Building facade unmanned aerial vehicle takeoff and landing platform layout decision-making method and system

Abstract

The invention discloses a building facade unmanned aerial vehicle takeoff and landing platform layout decision-making method and system. The method comprises the steps of obtaining BIM, facade images, wind field simulation and internal demand data of a building; grid candidate points are generated on the facade based on BIM, and screening is carried out according to the structure and airspace security constraints; performing quantitative evaluation on the wind environment stability, the structural bearing capacity and the user demand responsivity on the candidate points; an optimization model with the weighted sum of three scores as a target is constructed, solving is carried out under the constraints of the number of platforms and the airspace, and a Pareto optimal scheme set is output; multi-factor coupling analysis is carried out through a structure bearing, wind environment adaptability and demand responsivity evaluation model, a Pareto optimal layout scheme set is formed through a multi-target optimization algorithm, and automatic and scientific decision-making of unmanned aerial vehicle take-off and landing platform site selection on a complex building elevation is achieved; and the safety, economy and applicability of deployment of the low-altitude economic infrastructure on the building scale are remarkably improved.

Description

Technical Field

[0001] This invention relates to the planning of smart buildings and low-altitude economic infrastructure, and more particularly to a method and system for making decisions on the layout of unmanned aerial vehicle (UAV) landing platforms on building facades. Background Technology

[0002] With the booming development of the urban low-altitude economy, the application of drones in logistics, inspection, and emergency response is deepening, creating an urgent need for high-density, networked drone take-off and landing site infrastructure in urban airspace. Currently, these sites are mainly planned on building rooftops or open ground areas, and their site selection technology relies on macro-geographic information and oblique photogrammetry modeling, focusing on the assessment of urban-scale factors such as airspace safety and noise impact. However, this model is difficult to effectively utilize the vast vertical facade space resources of existing mid-to-high-rise buildings, and it cannot respond to the future demand for distributed, high-frequency, and individualized drone services. When attempting to deploy take-off and landing platforms on the complex three-dimensional vertical interface of building facades, existing technical solutions reveal significant limitations: First, their data foundation relies on large-scale scanning of rooftops or ground surfaces, resulting in insufficient accuracy in identifying finely distributed obstacles such as air conditioning units, billboards, and window frames on building facades, failing to meet the stringent requirements of centimeter-level safety clearance. Second, a severe lack of decision-making dimensions ignores crucial engineering feasibility factors at the building's microscale, such as the compatibility of the platform installation point with the building's internal load-bearing structure, the turbulent stability of the local wind field on the facade, and the service efficiency matching the actual functional zoning within the building. These limitations render existing planning methods unsuitable for facade scenarios. Forcibly applying them would expose the layout scheme to extremely high safety risks, difficulties in engineering implementation, or low post-construction efficiency, severely restricting the safe, efficient, and rational development and utilization of urban vertical space resources. Summary of the Invention

[0003] Purpose of the invention: The purpose of this invention is to provide an automated, refined, and scientific method for making decisions on the layout of unmanned aerial vehicle (UAV) landing platforms for building facades; on the other hand, it provides a system for making decisions on the layout of unmanned aerial vehicle (UAV) landing platforms for building facades.

[0004] Technical solution: The method for determining the layout of unmanned aerial vehicle (UAV) landing platforms for building facades as described in this invention includes the following steps:

[0005] Acquire enhanced building information model data of the target building, high-definition image data of the building facade, computational fluid dynamics wind environment simulation data of the building's surroundings, and the base data of drone service demand for various functional spaces inside the building;

[0006] Based on the enhanced building information model data, an initial set of candidate points is generated on the building facade by dividing it into three-dimensional meshes, and then filtered according to preset hard constraints to obtain a set of valid candidate points.

[0007] For each candidate point in the set of valid candidate points, wind environment adaptability assessment, structural bearing capacity assessment, and user demand responsiveness assessment are performed to obtain corresponding quantitative scores. The wind environment adaptability assessment is achieved by calculating the wind environment stability index, the structural bearing capacity assessment is achieved by calculating the structural bearing capacity score, and the user demand responsiveness assessment is achieved by calculating the total score of user demand intensity.

