Method for installing and detecting building special-shaped multi-curve external envelope structure based on unmanned aerial vehicle three-dimensional reconstruction

Abstract

The invention discloses a building special-shaped multi-curve outer envelope structure installation detection method based on unmanned aerial vehicle three-dimensional reconstruction, and belongs to the technical field of construction methods. Performing precision rechecking on the corresponding positions of the structure installation positions of the risk sensitive areas of the live-action three-dimensional model and the BIM design model, and generating deviation data between the live-action model and the BIM design model through registration deviation analysis; marking a position with deviation on the live-action three-dimensional model; according to the method, the structure installation position affected by galloping and vortex-induced resonance and all external envelope structures affecting long-term stability can be recognized, precision rechecking is conducted on a live-action three-dimensional model and a BIM design model, and constructors are guided to rectify the quality of an installation project through registration deviation analysis data; the limitation of the traditional technology is overcome, and the construction quality management efficiency and precision are remarkably improved, so that the construction risk is reduced.

Description

Technical Field

[0001] The present invention belongs to the fields of computer-aided design technology and construction management technology, and specifically relates to a method for detecting the installation of a building's special-shaped multi-curved exterior enclosure structure through three-dimensional reconstruction by an unmanned aerial vehicle (UAV). Background Art

[0002] In the contemporary construction industry, buildings with unusually shaped, curved facades are becoming a key choice for urban landmarks due to their unique artistic expression and spatial creativity. These buildings break away from the regular forms of traditional architecture, featuring complex shapes such as curves, broken lines, and irregular volumes. Key areas such as roofs and facades often exhibit nonlinear spatial variations, placing extremely high demands on construction measurement accuracy and adaptability to in-depth design. When hyperbolic facade envelope structures and roof metal panel systems are applied to high-rise buildings, large public buildings (such as airports and stadiums), and special buildings (large, complex, and densely populated), external risk factors such as gravity loads and wind loads can cause vertical compression deformation between the structural layer and the outer protective layer. To guide airflow and reduce wind pressure and wind vibration on the exterior envelope (or to improve the roof's anti-overturning performance, etc.), hyperbolic facade designs are often based on aerodynamic optimization by the architect. Insufficient line-laying accuracy can lead to deviations between the actual installed curved surface and the designed model (local unevenness and discontinuous curvature in the actual molding). This deviation can disrupt the streamlined design, create local turbulence and other unfavorable factors, increase local wind loads, and even induce harmful vibrations. Furthermore, curved envelope systems are prone to cumulative errors during installation, and errors in skeleton installation will further affect the subsequent processing and production of curved panels. Traditional measurement methods are limited by the complex spatial angles and irregular shapes of special-shaped structures. They rely on ground-based measurement methods such as total stations or 3D scanning. These methods often have difficulty covering all key points due to blind spots, resulting in low measurement efficiency and difficulty accurately capturing the continuous changes in the structure, resulting in significant deviations between the acquired data and the actual structure. Summary of the Invention

[0003] The purpose of the present invention is to propose a method for detecting the installation of a building's irregular multi-curved exterior envelope structure using three-dimensional reconstruction by an unmanned aerial vehicle (UAV) to solve one or more computer-aided design problems existing in the prior art and to at least provide a beneficial option or create conditions.

[0004] To achieve the above objectives, according to one aspect of the present disclosure, a method for detecting the installation of a building's irregularly shaped, multi-curved exterior enclosure structure using three-dimensional reconstruction by an unmanned aerial vehicle is provided. The method comprises the following steps: S100, importing the BIM design model of the target building, marking the location of the external enclosure structure on the BIM design model as the structure installation location; S200, obtains the surface wind pressure distribution of the BIM design model through simulation with computer-aided design simulation software; S300: Under different wind speed conditions, the surface wind pressure distribution is loaded into the BIM design model to obtain the vibration displacement of each structural installation location under different average wind speeds. The structural installation locations with risk-sensitive areas are identified and marked on the BIM design model. S400, uses a drone to capture a sequence of images of a target building and performs 3D reconstruction on the image sequence to obtain a real-life 3D model of the target building; S500: During the main structure installation phase, the enclosure structure skeleton installation phase, and the enclosure structure panel installation phase, the accuracy of the corresponding positions of the structural installation positions in the risk-sensitive areas of the real-life 3D model and the BIM design model are reviewed, and deviation data between the on-site real-life model and the BIM design model is generated through registration deviation analysis; S600: Mark the location where the deviation occurs on the real-scene 3D model, and output and display the real-scene 3D model.

[0005] Furthermore, in S100, the BIM design model of the target building is a static BIM three-dimensional model of the target building constructed by using BIM technology according to the design drawings of the target building.

[0006] Among them, the external protective structure includes curtain wall system and metal roof system. The curtain wall system includes metal curtain wall or glass curtain wall, supporting structure and connection sealing system. The supporting structure includes connecting parts or purlins or purlin supports. The connection sealing system includes silicone glue, sealing strips or hinges and bolts.

[0007] Furthermore, in S200, the method for simulating and obtaining the surface wind pressure distribution of the BIM design model through computer-aided design simulation software is: importing the BIM design model into the computer-aided design simulation software, performing flow field analysis under given wind speed boundary conditions to obtain the surface wind pressure distribution of the BIM design model, and the steps of the flow field analysis specifically include: importing the BIM design model, surface segmentation, meshing, and mesh property definition; setting material parameters, geometric properties, wind speed boundary conditions, load conditions, and defining the connectivity of the units.

[0008] Preferably, the flow field analysis is specifically as follows: in the peripheral flow field of the BIM design model of the target building, the transient wind pressure distribution under different wind direction angles (0°~360°, interval 22.5°) is simulated, and the wind pressure data is coupled to the structural finite element model; dynamic wind pressure load is applied, time history analysis is performed and resonance areas are identified, and the time history data of the surface wind pressure coefficient of the BIM design model is exported (.csv / .dat format), and the wind pressure distribution data and displacement cloud map of the BIM design model are exported.

[0009] Furthermore, the BIM design model is in rvt format or 3dm format.

[0010] Furthermore, in S100, it also includes segmenting and extracting the three-dimensional grid of the wall of the BIM design model, calculating the three-dimensional support plane fitted by the least squares method; calculating the Euclidean distance between the three-dimensional grid and the three-dimensional support plane, and normalizing it to generate a depth map of the wall; based on the calibrated camera parameters, generating a texture image of the wall through three-dimensional space texture mapping; the length and width dimensions of the depth map and texture map of the wall are kept consistent.

