Holographic display system and method for household condition gridding six-dimensional elements based on AI

Abstract

The invention discloses an AI-based household condition gridding six-dimensional element holographic display system and method, and relates to the technical field of gridding social governance information, and the system comprises a data aggregation module which integrates data across departments and carries out cleaning processing; the data modeling module is used for constructing a space-time model based on a GIS (Geographic Information System) and integrating a multi-source map, and automatically generating and realizing dynamic division of various types of grids for social governance based on the multi-source map; the social governance module is used for collecting data through an AR mobile terminal, processing the data through a clustering algorithm and arranging events according to priorities; the user condition grid module is used for dividing grids, associating six-dimensional data and dynamically associating with workers at all levels; the six-dimensional element module is used for carrying out multi-dimensional statistics on six elements; the holographic display module is used for realizing grid drilling and information visualization; and the intelligent AI module analyzes the data and generates early warning. According to the invention, intelligent analysis and dynamic display of grid information are realized through AI and holographic technologies, and an AI intelligent scheme and an efficient digital scheme are provided for social governance.

Description

Technical Field

[0001] The present invention relates to the field of gridded social governance information technology, and in particular to an AI-based household status gridded six-dimensional element holographic display system and method. Background Art

[0002] With the acceleration of urbanization, the continuous expansion of communities, and the surge in population within jurisdictions, the amount of information involved in social governance is growing exponentially. Staff members face multiple challenges, including difficulty in data collection, complex processing procedures, and inefficient management. In reality, communities are generally plagued by a contradiction between insufficient staffing and a multitude of tasks. Furthermore, they lack efficient information-based and intelligent support methods, making it difficult to achieve refined management and services for the "people, places, events, emotions, and organizations" within their jurisdictions. Traditional management models are no longer adapted to the real needs of modern social governance.

[0003] Existing social governance technologies suffer from significant flaws. Each department's business systems operate independently, with data stored on separate platforms. Staff must repeatedly search across multiple systems, only obtaining partial information and failing to fully display comprehensive grid information on a single interface. This "information silo" phenomenon prevents data from being shared, severely wastes resources, and hinders the development of collaborative governance, significantly hindering improvements in social governance efficiency.

[0004] While some systems have attempted to incorporate digital management concepts, most remain at the level of simple data entry and display, lacking the deep integration and intelligent analysis of multi-source data. This is particularly true for grid-based household management, which cannot achieve the dynamic tracking requirements of "understanding the plot through images, understanding the household status through signs, understanding the people through houses, and understanding the information through people." This makes it difficult to meet the objective requirement of shifting social governance from passive response to proactive prevention. A holographic display system that integrates multi-dimensional data and possesses intelligent analysis capabilities is urgently needed. Summary of the Invention

[0005] The present invention proposes an AI-based household grid-based six-dimensional element holographic display system and method to solve the problems mentioned in the above-mentioned prior art.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions: An AI-based six-dimensional gridded holographic display system for household conditions, including: The data aggregation module integrates multi-source social governance data across departments and systems, and relies on AI algorithms and blockchain technology to achieve unified access and intelligent association of global data; The data modeling module builds an intelligent spatiotemporal data model based on the GIS platform, automatically generates and dynamically divides various types of social governance grids through multi-source maps, and uses a data space computing engine to display city-level real-scene models; The social governance module, based on AI technology, seamlessly integrates social governance information with household status grids, allowing all departments to have a comprehensive overview of six-dimensional social governance information. Combined with a spatiotemporal deduction model, it predicts event trends and enables real-time perception and analysis of social governance trends. Household Grid Module: This module dynamically sets organizational grids, responsibility grids, communities, and buildings, integrates building unit household maps with the grid, and displays each department's data resources and service resources in a six-dimensional manner at each grid level, establishing an intelligent grid division model and a model for dynamically optimizing grid boundaries. The six-dimensional element module builds a six-dimensional element structured matrix of people, places, events, objects, emotions, and organizations, and achieves multi-dimensional penetrating visualization through element decoupling and dynamic association engine; The holographic display module implements a top-down, layer-by-layer, holographic three-dimensional display in the order of institutional grid, responsibility grid, community grid, building, house, and person. It integrates the six-dimensional elements of the grid and 470 demographic and social affairs attributes in multiple dimensions, presenting them at each level through colors, lines, icons, animations, charts, and labels. It also generates a building matrix skeleton and overlays multi-source grid data. The intelligent AI module uses AI technology to establish intelligent associations between six-dimensional elements, multi-dimensional analysis, trend prediction, and natural language interaction.

[0007] Furthermore, in the data aggregation module, the missing values are filled with the mean formula With median filling, outliers are identified and corrected using the 3σ principle, where is the mean, n is the number of samples, is the i-th sample value; data cleaning also includes: using blockchain technology to record data sources and processing processes, introducing intelligent anomaly detection algorithms, real-time monitoring of various types of data, using generative adversarial networks to fill missing values, and generating fill-in data through the game between generators and discriminators. At the same time, applying federated learning technology, while protecting data privacy, jointly cleaning and preprocessing data from multiple parties, and using automated data verification technology to verify the cleaned data.

[0008] Furthermore, in the data modeling module, the unified WGS84 coordinate system is used, and the chi-square test formula is used. Screening the six-dimensional data correlation characteristics of people, places, things, emotions, and organizations, is the chi-square statistic, is the actual observed value, is the theoretical expected value; adaptive coordinate transformation algorithm is used when integrating multi-source map services in, It represents the coordinate value after adaptive coordinate transformation, T(P) is the result of transforming the original coordinate P based on the basic transformation matrix T, κ=0.001 is the distance correction coefficient, d is the current point distance, and D is the reference distance. The GARCH-LSTM model is used to predict the frequency of caring people’s visits.

