Building intelligent monitoring system based on AI vision

Abstract

The invention, which belongs to the technical field of building monitoring, discloses an AI-vision-based intelligent building monitoring system comprising an acquisition end, an edge end and a cloud end. The acquisition end is used for being in butt joint with an initial monitoring system in a building; acquiring building monitoring data acquired by the initial monitoring system in real time; the edge end is used for preprocessing the building monitoring data; the cloud comprises a monitoring demand module, an AI vision module and a display module; the monitoring demand module analyzes the monitoring demand, obtains the monitoring demand of a user in real time, and sets a corresponding monitoring purpose according to the monitoring demand; the AI vision module is used for carrying out AI vision monitoring, identifying the monitoring purpose of the building monitoring data, calling an AI vision model according to the monitoring purpose to analyze the building monitoring data, and obtaining a vision analysis result; the display module is used for displaying data, setting a monitoring display interface and setting a corresponding access authority for the monitoring display interface; and inputting the visual analysis result into a monitoring display interface for display.

Description

Technical Field

[0001] This invention belongs to the field of building monitoring technology, specifically an AI vision-based intelligent building monitoring system. Background Technology

[0002] In the field of building security monitoring, traditional monitoring methods mainly rely on manual monitoring and video analysis based on fixed rules. With the rise of artificial intelligence technology, AI vision-based monitoring systems are gradually becoming mainstream. However, existing AI visual recognition technologies still have significant limitations in building monitoring scenarios, the core issue being the rigidity of the recognition objectives. Existing AI visual monitoring systems are typically designed around specific targets. For example, intrusion detection systems use trained models to identify pre-defined intrusion behaviors such as climbing over walls and breaking doors and windows, while fire monitoring systems focus on extracting image features of flames and smoke. The technical architecture of such systems is highly dependent on the pre-defined recognition tasks.

[0003] The rigid nature of existing systems makes them ill-suited to the dynamic changes in building environments. Human activity patterns and spatial layouts within buildings evolve over time, and existing systems struggle to adapt flexibly. For instance, if a building temporarily converts some public areas into exhibition and sales zones during holidays, the presence of flammable and explosive materials creates new monitoring needs. Existing systems struggle to intelligently monitor these needs, necessitating reliance on traditional manual monitoring methods.

[0004] In order to solve the above problems, the present invention provides an AI vision-based intelligent building monitoring system. Summary of the Invention

[0005] To address the problems of the aforementioned solutions, this invention provides an AI-based intelligent building monitoring system.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] An AI vision-based intelligent building monitoring system includes a data acquisition terminal, an edge terminal, and a cloud terminal.

[0008] The acquisition terminal is used to connect to the initial monitoring system within the building; to acquire building monitoring data collected by the initial monitoring system in real time; and to send the building monitoring data to the edge terminal.

[0009] The edge device is used to preprocess the received building monitoring data and send the preprocessed building monitoring data to the cloud.

[0010] Furthermore, the preprocessing refers to basic preprocessing, which includes noise reduction, compression, and frame rate adjustment.

[0011] Furthermore, the received building monitoring data undergoes preprocessing, including:

[0012] Configure the corresponding data deep processing model for the edge device; perform basic preprocessing on the building monitoring data, and then further process the preprocessed building monitoring data through the data deep processing model to complete the preprocessing of the building monitoring data.

[0013] The cloud platform includes a monitoring requirement module, an AI vision module, and a display module;

[0014] The monitoring requirement module performs monitoring requirement analysis, obtains the user's monitoring requirements in real time, sets the corresponding monitoring objectives based on the user's monitoring requirements, and sends the monitoring objectives to the AI ​​vision module.

[0015] Furthermore, the setting of monitoring objectives includes:

[0016] Establish a demand database to store various potential monitoring purposes for different buildings; identify the building information corresponding to the user, and match the corresponding potential monitoring purpose from the demand database based on the building information;

[0017] Identify the user's monitoring needs, calibrate each potential monitoring objective based on the monitoring needs, and determine whether the corresponding potential monitoring objectives meet the monitoring needs;

[0018] Potential monitoring objectives that meet the monitoring requirements are displayed to the user; the user then selects a monitoring objective from the potential objectives.

[0019] Furthermore, the platform can eliminate various potential monitoring objectives stored in the demand database based on operational needs.

