Video monitoring intelligent patrol and polling scheduling system and method fused with AR tag

Abstract

The invention provides a video monitoring intelligent patrol and polling scheduling system and method fused with an AR tag. The method comprises the following steps: deploying the AR tag and associating physical space and monitoring equipment information; collecting video data, AR label interaction data and environment perception data of a monitoring area; performing quantitative evaluation on the state of the monitoring area based on the multi-dimensional fusion evaluation model to obtain an area state evaluation index; intelligently generating a patrol task plan and a monitoring polling strategy according to the index and a preset scheduling rule; tasks and strategies are executed through the patrol terminal and the monitoring device, and the execution state is fed back in real time to dynamically adjust the scheduling scheme. According to the invention, accurate mapping of the physical space and the monitoring data is realized through the AR tag, and the problems of low efficiency and non-intelligent monitoring polling of traditional patrol are solved by combining a multi-dimensional data fusion evaluation and dynamic scheduling algorithm, so that intelligent and accurate scheduling of patrol and monitoring polling is realized, and the response speed and the coverage quality of safety prevention and control are improved.

Description

Technical Field

[0001] This invention relates to the fields of video surveillance technology, AR technology and intelligent scheduling, specifically to a video surveillance intelligent patrol and rotation scheduling system and method that integrates AR tags. Background Technology

[0002] In the field of security monitoring and facility inspection, patrol and video surveillance rotation are core means of ensuring security. Traditional patrol methods mainly rely on manual recording. Patrol personnel need to arrive at designated locations according to preset routes and confirm the completion of the patrol by checking in and signing. This has problems such as low efficiency, high cost, easy to miss inspections, and the possibility of cheating. At the same time, manual patrol cannot respond to sudden abnormal events in real time, and also poses certain risks to the safety of patrol personnel.

[0003] With the development of video surveillance technology, the traditional video surveillance patrol mode has gradually replaced some manual patrols. However, existing surveillance patrols mostly adopt fixed cycles and fixed routes, lacking dynamic perception of the actual status of the monitored area. For example, some areas have already shown abnormal risks (such as excessive smoke concentration or unauthorized entry), but the monitoring equipment still patrols according to the regular cycle, resulting in untimely detection of abnormal events; while some low-risk areas suffer from a waste of monitoring resources.

[0004] Furthermore, in existing technologies, there is a lack of precise correlation mapping between physical spatial locations, monitoring equipment, and patrol tasks. When monitoring equipment detects an anomaly, it is difficult to quickly locate the physical coordinates of the anomaly location and the corresponding patrol responsibility area, resulting in a delay in anomaly response. At the same time, patrol route planning does not fully consider multi-dimensional information such as regional risk level and environmental factors, resulting in insufficient route rationality and further reducing the overall efficiency of patrol and monitoring rotation.

[0005] Therefore, how to integrate emerging technologies to achieve precise correlation between physical space and monitoring data, dynamically adjust patrol and monitoring rotation strategies based on regional status, and improve the intelligence level and response speed of scheduling has become a technical problem that the industry urgently needs to solve. Summary of the Invention

[0006] This invention provides a video surveillance intelligent patrol and rotation scheduling system and method that integrates AR tags, aiming to solve the problems of low efficiency and unintelligent monitoring rotation in traditional patrols, realize intelligent and precise scheduling of patrols and monitoring rotation, and improve the response speed and coverage quality of security and prevention.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solution:

[0008] Firstly, a method for intelligent patrol and rotation scheduling of video surveillance integrating AR tags is provided, including the following steps:

[0009] Deploy AR tags and associate them with physical space and monitoring equipment information;

[0010] Collect video data, AR tag interaction data, and environmental perception data of the monitored area;

[0011] Based on the video data, AR tag interaction data and environmental perception data, the status of the monitored area is quantitatively evaluated through a multi-dimensional fusion evaluation model to obtain the area status evaluation index.

[0012] Based on the regional status assessment index and preset scheduling rules, intelligently generate patrol task planning and monitoring rotation strategies;

[0013] The patrol task and rotation strategy are executed through the patrol terminal and monitoring equipment, and the execution status is fed back in real time, and the scheduling scheme is dynamically adjusted.

[0014] As a further improvement to the technical solution of this invention, the deployment of AR tags and the association of physical space with monitoring equipment information specifically includes:

[0015] Deploy visual AR tags and near-field communication AR tags with built-in unique identifiers at key locations within the monitored area;

[0016] Establish a mapping table that associates AR tag identifiers with physical space coordinates, the ID of the monitoring device to which they belong, and the device parameters, including the pixel resolution, shooting angle, and effective monitoring radius of the monitoring device;

[0017] The association mapping table is stored in the edge computing node to enable real-time access to AR tag information and monitoring devices and physical spaces.

[0018] As a further improvement to the technical solution of this invention, the collection of video data, AR tag interaction data, and environmental perception data of the monitored area specifically includes:

[0019] Real-time video streams are collected by monitoring equipment, and features of moving targets, abnormal behavior, and scene changes are extracted from the video frames.

[0020] By interacting with AR tags through near-field communication, the system collects tag trigger time, trigger frequency, and patrol terminal location information.

[0021] Environmental data is collected by sensors deployed in the monitoring area, including light intensity, temperature and humidity, smoke concentration and vibration frequency.

[0022] The video data, AR tag interaction data, and environmental data are time-stamped and cleaned. A Gaussian filtering algorithm is used to remove noisy data and retain valid data samples.

[0023] As a further improvement to the technical solution of this invention, based on the video data, AR tag interaction data, and environmental perception data, a multi-dimensional fusion evaluation model is used to quantitatively evaluate the status of the monitored area, resulting in an area status evaluation index that specifically includes:

[0024] A multi-dimensional fusion evaluation model is constructed, and the input dimensions of the model include video anomaly feature weights, AR tag interaction activity, and environmental risk level.

[0025] The weight coefficients for each input dimension are determined using the Analytic Hierarchy Process (AHP): Video anomaly feature weights. AR tag interaction activity weight Environmental risk level weights ,satisfy Among them, the consistency test index CR of the judgment matrix of the analytic hierarchy process is <0.1;

[0026] Calculate the quantized values ​​for each dimension:

[0027] Video anomaly quantization value , ,in This is the severity coefficient (range 1-5) for the i-th type of abnormal behavior (intrusion, loitering, object movement, etc.). Let n be the detection confidence level for the i-th type of abnormal behavior (range 0-1), and n be the total number of abnormal behavior categories.

[0028] AR tag interaction activity ,in The importance coefficient for labels (λ=1.5 for key locations, λ=1.5 for ordinary locations) =1.0), The actual trigger duration per unit of time. The threshold value is 30 minutes (default), and k is the decay coefficient (value 0.8).

[0029] Environmental risk level , ,in The risk weights for the j-th type of environmental parameters are: smoke concentration b=0.4, vibration frequency b=0.3, temperature and humidity b=0.15, and light intensity b=0.15. Let be the normalized detection value of the j-th type of environmental parameter (range 0-1), and m be the total number of environmental parameter categories;

[0030] The regional status assessment index is calculated using a weighted fusion formula.

[0031] ,

[0032] in This is the covariance correction factor (values ​​range from 0.05 to 0.1). The trace of the covariance matrix of V, A, and E is used to correct for the correlation effects of multi-dimensional data. The regional state assessment index S ranges from 0 to 10.

