Digital twin design method, system, storage medium and terminal for garden landscape

Abstract

The present invention discloses a digital twin design method and system for garden landscape, the method includes S100: obtaining the geomorphic data of the planning area through drone aerial survey, and integrating it with GIS geographic information to obtain the physical module of the terrain data of the planning area; S200: obtaining the garden landscape design scheme according to the terrain data of the planning area, and establishing the BIM model of each garden landscape element; S300: establishing the landscape area spatial attribute DTLD model, realizing the fusion simulation of the BIM model of each garden landscape element and the terrain data of the corresponding planning block; S400: realizing the immersive experience of the fusion simulation design scheme through interactive equipment, and evaluating whether the design scheme meets the design requirements. The method of the present invention integrates and simulates the BIM model with the actual physical model, and obtains the simulation effect through real-time simulation calculation, which greatly improves the degree of approximation between the design scheme and the real construction effect, and solves a series of problems such as the difficulty of traditional two-dimensional or three-dimensional static landscape garden design to simulate the real construction effect.

Description

Technical Field

[0001] The present invention belongs to the technical field of garden landscape design, and more specifically, relates to a digital twin design method, system, storage medium and terminal for garden landscape. Background Art

[0002] In the design of urban garden landscape, in the overall layout arrangement of large geographical area, complex geological conditions, various functional zones, road network and the location of their ancillary garden facilities, in addition to the plane layout relationship that must meet the planning and design requirements, it should also take into account the overall planning of the garden, the overall plane layout and the coordination, ornamental, artistry, comfort and other uncertain factors between the landscapes such as trees, flowers, grass, stones, lakes, historical sites and buildings.

[0003] Traditional scenic spot design relies on two-dimensional contour maps to analyze and design the original site in the brain. The spatial relationship of landscape areas reflected by GIS information data is crucial for describing and understanding the physical environment of landscape design. The use of BIM technology can effectively control various problems that may arise in the design stage and promote the development of the landscape industry. At present, China's infrastructure construction is large and the construction industry is developing rapidly, but at the same time, the construction industry needs sustainable development, and construction companies are also facing more severe competition. In this context, the advantages of BIM technology such as visualization, coordination, simulation, and optimization are becoming more and more prominent.

[0004] However, the current conventional BIM 3D modeling method cannot fully describe and use this feature, especially it cannot match the landscape area GIS information with models of trees, flowers, grass, stones, lakes, monuments, buildings, etc. in real time for simulation analysis, making it difficult for designers to adjust the design plan in real time according to needs. Moreover, when the site is large and the terrain conditions are complex, it is difficult to coordinate the overall layout with manpower alone, which may lead to problems such as incoordination between the terrain design and the surrounding environment, poor drainage, large differences in earthwork excavation and filling volume, increased design costs, and reduced engineering efficiency. Summary of the invention

[0005] In view of the above defects or improvement needs of the prior art, the present invention provides a digital twin design method, system, storage medium and terminal for garden landscape, establishes node coding information of different landscape types, collects numerical attributes such as length, width, height and area of ​​broad-leaved forests, forest trails, lawns, flower beds, pavilions, streams and ponds, establishes BIM models of broad-leaved forests, forest trails, lawns, flower beds, pavilions, streams and ponds, and then establishes a topological association between the BIM model and the geographic information of the planned area surveyed by unmanned aerial vehicles to achieve the fusion of BIM models and GIS information of broad-leaved forests, forest trails, lawns, flower beds, pavilions, streams and ponds, and integrates the BIM model with the actual physical model for simulation. After the user makes a modification suggestion, the design scheme can be revised in real time, and the simulation calculation can be performed in real time to obtain the simulation effect, which greatly improves the degree of approximation between the design scheme and the actual construction effect, and solves a series of problems such as the difficulty of traditional two-dimensional or three-dimensional static landscape garden design in simulating the actual construction effect.

[0006] In order to achieve the above object, according to one aspect of the present invention, a digital twin design method for garden landscape is provided, comprising the following steps:

[0007] S100: Obtain the topographic data of the planning area through drone aerial survey and integrate it with GIS geographic information to obtain the physical module of the terrain data of the planning area;

[0008] S200: Obtain the garden landscape design plan based on the terrain data of the planning area and establish the BIM model of each garden landscape element;

[0009] S300: Establish a DTLD model of landscape area spatial attributes to achieve fusion simulation of the BIM model of each garden landscape element and the terrain data of the corresponding planning block;

[0010] S400: An immersive experience of the integrated simulation design scheme is achieved through interactive devices to evaluate whether the design scheme meets the design requirements. If not, steps two and three are repeated, feedback is provided to modify the BIM model and the integrated simulation calculation is re-performed and updated. If the requirements are met, the final design scheme is obtained.

