Method and system for generating digital twin model based on video monitoring

Abstract

The invention provides a method and a system for generating a digital twin model based on video monitoring. The method comprises the following steps: collecting physical space information; the collected physical space information is preprocessed; constructing a three-dimensional model by utilizing the processed data; performing space alignment on the monitoring video data and the three-dimensional model; fusing the real-time monitoring video data with the state information of the building model, and updating the actual change of the building model in real time; detecting and tracking an object in the video by using a machine learning model, and identifying an abnormal behavior and an event after analysis; and the identified information is processed, and a control instruction is generated according to a processing result to control corresponding equipment. The method has the beneficial effects that the object in the video is detected through the machine learning model, the abnormal behavior and event are identified, the control instruction is generated to control the related equipment, and the intelligent response capability and safety of the system are improved.

Description

Technical Field

[0001] The present invention belongs to the field of digital twin technology, and in particular, relates to a method and system for generating a digital twin model based on video surveillance. Background Art

[0002] With the development of the construction industry, digital twin models have received widespread attention as an emerging technology application. Through digital modeling and simulation technology, digital twin models map and synchronize the physical information of actual buildings with virtual models in real time to achieve full life cycle management and optimized operation of buildings. At the same time, the popularization and development of intelligent building systems have also provided rich data sources and intelligent control requirements for the application of building digital twin models.

[0003] At present, the field of digital twin technology has the following problems: different intelligent subsystems may use different data formats and communication protocols, which increases the complexity of system integration; video streams and digital twin models may be out of sync in time, resulting in inconsistency between monitoring data and model status; there may be errors in the spatial mapping of video monitoring images and models, affecting the monitoring effect and the response of intelligent subsystems. Therefore, a digital twin model construction method and system that can solve the above problems is urgently needed. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides a method and system for generating a digital twin model based on video surveillance, which is particularly suitable for scenarios that require efficient linkage between video surveillance and digital twin models to identify abnormal behaviors.

[0005] The technical solution adopted by the present invention is: In the first aspect, a method for generating a digital twin model based on video surveillance is provided, comprising:

[0006] Collect physical space information;

[0007] Preprocessing the collected physical space information;

[0008] Using the processed data to construct a three-dimensional model;

[0009] Spatial alignment of surveillance video data with the 3D model;

[0010] Integrate real-time surveillance video data with the status information of the building model to update the actual changes of the building model in real time;

[0011] Use machine learning models to detect and track objects in videos and analyze them to identify unusual behaviors and events;

[0012] Process the identified information and generate control instructions to control the corresponding equipment according to the processing results.

[0013] Furthermore, in spatially aligning the surveillance video data with the three-dimensional model, a feature matching algorithm is used to extract feature points in the video image, and the video image is aligned with the building model through geometric transformation.

[0014] Furthermore, collecting physical space information includes the following steps:

[0015] Collect geographic information and building internal environment information through video surveillance;

[0016] Collect architectural geometry information from design data;

[0017] Import device information into the device management system through IoT coding;

[0018] Get environmental information from sensors.

[0019] Furthermore, the spatial optimization search algorithm and the heuristic algorithm are used to optimize the way of collecting data through video surveillance.

[0020] Furthermore, constructing a three-dimensional model using the processed data includes the following steps:

[0021] Conduct multi-dimensional modeling of physical models, geometric models, behavioral models and rule models;

[0022] Couple the physical model with the geometric model;

[0023] The model generated by the equipment information is bound to the building model.

[0024] Furthermore, a spatial optimization search algorithm is used to optimize the method of collecting data through video surveillance.

[0025] In a second aspect, a system for generating a digital twin model based on video surveillance is provided, comprising:

[0026] An information collection module, used to collect physical space information;

[0027] An information preprocessing module, used for performing data preprocessing after receiving physical space information;

[0028] A model building module, for receiving the processed data and generating a three-dimensional model according to the data;

[0029] A fusion processing module is used to spatially align and fuse the surveillance video data with the three-dimensional model;

[0030] Intelligent analysis module, used to receive surveillance video data from the information collection module, identify and process abnormal behaviors and events after analysis;

[0031] The interactive display module is used to visualize the digital twin model and interact with the user.