[0008] A multi-objective optimization model is constructed with the goal of maximizing the comprehensive fitness function, which is obtained by weighted summation of the wind environment stability index, structural bearing capacity score, and user demand intensity score. Under the constraints of satisfying the upper limit of the total number of platforms and no airspace conflict, a multi-objective optimization algorithm is used to solve the effective candidate point set and output the Pareto optimal layout scheme set.

[0009] Preferably, the enhanced building information model (BIM) data is an IFC format model, which supplements the identification of the location of major structural components (such as structural columns, load-bearing walls, and structural beams) and window distribution information.

[0010] Preferably, the demand base for drone services is obtained through statistical data from the building management system and / or user questionnaire survey data for each room inside the building, and is constructed as a demand distribution matrix associated with the room location.

[0011] Preferably, the hard constraints include:

[0012] First constraint: Candidate points must be located within the projection area of ​​the load-bearing structural members marked in the enhanced building information model on the facade;

[0013] The second constraint is that within a cylindrical airspace centered on the candidate point and with a preset radius as the clearance range, there are no facade obstacles identified based on the high-definition image data; and this clearance range does not overlap with the clearance ranges of other candidate points during the initial screening; the identification is achieved through image recognition technology based on a deep learning instance segmentation model.

[0014] Preferably, the wind environment adaptability assessment is achieved by calculating the wind environment stability index of each candidate point. Its index The calculation method is as follows:

[0015] ;

[0016] in, The wind environment stability index. Candidate point locations obtained based on the computational fluid dynamics wind environment simulation data. average wind speed, Candidate point location Gust wind speed, Candidate point location turbulence intensity, The reference wind speed is set according to the wind resistance level of the drone. This is the turbulence correction factor.

[0017] Preferably, the computational fluid dynamics (CFD) wind environment simulation data is obtained through ANSYS Fluent software simulation, and the output is a cloud map of wind speed and turbulence intensity distribution at various points on the building facade.

[0018] Preferably, the structural bearing capacity score is calculated as follows:

[0019] Define the structural bearing capacity score for:

[0020] ;

[0021] in, Candidate points Structural bearing capacity scoring, max function iterates through candidate points Preset all major load-bearing components within the adjacent space; The weighting coefficients are preset according to the type of load-bearing components, with the main structural columns having the highest weight and non-load-bearing components having a weight of zero. Candidate points The shortest distance in three-dimensional space to the currently traversed load-bearing component; This is the distance attenuation coefficient.

[0022] Preferably, in calculating the structural bearing capacity score At that time, the component type weight The assignment strategy is as follows: main structural columns are assigned 1.0, load-bearing walls are assigned 0.9, structural beams are assigned 0.7, and non-load-bearing components are assigned 0.

[0023] Preferably, the total score for user demand intensity is calculated as follows:

[0024] Define the total score of the user demand intensity. for:

[0025] ,

[0026] in, Candidate points Total score of user demand intensity, summation function Traverse candidate points All the functional spaces inside the building that can be spatially connected ; For functional space The base of demand for drone services; Candidate points To functional space Spatial accessibility distance; Based on The distance decay function is of the form:

[0027] ,

[0028] in, This is the distance attenuation coefficient.

[0029] Preferably, the maximization of the comprehensive fitness function Defined as:

[0030] ;

[0031] in, This is the overall fitness function; , , These are the weighting coefficients for user demand responsiveness, wind environment adaptability, and structural bearing capacity, respectively. ; As a decision variable, when candidate points The value is 1 when selected, and 0 otherwise. , , Candidate points The total score of user demand intensity, wind environment stability index and structural bearing capacity score; The total number of valid candidate points;

[0032] The constraints include: ,in The platform is required to set a maximum number of points; and all selected candidate points must meet the airspace corridor conflict-free constraint.

