Low-altitude airspace three-dimensional modeling method based on adaptive mesh and multi-source data fusion

Abstract

The application belongs to the technical field of low-altitude three-dimensional modeling, and discloses a low-altitude airspace three-dimensional modeling method based on adaptive grid and multi-source data fusion, which comprises the following steps: acquiring a digital elevation model, building outlines and real-time weather data, obtaining multi-source fusion data through space-time registration and fusion; simulating air flow by using a computational fluid dynamics model to obtain the background air flow field distribution of wind speed, turbulence intensity and air pressure change; performing initial grid division according to the air flow field, terrain and building density information; simulating the interaction between an aircraft and air flow on the initial grid to obtain an aircraft-induced disturbed air flow field, dynamically adjusting the grid density and generating a dynamic three-dimensional grid model; and finally, fusing the background air flow field and the disturbed air flow field to construct a low-altitude airspace three-dimensional model of the aircraft-air flow coupling effect. The method can accurately simulate the air flow change in the low-altitude airspace by dynamically adjusting the grid resolution, and improve the safety of aircraft path planning and the accuracy of airspace management.

Description

Technical Field

[0001] This invention relates to the field of low-altitude 3D modeling technology, specifically to a low-altitude airspace 3D modeling method based on adaptive mesh and multi-source data fusion. Background Technology

[0002] With the gradual opening of low-altitude airspace and the widespread application of low-altitude aircraft (such as drones and helicopters), low-altitude airspace modeling technology has gradually become one of the key technologies in aircraft path planning, flight control, and airspace management. In low-altitude airspace, aircraft are affected by multiple factors such as terrain, meteorological conditions (such as wind speed and air pressure), and building density. The dynamic changes of these factors are of great significance to the stability and path planning of aircraft.

[0003] However, existing low-altitude airspace modeling methods mostly employ static mesh generation techniques, typically using static data such as digital elevation models (DEMs) and building outlines. These methods suffer from insufficient accuracy and poor real-time responsiveness when dealing with dynamic changes in low-altitude airspace, especially in areas with drastic weather changes and complex terrain. Traditional methods often fail to adequately consider the hydrodynamic effects of airflow, turbulence, and wind speed changes in low-altitude airspace, thus affecting aircraft flight path planning and flight safety.

[0004] Therefore, how to accurately reflect the comprehensive impact of factors such as meteorological conditions, topographic relief, and building density on the low-altitude environment when modeling low-altitude airspace is an urgent problem to be solved in current low-altitude airspace modeling technology. Summary of the Invention

[0005] In view of the above-mentioned shortcomings mentioned in the background art, the purpose of this invention is to provide a method and platform for low-altitude airspace 3D modeling based on adaptive mesh and multi-source data fusion.

[0006] A first aspect of the present invention provides a method for three-dimensional modeling of low-altitude airspace based on adaptive mesh and multi-source data fusion, the method comprising the following steps: S1 acquires digital elevation model data, building outline data, and real-time meteorological data, and obtains multi-source fused data through spatiotemporal registration and fusion processing; S2, based on multi-source fusion data, uses a computational fluid dynamics model to simulate airflow and obtain the background airflow field distribution including wind speed vector, turbulence intensity and air pressure changes; S3. Based on the background airflow field distribution and the terrain and building density information extracted from the multi-source fusion data, perform initial grid division of the low-altitude airspace. S4. Under the initial grid division, the interaction between the aircraft and the background airflow is simulated based on the background airflow field distribution to obtain the aircraft-induced disturbance airflow field. The airflow disturbance area around the aircraft is identified according to the disturbance airflow field. The grid of the area is dynamically refined, and the grid of the area far away from the aircraft and with stable airflow is sparsed to generate a dynamic low-altitude airspace three-dimensional grid model. S5. Based on the dynamic low-altitude airspace three-dimensional grid model, the background airflow field distribution and the disturbed airflow field are fused to construct a low-altitude airspace three-dimensional model that includes the aircraft-airflow coupling effect.

[0007] A second aspect of the present invention provides an electronic device, comprising: a memory for storing instructions executable by one or more processors of the electronic device; and a processor for performing the method as described in any of the preceding claims.

[0008] A third aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in any of the preceding claims.

[0009] A fourth aspect of the present invention provides a computer program product comprising instructions that, when executed on an electronic device, cause the electronic device to perform the method described in any of the preceding claims.

[0010] Compared with existing technologies, this invention acquires and integrates digital elevation models, building outline data and meteorological data in real time, and combines them with computational fluid dynamics models to simulate airflow. This allows for accurate prediction of airflow changes and dynamic adjustment of grid resolution, thereby optimizing the modeling accuracy of low-altitude airspace.

[0011] Based on this, the method can not only improve the safety of aircraft path planning and prevent aircraft from entering unfavorable airflow areas, but also further refine the low-altitude airspace model through the coupling effect between the aircraft and the airflow, providing more accurate and real-time support for airspace management and flight safety control. Attached Figure Description

[0012] Figure 1 This is a schematic diagram of the overall process of a low-altitude airspace 3D modeling method based on adaptive mesh and multi-source data fusion disclosed in an embodiment of the present invention; Figure 2 This is a schematic diagram of the system architecture of the technical solution of the present invention; Figure 3 This is a schematic diagram of the structure of an electronic device disclosed in an embodiment of the present invention. Detailed Implementation

[0013] The following specific embodiments illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0014] The solution in this embodiment can be deployed on the server, edge computing node, or flight control system corresponding to the low-altitude flight mission planning platform to generate safe and recoverable flight paths for UAVs or other low-altitude aircraft. See also... Figure 2 As shown, this low-altitude flight mission planning platform integrates a 3D modeling system for low-altitude airspace and a flight path planning system. It can analyze the flight environment in real time and provide accurate flight path planning and optimization solutions. Specifically, the flight path generation process considers factors such as terrain, weather conditions, building density, and the interaction between the aircraft and airflow to ensure the aircraft's flight safety in complex environments.

