Regional personnel intelligent management method, related equipment and dynamic coding tag

Abstract

The invention discloses an intelligent management method for regional personnel, related equipment and a dynamic coding tag, which is realized on the basis of the dynamic coding tag configured for the regional personnel, a radio frequency positioning network and multi-modal visual identification equipment, and comprises the following steps: acquiring a rough position through radio frequency positioning and scheduling visual equipment; and identifying personnel attributes and accurate positions in combination with the dynamic coding labels, calculating overall and attribute sub-densities, comparing the overall and attribute sub-densities with a dynamic threshold value, judging a density abnormal situation and executing corresponding management operation. The dynamic coding label contains visible light and fluorescent hidden parts which are periodically updated, and can prevent counterfeiting and assist precise positioning and attribute quick matching; the multi-modal visual equipment reduces invalid acquisition through radio frequency scheduling, and improves the accuracy and efficiency in a complex environment in combination with tag feature recognition. According to the scheme, through double statistics of overall and attribute sub-density, misjudgment is avoided by matching with a dynamic threshold rule base, and quick handling of abnormities is realized by linkage of real-time data processing and a plan base, so that powerful support is provided for regional safety and efficient operation.

Description

Technical Field

[0001] This application relates to the field of intelligent management technology, and more specifically, to a regional intelligent management method, related equipment, and dynamic coding tags. Background Technology

[0002] With the acceleration of urbanization, densely populated areas such as industrial parks, commercial complexes, and transportation hubs are increasing, making precise and efficient management of personnel within these areas a core requirement for ensuring public safety and improving operational efficiency. Especially in scenarios involving public emergencies and the security of important events, quickly grasping the distribution, attributes, and density of personnel is crucial for scientific scheduling and risk warning. Therefore, developing efficient intelligent methods for regional personnel management is of paramount practical necessity.

[0003] In existing technologies, regional personnel management typically employs visual recognition technology, using surveillance cameras to capture images for personnel detection. However, this method is easily affected by factors such as lighting, occlusion, and complex backgrounds, resulting in significant fluctuations in recognition accuracy. Furthermore, it cannot efficiently correlate personnel attribute information, making categorized management difficult. Another approach combines manual statistics with traditional monitoring, relying on on-site inspections or review of surveillance data by management personnel. However, this method suffers from poor real-time performance and cannot perform large-scale population density and attribute classification statistics, making it difficult to adapt to dynamically changing management needs. These shortcomings result in deficiencies in personnel positioning accuracy, recognition security, statistical comprehensiveness, and management response timeliness. When faced with sudden increases in population flow or clusters of individuals with specific attributes, problems such as delayed warnings and improper handling can easily occur, failing to meet the high-precision and intelligent requirements of modern regional management.

[0004] Therefore, there is an urgent need for a new intelligent management method for regional personnel to overcome the shortcomings of existing technologies and achieve the intelligent management goals of accurate positioning, efficient identification, comprehensive statistics, and timely response to anomalies for regional personnel. Summary of the Invention

[0005] This application provides a regional intelligent personnel management method, related equipment, and dynamic coding tags. Through the collaboration of dynamic coding tags, radio frequency positioning networks, and multimodal visual recognition equipment, it enables precise personnel positioning, efficient attribute identification, rapid judgment of abnormal density situations, and automated handling, providing support for regional safe operation.

[0006] A regional intelligent personnel management method is based on dynamically coded tags configured for each person in the region, a radio frequency positioning network deployed in the region, and multimodal visual recognition equipment. The method includes:

[0007] The radio frequency positioning network is used to obtain the approximate location information of the personnel, and the corresponding visual recognition device is dispatched based on the approximate location information. The attribute category and location information of the personnel in the area are determined by recognizing the dynamic coded tag. The coded information of the dynamic coded tag includes a periodically dynamically updated visible light part and / or a fluorescent hidden part, and is associated with the personnel attribute category information in the background database.

[0008] Based on the identified attribute categories and location information of the personnel, the overall personnel density of the target area is calculated in real time, and at the same time, at least one attribute sub-density based on the personnel attribute classification is calculated. The attribute sub-density is the density obtained by statistically analyzing the personnel with the same attribute category in the target area.

[0009] The overall personnel density and the attribute sub-density calculated in real time are compared with a preset dynamic threshold rule library to determine whether there is an abnormal density situation. The dynamic threshold rule library includes attribute sub-density thresholds set based on different attribute category combinations and / or different spatiotemporal scenarios.

[0010] If an abnormal density situation is detected, corresponding management operation instructions will be generated and executed based on the anomaly type and the preset management contingency plan library.

[0011] Optionally, the radio frequency positioning network is used to obtain a rough area location information of a person, and a corresponding visual recognition device is dispatched based on the rough area location information. The attribute category and location information of the person within the area are determined by recognizing the dynamically coded tag, including:

[0012] The radio frequency positioning network periodically detects signals to obtain rough location information of each person in the area;

[0013] Based on the rough area location information, visual recognition devices deployed above the corresponding area are scheduled to acquire images of the target area;

[0014] During the image acquisition process, the images are analyzed to detect areas where people block each other or where dynamically coded tags overlap.

[0015] If present, at least two visual recognition devices with different spectral characteristics are scheduled to perform collaborative image acquisition and multimodal image fusion processing on the target area to resolve the occluded dynamic coded label.

[0016] From the acquired images and the fused images, locate and identify the current encoding information of the dynamic coded tag worn by each person;

[0017] The identified current encoding information is matched with the background database to obtain the corresponding personnel attribute category, and the personnel's location information is determined by combining image and radio frequency signal analysis.

[0018] Optionally, before scheduling the visual recognition device, the method further includes extracting and analyzing multi-dimensional radio frequency features of the radio frequency positioning network signal to optimize the visual recognition strategy.

[0019] The multi-dimensional radio frequency features include at least multipath delay difference features, signal strength gradient features, and signal dynamic change rate features. The multipath delay difference features are used to characterize the richness of signal propagation paths, the signal strength gradient features are used to characterize the spatial variation of the received signal strength at different positioning anchor points, and the signal dynamic change rate features are used to characterize the frequency and amplitude of signal strength changes over time.

