Internet of Things-based security monitoring methods and systems

Abstract

This invention relates to the technical field of security monitoring, and discloses a security monitoring method and system based on the Internet of Things. The security monitoring system monitors a designated area to obtain raw monitoring data from each monitoring node within the system. Based on a local terminal, the raw monitoring data from each monitoring node is aggregated to obtain overall monitoring data for the designated area. An AI perception model deployed on the local terminal analyzes the overall monitoring data to generate surface-level risk perception information for the designated area. The overall monitoring data is then uploaded to a cloud server to identify temporal characteristics of the overall monitoring data at different times, thereby obtaining potential-level risk perception information for the designated area.

Description

Technical Field

[0001] This invention relates to the technical field of security monitoring, and in particular to a security monitoring method and system based on the Internet of Things. Background Technology

[0002] Work safety is of paramount importance in factories. Currently, IoT technology enables the fusion and processing of large amounts of data from various types of sensors. By integrating data from image sensors, smoke sensors, temperature and humidity sensors, a more comprehensive understanding of the safety status of a designated area can be achieved. However, with the increase in monitoring units and the expansion of monitoring projects, traditional intelligent analysis methods are inefficient and prone to omissions when faced with massive amounts of monitoring data. Summary of the Invention

[0003] The purpose of this invention is to provide a security monitoring method and system based on the Internet of Things, aiming to solve the problem of low efficiency in analyzing large amounts of security monitoring data in the prior art.

[0004] The present invention is implemented as follows: Firstly, the present invention provides a security monitoring method based on the Internet of Things, comprising: Information is monitored in a designated area through a security monitoring system to obtain the raw monitoring data of each monitoring node within the system; The original monitoring data of each monitoring node is summarized based on the local terminal to obtain the overall monitoring data of the specified area. The overall monitoring data is analyzed by an AI perception model deployed on a local terminal to generate risk perception information at the surface level of the specified area. The overall monitoring data is uploaded to a cloud server to identify the temporal characteristics of the overall monitoring data at each time point, thereby obtaining risk perception information of the potential level of the specified area.

[0005] Secondly, the present invention provides an Internet of Things (IoT)-based security monitoring system for implementing the IoT-based security monitoring method described in any one of the first aspects, comprising: The data monitoring module is used to monitor information in a designated area through the security monitoring system in order to obtain the raw monitoring data of each monitoring node in the system. The data aggregation module is used to aggregate the raw monitoring data of each monitoring node based on the local terminal to obtain the overall monitoring data of the specified area. The surface perception module is used to analyze the overall monitoring data through an AI perception model deployed on a local terminal, and generate risk perception information of the surface layer of the specified area. The potential perception module is used to upload the overall monitoring data to the cloud server to identify the temporal characteristics of the overall monitoring data at each time point, and obtain the potential level risk perception information of the specified area.

[0006] This invention provides a security monitoring method based on the Internet of Things, which has the following beneficial effects: This invention leverages a security monitoring system to comprehensively acquire raw data, providing a foundation for subsequent analysis. Local terminals aggregate data and generate surface risk information through AI perception models, enabling timely detection of obvious hidden dangers. Data is then uploaded to a cloud server, utilizing its powerful computing capabilities to identify temporal characteristics and uncover potential risks. This combination of local and cloud-based approaches ensures both real-time performance and in-depth data analysis, achieving comprehensive monitoring of the surface and potential risks in a designated area and enhancing the level of security. Attached Figure Description

[0007] Figure 1 This is a schematic diagram illustrating the steps of a security monitoring method based on the Internet of Things provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of a security monitoring system based on the Internet of Things provided in an embodiment of the present invention. Detailed Implementation

[0008] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0009] The implementation of the present invention will be described in detail below with reference to specific embodiments.

[0010] Reference Figure 1 , Figure 2 The diagram shows a preferred embodiment of the present invention.

[0011] In a first aspect, the present invention provides a security monitoring method based on the Internet of Things, comprising: S1: Use a security monitoring system to monitor information in a designated area to obtain raw monitoring data from each monitoring node within the system; S2: Based on the local terminal, the original monitoring data of each monitoring node is summarized to obtain the overall monitoring data of the specified area; S3: Analyze the overall monitoring data using an AI perception model deployed on a local terminal to generate risk perception information at the surface level of the specified area; S4: Upload the overall monitoring data to the cloud server to identify the temporal characteristics of the overall monitoring data at each time point, and obtain the potential level of risk perception information for the specified area.

[0012] Specifically, in step S1 of the embodiment provided by the present invention, the monitoring nodes in the security monitoring system include image sensors, smoke sensors, temperature and humidity sensors, door and window electromagnetic sensors, etc. According to the characteristics of the designated area and the monitoring requirements, appropriate types of sensors are selected. For example, in a warehouse, smoke sensors and temperature and humidity sensors need to be deployed in order to monitor fire hazards and the storage environment of goods; in an office area, in addition to temperature and humidity monitoring, image sensors are also needed to monitor personnel activities and door and window electromagnetic sensors are needed to monitor the opening and closing status of doors and windows.

[0013] More specifically, based on the layout of the designated area and the monitoring objectives, the locations of each monitoring node should be arranged reasonably. For example, image sensors should be installed in locations that can cover key areas and passages to ensure comprehensive capture of the activities of people and objects; smoke sensors are usually installed on the ceiling to detect smoke in a timely manner; temperature and humidity sensors should be placed in locations that can reflect the overall temperature and humidity of the area; and door and window electromagnetic sensors are installed on the edges of doors and windows to detect the opening and closing status of doors and windows.

[0014] More specifically, different types of monitoring nodes can collect information from a designated area from multiple dimensions. Image sensors can provide intuitive visual information, helping monitoring personnel understand the activities of people and the status of objects in the area; smoke sensors can detect fire hazards in a timely manner; temperature and humidity sensors can monitor environmental conditions to ensure that equipment and items in the area are stored in a suitable environment; and door and window electromagnetic sensors can prevent illegal intrusion. By comprehensively utilizing multiple sensors, comprehensive monitoring of a designated area can be achieved, improving the level of security. Reasonable planning of the location of monitoring nodes can ensure effective coverage of key parts and important passages in the designated area, thus avoiding monitoring blind spots and timely detection of potential security risks.

[0015] More specifically, each monitoring node periodically collects data at preset time intervals. For example, an image sensor can collect one frame of image per second, a smoke sensor detects smoke concentration every 10 seconds, a temperature and humidity sensor records temperature and humidity data every minute, and a door and window electromagnetic sensor monitors the opening and closing status of doors and windows in real time and records the time points when the data changes.