[0009] Furthermore, in the household grid module, when the load factor When it exceeds 1.2, the genetic algorithm optimized grid re-division is triggered. The six-dimensional elements of people, places, things, objects, emotions and organizations within the grid. is the global mean, is the global mean of the six-dimensional elements of people, places, things, objects, emotions, and organizations within the grid. is the grid G i The load factor, It represents the six-dimensional elements of people, places, events, objects, emotions, and organizations within the grid. Grid division also incorporates spatial syntax features as density calculation factors, enabling millisecond-level queries of tens of millions of entity relationships through a distributed graph database, supporting concurrent access by 500 grid members online at the same time.

[0010] Furthermore, in the intelligent AI module, text classification uses the BERT-wwm model that integrates domain knowledge and uses knowledge graph technology to sort out household-related domain knowledge and integrate it into the model. It introduces real-time event monitoring technology and combines it with IoT sensor data to quickly identify and trigger early warnings through multimodal information fusion analysis of emergency events.

[0011] Furthermore, the method of the AI-based household grid six-dimensional element holographic display system includes the following steps: Data processing and database construction steps: cross-departmental interfaces connect to information resources and service resources of various departments, clean data through ETL tools, use mean and median to fill missing values, and build a six-dimensional factor database; Spatiotemporal model and feature processing steps: Build a spatiotemporal model based on the GIS platform, integrate multi-source map services, use chi-square test and information gain to screen features, and generate event heat maps; Dynamic meshing steps: Dynamic meshing is performed using an improved Delaunay algorithm, with the coefficient of variation calculated to assess balance. Re-meshing occurs when the load factor exceeds 1.2. An intelligent meshing model is established by analyzing key data such as organizational grids, responsibility grids, community grids, buildings, and houses, as well as population density, building density, and event density. The modeling and meshing boundaries of the mesh are dynamically optimized. Geographic location is also used to automatically associate the six-dimensional elements of "people, places, events, objects, emotions, and organizations" with the corresponding meshes. Holographic display steps: A top-down, layer-by-layer, holographic three-dimensional display is achieved in the order of institutional grid, responsibility grid, community grid, building, house, and person. The six-dimensional elements of the grid and 470 demographic and social affairs attributes are multi-dimensionally integrated and presented intuitively at each level through colors, lines, icons, animations, charts, and labels. Multi-source grid data is superimposed by generating a building matrix skeleton. Risk warning steps: Use LSTM-Transformer to analyze events, calculate risks using situational awareness formulas, and automatically push warning events with response plans.

[0012] Furthermore, in the dynamic grid division step, the real-time updated population flow heat map and POI data are combined to identify high-density areas through the improved DBSCAN algorithm.

[0013] Furthermore, in the risk warning step, situational awareness calculations are triggered for abnormal activities of special groups, and warnings are generated in combination with risk disposal formulas. Disposal plans are simultaneously pushed to the grid worker's mobile terminal, and the average response time is shortened to less than 2 hours.

[0014] Compared with the existing technology, the beneficial effects of the present invention are: The system integrates data resources from multiple departments, realizes the full collection and dynamic update of the six-dimensional elements of "people, places, things, emotions, and organizations", forms a unified data model, completely solves the problem of information islands, and makes governance data more comprehensive, timely, and accurate.

[0015] Leveraging AI technology and holographic visualization, the system enables three-dimensional display and intelligent analysis of grid information. This system effectively integrates and tightly integrates dynamic grids, building and household color maps, and six-dimensional elements to achieve a dynamic, visual, and coordinated display of data information. Furthermore, through dynamic grid division, event heat map presentation, and risk warning capabilities, it accurately identifies key governance areas and potential issues, proactively identifies hidden dangers, and initiates actionable responses, transforming reactive response into proactive prevention, significantly improving problem resolution efficiency and governance effectiveness.

[0016] The system innovatively adopts a user status grid management model, which enables governance personnel to quickly grasp the details of their jurisdiction through intuitive icon identification and penetrating information display. At the same time, the system supports cross-departmental collaboration, unifies data platforms and business processes, improves the integration and utilization of governance resources, reduces management costs, and provides efficient and intelligent digital solutions for modern social governance. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a schematic block diagram of an AI-based household grid-based six-dimensional element holographic display system proposed by the present invention; Figure 2This is a schematic diagram of an AI-based six-dimensional gridded holographic display method for household conditions proposed by the present invention; Figure 3 This is a bar chart comparing the six-dimensional factor data integrity rates of the traditional system and this system; Figure 4 A line chart comparing the grid anomaly recognition time of the traditional system and this system under different abnormal events; Figure 5 A radar comparison chart of care coverage for special groups; Figure 6 This is a grid diagram of household conditions; Figure 7 This is an electronic household map. DETAILED DESCRIPTION

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0019] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like to indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0020] In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the said features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined. In addition, the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be a connection between the two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. The present invention will be further described in detail below with reference to the accompanying drawings.

[0021] Reference Figures 1 to 7 : An AI-based six-dimensional gridded holographic display system for household conditions, including the following modules: Data aggregation module: Through cross-departmental standardized interfaces, cross-platform data governance robots (real-time cross-departmental data exchange) and API calls (such as calling the civil affairs department's population data API), real-time docking is established with relevant departments' population management systems, civil affairs social assistance systems, housing and construction real estate registration systems, etc., and the incremental data synchronization mechanism is triggered at 2:00 a.m. every day to integrate structured data such as household registration data of relevant departments, civil affairs subsistence allowance information, housing and construction property rights, and unstructured data such as grid workers' patrol texts.

[0022] For unstructured data, the BERT-wwm model in NLP technology is used to extract keywords. For example, from "2025-07-10 09:30 Grid 7 Elderly man living alone, Mr. Li's water pipe broke", entity labels such as "elderly man living alone", "water pipe broke", "Grid 7" are extracted and mapped to the "event-emergency" and "person-special population" dimensions.