[0015] Optionally, the low-altitude flight mission planning platform can also provide interfaces for 3D environment display, flight path planning, path risk assessment, and path optimization. Through these functions, operators can view a 3D model of the low-altitude airspace in real time, assess potential risks in the planned path, and dynamically optimize the flight path based on environmental changes. This provides aircraft with real-time path adjustment options to ensure flight safety and flexibility.

[0016] Furthermore, the technical features involved in the different embodiments of this application described below can be combined with each other as long as they do not conflict with each other.

[0017] Please see Figure 1 This invention provides a low-altitude airspace 3D modeling method based on adaptive mesh and multi-source data fusion. This method combines a fluid dynamics model with multi-source data fusion, computational fluid dynamics simulation, and aircraft-induced airflow disturbance simulation to dynamically adjust the low-altitude airspace mesh resolution, ensuring high-precision modeling and flight path planning in complex environments. The implementation steps and corresponding technical details are described in detail below.

[0018] S1 acquires digital elevation model data, building outline data, and real-time meteorological data, and obtains multi-source fused data through spatiotemporal registration and fusion processing; In this step, low-altitude airspace data from multiple sources are acquired, including Digital Elevation Model (DEM) data, building outline data, and real-time meteorological data, and then spatiotemporally registered and fused. To ensure data consistency and usability, the multi-source data needs to be cleaned, calibrated, and registered to eliminate data inconsistencies caused by time differences or incompatible spatial coordinate systems.

[0019] As an example, multi-source fused data is obtained through spatiotemporal registration and fusion processing, including: S11, the real-time meteorological data is time-synchronized according to a preset time interval, and missing meteorological data is completed by spatiotemporal interpolation to obtain a time-continuous meteorological dataset. Real-time meteorological data originates from weather stations, sensors mounted on drones, or satellite remote sensing. Their sampling frequencies may differ, and data gaps may occur. To ensure the continuity and integrity of meteorological data over time, all meteorological data is first synchronized according to a unified time benchmark. For example, a time synchronization interval of 5 minutes is set, and meteorological observations from various sources are aligned to this time grid using linear interpolation or cubic spline interpolation methods.

[0020] For missing data points, spatiotemporal interpolation methods can be used to complete them. Specifically, in the time dimension, linear interpolation is performed using observations from previous and subsequent times; in the spatial dimension, spatial interpolation can be performed by combining data from nearby meteorological stations using inverse distance weighted interpolation or Kriging interpolation. Through the above processing, a meteorological dataset with continuous time and complete spatial coverage is generated.

[0021] S12, unify the digital elevation model data, the building outline data, and the time-continuous meteorological dataset into the same spatial coordinate system, and perform spatial resolution matching; Multi-source data typically uses different spatial coordinate systems and resolutions. For example, DEM data uses the UTM projected coordinate system with a resolution of 10 meters; building outline data is in vector format under WGS-84 latitude and longitude coordinates; meteorological data is a NetCDF file with a 1-kilometer grid. To achieve spatial alignment, all data needs to be unified into the same spatial reference system, typically the WGS-84 geographic coordinate system or a plane coordinate system suitable for low-altitude projection. Specifically, coordinate transformation tools (such as the PROJ library) can be used to transform DEM and meteorological data to a coordinate system consistent with the building outlines.

[0022] Simultaneously, spatial resolution is matched; specifically, meteorological data is downscaled to the DEM resolution (e.g., 10 meters) using bilinear interpolation or nearest-neighbor interpolation methods, or the DEM and building data are resampled to a grid matching the meteorological data, selecting an appropriate resolution based on modeling requirements. To preserve terrain and building details, this embodiment preferably downscales the meteorological data to a 10-meter resolution.

[0023] S13, interpolate and fill the empty areas in the digital elevation model data, extract the building height and outline information from the building outline data and generate a three-dimensional building model; DEM data may contain voids caused by sensor occlusion or signal loss. To ensure the continuity of the terrain surface, these voids need to be filled. Common methods include inverse distance weighted interpolation based on surrounding elevation values, spline interpolation, or filling algorithms based on terrain slope aspect. For example, using inverse distance weighted interpolation, the elevation values ​​of each point within the void are calculated using distance weighting based on the elevation values ​​of the effective pixels surrounding the void, thus obtaining a complete and continuous digital elevation model.

[0024] For building outline data, the two-dimensional outline polygons of the buildings are extracted from remote sensing images or GIS databases, and the corresponding building height attributes are obtained (which can be obtained through LiDAR point clouds or surveying data). Then, using 3D modeling technology, the two-dimensional outlines are stretched vertically to the corresponding height to generate a 3D solid model of the building, which includes the geometric shape and spatial location information of each building.

[0025] S14, the matched digital elevation model data, the three-dimensional building model and the meteorological dataset are overlaid and fused to generate the multi-source fused data containing terrain, building and meteorological information.

[0026] After completing the above preprocessing, the DEM data, 3D building model, and meteorological dataset are overlaid and fused on a unified spatial grid. Specifically: Based on the terrain surface constructed from DEM data, a three-dimensional building model is superimposed on it to form a geometric model that includes the land surface and buildings. At the same time, parameters such as wind speed, wind direction, air pressure, and temperature from the meteorological dataset are interpolated into each grid node as a three-dimensional spatial field, thereby realizing the fusion of terrain, buildings and meteorological information.

[0027] For example, at a certain grid point (x, y, z), both the identifier of whether the point belongs to a building or terrain surface and the meteorological parameter value of the point are stored, which is convenient for subsequent calculations.