[0011] Preferably, the grid division method is a sweeping method (Sweep) grid division method.

[0012] The wind speed boundary conditions are given as 5m / s to 36.9m / s.

[0013] Preferably, different wind speeds are set at different heights of the BIM design model of the target building for simulation, and given wind speed boundary conditions are used for simulation: the wind speed in the area of ​​0-10 meters corresponding to the BIM design model of the target building is set to 5m / s, the wind speed at a height of 10-30 meters is set to 8.7m / s, and the wind speed at a height of 30-90 meters is set to 15m / s; the wind speed above 90 meters is set to 36.9m / s.

[0014] Note: The Code for Loads on Building Structures (GB 50009-2012) stipulates that high-rise buildings require a wind speed of force 12 (32.7-36.9 m / s) as the design basis; while the wind speed at a height of 10 meters is 5 m / s, at a height of 30 meters it reaches 8.7 m / s, and at a height of 90 meters it rises to 15 m / s. See reference: Fang Xilin, Sun Pengou. Construction Engineering Disaster Prevention Reader [M]. China Environmental Press, 2014.

[0015] Preferably, the wind speed boundary condition is set according to local meteorological data (wind speed and wind direction probability distribution, refer to the Code for Loads on Building Structures (GB 50009-2012)). Preferably, the computer-aided design simulation software is open source OpenFOAM software or Elmer FEM software.

[0016] Further, in S300 , the different wind speed conditions are a wind speed range of 5 m / s to 36.9 m / s.

[0017] Furthermore, in S300, the method for identifying and marking the structural installation locations where risk-sensitive areas exist on the BIM design model is: Mark the wind pressure direction area in the adjacent structure installation position of the current structure installation position; According to the linear wind pressure conditions, the structural installation locations in the risk-sensitive areas are screened out in the wind pressure direction area.

[0018] Furthermore, the wind pressure direction area is marked in the adjacent structure installation position of the current structure installation position, specifically: Based on the surface wind pressure distribution, obtain the wind pressure values ​​corresponding to all points at each structural installation location in the BIM design model. Record the vibration displacement of the geometric center of gravity of each structural installation location from the minimum wind pressure value to the maximum wind pressure value as the wind pressure limit displacement. (When subjected to wind pressure at different locations and heights in the BIM design model, the wind pressure limit displacement will be different due to differences in material parameters, geometric properties, wind speed boundary conditions, and load conditions). The other structural installation positions adjacent to the current structural installation position are recorded as the adjacent structural installation positions of the current structural installation position (regardless of whether the external envelope structure is a component-type building curtain wall, a unitized curtain wall, a glass curtain wall, a stone curtain wall, a metal plate curtain wall, or a full-glass curtain wall, they are all connected to each other at a single structural installation position. Therefore, each structural installation position is surrounded by other curtain walls. Adjacent means: the other structural installation position with the shortest distance to the current structural installation position); Denote the geometric centroid of the structural installation location with the largest wind pressure limit displacement among all adjacent structural installation locations of the current structural installation location as MaxCUP; denote the geometric centroid of the structural installation location with the smallest wind pressure limit displacement among all adjacent structural installation locations of the current structural installation location as MinCUP; then, a ray CUPL is formed from MinCUP to MaxCUP; all structural installation locations sequentially traversed by CUPL are recorded as the wind pressure orientation area of ​​the current structural installation location; (all wind pressure orientation areas sequentially traversed by CUPL are areas with the largest limit displacement, continuous wind vibration, or high wind load risk, passing through the current structural installation location. Since the unit panels of the glass curtain wall are typically installed interconnected to ensure the airtightness, watertightness, and wind pressure resistance of the overall structure, the connection method, force transmission, and deformation adaptation between the individual glass curtain wall units are typically affected by the adjacent structural installation locations. Due to the influence of continuous wind vibration, the dynamic deformation of the structural displacement under the action of wind vibration will cause stress to gradually diffuse and transmit from the area with large wind pressure limit displacement to the area with low wind pressure limit displacement. The range of diffusion and transmission is the wind pressure orientation area marked in the above steps). Furthermore, based on the linear wind pressure conditions, the structural installation locations with risk-sensitive areas are screened out in the wind pressure direction area, specifically: Let WV(j) be the j-th wind pressure direction area of ​​the current structure installation position; j is the sequence number, and the order of the wind pressure direction areas is the order in which CUPL passes through the wind pressure direction areas in sequence; Determine each WV(j). If WV(j) meets the linear wind pressure condition, mark WV(j) as a structural installation location with a risk-sensitive area. The linear wind pressure condition is: MV(WV(j-1))>MV(WV(j)) and MV(WV(j))<MV(WV(j+1)), where MV(WV(j)) is the wind pressure limit displacement corresponding to WV(j); Among them, the above method quickly locates the structural installation positions with risk-sensitive areas by determining the wind pressure limit displacement of the discontinuous structural installation positions. These risk-sensitive areas cannot be transferred by the unloading of adjacent adjacent structural installation positions, and thus may be the areas most affected by continuous wind vibration. However, if the structural installation position can only detect a discrete point area, when a continuous structural installation position affected by wind vibration appears, the above method has low recognition accuracy and low accuracy. Therefore, it is necessary to use the following linear wind pressure conditions for identification, specifically: Preferably, the linear wind pressure condition is: DE(WV(j-1))>DE(WV(j)) and DE(WV(j))<DE(WV(j+1)), Where DE(WV(j)) is the inertial wind pressure index of the j-th wind pressure direction area WV(j) at the current structure installation location; Preferably, the calculation method of the inertial wind pressure index DE(WV(j)) is: DE(WV(j))=|MinA1-MinA2|÷|MaxA1-MaxA2|; Among them, MaxA1 is the maximum value of the wind pressure limit displacement corresponding to the wind pressure direction areas from the 1st to the j-1th; MinA1 is the minimum value of the wind pressure limit displacement corresponding to the wind pressure direction areas from the 1st to the j-1th; MaxA2 is the maximum value of the wind pressure limit displacement corresponding to the wind pressure direction areas from the j+1th to the SuMth; MinA2 is the minimum value of the wind pressure limit displacement corresponding to the wind pressure direction areas from the j+1th to the SuMth; SuM is the number of wind pressure direction areas.