[0033] As a further improvement to the technical solution of the present invention, based on the regional state assessment index and preset scheduling rules, the intelligent generation of patrol task planning and monitoring rotation strategy specifically includes:

[0034] Preset status assessment threshold range: Low risk range Medium-risk area High-risk areas ;

[0035] When the regional status assessment index S falls within the low-risk range, a regular patrol route is generated, employing a uniform rotation strategy with a rotation cycle of [missing information]. ,in The basic patrol cycle (default 60 minutes);

[0036] When S falls within the medium-risk range, an optimized patrol route is generated, increasing the patrol frequency of key locations (by 50% compared to the regular route). A weighted rotation patrol strategy is adopted, with a rotation cycle of... ,in This is a risk correction factor (with a value of 0.6-0.8).

[0037] When S belongs to a high-risk zone, an emergency patrol route is generated, prioritizing the dispatch of patrol terminals closest to the target area, and employing a dense patrol strategy with a patrol cycle of [missing information]. ,in This is the emergency response coefficient (value range: 0.3-0.5).

[0038] Based on the physical spatial coordinates associated with AR tags, by improving Algorithm for planning patrol routes, path cost function

[0039] ,

[0040] in Let k be the length of the path segment k (in meters). This is the regional status assessment index corresponding to the k-th path segment. This is the path risk weight (value 0.3). This represents the estimated travel time (in seconds) for the k-th segment of the path. The speed of the patrol terminal (unit: meters per second). The time cost weight is 0.4. Let be the obstacle density of the k-th path segment (values ​​0-1). This is the obstacle avoidance coefficient (values ​​range from 1.2 to 2.0). The obstacle weight is 0.3, and I is the total number of path segments.

[0041] As a further improvement to the technical solution of this invention, the patrol task and rotation strategy are executed through the patrol terminal and monitoring equipment, and the execution status is fed back in real time. The scheduling scheme is dynamically adjusted, specifically including:

[0042] The patrol terminal confirms arrival at patrol points through dual verification via visual recognition of AR tags (QR code parsing) and near-field communication (UWB positioning), with a verification success rate of ≥99.5%.

[0043] The monitoring equipment adjusts the shooting parameters, including focal length, exposure and shooting angle, according to the patrol strategy, and performs close-up shooting of high-risk areas (resolution ≥1080P, frame rate ≥30fps).

[0044] The patrol terminal and monitoring equipment upload execution data in real time through the 5G edge computing network, including patrol completion rate, monitoring screen clarity, and abnormal event handling results;

[0045] The regional status assessment index is updated based on real-time feedback data. The patrol route and patrol cycle are dynamically adjusted through a closed-loop control algorithm. The adjustment formula is as follows: ,in This is the updated state assessment index. For the updated patrol cycle, To adjust the sensitivity coefficient (value 0.1-0.2), R is the execution status feedback coefficient (R=0.8 when the patrol completion rate is ≥95%; R=0.5 when the patrol completion rate is ≤80% and <95%; R=0.2 when the patrol completion rate is <80%).

[0046] As a further improvement to the technical solution of the present invention, the visual AR tag is a composite tag combining QR code and infrared marker, and the near-field communication AR tag supports dual-mode communication of NFC and UWB. The positioning accuracy of UWB communication is ≤5 cm, and the effective distance of NFC communication is ≤10 cm.

[0047] A second aspect of the present invention provides a video surveillance intelligent patrol and rotation scheduling system integrating AR tags, comprising:

[0048] The tag deployment and association unit is used to deploy AR tags and associate them with physical space and monitoring equipment information, establish an association mapping table and store it to the edge computing node;

[0049] The data acquisition unit is used to collect video data, AR tag interaction data and environmental perception data of the monitored area, and to perform timestamp alignment and data cleaning.

[0050] The status assessment unit is used to quantitatively assess the status of the monitored area based on the video data, AR tag interaction data and environmental perception data, through a multi-dimensional fusion assessment model, and obtain the area status assessment index.

[0051] The scheduling generation unit is used to improve the scheduling based on the regional state evaluation index and preset scheduling rules. The algorithm generates patrol task plans and determines the monitoring round-robin strategy;

[0052] The execution and feedback unit is used to execute the patrol task and rotation strategy through the patrol terminal and monitoring equipment, provide real-time feedback on the execution status, and dynamically adjust the scheduling scheme based on the closed-loop control algorithm.

[0053] A third aspect of the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the computer program to implement the above-described intelligent patrol and rotation scheduling method for video surveillance with fused AR tags.

[0054] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described intelligent patrol and rotation scheduling method for video surveillance fused with AR tags.

[0055] The technical solution of the present invention has the following advantages over the prior art:

[0056] This invention establishes a precise association between physical space, monitoring equipment, and patrol terminals by deploying AR tags, breaking through the limitations of information fragmentation and ambiguous positioning in traditional systems, and enabling rapid tracing and location of abnormal events. Simultaneously, it collects video data, AR tag interaction data, and environmental perception data, combining them with a multi-dimensional fusion evaluation model to quantitatively assess the status of the monitored area, avoiding the one-sidedness of single-data-dimensional assessments and ensuring the accuracy of regional risk perception. Then, based on the regional status assessment index and preset scheduling rules, it intelligently generates patrol task planning and monitoring rotation strategies, changing the resource waste or response delays caused by traditional fixed-route, fixed-cycle scheduling, and achieving "on-demand scheduling." Finally, it dynamically adjusts the scheduling scheme through real-time feedback of execution status, forming a closed-loop mechanism of collection-assessment-scheduling-execution-feedback-adjustment, ensuring that the system can adapt to changes in the status of the monitored area. This overall upgrade achieves intelligent and precise patrol and monitoring rotation, effectively improving the response speed, coverage quality, and resource utilization efficiency of security and prevention. It can be widely applied to multiple fields such as park security, factory inspection, and building monitoring, possessing significant practical value and promising prospects for promotion. Attached Figure Description

[0057] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0058] Figure 1 This is a schematic diagram of the framework of a video surveillance intelligent patrol and rotation scheduling method integrating AR tags according to an embodiment of the present invention;

[0059] Figure 2 This is a schematic diagram of the module framework of a video surveillance intelligent patrol and rotation scheduling system integrating AR tags according to an embodiment of the present invention;

[0060] Figure 3 This is a schematic diagram of the composition of a computing device according to an embodiment of the present invention. Detailed Implementation

[0061] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0062] The present invention will be further described in detail below with reference to the accompanying drawings.

[0063] Reference Figure 1 Firstly, a method for intelligent patrol and rotation scheduling of video surveillance integrating AR tags is provided, including the following steps:

[0064] Deploy AR tags and associate them with physical space and monitoring equipment information;

[0065] Collect video data, AR tag interaction data, and environmental perception data of the monitored area;

[0066] Based on the video data, AR tag interaction data and environmental perception data, the status of the monitored area is quantitatively evaluated through a multi-dimensional fusion evaluation model to obtain the area status evaluation index.

[0067] Based on the regional status assessment index and preset scheduling rules, intelligently generate patrol task planning and monitoring rotation strategies;

[0068] The patrol task and rotation strategy are executed through the patrol terminal and monitoring equipment, and the execution status is fed back in real time, and the scheduling scheme is dynamically adjusted.

