AI-based old community reconstruction terrain simulation display method and system

Abstract

The invention provides an AI-based old community reconstruction terrain simulation display method and system, and relates to the field of terrain simulation, and the method comprises the steps: obtaining the multi-source high-precision terrain data of an old community, and carrying out the cleaning and fusion; according to the topographic data set, identifying topographic key features and classifying land cover types in combination with a semantic segmentation technology so as to obtain a digital topographic feature map with semantic tags; resident investigation and survey data, policy specifications and economic cost constraints are obtained, constraint conditions are quantified, and a structured demand constraint matrix is generated; constructing a multi-objective optimization model, and iteratively generating and verifying a terrain reconstruction scheme set by adopting deep reinforcement learning; an immersive terrain simulation environment is constructed, a scheme is adjusted in real time, user behavior data is recorded, and a final optimization scheme is output; and performing visual achievement and dynamic demonstration to obtain a multi-terminal compatible transformation scheme package. According to the method, the defect that intelligent tool application in the prior art flows in the form and cannot be deeply combined with a core transformation link is overcome.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of terrain simulation, in particular to an AI-based old community reconstruction terrain simulation display method and system. BACKGROUND

[0002] Old city reconstruction should take digital technology and physical economy as the main line, strengthen digital infrastructure construction, perfect digital economic governance system, and promote digital industrialization and industrial digitization, and empower traditional industry transformation and upgrading; old community reconstruction in cities and towns is a major livelihood project and development project, which is of great significance to meet the needs of the people for a better life, promote the expansion of domestic demand, promote urban renewal and development, and promote the transformation of economic development.

[0003] Nowadays, the limitations of traditional reconstruction methods are more and more prominent; traditional manual diagnosis is low in efficiency and high in cost, and it is difficult to accurately capture the real demands of residents; the design of the reconstruction scheme depends on manual experience, and is highly repetitive and poorly adaptable, making it difficult to quickly respond to diversified project scenarios.

[0004] In recent years, intelligent technology and software have gradually popularized in the field of architecture and urban planning, but there are still obvious deficiencies in their practical application in old city reconstruction; although design institutes and construction units have introduced related technologies, their application is mostly formal and has not been deeply combined with specific projects; the potential of intelligent tools in demand analysis, scheme design, and effect optimization has not been fully utilized, and it is difficult to effectively solve the core problems in old city reconstruction; therefore, promoting the deep application of AI technology in old city reconstruction is not only an inevitable trend of industry development, but also an important path for enterprises to realize transformation and upgrading. SUMMARY

[0005] The application provides an AI-based old community reconstruction terrain simulation display method and system to solve the problem that intelligent tools are only used in a formal way and cannot be deeply combined with core reconstruction links in the prior art.

[0006] In one aspect, the application provides an AI-based old community reconstruction terrain simulation display method, which comprises:

[0007] Obtain multi-source high-precision terrain data of the old community, and clean and fuse the data to obtain a terrain data set, wherein the high-precision terrain data includes elevation, slope, building contour, and road network;

[0008] According to the terrain data set, identify key terrain features and classify the ground cover types by combining a semantic segmentation technology to obtain a digital terrain feature map with semantic labels;

[0009] Obtain resident survey data, policy specifications and economic cost constraints, quantify the constraint conditions, and generate a structured demand constraint matrix;

[0010] Based on the digital terrain feature map with semantic labels and the structured demand constraint matrix, a multi-objective optimization model is constructed, and deep reinforcement learning is used to iteratively generate and verify the terrain transformation scheme set;

[0011] Based on the terrain transformation scheme set, an immersive terrain simulation environment is constructed and the scheme is adjusted in real time, user behavior data is recorded and the final optimization scheme is output;

[0012] Based on the final optimization scheme, a visualized result and dynamic demonstration are performed to obtain a multi-terminal compatible transformation scheme package.

[0013] Further, obtain multi-source high-precision terrain data of old communities, and perform cleaning and fusion to obtain a terrain data set, the high-precision terrain data including elevation, slope, building contour, road network, comprising:

[0014] Obtain multi-source high-precision terrain data by unmanned aerial vehicle oblique photography, LiDAR, satellite remote sensing and GIS topographic map;

[0015] Filter and denoise the multi-source high-precision terrain data, complete missing data, and unify the coordinate system to obtain cleaned multi-source data;

[0016] Align the multi-source data and construct a digital elevation model to obtain a terrain data set.