[0020] The process of optimizing the visual recognition strategy includes:

[0021] Based on the multipath delay difference characteristics, it is determined whether there is a densely populated sub-region within the approximate location. If so, the visual recognition device is scheduled to prioritize the multimodal collaborative acquisition and fusion recognition of the region.

[0022] Based on the signal strength gradient characteristics, the relative distance distribution between different people in the approximate area is calculated to determine the focusing and field of view segmentation parameters when the visual recognition device acquires data.

[0023] Based on the dynamic change rate characteristics of the signal, the visual acquisition frame rate is dynamically adjusted, with a high frame rate for moving personnel and a low frame rate for stationary personnel.

[0024] Optionally, the attribute sub-density includes at least one of the function sub-density, behavior state sub-density, and spatiotemporal compliance sub-density;

[0025] The functional sub-density is obtained by classifying and statistically analyzing the job functions of personnel;

[0026] The behavioral state sub-density is obtained by classifying and statistically analyzing the human behavioral states identified through visual posture analysis or radio frequency signal features.

[0027] The spatiotemporal compliance sub-density is obtained by classifying and statistically analyzing the matching results of personnel's scheduling information, access history, and real-time location information.

[0028] Optionally, the real-time calculated overall population density and attribute sub-density are compared with a preset dynamic threshold rule base to determine whether there is an abnormal density trend, including:

[0029] The overall population density is compared with the maximum carrying density threshold calculated based on the regional three-dimensional terrain structure, passage capacity and real-time event status. When the overall population density exceeds the threshold and the duration exceeds the preset time, it is determined to be an abnormal regional cluster density.

[0030] Based on the functional sub-density and the spatiotemporal compliance sub-density, a dual compliance detection method, including job absence detection and unauthorized trajectory analysis, is used to determine whether there is an abnormal functional compliance density.

[0031] Monitor the instantaneous change rate and spatial distribution entropy value of the specific functional sub-density. When the instantaneous change rate generated by the concentration of specific functional personnel exceeds the dynamic risk threshold in a short period of time, and the spatial distribution entropy value changes abruptly from a uniform distribution to a highly clustered distribution, it is determined that the functional cluster density is abnormal, and cross-validation based on the behavioral state sub-density within the region is triggered.

[0032] The specific density of the behavior state is compared in real time with the state alarm threshold preset based on the scene safety rules. At the same time, the spatial propagation speed of the behavior state and the direction of personnel flow are monitored. When the state alarm threshold is exceeded and a trend of spreading to key areas is shown, it is determined that the state aggregation density is abnormal.

[0033] Optionally, the process of generating management operation instructions for the evacuation guidance triggered by the abnormal regional agglomeration density based on multi-dimensional attribute collaboration includes:

[0034] The movement status of people in the area is classified based on the identified behavioral state sub-density, and the people are grouped by function based on the identified functional sub-density.

[0035] By calculating the optimal path load for people to reach each safety exit within the area, and combining this with the abnormal dwelling areas identified by the spatiotemporal compliance sub-density, differentiated evacuation route planning is generated.

[0036] Based on the differentiated evacuation route plan, evacuation instructions with different priorities and different route guidance are sent to the smart terminals held by different functional groups.

[0037] Optionally, the execution process of the dual compliance detection includes:

[0038] The functional sub-density of each key area is compared with the expected functional density distribution map dynamically generated based on scheduling information and workflow dependencies. The job shortage detection result is determined by detecting whether the functional sub-density corresponding to the necessary functions is lower than the minimum configuration threshold and whether the duration exceeds the business process tolerance window.

[0039] The spatiotemporal compliance sub-density is matched with a predefined three-dimensional permission matrix. When an unauthorized target is detected whose real-time location information does not match the authorized spatiotemporal range, the historical trajectory and current activity pattern of the unauthorized target are back-analyzed to construct an abnormal behavior sequence in order to determine the unauthorized trajectory analysis results.

[0040] Optionally, the cross-validation process based on the density of the behavioral state sub-densities within the region includes:

[0041] When the functional aggregation density is determined to be abnormal, the behavioral state sub-density is obtained within the same time period and spatial range as the abnormal functional aggregation, and the time series change curve of the functional sub-density is aligned and the correlation is calculated with the concurrent change curve of the behavioral state sub-density.

[0042] If the calculation results show that the occurrence time of functional aggregation is highly correlated with the outbreak time of abnormal behavior and the spatial range highly overlaps, then the confidence level of the functional aggregation density anomaly is increased, and it is determined to be a high-probability emergency response event; otherwise, the confidence level of the functional aggregation density anomaly is decreased.

[0043] Optionally, when determining anomaly patterns in density, the following may also be included:

[0044] Real-time monitoring and comparison of the changing trends of attribute sub-densities of the same attribute category across different regions;

[0045] When a specific attribute sub-density in the first region is detected to decrease significantly within a certain period of time, and the same attribute sub-density in the second region adjacent to the first region increases significantly within a similar period of time, and the changes are correlated, it is determined that a cross-regional personnel transfer event has occurred, and the root cause is inferred by combining the other sub-density trends of the first region and the second region.

[0046] Optionally, after determining the attribute category and location information of the person within the area by identifying the dynamic coded tag, the method further includes:

[0047] Obtain a causal knowledge graph containing prior business rules for the scenario;

[0048] The attribute categories and location information of the personnel within the area determined by identifying the dynamic coding tags, along with the causal knowledge graph, are input into the causal inference engine for logical consistency verification.

[0049] If the logical consistency check finds a conflict, an identity verification process is triggered, which includes:

[0050] Re-identification and verification of dynamic coded tags of relevant personnel based on multimodal image fusion;

[0051] If the verification is correct, based on the personnel's historical behavior data and shift information, combined with the causal knowledge graph, the conflicting attribute categories are probabilistically corrected until the attribute category matching result that conforms to the prior business logic of the scenario is output.