[0069] In practice, firstly, AR tags are deployed in the monitored area to establish a relationship between the AR tags and physical space coordinates, as well as monitoring equipment information (including device ID, pixel resolution, shooting angle, and other parameters). Secondly, video data, AR tag interaction data (such as trigger time and terminal location), and environmental perception data (such as temperature, humidity, and smoke concentration) of the monitored area are collected through monitoring equipment, patrol terminals, and environmental sensors. Next, a multi-dimensional fusion evaluation model is constructed, and the collected multi-dimensional data is input to obtain an area status evaluation index through quantitative calculation. Then, based on the evaluation index and a preset risk threshold range, a patrol task plan (such as route and frequency) and a monitoring rotation strategy (such as rotation cycle) matching the area's risk level are generated. Finally, the patrol terminal and monitoring equipment execute the relevant tasks and strategies, provide real-time feedback on the execution status (such as patrol completion rate and image clarity), and dynamically adjust the scheduling scheme based on the feedback data to form a complete scheduling closed loop.

[0070] This invention achieves a deep integration of AR tags, multi-dimensional data acquisition, status assessment, and dynamic scheduling, breaking through the limitations of information fragmentation and fixed scheduling in traditional patrol and monitoring rotation. By associating physical space and equipment information with AR tags, the problem of ambiguous abnormal event location is solved; multi-dimensional data acquisition and fusion assessment ensure the comprehensiveness and accuracy of area status perception; dynamic scheduling based on assessment indices avoids resource waste or response lag; and the real-time feedback adjustment mechanism improves the system's adaptability to changes in area status. Ultimately, it realizes the intelligent and precise upgrade of patrol and monitoring rotation, effectively improving the response speed, coverage quality, and resource utilization efficiency of security and prevention, and can be widely applied to multiple fields such as park security, factory inspection, and building monitoring.

[0071] In some embodiments, deploying AR tags and associating them with physical space and monitoring device information specifically includes:

[0072] Deploy visual AR tags and near-field communication AR tags with built-in unique identifiers at key locations within the monitored area;

[0073] Establish a mapping table that associates AR tag identifiers with physical space coordinates, the ID of the monitoring device to which they belong, and the device parameters, including the pixel resolution, shooting angle, and effective monitoring radius of the monitoring device;

[0074] The association mapping table is stored in the edge computing node to enable real-time access to AR tag information and monitoring devices and physical spaces.

[0075] It should be noted that two types of AR tags are deployed at key locations within the monitored area (such as equipment rooms and fire exits): one type is visual AR tags, which use a composite form combining QR codes and infrared markers; the other type is near-field communication AR tags, supporting NFC and UWB dual-mode communication. Subsequently, a mapping table is established linking the unique identifier of each AR tag with its physical spatial coordinates (using the WGS-84 coordinate system), the ID of the associated monitoring device, and the device parameters (pixel resolution, shooting angle, effective monitoring radius). Finally, this mapping table is stored on an edge computing node to reduce data transmission latency and ensure real-time access to AR tag information, physical spatial information, and monitoring device information.

[0076] This invention, through the deployment of dual-mode AR tags, ensures effective tag recognition in complex environments (such as low light and occlusion) while achieving centimeter-level positioning accuracy and near-field identity verification, avoiding the limitations of a single tag type. The establishment of the association mapping table realizes a three-in-one association between physical space, monitoring equipment, and AR tags, providing a foundation for rapid tracing of subsequent abnormal events and precise equipment scheduling. The storage design of the edge computing nodes improves information retrieval efficiency, reduces data transmission loss, and ensures real-time system response.

[0077] In some embodiments, the collection of video data, AR tag interaction data, and environmental perception data of the monitored area specifically includes:

[0078] Real-time video streams are collected by monitoring equipment, and features of moving targets, abnormal behavior, and scene changes are extracted from the video frames.

[0079] By interacting with AR tags through near-field communication, the system collects tag trigger time, trigger frequency, and patrol terminal location information.

[0080] Environmental data is collected by sensors deployed in the monitoring area, including light intensity, temperature and humidity, smoke concentration and vibration frequency.

[0081] The video data, AR tag interaction data, and environmental data are time-stamped and cleaned. A Gaussian filtering algorithm is used to remove noisy data and retain valid data samples.

[0082] It should be noted that, regarding video data acquisition, real-time video streams are obtained through monitoring equipment, and the YOLOv8 algorithm is used to extract moving target features (such as target type and trajectory), abnormal behavior features (such as intrusion and loitering), and scene change features (such as door and window status) from the video frames. For AR tag interaction data acquisition, near-field communication between the patrol terminal and the AR tags is used to obtain tag trigger time, trigger frequency, and patrol terminal location information (UWB positioning data update frequency is 10Hz). For environmental perception data acquisition, corresponding environmental parameters are obtained through deployed sensors (smoke, vibration, temperature and humidity, and light sensors). Subsequently, the three types of data are timestamped (using unified UTC time with millisecond accuracy), and a Gaussian filtering algorithm is used to remove noise data, retaining valid data samples. The collection of multi-dimensional data covers core dimensions such as abnormal behavior, personnel activities, and environmental safety in the monitored area, providing comprehensive data support for subsequent status assessment. The application of the YOLOv8 algorithm ensures the accuracy and efficiency of abnormal feature extraction from the video. The high-frequency update of UWB positioning ensures the real-time nature of AR interactive data. Timestamp alignment and Gaussian filtering preprocessing effectively improve data quality, avoid interference from noisy data on subsequent assessment models, and ensure the reliability of assessment results.

[0083] In some embodiments, based on the video data, AR tag interaction data, and environmental perception data, a multi-dimensional fusion evaluation model is used to quantitatively evaluate the status of the monitored area, resulting in an area status evaluation index, which specifically includes:

[0084] A multi-dimensional fusion evaluation model is constructed, and the input dimensions of the model include video anomaly feature weights, AR tag interaction activity, and environmental risk level.

[0085] The weight coefficients for each input dimension are determined using the Analytic Hierarchy Process (AHP): Video anomaly feature weights. AR tag interaction activity weight Environmental risk level weights ,satisfy Among them, the consistency test index CR of the judgment matrix of the analytic hierarchy process is <0.1;

[0086] Calculate the quantized values ​​for each dimension:

[0087] Video anomaly quantization value , ,in This is the severity coefficient (range 1-5) for the i-th type of abnormal behavior (intrusion, loitering, object movement, etc.). Let n be the detection confidence level for the i-th type of abnormal behavior (range 0-1), and n be the total number of abnormal behavior categories.

[0088] AR tag interaction activity ,in The importance coefficient for labels (λ=1.5 for key locations, λ=1.5 for ordinary locations) =1.0), The actual trigger duration per unit of time. The threshold value is 30 minutes (default), and k is the decay coefficient (value 0.8).

[0089] Environmental risk level , ,in The risk weights for the j-th type of environmental parameters are: smoke concentration b=0.4, vibration frequency b=0.3, temperature and humidity b=0.15, and light intensity b=0.15. Let be the normalized detection value of the j-th type of environmental parameter (range 0-1), and m be the total number of environmental parameter categories;

[0090] The regional status assessment index is calculated using a weighted fusion formula.

[0091] ,

[0092] in This is the covariance correction factor (values ​​range from 0.05 to 0.1). The trace of the covariance matrix of V, A, and E is used to correct for the correlation effects of multi-dimensional data. The regional state assessment index S ranges from 0 to 10.

