Real estate data query system for people stream

Abstract

The invention discloses a real estate data query system for people flow, relates to the technical field of real estate query, and aims at constructing a basic data resource pool by deploying a plurality of sensors in a real estate area to collect people flow and environment data in real time and extracting information such as real estate attributes from a related database. A distributed computing framework is used for data preprocessing, potential association is mined, and a multi-dimensional fusion data set is generated. And constructing a virtual scene model based on the data set, and developing an AR program with an information enhancement display function and an immersive VR virtual roaming program. The system provides intelligent recommendation according to historical query records and preferences of the user, and supports the user to acquire detailed real estate information through an AR / VR mode. And meanwhile, recording user behavior data in real time to optimize a recommendation algorithm and update a fusion model. According to the method, multi-dimensional query and immersive experience of real estate data are realized, and the query efficiency and the user experience are effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of real estate query technology, and in particular to a real estate information query system targeting human flow. Background Art

[0002] With the development of information technology, real estate information query systems are gradually shifting from traditional single-data query models to multi-dimensional, dynamic query methods. Most existing query systems only provide basic real estate attribute information and simple transaction record queries, such as building area and geographic location. These query systems primarily rely on static database storage and retrieval, lacking real-time perception and analysis of pedestrian flow data.

[0003] Existing technologies lack the ability to deeply integrate and analyze real-time dynamic data on pedestrian flows with multi-dimensional real estate information. Traditional query systems only display static real estate information, failing to intuitively present the relationship between the property, its surroundings, and pedestrian flows. They also struggle to provide personalized query recommendations. Furthermore, existing query methods are mostly passive, requiring users to actively enter keywords and other information to access real estate data. The query results are presented in a relatively simple format, lacking intuitiveness and interactivity, and thus failing to meet users' needs for a comprehensive and in-depth understanding of real estate information. Summary of the Invention

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

[0005] Therefore, the present invention provides a real estate information query system for pedestrian flow to solve the problems of existing real estate information query methods, such as the inability to integrate pedestrian flow dynamic data with multi-dimensional real estate information, the lack of personalized recommendation services, and how to provide multi-dimensional, dynamic and personalized real estate information query services.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] The present invention provides a real estate information query system targeting human traffic, which is characterized by comprising the following steps:

[0008] S1: Deploy mobile positioning sensors, video surveillance, and environmental sensing equipment in real estate areas to collect real-time data on the location, trajectory, duration, and environment of pedestrian flows. Simultaneously, extract real estate attributes, supporting facilities, and transaction data from the real estate registration system, surrounding supporting databases, and historical transaction records. All data is stored in a distributed storage system in a unified format to build a basic data resource pool.

[0009] S2: Using a distributed computing framework, the data in the basic data resource pool is cleaned, normalized, and feature extracted to remove noise and outliers, unify data dimensions, extract key features to form a standardized data set, and based on association rule mining and machine learning clustering algorithms, explore the potential association between pedestrian flow and real estate data, establish a data fusion model, and generate a multi-dimensional fusion data set;

[0010] S3: Based on the fused dataset, a virtual scene model of the real estate and its surrounding environment is constructed using 3D modeling and GIS technology, covering elements such as the building's exterior, interior layout, surrounding roads and green spaces. Furthermore, an AR program with information-enhanced display capabilities and an immersive VR virtual tour program are developed by combining the AR development platform and VR engine, and loaded onto the user's query terminal.

[0011] S4: In response to the user's login request through the query terminal, the system uses the intelligent recommendation algorithm to filter out real estate information that meets the user's needs from the fused data set based on the user's historical query records and preference settings. The system displays the information on the terminal. After the user selects an item to enter the details page, the system provides an option to switch between AR / VR modes to guide the user to obtain detailed real estate information.

[0012] S5: The user selects AR mode and scans the site with the terminal camera. The AR program recognizes the scene and overlays detailed information about the real estate, including ownership, supporting facilities, structure, and transaction history. The user selects VR mode and enters a virtual roaming environment, allowing for a comprehensive immersive viewing of the interior and surroundings of the real estate.

[0013] S6: During the user query process, the system records the user query behavior data in real time, including keywords, browsing time and preference adjustments, and transmits it back to the distributed storage system as input for optimizing the recommendation algorithm and updating the fusion model. At the same time, the query result display content is dynamically adjusted according to the user's real-time interactive instructions.

[0014] As a preferred solution of the real estate information query system for pedestrian flow described in the present invention, the system: uses an Internet of Things sensor network deployed in the real estate area to perceive the biometric information and behavioral pattern data of pedestrian flow in real time, and integrates them with the location, trajectory, stay time and environmental data to generate a comprehensive pedestrian flow feature data set. At the same time, data mining technology is used to extract potential market trend data from the historical transaction records of real estate, and together with the attributes and supporting data, it constitutes an extended real estate data set, which is then associated and stored in a distributed storage system.