[0020] Furthermore, when users have new monitoring needs, the corresponding monitoring objectives are determined based on these new needs.

[0021] Furthermore, the various potential monitoring objectives are calibrated according to monitoring requirements, including:

[0022] Set appropriate adaptive feature ranges for each potential monitoring objective in the requirements library;

[0023] Identify the adaptation feature range corresponding to the matching potential monitoring objectives, integrate the adaptation feature range and monitoring requirements into input data, and input them into a preset calibration model for analysis to obtain the calibration value for the corresponding potential monitoring objective. The expression of the calibration model is:

[0024]

[0025] In the formula: (s, U) are the input data, where s represents the monitoring requirement and U represents the range of adaptability characteristics of the corresponding potential monitoring purpose; the output data is the calibration value HK(s, U), and the calibration value is 1 or 0.

[0026] When the calibration value is 0, the corresponding potential monitoring objective does not meet the monitoring requirements;

[0027] When the calibration value is 1, the corresponding potential monitoring objective meets the monitoring requirements.

[0028] The AI ​​vision module is used for AI visual monitoring. It receives building monitoring data sent from various edge devices, identifies the monitoring purpose of the building monitoring data, calls the corresponding AI vision model according to the monitoring purpose, analyzes the corresponding building monitoring data through the AI ​​vision model, obtains the visual analysis result of the corresponding monitoring purpose, and sends the visual analysis result to the corresponding display module.

[0029] Furthermore, identify the monitoring purpose of building surveillance data, including:

[0030] Receive the monitoring objectives of each user, generate the corresponding monitoring directory based on the monitoring objectives of each user, and dynamically update the monitoring directory according to changes in the monitoring objectives of users.

[0031] Identify the user corresponding to the received building monitoring data, and match the corresponding monitoring purpose from the monitoring directory based on the user.

[0032] The display module is used to display data, identify the monitoring purpose of each user, generate the monitoring display interface of the user according to the monitoring purpose, and set the corresponding access permissions for the monitoring display interface of the corresponding user.

[0033] It receives visual analysis results from relevant users in real time and inputs these results into the monitoring display interface for display.

[0034] Furthermore, the cloud also includes an edge analytics module;

[0035] The edge analysis module performs edge analysis, determines the data processing requirements of the edge, establishes a deep processing model based on the data processing requirements, and deploys the deep processing model in the edge.

[0036] Furthermore, the data processing requirements at the edge are determined, including:

[0037] Connect to the demand database to obtain building information. Based on the building information, match the corresponding potential monitoring objectives from the demand database. Assuming each potential monitoring objective is a monitoring objective, perform data processing simulation to determine the available processing requirements.

[0038] Perform priority sorting on the candidate processing requirements, and use the candidate processing requirement with the highest priority as the data processing requirement of the edge side.

[0039] Further, before performing priority sorting on the candidate processing requirements, screen the candidate processing requirements according to the edge side information.

[0040] Further, performing priority sorting on the candidate processing requirements includes:

[0041] Perform simulation analysis on the candidate processing requirements to obtain the transmission efficiency corresponding to the corresponding candidate processing requirements and the data analysis efficiency for each potential monitoring purpose;

[0042] Mark the potential monitoring purpose as i, i = 1, 2,..., n, where n is the number of potential monitoring purposes;

[0043] Remove the dimension and take its numerical value for calculation. Calculate the priority value of the corresponding candidate processing requirement according to the preset priority value formula. The priority value formula is:

[0044]

[0045] In the formula: QW is the priority value; b1 and b2 are both proportionality coefficients, and the value range is 0 < b1 ≤ 1, 0 < b2 ≤ 1; CL is the transmission efficiency; λ i represents the weight coefficient of the corresponding potential monitoring purpose; FL i represents the analysis efficiency of the corresponding potential monitoring purpose;

[0046] Sort the candidate processing requirements in descending order according to the priority value.