[0023] In terms of unified data format processing, the date field "July 1, 2025" in the relevant department system is converted to "2025-07-01", and the "building area: one hundred and twenty square meters" in the housing and construction system is converted to the value "120.0㎡". The format legitimacy of key data such as ID number and mobile phone number is verified through regular expressions. Abnormal data is automatically marked in red and pushed to the data administrator for review.

[0024] Use ETL tools to unify the date format to "YYYY-MM-DD" and convert string numbers to numeric types. For missing values, use the mean filling formula (in represents the mean, n is the number of samples, The formula calculates the mean (the sum of all sample values divided by the number of samples) and fills missing values with the median (the middle value is used for odd-numbered samples, and the mean of the two middle values is used for even-numbered samples). Outliers are identified and corrected using the 3σ principle (data with a deviation exceeding three standard deviations from the mean is marked as outliers. The standard deviation measures the degree of dispersion in the data and can effectively identify data that deviates from the normal range).

[0025] Blockchain technology is also used to record data sources and processing processes, ensuring data authenticity and traceability. An intelligent anomaly detection algorithm is introduced to monitor various data types, including IoT card traffic, in real time, accurately identifying abnormal periods. A generative adversarial network (GAN) is used to fill missing values. Through a game between a generator and a discriminator, appropriate filler data is generated. Federated learning techniques are applied to clean and preprocess data from multiple sources while protecting data privacy, further improving data quality. Automated data verification technology is used to perform multiple rounds of verification on the cleaned data to ensure accuracy and consistency, ultimately building a distributed database encompassing "people, places, events, objects, emotions, and organizations."

[0026] The data aggregation module, through a cross-departmental and cross-system data integration mechanism, unifies and centrally manages social governance data scattered across various sources, including departments, systems, community organizations, and public services. This module utilizes standardized interfaces, cross-platform data governance robots, API calls, ETL data cleansing, and distributed storage to aggregate multi-source data covering six dimensions: people, places, events, objects, emotions, and organizations. By establishing data sharing and exchange, the full collection and dynamic updating of these six dimensions is achieved, ultimately realizing an intelligent analysis platform for social governance that provides "clear information, a clear baseline, early detection, early feedback, and early resolution." This platform provides data support for precise policy implementation, intelligent early warning, and scientific decision-making, comprehensively enhancing the coordination, precision, and timeliness of social governance.

[0027] Data modeling module: First, the spatiotemporal data model is constructed. Based on the GIS platform, a 3D urban geographic spatial model is constructed, and the geometric and semantic attributes of entities such as buildings, roads, and green spaces are defined according to the CityGML standard. At the same time, multi-source map services such as Baidu Maps (using Web Mercator projection) and Amap (using GCJ-02 coordinate system) are integrated, and adaptive coordinate conversion algorithms are used to convert the data into a 3D urban geographic spatial model. Achieve unified coordinate transformation, where T is the basic transformation matrix, which determines the basic rules and transformation relationship of coordinate transformation; κ = 0.001 is the distance correction coefficient, which is used to further correct and adjust the coordinate transformation result based on the relationship between the current point distance d and the reference distance D; d is the distance from the current point to the reference point; D = 10,000 meters is the reference distance, to ensure that the coordinate deviation of the two-dimensional and three-dimensional scenes is less than 0.5 meters.

[0028] The data modeling module builds a social governance thematic data modeling system based on the GIS platform, and integrates multi-dimensional social governance data such as "people, places, things, emotions, and organizations" in accordance with the grid management model and household status map color display format for linkage holographic display, and establishes a spatiotemporal data model with a real-time update mechanism. At the same time, it uses spatial linkage technology to achieve synchronous interaction of two- and three-dimensional scenes, and integrates multi-source online map services such as Baidu Maps, Amap, Tencent Maps, Tiandi Maps and Google Maps, supports dynamic switching and mixed loading on the server side, and realizes multi-scale visualization analysis in a unified coordinate system; by constructing thematic analysis models such as event heat maps and resource distribution models, combined with machine learning algorithms to carry out trend prediction and resource allocation optimization, it provides dynamic monitoring and scientific decision-making support for social governance scenarios such as household status gridding and six-dimensional element holographic display, and ultimately realizes real-time perception and accurate judgment of social governance trends.

[0029] Spatial linkage technology is used to achieve interactive synchronization. When you click "Building 3 of XX Community" on the 2D map, the 3D model automatically rotates to the building's perspective, highlighting the unit entrance and popping up a property window (including the building area, number of households, occupancy rate, etc.). At the same time, the LSTM-Attention model is triggered. This model is trained based on historical data and is used to predict the number of maintenance events for the building in the next week.

[0030] In terms of feature screening and thematic model construction, the chi-square test formula is used. Analyze feature correlation, where O ij represents the actual observed value, E ij Indicates the theoretical expected value. For example, when analyzing the correlation between "caring for and paying attention to the people" and "frequency of using community services", 500 sample data are collected and calculated. (the degree of freedom is 1, the critical value is 3.84), and the two are determined to be significantly correlated, and then the planning weight of community service points is increased in the resource distribution model to care for and pay attention to the dense grid of the population.

[0031] Event heat map uses spatiotemporal decay model , where w i is the event weight, set according to the degree of urgency, such as fire w i =5. Neighborhood disputes i =2;d i is the spatial distance in meters; t is the current time, t i is the time of event occurrence; σ = 24 hours is the time decay coefficient. For example, a fire occurs in grid 4 at 10:00 on 2025-07-05 (w i =5), as of 10:00 on July 6, the impact value of this event on the heat map is The thermal value 500 meters away from the fire point is 0.0138 (displayed in orange).

