Construction method and system of smart park digital twin based on big data

Abstract

The invention relates to the technical field of digital twinning, in particular to a smart park digital twinning construction method and system based on big data, and the method comprises the steps: constructing a multi-source heterogeneous data fusion collection system, deploying an Internet of Things sensor in a smart park to collect data, and carrying out the data preprocessing; constructing a distributed big data storage architecture to store the collected data, and extracting data depth features in combination with machine learning and deep learning algorithms; constructing a park high-precision digital twin three-dimensional model, and accurately associating the extracted data depth features with a model entity; a real-time data monitoring mechanism is set, a message queue and edge calculation are adopted for data transmission and processing, a multi-terminal human-computer interaction interface regular evaluation model is developed, and an optimization algorithm is used for iteratively upgrading the digital twinborn body. Through full-process data driving and model iteration, digital mapping of the smart park is realized, and the management efficiency and decision scientificity of the smart park are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of digital twinning, in particular to a method and system for constructing a digital twinning of a smart park based on big data. BACKGROUND

[0002] With the development of information technology, digital twinning technology has been widely used in the construction and management of smart parks. Digital twinning can digitally map the physical entities of a smart park, providing decision support for the planning, operation and management of the park. However, the existing methods for constructing digital twinning of a smart park still have many problems. On the one hand, the types and scope of data collection are limited, and it is not possible to comprehensively obtain information about equipment, environment, personnel and other aspects in the park, resulting in the digital twinning being unable to accurately reflect the true state of the physical park; on the other hand, in the data processing process, the traditional method lacks the ability to analyze and mine big data, making it difficult to achieve efficient use of data, resulting in poor real-time performance and accuracy of the digital twinning. In addition, the existing construction method lacks an effective processing mechanism for dynamic changes in data, and cannot update the digital twinning in a timely manner, making it difficult to meet the complex and changing management needs of the smart park. SUMMARY

[0003] The present application provides a method and system for constructing a digital twinning of a smart park based on big data to solve the problems of incomplete data collection, low processing efficiency and inability to adapt to dynamic changes in data in the prior art, and to achieve accurate digital mapping and efficient management of physical entities in a smart park.

[0004] The present application is achieved by the following technical solutions: A method for constructing a digital twinning of a smart park based on big data is provided, the method comprising the following steps: Step S10: Collecting multispectral images of the packaging glass bottle at 0°, 45° and 90°, separating the packaging glass bottle body and the packaging label area based on spectral differences, removing glare interference by polarization decomposition, and outputting the spectral features Step S10: Constructing a multi-source heterogeneous data fusion collection system through Internet of Things technology, deploying Internet of Things sensors inside the smart park to collect data comprehensively and at multiple levels, and performing data preprocessing; Step S20: Constructing a distributed big data storage architecture to store the collected data, and extracting deep features of the data using machine learning and deep learning algorithms; Step S30: Constructing a high-precision digital twinning of the park three-dimensional model, and accurately associating the extracted deep features of the data with the model entities; Step S40: Set up real-time data monitoring mechanism, use message queue and edge computing for high-throughput data transmission and fast processing, develop multi-terminal human-computer interaction interface, regularly evaluate the three-dimensional model of the park digital twin based on accuracy, real-time and stability indicators, and use optimization algorithm to iteratively upgrade the three-dimensional model of the park digital twin.