[0015] As a preferred solution of the real estate information query system for pedestrian flow described in the present invention, it adopts an adaptive data fusion algorithm based on deep learning to perform dynamic feature extraction on the standardized data set, automatically adjusts the weight of the feature vector according to the real-time changes of the data, and constructs a multi-layer correlation neural network model to explore the deep correlation between pedestrian flow and real estate data in time series and spatial distribution, and generate a fused data set containing spatiotemporal correlation features.

[0016] As a preferred solution of the real estate information query system for pedestrian flow described in the present invention, the virtual scene model is enhanced and rendered using mixed reality (MR) technology, virtual real estate information is accurately superimposed with physical objects in the real scene, and spatial positioning technology is combined to achieve dynamic interaction between the user's position in the physical space and the virtual scene, while supporting multiple users to perform real-time collaborative editing and annotation of the same virtual scene through different terminals.

[0017] As a preferred solution of the real estate information query system for pedestrian flow described in the present invention, the intelligent recommendation algorithm performs personalized login based on the user's behavioral biometric characteristics (such as fingerprint, facial recognition, etc.), calls the user's historical query data and preference settings, analyzes the user's potential intentions by constructing a user cognitive map, and combines real-time pedestrian flow characteristics and real estate market dynamics to generate a dynamically updated recommendation list, while supporting users to conduct interactive queries through eye control or brain-computer interface devices.

[0018] As a preferred solution of the real estate information query system for pedestrian flow described in the present invention, the AR mode adopts computer vision-based target recognition technology to perform real-time identification and information association of specific objects in the on-site environment (such as buildings, billboards, etc.), and provides dynamic interactive content related to surrounding commercial activities and cultural events while displaying detailed real estate information; the VR mode supports users to perform physical simulation operations in a virtual environment (such as opening and closing doors and windows, adjusting indoor layout, etc.), and provides real-time feedback on the impact of operation results on real estate value assessment.

[0019] As a preferred solution of the real estate information query system for human flow described in the present invention, the system uses federated learning technology to perform distributed analysis of user behavior data, and by constructing a multi-centered user behavior model, it realizes cross-regional and cross-platform user behavior data sharing and collaborative learning while protecting user privacy, and dynamically adjusts the parameters of the data fusion model and the weight of the recommendation algorithm based on the learning results. At the same time, it supports linkage with external smart contract systems and automatically triggers related real estate transactions or leasing processes according to preset rules.

[0020] The beneficial effects of the present invention are as follows: by integrating the dynamic data of pedestrian flow with the multi-dimensional information of real estate, the system can provide a more comprehensive and real-time real estate information query service to help users quickly understand the real estate and its surrounding environment. Secondly, the system uses AR and VR technology to provide users with a more intuitive and immersive experience, allowing users to gain an in-depth understanding of the detailed information of the real estate without having to visit the site. Thirdly, the intelligent recommendation algorithm based on user behavior and preferences can accurately recommend real estate information that meets their needs to users, thereby improving query efficiency. In addition, the real-time recording of user query behavior data and the optimization of the recommendation algorithm and the update of the fusion model enable the system to continuously improve and provide more personalized services. Finally, the protection of user privacy through federated learning technology, while achieving cross-regional and cross-platform data sharing and collaborative learning, enhances the practicality and innovation of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0022] Figure 1 This is a flow chart of the real estate information query system for human traffic in Example 1. DETAILED DESCRIPTION

[0023] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0024] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0025] Secondly, the term "one embodiment" or "embodiment" 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 various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0026] Example 1, with reference to Figure 1 , which is the first embodiment of the present invention, provides a real estate information query system for human traffic, including the following steps:

[0027] S1: Deploy mobile positioning sensors, video surveillance, and environmental sensing equipment in real estate areas to collect real-time data on the location, trajectory, duration, and environment of pedestrian flows. Simultaneously, extract real estate attributes, supporting facilities, and transaction data from the real estate registration system, surrounding supporting databases, and historical transaction records. All data is stored in a distributed storage system in a unified format to build a basic data resource pool.

[0028] S2: Using a distributed computing framework, the data in the basic data resource pool is cleaned, normalized, and feature extracted to remove noise and outliers, unify data dimensions, extract key features to form a standardized data set, and based on association rule mining and machine learning clustering algorithms, explore the potential association between pedestrian flow and real estate data, establish a data fusion model, and generate a multi-dimensional fusion data set;

[0029] S3: Based on the fused dataset, a virtual scene model of the real estate and its surrounding environment is constructed using 3D modeling and GIS technology, covering elements such as the building's exterior, interior layout, surrounding roads and green spaces. Furthermore, an AR program with information-enhanced display capabilities and an immersive VR virtual tour program are developed by combining the AR development platform and VR engine, and loaded onto the user's query terminal.