[0032] Furthermore, the information collection module includes:

[0033] A video monitoring unit, used to collect geographic information and building internal environment information through video monitoring equipment;

[0034] A sensor unit, used for collecting environmental information through sensors;

[0035] A data storage unit for inputting architectural geometry information in design data;

[0036] The device information unit is used to import device information encoded by the Internet of Things.

[0037] Furthermore, the model building module includes:

[0038] A three-dimensional modeling unit, used to construct a three-dimensional model using physical space information;

[0039] Model fusion unit, used to couple the physical model and the geometric model;

[0040] The equipment binding unit is used to bind the equipment model to the building model.

[0041] Furthermore, the fusion processing module includes:

[0042] A spatial alignment unit, used for spatially aligning the surveillance video data with the three-dimensional model;

[0043] An information fusion unit, used to fuse real-time monitoring video data with state information of the building model;

[0044] The model updating unit is used to synchronously update the building model according to the changes of the real-time monitoring video data.

[0045] Furthermore, the intelligent analysis module includes:

[0046] An analysis unit, used to analyze surveillance video data and identify abnormal behaviors and events;

[0047] The processing unit is used to generate control instructions to control corresponding devices based on processing results.

[0048] The advantages and positive effects of the present invention are as follows: due to the adoption of the above-mentioned technical scheme, real-time monitoring of the physical space is realized, and the monitoring data can be spatially aligned and data fused with the digital twin model, and the status information of the model can be updated in real time to ensure the real-time and accuracy of the model; objects in the video are detected through machine learning models, abnormal behaviors and events are identified, and control instructions are generated to control related equipment, thereby improving the intelligent response capability and safety of the system; information from multiple sources such as video surveillance, sensor data, design materials, etc. is integrated, and preprocessed and standardized integration is performed to ensure the quality and consistency of the data. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 Schematic diagram of a method for generating a digital twin model according to an embodiment of the present invention

[0050] Figure 2 Schematic diagram of the information collection method according to an embodiment of the present invention

[0051] Figure 3 Schematic diagram of a three-dimensional model construction method according to an embodiment of the present invention

[0052] Figure 4 Schematic diagram of the structure of a system for generating a digital twin model according to an embodiment of the present invention

[0053] Figure 5 Schematic diagram of the structure of a video acquisition module according to an embodiment of the present invention

[0054] Figure 6 Schematic diagram of the structure of a model building module according to an embodiment of the present invention

[0055] Figure 7 is a structural diagram of a fusion processing module according to an embodiment of the present invention

[0056] Figure 8 This is a schematic diagram of the structure of an intelligent analysis module according to an embodiment of the present invention. DETAILED DESCRIPTION

[0057] The present disclosure is described more fully below with reference to the accompanying drawings, in which exemplary embodiments of the present disclosure are described. The technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0058] like Figure 1 As shown, the present invention provides a method for generating a digital twin model based on video monitoring, comprising:

[0059] S100, collecting physical space information;

[0060] S200, preprocessing the collected physical space information;

[0061] The collected physical space information is used to remove noise and outliers in the collected data to ensure the accuracy and completeness of the data; data from different sources are converted into a unified format for subsequent processing; video images are annotated to mark key areas and objects, including but not limited to entrances, exits and equipment locations; the contrast and brightness of video images are enhanced through image processing technology to improve image quality.