[0017] Further, according to the terrain data set, identify key terrain features and classify surface cover types using semantic segmentation technology to obtain a digital terrain feature map with semantic labels, including:

[0018] According to the terrain data set, identify risk features such as steep slope areas and low-lying areas to obtain a terrain risk feature matrix;

[0019] According to the terrain data set, classify the ground and use conditional random fields to refine the edges to obtain a ground classification map matrix;

[0020] Multiply the terrain risk feature matrix and the ground classification map matrix to generate a digital terrain feature map with composite semantic labels.

[0021] Further, obtain resident survey data, policy specifications and economic cost constraints, quantify the constraint conditions, and generate a structured demand constraint matrix, including:

[0022] Integrate resident survey text, policy specification files and economic cost data, perform text cleaning, document parsing and cost standardization to obtain preprocessed resident survey data, policy specifications and economic cost constraint data;

[0023] Based on the pre-processed resident survey data, extract demand entities, combine resident identity weight to calculate priority, and get demand priority;

[0024] Based on the pre-processed policy specification, parse the specification document, regular match the key parameters and convert to algorithm executable constraint condition, get the digital policy specification;

[0025] Based on the pre-processed economic cost constraint, build unit cost database, dynamically check budget feasibility and adjust demand weight, get economic constraint model;

[0026] Based on the demand priority, digital policy specification and economic constraint model, generate structured demand constraint matrix.

[0027] Further, based on the digital terrain feature map with semantic label and structured demand constraint matrix, construct a multi-objective optimization model, and use deep reinforcement learning to iteratively generate and verify the terrain reconstruction scheme set, including:

[0028] Based on the digital terrain feature map with semantic label and structured demand constraint matrix, design a composite objective function, and use a dynamic weight adjustment strategy to get a multi-objective optimization model;

[0029] Based on the multi-objective optimization model, use deep reinforcement learning to construct a state space containing terrain parameters, define an action space, and guide the optimization direction through a reward function, iteratively generate multiple terrain reconstruction schemes;

[0030] Based on multiple terrain reconstruction schemes, real-time collision detection is performed to obtain detection results;

[0031] The qualified terrain reconstruction scheme is integrated into the terrain reconstruction scheme set.

[0032] Further, based on the terrain reconstruction scheme set, construct an immersive terrain simulation environment and adjust the scheme in real time, record user behavior data and output the final optimization scheme, including:

[0033] Based on the terrain reconstruction scheme set, build a virtual environment and develop an interactive parameter adjustment system to get the constructed virtual environment;

[0034] Based on the constructed virtual environment, use multi-modal interaction to adjust parameters in real time to get multiple adjustment schemes;

[0035] Based on multiple adjustment schemes, record user spatiotemporal trajectory and parameter modification behavior (frequency / restoration rate) to get collected behavior data;

[0036] Based on the collected behavior data, extract common preferences through cluster analysis to get the final optimization scheme.

[0037] Further, based on the final optimization scheme, visual results and dynamic demonstration are carried out to obtain a multi-terminal compatible reconstruction scheme package, including:

[0038] The final optimization scheme is converted into a lightweight and parameter format conversion to obtain standardized scheme data;

[0039] Based on the standardized scheme data, a slope heat map and an interactive reconstruction comparison chart are generated to generate a two-dimensional visual scheme:

[0040] Based on the standardized scheme data, three-dimensional modeling is carried out using BIM to obtain a three-dimensional visual scheme:

[0041] Based on the standardized scheme data, rain season drainage is simulated, and construction progress timing is displayed to obtain a dynamic demonstration scheme;

[0042] The two-dimensional visual scheme, the three-dimensional visual scheme and the dynamic demonstration scheme are classified and packaged to obtain a multi-terminal compatible reconstruction scheme package.

[0043] On the other hand, an AI-based old community reconstruction terrain simulation display system includes:

[0044] The acquisition module is used to acquire multi-source high-precision terrain data of the old community, and to clean and fuse to obtain a terrain data set, wherein the high-precision terrain data includes elevation, slope, building contour, and road network;

[0045] The processing module is used to identify terrain key features according to the terrain data set and classify the ground cover type by combining the semantic segmentation technology to obtain a digital terrain feature map with semantic labels; resident survey data, policy specifications and economic cost constraints are obtained, and the constraint conditions are quantized to generate a structured demand constraint matrix; based on the digital terrain feature map with semantic labels and the structured demand constraint matrix, a multi-objective optimization model is constructed, and a deep reinforcement learning is used to generate and verify a terrain reconstruction scheme set; based on the terrain reconstruction scheme set, an immersive terrain simulation environment is constructed and the scheme is adjusted in real time, user behavior data is recorded and an final optimization scheme is output; based on the final optimization scheme, visual results and dynamic demonstration are carried out to obtain a multi-terminal compatible reconstruction scheme package.