[0016] More specifically, after the monitoring node collects the raw monitoring data, it transmits the data to the local terminal in real time via wired or wireless communication. Common wireless communication methods include Wi-Fi, ZigBee, Bluetooth, etc., while wired communication methods include Ethernet, etc.

[0017] More specifically, periodic data collection ensures data continuity and timeliness, enabling monitoring personnel to promptly understand dynamic changes within the area. Real-time data transmission allows local terminals to obtain the latest monitoring information in a timely manner for subsequent data processing and analysis. For example, in the event of a fire, smoke sensors can quickly detect smoke and transmit the data to the local terminal in real time, thereby issuing an alarm in a timely manner and reducing losses.

[0018] More specifically, raw monitoring data forms the basis for subsequent data aggregation, analysis, and risk assessment. Accurate and comprehensive raw data can provide a reliable basis for subsequent processing steps, ensuring that the generated risk perception information is accurate and effective. If the raw data collection is incomplete or inaccurate, it will lead to deviations in the subsequent analysis results and affect the effectiveness of safety monitoring.

[0019] Specifically, in step S2 of the embodiment provided by the present invention, the local terminal receives raw monitoring data transmitted in real time from each monitoring node through a pre-established communication connection. For example, if the monitoring node uses Wi-Fi communication, the local terminal will turn on the corresponding Wi-Fi receiving module, listen to a specific port, and wait to receive data. According to the setting location and monitoring form of the monitoring node, the local terminal adds a data source mark to the received raw monitoring data. For example, data from the temperature and humidity sensor located in the southeast corner of the warehouse is marked as "Southeast corner of the warehouse - temperature and humidity sensor"; data from the door and window electromagnetic sensor at the first-floor entrance of the office building is marked as "First-floor entrance of the office building - door and window electromagnetic sensor", thereby obtaining the node monitoring data of each monitoring node.

[0020] More specifically, marking the source of raw monitoring data helps with subsequent data management and analysis. When it is necessary to view data from a specific location or a specific type of monitoring node, the required data can be located quickly and accurately. For example, when investigating abnormal temperature and humidity in a certain area, the data marked by the temperature and humidity sensors in that area can be directly filtered for analysis. Clarifying the data source can increase the reliability of the data. When data anomalies occur or the accuracy of the data needs to be verified, the specific monitoring node can be traced back to the marking to check whether the node is working properly.

[0021] More specifically, based on the location of each monitoring node within the security monitoring system, the relative positional relationships between them are analyzed. For example, in a large shopping mall monitoring system, the distance and orientation relationships between image sensors, smoke sensors, etc., at different locations on different floors are analyzed. Based on the relative positioning relationships, the interaction value analysis of the monitoring data obtained by each monitoring node is performed. Taking a shopping mall as an example, the data from adjacent image sensors and smoke sensors have high interaction value because the information on personnel activities captured by the image sensors is correlated with the information on fire hazards detected by the smoke sensors. By analyzing this interaction value, the monitoring interaction characteristics of each monitoring node are obtained.

[0022] More specifically, based on the monitoring interaction characteristics of each monitoring node, the feasibility of combining them for monitoring is evaluated. For example, if the data of two adjacent monitoring nodes are strongly correlated and can provide more valuable information after combination, such as the combination of adjacent temperature and humidity sensors and image sensors can analyze the impact of environmental temperature and humidity on human activities, then they are considered to have the feasibility of monitoring combination, and thus a data combination pattern between each monitoring node is generated.

[0023] More specifically, by analyzing the relative positioning relationships of monitoring nodes and the interactive value of monitoring data, potential correlations between data from different monitoring nodes can be discovered. For example, in a hospital monitoring system, there is a correlation between temperature and humidity sensor data near the operating room and equipment operating status data inside the operating room. By exploring this correlation, the smooth progress of surgery can be better guaranteed. By rationally combining monitoring nodes, scattered data can be integrated into more valuable information. For example, combining data from multiple adjacent image sensors can form a larger monitoring screen, providing a more comprehensive understanding of the activities of people and objects, and improving the monitoring effect.

[0024] More specifically, based on the location and monitoring method of each monitoring node within the security monitoring system, a pre-determined data combination pattern is retrieved for the monitoring data of each node. The monitoring data of each node is then combined in multiple ways according to the retrieved data combination pattern. For example, in the monitoring system of an industrial park, the data from image sensors, smoke sensors, and temperature and humidity sensors in the same workshop are combined to obtain the area monitoring data of that workshop. The area monitoring data of multiple workshops are further combined to obtain the area monitoring data of different areas of the entire industrial park. The area monitoring data of each level are then used together as the overall monitoring data of the designated area.

[0025] More specifically, combining regional monitoring data from different levels into overall monitoring data for a designated area can provide monitoring personnel with a comprehensive and integrated monitoring perspective. For example, in a city's traffic monitoring system, combining data such as traffic flow, vehicle speed, and traffic light status at various intersections into city-wide traffic monitoring data helps traffic management departments make informed decisions, such as adjusting traffic light timings and planning traffic diversion schemes. In complex monitoring scenarios, data from a single monitoring node cannot meet the needs of comprehensive monitoring. By combining multiple data sources, different monitoring scenarios can be adapted to support various application requirements.

[0026] Specifically, in step S3 of the embodiment provided by the present invention, several information interaction channels are constructed for the AI ​​perception model deployed on the local terminal. These channels are data transmission paths between the local terminal and the AI ​​perception model. Their function is to accurately transmit different types and levels of regional monitoring data to the AI ​​perception model for processing. For example, independent information interaction channels are constructed for image-type regional monitoring data, temperature and humidity-type regional monitoring data, smoke concentration-type regional monitoring data, etc., to ensure the stability and professionalism of data transmission.

[0027] More specifically, the parameters of the AI ​​perception model are adjusted in each information exchange channel. Based on the characteristics and needs of monitoring data in different areas, the relevant parameters of the model are adjusted so that the model's risk perception function can adapt to the perception mode of monitoring data in each area. For example, for image data, the threshold parameters for recognizing objects and behaviors in the model need to be adjusted; for temperature and humidity data, the parameters for judging risks need to be adjusted according to the set safety range, etc.

[0028] More specifically, the overall monitoring data within a designated area is diverse and complex. Different types of regional monitoring data (such as images, temperature and humidity, smoke concentration, etc.) differ in data format, characteristics, and risk manifestations. By constructing multiple information interaction channels and adjusting parameters separately, the AI ​​perception model can be optimized for different types of data, improving the model's adaptability to various types of data and ensuring accurate analysis and processing of various types of data.