[0032] The Household Grid Module: This module intelligently and dynamically sets organizational grids, responsibility grids, residential communities, and building blocks, seamlessly integrating building-unit household maps with the grids, and displaying departmental data and service resources in a six-dimensional fashion at each grid level. It establishes an intelligent grid division model and a dynamic grid boundary optimization model. By analyzing key data such as population density, building density, and event density, it automatically associates the six-dimensional elements of "people, places, events, objects, emotions, and organizations" with the corresponding grids using geographic location. For dynamic grid division and load balancing, the module utilizes an improved Delaunay triangulation algorithm, combined with real-time population flow heat maps generated with a one-hour sampling interval using operator signaling data, and POI data for 15 categories of points of interest, including supermarkets, schools, and hospitals. The DBSCAN algorithm parameters are set to a neighborhood radius of 500 meters and a minimum sample size of 30 people. Triangulation refinement is triggered when the population density in an area exceeds 1,000 people per square kilometer. For example, due to the construction of a new primary school, the number of people picking up and dropping off passengers in grid 8 increased sharply. Therefore, the algorithm was used to split the original grid into grid 8a (teaching area) and grid 8b (residential area), and the load factor was reduced from 1.5 to 0.98.

[0033] Load balancing is achieved through the formula Implementation, where N i Represents the grid population, =1,476 people is the average population of the entire district, which is used to measure the relative relationship between the grid population and the average population of the entire district; L i is the total length of roads within the grid, reflecting the road resource situation within the grid; max(L i ) and min(L i ) are the maximum and minimum road lengths in the region, is the average length of roads in the region. These parameters are used together to calculate the impact of road length differences on load. When the genetic algorithm optimization is triggered. For example, grid 11 (N i =2,000, L i =5 km, Load=1.3), by adjusting the boundary and moving 200 people to the adjacent grid, the load is reduced to 1.1.

[0034] In the six-dimensional element association modeling, the knowledge graph is used to construct the "people-place-event" association network. With "resident Zhang San" as the central node, it is associated with entities such as "House A (address: Grid 5-2 Building 301)", "Vehicle B (license plate number Anhui A12345)", and "Medical Insurance Reimbursement Event C (2025-06-15 Community Hospital)". The edge weights are calculated using cosine similarity. For example, the residential relationship weight between "Zhang San" and "House A" is 0.98. At the same time, the spatiotemporal graph neural network is used to update the formula ,in represents the hidden state of node i at time t; σ is the activation function, which is used to introduce nonlinear factors; W1 and W2 are weight matrices that control information transmission and conversion; N(i) is the set of neighbor nodes of node i; is the hidden state of neighbor node j at time t−1; a ij The attention weight is calculated based on the time interval between events, with a weight coefficient of 1.5 for events in the past 7 days, highlighting the importance of recent events in model updates.

[0035] Holographic Display Module: This module provides a top-down, layer-by-layer, holographic display of the organizational grid, responsibility grid, unit grid, community, building, house, and individual. It integrates the six-dimensional elements of the grid and 470 demographic and social attributes in a multi-dimensional manner, presenting each layer visually through color, lines, icons, animations, charts, and labels. By generating a building matrix skeleton and overlaying multi-source grid data, it achieves a clear understanding of the situation and the bottom line. GIS maps enable four levels of information drilling down from the city to the community level. On the GIS map, it supports drilling down from the provincial and city levels to the district, street (township), community grid, and responsible person grid. On the community-level interface, clicking on a specific responsible person grid, such as grid 6, will expand the distribution of the 18 buildings within that grid. Clicking on Building 5 will display floor plans for each floor. Clicking on a specific room, such as Room 302, will bring up a "one household, one file" pop-up, including the household head, Li Si (65 years old), caring group, focus group, house area of 90 square meters, abnormal water and electricity bills in July 2025 (automatically marked in red by the system), and the associated most recent grid worker visit record.

[0036] For special groups, binary coding is expanded to multi-dimensional identification. For example, a family has both "caring people" (b1=1) and "focused people" (b2=1), then the code B=1×2 0 +1×2 1 =3, displayed as a purple icon on the map, and detailed labels can be displayed by hovering the mouse.

[0037] Innovatively incorporating augmented reality (AR) technology, when users view property information using AR devices, the device presents a three-dimensional model of the property and its surroundings, based on real-world scenarios, enhancing the user's perception of the space. Blockchain technology is also incorporated to ensure data security and immutability during display, ensuring the authenticity and reliability of information such as "one household, one file." Utilizing virtual reality (VR) technology, users can immerse themselves in the community and freely browse through property information, creating an immersive experience.

[0038] In addition, adaptive visualization technology is used to automatically optimize the display effect based on the screen size and resolution of the user's device. Regarding housing status identification, the color spectrum is adaptively adjusted through the HSV color space. Originally, rental housing is identified as red (H=0, S=100%, V=80%). When the proportion of rental housing in a grid exceeds 30%, the saturation is automatically reduced to 60% (to avoid visual overload), and the display is displayed as light red (H=0, S=60%, V=80%). The data dashboard integrates time and space dimensions. The left side displays six-dimensional element statistics for the entire area. The upper right portion is an event heat map, and the lower portion is a time series trend. Clicking on a high-risk area in the heat map drills down to a list of specific events to view the progress of each event. For example, "Grid 3 garbage accumulation" has been assigned to the sanitation department and is expected to be resolved within 2 hours. Drilling down into the time and space dimensions supports analyzing event trends over the past 7 days.

