An AI vision-based building intelligent monitoring system

Abstract

The application discloses a kind of based on AI vision's building intelligent monitoring system, belongs to building monitoring technical field, including acquisition end, edge end and cloud end;Acquisition end is used to interface the initial monitoring system in building;Real-time acquisition initial monitoring system collected building monitoring data;Edge end is used to pre-process building monitoring data;Cloud end includes monitoring demand module, AI vision module and display module;Monitoring demand module carries out monitoring demand analysis, real-time acquisition user's monitoring demand, according to monitoring demand set corresponding monitoring purpose;AI vision module is used to carry out AI vision monitoring, identifies the monitoring purpose of building monitoring data, according to monitoring purpose call AI vision model and analyze building monitoring data, obtain vision analysis result;Display module is used to carry out data display, set monitoring display interface, set corresponding access authority for monitoring display interface;Visual analysis result is input into monitoring display interface and is shown.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of building monitoring, and specifically relates to a building intelligent monitoring system based on AI vision. BACKGROUND

[0002] In the field of building security monitoring, traditional monitoring methods mainly rely on manual duty and fixed rule video analysis. With the rise of artificial intelligence technology, AI vision-based monitoring systems have gradually become mainstream. However, existing AI vision recognition technology still has significant limitations in building monitoring scenarios. The core problem lies in the rigidity of the recognition purpose. Existing AI vision monitoring systems are usually designed around specific targets, for example, intrusion recognition systems identify pre-set intrusion behaviors such as climbing over walls and damaging doors and windows by training models, and fire monitoring systems focus on image feature extraction of flames and smoke. The technical architecture of such systems is highly dependent on the recognition task defined in advance.

[0003] The rigidity of existing systems makes it difficult to adapt to the dynamic changes of building environments. The personnel activity patterns and spatial function layouts in building scenarios continue to evolve over time, and existing systems are difficult to adapt flexibly; for example, a building temporarily transforms part of the public area into an exhibition area during holidays. At this time, because the activity has flammable and explosive materials, new monitoring needs arise, and existing systems are difficult to achieve intelligent monitoring of this monitoring need, resulting in the need for traditional manual duty and other methods.

[0004] Based on this, in order to solve the above problems, the application provides a building intelligent monitoring system based on AI vision. SUMMARY

[0005] In order to solve the problems existing in the above scheme, the application provides a building intelligent monitoring system based on AI vision.

[0006] The purpose of the application can be achieved through the following technical solutions:

[0007] A building intelligent monitoring system based on AI vision, comprising a collection end, an edge end and a cloud end.

[0008] The collection end is used to interface with the initial monitoring system in the building; real-time acquisition of building monitoring data collected by the initial monitoring system; sending the building monitoring data to the edge end.

[0009] The edge end is used to pre-process the received building monitoring data, and send the pre-processed building monitoring data to the cloud end.

[0010] Further, the pre-processing refers to basic preprocessing, which includes denoising, compression, frame rate adjustment.

[0011] Further, the received building monitoring data is preprocessed, including:

[0012] The edge end is configured with a corresponding data deep processing model; the building monitoring data is subjected to basic preprocessing, and the building monitoring data after basic preprocessing is subjected to reprocessing through the data deep processing model, so as to complete the preprocessing of the building monitoring data.

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

[0014] The monitoring demand module performs monitoring demand analysis, and acquires the monitoring demand of a user in real time; according to the monitoring demand of the user, a corresponding monitoring purpose is set, and the monitoring purpose is sent to the AI vision module.

[0015] Further, the setting of the monitoring purpose includes:

[0016] A demand library is established, which is used to store various potential monitoring purposes corresponding to various buildings; building information corresponding to a user is identified, and a corresponding potential monitoring purpose is matched from the demand library according to the building information;

[0017] The monitoring demand of the user is identified, and each matched potential monitoring purpose is calibrated according to the monitoring demand; whether the corresponding potential monitoring purpose meets the monitoring demand is judged;

[0018] The potential monitoring purpose meeting the monitoring demand is displayed to the user; the user selects a monitoring purpose from the potential monitoring purpose.