[0093] In practice, the process involves several steps. First, a multi-dimensional fusion evaluation model is constructed, with the input dimensions being video anomaly features, AR tag interaction activity, and environmental risk level. Second, an analytic hierarchy process (AHP) is used to invite security experts to construct a judgment matrix. After a consistency test (CR < 0.1), the weight coefficients for each dimension are determined (e.g., video anomaly feature weight 0.45, AR interaction activity weight 0.3, and environmental risk level weight 0.25). Next, the quantitative values ​​for each dimension are calculated: the video anomaly quantification value V is obtained by summing the product of the anomaly severity coefficient and the detection confidence; the AR tag interaction activity A is calculated by the tag importance coefficient, the trigger duration ratio, and the decay function; and the environmental risk level E is obtained by summing the product of the environmental parameter risk weight and the normalized detection value. Finally, the regional status evaluation index S is calculated using a weighted fusion formula (introducing a covariance correction term) to achieve a quantitative evaluation of the monitored area's status. Multi-dimensional input ensures the comprehensiveness of the assessment and avoids the one-sidedness of single-dimensional assessment; the analytic hierarchy process (AHP) determines the weight coefficients, and combined with expert experience and consistency checks, ensures the rationality of weight allocation; the design of quantitative formulas for each dimension realizes the transformation of qualitative characteristics into quantitative indicators, improving the objectivity of the assessment; the introduction of covariance correction terms corrects the correlation between multi-dimensional data, further improving the accuracy of the assessment index and providing a reliable quantitative basis for subsequent differentiated scheduling.

[0094] In some embodiments, the intelligent generation of patrol task planning and monitoring rotation strategies based on the regional status assessment index and preset scheduling rules specifically includes:

[0095] Preset status assessment threshold range: Low risk range Medium-risk area High-risk areas ;

[0096] When the regional status assessment index S falls within the low-risk range, a regular patrol route is generated, employing a uniform rotation strategy with a rotation cycle of [missing information]. ,in The basic patrol cycle (default 60 minutes);

[0097] When S falls within the medium-risk range, an optimized patrol route is generated, increasing the patrol frequency of key locations (by 50% compared to the regular route). A weighted rotation patrol strategy is adopted, with a rotation cycle of... ,in This is a risk correction factor (with a value of 0.6-0.8).

[0098] When S belongs to a high-risk zone, an emergency patrol route is generated, prioritizing the dispatch of patrol terminals closest to the target area, and employing a dense patrol strategy with a patrol cycle of [missing information]. ,in This is the emergency response coefficient (value range: 0.3-0.5).

[0099] Based on the physical spatial coordinates associated with AR tags, by improving Algorithm for planning patrol routes, path cost function

[0100] ,

[0101] in Let k be the length of the path segment k (in meters). This is the regional status assessment index corresponding to the k-th path segment. This is the path risk weight (value 0.3). This represents the estimated travel time (in seconds) for the k-th segment of the path. The speed of the patrol terminal (unit: meters per second). The time cost weight is 0.4. Let be the obstacle density of the k-th path segment (values ​​0-1). This is the obstacle avoidance coefficient (values ​​range from 1.2 to 2.0). The obstacle weight is 0.3, and I is the total number of path segments.

[0102] In practice, firstly, the risk threshold range of the regional status assessment index is preset: low risk range. Medium-risk area High-risk areas Secondly, corresponding scheduling strategies are generated based on the interval to which the assessment index belongs: Low-risk intervals employ regular patrol routes and a uniform rotation strategy, with the rotation cycle calculated using a base cycle and risk coefficient; medium-risk intervals employ optimized patrol routes (increasing the frequency of key points) and a weighted rotation strategy, with the rotation cycle adjusted using a risk correction coefficient; high-risk intervals employ emergency patrol routes (prioritizing scheduling near-terminal locations) and a dense rotation strategy, with the rotation cycle additionally incorporating an emergency response coefficient. Finally, based on the physical spatial coordinates associated with AR tags, an improved A* algorithm is used to plan the optimal patrol route by constructing a path cost function that includes path length, time cost, and obstacle factors. The division of risk threshold intervals enables hierarchical management of regional risks, providing a clear basis for differentiated scheduling; the patrol and rotation strategies corresponding to different risk intervals ensure "fewer patrols for low-risk, optimized patrols for medium-risk, and dense patrols for high-risk," avoiding resource waste and response delays; improvements... The algorithm's path planning incorporates multi-dimensional cost factors to ensure that patrol routes balance risk coverage, time efficiency, and accessibility, further improving the efficiency and rationality of patrol tasks.

[0103] In some embodiments, the patrol task and rotation strategy are executed through the patrol terminal and monitoring equipment, and the execution status is fed back in real time. The scheduling scheme is dynamically adjusted, specifically including:

[0104] The patrol terminal confirms arrival at patrol points through dual verification via visual recognition of AR tags (QR code parsing) and near-field communication (UWB positioning), with a verification success rate of ≥99.5%.

[0105] The monitoring equipment adjusts the shooting parameters, including focal length, exposure and shooting angle, according to the patrol strategy, and performs close-up shooting of high-risk areas (resolution ≥1080P, frame rate ≥30fps).

[0106] The patrol terminal and monitoring equipment upload execution data in real time through the 5G edge computing network, including patrol completion rate, monitoring screen clarity, and abnormal event handling results;

[0107] The regional status assessment index is updated based on real-time feedback data. The patrol route and patrol cycle are dynamically adjusted through a closed-loop control algorithm. The adjustment formula is as follows: ,in This is the updated state assessment index. For the updated patrol cycle, To adjust the sensitivity coefficient (value 0.1-0.2), R is the execution status feedback coefficient (R=0.8 when the patrol completion rate is ≥95%; R=0.5 when the patrol completion rate is ≤80% and <95%; R=0.2 when the patrol completion rate is <80%).

[0108] In practice, firstly, the patrol terminal performs dual verification through QR code parsing of the visual AR tag and UWB positioning of the near-field communication AR tag to confirm the arrival at the patrol point; secondly, the monitoring equipment adjusts the shooting parameters according to the patrol strategy, performing close-up shooting (resolution ≥1080P, frame rate ≥30fps) on high-risk areas; nextly, the patrol terminal and the monitoring equipment upload execution data (patrol completion rate, image clarity, and anomaly handling results) in real time through a 5G edge computing network; finally, based on the real-time feedback data, the area status assessment index is updated through a closed-loop control algorithm. ) and the cycle of rotation ( The system dynamically adjusts patrol routes and rotation strategies. A dual verification mechanism effectively avoids false check-ins in traditional patrols, ensuring the genuine execution of patrol tasks. Close-up photography of high-risk areas ensures clear capture of abnormal details, facilitating subsequent handling. The application of 5G edge computing networks improves the upload speed of execution data, ensuring real-time feedback. The introduction of a closed-loop control algorithm enables dynamic adjustment of the scheduling scheme, allowing the system to adapt to changes in the state of the monitored areas, improving the system's robustness and adaptability.

[0109] In some embodiments, the visual AR tag is a composite tag combining a QR code and an infrared marker. The near-field communication AR tag supports dual-mode communication of NFC and UWB. The positioning accuracy of UWB communication is ≤5 cm, and the effective distance of NFC communication is ≤10 cm. It should be noted that the visual AR tag is made of wear-resistant, waterproof, and corrosion-resistant materials to adapt to different application environments such as outdoor and computer room environments. The near-field communication AR tag supports dual-mode communication of NFC and UWB, where the positioning accuracy of UWB communication is ≤5 cm, and the effective distance of NFC communication is ≤10 cm. The tag also has a built-in low-power battery with a battery life of ≥1 year and supports a low-power wake-up mode, activating the communication function only when a patrol terminal is detected approaching.