[0011] Furthermore, step S100 includes:

[0012] S101: Survey the planned area, determine the time for drone aerial survey based on the survey results, plan the drone route, collect aerial photography data, test and process the collected photography data, and obtain the aerial survey terrain of the planned area;

[0013] S102: Determine the depth sounder detection time according to the survey results, collect the mountain range, terrain distribution, water range, water depth value and coordinates of the planned area to be measured, and obtain the plane coordinates and elevation of the planned area to be measured according to the coordinates;

[0014] S103: The mountain range, terrain distribution, water range, water depth and coordinates are integrated to obtain the overall terrain distribution of the planning area, which is then integrated with the GIS geographic information to obtain the final mapping results of the planning area.

[0015] Furthermore, step S101 includes testing the flight quality and image quality after the flight is completed, making up for unqualified routes and leaky areas, generating a three-dimensional real-scene model using aerial triangulation and regional network adjustment, and correcting the drone aerial survey results based on the measurement results selected on the three-dimensional real-scene model to obtain the aerial survey terrain of the planned area.

[0016] Further, step S200 includes:

[0017] S201: Establish node attributes of each landscape design element for:

[0018]

[0019] Where i represents the type of landscape design element, j represents the number of each landscape design element, G represents the attribute of the jth element in the i-th landscape design element, and [ID01]-[IDij] represents the coding of the landscape design element;

[0020] S202: Constructing numerical attributes of each landscape design element for:

[0021]

[0022] Among them, V represents the geometric attribute information of the jth element in the i-type landscape design element.

[0023] Further, step S300 includes:

[0024] S301: Analyze the spatial attributes of the planning area, including:

[0025]

[0026] Among them, Z p is the coordinate of the interpolated elevation point; n is the number of interpolated elevation points; P i is the weight of the i-th data point; Z i is the elevation of the ith data point;

[0027] S302: Correct the measurement results:

[0028]

[0029] Among them, Δh is the correction value of the measurement result; hf It is a drone that surveys terrain; h d The underwater terrain detected by the depth sounder.

[0030] Further, step S300 includes:

[0031] S303: Establish the topological relationship between the node attributes, numerical attributes and planning areas of each landscape design element, namely the DTLD model, which is defined as:

[0032]

[0033] Further, step S300 includes:

[0034] S304: setting the number of sampling with replacement is to initialize the new training set to an empty set, and perform the sampling operation number of times in a loop. Each time the sampling operation is performed, a pair of samples is randomly selected from the landscape BIM and the terrain data of the planning area, and a copy of the sample is copied and appended to the new training set;

[0035] S305: construct an algorithm parameter search set, each element in the set has a set of specific parameters that match the algorithm, randomly divide the samples in the training set into a training set with a ratio of 75% and a validation set with a ratio of 25%, and obtain the landscape BIM and planning area terrain topology function;

[0036] S306: defining a loss function machine evaluation function for matching landscape area terrain data with the BIM model, using the parameters to carry out training on the training set, and calculating the classification accuracy on the validation set;

[0037] S307: If the classification accuracy does not meet the requirements, return to step S304; if the classification accuracy meets the requirements, the training ends, and the parameters corresponding to the highest classification accuracy on the validation set are screened out; and the classification result value of the topological relationship between the landscape BIM and the planned area terrain is obtained.

[0038] According to a second aspect of the present invention, a digital twin design system for garden landscape is provided, which is applied to the digital twin design method for garden landscape described above, and includes a three-layer architecture of a data assurance layer, a modeling and calculation layer, and an immersive experience layer; wherein,

[0039] The data security layer includes a GIS geographic information data acquisition module, a landscape design data module, a high-data data transmission module, and a data management module. The GIS geographic information data acquisition module is used to realize aerial survey of the landscape planning area by means of drones and other means to obtain aerial survey terrain and geomorphic data. The landscape design data module is used to establish BIM models of various landscape types. The high-data data transmission module is used to realize real-time transmission and interaction of terrain data in the planning area and BIM model data of each landscape. The data management module stores, reads, modifies, and updates the PB-level big data generated by the digital twin design system.