[0019] Further, the platform party can eliminate various potential monitoring purposes stored in the demand library according to business needs.

[0020] Further, when the user has a new monitoring demand, a corresponding monitoring purpose is determined according to the new monitoring demand.

[0021] Further, the calibration of each matched potential monitoring purpose according to the monitoring demand includes:

[0022] A corresponding adaptive feature range is set for each potential monitoring purpose in the demand library;

[0023] The adaptive feature range corresponding to the matched potential monitoring purpose is identified; the adaptive feature range and the monitoring demand are integrated as input data and input into a preset calibration model for analysis, so as to obtain a calibration value of the corresponding potential monitoring purpose; the expression of the calibration model is:

[0024]

[0025] In the formula: (s, U) is input data, s represents monitoring requirements, and U represents an adaptive feature range of a corresponding potential monitoring purpose; output data is a calibration value HK(s, U), and the calibration value is 1 or 0;

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

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

[0028] The AI vision module is configured to perform AI vision monitoring, receive building monitoring data sent by each edge terminal, identify a monitoring purpose of the building monitoring data, call a corresponding AI vision model according to the monitoring purpose, analyze corresponding building monitoring data through the AI vision model, and obtain a vision analysis result of the corresponding monitoring purpose; and the vision analysis result is sent to a corresponding display module.

[0029] Further, identifying the monitoring purpose of the building monitoring data includes:

[0030] Receiving monitoring purposes corresponding to each user, generating a corresponding monitoring directory according to the monitoring purposes of each user, and dynamically updating the monitoring directory according to changes in the monitoring purposes of the users;

[0031] Identifying a user corresponding to received building monitoring data, and matching a corresponding monitoring purpose from the monitoring directory according to the user.

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

[0033] Real-time receiving of a vision analysis result of a corresponding user, and inputting the vision analysis result into the monitoring display interface for display.

[0034] Further, the cloud end further includes an edge analysis module.

[0035] The edge analysis module performs edge terminal analysis, determines data processing requirements of the edge terminal, establishes a deep processing model according to the data processing requirements, and arranges the deep processing model in the edge terminal.

[0036] Further, determining the data processing requirements of the edge terminal includes:

[0037] Accessing a requirement library, obtaining building information, matching corresponding potential monitoring purposes from the requirement library according to the building information, assuming that each potential monitoring purpose is a monitoring purpose for data processing simulation, and determining a to-be-selected processing requirement;

[0038] Prioritize the to-be-selected processing requirements, and take the to-be-selected processing requirement with the highest priority as the data processing requirement of the edge side.

[0039] Further, before prioritizing the to-be-selected processing requirements, the to-be-selected processing requirements are screened according to the edge side information.

[0040] Further, the prioritizing of the to-be-selected processing requirements comprises:

[0041] The to-be-selected processing requirements are simulated and analyzed to obtain the transmission efficiency corresponding to each to-be-selected processing requirement and the data analysis efficiency for each potential monitoring purpose;

[0042] The potential monitoring purposes are marked as i, i = 1, 2, …, n, and n is the number of potential monitoring purposes;

[0043] The dimension is removed, and the priority value of the corresponding to-be-selected processing requirement is calculated according to a preset priority value formula, and the priority value formula is:

[0044]

[0045] In the formula, QW is the priority value; b1 and b2 are both proportional coefficients, and the value range is 0 < b1 < 1 and 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] The to-be-selected processing requirements are sorted in descending order of priority value.