[0029] More specifically, monitoring data from different areas have unique characteristics and risk patterns. For example, the monitoring focus and risk types of office areas and warehouse areas are different. By adjusting the parameters to adapt the model to the perception patterns of monitoring data in each area, it is possible to more accurately identify the surface risks in each area and avoid misjudgment or omission of risks due to the use of uniform and inappropriate parameters.

[0030] More specifically, the monitoring data from each area of ​​the overall monitoring data is fed into designated information exchange channels. Following the previously established channel correspondence, different types and areas of monitoring data are accurately input into their respective channels to ensure the data can be correctly processed by the model. This allows the AI ​​perception model to perform risk perception processing on the monitoring data from each area. The model analyzes the input data based on pre-trained algorithms and adjusted parameters, identifying potential risk factors. For example, for image data, the model detects abnormal human behavior or improper placement of objects; for smoke concentration data, it determines whether safety thresholds are exceeded. Ultimately, it generates surface-level risk perception information for the designated area, which can be presented in the form of text descriptions, risk ratings, etc.

[0031] More specifically, by inputting monitoring data from each region into designated information exchange channels, the model ensures that the data can be processed in a targeted manner according to its own characteristics. This avoids interference between different types of data, improves the accuracy and efficiency of the model in processing data, and thus more accurately identifies risk information in the data. Through the analysis and processing of monitoring data from each region by the AI ​​perception model, obvious and direct risk factors in the designated area can be quickly identified, generating surface-level risk perception information. This information can help monitoring personnel understand the security situation in the area at the first moment and take corresponding measures in a timely manner, such as stopping violations and adjusting for environmental anomalies.

[0032] More specifically, performing AI model analysis on local terminals avoids uploading all the large amounts of overall monitoring data to the cloud for processing, reducing the pressure and latency of data transmission. This is especially beneficial for monitoring scenarios with high real-time requirements, such as fire warnings and personnel safety monitoring. Local processing can generate risk perception information more quickly and issue alerts in a timely manner. Local processing of some data can also reduce security risks during data transmission and storage. This is particularly true for monitoring data containing sensitive information, such as facial images of people or environmental data of specific areas. Analyzing and processing data locally can better protect data privacy and security.

[0033] More specifically, the raw monitoring data includes image monitoring data and monitoring data from other sources. By performing risk perception on user behavior using image monitoring data, corresponding behavioral risk perception information can be obtained as part of the surface-level risk perception information, thereby achieving efficient and accurate behavioral safety supervision.

[0034] Specifically, in step S4 of the embodiment provided by this invention, the local terminal uploads the aggregated overall monitoring data to the cloud server via a network connection. Secure and reliable communication protocols, such as HTTPS, can be used to ensure the security and integrity of the data during transmission. After receiving the overall monitoring data at each moment, the cloud server arranges the data in chronological order. For example, based on timestamps, it arranges the data from image sensors, smoke sensors, temperature and humidity sensors, etc., collected at different times sequentially to form a monitoring information stream for a specified area. This clearly shows how the data changes over time.

[0035] More specifically, the overall monitoring data contains a large amount of information from multiple monitoring nodes at different times. The data volume is enormous and time-series. Cloud servers have powerful computing resources and storage capabilities, enabling them to efficiently process and analyze massive amounts of monitoring data. Local terminals, due to limitations in computing and storage capabilities, find it difficult to complete such complex tasks. Potential risk perception models are usually complex machine learning or deep learning models that require a large amount of computing resources for training and inference. Cloud servers can provide a stable computing environment, ensuring that the model can accurately identify the time-series characteristics of potential risks.

[0036] More specifically, a potential risk perception model is deployed on a cloud server. This model, trained on a large amount of historical data, has the ability to identify the temporal characteristics of potential risks. When the monitoring information stream is input into the potential risk perception model, the model begins to analyze and process it. The potential risk perception model identifies the temporal characteristics of various potential risks in the monitoring information stream. For example, by analyzing the changing trends of temperature and humidity data over time, it can determine whether there are potential risks such as equipment overheating or dampness; by analyzing the data temporal sequence of electromagnetic sensors for doors and windows, it can detect whether there are abnormal door and window opening and closing patterns, indicating potential threats such as illegal intrusion. Finally, the temporal performance characteristics of various potential risks are obtained.

[0037] More specifically, many potential risks are not obvious in data at a single moment, but by identifying the temporal characteristics of the overall monitoring data at various moments, the patterns and trends of data changes over time can be discovered, thereby uncovering the potential risks hidden behind the data. For example, equipment failure may not cause obvious anomalies in the short term, but by monitoring the temporal changes of its operating parameters over a long period, the occurrence of failure can be predicted in advance. By accumulating and processing information and generating risk weight parameters, potential risk information in different time periods can be comprehensively considered, and the severity and probability of potential risks can be more accurately assessed. Combining risk weight parameters can comprehensively display the potential risk status in a specified area, providing strong support for safety decisions.

[0038] More specifically, the information accumulation processing of the temporal characteristics of various potential risks can be carried out by methods such as integration and weighted summation. The potential risk information in different time periods can be accumulated. For example, for the potential fire risk in a certain area, the characteristic information such as the increase in temperature and the change in smoke concentration over a period of time can be accumulated. Based on the result of information accumulation, risk weight parameters of various potential risks are generated. The risk weight parameters reflect the severity and probability of each potential risk. The higher the weight, the more attention the potential risk deserves.

[0039] More specifically, by combining the risk weight parameters of various potential risks and comprehensively considering the interrelationships and impacts between different potential risks, risk perception information of potential levels in a specified area can be obtained. This information can be presented in the form of risk reports, risk maps, etc., to intuitively show the distribution and degree of potential risks in the specified area.

[0040] More specifically, cloud servers enable data sharing, allowing different users (such as security managers and maintenance personnel) to access potential risk perception information in the cloud via the network. This facilitates collaborative analysis and decision-making. Cloud servers can collect vast amounts of monitoring data and risk assessment results to continuously improve potential risk perception models. As data accumulates and models are optimized, the model's ability to identify potential risks will continuously improve.