[0005] Preferably, the step S10 of constructing a multi-source heterogeneous data fusion collection system through Internet of Things technology includes the following steps: Data collection scenario and index planning: According to the functional zoning of the smart park, the data collection requirements are determined, including closed areas, open areas and boundary areas. The closed area collects temperature, humidity, carbon dioxide concentration and equipment operating current and voltage indicators, and sets the collection range, such as temperature collection range of 0-50℃, carbon dioxide concentration collection range of 0-5000ppm, equipment operating current of 0-50A, and equipment operating voltage of 0-380V. The open area collects vehicle information, personnel flow information and environmental lighting indicators, including license plate, vehicle type and entry and exit time, personnel location and movement trajectory, and environmental lighting indicators of 0-10000lux. The boundary area such as the fence and the entrance and exit position collects meteorological indicators such as wind speed and rainfall, and sets the wind speed collection range of 0-30m / s and the rainfall collection range of 0-50mm / h; Sensor selection and deployment: According to the scene requirements, Internet of Things sensors are selected. For environmental monitoring, temperature and humidity sensors and carbon dioxide sensors are used, with temperature and humidity sensor accuracy of ±0.5℃, ±2%RH, and carbon dioxide sensor accuracy of ±50ppm. In the machine room, additional smoke sensors are deployed to trigger an alarm when smoke is detected and the response time is less than 30s. For equipment monitoring, current sensors and voltage sensors are used, with current sensor accuracy of 0.01A and voltage sensor range of 0-500V. For personnel and vehicle monitoring, 4K high-definition cameras are used with frame rate of 60fps, RFID reader-writer with recognition distance of 0-5m, and infrared beam sensor to prevent personnel from climbing over the fence. The deployment density is classified according to the importance of the function, with an interval of 30 meters for key areas such as machine rooms, power distribution rooms, etc., an interval of 50 meters for ordinary office areas, and one group of cameras every 50 meters on main roads. Inside the building, according to the functional area division, temperature and humidity sensors, carbon dioxide concentration sensors are installed every 30-50 meters in office, corridor, machine room, etc. to monitor indoor environmental parameters in real time. Vibration sensors, current sensors and voltage sensors are deployed on elevators, air conditioners, power distribution equipment and other critical facilities to obtain physical parameters of equipment operation. 4K high-definition cameras and RFID reader-writers are installed at parking lot entrances, main road intersections and personnel-intensive areas to collect vehicle and personnel flow information; Sensor Internet of Things construction: for temperature and humidity sensors, carbon dioxide sensors, current sensors and voltage sensors, ZigBee networking is adopted, for 4K high-definition cameras, RFID readers and infrared beam sensors, WIFI / Ethernet direct connection is adopted, one Internet of Things gateway is set up every 5000 square meters, supporting protocol conversion such as ZigBee to TCP / IP, realizing sensor data aggregation, for fixed sensors, AC220V power supply is adopted, for mobile monitoring points such as temporary construction area, lithium battery combined with solar charging is adopted, lithium battery endurance ≥72 hours; Multi-source data access integration: internal data docking and external data docking are constructed, internal data docking includes data interconnection with park property management system, access control system and energy management system through API interface, among which device account is obtained from property management system, personnel ID is obtained from access control system, and electric meter data is obtained from energy management system, and data exchange is carried out in JSON format, external data docking includes docking meteorological platform and traffic department API through HTTPS protocol, among which hourly forecast is obtained from meteorological platform through HTTPS protocol, and surrounding road conditions are obtained from traffic department API, and next day data is automatically synchronized at 3 o'clock every morning, and a unique ID is allocated to all collection points, the format is area code+device type+serial number, and data traceability is ensured.

[0006] Preferably, the data preprocessing step in step S10 includes: Data cleaning: using rule-based cleaning method, setting various data value range threshold, eliminating noise data beyond reasonable range, using clustering algorithm to identify and remove duplicate data, using data interpolation algorithm to fill missing data; Format conversion: using ETL tool to convert different format data such as CSV, XML and JSON into Parquet format, improving data storage and processing efficiency; Normalization processing: Z-score standardization method is used for normalization processing of numerical value type data, so that the mean value of processed data is 0 and the standard deviation is 1, eliminating the influence of dimension, laying a foundation for subsequent data analysis, and non-numerical value type data is not normalized.

[0007] Preferably, the step of constructing distributed big data storage architecture to store collected data and extracting deep features of data by combining machine learning and deep learning algorithm in step S20 includes: Distributed big data storage architecture: A storage architecture combining Hadoop distributed file system and distributed database is adopted. The Hadoop distributed file system is used to store massive unstructured and semi-structured data, and the distributed database is used to store structured data. A data index and metadata management system is established. Apache Solr is used to build a full-text search index for quick data query. Apache Atlas is used for metadata management to record information such as data source, processing process and storage location, and realize full life cycle management of data. Data classification strategy: According to the importance, update frequency and usage frequency of data, the classification storage strategy is set. The high-frequency usage data such as real-time collected key equipment operation data and security monitoring data are stored in high-performance solid state disk SSD to ensure fast reading. The low-frequency usage data such as historical data and analysis result data are stored in low-cost mechanical hard disk HDD or cloud storage to reduce storage cost. Deep feature extraction of data: For image type data, an improved YOLOv5 convolutional neural network model based on TensorFlow framework is built to train and optimize the images collected by 4K high-definition camera, realizing high-precision identification of personnel behavior and vehicle type. For time series type data, a hybrid neural network model combining Transformer architecture and LSTM is used to analyze the time series data such as current and voltage of equipment operation, extract abnormal features of equipment operation, and predict the time and type of equipment failure. For text data, BERT pre-training model is used to extract features of text data such as equipment maintenance records and user feedback, and mine potential equipment problems and user demands.

[0008] Preferably, the step S30 of constructing a high-precision digital twin three-dimensional model of the park includes the step of precisely associating the extracted deep data features with the model entity. Construction of high-precision digital twin three-dimensional model of park: Professional three-dimensional modeling software is used to construct a high-precision three-dimensional geometric model of the park in combination with architectural design drawings, BIM models and field surveying data of the park, and a local update mechanism is set. In the modeling process, the appearance materials such as glass, stone and metal, internal structure such as floor layout and room function, and equipment form such as air conditioning unit, power distribution cabinet and elevator car of the building are finely described to ensure the authenticity and detail of the model. Visualization rendering: Based on Unity 3D or Unreal Engine game engine, a digital twin visualization platform is built. The constructed three-dimensional geometric model is imported into the digital twin visualization platform. Shader shader technology is used to realize realistic light and shadow effects and material rendering, and restore the real visual effect of the park. Data association: A data channel is established using C# or Python scripts to accurately associate the data features extracted in step S20 with the corresponding entities in the high-precision digital twin 3D model of the park.