[0033] Preferably, the conflict-free airspace corridor includes: generating a cylindrical clearance corridor centered on each candidate point with a preset radius, and detecting whether there is spatial interference between the corridor and all obstacle areas identified based on the image, as well as the clearance corridors of other candidate points.

[0034] Preferably, the multi-objective optimization algorithm is used to solve the problem, including using a non-dominated sorting genetic algorithm with an elitist strategy to iteratively calculate the optimization model containing the comprehensive fitness function and constraints, and finally output a Pareto optimal solution set that achieves a trade-off among multiple objectives.

[0035] Preferably, the image recognition technology based on the deep learning instance segmentation model includes using a model trained on the Mask R-CNN architecture to process the high-definition image data and output pixel-level mask and bounding box information of facade obstacles.

[0036] Preferably, the training dataset used for training the instance segmentation model based on the Mask R-CNN architecture includes labeled images of common obstacles on building facades (such as air conditioner outdoor units, billboards, window frames, etc.).

[0037] Preferably, the spacing of the three-dimensional mesh is 0.5 meters to 1.0 meters; the preset radius of the clearance range is 2.5 meters to 3.5 meters; in the structural bearing capacity scoring calculation, the preset neighborhood space is a 5-meter range around the candidate point, with a distance attenuation coefficient. The value is 0.5; in the calculation of the total score for user demand intensity, the distance attenuation coefficient is used. The value ranges from 0.1 to 0.3.

[0038] A decision-making system for the layout of unmanned aerial vehicle (UAV) landing platforms for building facades includes:

[0039] The multi-source data fusion module is used to acquire and integrate enhanced building information model data, high-definition image data of building facades, computational fluid dynamics wind environment simulation data, and user requirement data;

[0040] The candidate point generation and screening module is used to generate three-dimensional mesh candidate points on the facade based on the building information model, and to perform initial screening based on load-bearing area constraints and airspace obstacle constraints.

[0041] The multi-factor coupled analysis and evaluation module includes a wind environment evaluation unit, a structural bearing capacity evaluation unit, and a user demand evaluation unit, which are used to calculate the wind environment stability index, structural bearing capacity score, and user demand intensity total score of candidate points, respectively.

[0042] The multi-objective optimization decision module is used to construct the comprehensive fitness function and constraints, run the multi-objective optimization algorithm, and output the Pareto optimal layout scheme set.

[0043] Beneficial Effects: Compared with existing technologies, this invention has the following significant advantages: By integrating Building Information Modeling (BIM), high-precision image recognition, computational fluid dynamics simulation, and user demand data, and combining three major evaluation models—structural load-bearing capacity, wind environment adaptability, and demand responsiveness—for multi-factor coupled analysis, a Pareto-optimal layout scheme set is ultimately formed through a multi-objective optimization algorithm. This achieves automated and scientific decision-making for the site selection of UAV take-off and landing platforms on complex building facades. It not only improves the accuracy and return on investment of urban vertical spatial infrastructure planning, but also solves the problems of high implementation risk, insufficient safety redundancy, and low service efficiency caused by the single decision-making dimension and neglect of engineering feasibility and micro-environmental factors in existing technologies. This significantly enhances the safety, economy, and applicability of low-altitude economic infrastructure deployment at the building scale. Attached Figure Description

[0044] Figure 1 This is a flowchart illustrating the system operation of the present invention.

[0045] Figure 2 This is a schematic diagram of a building scan by an unmanned aerial vehicle (UAV) according to the present invention;

[0046] Figure 3 This is a diagram illustrating the building obstacle recognition method of the present invention.

[0047] Figure 4 This is a simulation diagram of the building wind environment according to the present invention;

[0048] Figure 5 This is a diagram of the Pareto front multi-objective optimization strategy of the present invention. Detailed Implementation

[0049] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0050] like Figure 1 As shown in the figure, this embodiment provides a method for making decisions on the layout of a UAV landing platform for a building facade, which includes the following steps:

[0051] Step 1: Multi-source data fusion and candidate point generation.

[0052] Taking a mid-to-high-rise office building that requires planning for a drone take-off and landing platform on its facade as an example, we obtained its relevant multi-source heterogeneous data.