[0030] S4: In response to the user's login request through the query terminal, the system uses the intelligent recommendation algorithm to filter out real estate information that meets the user's needs from the fused data set based on the user's historical query records and preference settings. The system displays the information on the terminal. After the user selects an item to enter the details page, the system provides an option to switch between AR / VR modes to guide the user to obtain detailed real estate information.

[0031] S5: The user selects AR mode and scans the site with the terminal camera. The AR program recognizes the scene and overlays detailed information about the real estate, including ownership, supporting facilities, structure, and transaction history. The user selects VR mode and enters a virtual roaming environment, allowing for a comprehensive immersive viewing of the interior and surroundings of the real estate.

[0032] S6: During the user query process, the system records the user query behavior data in real time, including keywords, browsing time and preference adjustments, and transmits it back to the distributed storage system as input for optimizing the recommendation algorithm and updating the fusion model. At the same time, the query result display content is dynamically adjusted according to the user's real-time interactive instructions.

[0033] It should be noted that mobile positioning sensors, video surveillance, and environmental sensing equipment are deployed within real estate areas. These devices collect real-time information on people's location, movement trajectory, duration of stay, and environmental data (such as temperature and humidity). Simultaneously, the system extracts real estate attribute data (such as floor area, unit type, etc.), surrounding supporting facilities information (such as schools, hospitals, etc.), and historical transaction data (such as past sales and rental prices) from the real estate registration system, surrounding supporting databases, and historical transaction records. All collected data is organized and encoded according to a unified data format and then stored in a distributed storage system, thus building a comprehensive basic data resource pool.

[0034] By deploying a variety of sensors and data collection devices, we achieve real-time monitoring of pedestrian flow and environmental conditions within real estate areas, as well as comprehensive collection of real estate-related data. This provides a rich, accurate, and uniformly formatted data foundation for subsequent data processing and analysis. The use of a distributed storage system ensures secure data storage and efficient retrieval, supporting rapid read and write operations on large amounts of data, laying a solid data foundation for the efficient operation of the entire query system. The construction of a basic data resource pool enables the system to integrate data resources from various channels, facilitating subsequent in-depth integration and analysis.

[0035] Using distributed computing frameworks (such as Hadoop and Spark), data stored in the basic data resource pool of the distributed storage system is preprocessed. The preprocessing process includes data cleaning (removing duplicate and erroneous data), data normalization (converting data of different dimensions to the same dimension), and feature extraction (extracting key features in the data, such as pedestrian density and transaction frequency). After preprocessing, a standardized data set is formed. Then, based on association rule mining algorithms (such as the Apriori algorithm) and machine learning clustering algorithms (such as the K-Means algorithm), potential correlations between pedestrian data and real estate data are mined. For example, the correlation between pedestrian density and surrounding real estate rents is analyzed. Using these algorithms, a data fusion model is established, and the mined correlations are expressed in the form of a mathematical model, ultimately generating a fused data set containing multi-dimensional correlation information.

[0036] Efficient preprocessing of large-scale data through a distributed computing framework ensures data quality and consistency, providing a reliable data foundation for subsequent data analysis. Data cleaning removes noise data and outliers, improving data accuracy; data normalization unifies data dimensions, enabling effective comparison and fusion of data from different sources; feature extraction highlights key information in the data, reduces data dimensionality, and improves data processing efficiency. The application of association rule mining and machine learning clustering algorithms deeply explores the potential correlation between pedestrian flow and real estate data and establishes a data fusion model. This enables the system to conduct comprehensive analysis of data from multiple dimensions, providing richer and more accurate information support for subsequent intelligent recommendations and virtual scene construction. The generation of a fused data set integrates the originally scattered data into an organic whole, providing the system with a comprehensive and in-depth data perspective, which helps to discover hidden information and value in the data.

[0037] Based on the information from the fused dataset, a virtual scene model of the property and its surroundings is constructed using 3D modeling software (such as 3ds Max and Maya) and Geographic Information System (GIS) technology. This model encompasses elements such as the building's exterior features, detailed interior layout (such as room layout and decor), surrounding road networks, and green space distribution. Furthermore, the virtual scene model is further developed and optimized using AR development platforms (such as ARKit and ARCore) and VR engines (such as Unity and UnrealEngine) to enable enhanced information display capabilities. For example, in an AR program, users can use their phone's camera to view detailed property information overlaid on a real-world scene; in a VR virtual tour program, users can immerse themselves in the property's interior and surroundings through a head-mounted display. The completed AR and VR programs are then loaded onto user terminals (such as smartphones, tablets, and VR devices) for subsequent query operations.

[0038] Utilizing 3D modeling and GIS technology to construct virtual scene models, abstract data is transformed into intuitive visual information, allowing users to more clearly understand the actual conditions of real estate and its surroundings. This not only enhances users' knowledge and understanding of real estate but also provides them with a richer and more vivid query experience. The development of AR and VR programs has further expanded the functions and application scenarios of the query system. AR programs integrate real-world scenes with virtual information, allowing users to obtain detailed information about real estate on-site or remotely, while VR programs provide users with a dynamic and immersive browsing experience. Once loaded onto the user query terminal, users can conduct queries anytime, anywhere, without being restricted by time or space, greatly improving the convenience and practicality of the query system.