[0009] Preferably, step S40 includes: Real-time data monitoring and updating: Message queues are used to achieve high-throughput data transmission and real-time distribution. Edge computing nodes are set up within the park to perform preliminary processing and filtering of the collected raw data, reducing data transmission volume and improving data processing efficiency. When the data collected by IoT sensors changes, the changed data is updated to the digital twin visualization platform in real time via the WebSocket protocol, triggering the local update mechanism of the park's high-precision digital twin 3D model to synchronize the data, ensuring that the park's high-precision digital twin 3D model and the actual park are synchronized within seconds. Interactive Interface Development: Develop a feature-rich and user-friendly human-computer interaction interface that supports access from multiple terminals, including PCs, mobile devices, and large screens. Functions include remote monitoring, equipment control, and report generation. Managers can remotely monitor the park through the human-computer interaction interface, view the park's operating status through a high-precision digital twin 3D model, remotely control equipment, and generate various statistical reports and visualization charts using the data analysis unit. A mechanism for evaluating and optimizing a high-precision digital twin 3D model of the park is established: The high-precision digital twin 3D model of the park is regularly evaluated based on assessment indicators such as accuracy, real-time performance, stability, and usability. Accuracy is evaluated by comparing the actual park status with the displayed status of the high-precision digital twin 3D model and calculating the error rate. Real-time performance is evaluated by the time interval between recording data changes and the completion of the update of the high-precision digital twin 3D model. Stability is evaluated by statistically analyzing the number of failures and runtime of the high-precision digital twin 3D model within a certain period. Usability is evaluated through user feedback and operation process analysis. Based on the evaluation results, model optimization algorithms are used to optimize the rendering performance and data transmission efficiency of the high-precision digital twin 3D model of the park, and the high-precision digital twin 3D model of the park is iteratively upgraded to continuously improve its performance and reliability.

[0010] Furthermore, to achieve the above objectives, this invention also proposes a system for constructing a smart park digital twin based on big data, wherein the system comprises: Park Data Acquisition and Preprocessing Module: Used to build a multi-source heterogeneous data fusion acquisition system through IoT technology, deploy IoT sensors in the smart park to collect data in an all-round and multi-level manner, and perform data preprocessing; The park's data storage and feature extraction module is used to build a distributed big data storage architecture to store the collected data and to extract deep features of the data by combining machine learning and deep learning algorithms. Digital Twin Modeling and Data Association Module: Used to construct a high-precision 3D digital twin model of the park, and to accurately associate the extracted data depth features with the model entities; Real-time monitoring and optimization module: This module is used to set up a real-time data monitoring mechanism, use message queues and edge computing for high-throughput data transmission and fast processing, develop multi-terminal human-computer interaction interfaces to periodically evaluate the 3D model of the park's digital twin based on accuracy, real-time performance, and stability indicators, and use optimization algorithms to iteratively upgrade the 3D model of the park's digital twin.

[0011] Furthermore, to achieve the above objectives, the present invention also proposes a device for constructing a smart park digital twin based on big data. The device includes: a memory, a processor, and programs such as an algorithm for constructing a smart park digital twin based on big data, stored in the memory and capable of running on the processor. The algorithm and programs for constructing a smart park digital twin based on big data are steps for implementing the method for constructing a smart park digital twin based on big data as described above.

[0012] In addition, to achieve the above objectives, the present invention also provides a computer program product, which includes programs such as an algorithm for constructing a smart park digital twin based on big data. When the algorithm for constructing a smart park digital twin based on big data is executed by a processor, it implements the method for constructing a smart park digital twin based on big data as described above.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. Advantages in data acquisition and processing: By integrating and acquiring multi-source heterogeneous data, it covers various types of data inside and outside the park. Compared with the traditional single data acquisition method, it greatly enriches the data sources of the park's high-precision digital twin 3D model. Combined with advanced data preprocessing and deep analysis technology, it effectively improves data utilization efficiency and increases the mapping accuracy of the park's high-precision digital twin 3D model to the actual park by more than 30%. 2. Real-time performance and dynamic adaptability: The real-time data monitoring and instant update mechanism can update the high-precision digital twin 3D model of the park within 1 second after the data changes. Compared with the existing technology, the response speed is improved by 5 times. It can effectively cope with the complex and ever-changing environment and equipment operation status in the park, and ensure that the high-precision digital twin 3D model of the park always maintains a high degree of synchronization with the actual park. 3. Upgraded Management Decision Support: The human-computer interaction interface and continuously optimized high-precision digital twin 3D model of the park provide managers with more intuitive and accurate park operation information, improving decision-making efficiency by 40%. In scenarios such as equipment maintenance, energy management, and security monitoring, problems can be predicted in advance and solutions can be developed, reducing park operating costs and security risks. 4. Technology Integration and Scalability: It integrates multiple advanced technologies such as the Internet of Things, big data, and artificial intelligence to build a modular digital twin framework, which facilitates the connection with future new systems and equipment in the park. It has good scalability and can meet the long-term development needs of the smart park. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a flowchart of a method for constructing a smart park digital twin based on big data, according to the present invention.