[0019] Among them, the inertial wind pressure index DE (WV (j)) calculated by the above method is used to judge whether the wind pressure influence has continuous inertia in the form of the ratio of the quantitative change of the convergent wind pressure difference and the transmitted wind pressure difference at the structural installation position, and whether it affects the risk-sensitive area of ​​the surrounding adjacent external protective structure due to vertical compression deformation, expansion deformation, and stress concentration caused by wind vibration. However, due to the characteristics of wind pressure, when the wind passes by the edge of the building, the Karman vortex street effect will occur periodically, resulting in galloping and vortex-induced resonance caused by alternating wind pressure. The external protective structure continues to absorb energy in the wind due to aerodynamic negative damping, resulting in an increase in amplitude. The above method cannot identify the risk-sensitive area of ​​the structural installation position affected by galloping and vortex-induced resonance, but can only identify the structural installation position affected by continuous wind vibration. In order to accurately identify the structural installation position affected by galloping and vortex-induced resonance, it is necessary to identify it through the following linear wind pressure conditions, specifically: Preferably, the calculation method of the inertial wind pressure index DE(WV(j)) is: Calculate the convergent wind pressure difference and the transfer wind pressure difference at all structural installation locations. The specific method is: The convergence wind pressure difference is the difference between the average value of the wind pressure limit displacements corresponding to the wind pressure direction area from the 1st to the j-1th structural installation positions and the corresponding wind pressure limit displacement at the structural installation position; the transfer wind pressure difference is the difference between the average value of the wind pressure limit displacements corresponding to the wind pressure direction area from the j+1th to the Nth structural installation positions and the corresponding wind pressure limit displacement at the structural installation position; (the transfer wind pressure difference indicates the degree to which the wind pressure limit displacement increases from the structural installation position along the wind pressure direction area, and the convergence wind pressure difference indicates the degree to which the wind pressure limit displacement decreases from the structural installation position along the wind pressure direction area). Traverse the wind pressure direction areas corresponding to each structural installation position, and use MaxS to represent the number of wind pressure direction areas where the convergent wind pressure difference of each wind pressure direction area corresponding to the structural installation position is greater than the transmission wind pressure difference; use MinS to represent the number of wind pressure direction areas where the convergent wind pressure difference of each wind pressure direction area corresponding to the structural installation position is less than the transmission wind pressure difference; then the calculation method of the inertial wind pressure index DE(WV(j)) is: inertial wind pressure index DE(WV(j)) = MaxS ÷ (MaxS + MinS).

[0020] Among them, the inertial wind pressure index DE (WV (j)) calculated by the above method is in the form of the ratio of the quantitative change of the convergent wind pressure difference and the transmitted wind pressure difference at the structural installation position, and judges the inertial diffusion degree of the wind pressure limit displacement caused by the galloping vibration and vortex-induced resonance caused by the alternating wind pressure generated by the Karman vortex street effect in certain continuous areas, thereby accurately quantifying and marking the structural installation positions with higher wind pressure limit displacements that will be affected by the wind pressure inertia and thus fail to effectively adapt to the dynamic deformation under the action of wind vibration, identifying the structural installation positions affected by galloping vibration and vortex-induced resonance, and affecting the long-term stability of various external protective structures.

[0021] Furthermore, in S400, the method for photographing the image sequence of the target building by the drone is as follows: using the drone as a carrier, generating a collection route, and the image sequence of the drone flying along the route ultimately forms the image sequence of the target building.

[0022] Furthermore, in S400 , performing three-dimensional reconstruction on the image sequence to obtain the real three-dimensional model of the target building includes: performing three-dimensional reconstruction on the image sequence by using a 3D Reconstruction or MeshRoomCL three-dimensional reconstruction tool to obtain the real three-dimensional model.

[0023] Furthermore, in S500, the accuracy of the corresponding positions of the structural installation positions of the real scene 3D model and the BIM design model in the risk-sensitive areas in the main structure installation stage, the enclosure structure skeleton installation stage, and the enclosure structure panel installation stage are respectively reviewed. The method for generating the deviation data between the on-site real scene model and the BIM design model through registration deviation analysis is as follows: The ICP algorithm is used to verify the accuracy of the corresponding positions of the structural installation positions in risk-sensitive areas between the real-life 3D model and the BIM design model, and the deviation data between the on-site real-life model and the BIM design model is generated through alignment deviation analysis.

[0024] The ICP algorithm is used for fusion analysis: the real-life 3D model is registered with the BIM design model, and the deviation data of the corresponding positions of the structural installation positions in risk-sensitive areas are calculated. The deviation data is color-coded to show the size deviation range of different areas. Among them, the deviation data includes overall deviation statistics (mean, standard deviation, maximum deviation), deviation heat map (color mapping distance value, intuitive display of error distribution), and structural installation position error analysis data in risk-sensitive areas: statistical deviations are grouped by structures (external protective structures) in risk-sensitive areas.

[0025] Among them, the ICP algorithm is the ICP registration algorithm in the open source 3D data processing library Open3D.

[0026] During the main structure installation phase, the corresponding positions of the structural installation positions in the real-life 3D model and the BIM design model include: steel structure / concrete main body (load-bearing core of special-shaped buildings), connecting parts or purlins or purlin supports, steel columns, steel beams or space trusses (need to be pre-bent into design curves), concrete shear walls or special-shaped core tubes (curved surface modeling is achieved through template customization).

[0027] Among them, the structural installation position in the above main structure installation stage provides stable support for the special-shaped external protective structure, ensuring the positioning reference of the subsequent skeleton and panel.

[0028] During the installation phase of the enclosure structure skeleton, the corresponding positions of the structural installation positions in the real-life 3D model and the BIM design model include: main purlin: usually a curved steel beam or aluminum profile, arranged along the normal direction of the curved surface; secondary purlin: perpendicular to the main purlin, forming a grid-like support (needs to be pre-bent into a hyperbolic shape); connection nodes: three-dimensional adjustable articulated supports (to adapt to multi-directional freedom adjustment) in the corresponding positions of the structural installation positions in the real-life 3D model and the BIM design model.

[0029] Among them, the structural installation position during the main structure installation phase is a secondary support system built to adapt to the special-shaped curved surface, providing an installation carrier for the panel.

[0030] During the installation phase of the enclosure structure panels, the corresponding positions of the structural installation positions in the real-life 3D model and the BIM design model include: metal plates (aluminum plates, titanium zinc plates): hyperbolic panels that need to be CNC stamped, glass curtain walls: using cold bending technology or triangular flat glass to splice fitting surfaces, composite panels (GFRP, ETFE): factory-prefabricated integral curved surface units, connection and sealing systems including silicone glue, sealing strips or hinges, and bolts in the corresponding positions of the structural installation positions in the real-life 3D model and the BIM design model.