[0039] The intelligent AI module: Based on six-dimensional grid feature data, it leverages natural language processing (NLP) technology to automatically discover and establish explicit and implicit connections between features, constructing maps of population social relationships and event spatiotemporal relationships. It identifies unusual patterns in these six-dimensional features (e.g., sudden crowd gatherings in a specific area and a surge in negative information) and predicts future trends (e.g., public welfare-related incidents, facility failure risks, event probability, and information trends). Users can use natural language to ask questions (e.g., "Show me the facilities and handling of the elderly fall incident in grid A"), and the AI interprets and presents the results. Leveraging AI algorithms and blockchain technology, it integrates multi-source social governance data across departments and systems, achieving unified access, governance, and intelligent correlation of global data. A comprehensive suite of advanced technologies is employed to build an intelligent analysis system. Text classification utilizes the BERT-wwm model, which incorporates domain knowledge embedding. It first extracts entities from 12 systems, including household registration, civil affairs assistance, and housing and construction, to construct a household knowledge graph containing 50,000 triples. Relationships in the graph are incorporated into the BERT-wwm word embedding layer to enhance text semantic understanding. Using semi-supervised learning, the model was initially trained using 2,000 labeled data points. The trained model was then used to predict 8,000 unlabeled data points. High-confidence data was selected and added to the training set. After iteration, the classification accuracy was significantly improved.

[0040] Event analysis utilizes a hybrid LSTM-Transformer architecture, with the LSTM layer processing temporal information and the Transformer layer capturing long-range dependencies. Federated learning technology is innovatively incorporated to combine data from multiple communities for model training. Privacy protection mechanisms are incorporated into local training within each community to ensure data privacy. A reinforcement learning mechanism is introduced, with a reward function that considers factors such as processing time, satisfaction rating, and processing cost. The model continuously optimizes event allocation strategies using a policy gradient algorithm.

[0041] A graph neural network (GNN) is used to construct a three-dimensional "building-house-population-event" correlation network, with nodes including entities such as houses, people, and events. Edge weights are calculated based on spatiotemporal co-occurrence. An attention mechanism is employed to enhance focus on key features. Integrating IoT technology, data from IoT devices such as smart water meters, door sensors, and smoke alarms is integrated. For emergencies, multimodal information fusion analysis is used. When multiple abnormalities occur simultaneously, the model comprehensively identifies high-risk events and immediately triggers the early warning process, ensuring a response time of less than 15 minutes.

[0042] Incorporating active learning technology, the model automatically filters out difficult-to-classify data after processing a certain number of events and pushes it to the expert annotation platform. Events with historical resolution cases are automatically pushed based on risk. For example, when a grid worker receives a "water pipe break" event, the system also pushes information such as the average resolution time, required materials, and historical successful cases for similar incidents, facilitating efficient decision-making.

[0043] The present invention also includes a social governance module: by integrating AR spatial anchor annotation, two-stage clustering algorithm and blockchain traceability technology, a full-process intelligent management system is constructed. Grid workers are equipped with explosion-proof mobile terminals equipped with Beidou + UWB dual-mode positioning. During the inspection process, the AR collection function on the terminal is turned on. The device automatically identifies the current location based on spatial anchor technology (positioning accuracy reaches 0.3 meters) and superimposes a virtual annotation box on the real-life picture. When the fire escape is found to be occupied, the grid worker triggers a rapid collection through the voice command "Record fire escape violation". The terminal automatically takes three photos of the scene from different angles, simultaneously records a 15-second on-site video, and transcribes the grid worker's oral description of the event in real time through NLP technology. All collected data is automatically associated with AR spatial anchors, including three-dimensional coordinates, collection timestamps, and collector IDs, and is pushed to the cloud in real time through the 5G network.

[0044] The social governance module, powered by AI technology, seamlessly integrates social governance information with a grid-based household situation. This allows all departments to gain a comprehensive overview of six dimensions of social governance information: "See the map to understand the situation in the area, see the signs to understand the household situation, see the house to understand the people, and see the people to understand the information." Combined with spatiotemporal deduction models to predict event trends, this module ultimately enables real-time perception and accurate assessment of social governance trends. Through grid-based governance and intelligent analysis technologies, this platform builds a "data-driven, intelligent decision-making" social governance system. Through dynamic data collection by grid workers and the integration of multidimensional data, the platform provides a comprehensive understanding of all factors: people, places, events, objects, emotions, and organizations. AI algorithms are used for intelligent analysis and risk warning, establishing a closed-loop solution mechanism of "monitoring-warning-action-feedback." Based on multidimensional analytical models, such as spatiotemporal and event-based, this module provides decision makers at all levels with accurate analysis and assessment, ensuring early detection and resolution of issues at the grid level. This shifts the governance model from passive response to proactive prevention, comprehensively enhancing the refinement and intelligence of social governance.

[0045] After receiving the event data, the cloud server first performs format verification, such as whether the photo size is compliant and whether the video length exceeds the limit. Then, a two-stage clustering algorithm combining DBSCAN and hierarchical clustering is used. In the first stage, the DBSCAN algorithm sets the neighborhood radius ε = 50 meters (based on the average building distance in the community) and the minimum sample number MinPts = 3, and performs preliminary clustering on events received within 10 minutes. For example, three environmental sanitation incidents within 500 meters of XX Street are classified into the "street cleaning" cluster. In the second stage, the hierarchical clustering algorithm calculates the semantic similarity between clusters. For example, the "fire passage violation" cluster and the "safety hazard" cluster are merged when the similarity reaches 0.85. After clustering is completed, the system automatically generates a visual cluster map, and different clusters are marked with different colors (for example, red represents high-risk clusters).

[0046] Event priority is calculated using the formula , The risk value, R, is the incident priority; α1, α2, and α3 are weighting coefficients; R is the risk value; U is the number of affected users; and T is the remaining time. The risk value, R, is automatically assessed by a deep learning model (based on the ResNet-50 architecture) by analyzing on-site photos and videos. If a fire escape is completely blocked, R = 8 (out of 10). The number of affected users, U, is obtained from housing registration data from the housing and construction department. If there are three residential buildings within 50 meters of the violation (each with 30 households), U = 90. The remaining time, T, is automatically set based on the incident type. For fire hazard incidents, T = 4 hours by default. Weighting coefficients, α1 = 0.5, α2 = 0.3, and α3 = 0.2 (dynamically adjustable through the backend management system). After calculating the priority value, the system pushes the incident to the grid worker's mobile device and generates a highlighted alert in the backend management system.