[0041] This invention provides a security monitoring method based on the Internet of Things, which has the following beneficial effects: This invention leverages a security monitoring system to comprehensively acquire raw data, providing a foundation for subsequent analysis. Local terminals aggregate data and generate surface risk information through AI perception models, enabling timely detection of obvious hidden dangers. Data is then uploaded to a cloud server, utilizing its powerful computing capabilities to identify temporal characteristics and uncover potential risks. This combination of local and cloud-based approaches ensures both real-time performance and in-depth data analysis, achieving comprehensive monitoring of the surface and potential risks in a designated area and enhancing the level of security.

[0042] Preferably, the security monitoring system includes various monitoring nodes pre-set in the designated area, and the monitoring nodes include image sensors, smoke sensors, temperature and humidity sensors, and door and window electromagnetic sensors.

[0043] Preferably, the step of aggregating the raw monitoring data of each monitoring node based on the local terminal to obtain the overall monitoring data of the specified area includes: S21: Receive the raw monitoring data from each monitoring node based on the local terminal, and mark the data source of the raw monitoring data according to the setting location and monitoring form of the monitoring node to obtain the node monitoring data of each monitoring node. S22: Based on the location and monitoring mode of each monitoring node within the security monitoring system, the monitoring data of each node is combined in multiple ways to obtain several regional monitoring data. S23: Combine the regional monitoring data from each level into the overall monitoring data for the specified region.

[0044] Specifically, the local terminal establishes a connection with each monitoring node through a pre-set communication protocol and receives the raw monitoring data transmitted by them in real time. The communication protocol can be a wired Ethernet protocol or a wireless protocol such as Wi-Fi or ZigBee. For example, in a security monitoring system of a smart building, the local terminal will open the corresponding network port to listen to the data sent by monitoring nodes such as image sensors and smoke sensors on each floor.

[0045] More specifically, after receiving the raw monitoring data, the local terminal adds tags to the monitoring nodes according to their set location and monitoring type. The set location can be accurate to the specific floor, room number, or geographical coordinates; the monitoring type is clearly defined as image monitoring, smoke monitoring, temperature and humidity monitoring, etc. For example, for temperature and humidity sensor data from the 3rd floor conference room, it is marked as "3rd floor conference room - temperature and humidity sensor", thus obtaining the node monitoring data of each monitoring node.

[0046] More specifically, marking data sources facilitates accurate data tracing later. When abnormal data is detected or detailed analysis of monitoring information at a specific location is required, the specific monitoring node can be quickly located, and its relevant data can be queried. For example, if abnormal temperature and humidity data is found in a certain area, marking can directly locate the corresponding temperature and humidity sensor data for in-depth investigation. Clear data source marking makes data classification more convenient, facilitating effective management of monitoring data from different types and locations by local terminals. This provides a good foundation for subsequent data combination and analysis, improving the efficiency and accuracy of data processing.

[0047] More specifically, based on the location of each monitoring node within the security monitoring system, the relative positional relationships and spatial layout between them are analyzed. At the same time, considering their monitoring methods, the interactive value analysis of the monitoring data is conducted to determine the feasibility of data combinations between nodes. For example, image sensor data and smoke sensor data from adjacent rooms have high interactive value because images can intuitively show the situation on site, while smoke data can reflect whether there are fire hazards. Combining the two can more comprehensively assess the safety status. Based on these analysis results, suitable multi-data combination patterns are generated.

[0048] More specifically, the monitoring data from each node is combined according to the established data combination pattern. This can be achieved through methods such as data splicing and correlation analysis. For example, image sensor data and temperature and humidity sensor data from several adjacent offices on the same floor can be spliced ​​together, and correlation analysis can be used to find the inherent relationships between the data, such as the relationship between personnel activity and changes in indoor temperature and humidity, thereby obtaining the area monitoring data for that floor.

[0049] More specifically, data from a single monitoring node often only reflects partial information. By combining multiple data points, we can uncover the correlations between data from different nodes and obtain more comprehensive and in-depth information. For example, by combining and analyzing electromagnetic sensor data and image sensor data from doors and windows, we can determine whether personnel entry and exit are properly matched with door and window opening and closing, and promptly detect abnormal situations such as illegal intrusions. Different monitoring areas have different monitoring focuses and needs. By combining multiple data points, we can flexibly generate regional monitoring data that meets various needs based on actual conditions. For example, for a large shopping mall, in order to meet fire monitoring requirements, we can combine and analyze smoke sensor data, temperature and humidity sensor data, and image sensor data from emergency exits on each floor.

[0050] More specifically, the regional monitoring data at each level are integrated into the overall monitoring data of the designated area. This regional monitoring data can be a collection of monitoring data from different floors, different functional areas, etc. For example, the regional monitoring data of different floors of a building are aggregated together to form the overall monitoring data of the entire building.

[0051] More specifically, a designated area contains multiple different sub-areas, and the monitoring data at each level can only reflect the situation in a part of the area. Combining these into overall monitoring data can provide monitoring personnel with a comprehensive and integrated perspective, fully displaying the security status and operational status of the entire designated area. Overall monitoring data facilitates unified analysis and decision-making. Security managers can conduct comprehensive risk assessments based on overall monitoring data, formulate more scientific and reasonable security management strategies and emergency response plans, and avoid decision-making errors caused by focusing only on local data.

[0052] Preferably, the step of combining the monitoring data of each monitoring node in multiple ways to obtain several areas of monitoring data, based on the location and monitoring method of each monitoring node within the security monitoring system, includes: S21: Based on the location and monitoring mode of each monitoring node within the security monitoring system, retrieve a pre-determined data combination pattern for the monitoring data of each node; S22: Based on the data combination mode, the monitoring data of each node is combined in multiple ways to obtain several regional monitoring data; The steps for pre-determining the data combination pattern include: S211: The relative positioning relationship of each monitoring node is obtained based on the positioning analysis of each monitoring node within the security monitoring system; S212: Based on the relative positioning relationship, perform interactive value analysis on the monitoring data obtained from the monitoring of each monitoring node to obtain the monitoring interaction characteristics of each monitoring node. S213: Based on the monitoring interaction characteristics of each monitoring node, evaluate the feasibility of monitoring combination of each monitoring node to generate a data combination pattern between each monitoring node.

[0053] Specifically, the location information of each monitoring node is collected through the deployment documents of the security monitoring system, geographic information system (GIS) records, or the positioning module carried by the node itself. For example, for a security monitoring system of a large warehouse, the installation coordinates of each monitoring node (such as smoke sensor, temperature and humidity sensor, etc.) can be determined from the building floor plan of the warehouse.

