Artificial intelligence-based urban health examination improvement scheme optimization method and system

Abstract

The application discloses an urban health check improvement scheme optimization method and system based on artificial intelligence, relates to the technical field of smart cities, and comprises the following steps: acquiring social media text data and urban planning GIS data, performing spatiotemporal alignment processing, and obtaining a structured data set; analyzing the structured data set, calculating building stress distribution and road bearing capacity, and obtaining a physical health heat map; performing sentiment analysis on the social media text data using a BERT model, combining geocoding technology to obtain a citizen emotion density distribution map, constructing a physically constrained neural network model, embedding the physical health heat map into a network loss function, and obtaining a three-dimensional structure health diagnosis report. The application innovatively couples and analyzes building stress distribution and citizen emotion density by constructing a physically constrained neural network model, utilizes a PDE loss function, realizes multi-modal deep fusion of structure health and social emotion, and can accurately identify high-stress-high-emotion coupling areas.

Description

Technical Field

[0001] This invention relates to the field of smart city technology, and in particular to an optimization method and system for improving urban health check-up schemes based on artificial intelligence. Background Technology

[0002] Currently, the smart city field has gradually adopted multi-source data fusion technology for urban health status assessment. This involves deploying vibration sensor networks to collect building structure response data, combining finite element analysis to calculate stress distribution, and using spatial interpolation algorithms to generate physical health heat maps. Some studies have attempted to introduce social media data to assist decision-making, using sentiment analysis models to classify citizen feedback and generating sentiment distribution maps based on kernel density estimation. Existing technologies have initially achieved parallel analysis of physical monitoring data and social perception data, and have formed a certain standardized process in data acquisition and visualization.

[0003] Existing technologies typically employ an independent analysis followed by overlay presentation, failing to establish a dynamic correlation mechanism between physical conditions and public sentiment. Physical health heatmaps only reflect the structural mechanical state, neglecting the potential impact of high-stress areas on public sentiment. The lack of interactive feedback between sentiment analysis results and engineering parameters makes it difficult to quantify the positive impact of improved sentiment on structural reinforcement benefits when optimizing planning schemes. When high stress concentration and negative sentiment accumulation occur simultaneously in a certain area, existing technologies cannot quantify the coupling effect between the two through mathematical models, relying solely on manual experience to adjust the scheme, which can easily lead to resource allocation deviations. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an optimization method for improving urban health checkup schemes based on artificial intelligence, which solves the problem of suboptimal planning schemes caused by the separation of physical state and citizen emotions in existing technologies.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides an optimization method for improving urban physical examination schemes based on artificial intelligence, which includes acquiring social media text data and urban planning GIS data and performing spatiotemporal alignment processing to obtain a structured dataset;

[0008] The structured dataset was analyzed to calculate the building stress distribution and road bearing capacity, resulting in a physical health heat map.

[0009] The BERT model is used to perform sentiment analysis on social media text data. Combined with geocoding technology, a citizen sentiment density distribution map is obtained. A physically constrained neural network model is constructed, and a physical health heat map is embedded into the network loss function to obtain a three-dimensional structural health diagnosis report.

[0010] Based on the 3D structural health diagnosis report, the CityEngine planning module is used to generate a set of feasible renovation plans.

[0011] The set of feasible renovation schemes is imported into the AnyLogic simulation module to simulate social benefits, and an evaluation report of the economic benefit prediction curve is obtained. A partial differential equation for emotion propagation is established, and the concentration of negative emotions is predicted based on the distribution map of citizens' emotions. The public space configuration ratio of the set of feasible renovation schemes is dynamically adjusted to obtain the optimized scheme.

[0012] As a preferred embodiment of the AI-based urban health check improvement scheme optimization method described in this invention, the following steps are included: acquiring social media text data and urban planning GIS data, performing spatiotemporal alignment processing to obtain a structured dataset.

[0013] Social media text data is obtained through API interface, and a coordinate transformation module is used to unify the spatial benchmark of social media text data. Urban planning GIS data is obtained through WFS service via the municipal geographic information public service module, combined with time series data of building vibration sensors.

[0014] The dynamic time warping algorithm was used to perform spatiotemporal alignment of time-series data from building vibration sensors, social media text data, and urban planning GIS data to obtain a structured dataset.

[0015] As a preferred embodiment of the optimization method for improving urban physical examinations based on artificial intelligence as described in this invention, the method includes the following steps: analyzing structured datasets, calculating building stress distribution and road bearing capacity to obtain a physical health heatmap.

[0016] The vibration signals in the structured dataset are subjected to Butterworth bandpass filtering to obtain denoised acceleration time series data. The building mass is obtained from the material density and building volume in the urban planning GIS data. The dynamic load is calculated to obtain the filtered dynamic load time series data.