[0016] Figure 2 This is a schematic diagram of the system structure for constructing a smart park digital twin based on big data, according to the present invention.

[0017] Figure 3 This is a schematic block diagram of an electronic device structure for constructing a smart park digital twin based on big data, according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] like Figure 1 As shown, in one embodiment of the present invention, a method for constructing a digital twin of a smart park based on big data includes the following steps: Step S10: Construct a multi-source heterogeneous data fusion and acquisition system through IoT technology, deploy IoT sensors in the smart park to collect data in an all-round and multi-level manner, and perform data preprocessing.

[0020] Specifically, step S10 involves constructing a multi-source heterogeneous data fusion and acquisition system using IoT technology, and deploying IoT sensors within the smart park to collect data comprehensively and at multiple levels. Data Collection Scenarios and Indicator Planning: Data collection needs are defined according to the functional zoning of the smart park, including closed areas, open areas, and boundary areas. Closed areas, such as offices and server rooms, will collect indicators such as temperature, humidity, carbon dioxide concentration, and equipment operating current and voltage. Collection ranges will be set, such as 0℃-50℃ for temperature, 0-5000ppm for carbon dioxide concentration, 0-50A for equipment operating current, and 0-380V for equipment operating voltage. Open areas, such as parking lots and main roads, will collect indicators such as vehicle information, personnel flow information, and ambient light. Vehicle information includes license plate number, vehicle type, and entry / exit time; personnel flow information includes personnel location and movement trajectory; and ambient light indicators will be 0-10000 lux. Boundary areas, such as walls and entrance / exit locations, will collect meteorological indicators such as wind speed and rainfall. Wind speed collection ranges will be set to 0-30m / s, and rainfall collection ranges to 0-50mm / h. Sensor Selection and Deployment: IoT sensors are selected based on scenario requirements. For environmental monitoring, temperature and humidity sensors and carbon dioxide sensors are used. The temperature and humidity sensors have an accuracy of ±0.5℃ and ±2%RH, while the carbon dioxide sensors have an accuracy of ±50ppm. Smoke sensors are additionally deployed in the server room, triggering an alarm with a response time of <30s upon smoke detection. For equipment monitoring, current and voltage sensors are used. The current sensor has an accuracy of 0.01A, and the voltage sensor has a range of 0-500V. For personnel and vehicle monitoring, 4K high-definition cameras (60fps), RFID readers (0-5m reading distance), and infrared beam sensors are used to prevent people from climbing over walls. Deployment density is based on functional importance, monitoring indoor environmental parameters in real time. Vibration sensors are deployed on critical facilities such as elevators, air conditioners, and power distribution equipment. In critical areas such as server rooms and power distribution rooms, cameras are deployed at 30-meter intervals; in general office areas, at 50-meter intervals; and along main roads, one set of cameras is deployed every 50 meters. Inside buildings, cameras are deployed every 30 meters in offices, corridors, and server rooms, according to functional zoning. Temperature and humidity sensors, carbon dioxide concentration sensors, motion sensors, current sensors, and voltage sensors are installed every 50 meters to obtain the physical parameters of equipment operation; 4K high-definition cameras and RFID readers are installed at parking lot entrances, main road intersections, and densely populated areas to collect information on vehicle and personnel movement. Sensor IoT Setup: ZigBee networking is used for temperature and humidity sensors, carbon dioxide sensors, current sensors, and voltage sensors. WIFI / Ethernet direct connection is used for 4K high-definition cameras, RFID readers, and infrared beam sensors. One IoT gateway is set up for every 5,000 square meters, supporting protocol conversion such as ZigBee to TCP / IP to achieve sensor data aggregation. AC220V power supply is used for fixed sensors, and lithium batteries combined with solar charging are used for mobile monitoring points such as temporary construction areas. The lithium battery has a battery life of ≥72 hours. Multi-source data access and integration: This involves building internal and external data interfaces. Internal data integration includes data exchange with the park's property management system, access control system, and energy management system via API interfaces. This includes obtaining equipment ledgers from the property management system, personnel IDs from the access control system, and electricity meter data from the energy management system, using JSON format for data exchange. External data integration includes connecting to the meteorological platform and transportation department APIs via HTTPS protocol. This includes obtaining hourly forecasts from the meteorological platform and surrounding road conditions from the transportation department API. Data for the next day is automatically synchronized at 3 AM daily. Each collection point is assigned a unique ID in the format of area code + device type + serial number, such as OF-TH-001 representing office temperature and humidity sensor 001, ensuring data traceability.