[0110] The special materials of the visual AR tags ensure their durability in complex environments and extend their service life; the dual-mode near-field communication tags achieve high-precision positioning through UWB and near-field authentication through NFC, ensuring the reliability of positioning and verification; the low-power design and long-lasting battery significantly reduce the frequency and cost of tag maintenance, reduce the overall system maintenance overhead, and improve the system's usability.

[0111] Reference Figure 2 The second aspect of the present invention provides a video surveillance intelligent patrol and rotation scheduling system integrating AR tags, comprising:

[0112] The tag deployment and association unit is used to deploy AR tags and associate them with physical space and monitoring equipment information, establish an association mapping table and store it to the edge computing node;

[0113] The data acquisition unit is used to collect video data, AR tag interaction data and environmental perception data of the monitored area, and to perform timestamp alignment and data cleaning.

[0114] The status assessment unit is used to quantitatively assess the status of the monitored area based on the video data, AR tag interaction data and environmental perception data, through a multi-dimensional fusion assessment model, and obtain the area status assessment index.

[0115] The scheduling generation unit is used to improve the scheduling based on the regional state evaluation index and preset scheduling rules. The algorithm generates patrol task plans and determines the monitoring round-robin strategy;

[0116] The execution and feedback unit is used to execute the patrol task and rotation strategy through the patrol terminal and monitoring equipment, provide real-time feedback on the execution status, and dynamically adjust the scheduling scheme based on the closed-loop control algorithm.

[0117] It should be noted that the system of this invention includes five core units: a tag deployment and association unit, consisting of a tag deployment module (planning and installing AR tags), an association mapping module (establishing associations between tags and physical spaces and devices), and a data storage module (storing the association mapping table to edge nodes); a data acquisition unit, consisting of a video acquisition module (extracting video features), an AR interaction acquisition module (acquiring tag interaction data), an environment acquisition module (acquiring environmental parameters), and a data preprocessing module (aligning and cleaning data); a status evaluation unit, consisting of a weight determination module (AHP determining weights), a dimension quantification module (calculating values ​​for each dimension), and a fusion evaluation module (generating an evaluation index); and a scheduling generation unit, consisting of a threshold judgment module (determining risk levels), a patrol path planning module (improving...). The system comprises an algorithm path planning module and a polling strategy generation module (determining polling cycle parameters); and an execution and feedback unit, consisting of a task execution module (controlling the terminal and equipment to execute tasks), a status feedback module (uploading execution data), and a dynamic adjustment module (closed-loop adjustment of the scheduling scheme). Each unit collaborates through data interaction, achieving intelligent operation throughout the entire process from tag deployment to dynamic scheduling. The modular design of the system clearly defines the responsibilities of each functional unit, facilitating later maintenance and functional expansion. The collaborative work between units achieves end-to-end intelligent operation from data acquisition to status assessment to schedule generation to execution feedback, avoiding the drawbacks of independent operation in traditional systems. Efficient data interaction between units ensures the smoothness and real-time performance of the overall system operation, further enhancing the intelligence level of patrol and monitoring rotation.

[0118] Reference Figure 3 The third aspect of this invention provides a computer device, including a memory and a processor. The memory stores a computer program, and the processor is configured to execute the computer program to implement the aforementioned intelligent patrol and rotation scheduling method for video surveillance integrating AR tags. This computer device provides a hardware platform for the aforementioned intelligent patrol and rotation scheduling method for video surveillance integrating AR tags, ensuring that the method can be implemented. The memory's storage of the program ensures the stability and reusability of the method's logic. The processor's computing power ensures multi-dimensional data processing and complex model calculations (such as multi-dimensional fusion evaluation and improvement). The high efficiency of the algorithm can adapt to the data processing needs of large-scale monitoring areas, providing support for the real-time performance and reliability of the system.

[0119] In some embodiments, the video surveillance intelligent patrol and rotation scheduling method integrating AR tags in the above embodiments can be implemented by a computer device, which includes at least one processor, a communication bus, a memory, and at least one communication interface.

[0120] A processor can be a general-purpose central processing unit (CPU) or an application-specific integrated circuit (ASIC).

[0121] A communication bus can be used to transmit information between the aforementioned components.

[0122] The memory can be read-only memory (ROM) or other types of static storage devices capable of storing static information and instructions, random access memory (RAM) or other types of dynamic storage devices capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, universal optical discs, Blu-ray discs, etc.), magnetic disks or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited to these. The memory can exist independently and be connected to the processor via a communication bus. The memory can also be integrated with the processor.

[0123] The memory stores the program code that executes the solution of this application, and its execution is controlled by the processor. The processor executes the program code stored in the memory. The program code may include one or more software modules. In the above embodiments, the video surveillance intelligent patrol and rotation scheduling method integrating AR tags can be implemented by the processor and one or more software modules in the program code in the memory.

[0124] A communication interface is a device that uses any transceiver or similar device to communicate with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0125] In a specific implementation, as one example, a computer device may include multiple processors, each of which may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. Here, a processor may refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).

[0126] The aforementioned computer device can be a general-purpose computer device or a special-purpose computer device. In specific implementations, the computer device can be a desktop computer, a portable computer, a network server, a handheld digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. This application does not limit the type of computer device.

[0127] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described intelligent patrol and rotation scheduling method for video surveillance fused with AR tags.

[0128] It should be noted that the computer-readable storage medium (such as hard disk, USB flash drive, optical disk, ROM, etc.) stores a computer program containing code instructions for implementing a video surveillance intelligent patrol and rotation scheduling method that integrates AR tags. When the computer program is read and executed by a processor (such as a computer or server processor), the processor will execute the following steps in sequence according to the logic in the program: AR tag deployment and association, multi-dimensional data collection and preprocessing, regional status quantitative assessment, patrol and rotation strategy generation, execution feedback and dynamic adjustment, thereby realizing the aforementioned video surveillance intelligent patrol and rotation scheduling method that integrates AR tags.

[0129] Computer-readable storage media facilitate the storage and dissemination of programs that implement the above scheduling methods, enabling the methods to be deployed and run on different computer devices, thus improving the practicality and portability of the methods. At the same time, the stability of the storage media ensures the long-term preservation of the programs, preventing program loss or damage, providing a guarantee for the continuous application of the methods, and lowering the threshold for the promotion and application of the methods.

[0130] To provide a clearer understanding of the invention, the invention is further described below:

[0131] In a first aspect, this application provides a video surveillance intelligent patrol and rotation scheduling method integrating AR tags, comprising the following steps:

[0132] Deploy AR tags and associate them with physical space and monitoring equipment information;

[0133] Collect video data, AR tag interaction data, and environmental perception data of the monitored area;

[0134] Based on the video data, AR tag interaction data and environmental perception data, the status of the monitored area is quantitatively evaluated through a multi-dimensional fusion evaluation model to obtain the area status evaluation index.

[0135] Based on the regional status assessment index and preset scheduling rules, intelligently generate patrol task planning and monitoring rotation strategies;

[0136] The patrol task and rotation strategy are executed through the patrol terminal and monitoring equipment, and the execution status is fed back in real time, and the scheduling scheme is dynamically adjusted.