[0031] Among them, the structural installation position during the enclosure structure panel installation phase is to complete the precise installation of special-shaped multi-curved panels to ensure the appearance and air tightness.

[0032] Furthermore, in S500 , the method for outputting and displaying the real-scene 3D model is: performing adaptive histogram equalization processing on the position where the deviation exists on the real-scene 3D model to enhance the local contrast.

[0033] Preferably, the method further includes transmitting the real-scene three-dimensional model to a client or a mobile device for display output.

[0034] Preferably, the method for transmitting the real scene three-dimensional model to a client or mobile device for display output comprises the following steps: processing the real scene three-dimensional model by an edge folding algorithm to obtain a simplified model; The simplified model is divided into 1m×1m blocks according to space and compressed separately; Dynamically push visible blocks based on the client or mobile device's field of view (using HTTP / 2 ServerPush).

[0035] When loading on the client or mobile device, detect the GPU performance and select the corresponding LOD level: High-end devices: LOD0 (original accuracy); Mid-range devices: LOD1 (50% mesh); Low-end devices: LOD2 (20% mesh + solid color heat map); The structural installation locations of risk-sensitive areas are highlighted in real time.

[0036] Among them, high-end devices: support hardware-accelerated ray tracing, floating-point computing power ≥ 3 TFLOPS, memory bandwidth ≥ 100GB / s; Mid-range devices: support Vulkan 1.1 and above API, floating point computing power of 1-3 TFLOPS; Low-end devices: only support OpenGL ES 3.0, floating point computing power < 1 TFLOPS.

[0037] The beneficial effects of the present invention are as follows: the present invention provides a method for detecting the installation of special-shaped and multi-curved external envelope structures of buildings based on three-dimensional reconstruction by drones, accurately quantifies and marks the structural installation positions at which the wind pressure limit displacement is higher, which will be affected by the wind pressure inertia and thus fail to effectively adapt to the dynamic deformation under the action of wind vibration, can identify the structural installation positions affected by galloping vibration and vortex-induced resonance, and the various external envelope structures that affect the long-term stability, conduct accuracy review of the real-life three-dimensional model and the BIM design model, accurately quantify the differences between construction and design, and guide construction personnel to rectify the quality of the installation project through alignment deviation analysis data; overcome the limitations of traditional technologies, significantly improve the efficiency and accuracy of construction quality management, and thus reduce construction quality risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 The figure shows a flow chart of the installation inspection method of a building's irregular-shaped multi-curved exterior envelope structure based on UAV 3D reconstruction. DETAILED DESCRIPTION

[0039] The following will be combined with the embodiments and drawings to clearly and completely describe the concept, specific structure and technical effects of the present disclosure to fully understand the purpose, scheme and effect of the present disclosure. It should be noted that the embodiments and features in the embodiments of this application can be combined with each other unless there is a conflict.

[0040] Example 1 Figure 1 The figure shows the flow chart of the installation detection method of the building's special-shaped multi-curved external enclosure structure reconstructed by UAV 3D. Figure 1 The following describes a method for detecting the installation of a building's irregular-shaped, multi-curved exterior enclosure structure using three-dimensional reconstruction by a drone according to an embodiment of the present invention. The method comprises the following steps: S100, importing the BIM design model of the target building, marking the location of the external enclosure structure on the BIM design model as the structure installation location; S200, obtains the surface wind pressure distribution of the BIM design model through simulation with computer-aided design simulation software; S300: Under different wind speed conditions, the surface wind pressure distribution is loaded into the BIM design model to obtain the vibration displacement of each structural installation location under different average wind speeds. The structural installation locations with risk-sensitive areas are identified and marked on the BIM design model. S400, uses a drone to capture a sequence of images of a target building and performs 3D reconstruction on the image sequence to obtain a real-life 3D model of the target building; S500: performing accuracy review on the corresponding positions of the structural installation positions in the risk-sensitive areas of the real-scene 3D model and the BIM design model, and generating deviation data between the real-scene model and the BIM design model through registration deviation analysis; S600: Mark the location where the deviation occurs on the real-scene 3D model, and output and display the real-scene 3D model.

[0041] Furthermore, in S100, the BIM design model of the target building is a static BIM three-dimensional model of the target building constructed by using BIM technology according to the design drawings of the target building.

[0042] Among them, the external protective structure includes curtain wall system and metal roof system. The curtain wall system includes metal curtain wall or glass curtain wall, supporting structure and connection sealing system. The supporting structure includes connecting parts or purlins or purlin supports. The connection sealing system includes silicone glue, sealing strips or hinges and bolts.

[0043] Furthermore, in S200, the method for simulating and obtaining the surface wind pressure distribution of the BIM design model through computer-aided design simulation software is: importing the BIM design model into the computer-aided design simulation software, performing flow field analysis under given wind speed boundary conditions to obtain the surface wind pressure distribution of the BIM design model, and the steps of the flow field analysis specifically include: importing the BIM design model, surface segmentation, meshing, and mesh property definition; setting material parameters, geometric properties, wind speed boundary conditions, load conditions, and defining the connectivity of the units.

[0044] Preferably, the flow field analysis is specifically as follows: in the peripheral flow field of the BIM design model of the target building, the transient wind pressure distribution under different wind direction angles (0°~360°, interval 22.5°) is simulated, and the wind pressure data is coupled to the structural finite element model; dynamic wind pressure load is applied, time history analysis is performed and resonance areas are identified, and the time history data (.csv format) of the surface wind pressure coefficient of the BIM design model is exported, and the wind pressure distribution data and displacement cloud map of the BIM design model are exported.

[0045] Furthermore, the BIM design model is in 3dm format.

[0046] Furthermore, in S100, it also includes segmenting and extracting the three-dimensional grid of the wall of the BIM design model, calculating the three-dimensional support plane fitted by the least squares method; calculating the Euclidean distance between the three-dimensional grid and the three-dimensional support plane, and normalizing it to generate a depth map of the wall; based on the calibrated camera parameters, generating a texture image of the wall through three-dimensional space texture mapping; the length and width dimensions of the depth map and texture map of the wall are kept consistent.

[0047] Preferably, the grid division method is a sweeping method (Sweep) grid division method.

[0048] The wind speed boundary conditions are given as 5m / s to 36.9m / s.

[0049] Preferably, different wind speeds are set at different heights of the BIM design model for simulation and given wind speed boundary conditions are used for simulation: the wind speed in the area of ​​0-10 meters corresponding to the BIM design model is set to 5m / s, the wind speed at a height of 10-30 meters is set to 8.7m / s, and the wind speed at a height of 30-90 meters is set to 15m / s; the wind speed above 90 meters is set to 36.9m / s.