[0053] First, high-resolution images of all facades of the building were obtained through drone aerial photography (such as...). Figure 2 (As shown). These images are processed using an instance segmentation model trained on a Mask R-CNN architecture. The model, trained on a dataset containing categories such as "air conditioner outdoor unit," "billboard," and "window frame," accurately identifies and outputs pixel-level masks and bounding box information for these obstacles, generating images like... Figure 3The diagram shows the identification of obstacles and windows, thus accurately defining their three-dimensional spatial occupancy.

[0054] At the same time, the building's enhanced building information model (BIM) data was imported. The model is in IFC format and has been supplemented with the spatial location information of the main load-bearing components (including structural columns, load-bearing walls, and structural beams) and the distribution of the building's exterior windows.

[0055] The ANSYS Fluent software was used to simulate the wind environment of the building and its surroundings, obtaining the average wind speed, gust wind speed, and turbulence intensity distribution cloud maps at various spatial locations on the building facade (e.g., Figure 4 (As shown).

[0056] By acquiring the functional attributes of rooms on each floor through a building management system (BMS) and combining this with questionnaires from room users, a demand base value for drone services is assigned to each functional room, forming a demand intensity matrix associated with the room's spatial coordinates. For example, a courier room could be assigned a value of 10.0, an office area 3.0, and an unmanned area 0.

[0057] Based on enhanced BIM data, an initial candidate point set is generated on the building facade through 3D mesh generation. The spacing of the 3D mesh ranges from 0.5 meters to 1.0 meters; in this embodiment, 0.8 meters is used. The generated 3D mesh nodes constitute the initial candidate point set. The initial candidate point set is then filtered according to preset hard constraints to obtain a valid candidate point set.

[0058] The hard constraints include two screening criteria that must be met simultaneously: First constraint (structural feasibility constraint): The candidate point must be located within the vertical projection area of ​​the load-bearing structural components (including main structural columns and load-bearing walls) identified in the enhanced BIM model on the building facade, or within a preset tolerance (0.5 meters in this embodiment) of their edge. Second constraint (airspace safety constraint): Within a cylindrical clear space with a preset radius of 3 meters centered on the candidate point, there must be no facade obstacles extracted based on image recognition technology; and during the initial screening, this clear space should not overlap with the clear space of other candidate points.

[0059] Step 2: Multi-factor coupled quantitative assessment.

[0060] For each candidate point in the valid candidate point set, wind environment adaptability assessment, structural bearing capacity assessment, and user demand responsiveness assessment are performed respectively, and the corresponding quantitative score is calculated: wind environment stability index. Structural bearing capacity score Total score of user demand intensity .

[0061] Wind environment adaptability assessment is performed by calculating the wind environment stability index of candidate sites. The specific calculation method is as follows:

[0062] ;

[0063] in, The wind environment stability index. Candidate point locations obtained based on the computational fluid dynamics wind environment simulation data. The average wind speed (m / s). Candidate point location Gust wind speed (m / s). Candidate point location turbulence intensity, The reference wind speed (m / s) is set according to the wind resistance level of the drone, for example, 12 m / s. This is the turbulence correction factor.

[0064] Structural bearing capacity assessment is performed by calculating the structural bearing capacity score of candidate points. The specific calculation method is as follows:

[0065] ;

[0066] in, Candidate points Structural bearing capacity scoring, max function iterates through candidate points All major load-bearing components within a predefined neighborhood space (such as a sphere with a radius of 5 meters); The assignment strategy is as follows, based on the pre-set weighting coefficients for the types of load-bearing components: main structural columns = 1.0, load-bearing walls = 0.9, structural beams = 0.7, and non-load-bearing components = 0. Candidate points The shortest distance (in meters) in three-dimensional space to the currently traversed load-bearing component; The distance attenuation coefficient is set to 0.5 in this embodiment. This model ensures that the score decreases with increasing distance and preferentially associates the component with the strongest load-bearing capacity.