[0137] Preferably, the deployment of AR tags and association of physical space with monitoring equipment information specifically includes: deploying visual AR tags and near-field communication AR tags with built-in unique identifiers at key locations in the monitoring area (such as equipment rooms, fire exits, hazardous materials storage areas, entrances and exits, etc.); wherein, the visual AR tags are composite tags combining QR codes and infrared markers to ensure effective identification even in complex environments such as low light and obstruction; the near-field communication AR tags support NFC and UWB dual-mode communication, with UWB communication achieving centimeter-level positioning and NFC communication used for near-field authentication, providing dual protection for positioning and verification accuracy.

[0138] Establish a mapping table that associates AR tag identifiers with physical space coordinates (using the WGS-84 coordinate system), the ID of the monitoring device to which they belong, and device parameters. The device parameters include the pixel resolution of the monitoring device (e.g., 4K, 1080P) and the shooting angle (e.g., horizontal). ,vertical The effective monitoring radius (e.g., 50 meters, 100 meters) is defined; the associated mapping table is stored in the edge computing node to reduce data transmission latency and enable real-time access to AR tag information and monitoring equipment and physical space.

[0139] Preferably, the acquisition of video data, AR tag interaction data, and environmental perception data of the monitored area specifically includes: acquiring real-time video streams through monitoring equipment (such as network cameras and PTZ cameras), and using the YOLOv8 algorithm to extract moving target features (such as target type, motion trajectory, and speed), abnormal behavior features (such as entering prohibited areas, prolonged loitering, and illegal movement of objects), and scene change features (such as changes in door and window status and equipment position shifts) from the video frames; acquiring tag trigger time, trigger frequency, and patrol terminal location information through near-field communication interaction between patrol terminals (such as AR glasses and smart handheld terminals) and AR tags, wherein the UWB positioning data update frequency is 10Hz to ensure the real-time nature of the location information; and acquiring environmental data through sensors deployed in the monitored area (such as smoke sensors, vibration sensors, temperature and humidity sensors, and light sensors), wherein the environmental data includes light intensity (unit: lux), temperature and humidity (unit: lux), and environmental data such as light intensity (unit: lux), temperature and humidity (unit: lux). Humidity unit: The video data, AR tag interaction data, and environmental data were timestamped (using UTC timestamps with millisecond accuracy) and cleaned. A Gaussian filtering algorithm was used to remove noisy data and retain valid data samples to ensure the quality of input data for subsequent evaluation models.

[0140] Preferably, the step of quantitatively evaluating the status of the monitored area based on the video data, AR tag interaction data, and environmental perception data using a multi-dimensional fusion evaluation model to obtain the area status evaluation index specifically includes: constructing a multi-dimensional fusion evaluation model, wherein the input dimensions of the model include video anomaly feature weights, AR tag interaction activity, and environmental risk levels, comprehensively covering core dimensions such as behavioral anomalies, personnel activities, and environmental safety in the monitored area; determining the weight coefficients of each input dimension using the analytic hierarchy process (AHP); inviting 5-10 security experts to construct a judgment matrix, comparing the importance of each input dimension pairwise, and ensuring the rationality of the weight allocation through a consistency test (CR < 0.1), ultimately determining the video anomaly feature weights. AR tag interaction activity weight Environmental risk level weights ,satisfy ; Calculate the quantized values ​​for each dimension:

[0141] Video anomaly quantization value , n represents the total number of abnormal behavior categories (in this example, n=5, including intrusion, loitering, object movement, climbing, and destruction). is the severity coefficient of the i-th type of abnormal behavior (intrude a=5, loiter a=3, move object a=4, climb a=5, vandalize a=5). The detection confidence score for the i-th type of anomalous behavior (output by the YOLOv8 algorithm, ranging from 0 to 1); for example, if an intrusion is detected ( ) and loitering behavior ( ),but ;

[0142] AR tag interaction activity : The importance coefficient of the label (critical locations such as hazardous materials storage areas) Ordinary locations such as corridors ), For a unit of time ( The actual trigger duration (in minutes), where k is a decay coefficient (valued at 0.8) used to avoid distortion in activity calculation when the trigger duration deviates excessively from the threshold; for example, the trigger duration of key points within 30 minutes. minutes, then ;

[0143] Environmental risk level , m represents the total number of environmental parameter categories (m=4). The risk weights for the j-th type of environmental parameters are: smoke concentration b=0.4, vibration frequency b=0.3, temperature and humidity b=0.15, and light intensity b=0.15. This represents the normalized detection value of the j-th type of environmental parameter (the original detection value is mapped to the 0-1 range through min-max normalization; the larger the value, the higher the risk). For example, if the normalized values ​​are: smoke concentration c=0.8, vibration frequency c=0.6, temperature and humidity c=0.3, and light intensity c=0.2, then... ;

[0144] The regional status assessment index is calculated using a weighted fusion formula.

[0145] : Covariance correction coefficient (in this embodiment) ), The trace of the covariance matrix of V, A, and E is used to correct for the correlation effects of multi-dimensional data (for example, abnormal behavior and smoke concentration may be positively correlated, and covariance correction can improve the accuracy of the assessment); continuing the example above, the trace of the covariance matrix ,but It falls within the medium-risk range.

[0146] Preferably, the step of intelligently generating patrol task planning and monitoring rotation strategy based on the regional state assessment index and preset scheduling rules specifically includes: a preset state assessment threshold range: a low-risk range. Medium-risk area High-risk areas The threshold is determined through statistical analysis of historical abnormal event data; when the regional status assessment index S belongs to the low-risk range, a regular patrol route is generated, covering all preset points, and a uniform rotation strategy is adopted, with a rotation cycle of... ( =60 minutes, if S=2, then =150 minutes); when S belongs to the medium-risk range, an optimized patrol route is generated by increasing key points (such as AR tag importance coefficient). The patrol frequency for points with a radius of 1.5 is increased by 50% compared to the regular route, using a weighted rotation patrol strategy, with a rotation cycle of... ;

[0147] in, =0.7, if S=3.64, then Minutes; when S is in a high-risk zone, an emergency patrol route is generated, and the patrol terminal closest to the target area (response time ≤ 5 minutes) is selected by GPS positioning. A dense patrol strategy is adopted, with a patrol cycle of [missing information]. ( If S=8, then (minutes); based on the physical spatial coordinates associated with AR tags, through improvements Algorithm for planning patrol routes, path cost function , : I represents the total number of path segments, Let k be the length of the path segment. This is the regional status assessment index corresponding to the k-th path segment. =0.3, Let be the estimated travel time for the k-th segment of the path. The speed at which the patrol terminal moves (default 1.2 m / s). =0.4, The obstacle density of the k-th path segment (calculated using environmental sensor data). This is the obstacle avoidance coefficient. =0.3; The improved A* algorithm uses this cost function to select the optimal path, ensuring that the path takes into account risk coverage, time efficiency and travel convenience.