[0047] Once the handling process is initiated, a grid worker receives the event (with navigation route planning) via their terminal. Upon arrival, they scan the violating vehicle's QR code (with built-in NFC chip) to automatically link the vehicle's owner information and take photos of the handling process (each photo includes a GPS watermark). If the handling process is not completed within 24 hours, the system automatically escalates the incident to the subdistrict office manager's terminal and triggers a text message notification. All operation records, including collection, dispatch, handling, and feedback, are encrypted and stored on the Hyperledger Fabric blockchain platform. Each block contains the Merkle root hash value of the previous 100 operation records, ensuring that the data cannot be tampered with. When querying, managers can view the complete event chain after verifying their identity using a digital certificate. The blockchain query response time is ≤3 seconds.

[0048] The present invention also includes a six-dimensional element module: a comprehensive analysis of "people, places, things, objects, emotions, and organizations". In terms of multidimensional statistical analysis, various types of data are collected, such as demographic information, geographic data, event records, etc. Semantic analysis technology is introduced, and professional natural language processing tools are used to conduct in-depth analysis of event text information. For example, for the text "A dispute occurred in the community because the pet dog was not tied with a leash", semantic labels such as "neighborhood dispute" and "pet management" are automatically affixed through lexical analysis, syntactic analysis and semantic role labeling. Based on these labels, events are marked on the map according to geographic location information, and a heat map of events with semantic labels is generated based on the number and density of events. For example, in a certain area, "environmental sanitation problems" events are concentrated, and the corresponding color of the area on the heat map is deepened to intuitively present the distribution situation.

[0049] Using knowledge graph technology, entities and relationships are extracted from large amounts of data to construct a six-dimensional element association network. For example, "people" and "places" are linked through residential relationships, and "events" and "objects" are connected through involvement relationships. Leveraging deep reinforcement learning algorithms, this association network is continuously trained and optimized. When the correlation strength between "pet-related incidents" and "low pet registration rates" reaches a set standard, a causal inference report is automatically generated, identifying possible causal relationships and assisting decision-making.

[0050] Using spatiotemporal big data analysis technology, combined with geographic information systems (GIS) and time series data, we analyze the temporal and spatial variations of six-dimensional elements. For example, we analyze the flow patterns of "people" elements and the frequency of "events" elements across different seasons and time periods, providing strong data support for community resource allocation, risk warning, and other work.

[0051] The six-dimensional element module utilizes a visual interactive interface, visually presenting the relationships between data dimensions on a grid through visual icons. Leveraging big data analytics and artificial intelligence, it achieves intelligent correlation and dynamic response across the six dimensions, forming a closed-loop management mechanism of "perception-analysis-decision-feedback." It places special emphasis on results-oriented design, using real-time holographic displays of data across the six dimensions within the grid to strengthen governance direction and enhance synergy between dimensions. The system innovatively transforms abstract governance elements into concrete, visual outcomes, maintaining the independence of each dimension while achieving integrated linkage, providing a new digital solution for modern governance.

[0052] The present invention comprises the following steps: Through standardized interfaces, cross-platform data robots, API interfaces, etc., we connect the data resources and information resources of various department systems, establish a data synchronization mechanism, and automatically obtain the latest data every morning. Use ETL tools to clean the data and unify the messy data formats, such as converting the date format of different departments to "YYYY-MM-DD" and converting string numbers into numeric types. For data with missing values, choose a processing method based on the data distribution. If the data is relatively uniform, use the average value of all data to fill it in; if the data is unevenly distributed and there are extreme values, use the median to fill it in, and finally build a six-dimensional element database containing "people, places, things, objects, emotions, and organizations"; Based on the GIS platform, a spatiotemporal model is built, and basic geographic information data such as terrain and buildings is imported to construct a three-dimensional scene. Multi-source map services such as Baidu Maps and Amap are also integrated to unify the coordinate systems of different maps. A chi-square test is used to analyze the degree of correlation between various elements, screening out elements that are closely related to the occurrence of events. Then, combined with the information gain algorithm, the key features that have a greater impact on the event are further identified. Based on these features, an event heat map is generated to intuitively present the distribution of events in different regions. Grid division is performed using an improved Delaunay algorithm, combined with real-time population flow heat maps (updated every 15 minutes) and POI data (covering various locations such as supermarkets and hospitals). Indicators such as population size and building density within each grid are calculated to assess the balance between grids. A load factor is set to measure grid workload. When a grid's load factor exceeds 1.2, indicating excessive load, the grid is re-divided to ensure that each grid has a reasonable workload. With the help of GIS maps, users can drill down to multiple levels of information, starting from the provincial, municipal, and district (county) levels, down to the street (township), community (village), responsible person grid, and finally to the building unit household status color map. For the status of the house, the Bayesian classification algorithm is used to combine the house's age, transaction records and other characteristics to determine whether the house is self-occupied, rented, or idle. Binary coding is used to identify special groups. For example, if a household has a group of people who are cared for or concerned about, the corresponding coding position is marked as 1. This information is summarized and displayed to generate a dynamic data dashboard, allowing users to quickly obtain the information they need. The LSTM-Transformer model analyzes event text, extracts key information, and determines the event type and urgency. A situational awareness model calculates the risk level based on factors such as event type, frequency, and impact. Once a high-risk event is identified, the system automatically matches the corresponding response plan from the response plan library and pushes the warning event, along with the response plan, to relevant personnel for timely resolution.