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

[0048] The building intelligent monitoring system based on AI vision of the present application effectively overcomes the problem of fixed identification purposes of existing AI vision recognition technology in building monitoring scenarios. The traditional system is designed around specific targets, and the technical architecture highly depends on the pre-defined identification tasks, which is difficult to adapt to the dynamic changes of the building environment. However, the system of the present application is no longer limited to the pre-set specific identification target, and can flexibly adjust the identification strategy according to the changes of the building environment and the emerging new monitoring requirements. For example, when part of the public area of the building is temporarily transformed into an exhibition area during holidays, and new monitoring requirements are generated due to the existence of flammable and explosive materials, the system of the present application can quickly adapt to and realize intelligent monitoring of this new scene without system modification or retraining of the model, greatly improving the adaptability and flexibility of the system to the changes of the building environment. BRIEF DESCRIPTION OF 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 an embodiment, the edge end pre-processes the received building monitoring data, which refers to basic preprocessing, so that the pre-processed building monitoring data has high universality and can be applied to various monitoring purposes for analysis and application; such as denoising, compression, frame rate adjustment, etc. The processed data still retains the complete information of the original video, only the storage and transmission efficiency is optimized; such data can be adapted to various monitoring purposes.

[0059] In an embodiment, in order to reduce the cloud analysis pressure and make full use of the edge resources, the basic preprocessing is further processed, while ensuring that the processed building monitoring data can be applied to various monitoring purposes that the building may have; the process is as follows:

[0060] The edge end is configured with a corresponding lightweight processing model, which is marked as a data deep processing model;

[0061] The building monitoring data after basic preprocessing is obtained; the building monitoring data is processed by the data deep processing model.

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

[0063] The monitoring demand module is used to set the corresponding monitoring purpose according to the user's monitoring demand, which can have multiple monitoring purposes at the same time, and sends the monitoring purpose to the AI vision module.

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

[0065] In an embodiment, for the special situation that the user changes the monitoring demand or adds the monitoring demand due to holding activities or changes in business purposes, in this embodiment, auxiliary analysis is performed according to the demand change situation to help the user set the monitoring purpose; the process is as follows:

[0066] Based on big data statistics or other statistical methods, a large amount of building historical monitoring data is analyzed to determine various potential monitoring purposes that various buildings have, or the platform party can directly set the potential monitoring purposes that various buildings have according to the business demand; various potential monitoring purposes corresponding to various buildings are summarized to establish a corresponding demand library;

[0067] Identify the corresponding building information of the user, such as office buildings, shopping malls, space layout and other related information; according to the building information, match various potential monitoring purposes that can be applied to the building from the demand library;

[0068] Real-time acquisition of user monitoring needs, such as avoiding high-altitude throwing, fire, stampede and other purposes, and adjusting and adding new monitoring needs at any time; according to the monitoring needs, the matched various potential monitoring purposes are calibrated to determine whether the corresponding potential monitoring purpose meets the monitoring needs, that is, whether it can solve the corresponding monitoring needs;

[0069] The potential monitoring purposes meeting the monitoring needs are displayed to the user; the user selects the monitoring purpose from the potential monitoring purposes.

[0070] In one embodiment, the platform party can exclude various buildings and potential monitoring purposes stored in the demand library according to business needs; for example, if the platform party believes that the market of this kind of building or monitoring purpose is small, it will not be included in the service range, then the corresponding building and monitoring purpose are excluded.

[0071] In one embodiment, the platform party can exclude various potential monitoring purposes stored in the demand library, which can be based on existing methods, such as common manual review exclusion, intelligent model review, etc.

[0072] In one embodiment, according to the monitoring needs, the matched various potential monitoring purposes are calibrated, which can use existing calibration models for calibration, such as establishing calibration models based on machine learning, deep learning algorithm, etc., using corresponding historical data to label the corresponding training set for training, the training set includes input data and output data, the input data is health needs and potential monitoring purposes, and the output data is calibration result.

