Interactive garden design method and system based on landscape garden simulation

Abstract

The invention discloses an interactive garden design method and system based on landscape garden simulation, particularly relates to the technical field of garden landscape design and computer graphic simulation, and relates to a plurality of modules of environmental data acquisition, three-dimensional modeling, plant growth simulation, water simulation, interactive design, microclimate prediction and the like. According to the method, terrain, climate and soil data are collected through a laser radar and multiple sensors, and a standardized environmental map is constructed; generating a plant topological structure based on a context sensitive type L system, and simulating a growth process by adopting a Logistic function in combination with an environmental factor; the method comprises the following steps: realizing dynamic simulation of a water body by utilizing an SPH particle method and a Navier-Stokes equation, and establishing a plant-water body feedback model; and real-time operation and design optimization are supported through a three-dimensional interaction interface. The system has the advantages of being high in ecological adaptability, high in visualization degree, perfect in feedback mechanism and the like, the intelligence and scientificity of garden design are remarkably improved, and the system is suitable for various scenes such as urban green lands and ecological restoration.

Description

Technical Field

[0001] The present invention relates to the technical field of garden landscape design and computer graphic simulation, and in particular to an interactive garden design method and system based on landscape garden simulation. Background Art

[0002] Current landscape design typically relies on CAD mapping, static 3D modeling, or simplified GIS systems. These systems lack the ability to model real-time environmental responses, dynamic plant growth, water changes, and user behavior feedback. This leads to the following main issues:

[0003] Low data collection granularity: Traditional methods rely heavily on satellite images and manual mapping, making it difficult to obtain dynamic information such as soil and microclimate; Static plant simulation: The plant models in mainstream garden design tools are static and cannot simulate dynamic growth based on the real environment; Lack of physical interaction simulation: The impact of natural factors such as water, wind, and light on landscape elements is not truly reflected; Delayed design feedback: The design effect lacks intuitive feedback, and it is impossible to form an iterative optimization mechanism centered on user behavior.

[0004] Therefore, there is an urgent need for a garden design system driven by multi-source perception and supported by a simulation engine, which integrates physics, ecology and human-computer interaction logic to improve the authenticity, scientificity and interactivity of digital garden landscape design. Summary of the Invention

[0005] The main purpose of the present invention is to provide an interactive garden design method and system based on landscape garden simulation, which forms a complete intelligent design closed loop through multi-source sensor data acquisition, three-dimensional modeling, plant dynamic growth simulation, water body physical response modeling, real-time interactive feedback and user behavior learning.

[0006] To achieve the above object, the technical solution adopted by the present invention is:

[0007] An interactive garden design method based on landscape garden simulation, the specific steps are as follows:

[0008] Step 1: Environmental data collection and preprocessing: Collect the topography, climate, and soil data of the garden site. Topography data is collected using lidar and RGB-D cameras, climate data is collected using a micro-weather station, and soil data is collected using conductivity, humidity, and pH sensors. The collected data is normalized and spatially registered to generate a structured environmental data map.

[0009] Step 2: 3D terrain modeling. Perform outlier removal, point cloud registration, and spatial interpolation on the point cloud data obtained in S1 to construct a 3D terrain mesh model with topological structure, texture coordinates, and geographic attributes, which serves as the basis for subsequent modeling.

[0010] Step 3: Plant modeling and growth simulation: Based on the three-dimensional terrain model, the L-System topology of the plant is constructed, and the plant growth model is calculated in combination with environmental factors. The logistic function is used to simulate the change of plant height, and the environmental adaptation function is calculated based on factors such as temperature, moisture, light and pH to adjust the plant growth behavior;

[0011] Step 4: Water body simulation and plant feedback modeling. A garden water body simulation model is constructed based on the terrain model and plant model. The Navier-Stokes equation is used to simulate the flow characteristics of the water body. The Smoothed Particle Hydrodynamics method is combined to perform particle simulation of the dynamic behavior of the water body. A feedback model of the water body on plant water absorption and morphological changes is established.

[0012] Step 5: Interactive design and feedback control. This provides a 3D interactive interface that allows users to adjust plant species, terrain height, irrigation paths, and layout density through mouse, touch, or voice input. The system provides real-time feedback on model changes based on these operations and optimizes rendering and computational efficiency through a level of detail management mechanism.