[0046] On the other hand, the present application also provides an electronic device, including a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the program as described above to realize the AI-based old community reconstruction terrain simulation display method.

[0047] In another aspect, the present application also provides a non-transitory computer-readable storage medium having stored thereon a computer program, which, when executed by a processor, implements the AI-based old community renovation terrain simulation display method of any of the above.

[0048] In another aspect, the present application also provides a computer program product comprising a computer program, which, when executed by a processor, implements the AI-based old community renovation terrain simulation display method of any of the above.

[0049] The AI-based old community renovation terrain simulation display method and system provided by the present application systematically improve old renovation efficiency and quality, realize full-chain AI empowerment from demand collection, scheme generation, and construction simulation, and compress the overall project cycle; through reinforcement learning algorithm, the contradictory objectives of space utilization rate, renovation cost, and old-age adaptation score are balanced, and the comprehensive benefits are improved compared with traditional design; The "knowledge graph + deep learning" architecture is created, the design experience rules are coded as computable nodes, the scheme compliance rate is improved, and professional control is retained at the same time; the personalized model library supports one-key generation of various regional architectural styles, the cost of effect picture production is reduced, and the old-age adaptation algorithm automatically adapts to the climate difference between the north and the south; Through VR immersive participation, the adoption rate of resident demands is improved, and the satisfaction after renovation is improved; a multi-type reusable AI old renovation tool chain is formed, the per capita output value of design institutes is improved, the industry is promoted from labor-intensive to technology-intensive, the risk of violation is reduced through standardized digital verification, the scheme package supports multiple terminals from AR inspection (mobile terminal) to BIM collaboration (workstation), the implementation adaptation cost is reduced, and the terrain simulation system can predict the climate impact for 30 years, so that the life cycle of the renovation scheme is extended; The present application realizes the paradigm change from experience dominance to digital empowerment in old city renovation by constructing a new mode of "data intelligent driving, man-machine collaborative decision-making, and full-cycle closed-loop management", and provides a replicable and generalizable technical model for urban renewal. BRIEF DESCRIPTION OF DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0054] Figure 1 is a flowchart of the AI-based old community renovation terrain simulation display method provided by the embodiment of the present application;

[0055] Figure 2 is a schematic diagram of an AI-based old community renovation terrain simulation display system provided by an embodiment of the present application.

[0056] Figure 3 is a structural schematic diagram of an electronic device. DETAILED DESCRIPTION

[0057] To make the objectives, technical solutions, and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below with reference to the drawings in the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0058] Figure 1 is one of the flow schematic diagrams of an AI-based old community renovation terrain simulation display method provided by an embodiment of the present application.

[0059] As shown in Figure 1 , the AI-based old community renovation terrain simulation display method provided by an embodiment of the present application mainly includes the following steps:

[0060] 11. Obtain multi-source high-precision terrain data of an old community, and perform cleaning and fusion to obtain a terrain data set, wherein the high-precision terrain data includes elevation, slope, building contour, and road network;

[0061] 12. According to the terrain data set, identify terrain key features and classify surface cover types in combination with a semantic segmentation technology to obtain a digital terrain feature map with semantic labels;

[0062] 13. Obtain resident survey data, policy specifications, and economic cost constraints, quantify the constraint conditions, and generate a structured demand constraint matrix;

[0063] 14. Based on the digital terrain feature map with semantic labels and the structured demand constraint matrix, construct a multi-objective optimization model, and use deep reinforcement learning to iteratively generate and verify a terrain renovation scheme set;

[0064] 15. Based on the terrain renovation scheme set, construct an immersive terrain simulation environment and adjust the scheme in real time, record user behavior data, and output a final optimization scheme;

[0065] 16. Based on the final optimization scheme, perform visualized achievement and dynamic demonstration to obtain a multi-terminal compatible renovation scheme package.

[0066] In the embodiment of the application, by integrating multi-dimensional terrain data such as elevation, slope, building contour, road network, a high-precision three-dimensional terrain model is constructed to eliminate the data island problem in traditional reconstruction and provide a reliable basement for scheme generation; combined with semantic segmentation technology, the ground cover type is finely classified to give terrain spatial semantic information and support accurate positioning of the reconstruction target; resident survey data, policy specifications and economic costs are converted into a structured constraint matrix to realize mathematical expression of unstructured demand and avoid subjective decision bias; through a multi-objective optimization model, reconstruction contradictions are balanced, and a deep reinforcement learning is used to automatically generate a Pareto optimal scheme set under complex constraints to improve the scientificity and feasibility of the scheme; a VR / AR simulation environment is constructed to support real-time adjustment of scheme parameters by users, and through behavior data collection, the model iteration is fed back to form a closed-loop optimization mechanism of "man-machine cooperation"; a post-reconstruction scene is dynamically simulated to predict potential problems in advance and reduce the risk of later rework.