[0039] When a user logs in to the system through a query terminal (such as a mobile phone or tablet), the system first verifies the user's identity and retrieves the user's historical query records and preferences. Historical query records include information such as the type, area, price range, etc. of real estate that the user has previously queried; preferences are query conditions pre-set by the user based on their own needs, such as preferences for specific conditions such as school district housing and proximity to the subway. Based on this information, the system calls intelligent recommendation algorithms (such as collaborative filtering algorithms, content-based recommendation algorithms, etc.) to filter out real estate information that highly matches the user's needs from the fused data set. This information is displayed to the user in a list on the terminal interface. Each real estate information in the list contains key information (such as pictures, prices, area, etc.). Users can select the real estate entry of interest from the list and click to enter the details view page. On the details page, the system provides an AR / VR mode switching option, and users can choose different modes to obtain more detailed real estate information.

[0040] By analyzing users' historical search history and preferences, intelligent recommendation algorithms accurately select real estate listings that meet their needs, eliminating the tedious process of searching through massive amounts of data and significantly improving search efficiency. Furthermore, the recommendation algorithm accurately tailors recommendations to users' individual needs, increasing the likelihood of finding satisfactory properties and enhancing the user experience. Displaying a list of properties on the terminal interface allows users to quickly browse through multiple options, making comparison and selection easier. Providing the option to switch between AR and VR modes further enriches user search methods and meets their diverse information needs. This not only increases user satisfaction and usage of the query system but also provides a more effective channel for real estate promotion and transactions.

[0041] When users select AR mode, they scan the property or surroundings using the terminal's camera. The AR program uses computer vision technology to identify and analyze the scene in real time, automatically matching the property's three-dimensional coordinates. The program then overlays detailed property information (such as ownership information, proximity to surrounding amenities, building structure, and historical transaction records) onto the real-world scene as a virtual overlay. Users can interact with the displayed information through the touchscreen or voice commands to view different content. If users select VR mode, they enter a virtual walkthrough environment. Within this environment, users can use VR devices (such as head-mounted displays or controllers) to fully explore the property's interior structure (e.g., the layout and decor of rooms like the living room, bedroom, and kitchen) and surrounding environment (e.g., nearby parks, shopping malls, and other facilities). Users can also perform physical simulations within the virtual environment, such as opening and closing doors and windows and adjusting the layout of interior furniture. The system provides real-time feedback on the results of these operations and dynamically adjusts the property's valuation based on their impact.

[0042] AR mode integrates virtual information with real-world scenes, allowing users to intuitively access detailed information about real estate in the real world, eliminating the disconnect between information and reality that plagues traditional search methods. Users can experience the property's actual location, surroundings, and related information on-site, enhancing their understanding and knowledge of the property. Interactive operations also allow users to access more information, making search more flexible and engaging. VR mode offers users an immersive experience, allowing them to gain a comprehensive virtual understanding of the property's interior and surroundings, as if they were there in person. This not only meets users' remote search needs but also avoids the inconvenience and cost of in-person visits. The physical simulation operation function allows users to personalize and plan real-world properties, gaining real-time insights into the impact of operations on value, providing stronger support for decision-making and further enhancing the user experience and system practicality.

[0043] During user queries, the system records user behavior data in real time, including the query keywords entered, the time spent browsing each page, the number of clicks on recommended results, and preference adjustments. This data is transmitted back to the distributed storage system via the network and serves as input for optimizing the recommendation algorithm and updating the data fusion model. Based on this behavioral data, the system uses machine learning algorithms (such as reinforcement learning) to dynamically adjust the parameters of the recommendation algorithm to more accurately predict user interests and needs. At the same time, the data fusion model is retrained and updated based on new data to reflect the latest data associations. In addition, the system dynamically adjusts the display of query results based on real-time user interaction commands (such as sliding the screen, clicking buttons, etc.) to ensure that the query results always meet the user's current needs.

[0044] By recording user behavior data in real time and feeding it back to the system, dynamic tracking and precise understanding of user needs are achieved. This enables the recommendation algorithm to self-optimize based on real-time user behavior, continuously improving the accuracy and personalization of recommendation results. Updates to the data fusion model ensure that the system promptly reflects the latest data changes and relationships, providing more accurate data support for subsequent data analysis and queries. Dynamically adjusting the query result display based on real-time user interaction commands improves the system's responsiveness and flexibility, ensuring that users always receive the most relevant query results. This not only enhances the user experience but also increases the system's intelligence, enabling the query system to better adapt to diverse user needs and market dynamics.