[0047] Compared with the prior art, the beneficial effects of the present invention are:

[0048] The building intelligent monitoring system based on AI vision of the present invention effectively overcomes the problem of fixed recognition purposes of the existing AI vision recognition technology in the building monitoring scenario. The traditional system is designed around specific targets, and its technical architecture highly depends on the previously defined recognition tasks, and it is difficult to adapt to the dynamic changes of the building environment. However, the system of the present invention is no longer limited to the preset specific recognition targets, and can flexibly adjust the recognition strategy according to the changes in the building environment and the emerging new monitoring requirements. For example, when a building temporarily transforms some public areas into exhibition areas during holidays and new monitoring requirements arise due to the presence of inflammable and explosive items, the system of the present invention can quickly adapt and achieve intelligent monitoring of this new scenario without system transformation or retraining the model, greatly improving the adaptability and flexibility of the system to the changes in the building environment. Brief Description of the Drawings

[0049] 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.

[0050] Figure 1 This is a block diagram illustrating the principle of the present invention. Detailed Implementation

[0051] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. 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 of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0052] like Figure 1 As shown, an AI vision-based intelligent building monitoring system includes a data acquisition terminal, an edge terminal, and a cloud terminal.

[0053] The acquisition terminal is used to connect to the building's monitoring system. In order to distinguish it from the monitoring system of this invention, the monitoring system is marked as the initial monitoring system; it acquires the building monitoring data collected by the initial monitoring system in real time; and it sends the building monitoring data to the edge terminal.

[0054] In one embodiment, the initial monitoring system is a pre-installed video, infrared, temperature, and other related monitoring system in the building, avoiding the waste of user resources caused by redeploying the monitoring system.

[0055] In one embodiment, if the building does not have a monitoring system or the existing monitoring system is incomplete, monitoring equipment can be deployed according to existing monitoring technologies to achieve monitoring coverage of the building.

[0056] The edge device is used to preprocess the received building monitoring data and send the preprocessed building monitoring data to the cloud.

[0057] In one embodiment, the building monitoring data is encrypted before being sent to the cloud to ensure data security. Specifically, encryption can be performed on the user end, the edge end, or both. At the same time, in order to ensure that the building monitoring data cannot be viewed by the platform, existing methods such as identity authentication and authorization, the principle of least privilege, data anonymization, and homomorphic encryption can also be used to solve this problem.

[0058] In one embodiment, the edge device preprocesses the received building monitoring data, which refers to basic preprocessing. This preprocessed data has high versatility and can be used for analysis and applications for various monitoring purposes, such as noise reduction, compression, and frame rate adjustment. The processed data still retains the complete information of the original video, only optimizing storage and transmission efficiency. This type of data can be adapted to various monitoring purposes.

[0059] In one embodiment, to reduce cloud analytics load and fully utilize edge resources, deep processing is performed on top of basic preprocessing, while ensuring that the processed building monitoring data is applicable to various monitoring purposes the building may have; the process is as follows:

[0060] Configure the corresponding lightweight processing model for the edge device and mark the lightweight processing model as the data deep processing model;

[0061] Acquire building monitoring data after basic preprocessing; process the building monitoring data using a data deep processing model.

[0062] The cloud platform includes a monitoring requirement module, an AI vision module, and a display module;

[0063] The monitoring requirement module is used to set corresponding monitoring objectives according to the user's monitoring requirements. It can have multiple monitoring objectives at the same time and send the monitoring objectives to the AI ​​vision module.

[0064] In one embodiment, the monitoring objective can be set by the user or by other existing methods.

[0065] In one embodiment, when a user changes or adds monitoring requirements due to special circumstances such as organizing an event or a change in business objectives, this embodiment performs auxiliary analysis based on the changes in requirements to help the user set monitoring objectives; the process is as follows:

[0066] Based on big data statistics or other statistical methods, a large amount of historical building monitoring data is analyzed to determine the various potential monitoring purposes of different buildings. Alternatively, the platform can directly set the potential monitoring purposes of different buildings according to operational needs. The various potential monitoring purposes corresponding to different buildings are then compiled to establish a corresponding demand database.

[0067] Identify the corresponding building information for each user, such as office buildings, shopping malls, spatial layout, and other related information; based on the building information, match various potential monitoring purposes that may be applied to that building from the demand database;

[0068] The system acquires users' monitoring needs in real time, such as those for preventing objects from being thrown from heights, fires, stampedes, and other purposes. It also adjusts and adds new monitoring needs as needed. Based on the monitoring needs, it calibrates each potential monitoring purpose to determine whether the corresponding potential monitoring purpose meets the monitoring needs, i.e., whether it can help solve the corresponding monitoring needs.