[0067] As shown in Figure 1 , 11, multi-source high-precision terrain data of the old community is obtained, and cleaning and fusion are performed to obtain a terrain data set, the high-precision terrain data including elevation, slope, building contour, road network, comprising:

[0068] 111, multi-source high-precision terrain data is obtained by unmanned aerial vehicle oblique photography, LiDAR, satellite remote sensing and GIS topographic map;

[0069] 112, the multi-source high-precision terrain data is filtered and denoised, missing data is completed, and the coordinate system is unified to obtain the cleaned multi-source data;

[0070] 113, the multi-source data is registered and aligned, and a digital elevation model is constructed to obtain the terrain data set.

[0071] In the embodiment of the application, building facade texture and detailed structure (such as roof slope, illegal construction) are obtained, with a resolution of centimeter level, making up for the vertical angle limitation of traditional orthographic images; LiDAR (laser radar) penetrates the vegetation canopy to obtain the true ground elevation, accurately identifies the terrain undulation (such as the underground garage entrance slope), and is suitable for densely wooded areas; satellite remote sensing provides background data for a large area (such as the connection of surrounding roads, regional terrain trend), which assists in analyzing the spatial relationship between the community and the surrounding environment; GIS topographic map integrates historical planning data (such as underground pipelines, historical protected building locations), which avoids damaging hidden infrastructure during reconstruction; through multi-technology cooperation, key parameters such as elevation (DEM), slope, building contour (including number of layers, facade material), road network (including width, trend) are synchronously obtained to form an integrated data system of "terrain-building-traffic";

[0072] The vegetation noise (such as outliers caused by tree branches swinging) in the LiDAR point cloud is removed by bilateral filtering or wavelet transform, and the true ground shape is retained; the elevation data of the cloud-shielded area of satellite remote sensing is filled by Kriging interpolation method, or the key nodes (such as building shadow area) are supplemented by unmanned aerial vehicle photography; different source data (WGS84, CGCS2000, etc.) are converted into unified projection coordinate system (such as UTM), and spatial analysis error is eliminated (error control within ±0.1 meters);

[0073] Based on feature point matching (such as building corner points, road intersections), multi-source data spatial alignment is realized, ensuring that the building contour corresponds to the elevation model strictly, fusing LiDAR point cloud and oblique photography three-dimensional model, and generating high-precision digital elevation model (resolution ≤0.1 meters), which can identify micro-terrain features such as steps and slopes;

[0074] The cleaned data set supports the semantic segmentation model to accurately distinguish roof, road, green land and other land types, providing basis for priority ranking of reconstruction; building contour data combined with policy specifications can automatically calculate the influence of added floors on the lighting of surrounding buildings, and generate a structured constraint matrix; high-precision DEM supports physical simulation of character walking and vehicle passing in immersive scenes, making scheme verification closer to real scenes.

[0075] As shown in Figure 1 According to the terrain data set, the terrain key features are identified and the ground cover types are classified by combining semantic segmentation technology to obtain a digital terrain feature map with semantic labels, including:

[0076] 121、According to the terrain data set, the risk features such as steep slope area and low-lying area are identified to obtain a terrain risk feature matrix;

[0077] 122、According to the terrain data set, the ground classification is carried out and the edges are refined using conditional random field to obtain a ground classification map matrix;

[0078] 123、The terrain risk feature matrix is multiplied with the ground classification map matrix to generate a digital terrain feature map with composite semantic labels.

[0079] In the embodiments of the present application, the slope is calculated based on the elevation data, the area with a slope > 15% is marked as a potential landslide risk area, and the risk level is further subdivided in combination with the soil type data (such as the proportion of clay / sand); low-lying area positioning: simulate surface runoff through hydrological analysis, identify low-lying areas with water accumulation depth > 0.3m, and mark them as waterlogging risk points to provide precise target areas for drainage system reconstruction; geological hazard marking: integrate geological survey data (such as fault zones and backfill areas) to generate a terrain risk feature matrix, and realize visual warning of geological disaster risk; the risk feature matrix supports sorting of the geological units according to the risk level (high / medium / low), and prioritizes processing of high-risk areas to ensure the safety of the reconstruction project; a network structure such as U-Net++ is used to classify oblique photography images at the pixel level, and distinguish between road, building, green land, and bare land; through modeling of the spatial relationship between pixels, the classification edge is optimized (such as distinguishing the fine cracks at the junction of road and building), so that the road network extraction accuracy is improved; the ground classification matrix supports the generation of land use status map, and clearly defines the potential areas for reconstruction (such as idle green land that can be converted into parking spaces), providing data support for function replacement;