[0040] The modeling and calculation layer includes a landscape area geography, landform environment physical module, a landscape object BIM module and a matching algorithm module. The landscape area geography, landform environment physical module is used to fuse the corresponding GIS information of the planning area, match and fuse the aerial survey terrain and landform data with the GIS position coordinates, and obtain the physical model of the actual terrain data. The landscape object BIM module is used to establish the BIM model of broad-leaved forests, forest trails, lawns, flower beds, pavilions, streams, and ponds. The matching algorithm module is used to fuse and simulate the BIM model with the actual physical model, and after the user proposes modification suggestions, the scheme design can be revised in real time, and the simulation calculation can be carried out in real time;

[0041] The immersive experience layer includes a three-dimensional virtual mapping module, a user interaction module and a virtual digital cluster module. The three-dimensional virtual mapping module is used to perform three-dimensional mapping between the physical model and the virtual model simulation calculation results to obtain a design scheme rendering that is close to the real environment. The user interaction module is used to interact with the design scheme through a virtual reality wearable device, modify the scheme, calculate in real time, update, and display. The virtual digital cluster module realizes multi-user collaborative interaction, design scheme review, modification, simulation calculation and update.

[0042] According to a third aspect of the present invention, there is provided an electronic device, comprising:

[0043] At least one processor, at least one memory and a communication interface; wherein,

[0044] The processor, memory and communication interface communicate with each other;

[0045] The memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to perform the method.

[0046] According to a fourth aspect of the present invention, there is provided a non-transitory computer-readable storage medium storing computer instructions, the computer instructions causing the computer to execute the method described.

[0047] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0048] 1. The method of the present invention establishes node coding information of different landscape types, collects numerical attributes such as length, width, height, and area of ​​broad-leaved forests, forest paths, lawns, flower beds, pavilions, streams, and ponds, establishes BIM models of broad-leaved forests, forest paths, lawns, flower beds, pavilions, streams, and ponds, and then establishes a topological association between the BIM model and the geographic information of the planned area surveyed by unmanned aerial vehicles to achieve the fusion of the BIM model of broad-leaved forests, forest paths, lawns, flower beds, pavilions, streams, and ponds with GIS information, and integrates the BIM model with the actual physical model for simulation. After the user makes a modification suggestion, the scheme design can be revised in real time, and the simulation calculation can be performed in real time to obtain the simulation effect, which greatly improves the degree of approximation between the design scheme and the actual construction effect, and solves a series of problems such as the difficulty of traditional two-dimensional or three-dimensional static landscape garden design in simulating the actual construction effect.

[0049] 2. The method of the present invention realizes the real-time transmission and interaction of the terrain data of the planning area and the BIM model data of each landscape, especially solves the problem of data delay, ensures the real-time fusion, interaction and update of the terrain data and the BIM model data of each landscape, solves the problems of high-speed data collection, transmission and calculation of the digital twin system, and at the same time cooperates with the data management module to store, read, modify and update the PB-level big data generated by the digital twin design system, providing key technical guarantee for the digital twin design system.

[0050] 3. In the method of the present invention, the matching algorithm module is used to integrate the BIM model with the actual physical model for simulation, and after the user makes a modification suggestion, the design scheme can be revised in real time, and the simulation calculation can be performed in real time to realize the modeling and high-speed calculation of the digital twin design system.

[0051] 4. In the method of the present invention, the three-dimensional virtual mapping module is used to perform three-dimensional mapping between the physical model and the simulation calculation results of the virtual model to obtain a design scheme rendering that is close to the real environment. The user interaction module is used to realize the interaction between the user and the design scheme, scheme modification, real-time calculation, update, display and other functions through virtual reality wearable devices such as VR glasses, VR helmets, etc. or intelligent interactive systems, and cooperate with the virtual digital cluster module to realize multi-user collaborative interaction, design scheme review, modification, simulation calculation and update, realize the integration of garden landscape aerial survey, GIS, BIM modeling, design, simulation, interaction and modification, and greatly improve the quality and efficiency of garden landscape design. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 This is a schematic diagram of the digital twin design system architecture of the garden landscape according to an embodiment of the present invention;

[0053] Figure 2 This is an overall flow chart of the digital twin design method for garden landscape according to an embodiment of the present invention;

[0054] Figure 3 This is a flow chart of the DTLD topology algorithm for the landscape area according to an embodiment of the present invention;

[0055] Figure 4 This is a general schematic diagram of a digital twin design case of a landscape area according to an embodiment of the present invention;

[0056] Figure 5 This is a schematic diagram of a top-down view of human-computer interaction in a digital twin design case of a landscape area according to an embodiment of the present invention;