[0021] Specifically, the data preprocessing steps in step S10 include: Data cleaning: Using rule-based cleaning methods, thresholds are set for the range of values ​​for various data types to remove noisy data that exceeds the reasonable range. Clustering algorithms such as DBSCAN are used to identify and remove duplicate data, and data interpolation algorithms such as Lagrange interpolation are used to fill in missing data. Format conversion: Use ETL tools such as Kettle / Talend to convert data of different formats such as CSV, XML and JSON into Parquet format to improve data storage and processing efficiency; Normalization: Numerical data such as temperature, humidity, voltage, and current parameters are normalized using the Z-score standardization method to ensure that the mean of the processed data is 0 and the standard deviation is 1, thus eliminating the influence of dimensions and laying the foundation for subsequent data analysis. Non-numerical data, such as vehicle models and personnel information, are not normalized.

[0022] Step S20: Construct a distributed big data storage architecture to store the collected data, and combine machine learning and deep learning algorithms to extract deep features of the data.

[0023] Specifically, step S20, which involves constructing a distributed big data storage architecture to store the collected data and extracting deep features from the data using machine learning and deep learning algorithms, includes: Distributed big data storage architecture: A storage architecture combining Hadoop Distributed File System (HDFS) and distributed database (HBase) is adopted. The Hadoop Distributed File System is used to store massive amounts of unstructured and semi-structured data, such as video files and log files, while the distributed database is used to store structured data, such as equipment operating parameters and personnel information, to meet the storage needs of different types of data. A data index and metadata management system is established. A full-text search index is built using Apache Solr to facilitate fast data retrieval. Apache Atlas is used for metadata management to record information such as the source, processing process, and storage location of data, realizing full lifecycle management of data. Data tiering strategy setting: Based on the importance, update frequency and usage frequency of the data, a tiered storage strategy is set. High-frequency data such as real-time collected key equipment operation data and security monitoring data are stored in high-performance solid-state drives (SSDs) to ensure fast reading, while low-frequency data such as historical data and analysis results are stored in low-cost hard disk drives (HDDs) or cloud storage to reduce storage costs. Deep feature extraction: For image-type data, an improved YOLOv5 convolutional neural network model was built based on the TensorFlow framework. This model was trained and optimized on images captured by 4K high-definition cameras to achieve high-precision recognition of human behaviors such as loitering and running, and vehicle types such as cars, trucks, and electric vehicles. For time-series data, a hybrid neural network model combining the Transformer architecture and LSTM was used to analyze time-series data such as current and voltage of equipment operation, extract abnormal features of equipment operation, and predict the time and type of equipment failure. For text data, a BERT pre-trained model was used to extract features from text data such as equipment maintenance records and user feedback to uncover potential equipment problems and user needs.

[0024] Step S30: Construct a high-precision digital twin 3D model of the park and accurately associate the extracted data depth features with the model entities.

[0025] Specifically, step S30, which involves constructing a high-precision digital twin 3D model of the park and accurately associating the extracted data depth features with the model entities, includes: Construction of a high-precision digital twin 3D model of the park: Using professional 3D modeling software such as 3ds Max, Maya and Blender, combined with the park's architectural design drawings, BIM model and on-site survey data, a high-precision 3D geometric model of the park is constructed, and a local update mechanism is set up. During the modeling process, the exterior materials of buildings such as glass, stone and metal, the internal structure such as floor layout and room function, and the equipment form such as air conditioning units, power distribution cabinets and elevator cars are meticulously depicted to ensure the model's realism and detail. Visualization and rendering: Based on the Unity 3D or Unreal Engine game engine, a digital twin visualization platform is built. The constructed 3D geometric model is imported into the digital twin visualization platform, and shader technology is used to achieve realistic lighting and shadow effects and material rendering to restore the real visual effect of the park. Data Association: A data channel is established using C# or Python scripts to accurately associate the data features extracted in step S20 with the corresponding entities in the high-precision digital twin 3D model of the park. For example, the operating parameters of air conditioning equipment, such as temperature, wind speed, and energy consumption, are bound to the air conditioning units in the 3D model. When the actual operating parameters of the equipment change, the air conditioning units in the high-precision digital twin 3D model of the park will display the corresponding status changes in real time, such as color changes indicating faults and dynamic updates of operating parameters. The location information of personnel is associated with the personnel model in the high-precision digital twin 3D model of the park to realize the real-time dynamic trajectory display of personnel in the park.

[0026] Step S40: Set up a real-time data monitoring mechanism, use message queues and edge computing for high-throughput data transmission and fast processing, develop a multi-terminal human-computer interaction interface to periodically evaluate the 3D model of the park's digital twin based on accuracy, real-time performance and stability indicators, and use optimization algorithms to iteratively upgrade the 3D model of the park's digital twin.