[0028] S2, based on multi-source fusion data, uses a computational fluid dynamics model to simulate airflow and obtain the background airflow field distribution including wind speed vector, turbulence intensity and air pressure changes; In this step, based on the multi-source fusion data generated above, a computational fluid dynamics (CFD) model is used to numerically simulate the airflow motion in the low-altitude airspace to obtain the background airflow field distribution that reflects the influence of terrain and buildings, providing a reference airflow environment for subsequent mesh generation and aircraft disturbance simulation.

[0029] As an example, based on multi-source fusion data, a computational fluid dynamics model is used to simulate airflow, obtaining the background airflow field distribution that includes wind speed, wind direction, turbulence intensity, and air pressure changes, including: S21. Based on the digital elevation model data in the multi-source fusion data and the three-dimensional building model, construct a three-dimensional geometric model of the low-altitude airspace and set the boundary conditions of the computational fluid dynamics model. A terrain surface grid is generated from DEM data in multi-source fusion data. DEM data typically provides elevation values ​​of ground points in the form of a regular grid, which can be connected to form a triangular or quadrilateral grid to create a geometric representation of the terrain.

[0030] Next, the 3D building model generated in step S13 is overlaid on the generated terrain surface mesh, embedding the building as an independent geometry into the terrain surface. Specifically, the 3D building model and the terrain surface are Boolean merged to ensure a seamless fit between the building's base and the terrain surface. The final result is a complete 3D geometric model containing surface undulations and building obstacles, covering the low-altitude airspace to be simulated, for example, 5km × 5km horizontally and from the ground to a height of 500m vertically.

[0031] Based on this, boundary conditions for the CFD simulation are set to ensure the physical rationality of the numerical solution. Specifically: Inlet boundary: The upwind surface of the computational domain is designated as the velocity inlet, and the inlet wind speed profile is given based on the prevailing wind direction and speed from real-time meteorological data. In the actual atmospheric boundary layer, wind speed variation with height typically follows an exponential or logarithmic law, for example, using an exponential law formula. ,in For reference height Wind speed at a location (e.g., 10m) This represents the ground roughness index (values ​​are assigned based on terrain type, e.g., 0.14 for open terrain, 0.25 for urban areas). Simultaneously, the turbulent kinetic energy at the entrance... and turbulent dissipation rate It can be given according to empirical formulas, such as , ,in For turbulence intensity, For turbulent scales; This is an empirical constant, for example, a value of 0.09.

[0032] Outlet boundary: Set the downwind side of the computational domain as the pressure outlet, set the relative static pressure to 0, and specify the recirculation conditions when recirculation is allowed.

[0033] Top and side boundaries: can be set as symmetrical boundaries or sliding walls to simulate an infinitely large region and avoid boundary reflections from interfering with the internal flow field.

[0034] Ground and building surfaces: These are set as non-slip walls, and the wall function parameters are set according to the surface roughness. For building surfaces, they are usually treated as smooth walls; for ground surfaces, the equivalent sand grain roughness height and roughness constant can be specified to reflect the frictional effects of different surfaces.

[0035] Through the above steps, a geometric model and boundary conditions suitable for CFD solutions are established.

[0036] S22, using real-time meteorological data from multi-source fusion data as initial conditions and input parameters, and based on a three-dimensional geometric model, by solving the Navier-Stokes equations and combining them with a turbulence model, the wind speed vector, turbulence intensity, and air pressure changes at various spatial locations in the low-altitude airspace are calculated to form the background airflow field distribution.

[0037] After completing the 3D geometric modeling and boundary condition setting, the wind speed, wind direction, air pressure, temperature, and other parameters in the time-continuous meteorological dataset generated in step S11 are used as initial conditions to initialize the flow field of the entire computational domain. For example, the entire computational domain can be initialized as a uniform wind speed field (the magnitude of which is the average inlet wind speed) and a static pressure field, or full-field interpolation can be performed based on the inlet profile.

[0038] Subsequently, the computational domain is discretized and solved using the finite volume method. First, the geometric model is divided into a finite number of control volumes (grid elements). The fluid motion control equations are integrated on each element, transforming them into a system of algebraic equations. During the solution process, the inlet boundary conditions can be dynamically updated based on the temporal changes in meteorological data to simulate unsteady airflow variations. For example, if the meteorological data is updated every 5 minutes, a time step of 0.1 seconds can be set. Iterative solutions are performed within each time step, and the inlet boundary conditions are adjusted based on the new meteorological data every 3000 time steps (corresponding to 5 minutes), continuing the calculation until the required time window (e.g., 30 minutes) is covered. It is understandable that this unsteady simulation can capture the evolution of airflow in complex terrain and building complexes, reflecting the dynamic characteristics of airflow under real meteorological conditions.

[0039] Understandably, when choosing a solver, one can use open-source CFD software (such as OpenFOAM) or commercial software (such as ANSYS Fluent), and set appropriate discretization schemes (such as second-order upwind schemes) and pressure-velocity coupled algorithms (such as SIMPLE algorithm or PISO algorithm) to ensure computational accuracy and stability.

[0040] In the solution process, the core governing equations are the Navier-Stokes equations, which describe the conservation of mass, momentum, and energy of the fluid. For atmospheric flow in the low-altitude airspace, the thermal effects in the energy equation can usually be ignored, and the focus is on solving the continuity and momentum equations.

[0041] Since airflow in low-altitude airspace is typically turbulent, directly solving the Navier-Stokes equations (Direct Numerical Simulation, DNS) is computationally too costly to meet engineering application requirements. Therefore, this embodiment preferably uses the Reynolds-Averaged Navier-Stokes (RANS) method to simulate turbulence, by introducing a turbulence model to close the Reynolds stress term. Specifically, the standard k-ε two-equation turbulence model is used, which simulates turbulence effects by solving the transport equations for turbulent kinetic energy k and turbulent dissipation rate ε. For the near-wall region, the standard wall function is used to reduce the requirement for near-wall mesh density while ensuring simulation accuracy.