[0057] Figure 6 This is a partially enlarged schematic diagram of human-computer interaction in a digital twin design case of a landscape area according to an embodiment of the present invention;

[0058] Figure 7 The figure is a schematic diagram of a design scheme of a garden landscape according to an embodiment of the present invention. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0060] like Figure 1As shown, the embodiment of the present invention provides a digital twin design system for garden landscape. The system generally includes a three-layer architecture of a data security layer, a modeling and calculation layer, and an immersive experience layer. The data security layer realizes the acquisition of aerial survey terrain data of the planning area and the fusion of GIS geographic information to obtain the terrain data of the planning area, establishes a BIM model of the landscape, and realizes the fusion simulation with the terrain data of the planning area through the DTLD topology algorithm. After the user proposes a modification suggestion, the scheme design can be revised in real time, and the simulation calculation can be carried out in real time to obtain the simulation effect, which greatly improves the degree of approximation between the design scheme and the real construction effect, and solves a series of problems such as the difficulty of traditional two-dimensional or three-dimensional static landscape garden design to simulate the real construction effect. Among them, the data security layer includes a GIS geographic information data acquisition module, a landscape design data module, a high-number data transmission module, and a data management module. The GIS geographic information data acquisition module is used to realize aerial survey of the landscape planning area through drones and other means to obtain aerial survey terrain data. The landscape design data module is used to establish BIM models of various landscape types, such as establishing node coding information of different landscape types, collecting the length, width, height, area and other numerical attributes of broad-leaved forests, forest trails, lawns, flower beds, pavilions, streams, ponds, etc. The high-data data transmission module is used to realize the real-time transmission and interaction of the terrain data of the planning area and the BIM model data of each landscape, especially to solve the problem of data delay, ensure the real-time fusion, interaction and update of terrain data and BIM model data of each landscape, solve the problem of high-speed data collection, transmission and calculation of digital twin system, and cooperate with the data management module to store, read, modify and update the PB-level big data generated by the digital twin design system, providing key technical guarantee for the digital twin design system. The modeling and calculation layer includes the landscape area geography, landform environment physical module, landscape object BIM module and matching algorithm module. Among them, the physical module of landscape area geography and landform environment is used to integrate the corresponding GIS information of the planning area, match and integrate the aerial survey terrain and landform data with the GIS location coordinates, and obtain the physical model of the actual terrain data; the landscape object BIM module is used to establish the BIM model of broad-leaved forests, forest trails, lawns, flower beds, pavilions, streams, and ponds. The matching algorithm module is used to integrate and simulate the BIM model with the actual physical model, and after the user proposes modification suggestions, the scheme design can be revised in real time, and the real-time simulation calculation can be realized to realize the modeling and high-speed calculation of the digital twin design system. The immersive experience layer includes a three-dimensional virtual mapping module, a user interaction module, and a virtual digital cluster module.Among them, the three-dimensional virtual mapping module is used to perform three-dimensional mapping of the physical model and the simulation calculation results of the virtual model to obtain a design scheme rendering that is close to the real environment. The user interaction module is used to realize the interaction between users and design schemes, scheme modification, real-time calculation, update, display and other functions through virtual reality wearable devices such as VR glasses, VR helmets, etc. or intelligent interactive systems, and cooperate with the virtual digital cluster module to realize multi-user collaborative interaction, design scheme review, modification, simulation calculation and update, and realize the integration of garden landscape aerial survey, GIS, BIM modeling, design, simulation, interaction and modification, greatly improving the quality and efficiency of garden landscape design.

[0061] like Figure 2 and Figure 3 As shown, an embodiment of the present invention provides a digital twin design method for garden landscape, comprising the following steps:

[0062] Step 1: Obtain the topographic data of the planning area through UAV aerial survey and integrate it with GIS geographic information to obtain the physical module of the terrain data of the planning area;

[0063] Survey the planning area, determine the time for drone aerial survey based on the survey results, plan the drone route, collect aerial photography data, test and process the collected photography data internally to obtain the aerial survey terrain of the planning area; determine the time for depth sounder detection based on the survey results, collect the mountain range, terrain distribution, water range, water depth and coordinates of the planning area to be measured, obtain the plane coordinates and elevation of the planning area to be measured based on the coordinates, integrate the mountain range, terrain distribution, water range, water depth and coordinates to obtain the overall terrain distribution of the planning area, and then integrate it with GIS geographic information to obtain the final mapping results of the planning area.