[0017] The building GIS outline is extracted from the urban planning GIS data. Based on the filtered dynamic load time series data and the building GIS outline, a three-dimensional finite element model is established in ANSYS. The load sequence in the filtered dynamic load time series data is applied to the floor nodes of the three-dimensional finite element model to calculate the building stress distribution.

[0018] Based on vibration data from urban planning GIS data, calculate the road deflection value;

[0019] The road deflection value and building stress distribution are overlaid and rendered in ArcGIS to obtain a physical health heat map.

[0020] As a preferred embodiment of the AI-based urban health checkup improvement scheme optimization method described in this invention, the method includes: using a BERT model to perform sentiment analysis on social media text data, combining geocoding technology to obtain a citizen sentiment density distribution map, constructing a physically constrained neural network model, embedding a physical health heatmap into the network loss function, and obtaining a three-dimensional structural health diagnosis report, comprising the following steps.

[0021] The BERT model was used to classify the sentiment of social media text data, resulting in labeled data with sentiment scores.

[0022] The urban area is extracted from the planning drawings, divided into grids and aligned with the building GIS outlines. The emotional density value of each grid is calculated to obtain the distribution map of citizens' emotional density.

[0023] Based on labeled data with emotion scores, along with emotion and physical health heatmaps, pixel-level spatial alignment is performed to obtain registered stress-emotion dual-channel images. A physical constraint neural network model is then constructed and trained using the PDE loss function.

[0024] The registered stress-emotion dual-channel images are input into a trained physical constraint neural network model to predict the global stress field and mark the high-emotion-high-stress coupling region, thus obtaining a three-dimensional structural health diagnosis report.

[0025] As a preferred embodiment of the AI-based urban health checkup improvement scheme optimization method described in this invention, the method involves: using the CityEngine planning module based on a three-dimensional structural health diagnosis report to generate a set of feasible renovation schemes, including the following steps:

[0026] Based on the three-dimensional structural health diagnosis report, voxel-level traversal screening is performed using dual threshold conditions, and the DBSCAN algorithm is used to merge adjacent high-risk areas to obtain a list of high-risk area boundary coordinates.

[0027] Based on the list of boundary coordinates of high-risk areas and building GIS outlines, CGA scripts are written in CityEngine to define dynamic reinforcement logic and form alternative solutions;

[0028] The cost, risk score, and mood improvement rate of the alternative solutions are calculated, and the optimal solution is obtained by using the NSGA-II algorithm to solve for the optimal solution.

[0029] Import the optimized solutions into Unity3D to simulate pedestrian flow and structural response, and generate a set of feasible modification solutions.

[0030] As a preferred embodiment of the AI-based urban health checkup improvement scheme optimization method described in this invention, the method includes the following steps: importing a set of feasible renovation schemes into the AnyLogic simulation module to simulate social benefits and obtain an evaluation report of the economic benefit prediction curve.

[0031] The building models in the feasible modification scheme set are converted into GLTF format, the road network is converted into GeoJSON format, a three-dimensional simulation environment is built in AnyLogic, the dynamic behavior logic of citizen intelligent agents and vehicle intelligent agents is set, and the initial simulation model is obtained.

[0032] Based on the distribution map of citizens' emotions, and configuring the behavioral rules of the intelligent agent;

[0033] The initial simulation model and agent behavior rules are loaded into the AnyLogic cloud computing module, Monte Carlo simulation is run, and data on pedestrian flow, commercial revenue, and public facility utilization rate are recorded to obtain raw simulation data. The commercial revenue growth rate and facility utilization Gini coefficient are calculated to obtain an evaluation report of the economic benefit prediction curve.

[0034] As a preferred embodiment of the AI-based urban health check improvement scheme optimization method described in this invention, the following steps are included: establishing a partial differential equation for emotion propagation, predicting the concentration of negative emotions based on a citizen emotion density distribution map, dynamically adjusting the public space configuration ratio of the feasible renovation scheme set, and obtaining the optimized scheme.

[0035] After establishing the reaction-diffusion partial differential equation based on the citizen emotion density map and solving it using the finite difference method, an emotion evolution model is obtained.

[0036] Using an emotion evolution model and social media text data, the equations are solved using the finite difference method to obtain the boundary of negative emotion hotspots;

[0037] Based on the area of ​​the hotspot within the boundary of negative emotion hotspots, the proportion of public space is dynamically increased to obtain a renovation plan;

[0038] The modification plan was imported into AnyLogic simulation to verify the improvement rate of mood and obtain the optimized plan.

[0039] Secondly, the present invention provides an optimization system for improving urban physical examination schemes based on artificial intelligence, including a data acquisition module that acquires social media text data and urban planning GIS data and performs spatiotemporal alignment processing to obtain a structured dataset;

[0040] The data analysis module analyzes structured datasets, calculates building stress distribution and road bearing capacity, and obtains physical health heat maps.