[0054] More specifically, mathematical methods (such as Euclidean distance formula to calculate distance, trigonometric functions to calculate azimuth angle) are used to analyze the relative distance and azimuth relationship between each monitoring node based on the acquired location information. For example, the straight-line distance between two adjacent temperature and humidity sensors and their azimuth relative to each other are calculated. The relative positioning relationship of each monitoring node is represented in the form of graphics (such as topology diagram) or data matrix to intuitively show the spatial layout between nodes. For example, in the topology diagram, nodes represent monitoring equipment, lines represent the relationship between nodes, and distance and azimuth information are marked.

[0055] More specifically, relative positioning relationships are an important foundation for subsequent data interaction value analysis and monitoring combination feasibility assessment. Data acquired by monitoring nodes in different locations have mutual influence and correlation. Understanding their relative positions helps to explore this correlation in depth. Based on relative positioning relationships, data combination methods can be planned more rationally, so that the combined data can more accurately reflect the actual situation of a specific area. For example, the data combination of two monitoring nodes in adjacent locations is more representative and can better depict the state of the local area.

[0056] More specifically, the types (such as images, numerical values, on / off status, etc.) and characteristics (such as trends, periodicity, etc.) of the data collected by each monitoring node should be clearly defined. For example, the characteristics of people's activity images collected by image sensors and the fluctuation range and frequency of temperature and humidity sensor data should be analyzed. Based on the relative positioning relationship of each monitoring node, the possible correlation between different types of data should be studied. For example, in a shopping mall monitoring system, there is a certain correlation between the flow of people near the entrance (image sensor statistics) and the temperature and humidity changes in the area (temperature and humidity sensor data). The degree of correlation can be verified by statistical analysis methods (such as correlation analysis).

[0057] More specifically, based on the results of data correlation analysis, the interaction value between data from each monitoring node is evaluated to determine which combinations can provide richer and more valuable information. For example, combining image sensor data inside the elevator with elevator operation status sensor data can provide a comprehensive understanding of elevator usage and passenger load, and this combination has high interaction value.

[0058] More specifically, data from different monitoring nodes may only reflect a single piece of information when viewed individually, but through interaction value analysis, hidden connections between them can be discovered, thereby uncovering more valuable information. For example, combining smoke sensor data and video surveillance data can more accurately determine the occurrence and development of a fire. After clarifying the interaction value of data from each monitoring node, targeted data combination strategies can be formulated, prioritizing the combination of nodes with high interaction value, thereby improving the quality and effectiveness of regional monitoring data.

[0059] More specifically, it is necessary to assess whether the data communication protocols between monitoring nodes are compatible, whether the data transmission bandwidth is sufficient to support the combined data processing, and whether the computing power of the local terminal can handle the corresponding data processing tasks. For example, if two monitoring nodes use different wireless communication protocols, protocol conversion or equipment replacement is required to achieve smooth data combination.

[0060] More specifically, consider the hardware costs, software development costs, and maintenance costs required to combine monitoring nodes, and compare them with the benefits that the combination can bring (such as improving the level of security monitoring and reducing accident losses). For example, if combining certain monitoring nodes requires a large investment in equipment upgrades and system modifications, but the additional benefits obtained are limited, then such a combination is not feasible.

[0061] More specifically, based on the feasibility assessment results of the monitoring combination, a reasonable data combination mode is formulated for each monitoring node. This can be represented by tree structure, mesh structure, or other methods. The monitoring nodes included in each combination and the rules for data transmission and processing are clearly defined. For example, in a factory safety monitoring system, multiple equipment-related monitoring nodes in the same workshop are combined into a subnet, and data is processed and transmitted according to preset priorities and logical relationships.

[0062] More specifically, through technical feasibility and cost-benefit analysis, we can avoid encountering unsolvable technical problems or unreasonable input-output ratios during actual implementation, ensuring the smooth implementation of data combination schemes. Generating reasonable data combination patterns can effectively utilize existing monitoring resources, improve resource utilization efficiency, and enable the security monitoring system to maximize cost-effectiveness while meeting monitoring needs.

[0063] More specifically, the pre-determined data combination patterns between each monitoring node are stored in the database of the local terminal or on the server to form a data combination pattern library, which is then classified and managed for easy querying and retrieval. When it is necessary to combine the monitoring data of the nodes, the local terminal performs a matching query in the data combination pattern library based on the setting, location and monitoring form information of each node, and retrieves the corresponding data combination pattern. For example, if the information "smoke sensor and camera on a certain floor" is entered, the system will automatically find the suitable combination pattern from the pattern library.

[0064] More specifically, pre-stored data combination patterns avoid the need for complex analysis and decision-making processes each time data is combined. They can quickly and accurately find suitable combination methods for node monitoring data, improve data processing efficiency, and call a unified database combination pattern to ensure consistency and standardization of data combinations at different times and under different conditions, avoiding combination differences caused by human factors.

[0065] More specifically, according to the data combination pattern, the corresponding raw data is collected from each monitoring node and the data is synchronized to ensure that the collected data is consistent in time. For example, data from multiple related monitoring nodes at the same time are collected simultaneously to enable accurate combination analysis. According to the data processing rules specified in the combination pattern (such as data splicing, weighted summation, logical operations, etc.), the collected node monitoring data is fused and processed. For example, the data from temperature sensors and humidity sensors are fused to calculate the comprehensive temperature and humidity index of the area.

[0066] More specifically, after data fusion and processing, several regional monitoring data are obtained. These data can more comprehensively reflect the safety status and operational status of a specific area, providing strong support for subsequent safety analysis and decision-making. For example, by combining the operational status data of multiple devices in a workshop, the overall operational status monitoring data of the workshop can be generated.

[0067] More specifically, by combining scattered node monitoring data into more valuable regional monitoring data through multiple combinations, a more comprehensive and macro-level perspective can be provided to observe and analyze the situation in a designated area. This helps to identify potential security risks and problems in a timely manner. Regional monitoring data can provide security monitoring system managers with more accurate and detailed information to help them make scientific and reasonable decisions. For example, based on the combined regional monitoring data, production plans can be adjusted in a timely manner and security precautions can be taken.

[0068] Preferably, the step of analyzing the overall monitoring data using an AI perception model deployed on a local terminal to generate surface-level risk perception information for the designated area includes: S31: Construct several information interaction channels for the AI ​​perception model deployed on the local terminal, and adjust the parameters of the AI ​​perception model in each of the information interaction channels to adjust the risk perception function of the AI ​​perception model to a perception mode that is adapted to the monitoring data of each region. S32: Substitute the monitoring data of each area of ​​the overall monitoring data into the designated information interaction channel, so that the AI ​​perception model can perform risk perception processing on the monitoring data of each area and generate risk perception information at the surface level of the designated area.