[0042] The above system of equations is solved iteratively until the residuals converge (e.g., the residuals of all variables decrease to 10). -4 The following steps are used to obtain the wind speed vector components (u, v, w), pressure p, turbulent kinetic energy k, and turbulent dissipation rate ε for each grid cell. Based on this, the turbulence intensity is further calculated. It is defined as the ratio of the root mean square of the fluctuating velocity to the average wind speed, and can be estimated by the value of k: ,in This represents the local average wind speed. Pressure changes are usually expressed as relative pressure, which is the pressure difference relative to a reference atmospheric pressure.

[0043] Ultimately, three-dimensional distribution data of the background airflow field covering the entire low-altitude airspace was obtained, including: wind speed vector. turbulence intensity air pressure change This data field is stored centered around grid nodes or grid cells. For example, at a certain grid point (x, y, z), the specific wind speed value (e.g., 5.2 m / s, southeast direction), turbulence intensity (e.g., 12%), and pressure value relative to standard atmospheric pressure (e.g., -15 Pa) can be obtained, thus comprehensively reflecting the airflow characteristics at that location.

[0044] S3. Based on the background airflow field distribution and the terrain and building density information extracted from the multi-source fusion data, perform initial grid division of the low-altitude airspace. In this step, the initial mesh generation aims to match the mesh resolution with the intensity of airflow changes, the degree of terrain undulation, and the density of buildings. That is, a finer mesh is used in areas with drastic airflow changes, complex terrain, or dense buildings, while a coarser mesh is used in areas with stable airflow, flat terrain, and sparse buildings, thereby improving computational efficiency while ensuring simulation accuracy.

[0045] As an example, based on the background airflow field distribution and the terrain and building density information extracted from multi-source fusion data, an initial grid division is performed on the low-altitude airspace, including: S31, extract the wind speed gradient and turbulence intensity gradient from the background airflow field distribution as characteristic parameters of airflow change; First, the background airflow field is post-processed to calculate the wind speed gradient magnitude and turbulence intensity gradient magnitude at each grid point, in order to quantify the intensity of airflow changes in space.

[0046] Specifically, for any point in space Its wind speed vector The wind speed gradient tensor is Its magnitude can be represented using the Frobenius norm of a second-order tensor: .

[0047] In practical calculations, partial derivatives can be approximated using the finite difference method, such as the central difference scheme. Similarly, the magnitude of the turbulence intensity gradient... A similar calculation method can be used, that is, for The amplitudes were calculated by taking the partial derivatives in the three directions. To facilitate subsequent fusion processing, the calculated wind speed gradient amplitude and turbulence intensity gradient amplitude were normalized to the [0,1] interval, resulting in normalized airflow change characteristic parameters: the amplitude of the wind speed gradient. and the magnitude of the turbulence intensity gradient .

[0048] S32, using the terrain information extracted from the multi-source fusion data to calculate the standard deviation of terrain elevation as a terrain complexity index, and using the building outline data to calculate the building coverage per unit area as a building density index. To account for terrain complexity, the entire low-altitude airspace is divided into several regular sub-regions (e.g., 100m × 100m grids) in the horizontal direction. Within each sub-region, the standard deviation of the elevation values ​​of all grid points in the multi-source fused data is calculated. A larger standard deviation indicates more dramatic terrain undulations and higher terrain complexity. Normalizing this standard deviation to the [0,1] interval yields the terrain complexity index. For different height layers within the same sub-region, this index can remain constant along the vertical direction or be appropriately attenuated according to the vertical height (e.g., decrease with increasing height) to reflect the diminishing effect of terrain on the upper atmosphere.

[0049] Regarding building density, within the same horizontal sub-region, the proportion of the projected area of ​​buildings to the total area of ​​the sub-region is calculated based on a 3D building model. This refers to building coverage. A higher building coverage ratio indicates a denser building density, resulting in more significant obstruction and disturbance of airflow. Normalized to an interval, the building density index is obtained. Similarly, appropriate attenuation can be applied along the vertical direction to reflect the decreasing characteristics of the building's impact on the upper floors.

[0050] S33, by combining airflow change characteristic parameters, terrain complexity index and building density index, the grid refinement priority coefficient is obtained through analysis; To more accurately reflect the comprehensive requirements of multi-factor coupling for mesh refinement, this embodiment uses a nonlinear coupling method to construct the mesh refinement priority coefficient.

[0051] As an example, by combining airflow variation characteristics, terrain complexity, and building density, the mesh refinement priority coefficients are analyzed and obtained, including: S331, based on airflow change characteristic parameters, terrain complexity index, and building density index, respectively construct airflow change influence field function, terrain complexity influence field function, and building density influence field function; wherein, each of the influence field functions takes spatial location as independent variable, and respectively characterizes the distribution characteristics of the intensity of the influence of a single factor on grid refinement as distance decreases; Each influence field function uses spatial location as the independent variable to characterize the influence intensity of a single factor on mesh refinement, and the influence intensity decreases as the distance from the center point of the factor increases.

[0052] The characteristic parameters of airflow variation can be viewed as a set of source points, each corresponding to a location with a high airflow gradient. A Gaussian influence field function is constructed centered on each source point: ;in, For the first The spatial coordinates of the source point These are the normalized airflow variation characteristic parameters at this point. The attenuation radius (can be set empirically, for example, 50m). The influence fields of all source points are superimposed to obtain the total airflow change influence field function: .

[0053] Similarly, based on terrain complexity index and building density index Constructing the field function of terrain complexity influence and the influence field function of building density Each sub-region can be considered as a source point, and the influence intensity is taken from the terrain complexity index of that sub-region. and building density index And spatial diffusion is performed using the same Gaussian function, as follows: For the There are 1 horizontal sub-regions, and the coordinates of their center points are given as follows: Then the influence function of terrain complexity is the sum of the contributions of each sub-region: ;in The attenuation radius of the terrain influence can be set according to the scale of the terrain features (e.g., 1-2 times the size of the sub-region).