[0013] Step 6: Natural response simulation: sunshine, wind field, and microclimate simulation are performed for the design scenario. Sunlight is simulated using the sun position algorithm, wind field simulation uses the k-ε turbulence model, and microclimate prediction is based on graph neural network to model local temperature, humidity, and light changes. Plant distribution and layout structure are adjusted according to the simulation results.

[0014] Step 7: Design evaluation and result output: Conduct a comprehensive evaluation of the ecology, landscape visibility, and visitor behavior paths of the garden design, export the evaluation report in PDF format, geographic information data in GeoJSON format, and 3D model in GLTF format, and integrate and synchronize with external BIM or GIS platforms through RESTful API.

[0015] Preferably, the point cloud data is anomaly eliminated by adopting a spatial anomaly recognition method based on a random sampling consistency algorithm, the point cloud registration adopts an iterative closest point algorithm to unify local and global coordinates, and the interpolation method adopts Kriging spatial interpolation to generate a continuous terrain surface.

[0016] Preferably, the L-System topology of the plant is a context-sensitive L-system, which branches and evolves according to the set growth rules and time steps, and each branch unit establishes a mapping relationship with the corresponding light intensity and nutrient concentration to control the direction of morphological expansion.

[0017] Preferably, the Logistic function is used to simulate the change of height H(t) during plant growth, and the specific form is:

[0018]

[0019] Among them, H(t) is the plant height at time t, H max is the maximum height, k is the growth rate coefficient, and t0 is the growth inflection point time.

[0020] Preferably, the water body simulation uses the Smoothed Particle Hydrodynamics method to calculate the momentum and pressure between liquid particles, and uses the Navier-Stokes equations to constrain the velocity field and pressure field to ensure flow continuity and incompressibility.

[0021] Preferably, the feedback mechanism between the water body and the plant is based on the plant root water absorption model and the water flow field contact probability model, and by updating the plant moisture content parameter θ(t), its leaf area index and transpiration coefficient are affected in real time.

[0022] Preferably, an independent operation mode switching mechanism is set in the three-dimensional interactive interface, which supports four types of operations: terrain editing, plant configuration, path optimization and water body layout, and the interface integrates a heat map overlay module to display the distribution results of light, humidity and crowd density.

[0023] Preferably, the GeoJSON file generated in step 7 contains the spatial coordinates, species identification, growth status value and microclimate type of each plant unit, and the GLTF file contains complete texture, normal vector, LOD index and physical property label for three-dimensional visual engine rendering.

[0024] The present invention also discloses an interactive garden design system based on landscape garden simulation, comprising:

[0025] The environmental perception module is used to collect the three-dimensional terrain, climate data, and soil environmental data of the garden area. The terrain data is obtained by lidar and RGB-D camera, the climate data is obtained by meteorological sensor nodes, and the soil data is obtained by pH, conductivity, and moisture content sensors. The raw data is then structured and converted.

[0026] The terrain modeling module is used to perform anomaly removal, registration, and spatial interpolation operations on the point cloud data provided by the environment perception module, and output a terrain mesh with spatial topology and texture properties;

[0027] Plant simulation module, including the L-System growth rule editor and the Logistic growth calculation engine, is used to construct and evolve the three-dimensional structure of plants and dynamically adjust growth strategies based on environmental factors;

[0028] The water simulation module builds a water flow simulation framework based on the SPH particle system and embeds the Navier-Stokes physics solver to achieve continuous water behavior modeling;

[0029] Feedback modeling module, used to capture the effect of water on plant water absorption and morphological changes, and pass feedback parameters to the plant simulation module;

[0030] An interactive visualization module, including a 3D modeling interface, a heat map analysis window, and a parameter control panel, allows users to edit terrain, plants, and water configurations in real time, and provides operational feedback through a GPU rendering engine.

[0031] The natural environment response module is used to perform microclimate calculation tasks based on sunshine prediction models, wind field turbulence models, and graph neural networks, and output environmental data for adjusting landscape layout;

[0032] The evaluation and integration module is used to output ecological performance evaluation reports, tourist path visualization data, and data interface communication with the BIM / GIS platform.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] 1. This invention constructs a plant topology based on a context-sensitive L-system and dynamically simulates the plant growth process by combining environmental response factors (such as light, moisture, temperature, and soil properties). This achieves a close coupling of plant structure and natural conditions, making the garden design not only ecologically adaptable but also structurally controllable and diverse in expression, significantly improving the scientific nature and sustainability of the design.