[0064] When using drone aerial photogrammetry for topographic mapping, 4RTK drones can be used for topographic mapping, and the flight implementation is designed according to the 1:500 scale mapping requirements and drone flight related requirements. The flight altitude is designed to be 100 meters, the heading overlap is 80%, and the lateral overlap is 70%. When mapping the water terrain, the GNSS-RTK+echo sounder combination is used for underwater topography measurement. For example, the R10 GNSS receiver and the HY1602 dual-frequency echo sounder can be used, and the echo sounder probe is installed on the ship's side. During the operation, the echo sounder needs to continuously collect and save the water depth value; the computer synchronously collects the water depth value and positioning data, and the computer water depth collection record is accurate to 0.01m. After the operation is completed, the water depth data is processed by post-processing software to eliminate gross errors and some erroneous water depth values, and the processed water depth value is fused with the GNSS-RTK elevation to obtain the bottom elevation value. The GNSS-RTK plane coordinates are added to the bottom elevation value to obtain the final underwater topography results. Determine the weather conditions on the day of the operation for aerial surveying. After the flight is completed, check the flight quality and image quality, make up for unqualified routes and leaky areas, use aerial triangulation and regional network adjustment to generate a three-dimensional real-life model, and use the measurement results selected on the three-dimensional real-life model as a benchmark to correct the drone aerial survey results and obtain the aerial survey terrain of the planning area. Generate the drone aerial survey terrain in accordance with the technical requirements of the aerial survey industry. Using the drone aerial survey terrain as a benchmark, calculate the correction value of the underwater terrain measurement results based on the overlapping area of ​​the aerial survey terrain and the underwater terrain, and splice and merge the corrected mountain terrain and water terrain with the GIS geographic location information to obtain the final mapping results of the planning area.

[0065] Step 2: Obtain the landscape design plan based on the terrain data of the planning area and establish the BIM model of each landscape element, including:

[0066] Step (1): Establish node attributes of each landscape design element

[0067] Since the BIM model has information completeness, it is very important to extract reasonable landscape design elements for the construction of the BIM model retrieval and classification platform. The basic landscape area geographical element information extracted by the present invention is shown in Table 1, where: ID is the identification number of each element in the BIM model; coordinates are used to judge the relative position of each element in the model, so for objects such as mountains, water bodies, and soil bodies, attention should be paid to the horizontal and vertical axis coordinates. The extraction of the above elements can help identify the geographic spatial characteristics of the landscape area and provide necessary data support for constructing topological attributes and domain attributes of fused topology.

[0068]

[0069]

[0070] The DTLD model defines the node attributes of each landscape design element as follows:

[0071]

[0072] Among them, i represents the type of landscape design element, j represents the number of each landscape design element, G represents the attribute of the jth element in the i-th landscape design element, and [ID01]-[IDij] represents the code of the landscape design element. Among them, building mountains includes but is not limited to garden mountains, hall mountains, building mountains, pavilion mountains, study mountains, pool mountains, inner room mountains, cliff mountains, rock pools, peaks, mountains, rocks, caves, streams, winding water, waterfalls and other forms. There are many types of seedlings, which can be flexibly selected and modeled, such as red maple leaves, verdant green bamboo, white magnolia, purple crape myrtle, spotted grain elm, etc.; buildings can refer to a variety of forms including palaces, pavilions, buildings, halls, halls, pavilions, and studios.

[0073] Step (2): Construct the numerical attributes of each landscape design element

[0074] The numerical attribute definition formula for constructing each landscape design element is:

[0075]

[0076] Among them, V represents the geometric properties (such as perimeter, area, coordinates, etc.) of the jth element in the i-type landscape design element.

[0077] Step 3: Establish a DTLD model of landscape area spatial attributes to achieve the fusion simulation of the BIM model of each garden landscape element and the terrain data of the corresponding planning block;

[0078] Therefore, in the embodiment of the present invention, a DTLD (Digital twin models for landscape design) model of landscape area spatial attributes is established to realize real-time matching simulation analysis between landscape area GIS information and models of trees, flowers, grass, stones, lakes, monuments, buildings, etc. Figure 3 As shown:

[0079] Step (1): Analyze the spatial attributes of the planning area

[0080] As an optional embodiment, the topographic map of the planning area is obtained based on drone aerial survey provided in the embodiment of the present invention, and the spatial attributes of the underwater terrain and the aerial survey terrain are interpolated one by one for the grid points using the weighted average method, including:

[0081]

[0082] Among them, Z pis the coordinate of the interpolated elevation point; n is the number of interpolated elevation points; P i is the weight of the i-th data point; Z i is the elevation of the ith data point.