[0054] The influence function of building density is: ;in This represents the attenuation radius of the building's influence. The above function also follows Gaussian attenuation in the vertical direction, meaning the influence intensity decreases the farther away from the center of the sub-region.

[0055] S332 nonlinearly couples the field function of airflow change, the field function of terrain complexity, and the field function of building density, and obtains the grid refinement priority coefficient for each spatial location by solving the multi-factor coupling equation; wherein, the multi-factor coupling equation introduces the interaction term between the factors to reflect the combined effect of the enhancement or attenuation of airflow in areas with drastic airflow changes due to the presence of terrain and buildings.

[0056] To reflect the combined effects of enhanced or diminished airflow due to topography and buildings in areas of dramatic airflow change, this embodiment introduces a nonlinear coupling equation to fuse the three influencing field functions mentioned above. A preferred coupling method is product-type coupling, which introduces an interaction term: ;in, and This is the coupling strength coefficient, used to adjust the amplification effect of topographic and architectural factors on airflow variations. For example, if a region has a large airflow gradient and also has complex topography or dense architecture, then... or The larger the priority coefficient, the better the coupling. This significantly increases the size, allowing for the use of finer grids in that area. Conversely, if the airflow gradient is large but the terrain is flat and there are no buildings, then... Only by The decision was made that the mesh refinement level is relatively low.

[0057] Through this nonlinear coupling calculation, the mesh refinement priority coefficients for each spatial location within the entire low-altitude airspace are obtained. Understandably, this coefficient comprehensively reflects the combined needs of airflow variations, terrain complexity, and building density for grid refinement.

[0058] S34. Based on the grid refinement priority coefficient, the low-altitude airspace is divided into different levels of grid refinement regions to complete the initial grid division.

[0059] Multiple priority thresholds can be set to divide the low-altitude airspace into different levels of grid refinement regions. For example, three levels can be set: Level 1 Refinement Area (High Priority): This area uses the finest grid size, for example... ; Secondary refinement area (medium priority): This area uses a medium grid size, for example ; Level 3 Refinement Area (Low Priority): This area uses a coarser grid size, for example... .

[0060] In the specific partitioning process, an octree mesh generation method can be used to dynamically adjust the subdivision level of the mesh cells based on the priority coefficient of each region, ensuring that the mesh resolution matches the priority coefficient. For example, in regions with high priority coefficients, the mesh cells are recursively subdivided until the target mesh size is reached; in regions with low priority coefficients, a coarser mesh level is maintained.

[0061] S4. Under the initial grid division, the interaction between the aircraft and the background airflow is simulated based on the background airflow field distribution to obtain the aircraft-induced disturbance airflow field. The airflow disturbance area around the aircraft is identified according to the disturbance airflow field. The grid of the area is dynamically refined, and the grid of the area far away from the aircraft and with stable airflow is sparsed to generate a dynamic low-altitude airspace three-dimensional grid model. In this step, based on the initial mesh generation and background airflow distribution, the interaction between the aircraft and the background airflow is simulated to obtain the aircraft-induced perturbation airflow field, and the mesh is dynamically adjusted accordingly. This process aims to capture the perturbation effect of the aircraft on the surrounding airflow during operation, automatically refine the mesh in areas of strong perturbation to improve simulation accuracy, and appropriately sparse the mesh in areas far from the aircraft and with stable airflow to save computational resources, thereby generating an adaptive mesh model that can dynamically reflect the aircraft-airflow coupling effect.

[0062] As an example, the interaction between the aircraft and the background airflow is simulated based on the background airflow field distribution to obtain the aircraft-induced disturbed airflow field. The disturbed airflow field is then used to identify the airflow disturbance region around the aircraft, including: S41, Construct an aircraft model based on the aircraft's geometric parameters and flight state parameters, and determine the spatial position of the aircraft model within the initial mesh division; A 3D model is constructed based on the actual geometry of the aircraft to be simulated. The geometric parameters of the aircraft include fuselage dimensions, wingspan, wing shape, fuselage length, and tail configuration. For example, for a certain type of quadcopter UAV, its geometric parameters might include: fuselage diameter 0.5m, arm length 0.3m, rotor diameter 0.2m, and an overall cross-shaped symmetrical structure. These parameters can be obtained from aircraft design drawings, 3D scanning data, or standard model libraries.

[0063] Simultaneously, the flight status parameters of the aircraft are acquired, including its current position coordinates, flight speed, attitude angles (pitch, roll, yaw), and rotor speed. These parameters can be acquired in real time through the flight control system or obtained by interpolation based on a preset flight path.

[0064] For example, setting the location of the spacecraft at a certain moment at coordinates With speed Fly horizontally in a due east direction, maintaining a horizontal attitude.

[0065] Based on the aforementioned geometric and flight state parameters, the specific spatial position and orientation of the aircraft model are determined during the initial mesh generation. In practice, the aircraft model can be embedded into the initial mesh as a 3D surface mesh, with the aircraft surface serving as the new wall boundary condition. The mesh region surrounding the aircraft will be marked as the region of interest for subsequent perturbation simulations. If the aircraft position changes over time (e.g., during flight), its position within the mesh needs to be dynamically updated according to the time step to simulate the moving boundary.

[0066] S42, based on the background airflow field distribution under the initial grid division, the computational fluid dynamics solution is performed on the flow field containing the aircraft model to simulate the disturbance of the background airflow on the surface of the aircraft and its surrounding area, and the velocity increment and pressure increment of each spatial point relative to the background airflow are obtained to form the disturbance airflow field induced by the aircraft. After embedding the aircraft model, the background airflow field distribution was used as the initial flow field condition. Based on the initial mesh generation, the entire flow field containing the aircraft model was solved again using CFD. The core objective of this solution was to capture the disturbance effect of the aircraft on the background airflow, including phenomena such as the obstruction of airflow by the aircraft surface, the induced velocity generated by rotor rotation, and wingtip vortices.