[0035] 2. This invention integrates functional modules such as 3D terrain modeling, water physics simulation, interactive visualization design, and microclimate feedback prediction to form a complete closed-loop garden digital twin system. This system supports real-time human-computer interaction and design optimization, outputs standardized data formats, and integrates interfaces across multiple platforms, improving the efficiency, precision, and intelligence of garden design. It is suitable for a variety of application scenarios, including urban green spaces, ecological restoration, and landscape education. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a flow chart of the working steps of the present invention;

[0037] Figure 2 It is a flowchart of the workflow of each module of the present invention. DETAILED DESCRIPTION

[0038] In order to make the technical means, creative features, objectives and effects achieved by the present invention easier to understand, the present invention is further described below in conjunction with specific implementation methods.

[0039] This invention discloses an interactive garden design method based on landscape garden simulation. This method aims to reconstruct the logical chain of digital garden design through precise perception, intelligent simulation, real-time feedback, and evaluation and prediction. The method steps are described below, with close technical and logical connections and data flow between each step.

[0040] Specifically, the present invention is implemented according to the following steps:

[0041] Step 1: Multi-source perception and collection of site environmental data: This step, as the starting point of the method, mainly completes the comprehensive perception and data initialization of the multi-dimensional environmental information of the design area.

[0042] The first step is to collect terrain and visual data: deploying LiDAR and RGB-D camera arrays to achieve 3D point cloud modeling of the park and high-precision image texture collection. The point cloud resolution reaches 1cm, ensuring the accuracy of subsequent terrain modeling.

[0043] Then, climate and soil data are acquired: a micro-meteorological station and a soil multi-parameter sensor are integrated to obtain real-time temperature and humidity (T, H), wind speed (V), light intensity (I), pH value, moisture content (W), etc.

[0044] Based on the data obtained above, the raw data is preprocessed: all sensor data is uploaded to the CUDA acceleration module for cleaning, format standardization, and preliminary visualization integration to form a unified "environmental data map." The output of this step directly serves as the input for terrain modeling and plant growth simulation (steps 2 and 3).

[0045] This step is the foundation of the entire garden design process and is responsible for collecting and processing multidimensional environmental data of the design area. Comprehensive terrain, climate, and soil information is acquired through various devices such as lidar, RGB-D cameras, and meteorological sensors. This data will provide basic input for subsequent terrain modeling (step two), plant growth simulation (step three), and water behavior simulation (step four). After data collection is completed, this raw information is converted into a standardized environmental data map through preprocessing, which serves as the initial condition for the design simulation system.

[0046] Step 2: 3D Terrain Modeling and Structural Reconstruction: After completing the original perception, this step focuses on building a terrain model to provide a spatial carrier for subsequent designs such as plant distribution and water direction. The details are as follows:

[0047] Anomaly removal and registration: Use RANSAC to remove noise points, and then use the ICP algorithm to accurately register multi-view point clouds to ensure model consistency;

[0048] Interpolation modeling: Kriging interpolation method is used to reconstruct and fill the low-density area point cloud to form a high-density continuous terrain;

[0049] Mesh generation: Construct a 3D Geo-Terrain mesh containing topology, texture coordinates, and GIS attribute fields;

[0050] Model visualization initialization: The rendering engine loads the model and connects to the LOD mechanism, serving as a three-dimensional container for subsequent simulation, interaction, and growth algorithm execution.

[0051] This step forms data continuity with step one, and the output terrain grid will serve as the spatial bearing basis for plant simulation in the next step. After completing the environmental data collection, step two processes the point cloud data to construct a high-precision three-dimensional terrain model. This step generates a continuous and detailed terrain grid model by eliminating abnormal data, aligning point clouds, and performing interpolation reconstruction. The model contains the spatial topology, texture coordinates, and attribute information of the terrain, and serves as the spatial basis for subsequent plant growth (step three) and water body simulation (step four). The accuracy of terrain modeling directly affects the feasibility and authenticity of subsequent designs, ensuring the feasibility of garden design plans.