[0041] The health diagnosis module uses the BERT model to perform sentiment analysis on social media text data, combines geocoding technology to obtain a citizen sentiment density distribution map, constructs a physically constrained neural network model, embeds the physical health heat map into the network loss function, and obtains a three-dimensional structural health diagnosis report.

[0042] The renovation module uses the CityEngine planning module based on the 3D structural health diagnosis report to generate a set of feasible renovation plans.

[0043] The optimization module imports the set of feasible renovation schemes into the AnyLogic simulation module to simulate social benefits, obtains an evaluation report of the economic benefit prediction curve, establishes a partial differential equation for emotion propagation, predicts the concentration of negative emotions based on the distribution map of citizens' emotions, dynamically adjusts the public space configuration ratio of the set of feasible renovation schemes, and obtains the optimized scheme.

[0044] Thirdly, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, it implements any step of the optimization method for urban physical examination improvement scheme based on artificial intelligence as described in the first aspect of the present invention.

[0045] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the optimization method for improving urban physical examination schemes based on artificial intelligence as described in the first aspect of the present invention.

[0046] The beneficial effects of this invention are as follows: By constructing a physical constraint neural network model, this invention innovatively couples the distribution of building stress with the density of citizens' emotions. By using the PDE loss function, it achieves a multimodal deep integration of structural health and social emotion, which can accurately identify high stress-high emotion coupling areas. At the same time, based on the partial differential equation of emotion propagation, a dynamic feedback mechanism is established to transform the predicted value of negative emotion concentration into an optimized scheme for public space configuration, forming a closed loop from data perception to scheme optimization. Attached Figure Description

[0047] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a flowchart of an optimization method for an AI-based urban health checkup improvement program.

[0049] Figure 2A schematic diagram of an AI-based urban health checkup improvement system optimization scheme.

[0050] Figure 3 This is a schematic diagram of data acquisition and alignment.

[0051] Figure 4 This is a schematic diagram of a set of feasible modification schemes. Detailed Implementation

[0052] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0053] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0054] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0055] Reference Figures 1-4 This is one embodiment of the present invention, which provides an optimization method for improving urban physical examinations based on artificial intelligence, including the following steps:

[0056] S1. Obtain social media text data and urban planning GIS data, perform spatiotemporal alignment processing, and obtain a structured dataset.

[0057] S1.1. Obtain social media text data through API interface, use coordinate transformation module to unify spatial benchmark of social media text data, use WFS service to obtain urban planning GIS data through city-level geographic information public service module, and combine with time series data of building vibration sensor.

[0058] Furthermore, by calling the API interface provided by the social media platform to obtain social media text data containing geographic location information, a coordinate transformation module is used to uniformly transform the geographic coordinates in the social media text data to the WGS84 coordinate system; urban planning GIS data containing building outlines and road networks is obtained through the WFS service interface provided by the municipal geographic information public service platform, and acceleration time series data recorded by vibration sensors deployed at key nodes of buildings are collected simultaneously.

[0059] S1.2. Use the dynamic time warping algorithm to perform spatiotemporal alignment of building vibration sensor time series data, social media text data and urban planning GIS data to obtain a structured dataset.

[0060] Furthermore, the acceleration sampling time series of the building vibration sensor time series data is extracted, and the timestamps of the social media text data and the version update time of the urban planning GIS data are obtained. The dynamic time warping algorithm is used to calculate the time offset between the building vibration sensor time series data and the social media text data, and the time offset between the building vibration sensor time series data and the urban planning GIS data, respectively. The time reference of the social media text data and the urban planning GIS data is compensated and corrected according to the calculated time offset. At the same time, the physical coordinates of the building vibration sensor installation location, the geographic coordinates in the social media text data and the spatial reference system in the urban planning GIS data are uniformly converted to the WGS84 coordinate system to generate a structured dataset with time alignment and unified spatial reference.

[0061] S2. Analyze the structured dataset to calculate the building stress distribution and road bearing capacity, and obtain a physical health heat map.

[0062] S2.1. Perform Butterworth bandpass filtering on the vibration signals in the structured dataset to obtain denoised acceleration time series data. Obtain the building mass through the material density and building volume in the urban planning GIS data, calculate the dynamic load, and obtain the filtered dynamic load time series data.

[0063] Specifically, the expression is,

[0064] ;

[0065] in, For building structures in time Dynamic loads, For building quality, For the second derivative of the displacement, It is the square of the time derivative. For displacement, For time.