[0052] A regional personnel intelligent management device includes a memory and a processor;

[0053] The memory is used to store programs;

[0054] The processor is used to execute the program to implement the various steps of the regional personnel intelligent management method as described in any of the above claims.

[0055] A readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the regional personnel intelligent management method as described in any of the preceding claims.

[0056] A computer program product includes a computer program that, when run by a processor, executes the steps of the regional personnel intelligent management method as described in any of the preceding claims.

[0057] A dynamically coded tag for use in the regional personnel intelligent management method as described in any of the preceding claims, comprising:

[0058] Label base;

[0059] A visible light coding area is disposed on the label substrate, the visible light coding area containing a combination of numbers and letters that are periodically and dynamically updated;

[0060] A fluorescent material layer disposed on or within the label substrate, the fluorescent material layer exhibiting a hidden coded pattern under excitation light of a specific wavelength;

[0061] A microcontroller is used to control the encoding of the visible light encoding area to be synchronously updated according to the personnel attribute category information in the background database.

[0062] As can be seen from the above technical solutions, the regional personnel intelligent management method, related equipment and dynamic coding tags provided in this application embodiment are based on dynamic coding tags configured for regional personnel, radio frequency positioning network and multimodal visual recognition equipment. The method obtains the rough position through radio frequency positioning and schedules the visual equipment. Combines the dynamic coding tags to identify personnel attributes and precise positions. After calculating the overall and attribute sub-densities, it compares them with dynamic thresholds to determine the density abnormality and execute corresponding management operations.

[0063] This application achieves multiple beneficial effects by addressing the shortcomings of existing technologies:

[0064] Firstly, it solves the problems of coarse positioning and insecure coding in existing radio frequency technology. The dynamic coding tag contains periodically updated visible light and fluorescent hidden parts, which not only avoids the risk of forgery of fixed codes, but also improves the location recognition from coarse area to precise positioning by coordinating radio frequency and vision. At the same time, the rapid matching of personnel attributes is achieved by associating the tag with the backend database.

[0065] Secondly, it makes up for the shortcomings of poor environmental adaptability and weak attribute association of single visual recognition. Multimodal visual recognition equipment reduces invalid image acquisition based on precise scheduling of radio frequency positioning. Combined with feature recognition of dynamic coded tags, it greatly improves the recognition accuracy in complex environments. Moreover, it can directly obtain attribute information through tags without relying on complex image feature extraction, thus improving the efficiency of attribute recognition.

[0066] Third, it overcomes the problems of one-sided density statistics and rigid thresholds in existing technologies. This application calculates the overall personnel density and sub-densities based on attribute classification, and combines them with a dynamic threshold rule library containing different attribute combinations and spatiotemporal scenarios, so that the density anomaly judgment is more in line with actual management needs and avoids misjudgment and omission under fixed thresholds in different scenarios.

[0067] Fourth, through real-time data collection, analysis, and linkage with the contingency plan database, it has achieved rapid identification and automated handling of abnormal situations, greatly improving management response speed and providing reliable support for regional security and efficient operation. Attached Figure Description

[0068] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0069] Figure 1 This is a flowchart of a regional personnel intelligent management method disclosed in an embodiment of this application;

[0070] Figure 2 This is a flowchart illustrating the architecture used in a regional personnel intelligent management method disclosed in an embodiment of this application.

[0071] Figure 3 This is a hardware structure block diagram of a regional personnel intelligent management device disclosed in an embodiment of this application. Detailed Implementation

[0072] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0073] The following section introduces the solution proposed in this application. The technical solution is as follows, and details are provided below.

[0074] Figure 1 This is a flowchart of a regional personnel intelligent management method disclosed in an embodiment of this application.

[0075] Figure 2 This is a flowchart illustrating the architecture of a regional personnel intelligent management method disclosed in an embodiment of this application.

[0076] The regional intelligent personnel management method is based on dynamic coded tags configured for each person in the region, radio frequency positioning networks deployed in the region, and multimodal visual recognition devices.

[0077] The dynamic encoding tag consists of the following components:

[0078] Label base;

[0079] A visible light coding area is disposed on the label substrate, the visible light coding area containing a combination of numbers and letters that are periodically and dynamically updated;

[0080] A fluorescent material layer disposed on or within the label substrate, the fluorescent material layer exhibiting a hidden coded pattern under excitation light of a specific wavelength;

[0081] A microcontroller is used to control the encoding of the visible light encoding area to be synchronously updated according to the personnel attribute category information in the background database.

[0082] Specifically, the regional personnel intelligent management method disclosed in this application relies on three key components: a dynamic coding tag uniformly configured for each person in the region, a radio frequency positioning network deployed in a grid-like manner across the entire region, and multi-modal visual recognition devices deployed in multiple dimensions. The three components work together through data interaction and command linkage to support the entire process management of personnel attribute identification, location positioning, density monitoring, and anomaly handling. Among them, the dynamic coding tag serves as the core carrier of personnel identity and attribute information, and its structural design and functional implementation directly determine the accuracy and security of the entire management method. The specific composition and working logic are as follows.

[0083] The dynamic coding tag uses a tag substrate as its basic support structure. This substrate is made of a wear-resistant and interference-resistant flexible material, adaptable to the physical requirements of different wearing scenarios, providing a stable installation and protective foundation for the tag's various functional areas and components. On the surface of the tag substrate, a visible light coding area is integrated. This area uses a high-contrast combination of numbers and letters as its basic coding format. Its core feature is the ability to periodically and dynamically update the coded information. It can automatically switch the coded content according to a preset time period or background command. This allows for rapid capture in conventional visual recognition scenarios while mitigating the security risks of fixed codes being easily counterfeited through dynamic coding changes. A fluorescent material layer is also embedded on or inside the tag substrate. This material layer has no obvious visual characteristics under natural light, but only reveals a preset hidden coding pattern under excitation light of a specific wavelength. This hidden code forms a complementary verification relationship with the content of the visible light coding area, providing secondary information support for personnel identity verification in special scenarios such as when the visible light code is obscured or tampered with.