[0052] Step 3: Plant growth behavior simulation and environmental response calculation: This step relies on the aforementioned environmental perception and three-dimensional space to build a plant simulation system that can respond to dynamic microclimate changes.

[0053] System structure generation: Constructing plant topology based on context-sensitive L-systems.

[0054] L-System, short for Lindenmayer System, is a formal language (Formal Grammar) for simulating plant growth. It was proposed by botanist Aristid Lindenmayer and is particularly suitable for describing the repetitive, recursive, and fractal features in plant structures. Using a context-sensitive L-system to model plant structures: when generating three-dimensional structures such as branches and leaves, the positional relationship of the current structural unit in the overall plant topology is taken into account. For example, new branches or leaves will only be generated when the light, water, or spatial structure around a branch meet specific conditions.

[0055] "Topology" refers to the way the various parts of a plant's structure are connected, such as the structural hierarchy of "trunk → first-order branches → second-order branches → leaves." This modeling approach is closer to the growth logic of real plants because it can simulate complex ecological feedback loops, such as "shading by upper branches and leaves affects growth in the lower layers."

[0056] Imagine you're modeling a crape myrtle tree: if a branch has a trunk on the left and a sunny area on the right, it will sprout a flower; otherwise, the branch will remain unchanged or wither. This type of growth logic based on the environment and structure requires a context-sensitive L-system to accurately express.

[0057] Growth process modeling: Logistic equation is used to simulate plant growth: the specific form is:

[0058]

[0059] Among them, H(t) is the plant height at time t, H max is the maximum height, k is the growth rate coefficient, and t0 is the growth inflection point time.

[0060] Environmental adaptation function embedding: f(E) = α·T(t)·S(W)·L(I)·F(Ph); where T(t) is the temperature function, S(W) is the moisture function, L(I) is the light function, F(Ph) is the pH adaptability, and α is the regulation coefficient.

[0061] Feedback mechanism: The system dynamically determines the number of branches, leaf density, and growth direction based on f(E), realizing personalized growth strategies for plants in different plots.

[0062] This step directly acts on the spatial model output from step 2 to realize the "grounding" simulation of digital plants. Relying on the three-dimensional terrain model and real-time climate data generated in the previous stage, the dynamic growth simulation of plants is carried out. The topological structure of the plant is constructed through the L-system, and the logistic growth equation is used to simulate the growth process of the plant. In addition, the environmental response function is introduced to adjust the growth strategy of the plant in real time by combining factors such as temperature, moisture, light and soil conditions. The core of this step is to ensure that the plants can grow reasonably in the designed environment and respond to environmental changes in real time. The accuracy of the simulation provides the basis for subsequent design optimization, interactive modification and ecological evaluation.

[0063] Step 4: Water behavior simulation and plant interaction feedback modeling: Considering the importance of the waterscape system in the garden, this step adds hydrodynamic behavior based on the aforementioned terrain and simulates its interaction with the plant system.

[0064] Hydrodynamic modeling: Based on the Navier-Stokes equations, it describes the behavior of water flow:

[0065]

[0066] in, is the water velocity, p is the water pressure, ρ is the water density, ν is the dynamic viscosity, is the external force term for plants, wind, etc. This equation expresses the conservation of mass of the fluid, which means that the density of the fluid and the divergence of the velocity field must meet certain conditions.

[0067] SPH simulation particle system: divides water flow into particle units to simulate complex behaviors such as waves, disturbances, and water surface ripples;

[0068] Water-plant interaction mechanism: When the plant structure is located in the water flow area, the system automatically calculates the resistance feedback and adjusts the plant's water intake efficiency and morphological swing.

[0069] Optimize water visual performance: Combine FFT+PerlinNoise to dynamically generate disturbance textures, and use MarchingCubes to generate flowing water surface meshes to improve realism.

[0070] This step connects step three (plants) with step two (terrain), forming a unified simulation closed loop for the three. The physical behavior of water flow is simulated by the Navier-Stokes equations, and combined with the Smoothed Particle Hydrodynamics (SPH) particle simulation technology, the flow, fluctuation and disturbance of water bodies are reproduced in detail. This step also introduces an interactive feedback mechanism between water bodies and plants to simulate the impact of water flow on plant morphology, growth and water absorption. Through these dynamic simulations, designers can fully understand the impact of water bodies on garden landscapes and ensure the synergy between water bodies and plant systems.