[0073] In one embodiment, according to the monitoring needs, the matched various potential monitoring purposes are calibrated, including:

[0074] In the demand library, the corresponding adaptive feature range of each potential monitoring purpose is set, which is used to indicate which monitoring needs, monitoring problems, etc. the potential monitoring purpose is suitable for, which can be counted according to the historical monitoring records of the potential monitoring purpose, and the monitoring purposes, needs, problems, etc. that can be solved are summarized;

[0075] The adaptive feature range corresponding to the matched potential monitoring purpose is identified, the adaptive feature range and the monitoring needs are integrated as input data to the preset calibration model for analysis, and the calibration value of the corresponding potential monitoring purpose is obtained, and the expression of the calibration model is:

[0076]

[0077] In the formula: (s, U) is input data, s represents monitoring demand, and U represents the adaptive feature range of the corresponding potential monitoring purpose; s∈U indicates that the corresponding adaptive feature range includes the monitoring demand or part of the monitoring demand, that is, it can completely or partially solve the monitoring demand, and the training set is marked using corresponding historical data, such as a training set formed by monitoring demand, adaptive feature range, and calibration value; the output data is a calibration value HK(s, U), and the calibration value is 1 or 0;

[0078] When the calibration value is 0, the corresponding potential monitoring purpose does not meet the monitoring demand.

[0079] When the calibration value is 1, the corresponding potential monitoring purpose meets the monitoring demand.

[0080] The AI vision module is configured to perform AI vision monitoring, receive building monitoring data sent by each edge terminal, identify a monitoring purpose corresponding to the building monitoring data, call a corresponding AI vision model according to the monitoring purpose, analyze the building monitoring data through the AI vision model, and obtain a visual analysis result of the monitoring purpose; and send the visual analysis result to the display module.

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

[0082] Receiving the monitoring purpose of each user, generating a corresponding monitoring directory according to the monitoring purpose of each user, and counting the monitoring purpose of each user and corresponding edge terminal information, so as to facilitate identification of the corresponding user according to the received building monitoring data, and dynamic updating of the monitoring directory according to changes in the monitoring purpose of the user.

[0083] Identifying the user corresponding to the received building monitoring data, and matching the corresponding monitoring purpose of the user from the monitoring directory.

[0084] In one embodiment, calling the corresponding AI vision model according to the monitoring purpose includes:

[0085] Accessing a demand library, identifying various potential monitoring purposes corresponding to various buildings, establishing an AI vision model according to the potential monitoring purpose of the corresponding building, and using the AI vision model to analyze the corresponding building monitoring data to achieve the purpose of monitoring and analyzing the corresponding potential monitoring purpose, such as intrusion identification. The AI vision model corresponding to the potential monitoring purpose can analyze the received building monitoring data to determine the intrusion result. The AI vision model is established according to existing AI vision technology.

[0086] Subsequently, the corresponding AI vision model is called according to the monitoring purpose.

[0087] The display module is used for data display, identifying the monitoring purpose of each user, generating a monitoring display interface of the user according to the monitoring purpose, and setting the monitoring display interface according to the user's needs, which can meet the user's personalized needs, and subsequent corresponding supplements according to the visual analysis results, and setting the corresponding access rights for the monitoring display interface of the corresponding user, that is, only the user or the authorized person of the user can access, check, copy and other operations;

[0088] Real-time receiving of visual analysis results of the corresponding user, inputting the visual analysis results into the monitoring display interface for display.

[0089] In one embodiment, the cloud further includes an edge analysis module,

[0090] The edge analysis module performs edge analysis to determine the data processing requirements of the edge, which indicates the degree of processing of the building monitoring data and the formation of corresponding data such as motion trajectory, color distribution, edge contour, etc. A deep processing model is established according to the data processing requirements and arranged in the edge.

[0091] In one embodiment, the determination of the data processing requirements of the edge includes:

[0092] Obtaining building information, interfacing with a demand library, matching corresponding potential monitoring purposes from the demand library according to the building information, assuming that each potential monitoring purpose is a monitoring purpose for data processing simulation, that is, the preprocessed building monitoring data needs to fully meet the analysis requirements of the above monitoring purposes, determining the optional data processing requirements, and marking them as selected processing requirements, that is, the building monitoring data processed to meet the selected processing requirements can be used for analysis of all assumed monitoring purposes.