[0084] To achieve precise control and synchronization of coded information, the dynamic coded tag incorporates a built-in microcontroller. This controller, as the tag's core control unit, maintains real-time data communication with the backend database, automatically receiving personnel attribute category information update commands from the backend. Simultaneously, it drives the visible light coding area to periodically update its coded content based on these commands, ensuring consistency between the hidden coding pattern in the fluorescent material layer and the update logic of the visible light coding area. This achieves precise association and binding between the coded information and the personnel attribute category information in the backend database. When personnel enter the managed area, the dynamic coded tag can simultaneously have its signal range identified by the radio frequency positioning network and its coded features captured by multimodal visual recognition devices, providing a reliable information source for subsequent location positioning and attribute identification.

[0085] In the actual management process, the dynamic coded tag is first captured by the radio frequency positioning network to obtain the personnel's rough location information. The back-end system will then dispatch the corresponding multimodal visual recognition device to start the data collection based on the rough location. The visual recognition device will simultaneously identify the real-time code of the visible light coding area and the hidden code of the fluorescent material layer, and match the identified coding information with the back-end database to accurately associate it with the personnel's attribute category information. Combined with image analysis and radio frequency signal calibration, the precise location of the personnel is finally determined, laying the data foundation for subsequent density calculation and anomaly judgment.

[0086] like Figure 1 and Figure 2 As shown, the regional personnel intelligent management method described in the application relies on the coordinated linkage of dynamic coded tags, radio frequency positioning networks, and multimodal visual recognition devices to achieve accurate perception and intelligent control of regional personnel. This method may include:

[0087] Step S1: Obtain the approximate location information of the personnel using the radio frequency positioning network, and schedule the corresponding visual recognition device based on the approximate location information. Determine the attribute category and location information of the personnel in the area by recognizing the dynamic coded tag. The coded information of the dynamic coded tag includes a periodically dynamically updated visible light portion and / or a fluorescent hidden portion, and is associated with the personnel attribute category information in the background database.

[0088] Specifically, the radio frequency (RF) positioning network periodically transmits detection signals into the managed area. By receiving RF signals from dynamically coded tags worn by personnel and combining this with multi-anchor signal convergence analysis, the system can initially pinpoint a rough location of the personnel. Simultaneously, while acquiring this rough location information, the system also extracts multi-dimensional features of the RF signals, including multipath delay difference characteristics representing the richness of signal propagation paths, signal strength gradient characteristics reflecting spatial variations in signal strength at different anchor points, and signal dynamic change rate characteristics reflecting the temporal variation of signal strength. These features are used to optimize the visual recognition strategy.

[0089] Specifically, the system will schedule multimodal visual recognition devices deployed in the corresponding areas to start image acquisition based on the approximate location of people. For densely populated sub-regions determined by multipath delay difference characteristics, vision devices with multispectral characteristics will be prioritized for collaborative acquisition. Based on the relative distance distribution of people calculated from signal intensity gradient characteristics, the focusing parameters and field of view segmentation range of the vision devices will be set. The acquisition frame rate will be adjusted according to the dynamic change rate characteristics of the signal, using a high frame rate for moving people and a low frame rate for stationary people. During the image acquisition phase, if personnel occlusion or tag overlap is detected, multimodal image fusion processing will be initiated to parse the occluded dynamic coded tags. Subsequently, the real-time alphanumeric combination of the visible light coded area will be identified from the acquired image. At the same time, a fluorescent hidden coded pattern will be awakened and identified by excitation light of a specific wavelength. The two types of coded information will be matched with the background database to obtain the attribute category of the person. Combined with image spatial positioning and radio frequency signal calibration, the precise location of the person will be determined. In addition, the system will input personnel attributes and location information into a causal knowledge graph that carries scenario-based prior business rules. The causal inference engine will then perform a logical consistency check. If an information conflict is found, the system will trigger an identity correction process, which will re-identify the user through tags and review historical data to output a matching result that conforms to the business logic.

[0090] It is worth considering that, after determining the attribute category and location information of the personnel within the area by identifying the dynamic coded tags, in order to further ensure the accuracy of personnel identity and attribute information and to align with the scenario's business logic, this application also adds a logical verification and identity correction process based on causal knowledge graphs, as follows:

[0091] ① Obtain a causal knowledge graph containing prior business rules for the scenario;

[0092] ② The attribute categories and location information of the personnel within the area determined by identifying the dynamic coding tags, as well as the causal knowledge graph, are input into the causal inference engine for logical consistency verification;

[0093] ③ If the logical consistency check finds a conflict, the identity verification process is triggered, which includes:

[0094] ④ Based on multimodal image fusion, the dynamic coded tags of relevant personnel are re-identified and verified;

[0095] ⑤ If the verification is correct, based on the personnel's historical behavior data and shift information, and combined with the causal knowledge graph, the conflicting attribute categories are probabilistically corrected until the attribute category matching result that conforms to the prior business logic of the scenario is output.

[0096] Specifically, the first step is to acquire a causal knowledge graph containing prior business rules for specific scenarios. This causal knowledge graph is constructed based on prior information such as historical business data from regional management, pre-defined job specifications, personnel access permission rules, and emergency response procedures. Its nodes cover core elements such as personnel attribute categories, job functions, authorized access areas, shift schedules, and business process dependencies. The edges between nodes represent the causal logic of each element. For example, business rules such as "personnel with specific functions must be stationed in designated areas during their shifts" and "visitors can only enter non-core areas during authorized times" are embedded in the knowledge graph in the form of causal relationships, providing a unified business judgment benchmark for subsequent logic verification.

[0097] Subsequently, the attribute categories and location information of the personnel within the area determined by the dynamically encoded tags, along with the causal knowledge graph, are input into the causal inference engine for logical consistency verification. The causal inference engine first structurally decomposes the real-time attributes and location information of the personnel, then matches them with the corresponding nodes in the knowledge graph. By traversing the causal relationship edges between nodes, it verifies whether the real-time information conforms to the scenario's prior business rules. For example, if the knowledge graph specifies that "security personnel must be stationed in the entrance and exit area from 8:00 AM to 6:00 PM on weekdays," and a security personnel's attribute category is identified as security, but their real-time location is in an unauthorized core office area and the time period is during working hours, the engine will determine that there is a logical conflict between the personnel's attribute and location information. If the personnel's attribute is visitor, but their real-time location is in an unauthorized equipment room, a logical conflict warning will also be triggered. Otherwise, the information is determined to conform to the business logic, and no further correction process needs to be initiated.