[0080] The risk feature matrix and the ground classification matrix are subjected to Hadamard product operation to generate a composite label containing both risk level and ground type (such as "high risk-road" and "low risk-green land"); the geological units are assigned with three attributes of risk level, ground type, and elevation value, supporting complex queries; in the immersive simulation environment, the system can automatically recommend low-risk-green land as a candidate area for adding a children's activity area, or mark high-risk-building areas that need to be reinforced first, and estimate the difficulty of reconstruction in combination with the risk level and ground type.

[0081] As shown in Figure 1 , 13, the resident survey data, policy specifications and economic cost constraints are obtained, the constraint conditions are quantified, and a structured demand constraint matrix is generated, including:

[0082] 131, integrate the resident survey text, policy specification file and economic cost data, perform text cleaning, document parsing and cost standardization to obtain preprocessed resident survey data, policy specification and economic cost constraint data;

[0083] 132, based on the preprocessed resident survey data, extract demand entities, calculate priority in combination with resident identity weight, and obtain demand priority;

[0084] 133, based on the preprocessed policy specification, parse the specification document, regularly match the key parameters and convert them into algorithm executable constraint conditions to obtain digitized policy specification;

[0085] 134. Construct a unit cost database based on the pre-processed economic cost constraints, dynamically check the budget feasibility and adjust the demand weight to obtain an economic constraint model;

[0086] 135. Generate a structured demand constraint matrix based on the demand priority, digitized policy specification and economic constraint model.

[0087] In the embodiments of the present application, NLP technology (such as BERT-CRF model) is used to extract demand entities from resident survey texts, and the recognition rate is improved; demand priority weights are assigned based on resident attributes to ensure that the needs of vulnerable groups are met first; key parameters are extracted from policy documents and converted into algorithm executable constraint conditions; policy constraints are embedded in the optimization model to automatically exclude illegal schemes and reduce the rework rate of schemes; a standardized database containing demolition compensation and other costs is constructed to support fast cost estimation; demand weights are dynamically adjusted through Monte Carlo simulation to generate a set of cost-controllable schemes;

[0088] The matrix integrates demand priority, policy constraints and economic restrictions in three dimensions to provide standardized input for multi-objective optimization models, supports flexible adjustment of constraint weights, and realizes dynamic adaptation of schemes; the matrix visualizes the conflict relationship of each constraint condition to assist decision makers in weighing and choosing; key constraints are identified through sensitivity analysis to focus on cost control.

[0089] As shown in Figure 1 , 14, based on the digital terrain feature map with semantic labels and the structured demand constraint matrix, a multi-objective optimization model is constructed, and deep reinforcement learning is used to iteratively generate and verify a set of terrain transformation schemes, including:

[0090] 141. Based on the digital terrain feature map with semantic labels and the structured demand constraint matrix, a composite objective function is designed, and a dynamic weight adjustment strategy is used to obtain a multi-objective optimization model;

[0091] 142. Based on the multi-objective optimization model, deep reinforcement learning is used to construct a state space containing terrain parameters, define an action space, and guide the optimization direction through a reward function to iteratively generate multiple terrain transformation schemes;

[0092] 143. Based on the multiple terrain transformation schemes, real-time collision detection is performed to obtain a detection result;

[0093] 144. The qualified terrain transformation schemes are integrated into a set of terrain transformation schemes.

[0094] In the embodiments of the present application, in combination with the digital terrain feature map, the slope optimization and the drainage efficiency improvement are set as the goals; based on the structured constraint matrix, the resident demand priority, the policy compliance and the economic cost are quantified as the constraints; the entropy weight method or the analytic hierarchy process is used to dynamically adjust the target weight according to the feedback of the stakeholders, so as to realize the demand-oriented optimization; the policy specification, the economic cost and the resident demand are converted into hard constraints or soft constraints and embedded into the optimization model; the digital terrain feature map is taken as the input, and the parameters such as elevation, slope and ground type are coded; the terrain reconstruction operation is defined, and each operation is accompanied by parameters; the multi-dimensional reward is designed to guide the model to generate a scheme that meets the constraints; the PPO (Proximal Policy Optimization) algorithm is used to learn by trial and error (for example, after trying different filling and digging schemes, the action probability is adjusted according to the reward feedback), so that the model gradually converges to the optimal strategy; the noise disturbance is introduced (for example, Gaussian noise is added to the action parameters), so as to avoid falling into local optimum and generate a scheme set containing 50-100 schemes;

[0095] Based on the three-dimensional terrain model, the spatial conflicts in the reconstruction scheme are detected, such as the collision between the newly built building and the existing pipeline, and the occupation of the fire access by the road widening; it is verified whether the scheme violates the policy constraints or the demand priority, and the unqualified scheme is filtered; from the collision detection qualified scheme, the Pareto optimal solution that cannot be simultaneously optimized by other schemes in multiple targets is selected; a scheme set containing 3-5 representative schemes is generated for the decision maker to select.