[0069] Specifically, based on the type (such as image data, temperature and humidity data, smoke concentration data, etc.) and characteristics of regional monitoring data, different types of information interaction channels are constructed for the AI ​​perception model. For example, a visual information interaction channel is constructed for image data, and a numerical information interaction channel is constructed for numerical temperature, humidity and smoke concentration data. A dedicated data interface is designed for each information interaction channel to ensure that the monitoring data of each region can be accurately and efficiently transmitted to the AI ​​perception model. These interfaces need to be compatible with the data formats and communication protocols of different monitoring nodes. For example, for sensor data using different communication standards, corresponding conversion interfaces are designed. The constructed information interaction channels are uniformly managed, including the opening and closing of channels, data flow monitoring, etc. Automated management can be achieved through software programs to ensure the stable operation of the channels.

[0070] More specifically, a deep analysis of the characteristics of monitoring data in each region is conducted, such as the data distribution range, variation patterns, and manifestations of anomalies. For example, analyzing the fluctuation range and normal value range of temperature and humidity data in a certain region across different seasons, and selecting appropriate parameter adjustment methods based on data characteristics and the type of AI perception model. Common methods include grid search, random search, and Bayesian optimization. For instance, for AI perception models based on neural networks, a random search method can be used to find the optimal parameter combination within a certain parameter range.

[0071] More specifically, in each information interaction channel, the parameters of the AI ​​perception model are adjusted using the selected adjustment method. By continuously trying different parameter values, the risk perception function of the model is adapted to the characteristics of the monitoring data in each region. For example, the size and number of convolution kernels in the image recognition model are adjusted to improve the accuracy of identifying abnormal targets in images of specific regions.

[0072] More specifically, the overall monitoring data of a designated area includes regional monitoring data of various types and characteristics. Constructing different information interaction channels can enable different types of data to interact with the AI ​​perception model independently and efficiently, avoiding interference between data. At the same time, parameter adjustment can optimize the model for different data features, improving the model's adaptability and processing capabilities for various types of data.

[0073] More specifically, the risk manifestations and characteristics of monitoring data in different regions vary. By adjusting the parameters of the AI ​​perception model in a targeted manner, the model can more accurately identify risk information in the monitoring data of each region, thereby improving the accuracy and reliability of surface-level risk perception. Building information interaction channels and adjusting parameters enable the system to flexibly respond to monitoring data of different types and characteristics. When a new monitoring node or data type is added, it can be incorporated into the system for risk perception simply by building the corresponding information interaction channels and adjusting the parameters, thus enhancing the scalability of the system.

[0074] More specifically, based on the type and characteristics of the regional monitoring data, it is matched with pre-built information interaction channels. For example, image-based regional monitoring data is substituted into the visual information interaction channel, and temperature and humidity data is substituted into the numerical information interaction channel. Before substituted into the information interaction channels, necessary preprocessing is performed on the data, such as normalization, filtering, and noise reduction. These operations can improve the quality of the data and enable the AI ​​perception model to process the data better. For example, image data is denoised and brightness is adjusted, and numerical data is normalized to ensure that its value range is within the range that the model can process. The preprocessed regional monitoring data is then substituted into the AI ​​perception model according to the matched information interaction channels to ensure the integrity and accuracy of the data during transmission.

[0075] More specifically, after receiving regional monitoring data, the AI ​​perception model performs inference calculations based on pre-trained algorithms and adjusted parameters. The model analyzes the features and patterns in the data to identify potential risk factors. For example, the image recognition model performs target detection and classification on the input image to determine whether there are abnormal personnel or objects; the numerical analysis model performs threshold judgment and trend analysis on temperature and humidity data to detect whether there are environmental anomalies. Based on the model's inference results, it generates risk perception information at the surface level of the specified area. This information can be presented in the form of text descriptions, risk level assessments, and visualization charts. For example, it may output the text message "The temperature in a certain area exceeds the normal range, there is a fire risk, and the risk level is high," or generate a heat map showing the risk distribution of each area.

[0076] More specifically, inputting monitoring data from various regions into designated information exchange channels and performing risk perception processing is a key step in using AI perception models to analyze and judge data. Through the inference and calculation of the model, surface-level risk information can be identified from a large amount of monitoring data, providing a basis for safety monitoring and decision-making.

[0077] More specifically, different types of regional monitoring data require different processing methods and model parameters. Substituting the data into a designated information exchange channel can ensure that the model performs specialized processing for each type of data, improving the effectiveness and accuracy of risk perception. By timely processing regional monitoring data for risk perception and generating risk perception information, safety monitoring personnel can promptly understand the safety status of designated areas. Once a risk is detected, corresponding measures can be taken quickly to reduce the possibility of accidents and losses.

[0078] Preferably, the step of uploading the overall monitoring data to a cloud server to identify the temporal characteristics of the overall monitoring data at each time point and obtaining the potential level of risk perception information for the specified area includes: S41: Upload the overall monitoring data to the cloud server, and have the cloud server perform time-series processing on the overall monitoring data at each moment to obtain the monitoring information stream of the specified area; S42: The potential risk perception model deployed on the cloud server is used to identify the temporal characteristics of various potential risks in the monitoring information stream, so as to obtain the temporal performance characteristics of each potential risk. S43: Accumulate and process information on the temporal performance characteristics of each potential risk to generate risk weight parameters for each potential risk; S44: Combine the risk weight parameters of each potential risk to obtain the potential level risk perception information of the specified area.

[0079] Specifically, local terminals upload overall monitoring data to the cloud server via a secure and reliable network connection (such as VPN, HTTPS, etc.). Batch upload or real-time streaming upload can be used, depending on the data volume and monitoring requirements. For example, batch upload can be used for small datasets with low real-time requirements; real-time streaming upload is used for data requiring real-time analysis. After receiving the overall monitoring data at each moment, the cloud server sorts the data based on the timestamp information. This can be done using database sorting functions or by writing a dedicated sorting algorithm to arrange the data in chronological order, forming a monitoring information stream for a specified area. For example, SQL statements can be used to sort the monitoring data stored in the database in ascending order by the time field.

[0080] More specifically, cloud servers have powerful computing and storage capabilities, enabling them to process large-scale monitoring data. Uploading the overall monitoring data to the cloud allows for full utilization of its advantages, enabling more complex data analysis and processing. Through time-series processing, discrete monitoring data can be transformed into a continuous monitoring information stream, clearly showing how the data changes over time. This helps in the subsequent identification and analysis of the temporal characteristics of potential risks.