[0066] It should be noted that the raw vibration signals recorded by the time series data of building vibration sensors in the structured dataset are processed by frequency domain filtering using a Butterworth bandpass filter to filter out environmental noise below the building's natural frequency and high-frequency interference above the sampling frequency, retaining the effective frequency band signals reflecting the dynamic characteristics of the structure, and obtaining denoised acceleration time series data. Based on the building material density parameters and building volume information recorded in the urban planning GIS data, the building mass is calculated by multiplying the volume and density. Based on the denoised acceleration time series data and the building mass, the dynamic load is calculated according to Newton's second law, where the acceleration value is taken as the second derivative result of the acceleration time series data, and finally output dynamic load data containing time series.

[0067] S2.2 Extract the building GIS outline from the urban planning GIS data. Based on the filtered dynamic load time series data and the building GIS outline, establish a three-dimensional finite element model in ANSYS. Apply the load sequence in the filtered dynamic load time series data to the floor nodes of the three-dimensional finite element model and calculate the building stress distribution.

[0068] Specifically, the expression is,

[0069] ;

[0070] in, For the distribution of building stress, For building structure in Normal stress in the direction, For building structure in Normal stress in the direction, For building structure in direction and Shear stress in the directional plane.

[0071] S2.3. Calculate the road deflection value based on the vibration data in the urban planning GIS data.

[0072] Specifically, the expression is,

[0073] ;

[0074] in, This represents the road deflection value. For dynamic loads, The equivalent radius, For elastic modulus, It is Poisson's ratio.

[0075] S2.4 Overlay and render the road deflection value and building stress distribution in ArcGIS to obtain a physical health heat map.

[0076] Furthermore, a spatial layer containing road networks and building outlines is created in the ArcGIS platform. The calculated road deflection values ​​are spatially interpolated according to the road centerline position to generate a continuous road health raster surface. At the same time, the calculated building stress distribution results are spatially correlated with the building GIS outline to generate a building stress intensity surface layer. Multi-band rendering technology is used to spatially overlay the road health raster surface and the building stress intensity surface layer. The road deflection values ​​use a gradient from green to red to represent the health level, and the building stress distribution uses a gradient from blue to yellow to represent the stress level. By setting the transparency parameter, the two layers remain visible when displayed in an overlay. Finally, a physical health heatmap that reflects both the road bearing capacity and the building structural stress is generated.

[0077] S3. Use the BERT model to perform sentiment analysis on social media text data, combine geocoding technology to obtain a citizen sentiment density distribution map, construct a physical constraint neural network model, embed the physical health heat map into the network loss function, and obtain a three-dimensional structural health diagnosis report.

[0078] S3.1 Use the BERT model to classify the sentiment of social media text data to obtain labeled data with sentiment scores.

[0079] Furthermore, a pre-trained BERT model is used to perform sentiment classification on the social media text data contained in the structured dataset. Each piece of social media text data is input into the BERT model to obtain a text feature vector. Then, a fully connected layer maps the feature vector to the sentiment classification label space, outputting the sentiment category and confidence score corresponding to each piece of social media text data. The sentiment categories are divided into three types: positive, neutral, and negative, and the confidence score ranges from 0 to 1 to represent the sentiment intensity. The classification results are associated with the geographical location information and timestamps of the original social media text data to form labeled data containing sentiment labels, sentiment scores, geographical coordinates, and timestamps. The labeled data retains the same spatiotemporal reference as the structured dataset.

[0080] S3.2 Extract urban areas from the planning drawings, divide them into grids, align them with the building GIS outlines, calculate the emotional density value of each grid, and obtain a distribution map of citizens' emotional density.

[0081] Specifically, the expression is,

[0082] ;

[0083] in, Emotional density value For the first Sentiment analysis results of the text. For bandwidth, For the first The distance from the text to the target point. The total number of texts. For text indexing.

[0084] S3.3. Based on the labeled data with emotion scores, including emotion and physical health heatmaps, pixel-level spatial alignment is performed to obtain the registered stress-emotion dual-channel image. A physical constraint neural network model is constructed and trained using the PDE loss function.

[0085] Specifically, the expression is,

[0086] ;

[0087] in, The PDE loss function is... The three-dimensional stress field predicted by the neural network. The stress value is measured by a fiber optic sensor. For the Laplace operator of the stress field, The material stiffness coefficient, For the emotion weight matrix, For PDE constraint weights, The strength of the emotion weight constraint.

[0088] S3.4 Input the registered stress-emotion dual-channel image into the trained physical constraint neural network model to predict the global stress field and mark the high emotion-high stress coupling region to obtain a three-dimensional structural health diagnosis report.