[0067] The solution process still employs the Navier-Stokes equations combined with a turbulence model, but requires additional treatment of the aircraft surface boundary conditions. The aircraft surface is set as a no-slip wall boundary. For moving components such as rotors, a multiple reference frame (MRF) model or a slip mesh model can be used to simulate their rotational effects. The solver settings are similar to those in step S22, but smaller iteration steps or higher convergence criteria can be used for local regions around the aircraft to ensure the accuracy of the perturbation field calculation.

[0068] By iteratively solving the problem, the velocity increment of each grid cell relative to the background airflow field is obtained. and pressure increment The velocity increment is defined as the vector difference between the current flow field wind speed and the background airflow field wind speed. ,in Background wind speed excluding aircraft. Pressure increment. It can be defined as the difference between total pressure and background pressure.

[0069] These incremental data constitute the perturbation airflow field induced by the aircraft. For example, directly in front of the aircraft, due to airflow obstruction, positive pressure increments (increased pressure) and decreased velocity (negative velocity increments) may occur; behind the aircraft, a wake region may form, manifested as negative pressure increments and velocity fluctuations; below the rotor, due to the rotor downwash, downward velocity increments and positive pressure increments may occur. This perturbation airflow field is stored in the form of a three-dimensional field, containing the velocity increment vector and pressure increment scalar for each grid point.

[0070] S43, extract spatial points from the disturbed airflow field whose velocity increment exceeds a preset disturbance threshold, and identify the area covered by these spatial points as the airflow disturbance area around the aircraft.

[0071] After obtaining the disturbed airflow field, it is necessary to quantitatively identify the areas where the aircraft significantly affects the surrounding airflow in order to dynamically adjust the grid. For this purpose, a disturbance threshold is set. It is used to determine whether the velocity increment is significant.

[0072] The disturbance threshold can be set based on the characteristics of the background airflow field and the type of aircraft. A preferred setting method is a relative threshold, which is a certain proportion of the background wind speed. For example, setting... In other words, when the velocity increment at a point exceeds 10% of the local background wind speed, that point is considered to be significantly disturbed by the aircraft. For areas with low background wind speeds (such as less than 1 m / s), an absolute threshold can be used as a supplement, for example, setting the absolute threshold to 0.1 m / s.

[0073] The specific identification process is as follows: traverse all grid points and calculate the velocity increment amplitude at each point. If this value is greater than the disturbance threshold If a point is identified as a disturbance point, it is marked as such. The connected region formed by all disturbance points constitutes the airflow disturbance region around the aircraft. This region is typically distributed in three-dimensional space and may include the compression zone in front of the aircraft, the surrounding flow zone above it, the wake zone behind it, and the downwash zone below the rotor, etc.

[0074] After identifying the disturbance region, the mesh is dynamically adjusted based on this region. Specifically, the mesh within the disturbance region is dynamically refined to improve the simulation accuracy of aircraft-airflow interaction; the mesh in areas far from the aircraft and with stable airflow (i.e., areas not marked as disturbance points) is thinned to save computational resources. For example, the mesh size within the disturbance region can be refined to 1 / 2 or 1 / 4 of the initial mesh size, while the mesh size in the stable region can be widened to twice the initial mesh size. Through this dynamic adjustment, a dynamic low-altitude airspace 3D mesh model that can adapt to the aircraft position and disturbance range is generated.

[0075] S5. Based on the dynamic low-altitude airspace three-dimensional grid model, the background airflow field distribution and the disturbed airflow field are fused to construct a low-altitude airspace three-dimensional model that includes the aircraft-airflow coupling effect.

[0076] In this step, a dynamic low-altitude airspace 3D mesh model is used as the basis for spatial discretization. The background airflow field distribution and the aircraft-induced disturbance airflow field are fused to construct a low-altitude airspace 3D model that can fully reflect the interaction between the aircraft and the airflow. It can be understood that this fused model integrates natural environmental factors such as terrain, buildings, and meteorological conditions, as well as the disturbance effects generated by aircraft operation, providing accurate data support for aircraft path planning, flight stability analysis, and airspace management.

[0077] As an example, based on the dynamic low-altitude airspace three-dimensional mesh model, the background airflow field distribution and the disturbed airflow field are fused to construct a low-altitude airspace three-dimensional model that includes the aircraft-airflow coupling effect, including: S51, using the dynamic low-altitude airspace three-dimensional grid model as the basis for spatial discretization, the background airflow field distribution and the disturbed airflow field are correlated at the same spatial coordinate points; First, a dynamic low-altitude airspace 3D mesh model is used as a unified spatial framework. After dynamic adjustment, this mesh model has a high mesh resolution in the airflow disturbance region around the aircraft, while the mesh is coarser in the stable region far away from the aircraft. This allows for accurate description of key areas of aircraft-airflow interaction at a reasonable computational cost.

[0078] Based on this, the background airflow field distribution and the disturbed airflow field are interpolated to each grid node or grid cell center of the dynamic grid model, respectively. Since the background airflow field and the disturbed airflow field were initially calculated under different grid systems (i.e., the background airflow field is based on the initial grid, and the disturbed airflow field is based on the grid after embedding the aircraft), spatial correspondence processing is required. Specifically, a trilinear interpolation method can be used to interpolate the wind speed vector in the background airflow field. and air pressure Interpolate to the nodes of the dynamic grid; similarly, interpolate the velocity increments in the perturbed airflow field. and pressure increment It is also interpolated to the same node position. In this way, it ensures that at each spatial coordinate point... It simultaneously stores background airflow field data and disturbed airflow field data.