[0027] Specifically, step S40 includes: Real-time data monitoring and updating: Message queues such as Kafka are used to achieve high-throughput data transmission and real-time distribution. Edge computing nodes are set up within the park to perform preliminary processing and filtering of the collected raw data, reducing data transmission volume and improving data processing efficiency. When the data collected by IoT sensors changes, the changed data is updated to the digital twin visualization platform in real time via the WebSocket protocol, triggering the local update mechanism of the park's high-precision digital twin 3D model to synchronize the data, ensuring that the park's high-precision digital twin 3D model and the actual park are synchronized within seconds. Interactive Interface Development: Develop a feature-rich and user-friendly human-computer interaction interface that supports access from multiple terminals including PCs, mobile devices, and large screens. Functions include remote monitoring, equipment control, and report generation. Remote monitoring is used for monitoring equipment status and personnel distribution. Equipment control is used for adjusting lighting and air conditioning. Report generation includes daily energy consumption reports and fault statistics. Managers can remotely monitor the park through the human-computer interaction interface, view the park's operational status (including equipment operating parameters, personnel and vehicle distribution, and environmental indicators) using a high-precision digital twin 3D model, and remotely control equipment (e.g., switching lighting on / off, adjusting air conditioning temperature, initiating elevator maintenance procedures). The data analysis unit generates various statistical reports and visualization charts, such as daily energy consumption reports and equipment fault frequency statistics, and visualization charts like heat maps showing densely populated areas and line graphs analyzing equipment operating trends, providing support for decision-making. A mechanism for evaluating and optimizing a high-precision digital twin 3D model of the park is established: The high-precision digital twin 3D model of the park is comprehensively evaluated regularly based on evaluation indicators such as accuracy, real-time performance, stability, and usability, such as weekly and monthly. Accuracy is evaluated by comparing the actual park status with the displayed status of the high-precision digital twin 3D model and calculating the error rate. Real-time performance is evaluated by the time interval between recording data changes and the completion of the update of the high-precision digital twin 3D model. Stability is evaluated by statistically analyzing the number of failures and runtime of the high-precision digital twin 3D model within a certain period. Usability is evaluated through user feedback and operation process analysis. Based on the evaluation results, model optimization algorithms, such as genetic algorithms and particle swarm optimization algorithms, are used to adjust the parameters of the machine learning model, optimize the rendering performance and data transmission efficiency of the high-precision digital twin 3D model, and iteratively upgrade the high-precision digital twin 3D model to continuously improve its performance and reliability.

[0028] In addition, such as Figure 2 As shown, in one embodiment of the present invention, a system for constructing a smart park digital twin based on big data is proposed. The system includes: Park Data Acquisition and Preprocessing Module: Used to build a multi-source heterogeneous data fusion acquisition system through IoT technology, deploy IoT sensors in the smart park to collect data in an all-round and multi-level manner, and perform data preprocessing; The park's data storage and feature extraction module is used to build a distributed big data storage architecture to store the collected data and to extract deep features of the data by combining machine learning and deep learning algorithms. Digital Twin Modeling and Data Association Module: Used to construct a high-precision 3D digital twin model of the park, and to accurately associate the extracted data depth features with the model entities; Real-time monitoring and optimization module: This module is used to set up a real-time data monitoring mechanism, use message queues and edge computing for high-throughput data transmission and fast processing, develop multi-terminal human-computer interaction interfaces to periodically evaluate the 3D model of the park's digital twin based on accuracy, real-time performance, and stability indicators, and use optimization algorithms to iteratively upgrade the 3D model of the park's digital twin.

[0029] This application provides a system for constructing a smart park digital twin based on big data. It employs a method for constructing a smart park digital twin based on big data as described in the above embodiments, and can solve the technical problems of incomplete data collection, low processing efficiency, and inability to adapt to dynamic data changes in traditional methods for constructing smart park digital twins based on big data. Compared with the prior art, the beneficial effects of the system for constructing a smart park digital twin based on big data provided in this application are the same as those of the method for constructing a smart park digital twin based on big data provided in the above embodiments. Furthermore, other technical features of the system for constructing a smart park digital twin based on big data are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0030] This application provides a device for constructing a smart park digital twin based on big data. The device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which are executed by the at least one processor to enable the at least one processor to execute the method for constructing a smart park digital twin based on big data as described in Embodiment 1 above.

[0031] like Figure 3 As shown in the illustration, in one embodiment of the present invention, a structural schematic diagram of a device suitable for implementing a big data-based smart park digital twin is presented. The device for constructing a big data-based smart park digital twin in this embodiment may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 3 The device for constructing a smart park digital twin based on big data shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments in this application.