[0050] Preferably, the computer-aided design simulation software is open source OpenFOAM software.

[0051] Further, in S300 , the different wind speed conditions are a wind speed range of 5 m / s to 36.9 m / s.

[0052] Furthermore, in S300, the method for identifying and marking the structural installation locations where risk-sensitive areas exist on the BIM design model is: Obtain the wind pressure values ​​corresponding to all points in each structural installation location of the BIM design model based on the surface wind pressure distribution, and record the vibration displacement of the geometric center of gravity of each structural installation location from the minimum wind pressure value to the maximum wind pressure value as the wind pressure limit displacement; Record the other structure installation positions adjacent to the current structure installation position as the adjacent structure installation positions of the current structure installation position; MaxCUP; MinCUP; CUPL; WV(j) and j; WV(j) is the jth wind pressure orientation area of the current structural installation position; j is the serial number, and the order of the wind pressure orientation area is the order of CUPL sequentially passing through the wind pressure orientation area; Determine each WV(j), and mark WV(j) as a structural installation position with a risk-sensitive area if WV(j) meets the linear wind pressure condition; wherein the linear wind pressure condition is MV(WV(j-1))>MV(WV(j)) and MV(WV(j))<MV(WV(j+1)), and MV(WV(j)) is the wind pressure limit displacement corresponding to WV(j).

[0053] wherein the key source code of the C# language description of the specific implementation of the method for identifying and marking the structural installation position with a risk-sensitive area on the BIM design model in Embodiment 1 is as follows: using System; using System.Collections.Generic; using System.Linq; using UnityEngine; / / Need the mathematical library of Unity engine, replace with other three-dimensional mathematical library public class CurtainWallRiskAnalyzer { / / Structural installation position data structure public class CurtainWallRegion {public int RegionId { get; set;} public Vector3 GeometricCenter { get; set;} / / Geometric center public List <vector3>SurfacePoints { get; set;} / / surface point set public float[] WindPressureValues ​​{ get; set;} / / Wind pressure values ​​at each point public float MaxWindPressureDisplacement { get; set;} / / Wind pressure limit displacement public List <int>AdjacentRegions { get; set;} / / Adjacent region ID list public bool IsAtRisk { get; set;} / / Risk flag} / / / Identify and mark structural installation locations where there are risk-sensitive areas / / / <param name="allRegions"> All structural installation location data public static void IdentifyRiskRegions(List <curtainwallregion>allRegions) { / / Step 1: Calculate the wind pressure limit displacement of all areas CalculateMaxDisplacements(allRegions); / / Step 2: Determine the wind pressure direction for each area Dictionary <int, List <curtainwallregion>>directionRegionsMap = new(); foreach (var region in allRegions) {var directionRegions = GetWindDirectionRegions(region, allRegions); directionRegionsMap.Add(region.RegionId, directionRegions);} / / Step 3: Check linear wind pressure conditions and mark risks foreach (var region in allRegions) {if (directionRegionsMap.TryGetValue(region.RegionId, out varwvList)) {for (int j = 1; j <wvList.Count - 1; j++) {if (CheckLinearCondition(wvList, j)) {wvList[j].IsAtRisk = true;Debug.Log($"Risk tag: Region {wvList[j].RegionId}"); }}}}} / / / Calculate the wind pressure limit displacement (center of gravity displacement from minimum wind pressure to maximum wind pressure) private static void CalculateMaxDisplacements(List <curtainwallregion>regions) {foreach (var region in regions) {if(region.WindPressureValues==null||region.WindPressureValues.Length==0) continue; / / Simulate displacement changes under different wind pressures float minPressure = region.WindPressureValues.Min(); float maxPressure = region.WindPressureValues.Max(); region.MaxWindPressureDisplacement = (maxPressure - minPressure) *0.1f; }} / / / Get the wind pressure direction area of ​​the current area (the area along the MinCUP->MaxCUP ray) private static List <curtainwallregion>GetWindDirectionRegions( CurtainWallRegion currentRegion, List <curtainwallregion>allRegions) { / / Find the extreme barycenter in the adjacent regions var adjacentRegions = allRegions.Where(r => currentRegion.AdjacentRegions.Contains(r.RegionId)).ToList(); if (adjacentRegions.Count == 0) return new List <curtainwallregion>(); var maxCUP = adjacentRegions.OrderByDescending(r =>r.MaxWindPressureDisplacement) .First().GeometricCenter; var minCUP = adjacentRegions.OrderBy(r =>r.MaxWindPressureDisplacement) .First().GeometricCenter; / / Construct ray direction Vector3 rayDirection = (maxCUP - minCUP).normalized; Ray cupRay = new Ray(minCUP, rayDirection); / / Ray detection of all areas passed through var piercedRegions = new List <curtainwallregion>(); foreach (var region in allRegions) {if (RayIntersectsRegion(cupRay, region)) piercedRegions.Add(region);} / / Sort by ray traversal order (simplified to sort by center of gravity distance) return piercedRegions.OrderBy(r => Vector3.Distance(minCUP, r.GeometricCenter)).ToList();} / / / Detect intersection between ray and structure installation position (using bounding box to simplify calculation) private static bool RayIntersectsRegion(Ray ray, CurtainWallRegionregion) {Bounds bounds = new Bounds(region.GeometricCenter, Vector3.zero); foreach (var point in region.SurfacePoints) bounds.Encapsulate(point); return bounds.IntersectRay(ray);} / / / Check the linear wind pressure condition MV(WV(j-1))>MV(WV(j)) <MV(WV(j+1)) private static bool CheckLinearCondition(List <curtainwallregion>wvRegions, int j) {if (wvRegions.Count<3 || j<= 0 || j>= wvRegions.Count - 1) return false; float prevMV = wvRegions[j - 1].MaxWindPressureDisplacement; float currentMV = wvRegions[j].MaxWindPressureDisplacement; float nextMV = wvRegions[j + 1].MaxWindPressureDisplacement; return prevMV>currentMV&¤tMV <nextMV;}}。

[0054] Furthermore, in S500, the accuracy of the corresponding positions of the structural installation positions of the real scene 3D model and the BIM design model in the risk-sensitive areas in the main structure installation stage, the enclosure structure skeleton installation stage, and the enclosure structure panel installation stage are respectively reviewed. The method for generating the deviation data between the on-site real scene model and the BIM design model through registration deviation analysis is as follows: The ICP algorithm is used to verify the accuracy of the corresponding positions of the structural installation positions in risk-sensitive areas between the real-life 3D model and the BIM design model, and the deviation data between the on-site real-life model and the BIM design model is generated through alignment deviation analysis.