[0062] S300, constructing a three-dimensional model using the processed data;

[0063] S400, spatially aligning the surveillance video data with the three-dimensional model;

[0064] The digital twin model architecture of the relationship between the video surveillance system and the building is established, and its expression is:

[0065] M DT ={S PE ,S VM ,S c ,ST D, S SP}

[0066] Among them, S PE For monitoring entities, S VM is a virtual model, S C For connection, S TD is the twin database, S SP It provides operation and maintenance services for the intelligent linkage system. Connection is an intelligent transit station for communicating monitoring entities and digital twin models, which can help realize the conversion, transmission, connection interaction, and mapping drive between virtual and real models. The twin database is an important foundation for realizing digital twins. It integrates the parameters of the monitoring entity, the structure and functional parameters of the intelligent linkage system, and its response parameters for spatial linkage. The complete dynamic twin database is conducive to the two-way collection, transmission and feedback update between the system monitoring entity and the virtual model. The operation and maintenance service of the intelligent linkage system refers to the analysis and monitoring of changes in entity operating parameters through the model library and database of the virtual space based on the actual needs of the physical space, and then forming a responsive three-dimensional display.

[0067] In the physical space, static information such as the geometric shape and spatial position of the building and dynamic information of the monitoring entity are collected, and a data integration module for physical information perception is established. After the data collection and integration are completed, the standardized data information is uniformly transmitted to the mapping layer for analysis. In the virtual space, firstly, a multi-dimensional modeling based on physics-geometry-behavior-rules is performed on the monitoring entity obtained by the mapping layer. On this basis, the intelligent linkage system is divided into three levels: components, systems (buildings, monitoring entities) and global according to the progressive law of the virtual model function realization, and refined modeling is performed to form the construction and operation logic of the underlying digital twin three-dimensional model of the virtual space. Based on the above multi-dimensional modeling elements, a multi-dimensional coupled digital twin virtual model can be obtained, and its expression is:

[0068] M DT = {P vm ,G vm ,B vm ,R vm}

[0069] Among them, P vm is the physical model, G vm is the geometric model, B vm is the behavioral model, R vm The physical model and geometric model are mainly used to represent the geometric form, spatiotemporal relationship and physical properties of physical entities in virtual space. The modeling subject is the specific constraint object of the behavior and rule model. vm It is mainly used to map the existence state of physical entities in the real world. Its mathematical expression is:

[0070] Pvm = {P G ,P L ,P R ,P E}

[0071] Among them, P G is the geometric characteristics of the physical entity, P L is the spatial position of the physical entity, P r is the response of the physical entity during monitoring, P EIt is the environmental information around the physical entity. The digital twin model of the intelligent linkage system is the basis for realizing the digital mapping of the building scene, and at the same time, it establishes a mechanism to strengthen the communication and operation between the systems. In order to establish a global three-dimensional model, the intelligent linkage system is regarded as a complex spatial data set integrating multiple factors through a step-by-step modeling method. During the operation of the intelligent linkage system, metadata and ontology are used to realize the standardized integration and conversion of multi-source heterogeneous data. At the same time, a unified neural network mapping technology is selected to fuse the three sub-models of spatial classifier, time classifier and image classifier to establish a unified high-precision data management system; finally, the information coupling between the data source and the geometric model is realized through sensor modeling and equipment function definition. The establishment of the three-dimensional model uses BIM software as the main modeling tool, and adopts a combination of system family and loadable family to establish a device-level model according to the spatial position and semantic relationship in the drawing. At the functional implementation level, the digital twin model can be simplified into the following form, and its expression is:

[0072] M DT =S PE +S VM =P input ∪P inter ∪P output

[0073] Among them, M DT is the digital twin model, S PE is a physical entity, S VM is a virtual model, P input is the input parameter, P inter is the intermediate parameter, P output is the output parameter. According to the functional structure and application requirements of the digital twin, the formation process of the three sets of parameters and the corresponding functions achieved by their mutual cooperation are as follows: the data of the physical space entity model is collected and aggregated to form input parameters; the interactively obtained data can be classified and cleaned through the analysis and integration of the twin database to form intermediate parameters; after the three-dimensional modeling optimization of the virtual model and the analysis of the monitoring entity data, the collected data can be converted into the output parameters of the virtual space. Through the coupling and transmission of the data and information of the input parameters, intermediate parameters and output parameters, the cyclic extraction analysis and orderly mapping between the internal (virtual model) and external (physical entity) elements of the digital twin can be realized.