[0096] As shown in Figure 1 Based on the terrain reconstruction scheme set, an immersive terrain simulation environment is constructed and the scheme is adjusted in real time, the user behavior data is recorded and the final optimization scheme is output, including:

[0097] 151, based on the terrain reconstruction scheme set, a virtual environment is built and an interactive parameter adjustment system is developed to obtain the constructed virtual environment;

[0098] 152, based on the constructed virtual environment, the parameters are adjusted in real time using multi-modal interaction to obtain multiple adjustment schemes;

[0099] 153, based on the multiple adjustment schemes, the user spatiotemporal trajectory and parameter modification behavior (frequency / restoration rate) are recorded to obtain the collected behavior data;

[0100] 154, based on the collected behavior data, the common preference is extracted through cluster analysis to obtain the final optimization scheme.

[0101] In the embodiments of the present application, based on the set of terrain modification schemes, a high-precision 3D model is generated to support seamless access of VR / AR devices; adjustable controls are developed to enable users to modify parameters in real time and immediately view the effects; multi-modal input such as gestures, voice, and somatosensory is supported to reduce the operation threshold; real-time rendering of lighting, shadows, and materials is achieved using engines such as Unreal Engine 5 to enhance immersion; eye trackers and position trackers are used to capture user visual focus and walking paths to identify areas of interest; parameter adjustment frequency and restoration rate are recorded to quantify demand intensity; K-means or DBSCAN algorithms are used to divide users into conservative, functional, and ecological groups based on behavior data; high-frequency modification parameters and high-attention areas are identified to generate group preference profiles; after users adjust the parameters, the system automatically recalculates the constraints and updates the scheme; when user operations result in policy violations or cost overruns, real-time warnings are popped up; the clustering analysis results are fed back to the multi-objective optimization model to generate a scheme that takes into account the common needs of the group and the individual differences; a comparison table containing key indicators is automatically generated to assist decision-makers in making choices.

[0102] As shown in Figure 1 , 16, based on the final optimized scheme, visual results and dynamic demonstrations are carried out to obtain a multi-terminal compatible modification scheme package, including:

[0103] 161, convert the final optimized scheme into a lightweight and parameter format to obtain standardized scheme data;

[0104] 162, based on the standardized scheme data, generate a slope heat map and an interactive modification comparison chart to generate a two-dimensional visual scheme:

[0105] 163, based on the standardized scheme data, use BIM for three-dimensional modeling to obtain a three-dimensional visual scheme:

[0106] 164, based on the standardized scheme data, simulate rain season drainage and display construction progress timing to obtain a dynamic demonstration scheme;

[0107] 165, classify and package the two-dimensional visual scheme, the three-dimensional visual scheme, and the dynamic demonstration scheme to obtain a multi-terminal compatible modification scheme package.

[0108] In the embodiments of the present application, the final scheme is decomposed into independent parameters and converted into a universal format, ensuring seamless data calling in GIS, BIM and other software; Draco compression algorithm is used to reduce the volume of three-dimensional model files, supporting fast loading on the web side; two-dimensional drawings and three-dimensional models of different resolutions are generated, adapting to multiple scenarios such as large-screen decision-making systems, tablet display terminals, and mobile APPs; the slope value is encoded with a color gradient to intuitively display the landslide risk area and assist decision-makers in positioning key reinforcement; superimpose the heat maps before and after the transformation to quantify the slope optimization effect and enhance the persuasiveness of the scheme; support one-key switching between "current situation" and "scheme" views to compare the changes in key indicators; embed text annotations and arrow indicators to make the scheme interpretation more intuitive; integrate building structures, mechanical and electrical pipelines, and landscape elements to support collision detection; apply PBR material systems to make the model appearance highly consistent with the actual construction effect; support arbitrary angle cutting of the model to check the spatial rationality; calculate the annual sunshine trajectory based on astronomical algorithms to verify whether the scheme meets the "Sunshine ≥ 2 hours on the Da Han Day" specification; simulate the surface runoff path based on terrain DEM and rainfall data to verify the design effect of the sponge city; risk warning: highlight the points of waterlogging risk to guide the optimization of the drainage system; associate the construction plan with the three-dimensional model to generate a dynamic progress bar; show the traffic diversion plan during construction to predict the impact on residents' travel; display the three-dimensional model and dynamic demonstration through a large-screen system to assist high-level decision-making; provide BIM models for engineers to further design and output construction drawings; and show the scheme to residents through a mobile APP or an exhibition hall touch screen to enhance their sense of participation.