[0071] Step 5: User interaction design and model feedback control: To achieve efficient human-computer collaborative design, the system introduces a visual interaction module and behavior-driven mechanism.

[0072] UI interaction framework: The system provides a multi-mode interface, including parameter panel, 3D visual window, and heat map overlay area;

[0073] Parameter response control: Users can make real-time adjustments to plant species, planting density, irrigation paths, and terrain morphology, and the system automatically triggers model updates;

[0074] LOD scheduling mechanism: Through octree indexing combined with screen space error control, resources are loaded on demand to ensure smooth interaction.

[0075] Behavior Recording Module: The system tracks user click trajectories, lens preferences, and modification frequency for subsequent recommendations and adaptive optimization of simulation strategies.

[0076] This step performs controllable operations on the first three steps (modeling, plants, and water bodies) to establish a human-machine-model linkage mechanism. It allows users to edit and adjust the model in real time during the design process. Through a variety of interactive input methods (such as mouse, touch, voice, etc.), users can flexibly adjust plant species, terrain, irrigation, etc. The system will update the model in real time based on the user's operation and provide feedback on the design effect. At the same time, the system has a built-in LOD (level of detail) scheduling mechanism to optimize rendering efficiency. The purpose of this step is to enhance the interactivity and convenience of the design process, ensure that designers can view and adjust the design effects in real time, and provide efficient decision support.

[0077] Step 6: Dynamic response of natural environment and microclimate simulation: In this stage, real natural elements are integrated into the simulation system to provide climate adaptability testing for the design scheme.

[0078] Sunlight prediction system: Calculates the sun's altitude angle and projection trajectory based on the SPA algorithm for shadow simulation and plant illumination evaluation;

[0079] Wind field disturbance analysis: The k-ε turbulence equation is used to construct a three-dimensional wind velocity field, which is fed back into the plant structure model to simulate the swaying of branches and leaves and changes in wind resistance.

[0080] Microclimate trend prediction: Use graph neural networks (GNNs) to make short-term predictions of local temperature, humidity, and light intensity, and feed the results back to the plant model for dynamic adjustment.

[0081] The core of step six is ​​to verify the design's adaptability in the real world through high-precision natural environment simulation. This step includes dynamic predictions of sunlight, wind patterns, and microclimate, enabling assessment of plant growth and water body trends under different time periods and climate conditions. By predicting environmental factors using graph neural networks (GNNs) and turbulence models (such as the k-ε model), the system provides real-time feedback and design adjustments, ensuring that the landscape maintains its optimal ecological and user-friendly environment under various natural conditions.

[0082] Step 7: Design evaluation, export and platform integration: After the design is completed, the system can generate a professional evaluation report and export documents in multiple formats, supporting solution archiving, reporting and platform integration.

[0083] Ecological and visual indicator analysis: Shannon-Wiener diversity index; visibility heat map, visual transparency analysis;

[0084] Tour simulation and traffic analysis: A* path algorithm simulates tourist flow; congestion factor modeling evaluates attraction density;

[0085] Export support: PDF format assessment documents; GeoJSON geographic data interface; GLTF model file docking BIM / GIS platform;

[0086] Platform linkage: Provides RESTful API; supports cloud collaboration, authorization authentication and encrypted transmission (TLS1.3+OAuth2.0).

[0087] Finally, step seven evaluates and outputs the entire design plan, generating a detailed report and export files. Through analysis across multiple dimensions, including ecological assessment, visitor experience simulation, and visual rendering, the system comprehensively assesses the feasibility and optimization potential of the design plan. Furthermore, it supports export in formats such as GeoJSON and GLTF, and integrates with external platforms via a RESTful API. This step ensures that the design plan can not only be stored and transmitted, but also be updated and optimized later, forming a complete design closed loop.

[0088] Based on the above steps, the present invention further discloses an interactive garden design system based on landscape garden simulation, which is used to execute the above method steps, specifically comprising:

[0089] The environmental perception module is used to collect the three-dimensional terrain, climate data, and soil environmental data of the garden area. The terrain data is obtained by lidar and RGB-D camera, the climate data is obtained by meteorological sensor nodes, and the soil data is obtained by pH, conductivity, and moisture content sensors. The raw data is then structured and converted.