[0067] User demand responsiveness assessment is achieved by calculating the total score of user demand intensity for candidate points. The specific calculation method is as follows:

[0068] ,

[0069] in, Candidate points Total score of user demand intensity, summation function Traverse candidate points All interior functional spaces of the building that are spatially connected (usually referring to those closest in vertical projection or accessible through windows). ; For functional space The base of demand for drone services; Candidate points To functional space Spatial accessibility distance (meters); Based on The distance decay function is of the form:

[0070] ,

[0071] in, The distance attenuation coefficient ranges from 0.1 to 0.3, and is set to 0.2 in this embodiment. This model aggregates discrete room requirements onto candidate facade points using a spatial attenuation function.

[0072] Step 3: Multi-objective optimization decision making.

[0073] A multi-objective optimization model is constructed with the goal of maximizing the comprehensive fitness function F. The comprehensive fitness function is obtained by weighted aggregation of the wind environment stability index, structural bearing capacity score and user demand intensity total score of all candidate points. Under the constraints of satisfying the upper limit of the total number of platforms and the absence of conflicts in the global airspace, a multi-objective optimization algorithm is used to solve the effective candidate point set and output a set of Pareto optimal layout schemes.

[0074] The maximum fitness function maxF is defined as follows:

[0075] ;

[0076] in, This is the overall fitness function; , , These are the weighting coefficients for user demand responsiveness, wind environment adaptability, and structural bearing capacity, respectively. Its specific value can be set by the decision-maker according to the project's focus; in this embodiment, it is set to [value]. =0.5, =0.3, =0.2; As a decision variable, when candidate points The value is 1 when selected, and 0 otherwise. , , Candidate points The total score of user demand intensity, wind environment stability index and structural bearing capacity score; This represents the total number of valid candidate points.

[0077] The constraints of the optimization model include:

[0078] Platform total limit: ,in The maximum number of platforms allowed by the project budget or facade space.

[0079] Global spatial conflict-free constraint: For all... The candidate points, whose corresponding cylindrical clearance corridors with radius R must not intersect each other in three-dimensional space, and none of them may intersect with any of the obstacle areas identified in step 1.

[0080] A multi-objective optimization algorithm is used to solve the problem, specifically the Non-Dominated Ranking Genetic Algorithm with Elite Strategy (NSGA-II). This algorithm uses decision variables... The encoding is represented as an individual, and iterative evolution is performed with the fitness function F as the objective, while satisfying the above constraints. Each generation generates a new population through selection, crossover, and mutation, and maintains solution diversity and approximation to the Pareto front by using non-dominated sorting and crowding comparison. The final output of the optimization solution is a Pareto optimal solution set (its front is illustrated in Figure 1). Figure 5 As shown in the diagram, each solution represents a set of selected candidate points (i.e., a layout scheme), and among all solutions, no single solution is superior to another in all three sub-objectives of F. Decision-makers can select the final implementation scheme from this solution set based on actual project budget, safety redundancy requirements, or operational strategies.

[0081] Step 4: Solution output and system configuration.

[0082] The selected final layout scheme is output as a construction guidance document, which includes the three-dimensional coordinates of each selected platform point, the corresponding load-bearing component number, the list of rooms to be covered by the service, and the clear corridor range diagram.

[0083] This embodiment also provides a building facade UAV landing platform layout decision system, including:

[0084] The multi-source data fusion module includes a BIM data parsing unit, an image recognition unit, a CFD data interface, and a requirement data input unit, which completes the acquisition, processing, and fusion of multi-source data.

[0085] The candidate point generation and initial screening module has a built-in grid partitioning algorithm and spatial conflict detection engine.

[0086] The multi-factor coupling analysis and evaluation module includes: a wind environment assessment unit, used to call CFD data and calculate based on the stated index formula. The structural load-bearing capacity assessment unit is configured to read BIM structural information and calculate based on the scoring model. The user demand assessment unit is configured to calculate based on the demand matrix and spatial topology. .

[0087] The multi-objective optimization decision module includes an optimization model building unit and an NSGA-II algorithm solver, which is responsible for setting weight parameters, defining constraints, running optimization iterations, and outputting the Pareto optimal solution set.