[0055] The ICP algorithm is used for fusion analysis: the real-life 3D model is registered with the BIM design model, and the deviation data of the corresponding positions of the structural installation positions in risk-sensitive areas are calculated. The deviation data is color-coded to show the size deviation range of different areas. Among them, the deviation data includes overall deviation statistics (mean, standard deviation, maximum deviation), deviation heat map (color mapping distance value, intuitive display of error distribution), and structural installation position error analysis data in risk-sensitive areas: statistical deviations are grouped by structures (external protective structures) in risk-sensitive areas.

[0056] Among them, the ICP algorithm is the ICP registration algorithm in the open source 3D data processing library Open3D.

[0057] Furthermore, in S500 , the method for outputting and displaying the real-scene 3D model is: performing adaptive histogram equalization processing on the position where the deviation exists on the real-scene 3D model to enhance the local contrast.

[0058] Preferably, the method further includes transmitting the real-scene three-dimensional model to a client or a mobile device for display output.

[0059] Preferably, the method for transmitting the real scene three-dimensional model to a client or mobile device for display output comprises the following steps: processing the real scene three-dimensional model by an edge folding algorithm to obtain a simplified model; The simplified model is divided into 1m×1m blocks according to space and compressed separately; Dynamically push visible blocks based on the client or mobile device's field of view (using HTTP / 2 ServerPush).

[0060] When loading on the client or mobile device, detect the GPU performance and select the corresponding LOD level: High-end devices: LOD0 (original accuracy); Mid-range devices: LOD1 (50% mesh); Low-end devices: LOD2 (20% mesh + solid color heat map); The structural installation locations of risk-sensitive areas are highlighted in real time.

[0061] Among them, high-end devices: support hardware-accelerated ray tracing, floating-point computing power ≥ 3 TFLOPS, memory bandwidth ≥ 100GB / s; Mid-range devices: support Vulkan 1.1 and above API, floating point computing power of 1-3 TFLOPS; Low-end devices: only support OpenGL ES 3.0, floating point computing power < 1 TFLOPS.

[0062] Example 2 In this embodiment 2, the linear wind pressure condition is replaced on the basis of embodiment 1, specifically: Preferably, the linear wind pressure condition is: DE(WV(j-1))>DE(WV(j)) and DE(WV(j))<DE(WV(j+1)), Where DE(WV(j)) is the inertial wind pressure index of the j-th wind pressure direction area WV(j) at the current structure installation location; Preferably, the calculation method of the inertial wind pressure index DE(WV(j)) is: DE(WV(j))=|MinA1-MinA2|÷|MaxA1-MaxA2|; Among them, MaxA1 is the maximum value of the wind pressure limit displacement corresponding to the wind pressure direction areas from the 1st to the j-1th; MinA1 is the minimum value of the wind pressure limit displacement corresponding to the wind pressure direction areas from the 1st to the j-1th; MaxA2 is the maximum value of the wind pressure limit displacement corresponding to the wind pressure direction areas from the j+1th to the SuMth; MinA2 is the minimum value of the wind pressure limit displacement corresponding to the wind pressure direction areas from the j+1th to the SuMth; SuM is the number of wind pressure direction areas.

[0063] The key source code of the C# language description of the specific implementation of the linear wind pressure condition of this embodiment 2 is: public class WindPressureOptimizer { / / / Check whether the linear wind pressure condition is met / / / DE(WV(j-1))>DE(WV(j)) and DE(WV(j))<DE(WV(j+1)) / / / <param name="deValues"> Array of inertial wind pressure indices for all regions (index = region number - 1) / / / <param name="currentRegion"> Current region j (starting from 1) public static bool IsLinearWindPressureCondition(double[] deValues,int currentRegion) {if (deValues ​​== null || currentRegion<2 || currentRegion>=deValues.Length) return false; return deValues[currentRegion - 2]>deValues[currentRegion - 1]&& deValues[currentRegion - 1] <deValues[currentRegion];} / / / Calculate the improved inertial wind pressure index DE(WV(j)) = |MinA1-MinA2| / |MaxA1-MaxA2| / / / <param name="windLoadLimits"> Wind pressure limit displacement array (index = area number - 1) / / / <param name="windDirectionRegions"> Wind pressure direction area dictionary (Key=area number, Value=corresponding direction area list) / / / <param name="currentRegion"> Current region j (starting from 1) public static double CalculateEnhancedInertialIndex( double[] windLoadLimits, Dictionary <int, List <int>>windDirectionRegions, int currentRegion) {if (windLoadLimits == null || windLoadLimits.Length == 0) throw new ArgumentException("Invalid wind pressure limit displacement data"); if (currentRegion<1 || currentRegion>windLoadLimits.Length) throw new ArgumentException("Area number out of range"); / / Get all wind pressure direction areas in the current area var currentDirectionRegions = windDirectionRegions[currentRegion]; if (currentDirectionRegions == null || currentDirectionRegions.Count== 0) return 0; / / Step 1: Extract the displacement values ​​of the previous j-1 regions belonging to the current direction region var previousRegions = Enumerable.Range(0, currentRegion - 1) .Where(i =>windDirectionRegions[i + 1].Intersect(currentDirectionRegions).Any()).Select(i =>windLoadLimits[i]) .ToList(); / / Step 2: Extract the displacement values ​​of the current direction area from the next j+1 to SuM areas var nextRegions = Enumerable.Range(currentRegion, windLoadLimits.Length - currentRegion).Where(i => windDirectionRegions[i + currentRegion].Intersect(currentDirectionRegions).Any()).Select(i => windLoadLimits[i]).ToList(); / / Calculate extreme values double minA1 = previousRegions.Any()? previousRegions.Min() : 0; double maxA1 = previousRegions.Any()? previousRegions.Max() : 0; double minA2 = nextRegions.Any()? nextRegions.Min() : 0; double maxA2 = nextRegions.Any()? nextRegions.Max() : 0; / / Prevent division by zero (return 0 or a specific default value when MaxA1 == MaxA2) double denominator = Math.Abs(maxA1 - maxA2); if (denominator < 1e-10) / / floating point precision tolerance return 0; return Math.Abs(minA1 - minA2) / denominator;} public static List <int>FindOptimalRegions(double[] windLoadLimits,Dictionary <int, List <int>>directionRegions) {var deValues = new double[windLoadLimits.Length]; for (int j = 1; j<= windLoadLimits.Length; j++) {deValues[j-1]=CalculateEnhancedInertialIndex(windLoadLimits,directionRegions, j);} var optimalRegions = new List <int>(); for (int j = 2; j <windLoadLimits.Length; j++) {if (IsLinearWindPressureCondition(deValues, j)) optimalRegions.Add(j);} return optimalRegions;}}.