[0083] Specifically, the interpolated elevation of the grid point is calculated by calculating the weighted average of all points in a circle with the grid point as the center and R as the radius, and the weighted average calculation is performed using the square of the distance as the inverse of the weight. The formula is shown in formula (1).

[0084] The measurement result correction values ​​include:

[0085]

[0086] Among them, Δh is the correction value of the measurement result; h f It is a drone that surveys terrain; h d The underwater terrain detected by the depth sounder.

[0087] Specifically, based on the measurement results of the overlapping area and the measurement results of the UAV flight as the benchmark, the measurement correction values ​​of the interpolated depth sounder measurement results and the UAV flight survey results on the grid points are calculated as shown in formula (2). After calculating the measurement correction values, the depth sounder underwater topographic measurement results are corrected and then spliced ​​and fused with the UAV flight measurement results. A topographic map is generated according to the 1:500 scale mapping requirements to obtain the spatial attributes of the planning area.

[0088] Step (2): Establish the topological relationship between the node attributes, numerical attributes and planning areas of each landscape design element, namely the DTLD model, which is defined as:

[0089]

[0090] Step 4: Use interactive devices to achieve an immersive experience of the integrated simulation design solution, and evaluate whether the design solution meets the design requirements. If not, repeat steps 2 and 3, provide feedback to modify the BIM model, re-calculate the integrated simulation, and update. If it meets the requirements, the final design solution is obtained.

[0091] Furthermore, if Figure 4 As shown, in an embodiment of the present invention, after the DTLD model is established, a machine learning algorithm is used to train the model, including the following steps:

[0092] S304: setting the number of sampling with replacement is to initialize the new training set to an empty set, and perform the sampling operation number of times in a loop. Each time the sampling operation is performed, a pair of samples is randomly selected from the landscape BIM and the terrain data of the planning area, and a copy of the sample is copied and appended to the new training set;

[0093] S305: construct an algorithm parameter search set, each element in the set has a set of specific parameters that match the algorithm, randomly divide the samples in the training set into a training set with a ratio of 75% and a validation set with a ratio of 25%, and obtain the landscape BIM and planning area terrain topology function;

[0094] S306: defining a loss function machine evaluation function for matching landscape area terrain data with the BIM model, using the parameters to carry out training on the training set, and calculating the classification accuracy on the validation set;

[0095] S307: If the classification accuracy does not meet the requirements, return to step S304; if the classification accuracy meets the requirements, the training ends, and the parameters corresponding to the highest classification accuracy on the validation set are screened out; and the classification result value of the topological relationship between the landscape BIM and the planned area terrain is obtained.

[0096] The present invention uses multiple machine learning classification algorithms to be trained in parallel on the labeled landscape BIM and planning area terrain topological relationship classification data sets. When training each machine learning classification model, the original full data set is not directly used, but the original full data set is randomly sampled with replacement, and the specific machine learning classification model is trained again on the new data set obtained by random sampling. Due to the randomness of sampling with replacement, the training sets of different machine learning classification models are different, ensuring the independence of the classification results, thereby better matching the landscape BIM with the planning area terrain, obtaining a high-precision design plan, and laying a good algorithm foundation for the garden landscape digital twin design system of the present invention.

[0097] like Figure 5-Figure 7As shown, in one embodiment of the present invention, a garden scheme is designed by using a digital twin design system for garden landscape. First, drone aerial survey is used to obtain a planned area of ​​2,300 square meters. The entire area is in a rectangular layout, including land and water parts, of which the land area is 1,600 square meters and the water part is 7,000 square meters. The water part is the center, and the land part is designed into a classical garden landscape combining water systems and Chinese architecture. The important architectural style is pure Chinese classical garden style, which includes various traditional Chinese buildings, including main halls, corridors and hexagonal pavilions. The water system is combined with the courtyard to form an overall water system that is interspersed in the entire garden system. Broad-leaved forests, forest paths, lawns, flower beds, pavilions, streams, ponds and other landscapes are designed for the land blocks in the planned area. Node coding information of different landscape types is established, and the broad-leaved forests, forest paths, lawns are collected. The length, width, height, area and other numerical attributes of flower beds, pavilions, streams and ponds are used to establish BIM models of broad-leaved forests, forest trails, lawns, flower beds, pavilions, streams and ponds. Then, the BIM model is topologically associated with the geographic information of the planned area surveyed by the drone through the above formula (5) to achieve the integration of the BIM model of broad-leaved forests, forest trails, lawns, flower beds, pavilions, streams and ponds with GIS information. In the embodiment of the present invention, M different machine learning algorithms are used to train the landscape BIM and planned area topological relationship model in parallel, and it is required that none of the M machine learning algorithms are multi-layer perceptron neural network classification models; the M machine learning algorithms are respectively recorded as Algorithm1, Algorithm2, ..., Algorithm M ;