[0088] The solution visualization and output module is used to visualize the optimization results in three dimensions and generate drawings and data reports that can be used for construction.

[0089] The method and system for making decisions on the layout of unmanned aerial vehicle (UAV) landing platforms for building facades provided by this invention have achieved the following significant progress and technical effects compared with the prior art:

[0090] This invention constructs a site selection decision framework for UAV take-off and landing platforms oriented towards the complex vertical space of building facades by integrating enhanced BIM, high-precision image recognition, CFD wind field simulation, and refined user demand data. This method innovatively proposes a wind environment stability index, a distance-attenuation-based structural bearing capacity scoring model, and a user demand spatial aggregation model, unifying and quantifying the three dimensions of engineering feasibility, physical environmental safety, and human-centered usage efficiency. This fundamentally overcomes the shortcomings of existing macro-site selection methods, which suffer from a single decision-making dimension at the micro-scale of buildings and neglect engineering constraints and real-world needs.

[0091] Furthermore, by constructing a multi-objective optimization model with the goal of maximizing the comprehensive fitness function and employing advanced algorithms such as NSGA-II for solution, the system automatically selects a set of Pareto-optimal layout schemes from a massive pool of candidate locations, achieving the best trade-off among multiple objectives such as safety, feasibility, and efficiency. This process not only significantly improves the scientific rigor and automation of the planning process, but its output solution set also provides decision-makers with flexible, data-driven comparison criteria, significantly enhancing the return on investment and the success rate of infrastructure planning.

[0092] Of particular note is that this invention, by introducing a Mask R-CNN-based pixel-level obstacle recognition system and a cylindrical clearance corridor conflict detection mechanism, provides centimeter-level precision assurance for the take-off and landing safety of UAVs in dense urban facade environments, solving the resolution deficiency problem of traditional roof scanning technology in this scenario. Simultaneously, its modular system design provides excellent scalability, easily incorporating more evaluation factors such as lighting impact and visual aesthetics.

[0093] Ultimately, this invention forms a complete technical closed loop from data acquisition, intelligent analysis, automatic optimization to solution output, providing a practical core decision-making tool for the efficient, safe, and rational development of low-altitude economic vertical space resources in existing buildings.

Claims

1. A method for determining the layout of unmanned aerial vehicle (UAV) landing platforms for building facades, characterized in that, Includes the following steps: Acquire enhanced building information model data of the target building, high-definition image data of the building facade, computational fluid dynamics wind environment simulation data of the building's surroundings, and the base data of drone service demand for various functional spaces inside the building; Based on the enhanced building information model data, an initial set of candidate points is generated on the building facade by dividing it into three-dimensional meshes, and then filtered according to preset hard constraints to obtain a set of valid candidate points. For each candidate point in the set of valid candidate points, wind environment adaptability assessment, structural bearing capacity assessment, and user demand responsiveness assessment are performed to obtain corresponding quantitative scores. The wind environment adaptability assessment is achieved by calculating the wind environment stability index, the structural bearing capacity assessment is achieved by calculating the structural bearing capacity score, and the user demand responsiveness assessment is achieved by calculating the total score of user demand intensity. A multi-objective optimization model is constructed with the goal of maximizing the comprehensive fitness function, which is obtained by weighted summation of the wind environment stability index, structural bearing capacity score, and user demand intensity score. Under the constraints of satisfying the upper limit of the total number of platforms and no airspace conflict, a multi-objective optimization algorithm is used to solve the effective candidate point set and output the Pareto optimal layout scheme set.

2. The method according to claim 1, characterized in that, The hard constraints include: First constraint: Candidate points must be located within the projection area of ​​the load-bearing structural members marked in the enhanced building information model on the facade; The second constraint is that within a cylindrical airspace centered on the candidate point and with a preset radius as the clearance range, there are no facade obstacles identified based on the high-definition image data; and this clearance range does not overlap with the clearance ranges of other candidate points during the initial screening; the identification is achieved through image recognition technology based on a deep learning instance segmentation model.