[0098] If a conflict is detected during logical consistency verification, the identity correction process is automatically triggered. Based on multimodal image fusion, the dynamic coded tags of relevant personnel are re-identified and verified. The system prioritizes visual recognition devices with multispectral acquisition capabilities to perform secondary image acquisition on the areas where conflicting personnel are located. Addressing potential issues such as tag occlusion and angular deviations, the system enhances the feature extraction effect of dynamic coded tags through the fusion processing of visible light, infrared, and fluorescence multimodal images. Subsequently, the visible light coded area and the fluorescence hidden coded area of ​​the tag are re-identified. After obtaining the coded information, it is again accurately matched with the backend database to confirm whether there are any recognition deviations in the personnel identity and attribute information corresponding to the tag code. If the verification finds that the tag recognition is incorrect, the personnel attribute information is directly corrected, and the verification loop is completed.

[0099] If the verification is correct, meaning there is no deviation between the tag code and the personnel attribute, the system will use the personnel's historical behavior data and shift information, combined with the causal knowledge graph, to probabilistically correct conflicting attribute categories until an attribute category matching result that conforms to the scenario's prior business logic is output. The system will retrieve the personnel's past job performance trajectory, shift attendance records, frequency of permission usage, and other historical behavior data over a period of time. Combining this with the business rules in the causal knowledge graph, a probabilistic correction model will be constructed. For example, if the personnel's historical data shows that they have a high frequency of behavior such as "temporarily dispatched to support the core area," and there is a corresponding dispatch approval record in the current time period, then the system will determine that their current attribute category is "temporary support position" rather than the originally identified regular position with a high probability, thereby correcting the attribute category. If the historical data shows that they have no record of unauthorized access to areas and no temporary dispatch instructions, the system will combine the knowledge graph rules to check whether the permission information has not been updated synchronously, and complete the adaptation and correction of permissions and attributes. After multiple rounds of probabilistic iteration and logical verification, the final attribute category matching result that conforms to the scenario's prior business logic will be output, ensuring the accuracy of personnel information and the consistency of business logic.

[0100] Step S2: Based on the identified attribute categories and location information of the personnel, calculate the overall personnel density of the target area in real time, and simultaneously calculate at least one attribute sub-density based on the personnel attribute classification. The attribute sub-density is the density obtained by statistically analyzing the personnel with the same attribute category in the target area.

[0101] Specifically, the system uses the preset geographical boundaries of the target area as the statistical range. By aggregating the location information of all personnel within the area and combining it with the area's spatial area, it calculates the overall personnel density in real time, reflecting the overall degree of personnel aggregation within the area. For the calculation of attribute sub-densities, the system conducts targeted statistics according to preset attribute classification dimensions, mainly covering three core sub-densities: first, functional sub-density, which classifies personnel based on their job function tags and calculates the distribution density of the same functional group within the target area; second, behavioral state sub-density, which combines visual posture analysis and dynamic characteristics of radio frequency signals to identify and classify personnel's behavioral states, calculating the density of specific behavioral state groups; and third, spatiotemporal compliance sub-density, which matches personnel's real-time locations with preset scheduling information and access permission ranges to calculate the distribution density of compliant and non-compliant personnel. These three types of sub-densities can be calculated individually or in combination, providing multi-dimensional data support for subsequent anomaly detection.

[0102] Step S3: Compare the real-time calculated overall personnel density and attribute sub-density with a preset dynamic threshold rule library to determine whether there is an abnormal density situation. The dynamic threshold rule library includes attribute sub-density thresholds set based on different attribute category combinations and / or different spatiotemporal scenarios.

[0103] Specifically, the dynamic threshold rule library pre-stores density judgment standards adapted to different scenarios, and performs data comparison and anomaly identification across dimensions. For overall personnel density, it is compared with the maximum carrying capacity threshold generated by combining regional terrain structure, passageway capacity, and real-time event status. If the overall density exceeds the threshold and the duration reaches a preset window, it is determined to be an abnormal regional cluster density. For functional sub-density and spatiotemporal compliance sub-density, dual compliance checks are performed using job absence detection and unauthorized trajectory analysis to match the actual functional density with the expected functional density distribution map, while verifying the consistency of personnel location with authorized spatiotemporal range, thereby determining whether there is an abnormal functional compliance density. For specific functional sub-density, its instantaneous change rate and spatial distribution entropy value are monitored. If there is a sudden increase in the instantaneous change rate and the distribution state changes from uniform to highly clustered, cross-validation of behavioral state sub-density is triggered to confirm whether it is an abnormal functional cluster density. For behavioral state sub-density, it is compared with the scene safety alarm threshold, while monitoring the state propagation speed and personnel flow direction. If it exceeds the threshold and spreads to critical areas, it is determined to be an abnormal state cluster density. In addition, the system will monitor the linkage changes of attribute sub-densities across regions. If the same attribute sub-density in adjacent regions shows reverse correlation fluctuations, it will be determined as a cross-regional personnel transfer event and root cause inference will be carried out.

[0104] Furthermore, considering that the movement of people within a managed area often exhibits cross-regional linkage characteristics, density monitoring in a single area may overlook the potential risk of localized relocation leading to aggregation in another area. Relying solely on density data from a single area to identify anomalies may result in delayed or misjudgments in early warning. Therefore, when assessing abnormal density trends, it is necessary to add a cross-regional personnel transfer monitoring component. This involves multi-regional data linkage analysis to achieve a more comprehensive situational awareness, specifically including:

[0105] ① Monitor and compare the changing trends of attribute sub-densities of the same attribute category in real time across different regions;

[0106] ② When a specific attribute sub-density in the first region is detected to decrease significantly within a certain period of time, and the same attribute sub-density in the second region adjacent to the first region increases significantly within a similar period of time, and the changes are correlated, it is determined that a cross-regional personnel transfer event has occurred, and the root cause is inferred by combining the other sub-density trends of the first region and the second region.