[0098] For a specific algorithm among the selected M machine learning algorithms j

[0099] Algorithm j ∈{Algorithm1,Algorithm2,…,Algorithm M}, the process of constructing a training set for the algorithm, selecting the optimal parameters, and training the model based on the optimal parameters is:

[0100] Step (3): Construct a training set; set the number of replacement sampling to L, and initialize the new training set to an empty set The sampling operation is performed L times in a loop. Each time the sampling operation is performed, a pair of samples is randomly selected from the landscape BIM and the terrain data of the planning area, and a copy of the sample is added to the new training set S. After the L sampling operations are completed, the sample capacity of the new training set S is L.

[0101] Step (4): Select the optimal parameters. j, j = 1, 2, ..., M, construct an algorithm parameter search set W for it, each element in W is a set of specific algorithm Algorithm j Algorithm is trained based on different parameters in W. j , the final classification accuracy is different, so we need to find a parameter from W that makes the classification accuracy the highest. The samples in W are randomly divided into a training set accounting for 75% and a validation set accounting for 25%. For each specific parameter in W, the model is trained on the training set using the parameter, and the classification accuracy is calculated on the validation set. The parameter with the highest classification accuracy on the validation set is selected and recorded as

[0102] Step (5): Train the model based on the optimal parameters. exist The landscape BIM and planning area topological relationship model is trained to obtain a landscape BIM and planning area terrain topological function The independent variable x represents a gray value vector of length 3×H×W, and the output value of the classification function is a classification result value representing the topological relationship between the landscape BIM and the terrain of the planned area. The BIM model is integrated with the actual physical model for simulation, and after the user proposes modification suggestions, the design scheme can be revised in real time, and the simulation calculation can be performed in real time to obtain the simulation effect, which greatly improves the degree of closeness between the design scheme and the actual construction effect, and solves a series of problems such as the difficulty of traditional two-dimensional or three-dimensional static landscape design to simulate the actual construction effect.

[0103] The method of the embodiment of the present invention is implemented by relying on electronic devices, so it is necessary to introduce the relevant electronic devices. For this purpose, an embodiment of the present invention provides an electronic device, which includes: at least one processor, a communication interface, at least one memory, and a communication bus, wherein at least one processor, a communication interface, and at least one memory communicate with each other through the communication bus. At least one processor can call the logic instructions in at least one memory to execute all or part of the steps of the method provided by the aforementioned various method embodiments.

[0104] In addition, the logic instructions in the at least one memory mentioned above can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each method embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0105] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A digital twin design method for garden landscape, characterized in that: The steps include: S100: Obtain the topographic data of the planning area through drone aerial survey and integrate it with GIS geographic information to obtain the physical model of the terrain data of the planning area; S200: Obtain the garden landscape design plan based on the terrain data of the planning area and establish the BIM model of each garden landscape element; S300: Establishing a DTLD model of landscape area spatial attributes to achieve fusion simulation of the BIM model of each garden landscape element and the terrain data of the corresponding planning block. The step S300 includes: S301: Analyze the spatial attributes of the planning area, including: Among them, Z p is the coordinate of the interpolated elevation point; n is the number of interpolated elevation points; P i is the weight of the i-th data point; Z i is the elevation of the ith data point; S302: Correct the measurement results: Among them, Δh is the correction value of the measurement result; h f It is a drone that surveys terrain; h d underwater topography detected by the depth sounder; S303: Establish the topological relationship between the node attributes, numerical attributes and planning areas of each landscape design element, namely the DTLD model, which is defined as: S304: setting the number of sampling with replacement is to initialize the new training set to an empty set, and perform the sampling operation number of times in a loop. Each time the sampling operation is performed, a pair of samples is randomly selected from the landscape BIM and the terrain data of the planning area, and a copy of the sample is copied and appended to the new training set; S305: construct an algorithm parameter search set, each element in the set has a set of specific parameters that match the algorithm, randomly divide the samples in the training set into a training set with a ratio of 75% and a validation set with a ratio of 25%, and obtain the landscape BIM and planning area terrain topology function; S306: defining a loss function machine evaluation function for matching landscape area terrain data with the BIM model, using the parameters to carry out training on the training set, and calculating the classification accuracy on the validation set; S307: If the classification accuracy does not meet the requirements, return to step S304; if the classification accuracy meets the requirements, the training ends, and the parameters corresponding to the highest classification accuracy on the validation set are selected; the classification result value of the topological relationship between the landscape BIM and the planned area terrain is obtained; S400: An immersive experience of the integrated simulation design scheme is achieved through interactive devices to evaluate whether the design scheme meets the design requirements. If not, steps two and three are repeated, feedback is provided to modify the BIM model and the integrated simulation calculation is re-performed and updated. If the requirements are met, the final design scheme is obtained.