3. The method according to claim 1, characterized in that, The wind environment adaptability assessment is achieved by calculating the wind environment stability index of each candidate point. Its index The calculation method is as follows: ； in, The wind environment stability index. Candidate point locations obtained based on the computational fluid dynamics wind environment simulation data. average wind speed, Candidate point location Gust wind speed, Candidate point location turbulence intensity, The reference wind speed is set according to the wind resistance level of the drone. This is the turbulence correction factor.

4. The method according to claim 1, characterized in that, The structural bearing capacity score is calculated as follows: Define the structural bearing capacity score for: ； in, Candidate points Structural bearing capacity scoring, max function iterates through candidate points Preset all major load-bearing components within the adjacent space; The weighting coefficients are preset according to the type of load-bearing components, with the main structural columns having the highest weight and non-load-bearing components having a weight of zero. Candidate points The shortest distance in three-dimensional space to the currently traversed load-bearing component; This is the distance attenuation coefficient.

5. The method according to claim 1, characterized in that, The total score for user demand intensity is calculated as follows: Define the total score of the user demand intensity. for: ， For the total user demand intensity score of candidate point i, the summation function is... Traverse all interior functional spaces of the building that can be spatially associated with candidate point i. ; For functional space The base of demand for drone services; For candidate point i to function space Spatial accessibility distance; For based on The distance decay function is of the form: ， in, This is the distance attenuation coefficient.

6. The method according to claim 1, characterized in that, The maximization of the comprehensive fitness function Defined as: ； in, This is the overall fitness function; in, This is the overall fitness function; , , These are the weighting coefficients for user demand responsiveness, wind environment adaptability, and structural bearing capacity, respectively. , is a decision variable, which takes the value 1 when candidate point i is selected, and 0 otherwise; , , These are the total user demand intensity score, wind environment stability index, and structural bearing capacity score for candidate point i, respectively. The total number of valid candidate points; the constraints include: ,in The platform is required to set a maximum number of points; and all selected candidate points must meet the airspace corridor conflict-free constraint.

7. The method according to claim 1, characterized in that, The multi-objective optimization algorithm is used to solve the problem, including the use of a non-dominated sorting genetic algorithm with an elitist strategy to iteratively calculate the optimization model containing the comprehensive fitness function and constraints, and finally output a Pareto optimal solution set that achieves a trade-off among multiple objectives.

8. The method according to claim 2, characterized in that, The image recognition technology based on the deep learning instance segmentation model includes using a model trained on the Mask R-CNN architecture to process the high-definition image data and output pixel-level mask and bounding box information of facade obstacles.

9. The method according to claim 1, characterized in that, The spacing of the three-dimensional mesh is 0.5 meters to 1.0 meter; the preset radius of the clearance range is 2.5 meters to 3.5 meters; in the structural bearing capacity scoring calculation, the preset neighborhood space is a 5-meter range around the candidate point, with a distance attenuation coefficient. The value is 0.5; in the calculation of the total score for user demand intensity, the distance attenuation coefficient is used. The value ranges from 0.1 to 0.

3.

10. A decision-making system for the layout of unmanned aerial vehicle (UAV) landing platforms for building facades, characterized in that, include: The multi-source data fusion module is used to acquire and integrate enhanced building information model data, high-definition image data of building facades, computational fluid dynamics wind environment simulation data, and user requirement data; The candidate point generation and screening module is used to generate three-dimensional mesh candidate points on the facade based on the building information model, and to perform initial screening based on load-bearing area constraints and airspace obstacle constraints. The multi-factor coupled analysis and evaluation module includes a wind environment evaluation unit, a structural bearing capacity evaluation unit, and a user demand evaluation unit, which are used to calculate the wind environment stability index, structural bearing capacity score, and user demand intensity total score of candidate points, respectively. The multi-objective optimization decision module is used to construct the comprehensive fitness function and constraints, run the multi-objective optimization algorithm, and output the Pareto optimal layout scheme set.

Images