[0053] In the present invention, in the dynamic grid division step, the system receives a population flow heat map generated by the operator's base station data and mobile phone signaling analysis every 10 minutes, and integrates POI data of 12 types of points of interest including supermarkets, hospitals, schools, etc. Based on the improved DBSCAN algorithm, the neighborhood radius is set to 500 meters and the minimum sample number is set to 30 people, and the community area is scanned. When the number of people gathered within 500 meters in a certain area exceeds 30 in two consecutive data updates, the algorithm identifies it as a high-density area. For example, in the vicinity of a newly built residential area, the opening of supporting commercial entities attracts a large number of people. The system promptly detects that the population density in the area is abnormal and triggers grid division adjustments. During the division process, the system calculates the population of each grid in real time, takes the average population as a benchmark, and automatically fine-tunes the grid boundaries. By splitting or merging grids, it ensures that the population size of each grid is balanced.

[0054] In the present invention, in the steps of risk warning for preset events and preset groups, the system continuously monitors the information of the people it cares for and pays attention to, such as recording the entry and exit of the people it cares for and pays attention to through smart door magnetic devices. When it is monitored that the people it cares for and pays attention to have not gone out for 48 consecutive hours, the system immediately triggers the situational awareness calculation program. The program integrates information such as the residents' archival information, recent medical records, and the distribution of surrounding medical resources, and combines it with a preset risk assessment model to judge the risk level of the event. If the risk is assessed to be high, the system automatically retrieves the corresponding disposal plan from the plan library, such as contacting the community hospital for on-site inspections, notifying family members, etc., and pushes the warning information and disposal plan in the form of a pop-up window to the grid platform and grid member mobile terminal responsible for the area, and at the same time makes a secondary reminder via text message. After receiving the push, the grid member must respond and go to handle it within 2 hours. During the processing, the on-site situation can be uploaded to the system in real time to ensure that the risk disposal process is traceable.

[0055] In this invention, the household status grid module intelligently and dynamically sets organizational grids, responsibility grids, communities, buildings, etc., seamlessly integrates building unit household status maps with the grid, and displays the data resources and service resources of each department in a six-dimensional manner at each grid level. It establishes an intelligent grid division model and a model for dynamically optimizing grid boundaries. It deeply integrates the physical space grid with household status information data, clarifies the job responsibilities and content of grid responsible persons, and intuitively displays the six-dimensional elements of "people, places, things, objects, emotions, and organizations" through holographic visualization. It constructs a new picture of refined social governance with "full spatial coverage, full information integration, full closed-loop management, and comprehensive services." It achieves precision and automation of social governance through multimodal intelligent algorithms. At the grid division level, the system uses an improved adaptive Delaunay triangulation algorithm, combined with real-time updated GIS geographic information, population flow heat maps, and POI point of interest data, to achieve dynamic grid division. Based on an improved DBSCAN density clustering algorithm, the system automatically identifies the spatial distribution characteristics of the six-dimensional elements of "people, places, events, objects, emotions, and organizations," ensuring balanced data load across all units through a flexible grid adjustment mechanism. Regarding data association, the system has built an intelligent association engine that integrates knowledge graphs and spatiotemporal graph neural networks. It utilizes a hybrid BERT-wwm+GraphTRM model to achieve semantic understanding and association mining across cross-domain data, significantly improving association accuracy. The dynamic optimization module integrates multidimensional urban vital signs and uses a spatiotemporal Transformer model to predict changing trends in grid management priorities. The system utilizes a distributed graph database architecture, supporting millisecond-level queries and visual analysis of tens of millions of entity relationships.

[0056] refer to Figure 6Based on data services in the field of social governance, this system analyzes data on "people, places, events, emotions, and organizations," establishing a spatiotemporal data model with a real-time update mechanism. It uses spatial linkage technology to achieve synchronous interaction between two- and three-dimensional scenes, integrates multiple online map services from Baidu Maps, AutoNavi Maps, Tencent Maps, Tiandi Map, and Google Maps, supports dynamic switching and mixed loading on the server side, and enables multi-scale visualization analysis within a unified coordinate system. This system achieves comprehensive information visualization across industries, platforms, and departments. It horizontally integrates grid business data from multiple departments, including urban management, relevant departments, civil affairs, health, emergency response, environmental protection, and safety supervision, ensuring that all aspects of each department's business operations are integrated into the grid and eliminating management gaps. Vertically, it runs through all levels, from provinces, cities, districts (counties), streets (towns), communities (villages), and responsible persons, down to buildings, units, and houses, dividing large areas into small grids and refining responsibility areas, extending governance tentacles to the grassroots level. For example, a district (county) could be divided into a large grid, a street (town) into a secondary grid, and a community (village) into a smaller grid. Each community (village) could then be divided into several grid responsibility areas, with a designated person responsible for each grid responsibility area. This allows for comprehensive management of the entire jurisdiction, ensuring that everyone within the grid has a role to play and everyone is accountable. Simultaneously, each grid dynamically displays information on the six dimensions of "people, places, events, objects, emotions, and organizations."

[0057] refer to Figure 7 It is an electronic household status map that displays the family situation, distribution and related characteristics of different houses in the form of intuitive graphics and icons. It includes health information, social management, people's livelihood projects, house rentals, visit information, talent services, elderly care information, precise governance, property information, smart home, credit points, event information and other data for analysis. Through different colors and marks, it clearly distinguishes the attributes of people, such as: permanent population, floating population, caring population, and concerned population.