[0074] S500, integrating the real-time monitoring video data with the state information of the building model, and updating the actual changes of the building model in real time;

[0075] S600, using machine learning models to detect and track objects in the video, and identify abnormal behaviors and events after analysis;

[0076] S700: Process the identified information and generate a control instruction to control a corresponding device according to the processing result.

[0077] In one embodiment, in spatially aligning the monitoring video data with the three-dimensional model, a feature matching algorithm is used to extract feature points in the video image, and the video image is aligned with the building model through geometric transformation.

[0078] Use feature matching algorithms to extract feature points in video images to ensure the accuracy and robustness of feature points; align feature points in video images with corresponding points in building models through geometric transformations, including but not limited to affine transformations and perspective transformations, to achieve spatial alignment of video images and models; perform spatial correction based on the alignment results to ensure spatial consistency between the video image and the model.

[0079] In one embodiment, if Figure 2 As shown, the collection of physical space information includes the following steps:

[0080] S101, collecting geographic information and building internal environment information through video surveillance;

[0081] The video surveillance system is used to collect geographic information about the building and its surroundings, including but not limited to topography, landforms and road layout, and to photograph various areas inside the building, including but not limited to room layouts, corridors and stairs.

[0082] S102, collecting architectural geometry information from design data;

[0083] Collect architectural design drawings and documents, including but not limited to the building's geometric information, structural information, and material information.

[0084] S103, importing device information into the device management system by means of Internet of Things coding;

[0085] The device information imported into the device management system through IoT coding includes but is not limited to the location, type, model and status of the device.

[0086] S104: Acquire environmental information from a sensor.

[0087] Environmental information inside and outside the building is collected through a sensor network, including but not limited to temperature, humidity, light intensity and air quality.

[0088] In one embodiment, a spatial optimization search algorithm and a heuristic algorithm are used to optimize the method of collecting data through video surveillance.

[0089] Spatial optimization search algorithms include but are not limited to the A* algorithm, the Dijkstra algorithm, and the greedy algorithm. The A* algorithm is used to find the shortest path or the best solution. The total cost f(n) of node (n) is obtained by the equation f(n) = g(n) + h(n), where g(n) represents the actual cost from the starting point to node (n), and h(n) represents the heuristic estimated cost from node (n) to the target. The Dijkstra algorithm is used to calculate the single-source shortest path. The shortest path distance d(v) from the starting point to node (v) is obtained by the equation d(v) = mind(v) + w(u,v), where w(u,v) represents the edge weight from node (u) to node (v). The greedy algorithm is used to select the option that seems to be the best at the moment at each step. The estimated total cost f(n) of node (n) is obtained by the formula f(n) = h(n), and the heuristic function h(n) is directly used as the evaluation criterion. Optimize the layout of cameras and monitoring paths to improve monitoring efficiency; use heuristic algorithms to find the best viewing angle or area in the model to improve monitoring accuracy; dynamically adjust the monitoring area and camera parameters according to real-time data to optimize resource allocation. The objective function includes but is not limited to contrast loss, triple loss, and center loss. Contrast loss is used to minimize the distance between similar sample pairs and maximize the distance between dissimilar sample pairs. Get L ce , where D represents the Euclidean distance between sample pairs, y represents the label of the sample pair, and m represents the set distance threshold. The triplet loss is used to ensure that the distance between the anchor sample and the positive sample is smaller than the distance between the anchor sample and the negative sample, and the difference is greater than the set margin. t =max(0,d ap -d an +a) to get L t , where d ap Represents the distance between the anchor point and the positive sample, d an represents the distance between the anchor point and the negative sample, and a represents the margin. The center loss is used to improve the intra-class consistency by minimizing the distance between the sample feature and its category center. Get L cr , where f(x i ) represents the feature vector of sample (x_i), c yiRepresents the feature center of the category (y_i). Constraints include but are not limited to spatial consistency constraints, detection box accuracy constraints, computational complexity constraints, and category balance constraints. Spatial consistency constraints are used to ensure consistent building detection results at different perspectives or scales. This is done by aligning the building positions at different image scales or perspectives and using geometric transformations or affine transformations. Detection box accuracy constraints are used to ensure the regression accuracy of the detection box. Regression losses are used to optimize the position and size of the detection box to maximize its intersection over union (IoU) with the true box. Computational complexity constraints are used to limit the use of computing resources in the network to meet real-time requirements, control the number of parameters, computational complexity, or inference speed of the network, and achieve this by simplifying the model architecture or optimizing computational efficiency. Category balance constraints are used to ensure that the training and detection of samples of each category are balanced. Class imbalance problems are handled through weighted loss functions or data augmentation techniques.