[0148] Preferably, the execution of the patrol task and rotation strategy through the patrol terminal and monitoring equipment, and the real-time feedback of the execution status, and the dynamic adjustment of the scheduling scheme specifically include: the patrol terminal confirming the arrival of patrol points through dual verification of AR tag visual recognition (QR code parsing success rate ≥99%) and near-field communication (UWB positioning) to avoid false check-ins; the monitoring equipment adjusting shooting parameters according to the rotation strategy, performing close-up shooting (resolution ≥1080P, frame rate ≥30fps) on high-risk areas to ensure that abnormal details are clearly identifiable; the patrol terminal and monitoring equipment uploading execution data in real time through the 5G edge computing network, including patrol completion rate (actual completed points / planned points), monitoring image clarity (evaluated by image entropy value, entropy value ≥5.0 is qualified), and abnormal event handling results (handled / not handled / handling); updating the area status evaluation index based on real-time feedback data, and dynamically adjusting the patrol route and rotation cycle through a closed-loop control algorithm, the adjustment formula being: , in R is the execution status feedback coefficient (R=0.8 when the patrol completion rate is ≥95%; R=0.5 when 80%≤patrol completion rate<95%; R=0.2 when patrol completion rate<80%). For example, if the original S=3.64 and the patrol completion rate=96% (R=0.8), then...

[0149] ,

[0150] minute.

[0151] Preferably, the visual label of the AR tag is made of wear-resistant, waterproof and corrosion-resistant material, which is suitable for different environments such as outdoors and computer rooms; the battery life of the near-field communication tag is ≥1 year, and it supports low power wake-up mode to reduce maintenance costs.

[0152] To better understand the technical solution of this application, the technical solution of this application will be described in detail below with reference to specific embodiments.

[0153] Example 1: Application of Park Security Scenarios

[0154] AR tags are deployed at 20 key locations within the park, including entrances and exits, equipment rooms, fire exits, hazardous materials warehouses, and parking lots. Each tag is assigned a unique identifier ID. A mapping is established between the tag ID and its physical coordinates (e.g., latitude). East longitude ), the ID of the monitoring device to which it belongs (e.g.) to ), Equipment parameters ( 4K resolution, horizontal The associated mapping table (shooting angle, effective monitoring radius 80 meters) is stored at the park edge computing node;

[0155] Data Acquisition: Real-time video streams were acquired through 20 monitoring devices, and the YOLOv8 algorithm was used to extract features of moving targets and abnormal behaviors; patrol personnel wore AR glasses (patrol terminals) and used AR tags for UWB positioning and NFC verification to collect trigger time and terminal location; sensors were deployed at various locations to collect data on light intensity, temperature and humidity, smoke concentration, and vibration frequency; all data were timestamped (UTC time, accuracy 1ms) and Gaussian filtering was used to remove noise;

[0156] State assessment: Weights are determined using the analytic hierarchy process. , , If, during a certain period, personnel are detected loitering in the hazardous materials warehouse area (s=0.9), then V=3×0.9=2.7; AR tag trigger duration. minute( ),but

[0157] After environmental data normalization: smoke concentration Vibration frequency Temperature and humidity Light intensity ,but Trace of covariance matrix , ,but ≈ This falls within a low-risk range.

[0158] Scheduling Generation: Generates a regular patrol route, covering 20 locations, with a rotation cycle. minutes; through improvement After the algorithm plans the path and calculates the path cost function, the optimal route is determined as follows:

[0159] (Sorted by lowest cost);

[0160] Execution and Feedback: Patrol personnel execute patrols according to the planned route; AR glasses confirm arrival through dual verification of tags; monitoring equipment rotates evenly every 180 minutes; real-time upload of patrol completion rate = 98% (R=0.8). , minute.

[0161] Example 2: Factory Inspection Scenario Application

[0162] Tag deployment: AR tags are deployed at 15 locations in the factory, including the production workshop, power distribution room, warehouse, and sewage outlet. The tags support UWB positioning (accuracy 3cm) and NFC communication; an association mapping table is established between the tags and physical coordinates, monitoring equipment (15 units), and equipment parameters.

[0163] Data Acquisition: Monitoring equipment collects video streams and detects abnormal equipment operation (such as machine oil leaks or parts falling off); patrol terminals (smart handheld terminals) collect AR tag interaction data; sensors collect workshop temperature and humidity, power distribution room vibration frequency, and sewage outlet smoke concentration.

[0164] Condition assessment: An abnormal vibration frequency was detected in the power distribution room during a certain period of time. No abnormal human behavior was observed (V=0); AR tag trigger duration minute( ), Environmental data Trace of covariance matrix ,

[0165] (Low risk) Subsequent detection showed a continuous increase in vibration frequency ( And people approached (V=5×0.85=4.25).

[0166] but (Still low risk) Further detection of smoke concentration ,but

[0167] ,

[0168] (Low risk) The final smoke concentration was detected. And people broke in ( ),but

[0169] ,

[0170] ,

[0171] (Approaching the medium-risk threshold);

[0172] Scheduling generation: Round-robin cycle Minutes; route planning prioritizes locations around the power distribution room;

[0173] Execution and Feedback: Upon arriving at the power distribution room, the patrol personnel discovered that the equipment connectors were overheating and reported that the abnormality was being handled (R=0.5). (Entering medium risk) Minutes later, the dispatch system increased the frequency of monitoring and patrolling the power distribution room and adjusted the shooting parameters to close-up mode.

[0174] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects:

[0175] Integrating AR tags to achieve multi-dimensional association: By deploying visual AR tags and near-field communication AR tags in a dual-mode manner, a three-in-one association mapping of physical space locations, monitoring equipment, and patrol terminals is established, which solves the problems of information fragmentation and inaccurate positioning in traditional systems, and enables rapid location and accurate response to abnormal events;

[0176] Multi-dimensional integrated assessment improves the accuracy of state perception: Construct a multi-dimensional assessment model covering video anomalies, AR interaction, and environmental parameters, introduce the analytic hierarchy process to determine weights and covariance correction mechanism, quantify the regional state assessment index, avoid the one-sidedness of single-dimensional assessment, and improve the accuracy of risk identification.

[0177] Dynamic scheduling strategy optimizes resource allocation: Based on the regional status assessment index, the patrol route, frequency and monitoring cycle are dynamically adjusted. An improved A* algorithm is used to plan the optimal patrol path to achieve differentiated scheduling of "low-risk less patrol, medium-risk optimized patrol, and high-risk intensive patrol" to avoid resource waste and improve the speed of anomaly response.

[0178] Closed-loop control ensures real-time scheduling: Through real-time status feedback from patrol terminals and monitoring equipment, the regional assessment index and scheduling scheme are dynamically updated, forming a closed-loop mechanism of collection-assessment-scheduling-execution-feedback-adjustment, which adapts to the dynamic changes in the status of the monitored area and improves the robustness and adaptability of the system.

[0179] In summary, this application's solution achieves intelligent and precise upgrades to patrol and monitoring rotation through the deep integration of AR tag fusion, multi-dimensional evaluation, and dynamic scheduling. It effectively solves the problems of low efficiency, slow response, and resource waste in traditional solutions, and can be widely applied in multiple fields such as park security, factory inspection, and building monitoring, demonstrating significant practical value and promising prospects for promotion.

[0180] The technical solutions provided by the embodiments of the present invention have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of the embodiments of the present invention. The descriptions of the embodiments above are only for helping to understand the principles of the embodiments of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the embodiments of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A video surveillance intelligent patrol and rotation scheduling method integrating AR tags, characterized in that, Includes the following steps: Deploy AR tags and associate them with physical space and monitoring equipment information; Collect video data, AR tag interaction data, and environmental perception data of the monitored area; Based on the video data, AR tag interaction data and environmental perception data, the status of the monitored area is quantitatively evaluated through a multi-dimensional fusion evaluation model to obtain the area status evaluation index. Based on the regional status assessment index and preset scheduling rules, intelligently generate patrol task planning and monitoring rotation strategies; The patrol task and rotation strategy are executed through the patrol terminal and monitoring equipment, and the execution status is fed back in real time, and the scheduling scheme is dynamically adjusted.