[0045] Specifically, through the Internet of Things sensor network deployed in the real estate area, the biometric information and behavioral pattern data of the crowd are perceived in real time, and integrated with the location, trajectory, stay time and environmental data to generate a comprehensive crowd feature data set. At the same time, data mining technology is used to extract potential market trend data from the historical transaction records of real estate, and together with the attributes and supporting data, it constitutes an extended real estate data set. The two are linked and stored in a distributed storage system.

[0046] It should be noted that an IoT sensor network is constructed within the real estate area. This network includes biometric sensors (such as facial recognition cameras and fingerprint scanners), behavioral monitoring devices (such as infrared thermal imagers and pressure-sensitive floor tiles), and environmental monitoring sensors (such as temperature and humidity sensors and light sensors). These devices collect real-time biometric information (such as facial features and fingerprint data), behavioral data (such as walking speed and frequency of changes in rest positions), and environmental data (such as on-site temperature and humidity, and light intensity). Simultaneously, the system establishes data connections with the real estate registration system, surrounding supporting databases, and historical transaction records. Data mining algorithms (such as time series analysis) are used to extract market trend data (such as housing price fluctuation curves and rental demand change rates) from historical transaction records. The collected biometric information and behavioral data are integrated with location, trajectory, duration of stay, and environmental data, and a comprehensive crowd flow feature dataset is generated using data fusion algorithms (such as Bayesian network fusion). Furthermore, this market trend data is linked with real estate attribute data (such as construction age and unit structure) and supporting data (such as transportation convenience score and commercial facility density) to form an extended real estate dataset. All data is transmitted to the distributed storage system through the data bus and stored according to the preset data structure and index rules to ensure data integrity and accessibility.

[0047] By deploying an IoT sensor network and incorporating data mining techniques, the system achieves precise perception of the biometrics and behavioral patterns of pedestrian flows. Market trend data is extracted from historical transaction records and combined with traditional real estate data to form an expanded dataset. This comprehensive data collection approach not only enriches data dimensions but also provides a more comprehensive foundation for subsequent in-depth data analysis and forecasting. Specifically, biometric and behavioral pattern data can reflect the real-time status and potential demand of pedestrian flows, while market trend data provides a forward-looking reference for real estate valuation and investment decisions. Correlating and storing this data in a distributed system enhances the reliability and scalability of data management, providing high-quality data support for subsequent intelligent recommendations and virtual scene construction, and improving the overall system's intelligence and service capabilities.

[0048] Specifically, an adaptive data fusion algorithm based on deep learning is used to dynamically extract features from the standardized data set, automatically adjust the weights of the feature vectors according to the real-time changes of the data, and construct a multi-layer correlation neural network model to explore the deep correlation between pedestrian flow and real estate data in time series and spatial distribution, and generate a fused data set containing spatiotemporal correlation features.

[0049] It should be noted that the standardized dataset is processed using an adaptive data fusion algorithm based on deep learning. First, the standardized data is input into a deep learning model consisting of multiple neural network layers, such as convolutional, pooling, and fully connected layers. During data processing, the model uses a dynamic feature extraction mechanism to analyze data trends in real time and automatically adjust the weights of each feature in the feature vector based on these changes. For example, if a feature (such as pedestrian density) significantly increases its impact on real estate value within a specific time period, the algorithm automatically increases the weight of that feature in the feature vector. Simultaneously, a multi-layer associative neural network model with multiple hidden layers is constructed to mine deep correlations between pedestrian and real estate data across time series (e.g., daily, weekly, and monthly changes) and spatial distribution (e.g., across different regions and building types). By training the neural network model and continuously optimizing its parameters, the model accurately captures complex correlation patterns between the data, ultimately generating a fused dataset containing spatiotemporal correlation features that more comprehensively reflects the dynamic relationship between pedestrian and real estate.

[0050] By employing an adaptive data fusion algorithm based on deep learning, dynamic feature extraction and weight adjustment are achieved for standardized datasets, automatically optimizing feature vectors based on real-time data changes. This not only improves the accuracy and adaptability of data fusion, but also ensures that the fused data better reflects actual conditions. The constructed multi-layer correlation neural network model effectively mines the deep connections between pedestrian and real estate data in both temporal and spatial dimensions, revealing complex relationships and underlying patterns that are difficult to detect using traditional methods. The resulting fused dataset contains rich temporal and spatial correlation features, providing more accurate and comprehensive data support for subsequent applications such as intelligent recommendations, virtual scene construction, and market trend forecasting, thereby enhancing the overall system's intelligence and decision-making capabilities.

[0051] Specifically, mixed reality (MR) technology is used to enhance the rendering of the virtual scene model, accurately superimpose virtual real estate information with physical objects in the real scene, and combine spatial positioning technology to achieve dynamic interaction between the user's position in the physical space and the virtual scene, while supporting multiple users to perform real-time collaborative editing and annotation of the same virtual scene through different terminals.