[0107] Specifically, using preset time slices as units, the system synchronously collects all attribute sub-density data for each sub-region within the management scope. For each attribute type (such as "security personnel," "visitors," and "technical staff"), a time series change curve for the attribute sub-density of each sub-region is plotted. To ensure the effectiveness of the comparison, the system standardizes the curves to eliminate density value deviations caused by differences in area between different regions, focusing on two core indicators: the magnitude of change and the rate of change. The former is measured by the ratio of the density difference between the current time period and the previous time period, while the latter is represented by the number of density changes per unit time. By overlaying and comparing the curves, the system intuitively presents the flow and correlation characteristics of the same attribute in different regions. At the same time, the system prioritizes marking areas where the attribute sub-density fluctuation exceeds a preset benchmark value, including them in the key monitoring scope to narrow the data dimensionality for subsequent correlation analysis.

[0108] The determination of a significant decrease or increase requires meeting two conditions: first, the magnitude of the density change exceeds the preset cross-regional flow trigger threshold in the dynamic threshold rule base (this threshold is dynamically adjusted based on the regional functional attributes; for example, the trigger threshold for office areas and rest areas is higher than that for two independent office areas); second, the duration of the change reaches at least two consecutive time slices, excluding instantaneous fluctuations. The correlation of the change is achieved by calculating the matching degree between the density decrease in the first region and the density increase in the second region. If the difference between the two is within the preset error range, and there is a temporal correlation that "the decrease in the first region precedes the increase in the second region by 0-1 time slices," then it can be clearly determined that the personnel with this attribute have moved from the first region to the second region.

[0109] After identifying a transfer event, root cause inference combines multi-dimensional sub-density data and scenario information from both regions: If the "employee sub-density during off-duty hours" in the first region decreases while the "employee sub-density in the cafeteria" in the second region increases, and there are no fluctuations in the sub-density of abnormal behavior states in either region, it is inferred to be a routine dining flow; if the "technical staff sub-density in the equipment room" in the first region drops sharply, while the "technical staff density at the emergency assembly point" in the adjacent second region rises sharply, and both regions show an increase in the sub-density of "emergency behavior states," then, based on the "equipment failure emergency response rules" in the causal knowledge graph, it is inferred to be an emergency transfer of technical personnel caused by equipment failure. Through this linked analysis, the system can distinguish between routine and emergency transfer events, providing a precise basis for whether to initiate abnormal handling procedures, avoiding over-responding to routine flows, and ensuring that emergency transfers receive timely attention.

[0110] Step S4: If an abnormal density situation is determined, generate and execute the corresponding management operation instructions based on the anomaly type and the preset management plan library.

[0111] Specifically, the management contingency plan library pre-stores handling strategies corresponding to various density anomalies. The system generates differentiated management operation instructions based on the type, level, and scope of impact of the anomaly. For area-wide cluster density anomalies, the system groups personnel based on their behavioral status and functional classification, and, combined with the load of each safety exit path and the distribution of abnormal dwelling areas, formulates differentiated evacuation route plans, sending priority evacuation guidance instructions to the smart terminals of different functional groups. For functional compliance density anomalies, if it is due to a missing post, a post replacement warning is pushed to the management terminal, and information on available personnel in the surrounding area is retrieved simultaneously; if it is an unauthorized trajectory, the trajectory monitoring of the unauthorized target is triggered, and anomaly location and handling prompts are sent to the security terminal. For functional or behavioral cluster density anomalies, if cross-verification confirms it as a high-probability emergency event, the system activates emergency broadcast equipment to issue a warning message, and simultaneously dispatches multimodal vision equipment to enhance real-time monitoring of the target area; for cross-regional personnel transfer events, the system sends a traffic warning to the management terminal of the transfer target area, and initiates dynamic allocation of regional carrying capacity in advance. After all instructions are issued, the system will continuously track the execution status and dynamically adjust the instructions based on real-time changes in personnel location and attributes to ensure the effectiveness of the response.

[0112] As can be seen from the above technical solutions, the regional personnel intelligent management method, related equipment and dynamic coding tags provided in this application embodiment are based on dynamic coding tags configured for regional personnel, radio frequency positioning network and multimodal visual recognition equipment. The method obtains the rough position through radio frequency positioning and schedules the visual equipment. Combines the dynamic coding tags to identify personnel attributes and precise positions. After calculating the overall and attribute sub-densities, it compares them with dynamic thresholds to determine the density abnormality and execute corresponding management operations.

[0113] This application achieves multiple beneficial effects by addressing the shortcomings of existing technologies:

[0114] Firstly, it solves the problems of coarse positioning and insecure coding in existing radio frequency technology. The dynamic coding tag contains periodically updated visible light and fluorescent hidden parts, which not only avoids the risk of forgery of fixed codes, but also improves the location recognition from coarse area to precise positioning by coordinating radio frequency and vision. At the same time, the rapid matching of personnel attributes is achieved by associating the tag with the backend database.

[0115] Secondly, it makes up for the shortcomings of poor environmental adaptability and weak attribute association of single visual recognition. Multimodal visual recognition equipment reduces invalid image acquisition based on precise scheduling of radio frequency positioning. Combined with feature recognition of dynamic coded tags, it greatly improves the recognition accuracy in complex environments. Moreover, it can directly obtain attribute information through tags without relying on complex image feature extraction, thus improving the efficiency of attribute recognition.

[0116] Third, it overcomes the problems of one-sided density statistics and rigid thresholds in existing technologies. This application calculates the overall personnel density and sub-densities based on attribute classification, and combines them with a dynamic threshold rule library containing different attribute combinations and spatiotemporal scenarios, so that the density anomaly judgment is more in line with actual management needs and avoids misjudgment and omission under fixed thresholds in different scenarios.