[0032] Figure 3The illustrated device for constructing a smart park digital twin based on big data may include a processing system 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to programs stored in read-only memory (ROM) 1002 or programs loaded from storage system 1003 into random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the device. The processing system 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input systems 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output systems 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage systems 1003 including, for example, magnetic tapes, hard disks, etc.; and communication systems 1009. Communication system 1009 allows a data-driven smart park digital twin construction device to communicate wirelessly or wiredly with other devices to exchange data. Although the figure shows a data-driven smart park digital twin construction device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented alternatively.

[0033] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication system, or installed from storage system 1003, or installed from read-only memory 1002. When the computer program is executed by processing system 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0034] This application provides a device for constructing a smart park digital twin based on big data. Employing a method for constructing a smart park digital twin based on big data as described in the above embodiments, it solves the technical problems of incomplete data collection, low processing efficiency, and inability to adapt to dynamic data changes in traditional methods for constructing smart park digital twins based on big data. Compared with the prior art, the beneficial effects of the device for constructing a smart park digital twin based on big data provided in this application are the same as those of the method for constructing a smart park digital twin based on big data provided in the above embodiments. Furthermore, other technical features of this device for constructing a smart park digital twin based on big data are the same as those disclosed in the previous embodiment, and will not be repeated here.

[0035] The various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0036] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method for constructing a smart park digital twin based on big data.

[0037] The computer program product provided in this application can solve the technical problems of incomplete data collection, low processing efficiency, and inability to adapt to dynamic data changes in traditional big data-based smart park digital twin construction methods. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the big data-based smart park digital twin construction method provided in the above embodiments, and will not be repeated here.

[0038] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for constructing a digital twin of a smart park based on big data, characterized in that, The method includes the following steps: Step S10: Construct a multi-source heterogeneous data fusion and acquisition system through IoT technology, deploy IoT sensors in the smart park to collect data, and perform data preprocessing; Step S20: Construct a distributed big data storage architecture to store the collected data, and combine machine learning and deep learning algorithms to extract deep features of the data; Step S30: Construct a 3D model of the park's digital twin, and accurately associate the extracted data depth features with the model entities; Step S40: Set up a real-time data monitoring mechanism, use message queues and edge computing for data transmission and processing, develop a multi-terminal human-computer interaction interface to regularly evaluate the 3D model of the digital twin of the park, and use optimization algorithms to iteratively upgrade the 3D model of the digital twin of the park.

2. The method for constructing a smart park digital twin based on big data according to claim 1, characterized in that, The steps in step S10, which involve constructing a multi-source heterogeneous data fusion and acquisition system using IoT technology and deploying IoT sensors to collect data within the smart park, include: Data collection scenarios and indicator planning: According to the functional zoning of the smart park, the data collection needs are defined, including closed areas, open areas and boundary areas. Closed areas collect temperature, humidity, carbon dioxide concentration and equipment operating current and voltage indicators. Open areas collect vehicle information, personnel flow information and ambient light. Boundary areas collect wind speed and rainfall. Sensor selection and deployment: Select IoT sensors according to scenario requirements. For environmental monitoring, use temperature and humidity sensors and carbon dioxide sensors. For equipment monitoring, use current sensors and voltage sensors. For personnel and vehicle monitoring, use 4K cameras, RFID readers and infrared beam sensors. Deployment density is graded according to functional importance. Sensor IoT setup: ZigBee networking is used for temperature and humidity sensors, carbon dioxide sensors, current sensors and voltage sensors, while WIFI / Ethernet direct connection is used for 4K cameras, RFID readers and infrared beam sensors. One IoT gateway is set up for every 5000 square meters. Multi-source data access and integration: Building internal and external data interfaces. Internal data interfaces include data exchange with the park's property management system, access control system, and energy management system through API interfaces. External data interfaces include connecting to the meteorological platform and transportation department APIs through the HTTPS protocol, and assigning a unique ID to each collection point.

3. The method for constructing a smart park digital twin based on big data according to claim 1, characterized in that, The data preprocessing steps in step S10 include: Data cleaning: Using rule-based cleaning methods, thresholds are set for the range of various data values ​​to remove noisy data that exceeds the reasonable range, clustering algorithms are used to identify and remove duplicate data, and data interpolation algorithms are used to fill in missing data; Format conversion: Use ETL tools to convert data of different formats into Parquet format; Normalization: Numerical data are normalized using the Z-score standardization method so that the mean is 0 and the standard deviation is 1. Non-numerical data are not normalized.