2. A digital twin design method for garden landscape according to claim 1, characterized in that: Step S100 includes: S101: Survey the planned area, determine the time for drone aerial survey based on the survey results, plan the drone route, collect aerial photography data, test and process the collected photography data, and obtain the aerial survey terrain of the planned area; S102: Determine the depth sounder detection time according to the survey results, collect the mountain range, terrain distribution, water range, water depth value and coordinates of the planned area to be measured, and obtain the plane coordinates and elevation of the planned area to be measured according to the coordinates; S103: The mountain range, terrain distribution, water range, water depth and coordinates are integrated to obtain the overall terrain distribution of the planning area, which is then integrated with the GIS geographic information to obtain the final mapping results of the planning area.

3. A digital twin design method for garden landscape according to claim 2, characterized in that: Step S101 includes testing the flight quality and image quality after the flight is completed, making up for unqualified routes and leaky areas, using aerial triangulation and regional network adjustment to generate a three-dimensional real-scene model, and using the measurement results selected on the three-dimensional real-scene model as a benchmark to correct the drone aerial survey results and obtain the aerial survey terrain of the planning area.

4. A digital twin design method for landscape architecture according to any one of claims 1 to 3, characterized in that: Step S200 includes: S201: Establish node attributes of each landscape design element for: Where i represents the type of landscape design element, j represents the number of each landscape design element, G represents the attribute of the jth element in the i-th landscape design element, and [ID01]-[IDij] represents the coding of the landscape design element; S202: Constructing numerical attributes of each landscape design element for: Among them, V represents the geometric attribute information of the jth element in the i-type landscape design element.

5. A digital twin design system for garden landscape, using the method according to any one of claims 1 to 4, characterized in that: It includes three layers of architecture: data security layer, modeling computing layer, and immersive experience layer. The data security layer includes a GIS geographic information data acquisition module, a landscape design data module, a high-data data transmission module, and a data management module. The GIS geographic information data acquisition module is used to realize aerial survey of the landscape planning area by means of drones and other means to obtain aerial survey terrain and geomorphic data. The landscape design data module is used to establish BIM models of various landscape types. The high-data data transmission module is used to realize real-time transmission and interaction of terrain data in the planning area and BIM model data of each landscape. The data management module stores, reads, modifies, and updates the PB-level big data generated by the digital twin design system. The modeling and calculation layer includes a landscape area geography, landform environment physical module, a landscape object BIM module and a matching algorithm module. The landscape area geography, landform environment physical module is used to fuse the corresponding GIS information of the planning area, match and fuse the aerial survey terrain and landform data with the GIS position coordinates, and obtain the physical model of the actual terrain data. The landscape object BIM module is used to establish the BIM model of broad-leaved forests, forest trails, lawns, flower beds, pavilions, streams, and ponds. The matching algorithm module is used to fuse and simulate the BIM model with the actual physical model, and after the user proposes modification suggestions, the scheme design can be revised in real time, and the simulation calculation can be carried out in real time; The immersive experience layer includes a three-dimensional virtual mapping module, a user interaction module and a virtual digital cluster module. The three-dimensional virtual mapping module is used to perform three-dimensional mapping between the physical model and the virtual model simulation calculation results to obtain a design scheme rendering that is close to the real environment. The user interaction module is used to interact with the design scheme through a virtual reality wearable device, modify the scheme, calculate in real time, update, and display. The virtual digital cluster module realizes multi-user collaborative interaction, design scheme review, modification, simulation calculation and update.

6. An electronic device, characterized in that: include: At least one processor, at least one memory and a communication interface; wherein, The processor, memory and communication interface communicate with each other; The memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to execute the method according to any one of claims 1 to 4.

7. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium stores computer instructions, which cause the computer to execute the method of any one of claims 1 to 4.

Images