[0117] Fourth, through real-time data collection, analysis, and linkage with the contingency plan database, it has achieved rapid identification and automated handling of abnormal situations, greatly improving management response speed and providing reliable support for regional security and efficient operation.

[0118] In some embodiments of this application, the process of step S1, which involves obtaining a rough area location information of a person using the radio frequency positioning network, scheduling a corresponding visual recognition device based on the rough area location information, and determining the attribute category and location information of the person within the area by recognizing the dynamic coded tag, is described. Specifically, it may include:

[0119] ① The radio frequency (RF) positioning network periodically detects signals to obtain rough location information of personnel within the area. The RF positioning network adopts a full-area grid deployment mode, deploying multiple RF positioning anchor points at preset intervals within the management area. Each anchor point synchronously and periodically emits a detection signal at a specific frequency, and the detection period can be dynamically adjusted according to the intensity of personnel movement in the area. When a person's dynamically coded tag enters the signal coverage area, it passively reflects the detection signal. After receiving the reflected signal, each anchor point extracts key parameters such as the signal arrival time difference and the received signal strength indication, and calculates the spatial coordinate range of the tag using a multi-anchor point signal convergence algorithm. Considering that RF signals are easily affected by obstacles such as walls and metal equipment, the system performs error compensation on the calculation results, ultimately outputting rough area location information represented by the combination of positioning anchor points plus the area grid number, such as "anchor point A3-A4-grid B2," ensuring that the location information not only covers the area where the personnel are located but also provides clear direction for subsequent visual equipment scheduling.

[0120] ② Based on the approximate location information, the system schedules visual recognition devices deployed above the corresponding area to acquire images of the target area. The system has a built-in mapping table between areas and devices, recording the installation location, monitoring coverage area, and operating status of each visual recognition device. Once the approximate location information of a person is obtained, the background scheduling module quickly matches the visual devices within the corresponding coverage area and issues acquisition commands containing parameters such as acquisition frame rate, focus range, and image resolution. For example, if the approximate location of a person is "grid B2," the system immediately schedules three panoramic visual cameras installed on the ceiling of area B2 to start working, locking the focus range to the ground area of ​​that grid, while simultaneously shutting down visual devices in areas not yet covered to reduce system energy consumption. If the primary visual device in the corresponding area is faulty, the system automatically switches to the backup device and pushes a device fault alarm to the management terminal to ensure uninterrupted acquisition.

[0121] ③ During image acquisition, the images are analyzed to detect areas of occlusion between people or overlapping dynamically coded tags. After the visual recognition device acquires images, it transmits them in real time to the image preprocessing module. The module uses target detection algorithms to extract and locate the contours of people in the image, marking the approximate areas of each person's head, torso, and dynamically coded tags. The system then calculates the percentage of overlap between each person's area. For example, if the overlap between the contour area of ​​one person and another person exceeds 30% of the former's area, occlusion between people is determined. Simultaneously, for tag areas, feature point matching technology is used to identify the rectangular boundaries of the tags. For example, if the intersection area of ​​two tags exceeds 20% of the area of ​​a single tag, tag overlap is determined. To improve detection accuracy, the system considers the changing trends of multiple consecutive frames to eliminate false occlusion detections caused by momentary overlap between people. Only when the occlusion or overlap persists for more than an acceptable number of frames is a problematic area confirmed to exist.

[0122] ④ If present, at least two visual recognition devices with different spectral characteristics are scheduled to perform collaborative image acquisition and multimodal image fusion processing on the target area to resolve the occluded dynamic coded tag. The system has a preset multispectral device combination scheme, commonly used combinations such as visible light camera and near-infrared camera, or visible light camera and ultraviolet excitation camera. Near-infrared cameras can penetrate some clothing materials, while ultraviolet excitation cameras can awaken the fluorescent hidden part in the dynamic coded tag. When a problem area is detected, the scheduling module immediately sends a collaborative acquisition command to the backup multispectral device in the corresponding area, ensuring that all devices start acquisition at the same timestamp to avoid image misalignment due to time differences. After acquisition, the multimodal image fusion module uses a pixel-level fusion algorithm to fuse the feature information of different spectral images. For example, the outline features of a person in the visible light image are superimposed with the outline features of the tag in the near-infrared image, and the fluorescent coding features in the ultraviolet image are complementary with the visible light coding features in the visible light image. Through feature enhancement, visual interference in the occluded area is eliminated, and the coding area of ​​the occluded tag is fully restored.

[0123] ⑤ From the acquired images and the fused images, the system locates and identifies the current encoding information of the dynamic coded tags worn by each person. First, the system uses image segmentation technology to accurately extract the independent region of each dynamic coded tag from the fused image, removing interfering elements such as clothing and background environment. For the visible light encoding area of ​​the tag, an optical character recognition algorithm is used to identify the combination of numbers and letters. Combined with the periodic update rules of the encoding, erroneous recognition results caused by image blurring are filtered out. For the fluorescent hidden encoding area, image thresholding technology is used to extract the contour features of the fluorescent pattern, which is then matched with a preset encoding pattern library to determine the hidden encoding content. To ensure recognition accuracy, the system cross-validates the recognition results of the two encoding methods. If the associated encoding sequences of the two are consistent, the final encoding information is output; if there are differences, a second acquisition and recognition process is initiated until a consistent result is obtained.

[0124] ⑥ The identified current encoded information is matched with the background database to obtain the corresponding personnel attribute categories, and the personnel's location information is determined by combining image and radio frequency signal analysis. The background database pre-stores a one-to-one correspondence between dynamic encoded tags and personnel attribute category data. The system uses the identified encoded information as search keywords to quickly match the corresponding records in the database and generate a data package containing the complete attribute categories of the personnel. In terms of location information determination, a fusion calibration strategy of visual positioning as the main method and radio frequency positioning as the auxiliary method is adopted: the precise coordinates of the personnel are obtained by converting the pixel distance between the tags in the visual image and the preset reference points (such as ground marking lines and equipment bases); at the same time, the coarse location information output by the radio frequency positioning network is combined with the Kalman filter algorithm to eliminate the error of the two positioning methods, and finally output the personnel location coordinates with higher accuracy. This precise location information and attribute category information will be synchronously stored in the real-time database to provide data support for the density calculation in step S2 and the monitoring of attribute sub-density change trends in subsequent cross-regional transfer determination. For example, the specific attribute sub-density required for subsequent cross-regional transfer determination comes from the personnel attributes and location information associated in this step.