[0069] Potential monitoring objectives that meet the monitoring requirements are displayed to the user; the user then selects a monitoring objective from the potential objectives.

[0070] In one embodiment, the platform can remove various buildings and potential monitoring targets stored in the demand database according to operational needs; if the platform believes that the market for such buildings or monitoring targets is small and not within the scope of services, then the corresponding buildings and monitoring targets will be removed.

[0071] In one embodiment, the platform can remove various potential monitoring objectives stored in the demand database. This removal can be done using existing methods, such as common manual review or intelligent model review.

[0072] In one embodiment, calibration is performed on each potential monitoring objective matched according to monitoring requirements. Existing calibration models can be used for calibration, such as those based on machine learning or deep learning algorithms. The corresponding training set is labeled with relevant historical data for training. The training set includes input data and output data. The input data are health requirements and potential monitoring objectives, and the output data is the calibration result.

[0073] In one embodiment, calibration is performed on each matching potential monitoring objective based on monitoring requirements, including:

[0074] In the requirement library, set the corresponding applicable feature range for each potential monitoring purpose. The applicable feature range is used to indicate which monitoring needs and monitoring problems the potential monitoring purpose is applicable to. It can be statistically analyzed based on the historical monitoring records of the potential monitoring purpose to summarize the monitoring purposes, needs, and problems that it can solve.

[0075] Identify the adaptation feature range corresponding to the matching potential monitoring objectives, integrate the adaptation feature range and monitoring requirements into input data, and input them into a preset calibration model for analysis to obtain the calibration value for the corresponding potential monitoring objective. The expression of the calibration model is:

[0076]

[0077] In the formula: (s, U) are the input data, s represents the monitoring requirement, and U represents the adaptive feature range of the corresponding potential monitoring purpose; s∈U means that the corresponding adaptive feature range includes the monitoring requirement or part of the monitoring requirement, that is, it can solve the monitoring requirement in whole or in part. The training is carried out using the corresponding historical data labeled training set, such as the training set formed by the monitoring requirement, the adaptive feature range and the calibration value; the output data is the calibration value HK(s, U), and the calibration value is 1 or 0.

[0078] When the calibration value is 0, the corresponding potential monitoring objective does not meet the monitoring requirements;

[0079] When the calibration value is 1, the corresponding potential monitoring objective meets the monitoring requirements.

[0080] The AI ​​vision module is used for AI visual monitoring. It receives building monitoring data sent from various edge devices, identifies the monitoring purpose corresponding to the building monitoring data, calls the corresponding AI vision model according to the monitoring purpose, analyzes the corresponding building monitoring data through the AI ​​vision model, obtains the visual analysis results of the corresponding monitoring purpose, and sends the visual analysis results to the display module.

[0081] In one embodiment, identifying the monitoring purpose of building monitoring data includes:

[0082] It receives the monitoring objectives of each user and generates corresponding monitoring directories based on the monitoring objectives of each user. This is used to collect statistics on the monitoring objectives of each user and the corresponding edge terminal information, so as to identify the corresponding user based on the received building monitoring data and to dynamically update the monitoring directory according to the changes in the user's monitoring objectives.

[0083] Identify the user corresponding to the received building monitoring data, and match the corresponding monitoring purpose from the monitoring directory based on the user.

[0084] In one embodiment, the appropriate AI vision model is invoked based on the monitoring purpose, including:

[0085] By connecting to the demand database, we identify the potential monitoring objectives for various buildings. Based on the potential monitoring objectives of each building, we establish an AI visual model. This AI visual model is used to analyze the monitoring data of the corresponding buildings and to monitor and analyze the potential monitoring objectives, such as intrusion detection. The AI ​​visual model corresponding to the potential monitoring objective can analyze the received building monitoring data and determine the intrusion result. This is done by establishing the model according to existing AI visual technologies.

[0086] The appropriate AI vision model will then be invoked based on the monitoring objective.

[0087] The display module is used to display data, identify the monitoring purpose of each user, generate the monitoring display interface for the user according to the monitoring purpose, and set the monitoring display interface according to the user's needs to meet the user's personalized needs. It can be supplemented according to the visual analysis results. The corresponding access permissions are set for the monitoring display interface of the corresponding user, that is, only the user or the person authorized by the user can access, view, copy and other operations.

