region. non-k H represents the waypoint density in non-critical areas, h(x,y) represents the UAV's flight altitude at the planar coordinates (x,y), and H... fusion (x,y) represents the merged elevation at the plane coordinates (x,y), and θ slope d represents the slope of the drone's flight path. safe (n) represents the safe distance between the UAV and the obstacle at the nth flight segment or waypoint, where n is the index of the flight segment or waypoint, and δ obs Let E(n) be the obstacle hazard coefficient, E(n) be the energy consumption of the UAV in the nth segment, m be the mass of the UAV, v be the flight speed of the UAV in this segment, g be the acceleration due to gravity, and Δz be the altitude change in this segment. wind E represents the wind speed in the flight area for this segment. max T represents the maximum total energy consumption of the drone. total Let v(n) be the total flight time of the UAV reconnaissance, v(n) be the flight speed of the UAV in the nth segment, and t be the total flight time of the UAV reconnaissance. stay (p) represents the dwell time of the UAV at the p-th waypoint, where p is the waypoint index, and T limit The maximum permissible total time for the exploration mission;

[0081] The final output is a sequence of route node coordinates {(x1,y1,z1),(x2,y2,z2),…} that satisfies all constraints.

[0082] S5: Based on the planned flight path and associated tasks, perform reconnaissance operations and dynamically calibrate to ensure waypoint accuracy. Specifically, when the UAV flies according to the flight path sequence, it receives waypoint coordinates and task parameters in real time. Upon reaching the preset waypoint, it sequentially captures eight-sided images, a panoramic view, an overall view of the tower and mast, a view of the platform and antenna, and a view of the covered target. It also calculates the image quality index in real time using the following formula:

[0083] Q total = 0.35D + 0.25C + 0.2E + 0.15S feat +0.05T k

[0084] In the formula, Q total The overall image quality index is defined by D, where D is the image sharpness score, C is the image contrast score, E is the image exposure score, and S is the image quality score. feat T represents the image feature recognition rate. k To quantify the parameters required for different types of exploration, Q totalA reshoot is triggered when the time is less than 85 seconds.

[0085] A UAV base station survey and planning system, characterized by comprising a UAV platform, a pilot terminal, and a task management module, a data acquisition and transmission module, a data processing module, a constraint control module, a flight path optimization module, and a quality assessment module deployed on a cloud or local server; the task management module allows administrators to create or import base station survey tasks, associates survey type requirements with quantitative parameters corresponding to existing base station surveys and new base station surveys, and sends task information to the pilot terminal; the pilot terminal receives task information, sends flight control commands to the UAV platform, and controls the UAV platform to collect survey data; the survey data includes base station image data and elevation-related data, the base station image data at least includes octagonal views, coverage target maps, and overall tower and mast maps, and the elevation-related data includes satellite remote sensing terrain elevation data, ground-measured elevation data corresponding to the base station coordinates, and historical elevation data corresponding to historical survey data; the data acquisition and transmission module collects the survey data and transmits it to the data processing module. The data processing module includes an elevation fusion unit and a model building unit. The elevation fusion unit calculates the fused elevation based on the weights of satellite remote sensing data, ground measured data, and historical data, combined with the corresponding elevation data. The model building unit constructs a 3D model of the base station or survey area. When surveying a new base station, a 3D model of the proposed location is performed, and during regional planning surveys, a 3D model of the entire area is performed. The constraint control module constructs a multi-dimensional constraint system based on the model with demand markings and the fused elevation, including demand-related constraints, terrain adaptation constraints, obstacle avoidance constraints, energy consumption control constraints, and time window constraints, to verify the flight parameters of the UAV platform. The flight path optimization module optimizes the UAV flight path based on the verification results of the multi-dimensional constraint system. During regional planning surveys, the flight path is also optimized based on the results of batch survey point selection within the area. The quality assessment module evaluates the image data quality of the base station based on the total image quality index. If the total image quality index is less than 85, a re-acquisition command is triggered.

[0086] The task management module also includes a task data preview unit, which displays basic information, survey information and base station image data. The basic information is the task data information that comes with the issued task, the survey information is the base station data output based on the UAV survey data, and the base station image data may also include at least one of the following: panoramic view, platform and antenna view, engineering parameter view, abnormal scene view, and proposed location view.