[0109] As shown in Figure 2 An AI-based old community transformation terrain simulation display system 20 includes:

[0110] An acquisition module 21 is configured to acquire multi-source high-precision terrain data of an old community, and clean and fuse the data to obtain a terrain data set, wherein the high-precision terrain data includes elevation, slope, building contour, and road network.

[0111] A processing module 22 is configured to identify terrain key features and classify surface cover types by combining a semantic segmentation technique based on the terrain data set, to obtain a digital terrain feature map with semantic labels; acquire resident survey data, policy specifications, and economic cost constraints, quantify the constraint conditions, and generate a structured demand constraint matrix; based on the digital terrain feature map with semantic labels and the structured demand constraint matrix, construct a multi-objective optimization model, and iteratively generate and verify a terrain transformation scheme set by using deep reinforcement learning; based on the terrain transformation scheme set, construct an immersive terrain simulation environment and adjust the scheme in real time, record user behavior data, and output a final optimized scheme; based on the final optimized scheme, perform visual achievement and dynamic demonstration, and obtain a multi-terminal compatible transformation scheme package.

[0112] Figure 3This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention.

[0113] like Figure 3 As shown, the electronic device may include a processor 610, a communications interface 620, a memory 630, and a communication bus 640. The processor 610, communications interface 620, and memory 630 communicate with each other via the communication bus 640. The processor 610 can call logical instructions from the memory 630 to execute an AI-based method for simulating and displaying terrain for the renovation of old residential areas.

[0114] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0115] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the AI-based terrain simulation and display method for the renovation of old residential areas provided by the above methods.

[0116] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the AI-based terrain simulation and display method for the renovation of old residential areas provided by the above methods.

[0117] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed to multiple network units. Part or all of the modules can be selected to achieve the purposes of the embodiments according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0118] Through the description of the above embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software and the necessary general hardware platform, and of course can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some parts of the embodiments.