[0090] The terrain modeling module is used to perform anomaly removal, registration, and spatial interpolation operations on the point cloud data provided by the environment perception module, and output a terrain mesh with spatial topology and texture properties;

[0091] Plant simulation module, including the L-System growth rule editor and the Logistic growth calculation engine, is used to construct and evolve the three-dimensional structure of plants and dynamically adjust growth strategies based on environmental factors;

[0092] The water simulation module builds a water flow simulation framework based on the SPH particle system and embeds the Navier-Stokes physics solver to achieve continuous water behavior modeling;

[0093] Feedback modeling module, used to capture the effect of water on plant water absorption and morphological changes, and pass feedback parameters to the plant simulation module;

[0094] An interactive visualization module, including a 3D modeling interface, a heat map analysis window, and a parameter control panel, allows users to edit terrain, plants, and water configurations in real time, and provides operational feedback through a GPU rendering engine.

[0095] The natural environment response module is used to perform microclimate calculation tasks based on sunshine prediction models, wind field turbulence models, and graph neural networks, and output environmental data for adjusting landscape layout;

[0096] The evaluation and integration module is used to output ecological performance evaluation reports, tourist path visualization data, and data interface communication with the BIM / GIS platform.

[0097] A specific implementation scenario is set below to further disclose the present invention:

[0098] Taking the "Public Green Space Landscape Renewal Project of a Southern Waterfront City" as the target area, combined with the garden design system provided by the present invention, the whole process design from environmental data collection, 3D modeling, growth simulation, water body simulation to environmental response assessment is completed, realizing visual, interactive, and multi-factor driven intelligent garden planning.

[0099] The specific process is as follows:

[0100] 1. Project Background and Initial Parameters

[0101] Project Name: Digital Landscape Renovation of Dongxi Waterfront Park;

[0102] Site area: approximately 35,000 square meters;

[0103] Geographical coordinates: 26.08°N, 119.31°E;

[0104] Climate characteristics: Subtropical humid climate, with an average annual temperature of about 19.5°C and an annual precipitation of about 1,350 mm;

[0105] Target plant configuration: including Cinnamomum camphora (evergreen tree), Crape myrtle (deciduous small tree), and Ophiopogon japonicus (ground cover plant).

[0106] 2. Environmental Data Collection and Preprocessing

[0107] The site was scanned using a vehicle-mounted lidar device (Velodyne VLP-16) to obtain approximately 450,000 pieces of 3D point cloud data. An Intel RealSense D455 RGB-D camera was also used to capture color images and depth maps, with a resolution of 1280×720 for key areas.

[0108] A weather microstation was installed to collect climate factors such as temperature (24.3°C), humidity (78%), wind speed (3.4 m / s), and sunshine (1800 lux). Soil sensors collected surface moisture content (35%), pH (6.1), and EC (0.9 dS / m).

[0109] Preprocessing includes spatial registration, noise filtering (RANSAC method is used to eliminate isolated point groups), and unified conversion to the EPSG:4490 projection coordinate system to form a standardized environmental data map.

[0110] 3.3D Terrain Modeling

[0111] The Kriging interpolation method is used to process the void areas of the point cloud. After removing abnormal height mutation points, a 1m precision triangulated network (TIN) terrain model is generated, which contains a total of 17234 grid cells. Each cell stores the terrain elevation, slope and texture map index.

[0112] 4. Plant Modeling and Growth Simulation

[0113] Construct context-sensitive L-System rules for camphor.

[0114] Combined with the field temperature (T = 24.3 ° C), light (I = 1800 lux), moisture (W = 35%), and pH = 6.1, the following adaptation function is used:

[0115] f(E)=0.95·T(t)·S(W)·L(I)·F(pH)

[0116] in: F(pH)=1-|6.1-6.5| / 2=0.8.

[0117] Calculation shows: f(E)≈0.95×0.81×0.7×0.6×0.8≈0.26.

[0118] Growth function: It indicates that the height of the camphor tree approaches 12 meters as the simulated number of days t increases.

[0119] 5. Water Simulation and Plant Feedback

[0120] An SPH particle water model was constructed based on the artificial stream on the east side of the site. A total of approximately 13,500 particles were generated, and the Navier-Stokes equation was used to control the water flow rate within the range of 0.8 m / s.