[0052] It should be noted that the constructed virtual scene model is enhanced and rendered using mixed reality (MR) technology. First, the virtual real estate information (such as the structure of the house and the layout of the surrounding facilities) is accurately superimposed with the physical objects in the real scene (such as actual buildings and roads). Through spatial positioning technology (such as SLAM algorithm), the user's position and movement trajectory in the physical space are tracked in real time, and this position information is mapped to the virtual scene to achieve dynamic interaction between the user and the virtual scene. For example, when the user moves in the physical space, the perspective in the virtual scene will be adjusted in real time according to the user's actual position and orientation. At the same time, a multi-user collaborative editing function is developed to allow multiple users to connect to the same virtual scene through different terminals (such as smartphones, tablets, MR glasses). Users can perform operations such as marking and annotation in the scene, and the system will synchronize these operations in real time, allowing other users to see the editing results immediately, enabling team members to collaboratively view and discuss the same real estate project.

[0053] By adopting mixed reality (MR) technology, the precise superposition of virtual real estate information and real-life scenes is achieved, enhancing the user's intuitive perception of real estate and its surrounding environment. Combined with spatial positioning technology, users can interact naturally and dynamically with virtual scenes in physical space, improving the authenticity and interactivity of the user experience. In addition, it supports multi-user real-time collaborative editing and annotation functions, allowing team members to transcend geographical restrictions and jointly conduct efficient collaboration and in-depth discussions on real estate projects. This function is particularly suitable for scenarios such as remote teamwork, project review and client presentation, improving work efficiency and collaboration effects. Overall, this step provides a more immersive and efficient solution for real estate information query and project collaboration through the combination of MR technology and multi-user collaborative functions.

[0054] Specifically, the intelligent recommendation algorithm performs personalized login based on the user's behavioral biometric characteristics (such as fingerprints, facial recognition, etc.), calls the user's historical query data and preference settings, analyzes the user's potential intentions by building a user cognitive map, and combines real-time traffic characteristics and real estate market dynamics to generate a dynamically updated recommendation list. At the same time, it supports users to conduct interactive queries through eye control or brain-computer interface devices.

[0055] It should be noted that the intelligent recommendation algorithm first uses behavioral biometrics (such as fingerprints and facial recognition) to perform personalized login to ensure accurate identification of the user. After login, the system accesses the user's historical query data and preferences. This data includes the user's past property search types, regions, price ranges, preferred tags (such as school district housing and subway housing), as well as query frequency and dwell time across different time periods. Based on this data, the system utilizes user cognitive graph technology to construct a cognitive graph consisting of user interest nodes (such as interest in a specific area), behavioral edges (such as frequent searches for a certain type of property), and semantic relationships (such as the importance attached to specific supporting facilities). This graph is analyzed using a graph neural network algorithm to identify the user's underlying intent. Simultaneously, the system captures real-time pedestrian flow characteristics (such as changes in pedestrian density in the current area) and real estate market dynamics (such as new property launches and policy adjustments). It combines this external dynamic information with the user's cognitive graph and dynamically adjusts recommendation weights using a reinforcement learning algorithm to generate a dynamically updated recommendation list. In addition, the system supports users to conduct interactive queries through eye-controlled devices (such as eye-tracking glasses) or brain-computer interface devices. Users can trigger query commands by gazing at specific areas of the screen or through EEG signals, and the system will respond and update recommendation results in real time.

[0056] Personalized login based on behavioral biometrics ensures the accuracy and security of user identification, avoiding the tediousness and potential risks of traditional account and password login. By analyzing users' potential intentions using the user cognitive graph and combining it with real-time traffic patterns and market trends, the system enables precise and dynamic updates of recommendation results, improving the personalization and timeliness of recommendations. For example, when traffic density in a certain area increases or new properties are released, the system can quickly adjust recommendations to prioritize properties that users may be interested in. Furthermore, interactive query methods that support eye contact control and brain-computer interface devices provide users with a more convenient and intuitive query experience, particularly suitable for users with limited mobility or those seeking efficient operations. This multimodal interaction not only enhances the user experience but also expands the applicable scenarios and user groups of the query system, enhancing the system's innovation and competitiveness.

[0057] Specifically, the AR mode uses computer vision-based target recognition technology to perform real-time identification and information association of specific objects in the on-site environment (such as buildings, billboards, etc.). While displaying detailed real estate information, it provides dynamic interactive content related to surrounding commercial activities and cultural events; the VR mode supports users to perform physical simulation operations in a virtual environment (such as opening and closing doors and windows, adjusting interior layout, etc.), and provides real-time feedback on the impact of the operation results on the real estate value assessment.

[0058] It should be noted that in AR mode, the system uses computer vision-based target recognition technology to identify specific objects in the on-site environment (such as buildings, billboards, etc.) in real time. Specifically, the system obtains on-site image data by querying the terminal's camera, and then uses pre-trained deep learning target detection models (such as YOLO, Faster R-CNN, etc.) to quickly locate and classify objects in the image. After identifying a specific object, the system associates it with detailed information about the real estate (such as property information, historical transaction records), as well as surrounding commercial activities (such as promotions, new stores), and cultural events (such as art exhibitions, community activities). This associated information is superimposed on the object through augmented reality technology, and users can view the details by touching the screen or using voice commands.