4. The method for constructing a smart park digital twin based on big data according to claim 1, characterized in that, The steps in step S20, including constructing a distributed big data storage architecture to store the collected data and extracting deep features of the data using machine learning and deep learning algorithms, include: Distributed big data storage architecture construction: A storage architecture combining Hadoop Distributed File System and distributed database is adopted. Hadoop Distributed File System is used to store unstructured and semi-structured data, and distributed database is used to store structured data. A data index and metadata management system is established. Full-text search index is built through Apache Solr, and metadata management is carried out using Apache Atlas to record the source, processing process and storage location information of the data. Data tiering strategy setting: Based on the importance, update frequency and usage frequency of the data, a tiered storage strategy is set. Real-time collected key equipment operation data and security monitoring data are stored in solid-state drives (SSDs), while historical data and analysis results are stored in hard disk drives (HDDs) or cloud storage. Deep feature extraction: For image data, an improved YOLOv5 convolutional neural network model was built based on the TensorFlow framework to train and optimize images captured by 4K cameras, enabling the recognition of personnel behavior and vehicle types. For time-series data, a hybrid neural network model combining the Transformer architecture and LSTM was used to analyze the current and voltage of the equipment, extract abnormal features of equipment operation, and predict the time and type of equipment failure. For text data, a BERT pre-trained model was used to extract features from equipment maintenance records and user feedback to uncover potential equipment problems and user needs.

5. The method for constructing a smart park digital twin based on big data according to claim 1, characterized in that, The step S30, which involves constructing a 3D digital twin model of the park and accurately associating the extracted data depth features with the model entities, includes: Construction of the digital twin 3D model of the park: Using professional 3D modeling software, combined with the park's architectural design drawings, BIM model and on-site survey data, a 3D geometric model of the park is constructed, and a local update mechanism is set up; Visualization and rendering: Based on the Unity 3D or Unreal Engine game engine, build a digital twin visualization platform, import the constructed 3D geometric model into the digital twin visualization platform, and use shader technology to realize lighting effects and material rendering; Data association: Establish a data channel using C# or Python scripts to accurately associate the data features extracted in step S20 with the corresponding entities in the 3D model of the digital twin of the park.

6. The method for constructing a smart park digital twin based on big data according to claim 1, characterized in that, Step S40 includes: Real-time data monitoring and updating: Message queues are used to transmit and distribute data in real time. Edge computing nodes are set up in the park to perform preliminary processing and filtering of the collected raw data. When the data collected by the IoT sensors changes, the changed data is updated to the digital twin visualization platform in real time through the WebSocket protocol, triggering the local update mechanism of the park's digital twin 3D model to synchronize the data. Interactive Interface Development: Develop a human-computer interaction interface that supports access from PC, mobile, and large screen terminals. Functions include remote monitoring, equipment control, and report generation. Managers can remotely monitor the park through the human-computer interaction interface, view the park's operating status through the park's digital twin 3D model, remotely control equipment, and generate various statistical reports and visualization charts using the data analysis unit. Establish a mechanism for evaluating and optimizing the 3D model of the digital twin of the park: Regularly conduct comprehensive evaluations of the 3D model of the digital twin of the park based on accuracy, real-time performance, stability, and usability. Accuracy is evaluated by comparing the actual state of the park with the displayed state of the 3D model of the digital twin of the park and calculating the error rate. Real-time performance is evaluated by recording the time interval between data changes and the completion of the update of the 3D model of the digital twin of the park. Stability is evaluated by statistically analyzing the number of failures and runtime of the 3D model of the digital twin of the park within a certain period of time. Usability is evaluated by analyzing user feedback and operation process. Based on the evaluation results, model optimization algorithms are used to optimize the rendering performance and data transmission efficiency of the 3D model of the digital twin of the park, and the 3D model of the digital twin of the park is iteratively upgraded.

7. A system for constructing a digital twin of a smart park based on big data, characterized in that, The method for constructing a smart park digital twin based on big data, as described in claim 1, includes: Park Data Acquisition and Preprocessing Module: Used to build a multi-source heterogeneous data fusion acquisition system through IoT technology, deploy IoT sensors in the smart park to collect data, and perform data preprocessing; The park's data storage and feature extraction module is used to build a distributed big data storage architecture to store the collected data and to extract deep features of the data by combining machine learning and deep learning algorithms. Digital Twin Modeling and Data Association Module: Used to construct a 3D model of the park's digital twin, and accurately associate the extracted data depth features with the model entities; Real-time monitoring and optimization module: This module is used to set up a real-time data monitoring mechanism, use message queues and edge computing for data transmission and processing, develop a multi-terminal human-computer interaction interface to periodically evaluate the 3D model of the digital twin of the park, and use optimization algorithms to iteratively upgrade the 3D model of the digital twin of the park.

8. A device for constructing a digital twin of a smart park based on big data, characterized in that, include: The system includes a memory, a processor, and a program for constructing a smart park digital twin based on big data, which is stored in the memory and can run on the processor. When the program for constructing a smart park digital twin based on big data is executed by the processor, it implements a method for constructing a smart park digital twin based on big data as described in any one of claims 1 to 6.

9. A computer program product, characterized in that, The computer program product includes a big data-based smart park digital twin construction program, which, when executed by a processor, implements a big data-based smart park digital twin construction method as described in any one of claims 1 to 6.

Images