[0093] Prioritizing the selected processing requirements, and selecting the selected processing requirement with the highest priority as the data processing requirement of the edge.

[0094] In one embodiment, before prioritizing the selected processing requirements, the selected processing requirements are screened according to the edge information, and the selected processing requirements with higher resource requirements than the actual conditions of the edge are removed.

[0095] In one embodiment, the establishment of the lightweight processing model (data deep processing model) is based on the existing lightweight CNN model technology, which realizes the processing of building monitoring data to meet the data processing requirements.

[0096] In one embodiment, the prioritization of the selected processing requirements can be based on the comprehensive calculation of the priority value of the transmission efficiency, resource utilization, analysis efficiency and other parameters; or other existing priority evaluation methods can be used for evaluation.

[0097] For example, the simulation analysis on the to-be-selected processing requirement is performed to obtain the transmission efficiency corresponding to the to-be-selected processing requirement and the data analysis efficiency of each potential monitoring purpose; the data transmission efficiency and the analysis efficiency under the condition can be determined by combining statistical analysis with the corresponding historical data; the historical efficiency requirements of each potential monitoring purpose are counted, and the corresponding weight coefficient of each potential monitoring purpose is set according to each historical efficiency requirement;

[0098] The potential monitoring purpose is marked as i, i = 1, 2, …, n, n is the number of potential monitoring purposes, and i refers to the potential monitoring purpose matched by the building information;

[0099] The dimension is removed to obtain the numerical calculation, the priority value of the corresponding to-be-selected processing requirement is calculated according to the preset priority value formula, and the priority value formula is:

[0100]

[0101] In the formula, QW is the priority value; b1 and b2 are proportional coefficients, and the value range is 0 < b1 ≤ 1 and 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] The to-be-selected processing requirements are sorted in the order from high to low according to the priority values.

[0103] The above formulas are all dimensionless numerical calculations, the formula is obtained by software simulation of a large amount of data to obtain a formula closest to the actual situation, and the preset parameters and the preset threshold in the formula are set by the person skilled in the art according to the actual situation or obtained by a large amount of data simulation.

[0104] The above examples are only used to illustrate the technical method of the present application but not limit the present application, although the present application is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical method of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical method of the present application.

Claims

1. An AI vision-based intelligent building monitoring system, characterized in that, The system comprises a collection end, an edge end and a cloud end; The collection end is used for interfacing with an initial monitoring system in a building, acquiring building monitoring data collected by the initial monitoring system in real time, and sending the building monitoring data to the edge end; The edge end is used for preprocessing the received building monitoring data and sending the preprocessed building monitoring data to the cloud end; The cloud end comprises a monitoring demand module, an AI vision module and a display module; The monitoring demand module performs monitoring demand analysis, acquires monitoring demands of users in real time, sets corresponding monitoring purposes according to the monitoring demands of the users, and sends the monitoring purposes to the AI vision module; The AI vision module is used for AI vision monitoring, receives building monitoring data sent by each edge end, identifies monitoring purposes of the building monitoring data, calls corresponding AI vision models according to the monitoring purposes, analyzes corresponding building monitoring data through the AI vision models, obtains vision analysis results of the corresponding monitoring purposes, and sends the vision analysis results to corresponding display modules; The display module is used for data display, identifies monitoring purposes of each user, generates a monitoring display interface of the user according to the monitoring purposes, and sets corresponding access permissions for the monitoring display interface of the corresponding user; The vision analysis results of the corresponding user are received in real time, and the vision analysis results are input into the monitoring display interface for display; The setting of the monitoring purposes comprises: A demand library is established, which is used for storing various potential monitoring purposes corresponding to various buildings; building information corresponding to a user is identified, and corresponding potential monitoring purposes are matched from the demand library according to the building information; The monitoring demands of the user are identified, and each matched potential monitoring purpose is calibrated according to the monitoring demands to determine whether the corresponding potential monitoring purpose meets the monitoring demands; The potential monitoring purposes meeting the monitoring demands are displayed to the user, and the user selects a monitoring purpose from the potential monitoring purposes.