[0089] Furthermore, the stress-emotion dual-channel images, which have been spatially aligned at the pixel level, are used as input data and fed into a pre-trained physical constraint neural network model for processing. The physical constraint neural network model first extracts stress distribution features and emotion density features through a convolutional neural network, and then uses a loss function calculation layer with PDE constraints to perform coupling analysis on the features, outputting a global three-dimensional stress field prediction result. At the same time, the coupling coefficient is calculated based on the product of the stress field prediction value and the emotion density value, and high emotion-high stress coupling regions with coupling coefficients exceeding a set threshold are automatically identified and marked, generating a three-dimensional structural health diagnosis report that includes a three-dimensional stress field cloud map, spatial coordinates of the coupling region, and coupling strength values.

[0090] S4. Based on the 3D structural health diagnosis report, use the CityEngine planning module to generate a set of feasible renovation schemes.

[0091] S4.1. Based on the three-dimensional structural health diagnosis report, set dual threshold conditions to perform voxel-level traversal screening, and use the DBSCAN algorithm to merge adjacent high-risk areas to obtain a list of high-risk area boundary coordinates.

[0092] Furthermore, based on the stress field values ​​and emotional coupling coefficients recorded in the 3D structural health diagnostic report, two screening conditions, stress threshold and emotional coupling threshold, are set respectively. Each voxel unit in the diagnostic report is traversed and detected. When the stress field value of a voxel unit exceeds the stress threshold and the emotional coupling coefficient exceeds the emotional coupling threshold, the voxel is marked as a potential high-risk unit. The spatial coordinates of all marked potential high-risk units are used as input, and the density-based DBSCAN spatial clustering algorithm is applied for processing. The DBSCAN algorithm identifies spatially densely connected unit clusters by calculating the Euclidean distance between units, and merges unit clusters with mutual distances less than the neighborhood radius and containing the fewest units into continuous high-risk regions. Finally, a list containing the 3D boundary coordinates of each high-risk region is output.

[0093] S4.2 Based on the list of boundary coordinates of high-risk areas and the building GIS outline, write CGA scripts in CityEngine to define dynamic reinforcement logic and form alternative solutions.

[0094] Furthermore, the list of boundary coordinates of high-risk areas and building GIS outline data are imported into the CityEngine 3D modeling environment. CGA rule scripts are written according to the building structure type and risk level. The CGA script first analyzes the spatial relationship between the boundary coordinates of high-risk areas and the building GIS outline, and defines differentiated reinforcement rules for areas with different risk levels: for frame structure buildings, the generation logic of adding shear walls or steel supports is set, and for masonry structure buildings, the generation rule of adding ring beams and structural columns is set; at the same time, the size parameters of the reinforcement components are adjusted according to the stress distribution gradient in the 3D structural health diagnosis report, and the reinforcement section is automatically increased in the stress concentration area of ​​the building; after executing the CGA script, a set of 3D models containing multiple reinforcement schemes is generated. Each model is accompanied by engineering parameters such as material usage and construction scope, forming a complete set of alternative schemes.

[0095] S4.3 Calculate the cost, risk score and emotion improvement rate of the alternative solutions, and use the NSGA-II algorithm to find the optimal solution to obtain the optimized solution.

[0096] Specifically, the expression is,

[0097] ;

[0098] in, To optimize the solution, For solid volume, This refers to the unit price of the materials.

[0099] S4.4. Import the optimized plan into Unity3D to simulate pedestrian flow and structural response, and generate a set of feasible modification plans.

[0100] Specifically, the expression is,

[0101] ;

[0102] in, This is a collection of feasible renovation plans. For the original acceleration, This refers to the acceleration after the modification.

[0103] S5. Import the set of feasible renovation schemes into the AnyLogic simulation module to simulate social benefits and obtain an evaluation report of the economic benefit prediction curve.

[0104] S5.1 Convert the building models in the feasible modification scheme set to GLTF format and the road network to GeoJSON format. Build a 3D simulation environment in AnyLogic, set the dynamic behavior logic of citizen agents and vehicle agents, and obtain the initial simulation model.

[0105] Furthermore, the 3D building models contained in the feasible renovation schemes are exported to GLTF format files using a format conversion tool, while the road network data is converted to GeoJSON format. A new 3D scene is created in the AnyLogic simulation platform, and the GLTF format building models and GeoJSON format road network data are imported to build the basic environment. The initial spatial distribution of citizen agents is set based on the citizen emotion density distribution map, and state machine logic including commuting, leisure, and shopping behaviors is configured for the citizen agents, where the behavior selection probability is related to the emotion score. A path planning algorithm and car-following model parameters based on the road network are defined for the vehicle agents. The interaction rules between citizen agents and vehicle agents are implemented through the agent programming interface provided by AnyLogic, including pedestrian crossing behavior and public transportation riding decisions, generating an initialization simulation model containing a 3D scene, agent behavior logic, and interaction rules.

[0106] S5.2. Based on the distribution map of citizens' emotions, configure the behavior rules of the intelligent agent.