[0088] It receives visual analysis results from relevant users in real time and inputs these results into the monitoring display interface for display.

[0089] In one embodiment, the cloud also includes an edge analytics module.

[0090] The edge analysis module performs edge analysis to determine the data processing requirements of the edge. The data processing requirements indicate the extent to which the building monitoring data will be processed and what corresponding data will be generated, such as feature data like motion trajectory, color distribution, and edge contour. Based on the data processing requirements, a depth processing model is established and deployed at the edge.

[0091] In one embodiment, determining the data processing requirements at the edge includes:

[0092] Obtain building information, connect to the demand database, match the corresponding potential monitoring objectives from the demand database based on the building information, and simulate data processing for each potential monitoring objective. That is, the pre-processed building monitoring data needs to fully meet the analysis requirements of the above monitoring objectives. Determine the optional data processing requirements and mark them as optional processing requirements. That is, after the building monitoring data is processed to meet the optional processing requirements, it can be used for the analysis of all assumed monitoring objectives.

[0093] Prioritize the pending processing requirements and select the highest priority pending processing requirement as the data processing requirement for that edge.

[0094] In one embodiment, before prioritizing the processing requirements, the processing requirements are screened based on edge information, and processing requirements that require more resources than the actual conditions at the edge are eliminated.

[0095] In one embodiment, the lightweight processing model (data deep processing model) is established based on existing lightweight CNN model technology to process building monitoring data into data that meets data processing requirements.

[0096] In one embodiment, prioritizing the processing requirements can be done by comprehensively calculating priority values ​​based on parameters such as transmission efficiency, resource utilization, and analysis efficiency; alternatively, other existing priority evaluation methods can be used.

[0097] Exemplarily, perform simulation analysis on the candidate processing requirements to obtain the transmission efficiency corresponding to the corresponding candidate processing requirements and the data analysis efficiency of each potential monitoring purpose; statistical analysis can be combined with the corresponding historical data to determine the data transmission efficiency and analysis efficiency under this condition; count the historical efficiency requirements of each potential monitoring purpose, and set corresponding weight coefficients for each potential monitoring purpose according to each historical efficiency requirement.

[0098] Mark the potential monitoring purpose as i, where i = 1, 2,..., n, and n is the number of potential monitoring purposes, referring to the potential monitoring purposes matched by the building information.

[0099] Remove the dimension and take its numerical value for calculation, and calculate the priority value of the corresponding candidate processing requirement according to the preset priority value formula. The priority value formula is:

[0100]

[0101] In the formula: QW is the priority value; b1 and b2 are both proportionality coefficients, and the value range is 0 < b1 ≤ 1, 0 < b2 ≤ 1; CL is the transmission efficiency; λ i represents the weight coefficient of the corresponding potential monitoring purpose; FL i represents the analysis efficiency of the corresponding potential monitoring purpose.

[0102] Sort the candidate processing requirements in descending order of the priority value.

[0103] The above formulas are all calculated by removing the dimension and taking its numerical value. The formula is obtained by collecting a large amount of data for software simulation to get a formula closest to the actual situation. The preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained by simulating a large amount of data.

[0104] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A building intelligent monitoring system based on AI vision, characterized in that, This includes the data acquisition terminal, the edge terminal, and the cloud. The acquisition terminal is used to connect to the initial monitoring system within the building; to acquire building monitoring data collected by the initial monitoring system in real time; and to send the building monitoring data to the edge terminal. The edge device is used to preprocess the received building monitoring data and send the preprocessed building monitoring data to the cloud. The cloud platform includes a monitoring requirement module, an AI vision module, and a display module; The monitoring requirement module performs monitoring requirement analysis, obtains the user's monitoring requirements in real time, sets the corresponding monitoring objectives according to the user's monitoring requirements, and sends the monitoring objectives to the AI ​​vision module. The AI ​​vision module is used for AI visual monitoring. It receives building monitoring data sent from various edge devices, identifies the monitoring purpose of the building monitoring data, calls the corresponding AI vision model according to the monitoring purpose, analyzes the corresponding building monitoring data through the AI ​​vision model, obtains the visual analysis result of the corresponding monitoring purpose, and sends the visual analysis result to the corresponding display module. The display module is used to display data, identify the monitoring purpose of each user, generate the monitoring display interface of the user according to the monitoring purpose, and set the corresponding access permissions for the monitoring display interface of the corresponding user. It receives visual analysis results from relevant users in real time and inputs these results into the monitoring display interface for display.