[0087] The system includes a storage module that stores historical data for all exploration tasks, including existing site exploration records from previous years and regional planning iteration models, and supports filtering and viewing by time range and exploration type.

[0088] The system supports marking and photographing abnormal scenes. If situations such as tower tilting or equipment damage are detected, abnormal scene images can be manually photographed and text annotations can be added. After being uploaded to the system, the system will automatically mark the abnormal location in the 3D model and trigger a maintenance warning.

[0089] Taking the survey of existing base stations as an example, the survey type is determined to be the survey of existing base stations. The corresponding quantitative parameter for the survey type is set to 0.65. The core task of the survey is to assess the aging status of existing base station equipment. It is necessary to focus on collecting data on the integrity of the tower and mast structure, the appearance damage of the equipment, and the distribution of surrounding obstacles. This clarifies that the model accuracy must meet the requirements for accurate identification of the aging area of ​​the equipment.

[0090] Satellite remote sensing topographic elevation data was collected, with a standard deviation of 0.5 meters and a time elapsed 30 days ago. A cloud cover coefficient of 0.1 was also included. Coordinates of 10 ground-based measured base stations and historical survey equipment parameters from 5 years ago for the same base station were also collected, with a standard deviation of 0.8 meters and a time elapsed 1825 days ago. Considering the survey type requirements, the accuracy and timeliness of each data type, the weights were calculated as follows: satellite remote sensing data approximately 0.42, ground-based measured data approximately 0.38, and historical data approximately 0.2, with a total weight of 1. The fused elevation data for each coordinate point in the survey area was obtained through weighted calculation with the corresponding elevation values. The fused elevation was 85.6 meters. The model was iterated with the objectives of minimizing model height deviation and optimizing model surface smoothness until the model height deviation was less than 0.04 times the sum of the elevations of all ground-based measured base stations. In this case, the sum of the elevations of all measured base stations was 850 meters, meaning the height deviation needed to be less than 34 meters and the model surface smoothness greater than 50°. Once the model meets the standards, key areas are marked in the model, and the area within 100 meters of the bottom of the tower mast is set as the equipment aging warning zone. At the same time, the area of ​​this warning zone is recorded as 0.01 square kilometers, as well as the coordinates of three tall tree obstacles in the surrounding area.

[0091] For key areas, the waypoint density is calculated based on the survey requirements and the area, resulting in 195 waypoints / square kilometer, meeting the requirement of no less than 0.8 waypoints / square kilometer. For non-key areas, the waypoint density is set at 0.2 waypoints / square kilometer, meeting the standard of no more than 0.3 waypoints / square kilometer. The flight altitude must be higher than the sum of the fused elevation and safety altitude coefficients, calculated using the survey requirements parameters to be 52.625 meters. For example, at the coordinates of the mast base mentioned above, the flight altitude must be no less than 138.225 meters. The flight slope must not exceed 24.05°. The safe distance from tree obstacles must be no less than 10.6 meters. The energy consumption per flight segment must not exceed 15% of the UAV's maximum total energy consumption of 1000 watt-hours, i.e., 150 watt-hours. The total flight time must not exceed 80% of the maximum allowable mission time of 120 minutes, i.e., 96 minutes.

[0092] Output a sequence of route node coordinates that meets all constraints, ensuring coverage of the equipment aging warning zone and surrounding critical areas.

[0093] The drone flew along the planned route, and upon reaching each waypoint, it sequentially took eight-sided photos, an overall view of the tower and mast, and detailed images of the equipment, calculating the image quality index in real time. Based on the survey requirements parameters, the calculated image quality index was approximately 86.5, exceeding the acceptable standard of 85, thus eliminating the need for re-shooting.