[0107] Furthermore, in the AnyLogic simulation module, the spatial grid data of the citizen emotion density distribution map is matched with the initial position of the agent. Each citizen agent is assigned a corresponding emotion score attribute, and behavioral rules for the citizen agents are set based on the emotion scores. The percentile of the emotion density distribution in the high emotion-high stress coupling area marked in the 3D structural health diagnosis report determines the threshold for the emotion score. Citizen agents with emotion scores below the threshold increase the probability of going to parks and recreational areas, while those with emotion scores above the threshold increase the frequency of commuting route changes and the time spent in commercial facilities. At the same time, an emotion propagation mechanism between citizen agents is configured. When two citizen agents meet within a set distance, they interact with each other according to the parameters of the emotion propagation partial differential equation. Speed ​​adjustment rules based on road network status are set for vehicle agents. The emotion density threshold of the road is set based on the 85th percentile of the emotion value in the road network area of ​​the citizen emotion density distribution map. When the emotion density of the road exceeds the road's emotion density threshold, the driving speed is automatically reduced.

[0108] S5.3 Load the initialization simulation model and agent behavior rules into the AnyLogic cloud computing module, run the Monte Carlo simulation, record the data of pedestrian flow, commercial revenue, and public facility utilization rate, obtain the original simulation data, calculate the commercial revenue growth rate and facility utilization Gini coefficient, and obtain an evaluation report of the economic benefit prediction curve.

[0109] Specifically, the expression is,

[0110] ;

[0111] in, For business revenue growth rate, For the original income, For the benefits after the renovation.

[0112] ;

[0113] in, The Gini coefficient is used for facilities. For the number of facilities, For the first The number of users of each facility during the simulation period. For facility indexing.

[0114] It should be noted that the initial simulation model, which includes the 3D scene and the intelligent agent's behavioral logic, is uploaded to the AnyLogic cloud computing module. Monte Carlo simulation parameters are configured and multiple rounds of simulation are started. During the simulation, real-time data on pedestrian traffic in commercial areas, revenue changes in commercial facilities, and usage data of public facilities such as parks are recorded for each time period. After the simulation is completed, the simulation results of each round are summarized, the commercial revenue growth rate before and after the implementation of the renovation plan is calculated, and the facility usage Gini coefficient is calculated based on the public facility usage data. The data on the changes in commercial revenue growth rate and facility usage Gini coefficient over time are visualized to generate an evaluation report reflecting the dynamic changes in economic and social benefits.

[0115] S6. Establish a partial differential equation for emotion propagation, predict the concentration of negative emotions based on the distribution map of citizens' emotions, dynamically adjust the public space configuration ratio of the feasible renovation scheme set, and obtain the optimized scheme.

[0116] S6.1. Based on the citizen emotion density map, the reaction-diffusion partial differential equation is established and solved by the finite difference method to obtain the emotion evolution model.

[0117] Specifically, the expression is,

[0118] ;

[0119] in, For the differential change in emotional concentration, For the differential change over time, Here, is the diffusion coefficient, and is the social sentiment weight. For social media public opinion, Emotional intensity.

[0120] S6.2 Using the emotion evolution model and social media text data, the equation is solved by the finite difference method to obtain the boundary of negative emotion hotspots.

[0121] Furthermore, the emotion evolution model is matched and calibrated with the emotion classification results of social media text data. The emotion evolution equation is solved discretically using the explicit finite difference method on the urban spatial grid, where the spatial derivative is discretized using a five-point difference scheme and the temporal derivative is discretized using a forward Euler scheme. The sentiment analysis results of social media text data are used as the emotion source term input into the finite difference method to solve the equation. The emotion concentration distribution for multiple future time periods is obtained through iterative calculation. The spatial distribution of emotion concentration is extracted using contour lines to identify the boundaries of continuous regions where the emotion concentration exceeds a set threshold. The output includes negative emotion hotspot boundary data containing the geometric boundaries of hotspot regions and peak emotion values.

[0122] S6.3. Based on the area of ​​the hotspot within the boundary of negative emotion hotspots, dynamically increase the proportion of public space to obtain the renovation plan.

[0123] Specifically, the expression is,

[0124] ;

[0125] in, To adjust the proportion of public space, This represents the original proportion of public space. To control the intensity of the impact of emotional hotspots on the configuration of public spaces, The total area of ​​the renovation zone, The area of ​​negative sentiment hotspots.

[0126] S6.4. Import the modification plan into AnyLogic simulation to verify the improvement rate of the mood and obtain the optimized plan.

[0127] Specifically, the expression is,

[0128] ;

[0129] in, For the rate of improvement in mood, To determine the concentration of negative emotions before the transformation, The concentration of negative emotions after modification.