[0081] More specifically, a pre-trained potential risk perception model is deployed on a cloud server. This model can be based on machine learning (such as neural networks, decision trees, etc.) or deep learning (such as recurrent neural networks, long short-term memory networks, etc.) and can effectively analyze time-series data. The monitoring information flow is input into the potential risk perception model, which analyzes the data segment by segment to identify the time-series characteristics of various potential risks. For example, by analyzing the trend of temperature data changes, it can determine whether there is a potential risk of equipment overheating; by analyzing the time distribution of personnel entry and exit frequency, it can detect whether there are abnormal personnel activity patterns. The model will extract features related to potential risks, such as rate of change, periodicity, abnormal fluctuations, etc., to obtain the time-series performance characteristics of various potential risks.

[0082] More specifically, many potential risks are not obvious in data at a single moment, but they will show specific characteristics in the temporal changes of the data. By analyzing the monitoring information flow through a potential risk perception model, potential risks hidden behind the data can be discovered, and security risks can be detected in advance. By using a specialized time series analysis model, the temporal characteristics of potential risks can be captured more accurately, avoiding misjudgments or omissions caused by the limitations of data at a single moment.

[0083] More specifically, based on the characteristics and importance of potential risks, appropriate information accumulation methods should be selected. Common methods include simple accumulation and weighted accumulation. For example, for certain key potential risks, a weighted accumulation method can be used, assigning higher weights to recent time-series performance characteristics and accumulating the time-series performance characteristics of each potential risk. For example, the rate of change of a potential risk over a period of time can be accumulated, or the number of abnormal fluctuations in different time periods can be statistically analyzed. Through information accumulation processing, the risk weight parameters of each potential risk can be obtained. These parameters reflect the severity and probability of the potential risk.

[0084] More specifically, a single temporal performance characteristic can only reflect one aspect of a potential risk. Through information accumulation and processing, all relevant characteristics over a period of time can be considered comprehensively to assess the severity and likelihood of potential risks more fully. Risk weight parameters quantify potential risks, enabling comparison and ranking between different potential risks. This helps safety managers take appropriate measures based on the priority of risks.

[0085] More specifically, based on the interrelationships and influences between potential risks, the combination rules for risk weight parameters are determined. Linear combination, nonlinear combination, and other methods can be adopted. For example, if there is a synergistic effect between certain potential risks, a nonlinear combination method can be used to consider their interactive influence. The risk weight parameters of each potential risk are combined according to the determined combination rules. Through combination processing, a comprehensive index is obtained, which is the risk perception information of the potential level of a specified area. This information can be presented in the form of risk score, risk level, etc., to intuitively reflect the potential risk status of the specified area.

[0086] More specifically, a single potential risk weight parameter can only reflect the situation of a certain type of potential risk. By combining them, information on all potential risks can be integrated to comprehensively present the potential risk status of a specified area. The comprehensive potential risk perception information provides safety managers with a clear risk view, which helps them to formulate scientific and reasonable safety strategies and decisions, and take effective measures to reduce potential risks.

[0087] Preferably, it further includes: performing security restriction analysis on a designated area based on potential level risk perception information to obtain the security range of monitoring data for each monitoring node of the security monitoring system for the designated area, and deploying a corresponding security monitoring trigger mechanism for each monitoring node according to the security range of monitoring data for each monitoring node, so as to trigger an alarm function when the monitoring node collects monitoring data that meets the standard.

[0088] Specifically, this involves a thorough interpretation of potential risk perception information, categorizing risks based on their type, severity, and scope of impact. For example, potential risks can be classified into fire risks, equipment failure risks, and personnel safety risks. The correlation between each type of potential risk and the monitoring nodes in the safety monitoring system is analyzed to determine which monitoring nodes' data reflects specific types of potential risks. For instance, smoke sensor data is closely related to fire risk, and equipment temperature sensor data is related to equipment failure risk. Combining the characteristics of potential risks with historical data, a safe range is determined for the monitoring data of each monitoring node. Statistical analysis methods can be used, such as calculating the mean and standard deviation of the data to determine the normal fluctuation range of the data; alternatively, safety thresholds can be set based on industry standards and experience. For example, for temperature and humidity sensors, a suitable temperature and humidity range can be determined as the safe range based on the requirements of cargo storage.

[0089] More specifically, by determining the security range of monitoring data for each monitoring node based on risk perception information at potential levels, potential risks can be identified more accurately. Different monitoring nodes monitor different types of potential risks, avoiding blind and over-monitoring, thus improving the efficiency and accuracy of security monitoring. A clear security range provides monitoring nodes with clear judgment criteria. When monitoring data exceeds the security range, anomalies can be detected in a timely manner, buying time to take measures and reducing the possibility of potential risks evolving into actual accidents.

[0090] More specifically, based on the safe range of the monitoring data of each monitoring node, corresponding safety monitoring trigger conditions are set. When the data collected by the monitoring node exceeds the safe range, the alarm function is triggered. The trigger condition can be the breach of a single data threshold or a combination of multiple data conditions. For example, when the smoke concentration detected by the smoke sensor exceeds the set threshold, or when the temperature detected by the temperature sensor rises sharply in a short period of time and exceeds a certain value, the alarm is triggered.

[0091] More specifically, determine the appropriate alarm method, such as audible and visual alarm, SMS alarm, email alarm, etc. Select one or more alarm methods according to different application scenarios and security requirements. For example, in an industrial production environment, audible and visual alarms and SMS alarms can be used simultaneously to ensure that relevant personnel can receive alarm information in a timely manner. Configure the parameters of the alarm function, such as alarm priority, alarm duration, alarm interval, etc. Set different alarm priorities according to the severity of potential risks to ensure that high-risk situations can be responded to in a timely manner. For example, for fire risk alarms, set the highest priority and continue to alarm until the risk is eliminated.

[0092] More specifically, deploying a security monitoring trigger mechanism, which triggers an alarm function when monitoring nodes collect monitoring data that meets the standards, ensures that relevant personnel are promptly informed of the existence of potential risks. The rapid response mechanism enables personnel to take swift action, such as evacuating personnel or activating emergency equipment, thereby reducing losses caused by accidents. By setting alarm parameters, such as duration and interval, continuous monitoring and early warning of potential risks can be achieved. Even when personnel are temporarily unable to handle the situation, they can be continuously reminded to pay attention to the risk situation, ensuring that the risk is properly handled.