Claims

1. A method for unmanned aerial base station surveying and planning, characterized in that, Comprise the following steps: S1: define the survey type requirements, according to the differentiated demand to quantify the different scene model accuracy requirements; S2: collect relevant data, based on demand, precision, timeliness calculation data weight calculation elevation; S3: the elevation as the input of the vertical axis to build the model, and embed the demand mark in the model; S4: based on the model with demand mark, build the constraint system, generate survey route; S5: based on the planned route and the bound task, execute the survey operation, dynamic calibration to ensure the accuracy of the waypoint. 2.The unmanned aerial vehicle base station survey planning method of claim 1, wherein, Step S1 defines the survey type requirement, including inventory base station survey, new base station survey, and regional planning survey, and sets the survey type requirement parameter T k Specifically, T1=0.65 for inventory base station, T2=0.85 for new base station, and T3=0.45 for regional planning. 3.The unmanned aerial vehicle base station survey planning method of claim 1, wherein, Step S2 collects relevant data, and calculates the data weight based on the demand, precision, and timeliness to calculate the height, specifically including collecting satellite remote sensing terrain height data H s (x, y), ground measured base station coordinates (x b,i ,y b,i ,z b,i ), historical survey equipment parameter data, and regional signal data Based on demand, precision, timeliness calculation data weight: In the formula, w s is the weight of satellite remote sensing terrain elevation data, w h is the weight of historical survey equipment design parameter data and regional signal data, and satisfies w s +w g +w h =1, T k is a survey type demand quantization parameter, σ s is the error standard deviation of satellite remote sensing terrain elevation data, t is the time of satellite remote sensing data from now, λ s is the cloud cover shielding coefficient of satellite remote sensing data, T k is a survey type demand quantization parameter, σ g is the error standard deviation of ground measured base station coordinate data, σ h is the error standard deviation of historical data, t h is the time of historical data from now. After fusion, the elevation is obtained: H fusion = w s H s + w g H g + w h H h In the formula, H fusion is the fused elevation data, w s is the satellite remote sensing data weight, H s is the satellite remote sensing terrain elevation data, w g is the ground measured data weight, H g is the ground measured elevation data, w h is the historical data weight, H h is the elevation information in the historical data. 4.The unmanned aerial vehicle base station survey planning method of claim 1, wherein, Step S3, the elevation as the input of the vertical axis to build the model, and embed the demand mark in the model, specifically including using improved particle swarm algorithm to build the model, the double objective function is the minimum height deviation combined with the optimization of surface smoothness, the formula is: where ΔZ is the model height bias, n is the number of ground measured base stations participating in the height bias calculation, z model,i is the elevation value of the model at the i-th ground measured base station position, z b,i is the actual elevation value of the i-th ground measured base station, w g is the ground measured data weight, S is the model surface smoothness, m is the number of calculation units divided on the model surface, k is the index of the k-th calculation unit, j is the index of the adjacent unit of the k-th calculation unit, θ k,j is the included angle between the k-th calculation unit and the j-th adjacent unit; Iterate the model to ΔZ < 0.04∑z b,i And S > 50 ° after embedding demand markers, mark the key areas in the model, including inventory base station Marking equipment aging warning area S old , New base station marking geological stable area S stab , Regional planning marking signal blind area S blind , Synchronous recording area S key , Obstacle coordinates (x o ,y o ,z o ) 5. The unmanned aerial vehicle base station surveying and planning method of claim 1, wherein, Step S4, based on the model with demand mark, build the constraint system, generate survey route, including: Focus area waypoint density: Non-focus area: p non-key ≤ 0.3 per km 2 Terrain adaptation constraint: flight height h(x, y) ≥ H fusion (x, y) + 50 x (1 + 0.15(1 - T k )) Flight slope: θ slope ≤ 25° - 3° x (1 - T k ) Safety distance d safe (n) ≥ 10 x (1 + 0.2δ obs ) Single leg route energy consumption: Time window constraint: total flight time wherein ρ key is the point density of the key area, T k is the survey type demand quantization parameter, S key is the area of the key area, ρ non-key is the point density of the non-key area, h(x, y) is the flight height of the UAV at the planar coordinates (x, y), H fusion (x, y) is the fused height at the planar coordinates (x, y), θ slope is the slope of the UAV flight route, d safe (n) is the safety distance of the UAV from the obstacle at the nth flight segment or point, n is the index of the flight segment or point, δ obs is the obstacle danger coefficient, E(n) is the energy consumption of the UAV at the nth flight segment, m is the mass of the UAV, v is the flight speed of the UAV at the flight segment, g is the gravitational acceleration, Δz is the height change amount of the flight segment, v wind is the wind speed of the flight region of the flight segment, E max is the maximum total energy consumption of the UAV, T total is the total flight time of the UAV survey, v(n) is the flight speed of the UAV at the nth flight segment, t stay (p) is the stay time of the UAV at the pth point, p is the point index, T limit is the maximum allowed total time of the survey task; The final output of the route node coordinate sequence {(x1, y1, z1), (x2, y2, z2), …} that satisfies all constraints.