[0130] This embodiment also provides an AI-based urban health check improvement scheme optimization system, including: a data acquisition module, which acquires social media text data and urban planning GIS data and performs spatiotemporal alignment processing to obtain a structured dataset;

[0131] The data analysis module analyzes structured datasets, calculates building stress distribution and road bearing capacity, and obtains physical health heat maps.

[0132] The health diagnosis module uses the BERT model to perform sentiment analysis on social media text data, combines geocoding technology to obtain a citizen sentiment density distribution map, constructs a physically constrained neural network model, embeds the physical health heat map into the network loss function, and obtains a three-dimensional structural health diagnosis report.

[0133] The renovation module uses the CityEngine planning module based on the 3D structural health diagnosis report to generate a set of feasible renovation plans.

[0134] The optimization module imports the set of feasible renovation schemes into the AnyLogic simulation module to simulate social benefits, obtains an evaluation report of the economic benefit prediction curve, establishes a partial differential equation for emotion propagation, predicts the concentration of negative emotions based on the distribution map of citizens' emotions, dynamically adjusts the public space configuration ratio of the set of feasible renovation schemes, and obtains the optimized scheme.

[0135] This embodiment also provides a computer device applicable to the optimization method of urban physical examination improvement scheme based on artificial intelligence, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the optimization method of urban physical examination improvement scheme based on artificial intelligence as proposed in the above embodiment.

[0136] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0137] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the optimization method for improving urban physical examinations based on artificial intelligence as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0138] In summary, this invention innovatively couples the distribution of building stress with the density of citizens' emotions by constructing a physical constraint neural network model. By using the PDE loss function, it achieves a multimodal deep integration of structural health and social emotion, accurately identifying high stress-high emotion coupling areas. At the same time, it establishes a dynamic feedback mechanism based on the partial differential equation of emotion propagation, transforming the predicted value of negative emotion concentration into an optimized scheme for public space configuration, thus forming a closed loop from data perception to scheme optimization.

[0139] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. An optimization method for urban physical examination improvement schemes based on artificial intelligence, characterized in that: include, Social media text data and urban planning GIS data are acquired and spatiotemporally aligned to obtain a structured dataset; The structured dataset was analyzed to calculate the building stress distribution and road bearing capacity, resulting in a physical health heat map. The BERT model is used to perform sentiment analysis on social media text data. Combined with geocoding technology, a citizen sentiment density distribution map is obtained. A physically constrained neural network model is constructed, and a physical health heat map is embedded into the network loss function to obtain a three-dimensional structural health diagnosis report. Based on the 3D structural health diagnosis report, the CityEngine planning module is used to generate a set of feasible renovation plans. The set of feasible renovation schemes is imported into the AnyLogic simulation module to simulate social benefits, and an evaluation report of the economic benefit prediction curve is obtained. A partial differential equation for emotion propagation is established, and the concentration of negative emotions is predicted based on the distribution map of citizens' emotions. The public space configuration ratio of the set of feasible renovation schemes is dynamically adjusted to obtain the optimized scheme. Sentiment analysis of social media text data was performed using the BERT model, and a citizen sentiment density distribution map was obtained by combining geocoding technology. A physically constrained neural network model was constructed, and a physical health heatmap was embedded into the network loss function to obtain a three-dimensional structural health diagnostic report. The process included the following steps. The BERT model was used to classify the sentiment of social media text data, resulting in labeled data with sentiment scores. The urban area is extracted from the planning drawings, divided into grids and aligned with the building GIS outlines. The emotional density value of each grid is calculated to obtain the distribution map of citizens' emotional density. Based on labeled data with emotion scores, along with emotion and physical health heatmaps, pixel-level spatial alignment is performed to obtain registered stress-emotion dual-channel images. A physical constraint neural network model is then constructed and trained using the PDE loss function. The registered stress-emotion dual-channel images are input into a trained physical constraint neural network model to predict the global stress field and mark the high emotion-high stress coupling region, thus obtaining a three-dimensional structural health diagnosis report. Based on the 3D structural health diagnostic report, the CityEngine planning module is used to generate a set of feasible renovation plans, including the following steps. Based on the three-dimensional structural health diagnosis report, voxel-level traversal screening is performed using dual threshold conditions, and the DBSCAN algorithm is used to merge adjacent high-risk areas to obtain a list of high-risk area boundary coordinates. Based on the list of boundary coordinates of high-risk areas and building GIS outlines, CGA scripts are written in CityEngine to define dynamic reinforcement logic and form alternative solutions; The cost, risk score, and mood improvement rate of the alternative solutions are calculated, and the optimal solution is obtained by using the NSGA-II algorithm to solve for the optimal solution. Import the optimized solutions into Unity3D to simulate pedestrian flow and structural response, and generate a set of feasible modification solutions.