2. The AI vision-based intelligent building monitoring system according to claim 1, wherein, The cloud end further comprises an edge analysis module; The edge analysis module performs edge end analysis, determines data processing requirements of the edge end, establishes a deep processing model according to the data processing requirements, and arranges the deep processing model in the edge end. 3.The AI vision-based intelligent building monitoring system according to claim 2, characterized in that, The preprocessing of the received building monitoring data comprises: A corresponding data deep processing model is configured for the edge end; the building monitoring data is subjected to basic preprocessing, and the building monitoring data subjected to the basic preprocessing is subjected to reprocessing through the data deep processing model, thereby completing the preprocessing of the building monitoring data.

4. The AI vision-based intelligent building monitoring system according to claim 2, wherein, Determining the data processing requirements of the edge end comprises: The demand library is accessed to acquire building information, corresponding potential monitoring purposes are matched from the demand library according to the building information, each potential monitoring purpose is assumed to be a monitoring purpose for data processing simulation, and the selected processing requirements are determined; The selected processing requirements are prioritized, and the selected processing requirement with the highest priority is taken as the data processing requirement of the edge end.

5. The AI vision-based intelligent building monitoring system according to claim 4, wherein, Before the selected processing requirements are prioritized, the selected processing requirements are filtered according to the edge end information.

6. The AI vision-based intelligent building monitoring system according to claim 4, wherein, Prioritizing the selected processing requirements comprises: The simulation analysis is performed on the selected processing requirements, and transmission efficiency corresponding to the selected processing requirements and data analysis efficiency for each potential monitoring purpose are obtained; The potential monitoring purposes are marked as i, i=1, 2, …, n, and n is the number of the potential monitoring purposes; The dimension is removed, and the priority value of the selected processing requirement is calculated according to a preset priority value formula, and the priority value formula is: In the formula, QW is a priority value; b1 and b2 are both proportional coefficients, and the value range is 0 < b1 ≤ 1 and 0 < b2 ≤ 1; CL is a transmission efficiency; λ i represents a weight coefficient of a corresponding potential monitoring purpose; FL i represents an analysis efficiency of the corresponding potential monitoring purpose; The selected processing requirements are sorted according to the priority value from high to low.

7. The AI vision-based intelligent building monitoring system according to claim 1, wherein, The matched potential monitoring purposes are calibrated according to the monitoring requirements, including: The corresponding adaptive characteristic ranges of the matched potential monitoring purposes are identified, the adaptive characteristic ranges and the monitoring requirements are integrated as input data, and the input data are input into a preset calibration model for analysis to obtain a calibration value of the corresponding potential monitoring purpose, and an expression of the calibration model is: In the formula, (s, U) is the input data, s represents the monitoring requirement, and U represents the adaptive characteristic range 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 purpose does not meet the monitoring requirement; When the calibration value is 1, the corresponding potential monitoring purpose meets the monitoring requirement. The monitoring purposes of the building monitoring data are identified, including: 8.The AI vision-based intelligent building monitoring system according to claim 1, wherein, The monitoring purposes corresponding to each user are received, the monitoring directory is generated according to the monitoring purposes of each user, and the monitoring directory is dynamically updated according to the change of the monitoring purposes of the user; The user corresponding to the received building monitoring data is identified, and the corresponding monitoring purpose is matched from the monitoring directory according to the user. The platform can eliminate the various potential monitoring purposes stored in the demand library according to the business requirements. 9.The AI vision-based intelligent building monitoring system according to claim 8, wherein, ​

Images