[0079] For example, for a node in a dynamic mesh The background wind speed at that point was obtained by interpolation. The background air pressure relative value is At the same time, the velocity increment at that point is obtained as The pressure increment is .

[0080] S52, for each grid cell or node, the wind speed vector in the background airflow field distribution is vector-superimposed with the velocity increment in the disturbed airflow field, and the pressure change in the background airflow field distribution is algebraically superimposed with the pressure increment in the disturbed airflow field to obtain the fused total wind speed vector field and total pressure field. After completing the data mapping, the physical field superposition calculation is performed for each grid node or grid cell center. Since wind speed is a vector field, vector addition is required for superposition; while air pressure is a scalar field, algebraic addition is sufficient.

[0081] For the wind speed field, the total wind speed vector It is obtained by superimposing the background wind speed vector and the disturbance flow velocity increment vector: , Therefore, the total wind speed at this node is: , That is, the total wind speed is approximately The direction was deflected to some extent, reflecting the change in the original airflow caused by the aircraft-induced disturbance.

[0082] For the pressure field, the total pressure It is obtained by superimposing the background pressure and the disturbance pressure increment: .

[0083] The total air pressure at this node is: This value reflects the pressure change at that point relative to the reference atmospheric pressure under the influence of aircraft disturbances.

[0084] By traversing all nodes in the dynamic mesh model and performing the vector superposition and algebraic superposition as described above, the total wind speed vector field covering the entire low-altitude airspace can be obtained. and total pressure field These two fields together constitute a complete flow field description that includes the aircraft-airflow coupling effect.

[0085] S53. Based on the dynamic low-altitude airspace three-dimensional grid model and the total wind speed vector field and total air pressure field, a low-altitude airspace three-dimensional model including the aircraft-airflow coupling effect is constructed.

[0086] After obtaining the total wind speed vector field and total pressure field, they are associated and stored with a dynamic low-altitude airspace 3D mesh model to construct the final low-altitude airspace 3D model that includes the aircraft-airflow coupling effect. This model uses a 3D mesh as its spatial framework, and each mesh node or mesh cell stores the following core physical quantities: Spatial location coordinates Total wind speed vector Total air pressure Furthermore, it can also include turbulence intensity (which can be estimated by combining total wind speed and background turbulence), wind speed gradient, etc. In addition, the model can retain separation information between the background airflow field and the disturbed airflow field to facilitate subsequent analysis of the specific contribution of aircraft disturbances. For example, additional storage is possible. , , , Fields such as...

[0087] Understandably, this low-altitude airspace 3D model can be directly applied to: Aircraft path planning: When planning the flight path, the total wind speed field provided by the model is used as input to calculate the airflow forces on the aircraft at different path points, evaluate the safety and energy efficiency of the path, and avoid areas with strong turbulence or unfavorable wind direction.

[0088] Flight stability analysis: By extracting the velocity gradient and air pressure distribution in the local area around the aircraft, the changes in the aerodynamic forces acting on the aircraft are analyzed, and its attitude stability in turbulent airflow is evaluated.

[0089] Airspace management and risk assessment: Based on the model, complex airflow areas in low-altitude airspace are identified, providing decision support for airspace delineation, traffic management and flight conflict resolution; at the same time, it can be combined with multi-aircraft scenarios to analyze the superposition effect of multi-aircraft disturbance fields and assess system-level operational risks.

[0090] For example, in drone logistics delivery missions, the model can be used to simulate the airflow environment along the flight path, identify the acceleration effect of turbulent vortices in the leeward area of ​​buildings or narrow channels, and dynamically adjust the flight altitude or route to ensure flight safety.

[0091] Please see Figure 3 The present invention also provides an electronic device, comprising: a memory for storing instructions executed by one or more processors of the electronic device; and a processor for performing the method as described in any of the preceding embodiments.

[0092] The processor executes program instructions stored in its memory to perform specific operations on the low-altitude airspace 3D modeling method, including multi-source data fusion processing, airflow simulation using computational fluid dynamics models, dynamic adjustment of mesh generation, simulation of the interaction between the aircraft and airflow, and path planning. This electronic device can respond to environmental changes in real time, providing accurate low-altitude airspace modeling and flight path planning support.

[0093] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in any of the preceding embodiments.

[0094] By loading the computer program into a storage medium, the method of the present invention can be executed on different devices, ensuring that dynamic modeling of low-altitude airspace and aircraft path planning can be achieved in a variety of application scenarios.

[0095] Continue reading Figure 3 The present invention also provides a computer program product, the computer program product including instructions, which, when executed on an electronic device, cause the electronic device to perform the method described in any of the preceding embodiments.

[0096] This program can be easily deployed on various electronic devices, including personal computers, servers, embedded systems, etc. By executing corresponding instructions, it realizes the technical solution of this invention and provides efficient technical support for low-altitude airspace modeling and flight path planning.