[0059] In VR mode, the system supports users to perform physical simulation operations in a virtual environment. After the user enters the virtual scene through VR equipment (such as a head-mounted display and a handle), he can interact with objects in the scene, such as opening and closing doors and windows, moving furniture, and adjusting the interior layout. The system uses a physics engine (such as Unity Physics, UnrealPhysics) to calculate in real time the impact of the user's operations on the virtual environment and feeds back these operation results to the user. At the same time, the system adjusts the value assessment parameters of the real estate based on the user's operations. For example, if the user opens the window, the lighting score may be improved, which will have a positive impact on the value assessment. The adjusted value assessment results are displayed in real time in the virtual scene, and the user can intuitively see the impact of the operation on the value of the real estate.

[0060] Using AR's object recognition and information association technology, users can obtain detailed information about a property and its surroundings, both on-site and remotely, while also understanding its connections to surrounding commercial and cultural activities. This interactive augmented reality experience not only enriches the user's search experience but also provides more comprehensive decision-making support, helping users better assess the property's overall value and potential returns.

[0061] In VR mode, users can intuitively experience the impact of different layout and functional adjustments on real estate value through physical simulation. This real-time feedback mechanism provides a more immersive experience, allowing users to preview and evaluate the effects of different design options in a virtual environment, enabling them to make more informed decisions. Furthermore, this interactive valuation method not only enhances users' understanding of real estate value but also strengthens user interaction with the system, improving its practicality and user satisfaction.

[0062] Specifically, the system uses federated learning technology to conduct distributed analysis of user behavior data. By building a decentralized user behavior model, it realizes cross-regional and cross-platform user behavior data sharing and collaborative learning while protecting user privacy. It dynamically adjusts the parameters of the data fusion model and the weight of the recommendation algorithm based on the learning results. It also supports linkage with external smart contract systems and automatically triggers related real estate transactions or leasing processes according to preset rules.

[0063] It should be noted that the system uses federated learning technology to perform distributed analysis of user behavior data. The specific operations are as follows. First, the system constructs local user behavior datasets containing information such as user query history, preferences, and interactions in AR / VR modes. The system then uses encryption technology to encrypt these local datasets to ensure data security during transmission and sharing. Next, the system constructs decentralized user behavior models. These models are independently trained on multiple servers or nodes, with each model processing only local datasets. Using technologies such as homomorphic encryption, the models are trained on encrypted data, generating encrypted model parameters. While protecting user privacy, the system utilizes secure multi-party computation technology to enable cross-regional and cross-platform user behavior data sharing and collaborative learning. Model parameters on each node are aggregated through a secure protocol to form a global user behavior model. Based on the learning results of the global model, the system dynamically adjusts the parameters of the data fusion model and the weights of the recommendation algorithm. For example, if the global model finds that users in a specific area are generally interested in a certain type of real estate, the system will increase the weight of that type of property in the recommendation list accordingly. The system also supports integration with external smart contract systems, automatically triggering relevant real estate transactions or leasing processes based on pre-set rules. For example, once a user completes the selection and conditions of a property in the system, the system automatically executes the transaction or leasing agreement through a smart contract, ensuring an automated and transparent process.

[0064] By utilizing federated learning technology, the system enables cross-regional and cross-platform sharing and collaborative learning of user behavior data while protecting user privacy. This not only improves data utilization efficiency but also enhances the model's generalization and accuracy. The system dynamically adjusts the parameters of the data fusion model and the weights of the recommendation algorithm based on the learning results of the global model, making recommendations more precise and better meeting the diverse needs of users. Furthermore, integration with the smart contract system automates the real estate transaction or leasing process, improving transaction efficiency, reducing the risk of human intervention, and enhancing the system's credibility and reliability. This innovative technical solution provides users with a more secure, efficient, and intelligent real estate information query and transaction experience.

[0065] In summary, the present invention achieves this by: integrating dynamic data on pedestrian flows with multi-dimensional information on real estate, so that the system can provide a more comprehensive and real-time real estate information query service, helping users to quickly understand real estate and its surrounding environment. Secondly, the system uses AR and VR technologies to provide users with a more intuitive and immersive experience, allowing users to gain an in-depth understanding of the detailed information of real estate without having to visit the site. Thirdly, the intelligent recommendation algorithm based on user behavior and preferences can accurately recommend real estate information that meets their needs to users, thereby improving query efficiency. In addition, the real-time recording of user query behavior data, optimization of the recommendation algorithm, and updating of the fusion model enable the system to continuously improve and provide more personalized services. Finally, the system's practicality and innovation are enhanced by protecting user privacy through federated learning technology, while achieving cross-regional and cross-platform data sharing and collaborative learning.