[0090] In one embodiment, if Figure 3 As shown, constructing a three-dimensional model using the processed data includes the following steps:

[0091] S301, performing multi-dimensional modeling of physical model, geometric model, behavior model and rule model;

[0092] Based on the architectural design data and collected geographic information, construct a physical model of the building, including but not limited to walls, roofs, doors and windows; use 3D modeling software to create a geometric model of the building, including but not limited to precise dimensions and shapes; the behavioral model refers to defining the functions and usage of each area in the building, including but not limited to office areas, rest areas and equipment areas; the rule model refers to setting the behavioral rules of the building model, including but not limited to the switching time of equipment and the control logic of lights.

[0093] S302, coupling the physical model and the geometric model;

[0094] S303: Bind the model generated by the equipment information to the building model.

[0095] In one embodiment, a spatial optimization search algorithm is used to optimize the method of collecting data through video surveillance.

[0096] like Figure 4 As shown, in order to facilitate the execution of the method for generating a digital twin model provided by the present disclosure, the present disclosure also provides a system for generating a digital twin model based on video monitoring, including:

[0097] The information collection module 800 is used to collect physical space information;

[0098] The information preprocessing module 900 is used to perform data preprocessing after receiving the physical space information;

[0099] A model building module 110 is used to receive the processed data and generate a three-dimensional model based on the data;

[0100] A fusion processing module 120 is used to spatially align and fuse the surveillance video data with the three-dimensional model;

[0101] Intelligent analysis module 130, used to receive surveillance video data from information collection module 800, analyze and identify and process abnormal behaviors and events;

[0102] The interactive display module 140 is used to visualize the digital twin model and interact with the user.

[0103] like Figure 5 As shown, in one embodiment, the information collection module 800 includes:

[0104] The video monitoring unit 801 is used to collect geographic information and building internal environment information through video monitoring equipment;

[0105] The sensor unit 802 is used to collect environmental information through sensors;

[0106] The data storage unit 803 is used to input the building geometry information in the design data;

[0107] The device information unit 804 is used to import device information encoded in the Internet of Things.

[0108] In one embodiment, if Figure 6 As shown, the model building module 110 includes:

[0109] A three-dimensional modeling unit 111, used to construct a three-dimensional model using physical space information;

[0110] A model fusion unit 112, used for coupling the physical model and the geometric model;

[0111] The device binding unit 113 is used to bind the device model to the building model.

[0112] In one embodiment, if Figure 7 As shown, the fusion processing module 120 includes:

[0113] A spatial alignment unit 121, used for spatially aligning the surveillance video data with the three-dimensional model;

[0114] An information fusion unit 122, used to fuse the real-time monitoring video data with the state information of the building model;

[0115] The model updating unit 123 is used to synchronously update the building model according to the changes of the real-time monitoring video data.