[0097] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for low-altitude airspace 3D modeling based on adaptive mesh and multi-source data fusion, characterized in that, The method includes the following steps: S1, acquiring digital elevation model data, building outline data, and real-time meteorological data, and obtaining multi-source fused data through spatiotemporal registration and fusion processing; S2, based on the multi-source fused data, performing airflow simulation using a computational fluid dynamics model to obtain a background airflow field distribution including wind speed vectors, turbulence intensity, and air pressure changes; S3, performing initial grid division of the low-altitude airspace based on the background airflow field distribution and the terrain information and building density information extracted from the multi-source fused data; S4, under the initial grid division, simulating the interaction between the aircraft and the background airflow based on the background airflow field distribution to obtain the aircraft-induced disturbance airflow field, identifying the airflow disturbance area around the aircraft based on the disturbance airflow field, dynamically refining the grid in the area, and sparsening the grid in the area far from the aircraft and with stable airflow, generating a dynamic low-altitude airspace three-dimensional grid model; S5, based on the dynamic low-altitude airspace three-dimensional grid model, fusing the background airflow field distribution with the disturbance airflow field to construct a low-altitude airspace three-dimensional model including the aircraft-airflow coupling effect; Based on the background airflow field distribution and the terrain and building density information extracted from the multi-source fusion data, the low-altitude airspace is initially gridded, including: S31, extracting wind speed gradient and turbulence intensity gradient from the background airflow field distribution as airflow change characteristic parameters; S32, calculating the terrain elevation standard deviation as a terrain complexity index using the terrain information extracted from the multi-source fusion data, and calculating the building coverage rate per unit area as a building density index based on building outline data; S33, analyzing and obtaining the grid refinement priority coefficient by combining the airflow change characteristic parameters, terrain complexity index, and building density index; S34, dividing the low-altitude airspace into different levels of grid refinement regions according to the grid refinement priority coefficient, thus completing the initial grid division; By comprehensively analyzing airflow change characteristic parameters, terrain complexity index, and building density index, the mesh refinement priority coefficient is obtained, including: S331, constructing airflow change influence field function, terrain complexity influence field function, and building density influence field function based on airflow change characteristic parameters, terrain complexity index, and building density index, respectively; wherein, each influence field function takes spatial location as independent variable, and respectively characterizes the distribution characteristics of the intensity of influence of a single factor on mesh refinement attenuating with distance; S332, nonlinearly coupling the airflow change influence field function, terrain complexity influence field function, and building density influence field function, and obtaining the mesh refinement priority coefficient for each spatial location by solving the multi-factor coupling equation; wherein, the multi-factor coupling equation introduces the interaction term between the factors to reflect the combined effect of enhancement or attenuation caused by the presence of terrain and buildings in areas with drastic airflow changes.

2. The low-altitude airspace 3D modeling method based on adaptive mesh and multi-source data fusion according to claim 1, characterized in that: Multi-source fused data is obtained through spatiotemporal registration and fusion processing, including: S11, performing time synchronization processing on the real-time meteorological data according to a preset time interval, and performing spatiotemporal interpolation to complete the missing meteorological data to obtain a time-continuous meteorological dataset; S12, unifying the digital elevation model data, the building outline data, and the time-continuous meteorological dataset to the same spatial coordinate system, and performing spatial resolution matching; S13, interpolating and filling the void areas in the digital elevation model data, extracting building height and outline information from the building outline data, and generating a three-dimensional building model; S14, overlaying and fusing the matched digital elevation model data, the three-dimensional building model, and the meteorological dataset to generate the multi-source fused data containing terrain, building, and meteorological information.

3. The low-altitude airspace 3D modeling method based on adaptive mesh and multi-source data fusion according to claim 2, characterized in that: Based on multi-source fusion data, airflow simulation is performed using a computational fluid dynamics model to obtain the background airflow field distribution, which includes changes in wind speed, wind direction, turbulence intensity, and air pressure. This includes: S21, constructing a three-dimensional geometric model of the low-altitude airspace based on the digital elevation model data and the three-dimensional building model from the multi-source fusion data, and setting the boundary conditions of the computational fluid dynamics model; S22, using real-time meteorological data from the multi-source fusion data as initial conditions and input parameters, and based on the three-dimensional geometric model, calculating the wind speed vector, turbulence intensity, and air pressure changes at various spatial locations in the low-altitude airspace by solving the Navier-Stokes equations and combining them with the turbulence model, thus forming the background airflow field distribution.

4. The low-altitude airspace 3D modeling method based on adaptive mesh and multi-source data fusion according to claim 1, characterized in that: The interaction between the aircraft and the background airflow is simulated based on the background airflow field distribution to obtain the aircraft-induced disturbance airflow field. The disturbance airflow field is used to identify the airflow disturbance region around the aircraft, including: S41, constructing an aircraft model based on the aircraft's geometric parameters and flight state parameters, and determining the spatial position of the aircraft model in the initial grid division; S42, performing computational fluid dynamics on the flow field containing the aircraft model under the initial grid division based on the background airflow field distribution, simulating the disturbance of the background airflow on the aircraft surface and its surrounding area, obtaining the velocity increment and pressure increment of each spatial point relative to the background airflow, forming the aircraft-induced disturbance airflow field; S43, extracting spatial points whose velocity increment exceeds a preset disturbance threshold from the disturbance airflow field, and identifying the area covered by these spatial points as the airflow disturbance region around the aircraft.

5. The low-altitude airspace 3D modeling method based on adaptive mesh and multi-source data fusion according to claim 1, characterized in that: Based on the dynamic low-altitude airspace three-dimensional mesh model, the background airflow field distribution and the disturbed airflow field are fused to construct a low-altitude airspace three-dimensional model containing the aircraft-airflow coupling effect, including: S51, using the dynamic low-altitude airspace three-dimensional mesh model as the basis for spatial discretization, the background airflow field distribution and the disturbed airflow field are mapped to the same spatial coordinate points; S52, for each mesh cell or node, the wind speed vector in the background airflow field distribution and the velocity increment in the disturbed airflow field are vector-superimposed, and the pressure change in the background airflow field distribution and the pressure increment in the disturbed airflow field are algebraically superimposed to obtain the fused total wind speed vector field and total pressure field; S53, based on the dynamic low-altitude airspace three-dimensional mesh model and the total wind speed vector field and total pressure field, a low-altitude airspace three-dimensional model containing the aircraft-airflow coupling effect is constructed.

6. An electronic device, characterized in that, include: Memory, used to store instructions executed by one or more processors of the electronic device; And a processor for performing the method as described in any one of claims 1 to 5.

7. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 5.

8. A computer program product, characterized in that, The computer program product includes instructions that, when executed on an electronic device, cause the electronic device to perform the method as described in any one of claims 1 to 5.

Images