[0121] The root zone radius of each plant was set to 0.3 m. The particle contact ratio was calculated and fed back into the plant model to update the plant water absorption rate θ(t). For example, the leaf area index of camphor trees increased by 8.7% under high humidity conditions for 72 consecutive hours.

[0122] 6. 3D Interaction Design and Optimization

[0123] Designers use the touchscreen interface to select "Plant Configuration" mode and arrange plants by dragging and dropping. A heat map on the interface displays the light distribution in real time. LOD levels can be set from 0 to 4, with lower-end devices automatically downgrading to LOD 2 or below to ensure smooth operation.

[0124] When the user adjusts the irrigation path, the system recalculates the plant water coverage in real time and automatically recommends an optimized irrigation plan.

[0125] 7. Natural Response Simulation and Adaptation

[0126] Sunshine simulation: Using the SPA model, the sunshine duration in the camphor tree canopy area during the summer solstice in June was calculated to be 8.2 hours;

[0127] Wind field prediction: Based on the k-ε turbulence model, the flow field distribution of southeast winds was simulated, and it was proposed to place tall plants at the upwind side to alleviate the wind pressure in the square;

[0128] Microclimate prediction: The graph neural network model predicts that the average daily temperature in the central part of the simulation area will increase by 1.8°C in the next 48 hours. It is recommended to appropriately increase the density of ground cover plants to reduce surface thermal radiation.

[0129] 8. Result Output and Evaluation (Corresponding to S7)

[0130] The system automatically generates the following content:

[0131] Ecological assessment report (PDF), including carbon sink estimates (4.3 tons / year) and water cycle efficiency (68%);

[0132] GeoJSON file, marking each plant unit (including UUID, coordinates, species, current growth height, transpiration coefficient);

[0133] GLTF 3D model file, used for Unity rendering;

[0134] The RESTful API interface successfully connects the BIM collaboration system Revit and the urban information platform CityGML.

[0135] 9. System Deployment Environment Description

[0136] Deployed on an NVIDIA RTX A6000 graphics workstation, equipped with CUDA 12.1, it supports 32-thread parallel water simulation tasks; remotely supports two urban designers and one botanical consultant to collaborate on modifying the design plan through an HTTPS encrypted channel; the version control system records all interactions and rollback operations, with a total of 12 versions.

[0137] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. An interactive garden design method based on landscape garden simulation, characterized in that: The specific steps are as follows: Step 1: Environmental data collection and preprocessing: Collect the topography, climate, and soil data of the garden site. Topography data is collected using lidar and RGB-D cameras, climate data is collected using a micro-weather station, and soil data is collected using conductivity, humidity, and pH sensors. The collected data is normalized and spatially registered to generate a structured environmental data map. Step 2: 3D terrain modeling. Perform outlier removal, point cloud registration, and spatial interpolation on the point cloud data obtained in S1 to construct a 3D terrain mesh model with topological structure, texture coordinates, and geographic attributes, which serves as the basis for subsequent modeling. Step 3: Plant modeling and growth simulation: Based on the three-dimensional terrain model, the L-System topology of the plant is constructed, and the plant growth model is calculated in combination with environmental factors. The logistic function is used to simulate the change of plant height, and the environmental adaptation function is calculated based on factors such as temperature, moisture, light and pH to adjust the plant growth behavior; Step 4: Water body simulation and plant feedback modeling. A garden water body simulation model is constructed based on the terrain model and plant model. The Navier-Stokes equation is used to simulate the flow characteristics of the water body. The Smoothed Particle Hydrodynamics method is combined to perform particle simulation of the dynamic behavior of the water body. A feedback model of the water body on plant water absorption and morphological changes is established. Step 5: Interactive design and feedback control. This provides a 3D visual interactive interface that allows users to adjust plant species, terrain height, irrigation paths, and layout density through mouse, touch, or voice input. The system provides real-time feedback on model changes based on these operations and optimizes rendering and computational efficiency through a level of detail management mechanism. Step 6: Natural response simulation: sunshine, wind field, and microclimate simulation are performed for the design scenario. Sunlight is simulated using the sun position algorithm, wind field simulation uses the k-ε turbulence model, and microclimate prediction is based on graph neural network to model local temperature, humidity, and light changes. Plant distribution and layout structure are adjusted according to the simulation results. Step 7: Design evaluation and result output: Conduct a comprehensive evaluation of the ecology, landscape visibility, and visitor behavior paths of the garden design, export the evaluation report in PDF format, geographic information data in GeoJSON format, and 3D model in GLTF format, and integrate and synchronize with external BIM or GIS platforms through RESTful API.