[0116] In one embodiment, if Figure 8 As shown, the intelligent analysis module 130 includes:

[0117] An analysis unit 131 is used to analyze the surveillance video data and identify abnormal behaviors and events;

[0118] The processing unit 132 is used to generate control instructions to control corresponding devices based on the processing results.

[0119] Based on the embodiments of the present disclosure, the present disclosure may be implemented in the form of an electronic device, a readable storage medium, and a computer program product.

[0120] Electronic devices are intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframes, and other suitable computers. Electronic devices may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0121] Various embodiments of the present disclosure may be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0122] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0123] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0124] The embodiments of the present invention are described in detail above, but the contents described are only preferred embodiments of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made according to the scope of application of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A method for generating a digital twin model based on video surveillance, characterized in that: include: Collect physical space information; Preprocessing the collected physical space information; Using the processed data to construct a three-dimensional model; Spatial alignment of surveillance video data with the 3D model; Integrate real-time surveillance video data with the status information of the building model to update the actual changes of the building model in real time; Use machine learning models to detect and track objects in videos and analyze them to identify unusual behaviors and events; Process the identified information and generate control instructions to control the corresponding equipment according to the processing results.

2. The method for generating a digital twin model based on video surveillance according to claim 1, characterized in that: In the spatial alignment of surveillance video data with the three-dimensional model, a feature matching algorithm is used to extract feature points in the video image, and the video image is aligned with the building model through geometric transformation.

3. The method for generating a digital twin model based on video surveillance according to claim 1, characterized in that: The collection of physical space information includes the following steps: Collect geographic information and building internal environment information through video surveillance; Collect architectural geometry information from design data; Import device information into the device management system through IoT coding; Get environmental information from sensors.

4. The method for generating a digital twin model based on video surveillance according to claim 3 is characterized in that: The spatial optimization search algorithm and heuristic algorithm are used to optimize the method of collecting data through video surveillance.

5. The method for generating a digital twin model based on video surveillance according to claim 1, characterized in that: The construction of a 3D model using the processed data includes the following steps: Conduct multi-dimensional modeling of physical models, geometric models, behavioral models and rule models; Couple the physical model with the geometric model; The model generated by the equipment information is bound to the building model.

6. A system for generating a digital twin model based on video surveillance, characterized in that: include: An information collection module, used to collect physical space information; An information preprocessing module, used for performing data preprocessing after receiving physical space information; A model building module, for receiving the processed data and generating a three-dimensional model according to the data; A fusion processing module is used to spatially align and fuse the surveillance video data with the three-dimensional model; Intelligent analysis module, used to receive surveillance video data from the information collection module, identify and process abnormal behaviors and events after analysis; The interactive display module is used to visualize the digital twin model and interact with the user.

7. The system for generating a digital twin model based on video surveillance according to claim 6, characterized in that: The information collection module includes: A video monitoring unit, used to collect geographic information and building internal environment information through video monitoring equipment; A sensor unit, used for collecting environmental information through sensors; A data storage unit for inputting architectural geometry information in design data; The device information unit is used to import device information encoded by the Internet of Things.

8. The system for generating a digital twin model based on video surveillance according to claim 6, characterized in that: Model building modules include: A three-dimensional modeling unit, used to construct a three-dimensional model using physical space information; Model fusion unit, used to couple the physical model and the geometric model; The equipment binding unit is used to bind the equipment model to the building model.

9. The system for generating a digital twin model based on video surveillance according to claim 6, characterized in that: The fusion processing module includes: A spatial alignment unit, used for spatially aligning the surveillance video data with the three-dimensional model; An information fusion unit, used to fuse real-time monitoring video data with state information of the building model; The model updating unit is used to synchronously update the building model according to the changes of the real-time monitoring video data.

10. The system for generating a digital twin model based on video surveillance according to claim 6, characterized in that: Intelligent analysis modules include: An analysis unit, used to analyze surveillance video data and identify abnormal behaviors and events; The processing unit is used to generate control instructions to control corresponding devices based on processing results.

Images