[0064] Example 3 In this embodiment 3, the calculation method of the inertial wind pressure index DE(WV(j)) is replaced on the basis of embodiment 1, specifically: Preferably, the calculation method of the inertial wind pressure index DE(WV(j)) is: Calculate the convergent wind pressure difference and the transfer wind pressure difference at all structural installation locations. The specific method is: The convergence wind pressure difference is the difference between the average value of the wind pressure limit displacements corresponding to the wind pressure direction area from the 1st to the j-1th structural installation positions and the corresponding wind pressure limit displacement at the structural installation position; the transfer wind pressure difference is the difference between the average value of the wind pressure limit displacements corresponding to the wind pressure direction area from the j+1th to the Nth structural installation positions and the corresponding wind pressure limit displacement at the structural installation position; (the transfer wind pressure difference indicates the degree to which the wind pressure limit displacement increases from the structural installation position along the wind pressure direction area, and the convergence wind pressure difference indicates the degree to which the wind pressure limit displacement decreases from the structural installation position along the wind pressure direction area). Traverse the wind pressure direction areas corresponding to each structural installation position, and use MaxS to represent the number of wind pressure direction areas where the convergent wind pressure difference of each wind pressure direction area corresponding to the structural installation position is greater than the transmission wind pressure difference; use MinS to represent the number of wind pressure direction areas where the convergent wind pressure difference of each wind pressure direction area corresponding to the structural installation position is less than the transmission wind pressure difference; then the calculation method of the inertial wind pressure index DE(WV(j)) is: inertial wind pressure index DE(WV(j)) = MaxS ÷ (MaxS + MinS).

[0065] The key source code of the C# language description of the specific implementation of the calculation method of the inertial wind pressure index DE(WV(j)) of this embodiment 3 is: public class WindPressureCalculator { / / / Calculate the inertial wind pressure index DE(WV(j)) at the specified structure installation location j / / / <param name="windLoadLimits"> Array of wind pressure limit displacement values ​​for all structural installation locations (subscript = area number - 1) / / / <param name="windDirectionRegions"> Wind pressure towards area dictionary for all structural mounting locations (Key=Area number, Value=List of corresponding towards areas) / / / <param name="j"> Current calculated structural mounting location number (starting from 1) public static double CalculateInertialIndex( double[] windLoadLimits, Dictionary<int, List <int>>windDirectionRegions,int j) { / / Parameter verification if (windLoadLimits == null || windLoadLimits.Length == 0) throw new ArgumentException("The wind pressure limit displacement array is invalid"); if (j<1 || j>windLoadLimits.Length) throw new ArgumentException("Structure installation position number is out of range"); int maxS = 0; / / The number of orientation areas where the convergent wind pressure difference is greater than the transmission wind pressure difference int minS = 0; / / The number of orientation areas where the convergent wind pressure difference is less than the transmission wind pressure difference / / Traverse all wind pressure direction areas in the current area j foreach (int directionRegion in windDirectionRegions[j]) { / / Calculate the convergence wind pressure difference (the difference between the average of the previous j-1 areas and the current area) double convergenceDiff = CalculateConvergenceDifference(windLoadLimits, j, directionRegion); / / Calculate the transmission wind pressure difference (the difference between the average of the next j+1 to N areas and the current area) double transmissionDiff = CalculateTransmissionDifference(windLoadLimits, j, directionRegion); / / Statistics MaxS and MinS if (convergenceDiff>transmissionDiff)maxS++; else if (convergenceDiff <transmissionDiff)minS++;} / / Calculate the inertial wind pressure index (avoid division by zero) return (maxS + minS) == 0 ? 0 : (double)maxS / (maxS + minS);} / / / Calculate the convergence wind pressure difference (the difference between the average wind pressure limit displacement of the previous j-1 areas and the current area j) private static double CalculateConvergenceDifference( double[] windLoadLimits, int currentRegion, int directionRegion) {double sum = 0; int count = 0; / / Accumulate the displacement values ​​of the previous 1 to j-1 areas belonging to the current direction area for (int i = 0; i <currentregion - 1; i++) 注意数组下标从0开始 {sum +="windLoadLimits[i];count++;}" double average="count">0 ? sum / count : 0; return average - windLoadLimits[currentRegion - 1]; / / Note array subscript conversion} / / / Calculate the transmitted wind pressure difference (the difference between the average wind pressure limit displacement of the next j+1 to N areas and the current area j) private static double CalculateTransmissionDifference( double[] windLoadLimits, int currentRegion, int directionRegion) {double sum = 0; int count = 0; / / Accumulate the displacement values ​​of the current direction area in the area from j+1 to the end for (int i = currentRegion; i <windLoadLimits.Length; i++) / / i对应区域编号i+1 { / / It is also necessary to determine whether area i+1 belongs to the current direction area sum += windLoadLimits[i];count++;} double average = count>0 ? sum / count : 0; return average - windLoadLimits[currentRegion - 1];}}.

[0066] Furthermore, construction engineers can mark the corresponding positions of the deviations on the real-life 3D model and perform any of the following constructions: setting expansion joint nodes, using offset connectors, and selecting highly elastic sealing materials as reinforced sealant applications.

[0067] Although the description of the present disclosure has been quite detailed and specifically describes several embodiments, it is not intended to be limited to any of these details or embodiments or any particular embodiment, so as to effectively cover the intended scope of the present disclosure. In addition, the above description of the present disclosure is based on the embodiments foreseen by the inventors, which is intended to provide a useful description, and those non-substantial changes to the present disclosure that have not yet been foreseen may still represent equivalent changes to the present disclosure.< / currentregion> < / int> < / int> < / int> < / int> < / int> < / curtainwallregion> < / curtainwallregion> < / curtainwallregion> < / curtainwallregion> < / curtainwallregion> < / curtainwallregion> < / curtainwallregion> < / curtainwallregion> < / int>