2. The building intelligent monitoring system based on AI vision according to claim 1, characterized in that, The cloud also includes an edge analytics module; The edge analysis module performs edge analysis, determines the data processing requirements of the edge, establishes a deep processing model based on the data processing requirements, and deploys the deep processing model in the edge.

3. The building intelligent monitoring system based on AI vision according to claim 2, characterized in that, The received building monitoring data is preprocessed, including: Configure the corresponding data deep processing model for the edge device; perform basic preprocessing on the building monitoring data, and then further process the preprocessed building monitoring data through the data deep processing model to complete the preprocessing of the building monitoring data.

4. The building intelligent monitoring system based on AI vision according to claim 2, characterized in that, Determine the data processing requirements at the edge, including: Connect to the demand database to obtain building information. Based on the building information, match the corresponding potential monitoring objectives from the demand database. Assuming each potential monitoring objective is a monitoring objective, perform data processing simulation to determine the available processing requirements. The pending processing requirements are prioritized, and the highest priority pending processing requirement is taken as the data processing requirement of the edge end.

5. A building intelligent monitoring system based on AI vision according to claim 4, characterized in that, Before prioritizing the processing requirements, the requirements are filtered based on the edge information.

6. The building intelligent monitoring system based on AI vision according to claim 4, characterized in that, Prioritize the processing requirements to be selected, including: Simulation analysis is performed on the processing requirements to obtain the corresponding transmission efficiency and data analysis efficiency for each potential monitoring objective. The potential surveillance objectives are labeled as i, i = 1, 2, ..., n, where n is the number of potential surveillance objectives; Dimensions are removed and numerical values ​​are used for calculation. The priority value of the corresponding candidate processing requirement is calculated according to a preset priority value formula, which is: Where: QW is the priority value; b1 and b2 are both proportionality coefficients, and the value ranges are 0 < b1 ≤ 1, 0 < b2 ≤ 1; CL is the transmission efficiency; λ i represents the weight coefficient for the corresponding potential monitoring purpose; FL i represents the analysis efficiency for the corresponding potential monitoring purpose; The pending processing requests are sorted in descending order of priority.

7. The building intelligent monitoring system based on AI vision according to claim 1, characterized in that, The monitoring objectives include: Establish a demand database to store various potential monitoring purposes for different buildings; identify the building information corresponding to the user, and match the corresponding potential monitoring purpose from the demand database based on the building information; Identify the user's monitoring needs, calibrate each potential monitoring objective based on the monitoring needs, and determine whether the corresponding potential monitoring objectives meet the monitoring needs; Potential monitoring objectives that meet the monitoring requirements are displayed to the user; the user then selects a monitoring objective from the potential objectives.

8. The building intelligent monitoring system based on AI vision according to claim 7, characterized in that, Calibrate each potential monitoring objective according to the monitoring requirements, including: Set appropriate adaptive feature ranges for each potential monitoring objective in the requirements library; Identify the adaptation feature range corresponding to the matching potential monitoring objectives, integrate the adaptation feature range and monitoring requirements into input data, and input them into a preset calibration model for analysis to obtain the calibration value for the corresponding potential monitoring objective. The expression of the calibration model is: In the formula: (s, U) are the input data, where s represents the monitoring requirement and U represents the range of adaptability characteristics of the corresponding potential monitoring purpose; the output data is the calibration value HK(s, U), and the calibration value is 1 or 0. When the calibration value is 0, the corresponding potential monitoring objective does not meet the monitoring requirements; When the calibration value is 1, the corresponding potential monitoring objective meets the monitoring requirements.

9. A building intelligent monitoring system based on AI vision according to claim 1, characterized in that, The purpose of identifying building surveillance data includes: Receive the monitoring objectives of each user, generate the corresponding monitoring directory based on the monitoring objectives of each user, and dynamically update the monitoring directory according to changes in the monitoring objectives of users. Identify the user corresponding to the received building monitoring data, and match the corresponding monitoring purpose from the monitoring directory based on the user.

10. A building intelligent monitoring system based on AI vision according to claim 7, characterized in that, The platform can eliminate various potential monitoring purposes stored in the demand database based on operational needs.

Images