2. The optimization method for improving urban physical examinations based on artificial intelligence as described in claim 1, characterized in that: The process of acquiring social media text data and urban planning GIS data, performing spatiotemporal alignment processing, and obtaining a structured dataset includes the following steps: Social media text data is obtained through API interface, and a coordinate transformation module is used to unify the spatial benchmark of social media text data. Urban planning GIS data is obtained through WFS service via the municipal geographic information public service module, combined with time series data of building vibration sensors. The dynamic time warping algorithm was used to perform spatiotemporal alignment of time-series data from building vibration sensors, social media text data, and urban planning GIS data to obtain a structured dataset.

3. The optimization method for improving urban physical examinations based on artificial intelligence as described in claim 2, characterized in that: The analysis of structured datasets, calculation of building stress distribution and road bearing capacity, and the resulting physical health heatmaps include the following steps. The vibration signals in the structured dataset are subjected to Butterworth bandpass filtering to obtain denoised acceleration time series data. The building mass is obtained from the material density and building volume in the urban planning GIS data. The dynamic load is calculated to obtain the filtered dynamic load time series data. The building GIS outline is extracted from the urban planning GIS data. Based on the filtered dynamic load time series data and the building GIS outline, a three-dimensional finite element model is established in ANSYS. The load sequence in the filtered dynamic load time series data is applied to the floor nodes of the three-dimensional finite element model to calculate the building stress distribution. Based on vibration data from urban planning GIS data, calculate the road deflection value; The road deflection value and building stress distribution are overlaid and rendered in ArcGIS to obtain a physical health heat map.

4. The optimization method for the urban physical examination improvement scheme based on artificial intelligence as described in claim 3, characterized in that: The set of feasible renovation schemes is imported into the AnyLogic simulation module to simulate social benefits and obtain an evaluation report of the economic benefit prediction curve. Includes the following steps, The building models in the feasible modification scheme set are converted into GLTF format, the road network is converted into GeoJSON format, a three-dimensional simulation environment is built in AnyLogic, the dynamic behavior logic of citizen intelligent agents and vehicle intelligent agents is set, and the initial simulation model is obtained. Based on the distribution map of citizens' emotions, and configuring the behavioral rules of the intelligent agent; The initial simulation model and agent behavior rules are loaded into the AnyLogic cloud computing module, Monte Carlo simulation is run, and data on pedestrian flow, commercial revenue, and public facility utilization rate are recorded to obtain raw simulation data. The commercial revenue growth rate and facility utilization Gini coefficient are calculated to obtain an evaluation report of the economic benefit prediction curve.

5. The optimization method for the urban physical examination improvement scheme based on artificial intelligence as described in claim 4, characterized in that: A partial differential equation for emotion propagation is established. Based on the distribution map of citizens' emotions, the concentration of negative emotions is predicted. The proportion of public space configuration in the set of feasible renovation schemes is dynamically adjusted to obtain an optimized scheme. This includes the following steps: After establishing the reaction-diffusion partial differential equation based on the citizen emotion density map and solving it using the finite difference method, an emotion evolution model is obtained. Using an emotion evolution model and social media text data, the equations are solved using the finite difference method to obtain the boundary of negative emotion hotspots; Based on the area of ​​the hotspot within the boundary of negative emotion hotspots, the proportion of public space is dynamically increased to obtain a renovation plan; The modification plan was imported into AnyLogic simulation to verify the improvement rate of mood and obtain the optimized plan.

6. An AI-based urban physical examination improvement scheme optimization system, based on the AI-based urban physical examination improvement scheme optimization method according to any one of claims 1 to 5, characterized in that: include, The data acquisition module acquires social media text data and urban planning GIS data, performs spatiotemporal alignment processing, and obtains a structured dataset. The data analysis module analyzes structured datasets, calculates building stress distribution and road bearing capacity, and obtains physical health heat maps. The health diagnosis module uses the BERT model to perform sentiment analysis on social media text data, combines geocoding technology to obtain a citizen sentiment density distribution map, constructs a physically constrained neural network model, embeds the physical health heat map into the network loss function, and obtains a three-dimensional structural health diagnosis report. The renovation module uses the CityEngine planning module based on the 3D structural health diagnosis report to generate a set of feasible renovation plans. The optimization module imports the set of feasible renovation schemes into the AnyLogic simulation module to simulate social benefits, obtains an evaluation report of the economic benefit prediction curve, establishes a partial differential equation for emotion propagation, predicts the concentration of negative emotions based on the distribution map of citizens' emotions, dynamically adjusts the public space configuration ratio of the set of feasible renovation schemes, and obtains the optimized scheme.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the optimization method for improving urban physical examination based on artificial intelligence as described in any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the optimization method for improving urban physical examination based on artificial intelligence as described in any one of claims 1 to 5.

Images