2. The video surveillance intelligent patrol and rotation scheduling method integrating AR tags according to claim 1, characterized in that: Deploying AR tags and associating them with physical space and monitoring equipment information specifically includes: Deploy visual AR tags and near-field communication AR tags with built-in unique identifiers at key locations within the monitored area; Establish a mapping table that associates AR tag identifiers with physical space coordinates, the ID of the monitoring device to which they belong, and the device parameters, including the pixel resolution, shooting angle, and effective monitoring radius of the monitoring device; The association mapping table is stored in the edge computing node to enable real-time access to AR tag information and monitoring devices and physical spaces.

3. The video surveillance intelligent patrol and rotation scheduling method integrating AR tags according to claim 1, characterized in that: The specific data collected, including video data, AR tag interaction data, and environmental perception data of the monitored area, includes: Real-time video streams are collected by monitoring equipment, and features of moving targets, abnormal behavior, and scene changes are extracted from the video frames. By interacting with AR tags through near-field communication, the system collects tag trigger time, trigger frequency, and patrol terminal location information. Environmental data is collected by sensors deployed in the monitoring area, including light intensity, temperature and humidity, smoke concentration and vibration frequency. The video data, AR tag interaction data, and environmental data are time-stamped and cleaned. A Gaussian filtering algorithm is used to remove noisy data and retain valid data samples.

4. The video surveillance intelligent patrol and rotation scheduling method integrating AR tags according to claim 1, characterized in that: Based on the video data, AR tag interaction data, and environmental perception data, a multi-dimensional fusion evaluation model is used to quantitatively evaluate the status of the monitored area, resulting in an area status evaluation index that specifically includes: A multi-dimensional fusion evaluation model is constructed, and the input dimensions of the model include video anomaly feature weights, AR tag interaction activity, and environmental risk level. The weight coefficients of each input dimension are determined using the analytic hierarchy process (AHP): video anomaly feature weights. AR tag interaction activity weight Environmental risk level weights ,satisfy Among them, the consistency test index CR of the judgment matrix of the analytic hierarchy process is <0.1; Calculate the quantized values ​​for each dimension: Video anomaly quantization value , ,in Let be the severity coefficient of the i-th type of abnormal behavior. Let n be the detection confidence score for the i-th type of abnormal behavior, and n be the total number of abnormal behavior categories. AR tag interaction activity ,in This is the label importance coefficient. The actual trigger duration per unit of time. The threshold value is k, where k is the attenuation coefficient. Environmental risk level , ,in Let the risk weight be the type j environmental parameter. Let be the normalized detection value of the j-th type of environmental parameter, and m be the total number of environmental parameter categories; The regional status assessment index is calculated using a weighted fusion formula. ， in This is the covariance correction factor. The trace of the covariance matrix of V, A, and E is used to correct for the correlation effects of multi-dimensional data. The regional state assessment index S ranges from 0 to 10.

5. The video surveillance intelligent patrol and rotation scheduling method integrating AR tags according to claim 1, characterized in that: Based on the regional status assessment index and preset scheduling rules, the intelligent generation of patrol task planning and monitoring rotation strategies specifically includes: Preset status assessment threshold range: Low risk range Medium-risk area High-risk areas ; When the regional status assessment index S falls within the low-risk range, a regular patrol route is generated, employing a uniform rotation strategy with a rotation cycle of [missing information]. ,in The basic rotation cycle; When S falls within the medium-risk range, an optimized patrol route is generated, increasing the patrol frequency at key locations, and a weighted rotation patrol strategy is adopted, with a rotation cycle of [missing information]. ,in This is a risk correction factor; When S belongs to a high-risk zone, an emergency patrol route is generated, prioritizing the dispatch of patrol terminals closest to the target area, and employing a dense patrol strategy with a patrol cycle of [missing information]. ,in This refers to the emergency response coefficient. Based on the physical spatial coordinates associated with AR tags, by improving Algorithm for planning patrol routes, path cost function ， in Let k be the length of the path segment. This is the regional status assessment index corresponding to the k-th path segment. For path risk weights, Let be the estimated travel time for the k-th segment of the path. To improve the movement speed of the patrol terminal, Weighted by time cost, Let K be the obstacle density of the k-th path segment. This is the obstacle avoidance coefficient. Here, I represents the obstacle weights, and I represents the total number of path segments.

6. The video surveillance intelligent patrol and rotation scheduling method integrating AR tags according to claim 1, characterized in that: The patrol tasks and rotation strategies are executed through patrol terminals and monitoring equipment, and the execution status is fed back in real time. The scheduling scheme is dynamically adjusted, specifically including: The patrol terminal confirms arrival at patrol points through dual verification of visual recognition and near-field communication using AR tags, with a verification success rate of ≥99.5%. The monitoring equipment adjusts the shooting parameters, including focal length, exposure and shooting angle, according to the patrol strategy, and performs close-up shooting of high-risk areas; The patrol terminal and monitoring equipment upload execution data in real time through the 5G edge computing network, including patrol completion rate, monitoring screen clarity, and abnormal event handling results; The regional status assessment index is updated based on real-time feedback data. The patrol route and patrol cycle are dynamically adjusted through a closed-loop control algorithm. The adjustment formula is as follows: ,in This is the updated state assessment index. For the updated patrol cycle, To adjust the sensitivity coefficient, R is the execution status feedback coefficient.

7. The video surveillance intelligent patrol and rotation scheduling method integrating AR tags according to claim 2, characterized in that: The visual AR tag is a composite tag that combines QR code and infrared marker. The near-field communication AR tag supports dual-mode communication of NFC and UWB. The positioning accuracy of UWB communication is ≤5 cm, and the effective distance of NFC communication is ≤10 cm.

8. A video surveillance intelligent patrol and rotation scheduling system integrating AR tags, characterized in that, include: The tag deployment and association unit is used to deploy AR tags and associate them with physical space and monitoring equipment information, establish an association mapping table and store it to the edge computing node; The data acquisition unit is used to collect video data, AR tag interaction data and environmental perception data of the monitored area, and to perform timestamp alignment and data cleaning. The status assessment unit is used to quantitatively assess the status of the monitored area based on the video data, AR tag interaction data and environmental perception data, through a multi-dimensional fusion assessment model, and obtain the area status assessment index. The scheduling generation unit is used to improve the scheduling based on the regional state evaluation index and preset scheduling rules. The algorithm generates patrol task plans and determines the monitoring round-robin strategy; The execution and feedback unit is used to execute the patrol task and rotation strategy through the patrol terminal and monitoring equipment, provide real-time feedback on the execution status, and dynamically adjust the scheduling scheme based on the closed-loop control algorithm.

9. A computer device, characterized in that, The system includes a memory and a processor, the memory storing a computer program, and the processor being configured to execute the computer program to implement the video surveillance intelligent patrol and rotation scheduling method with integrated AR tags as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the video surveillance intelligent patrol and rotation scheduling method with integrated AR tags as described in any one of claims 1 to 7.

Images