[0093] More specifically, the establishment of security limitation analysis and security monitoring triggering mechanisms is based on potential risk perception information and monitoring data. It is a data-driven security management approach that can provide a scientific basis for security management decisions, making the decisions more reasonable and effective. As potential risks change and monitoring data accumulates, the security scope of monitoring data and security monitoring triggering mechanisms of each monitoring node can be dynamically adjusted to continuously optimize security management strategies and adapt to different security needs and environmental changes.

[0094] Reference Figure 2 As shown, in a second aspect, the present invention provides an Internet of Things (IoT)-based security monitoring system for implementing the IoT-based security monitoring method described in any one of the first aspects, comprising: The data monitoring module is used to monitor information in a designated area through the security monitoring system in order to obtain the raw monitoring data of each monitoring node in the system. The data aggregation module is used to aggregate the raw monitoring data of each monitoring node based on the local terminal to obtain the overall monitoring data of the specified area. The surface perception module is used to analyze the overall monitoring data through an AI perception model deployed on a local terminal, and generate risk perception information of the surface layer of the specified area. The potential perception module is used to upload the overall monitoring data to the cloud server to identify the temporal characteristics of the overall monitoring data at each time point, and obtain the potential level risk perception information of the specified area.

[0095] In this embodiment, the specific implementation of each module in the above system embodiment is described in the above method embodiment, and will not be repeated here.

[0096] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A security monitoring method based on the Internet of Things, characterized in that, include: Information is monitored in a designated area through a security monitoring system to obtain raw monitoring data from each monitoring node within the system; wherein, the raw monitoring data includes image monitoring data and monitoring data from other sources. The original monitoring data of each monitoring node is summarized based on the local terminal to obtain the overall monitoring data of the specified area. The overall monitoring data is analyzed by an AI perception model deployed on a local terminal to generate surface-level risk perception information for the specified area; wherein, the surface-level risk perception information includes behavioral risk perception information obtained by analyzing the image monitoring data; The overall monitoring data is uploaded to a cloud server to identify the temporal characteristics of the overall monitoring data at each time point, thereby obtaining risk perception information of the potential level of the specified area.

2. The security monitoring method based on the Internet of Things as described in claim 1, characterized in that, The security monitoring system includes various monitoring nodes pre-installed in the designated area, and the monitoring nodes include image sensors, smoke sensors, temperature and humidity sensors, and door and window electromagnetic sensors.

3. The security monitoring method based on the Internet of Things as described in claim 1, characterized in that, The steps for aggregating the raw monitoring data of each monitoring node based on the local terminal to obtain the overall monitoring data of the specified area include: The system receives raw monitoring data from each monitoring node via a local terminal and marks the source of the raw monitoring data according to the location and monitoring method of the monitoring node to obtain the node monitoring data of each monitoring node. Based on the location and monitoring method of each monitoring node within the security monitoring system, the monitoring data of each node are combined in multiple ways to obtain several regional monitoring data. The monitoring data from each of the aforementioned areas are combined to form the overall monitoring data for the designated area.

4. The security monitoring method based on the Internet of Things as described in claim 3, characterized in that, Based on the location and monitoring method of each monitoring node within the security monitoring system, the steps of combining the monitoring data of each node to obtain monitoring data for several areas include: Based on the location and monitoring mode of each monitoring node within the security monitoring system, a predetermined data combination pattern is retrieved for the monitoring data of each node. The monitoring data of each node is combined in multiple ways according to the data combination pattern to obtain several regional monitoring data. The steps for pre-determining the data combination pattern include: The relative positioning relationship of each monitoring node is obtained based on the positioning analysis of each monitoring node within the security monitoring system. Based on the relative positioning relationship, the interaction value analysis of the monitoring data obtained from the monitoring of each monitoring node is performed to obtain the monitoring interaction characteristics of each monitoring node. The feasibility of monitoring combination of each monitoring node is evaluated based on the monitoring interaction characteristics of each monitoring node, so as to generate a data combination pattern between each monitoring node.

5. The IoT-based security monitoring method as described in claim 3, characterized in that, The steps of analyzing the overall monitoring data using an AI perception model deployed on a local terminal to generate surface-level risk perception information for the specified area include: Several information interaction channels are constructed for the AI ​​perception model deployed on the local terminal, and the parameters of the AI ​​perception model are adjusted in each of the information interaction channels to adjust the risk perception function of the AI ​​perception model to a perception mode that is adapted to the monitoring data of each region. The monitoring data of each region in the overall monitoring data is respectively substituted into the designated information interaction channel so that the AI ​​perception model can perform risk perception processing on the monitoring data of each region and generate surface-level risk perception information of the designated region.

6. The security monitoring method based on the Internet of Things as described in claim 1, characterized in that, The steps of uploading the overall monitoring data to a cloud server to identify the temporal characteristics of the overall monitoring data at each time point and obtaining the potential level of risk perception information for the specified area include: The overall monitoring data is uploaded to the cloud server, and the cloud server performs time-series processing on the overall monitoring data at each moment to obtain the monitoring information stream of the specified area. The potential risk perception model deployed on a cloud server is used to identify the temporal characteristics of various potential risks in the monitoring information stream, so as to obtain the temporal performance characteristics of each potential risk. Information on the temporal characteristics of each potential risk is accumulated and processed to generate risk weight parameters for each potential risk; By combining the risk weight parameters of each potential risk, the potential level of risk perception information for the specified area is obtained.

7. The security monitoring method based on the Internet of Things as described in claim 1, characterized in that, Also includes: Based on the risk perception information of the potential level, a security restriction analysis is performed on the designated area to obtain the security range of the monitoring data of each monitoring node of the security monitoring system for the designated area. According to the security range of the monitoring data of each monitoring node, a corresponding security monitoring trigger mechanism is deployed for each monitoring node to trigger an alarm function when the monitoring node collects monitoring data that meets the standard.

8. A security monitoring system based on the Internet of Things, characterized in that, A security monitoring method based on the Internet of Things as described in any one of claims 1-7 includes: The data monitoring module is used to monitor information in a designated area through the security monitoring system in order to obtain the raw monitoring data of each monitoring node in the system. The data aggregation module is used to aggregate the raw monitoring data of each monitoring node based on the local terminal to obtain the overall monitoring data of the specified area. The surface perception module is used to analyze the overall monitoring data through an AI perception model deployed on a local terminal, and generate risk perception information of the surface layer of the specified area. The potential perception module is used to upload the overall monitoring data to the cloud server to identify the temporal characteristics of the overall monitoring data at each time point, and obtain the potential level risk perception information of the specified area.

Images