[0119] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An AI-based old district renovation terrain simulation display method, characterized by, The method comprises the following steps: Obtain multi-source high-precision terrain data of the old community, and clean and fuse the data to obtain a terrain data set, wherein the high-precision terrain data comprises elevation, slope, building contour and road network; According to the terrain data set, identify key terrain features and classify the ground cover types by combining the semantic segmentation technology to obtain a digital terrain feature map with semantic labels; Obtain resident survey data, policy specifications and economic cost constraints, quantify the constraint conditions, and generate a structured demand constraint matrix; Based on the digital terrain feature map with semantic labels and the structured demand constraint matrix, a multi-objective optimization model is constructed, and deep reinforcement learning is used to iteratively generate and verify a terrain reconstruction scheme set; Based on the terrain reconstruction scheme set, an immersive terrain simulation environment is constructed, and the scheme is adjusted in real time, user behavior data is recorded, and a final optimized scheme is output; Based on the final optimized scheme, a visualized result and dynamic demonstration are performed to obtain a multi-terminal compatible reconstruction scheme package. 2.The AI-based old village renovation topography simulation display method of claim 1, wherein, Obtain multi-source high-precision terrain data of the old community, and clean and fuse the data to obtain a terrain data set, wherein the high-precision terrain data comprises elevation, slope, building contour and road network, comprising: Obtain multi-source high-precision terrain data by using unmanned aerial vehicle oblique photography, LiDAR, satellite remote sensing and GIS topographic map; Filter and denoise the multi-source high-precision terrain data, complete the missing data, and unify the coordinate system to obtain cleaned multi-source data; Register and align the multi-source data, and construct a digital elevation model to obtain a terrain data set. 3.The AI-based old village renovation topography simulation display method of claim 2, wherein, According to the terrain data set, identify key terrain features and classify the ground cover types by combining the semantic segmentation technology to obtain a digital terrain feature map with semantic labels, comprising: According to the terrain data set, identify risk features such as steep slope areas and low-lying areas to obtain a terrain risk feature matrix; According to the terrain data set, classify the ground and use conditional random field to refine the edges to obtain a ground classification matrix; Multiply the terrain risk feature matrix and the ground classification matrix to generate a digital terrain feature map with compound semantic labels. 4.The AI-based old village renovation topography simulation display method of claim 3, wherein, Obtain resident survey data, policy specifications and economic cost constraints, quantify the constraint conditions, and generate a structured demand constraint matrix, comprising: Integrate resident survey texts, policy specification files and economic cost data, perform text cleaning, document parsing and cost standardization to obtain preprocessed resident survey data, policy specifications and economic cost constraint data; Based on the preprocessed resident survey data, extract demand entities, calculate priorities in combination with resident identity weights to obtain demand priorities; Based on the preprocessed policy specifications, parse the specification documents, regularly match key parameters and convert them into algorithm-executable constraint conditions to obtain digital policy specifications; Based on the preprocessed economic cost constraints, construct a unit cost database, dynamically verify the budget feasibility and adjust the demand weights to obtain an economic constraint model; Based on the demand priorities, digital policy specifications and economic constraint model, generate a structured demand constraint matrix. 5.The AI-based old village renovation topography simulation display method of claim 4, wherein, Based on semantically labeled digital terrain feature maps and structured requirement constraint matrices, a multi-objective optimization model is constructed. Deep reinforcement learning is then used to iteratively generate and validate a set of terrain modification schemes, including: Based on semantically labeled digital terrain feature maps and structured requirement constraint matrices, a composite objective function is designed, and a dynamic weight adjustment strategy is adopted to obtain a multi-objective optimization model. Based on a multi-objective optimization model, deep reinforcement learning is used to construct a state space containing terrain parameters, define an action space, and guide the optimization direction through a reward function to iteratively generate multiple terrain modification schemes. Based on multiple terrain modification schemes, real-time collision detection is performed to obtain the detection results; The qualified terrain modification schemes were integrated into a terrain modification scheme set. 6.The AI-based old village renovation topography simulation display method of claim 5, wherein, Based on a set of terrain modification schemes, an immersive terrain simulation environment is constructed and the schemes are adjusted in real time. User behavior data is recorded and the final optimized scheme is output, including: Based on a set of terrain modification schemes, a virtual environment is built and an interactive parameter adjustment system is developed to obtain the constructed virtual environment; Based on the constructed virtual environment, parameters are adjusted in real time using multimodal interaction to obtain multiple adjustment schemes; Based on multiple adjustment schemes, record the user's spatiotemporal trajectory and parameter modification behavior (frequency / restore rate) to obtain the collected behavioral data; Based on the collected behavioral data, common preferences are extracted through cluster analysis to obtain the final optimization solution. 7.The AI-based old village renovation topography simulation display method of claim 6, wherein, Based on the final optimized solution, visualization and dynamic demonstrations were performed to obtain a multi-terminal compatible transformation solution package, including: The final optimized solution is converted to a lightweight form and the parameter format is transformed to obtain standardized solution data; Based on standardized scheme data, a slope heat map and an interactive modification comparison map are generated, resulting in a two-dimensional visualization scheme. Based on standardized solution data, BIM is used for 3D modeling to obtain a 3D visualization solution: Based on standardized scheme data, simulated rainy season drainage and display of construction progress sequence to obtain dynamic demonstration scheme; The 2D visualization solution, 3D visualization solution, and dynamic demonstration solution are categorized and packaged to obtain a multi-terminal compatible modification solution package. 8.An AI-based old district renovation terrain simulation display system, characterized by, include: The acquisition module is used to acquire multi-source high-precision terrain data of old residential areas, and to clean and fuse it to obtain a terrain dataset. The high-precision terrain data includes elevation, slope, building outline, and road network. The processing module is used to identify key terrain features based on the terrain dataset and classify land cover types using semantic segmentation technology to obtain a digital terrain feature map with semantic labels; acquire resident survey data, policy regulations, and economic cost constraints, quantify the constraints, and generate a structured demand constraint matrix; based on the digital terrain feature map with semantic labels and the structured demand constraint matrix, construct a multi-objective optimization model, and use deep reinforcement learning to iteratively generate and verify a set of terrain transformation schemes; Based on the terrain modification scheme set, an immersive terrain simulation environment is constructed and the scheme is adjusted in real time. User behavior data is recorded and the final optimized scheme is output. Based on the final optimized solution, the results are visualized and dynamically demonstrated to obtain a transformation solution package that is compatible with multiple terminals.

9. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor implements the AI-based old cell reconstruction topography simulation display method according to any one of claims 1 to 7 when executing the program.

10. A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program implements the AI-based old cell reconstruction topography simulation display method according to any one of claims 1 to 7 when executed by the processor.

Images