2. The interactive garden design method based on landscape garden simulation according to claim 1, characterized in that: The point cloud data is anomaly eliminated by using a spatial anomaly recognition method based on a random sampling consistency algorithm, the point cloud registration uses an iterative closest point algorithm to unify local and global coordinates, and the interpolation method uses Kriging spatial interpolation to generate a continuous terrain surface.

3. The interactive garden design method based on landscape garden simulation according to claim 1, characterized in that: The L-System topology of the plant is a context-sensitive L-system, which branches and evolves according to the set growth rules and time steps. Each branch unit establishes a mapping relationship with the corresponding light intensity and nutrient concentration to control the direction of morphological expansion.

4. The interactive garden design method based on landscape garden simulation according to claim 1, characterized in that: The Logistic function is used to simulate the change of height H(t) during plant growth, and its specific form is: Among them, H(t) is the plant height at time t, H max is the maximum height, k is the growth rate coefficient, and t0 is the growth inflection point time.

5. The interactive garden design method based on landscape garden simulation according to claim 1, characterized in that: The water simulation uses the Smoothed Particle Hydrodynamics method to calculate the momentum and pressure between liquid particles, and uses the Navier-Stokes equations to constrain the velocity field and pressure field to ensure flow continuity and incompressibility.

6. The interactive garden design method based on landscape garden simulation according to claim 1, characterized in that: In step 4, the feedback mechanism between the water body and the plant is based on the plant root water absorption model and the water flow field contact probability modeling, and by updating the plant moisture content parameter θ(t), its leaf area index and transpiration coefficient are affected in real time.

7. The interactive garden design method based on landscape garden simulation according to claim 1, characterized in that: The three-dimensional interactive interface is provided with an independent operation mode switching mechanism, which supports four types of operations: terrain editing, plant configuration, path optimization, and water body layout. The interface also integrates a heat map overlay module to display the distribution results of light, humidity, and crowd density.

8. The interactive garden design method based on landscape garden simulation according to claim 1, characterized in that: The GeoJSON file generated in step 7 contains the spatial coordinates, species identification, growth status value and microclimate type of each plant unit, and the GLTF file contains complete texture, normal vector, LOD index and physical property label for 3D visual engine rendering.

9. An interactive garden design system based on landscape garden simulation, used to execute the method according to any one of claims 1 to 8, characterized in that: include: The environmental perception module is used to collect the three-dimensional terrain, climate data, and soil environmental data of the garden area. The terrain data is obtained by lidar and RGB-D camera, the climate data is obtained by meteorological sensor nodes, and the soil data is obtained by pH, conductivity, and moisture content sensors. The raw data is then structured and converted. The terrain modeling module is used to perform anomaly removal, registration, and spatial interpolation operations on the point cloud data provided by the environment perception module, and output a terrain mesh with spatial topology and texture properties; Plant simulation module, including the L-System growth rule editor and the Logistic growth calculation engine, is used to construct and evolve the three-dimensional structure of plants and dynamically adjust growth strategies based on environmental factors; The water simulation module builds a water flow simulation framework based on the SPH particle system and embeds the Navier-Stokes physics solver to achieve continuous water behavior modeling; Feedback modeling module, used to capture the effect of water on plant water absorption and morphological changes, and pass feedback parameters to the plant simulation module; An interactive visualization module, including a 3D modeling interface, a heat map analysis window, and a parameter control panel, allows users to edit terrain, plants, and water configurations in real time, and provides operational feedback through a GPU rendering engine. The natural environment response module is used to perform microclimate calculation tasks based on sunshine prediction models, wind field turbulence models and graph neural networks, and output environmental data for adjusting landscape layout; the evaluation and integration module is used to output ecological performance evaluation reports, tourist path visualization data and data interface communication with the BIM / GIS platform.

Images