6. The unmanned aerial vehicle base station surveying and planning method of claim 1, wherein, Step S5, based on the planned route and the bound task, execute the survey operation, dynamic calibration to ensure the accuracy of the waypoint, specifically when the unmanned aerial vehicle flies according to the route sequence, real-time receives the waypoint coordinates and task parameters, arrives at the preset waypoint, and then shoots eight directions, panoramic, tower mast overall diagram, platform and antenna, cover target diagram, and real-time calculates the image quality index, the formula is: Q total = 0.35D + 0.25C + 0.2E + 0.15S feat + 0.05T k wherein Q total is the total image quality index, D is the image sharpness score, C is the image contrast score, E is the image exposure score, S feat is the image feature recognition rate, T k is the survey type requirement quantification parameter, Q total <85 triggers a re-shoot.

7. A drone base station survey planning system, comprising: The system comprises a UAV platform, a pilot terminal, a task management module deployed in the cloud or a local server, a data acquisition and transmission module, a data processing module, a constraint control module, a flight path optimization module, and a quality evaluation module. The task management module is used by an administrator to create or import base station survey tasks, and the task is associated with the survey type demand quantization parameter corresponding to the survey of the existing base station and the survey of the newly-built base station, and the task information is sent to the pilot terminal. The pilot terminal receives the task information, sends the flight control instruction to the UAV platform, and controls the UAV platform to collect survey data. The survey data includes base station image data and elevation-related data, and the base station image data at least covers eight directions, covers the target map, and the tower mast overall map, and the elevation-related data includes satellite remote sensing terrain elevation data, ground measured elevation data corresponding to the ground measured base station coordinates, and historical elevation data corresponding to the historical survey data. The data acquisition and transmission module acquires the survey data and transmits it to the data processing module. The data processing module comprises an elevation fusion unit and a model construction unit. The elevation fusion unit calculates the fusion elevation according to the weights of the satellite remote sensing data, the ground measured data, and the historical data, and combines the corresponding elevation data. The model construction unit constructs a three-dimensional model of the base station or the survey area, and performs 3D modeling on the proposed location when surveying the newly-built base station, and performs 3D modeling on the area when surveying the regional planning. The constraint control module constructs a multi-dimensional constraint system including demand-related constraints, terrain adaptation constraints, obstacle avoidance constraints, energy consumption control constraints, and time window constraints based on the model marked with the demand and the fusion elevation, and verifies the flight parameters of the UAV platform. The flight path optimization module optimizes the flight path of the UAV based on the verification result of the multi-dimensional constraint system, and also optimizes the flight path in combination with the batch survey point selection result in the regional planning when surveying the regional planning. The quality evaluation module evaluates the quality of the base station image data based on the image quality total index, and if the image quality total index is less than 85, a reacquisition instruction is triggered.

8. The unmanned aerial vehicle base station survey planning system of claim 7, wherein, The task management module is also provided with a task data preview unit to display basic information, survey information, and base station image data. The basic information is the task data information sent with the task, the survey information is the base station data output based on the UAV survey data, and the base station image data can also include at least one of the panoramic view, the platform and antenna diagram, the work parameter diagram, the abnormal scene diagram, and the proposed location diagram.

9. The unmanned aerial vehicle base station survey planning system of claim 7, wherein, The system comprises a storage module that stores the historical data of all survey tasks, including the survey records of the existing station in the past years and the iterative model of the regional planning, and supports filtering and viewing according to the time range and the survey type.

10. The unmanned aerial vehicle base station survey planning system of claim 7, wherein, The system supports abnormal scene marking and shooting. When the tower mast is found to be tilted or the equipment is damaged, an abnormal scene diagram can be manually shot and a text note can be added. After being uploaded to the system, the system automatically marks the abnormal position in the three-dimensional model and triggers a maintenance warning.

Images