[0066] 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 the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A real estate information query system for human traffic, characterized in that: The following steps are involved: S1: Deploy mobile positioning sensors, video surveillance, and environmental sensing equipment in real estate areas to collect real-time data on the location, trajectory, duration, and environment of pedestrian flows. Simultaneously, extract real estate attributes, supporting facilities, and transaction data from the real estate registration system, surrounding supporting databases, and historical transaction records. All data is stored in a distributed storage system in a unified format to build a basic data resource pool. S2: Using a distributed computing framework, the data in the basic data resource pool is cleaned, normalized, and feature extracted to remove noise and outliers, unify data dimensions, extract key features to form a standardized data set, and based on association rule mining and machine learning clustering algorithms, explore the potential association between pedestrian flow and real estate data, establish a data fusion model, and generate a multi-dimensional fusion data set; S3: Based on the fused dataset, a virtual scene model of the real estate and its surrounding environment is constructed using 3D modeling and GIS technology, covering elements such as the building's exterior, interior layout, surrounding roads and green spaces. Furthermore, an AR program with information-enhanced display capabilities and an immersive VR virtual tour program are developed by combining the AR development platform and VR engine, and loaded onto the user's query terminal. S4: In response to the user's login request through the query terminal, the system uses the intelligent recommendation algorithm to filter out real estate information that meets the user's needs from the fused data set based on the user's historical query records and preference settings. The system displays the information on the terminal. After the user selects an item to enter the details page, the system provides an option to switch between AR / VR modes to guide the user to obtain detailed real estate information. S5: The user selects AR mode and scans the site with the terminal camera. The AR program recognizes the scene and overlays detailed information about the real estate, including ownership, supporting facilities, structure, and transaction history. The user selects VR mode and enters a virtual roaming environment, allowing for a comprehensive immersive viewing of the interior and surroundings of the real estate. S6: During the user query process, the system records the user query behavior data in real time, including keywords, browsing time and preference adjustments, and transmits it back to the distributed storage system as input for optimizing the recommendation algorithm and updating the fusion model. At the same time, the query result display content is dynamically adjusted according to the user's real-time interactive instructions.

2. The real estate information query system for human traffic according to claim 1, characterized in that: Through the Internet of Things sensor network deployed in the real estate area, the biometric information and behavioral pattern data of the crowd are perceived in real time, and integrated with the location, trajectory, stay time and environmental data to generate a comprehensive crowd feature data set. At the same time, data mining technology is used to extract potential market trend data from the historical transaction records of real estate, and together with the attributes and supporting data, it constitutes an extended real estate data set. The two are linked and stored in a distributed storage system.

3. The real estate information query system for human traffic according to claim 2, characterized in that: An adaptive data fusion algorithm based on deep learning is used to dynamically extract features from the standardized data set, automatically adjust the weights of the feature vectors according to the real-time changes of the data, and construct a multi-layer association neural network model to explore the deep correlation between pedestrian flow and real estate data in time series and spatial distribution, and generate a fused data set containing spatiotemporal correlation features.

4. The real estate information query system for human traffic according to claim 3, characterized in that: Mixed reality (MR) technology is used to enhance the rendering of the virtual scene model, accurately superimposing virtual real estate information with physical objects in the real scene, and combining spatial positioning technology to achieve dynamic interaction between the user's position in the physical space and the virtual scene. At the same time, it supports real-time collaborative editing and annotation of the same virtual scene by multiple users through different terminals.

5. The real estate information query system for human traffic according to claim 4, characterized in that: The intelligent recommendation algorithm performs personalized login based on the user's behavioral biometric characteristics (such as fingerprints, facial recognition, etc.), calls the user's historical query data and preference settings, analyzes the user's potential intentions by building a user cognitive map, and combines real-time traffic characteristics and real estate market dynamics to generate a dynamically updated recommendation list. It also supports users to conduct interactive queries through eye control or brain-computer interface devices.

6. The real estate information query system for human traffic according to claim 5, characterized in that: The AR mode uses computer vision-based target recognition technology to perform real-time identification and information association of specific objects in the on-site environment (such as buildings, billboards, etc.). While displaying detailed real estate information, it provides dynamic interactive content related to surrounding commercial activities and cultural events; the VR mode supports users to perform physical simulation operations in a virtual environment (such as opening and closing doors and windows, adjusting indoor layout, etc.), and provides real-time feedback on the impact of the operation results on the real estate value assessment.

7. The real estate information query system for human traffic according to claim 6, characterized in that: The system uses federated learning technology to conduct distributed analysis of user behavior data. By building a decentralized user behavior model, it enables cross-regional and cross-platform user behavior data sharing and collaborative learning while protecting user privacy. It dynamically adjusts the parameters of the data fusion model and the weight of the recommendation algorithm based on the learning results. It also supports linkage with external smart contract systems and automatically triggers related real estate transactions or leasing processes according to preset rules.

Images