Claims

1. A method for detecting the installation of a building's irregular multi-curved exterior enclosure structure using three-dimensional reconstruction by an unmanned aerial vehicle, characterized in that: The method comprises the following steps: S100, importing the BIM design model of the target building, marking the location of the external enclosure structure on the BIM design model as the structure installation location; S200, obtains the surface wind pressure distribution of the BIM design model through simulation with computer-aided design simulation software; S300: Under different wind speed conditions, the surface wind pressure distribution is loaded into the BIM design model to obtain the vibration displacement of each structural installation location under different average wind speeds. The structural installation locations with risk-sensitive areas are identified and marked on the BIM design model. S400 uses a drone to capture a sequence of images of the target building and performs 3D reconstruction on the image sequence to obtain a realistic 3D model of the main structure of the target building. S500: During the main structure installation phase, the enclosure structure skeleton installation phase, and the enclosure structure panel installation phase, the accuracy of the corresponding positions of the structural installation positions in the risk-sensitive areas of the real-life 3D model and the BIM design model are reviewed, and deviation data between the on-site real-life model and the BIM design model is generated through registration deviation analysis; S600: Marking the locations where the deviations exist on the real-scene 3D model, and outputting and displaying the real-scene 3D model; In S300, the specific method for identifying and marking the structural installation locations with risk-sensitive areas on the BIM design model is as follows: Mark the wind pressure direction area in the adjacent structure installation position of the current structure installation position; Screen out the structural installation locations in risk-sensitive areas in the wind pressure direction area based on the linear wind pressure conditions; Among them, the linear wind pressure condition is: let WV(j) be the j-th wind pressure direction area of ​​the current structure installation position; j is the sequence number, MV(WV(j-1))>MV(WV(j)) and MV(WV(j))<MV(WV(j+1)), where MV(WV(j)) is the wind pressure limit displacement corresponding to WV(j).

2. The method for detecting the installation of a building's special-shaped and multi-curved external enclosure structure using three-dimensional reconstruction by an unmanned aerial vehicle according to claim 1, characterized in that: The method for marking the wind pressure direction area in the adjacent structure installation position of the current structure installation position is: Obtain the wind pressure values ​​corresponding to all points in each structural installation location of the BIM design model based on the surface wind pressure distribution, and record the vibration displacement of the geometric center of gravity of each structural installation location from the minimum wind pressure value to the maximum wind pressure value as the wind pressure limit displacement; Record the other structure installation positions adjacent to the current structure installation position as the adjacent structure installation positions of the current structure installation position; The geometric center of gravity of the structure installation position with the largest wind pressure limit displacement among all adjacent structure installation positions of the current structure installation position is recorded as MaxCUP; the geometric center of gravity of the structure installation position with the smallest wind pressure limit displacement among all adjacent structure installation positions of the current structure installation position is recorded as MinCUP; then a ray CUPL is formed from MinCUP to MaxCUP; all the structure installation positions sequentially passed by CUPL are recorded as the wind pressure direction area of ​​the current structure installation position.

3. The method for detecting the installation of a building's special-shaped and multi-curved external enclosure structure using three-dimensional reconstruction by an unmanned aerial vehicle according to claim 2, characterized in that: Replace the linear wind pressure condition with: DE(WV(j-1))>DE(WV(j)) and DE(WV(j))<DE(WV(j+1)), Where DE(WV(j)) is the inertial wind pressure index of the j-th wind pressure direction area WV(j) at the current structure installation location.

4. The method for detecting the installation of a building's special-shaped and multi-curved external enclosure structure using drone 3D reconstruction according to claim 3, characterized in that: The calculation method of inertial wind pressure index DE(WV(j)) is: DE(WV(j))=|MinA1-MinA2|÷|MaxA1-MaxA2|; Among them, MaxA1 is the maximum value of the wind pressure limit displacement corresponding to the wind pressure direction areas from the 1st to the j-1th; MinA1 is the minimum value of the wind pressure limit displacement corresponding to the wind pressure direction areas from the 1st to the j-1th; MaxA2 is the maximum value of the wind pressure limit displacement corresponding to the wind pressure direction areas from the j+1th to the SuMth; MinA2 is the minimum value of the wind pressure limit displacement corresponding to the wind pressure direction areas from the j+1th to the SuMth; SuM is the number of wind pressure direction areas.

5. The method for detecting the installation of a building's special-shaped and multi-curved external enclosure structure using three-dimensional reconstruction by an unmanned aerial vehicle according to claim 4, characterized in that: The calculation method of inertial wind pressure index DE(WV(j)) is replaced by: Calculate the convergent wind pressure difference and the transfer wind pressure difference at all structural installation locations. The specific method is: The convergence wind pressure difference is the difference between the average value of the wind pressure limit displacement corresponding to the wind pressure direction area from the 1st to the j-1th structure installation position and the wind pressure limit displacement corresponding to the structure installation position; The difference between the average value of the wind pressure limit displacement corresponding to the wind pressure direction area from the j+1th to the Nth structure installation position and the wind pressure limit displacement corresponding to the structure installation position is the transmitted wind pressure difference; Traverse the wind pressure direction areas corresponding to each structural installation position, and use MaxS to represent the number of wind pressure direction areas where the convergent wind pressure difference of each wind pressure direction area corresponding to the structural installation position is greater than the transmission wind pressure difference; use MinS to represent the number of wind pressure direction areas where the convergent wind pressure difference of each wind pressure direction area corresponding to the structural installation position is less than the transmission wind pressure difference; then the calculation method of the inertial wind pressure index DE(WV(j)) is: inertial wind pressure index DE(WV(j)) = MaxS ÷ (MaxS + MinS).

6. The method for detecting the installation of a building's special-shaped and multi-curved external enclosure structure using drone 3D reconstruction according to claim 1, characterized in that: In S500 , the method for outputting and displaying the real-scene three-dimensional model is: performing adaptive histogram equalization processing on the position where the deviation exists on the real-scene three-dimensional model to enhance the local contrast.

7. The method for detecting the installation of a building's special-shaped and multi-curved external enclosure structure using three-dimensional reconstruction by an unmanned aerial vehicle according to claim 1, characterized in that: It also includes transmitting the real-scene three-dimensional model to a client or mobile device for display output.

8. The method for detecting the installation of a building's special-shaped and multi-curved external enclosure structure using three-dimensional reconstruction by an unmanned aerial vehicle according to claim 7, characterized in that: The method for transmitting a real-scene three-dimensional model to a client or mobile device for display output comprises the following steps: processing the real-scene three-dimensional model by an edge folding algorithm to obtain a simplified model; Divide the simplified model into blocks according to space and compress them separately; Dynamically push visible blocks based on the client or mobile device's field of view; When loading on the client or mobile device, detect the GPU performance and select the corresponding LOD level.

Images