[0058] The above are only preferred specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solutions and inventive concepts of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. An AI-based household grid six-dimensional element holographic display system, characterized by: include: The data aggregation module integrates multi-source social governance data across departments and systems, and relies on AI algorithms and blockchain technology to achieve unified access and intelligent association of global data; The data modeling module builds an intelligent spatiotemporal data model based on the GIS platform, automatically generates and dynamically divides various types of social governance grids through multi-source maps, and uses a data space computing engine to display city-level real-scene models; The social governance module, based on AI technology, seamlessly integrates social governance information with household status grids, allowing all departments to have a comprehensive overview of six-dimensional social governance information. Combined with a spatiotemporal deduction model, it predicts event trends and enables real-time perception and analysis of social governance trends. Household Grid Module: This module dynamically sets organizational grids, responsibility grids, communities, and buildings, integrates building unit household maps with the grid, and displays each department's data resources and service resources in a six-dimensional manner at each grid level, establishing an intelligent grid division model and a model for dynamically optimizing grid boundaries. The six-dimensional element module builds a six-dimensional element structured matrix of people, places, events, objects, emotions, and organizations, and achieves multi-dimensional penetrating visualization through element decoupling and dynamic association engine; The holographic display module implements a top-down, layer-by-layer, holographic three-dimensional display in the order of institutional grid, responsibility grid, community grid, building, house, and person. It integrates the six-dimensional elements of the grid and 470 demographic and social affairs attributes in multiple dimensions, presenting them at each level through colors, lines, icons, animations, charts, and labels. It also generates a building matrix skeleton and overlays multi-source grid data. The intelligent AI module uses AI technology to establish intelligent associations between six-dimensional elements, multi-dimensional analysis, trend prediction, and natural language interaction.

2. The AI-based household grid six-dimensional element holographic display system according to claim 1 is characterized in that: In the data aggregation module, the missing values are filled with the mean formula With median filling, outliers are identified and corrected using the 3σ principle, where is the mean, n is the number of samples, is the i-th sample value; data cleaning also includes: using blockchain technology to record data sources and processing processes, introducing intelligent anomaly detection algorithms, real-time monitoring of various types of data, using generative adversarial networks to fill missing values, and generating fill-in data through the game between generators and discriminators. At the same time, applying federated learning technology, while protecting data privacy, jointly cleaning and preprocessing data from multiple parties, and using automated data verification technology to verify the cleaned data.

3. The AI-based household grid-based six-dimensional element holographic display system according to claim 1 is characterized in that: In the data modeling module, the unified WGS84 coordinate system is used, and the chi-square test formula is used. Screening the six-dimensional data correlation characteristics of people, places, things, emotions, and organizations, is the chi-square statistic, is the actual observed value, is the theoretical expected value; adaptive coordinate transformation algorithm is used when integrating multi-source map services in, It represents the coordinate value after adaptive coordinate transformation, T(P) is the result of transforming the original coordinate P based on the basic transformation matrix T, κ=0.001 is the distance correction coefficient, d is the current point distance, and D is the reference distance. The GARCH-LSTM model is used to predict the frequency of caring people’s visits.

4. The AI-based household grid-based six-dimensional element holographic display system according to claim 1 is characterized in that: In the household grid module, when the load factor When it exceeds 1.2, the genetic algorithm optimized grid re-division is triggered. The six-dimensional elements of people, places, things, objects, emotions and organizations within the grid. is the global mean, is the global mean of the six-dimensional elements of people, places, things, objects, emotions, and organizations within the grid. is the grid G i The load factor, It represents the six-dimensional elements of people, places, events, objects, emotions, and organizations within the grid. Grid division also incorporates spatial syntax features as density calculation factors, enabling millisecond-level queries of tens of millions of entity relationships through a distributed graph database, supporting concurrent access by 500 grid members online at the same time.

5. The AI-based household grid-based six-dimensional element holographic display system according to claim 1 is characterized in that: In the intelligent AI module, text classification uses the BERT-wwm model that integrates domain knowledge and uses knowledge graph technology to sort out household-related domain knowledge and integrate it into the model. It also introduces real-time event monitoring technology and combines it with IoT sensor data to quickly identify and trigger early warnings through multimodal information fusion analysis of emergency events.

6. A method for applying the AI-based household grid six-dimensional element holographic display system according to any one of claims 1 to 5, characterized in that: The following steps are involved: Data processing and database construction steps: cross-departmental interfaces are connected to relevant departments, civil affairs and other systems, data is cleaned through ETL tools, missing values are filled using mean and median, and a six-dimensional factor database is constructed; Spatiotemporal model and feature processing steps: Build a spatiotemporal model based on SuperMap and ArcGis platforms, integrate multi-source map services, use chi-square test and information gain to screen features, and generate event heat maps; Dynamic meshing steps: Dynamic meshing is performed using an improved Delaunay algorithm, with the coefficient of variation calculated to assess balance. Re-meshing occurs when the load factor exceeds 1.

2. An intelligent meshing model is established by analyzing key data such as organizational grids, responsibility grids, community grids, buildings, and houses, as well as population density, building density, and event density. Grid modeling and meshing boundaries are dynamically optimized. Geographic location is used to automatically associate the six-dimensional elements of "people, places, events, objects, emotions, and organizations" with the corresponding grids. Holographic display steps: A top-down, layer-by-layer, holographic three-dimensional display is achieved in the order of institutional grid, responsibility grid, community grid, building, house, and person. The six-dimensional elements of the grid and 470 demographic and social affairs attributes are integrated in multiple dimensions. At each level, they are presented through colors, lines, icons, animations, charts, and labels. Multi-source grid data is superimposed by generating a building matrix skeleton. Risk warning steps: Use LSTM-Transformer to analyze events, calculate risks using situational awareness formulas, and automatically push warning events with response plans.

7. The method of the AI-based household grid six-dimensional element holographic display system according to claim 6, characterized in that: In the dynamic grid division step, the real-time updated population flow heat map and POI data are combined to identify high-density areas through the improved DBSCAN algorithm.

8. The method of the AI-based household grid six-dimensional element holographic display system according to claim 6, characterized in that: In the risk warning step, abnormal activities of special groups of people trigger situational awareness calculations, and warnings are generated in combination with risk disposal formulas. Disposal plans are simultaneously pushed to the grid worker's mobile terminal, and the average response time is shortened to less than 2 hours.

Images