[0125] It is worth noting that the precise attribute and location data obtained in step S1 forms the basis for subsequent analysis and root cause inference in determining significant decreases or increases in cross-regional personnel transfers. For example, when a sharp drop in the "density of technical positions in the equipment room" is detected, data statistics need to be completed based on the "technical position" attribute category and "equipment room" location information identified in this step. Similarly, when inferring "equipment failure emergency response," the personnel attributes associated in this step need to be matched with the business rules of "equipment failure handling corresponding to technical positions" in the causal knowledge graph to ensure a closed-loop data flow throughout the entire management process.

[0126] Based on this, before scheduling the visual recognition device, the method further includes extracting and analyzing the multi-dimensional radio frequency features of the radio frequency positioning network signal to optimize the visual recognition strategy.

[0127] The multi-dimensional radio frequency features include at least multipath delay difference features, signal strength gradient features, and signal dynamic change rate features. The multipath delay difference features are used to characterize the richness of signal propagation paths, the signal strength gradient features are used to characterize the spatial variation of the received signal strength at different positioning anchor points, and the signal dynamic change rate features are used to characterize the frequency and amplitude of signal strength changes over time.

[0128] The process of optimizing the visual recognition strategy includes:

[0129] ① Based on the multipath delay difference characteristics, determine whether there is a densely populated sub-region within the approximate area location. If so, schedule the visual recognition device to prioritize multimodal collaborative acquisition and fusion recognition of the region.

[0130] ② Based on the signal strength gradient characteristics, the relative distance distribution between different people in the approximate area is calculated to determine the focusing and field-of-view segmentation parameters when the visual recognition device collects data;

[0131] ③ Based on the dynamic change rate characteristics of the signal, the visual acquisition frame rate is dynamically adjusted, with a high frame rate for moving personnel and a low frame rate for stationary personnel.

[0132] Specifically, the multipath delay difference characteristic is reflected by the time difference in receiving the same dynamically coded tag signal from different positioning anchor points. The richness of the signal propagation path is directly related to the number of people and the distribution of obstacles in the area. When this characteristic shows significant fluctuations, it indicates that the signal reaches the anchor point through multiple reflection paths during propagation, indirectly reflecting a dense population distribution in the area, which is prone to occlusion or tag overlap problems. This application interprets the multipath delay difference characteristic through a preset feature analysis model. If it is determined that there is a densely populated sub-area in the approximate area, the system will adjust the device scheduling order, prioritizing the use of visual recognition devices with multispectral acquisition capabilities in that area and its surroundings. These devices will be started synchronously according to the collaborative acquisition protocol, capturing images of the target area from different spectral dimensions such as visible light and specific wavelength excitation light. Subsequently, through multimodal image fusion technology, the feature information of each spectral image is complementary and superimposed, effectively penetrating the occlusion areas between people, completely resolving the covered dynamically coded tags, and avoiding recognition blind spots caused by single spectral acquisition.

[0133] Signal strength gradient features can intuitively reflect the spatial variation of signal strength received at different positioning anchor points. When people are densely distributed in an area, the signal strength difference between adjacent anchor points is small, and the gradient value is low; when people are sparsely distributed, the signal strength changes more significantly with spatial location, and the gradient value is high. This application utilizes the correlation between this feature and the relative distance between people to calculate the approximate relative position distribution of people within an area through a feature mapping model, thus identifying which areas have closer people and which areas have relatively sparse people. For densely populated sub-areas, the system configures the visual recognition device with small field of view and high focusing accuracy acquisition parameters to ensure clear and distinguishable tag images; for sparsely populated areas, a large field of view and dynamic focusing parameters are used to expand the coverage of a single device. Simultaneously, based on the relative distance distribution, the target area is divided into multiple independent acquisition sub-fields of view, each allocated dedicated image processing resources to improve the efficiency of tag positioning and recognition.

[0134] The dynamic rate of change of the signal is quantified by the frequency and amplitude of the change in radio frequency signal strength within a continuous period, directly related to the movement state of a person. When a person is moving, the relative position of the dynamic coded tag and the positioning anchor point changes continuously, the signal strength fluctuates frequently and with large amplitude, and the feature value is high; when the person is stationary, the signal strength tends to stabilize, and the feature value is low. This application establishes a linkage adjustment mechanism between the acquisition frame rate and the feature value by monitoring the change trend of this feature in real time: when a high dynamic rate of change feature value is detected, indicating that the person is moving, the acquisition frame rate of the vision device in the corresponding area is immediately increased to ensure complete capture of the coded information of the tag during movement, avoiding encoding ambiguity or missed recognition due to insufficient frame rate; when the feature value is low, indicating that the person is stationary, the acquisition frame rate is reduced, maintaining only periodic image acquisition to confirm the tag status. This dynamic adjustment method ensures the integrity of tag recognition for moving persons and avoids invalid high-frequency acquisition of stationary persons, significantly reducing the computing power consumption and data transmission pressure of the vision device.

[0135] In some embodiments of this application, such as Figure 2 As shown, the attribute sub-density includes at least one of the functional sub-density, behavioral state sub-density, and spatiotemporal compliance sub-density;

[0136] The functional sub-density is obtained by classifying and statistically analyzing personnel's job functions. Personnel attribute information identified and associated through dynamic coding tags includes explicit job function tags such as "security," "technical maintenance," "visitors," and "management." These tags are synchronized in real-time with personnel files in the backend database. During statistical analysis, the system extracts the functional tags of all personnel within the geographical boundary of the target area, groups them, summarizes the number of personnel in each group, and calculates the distribution density of personnel with different functions by combining this with regional spatial parameters. For example, in the production workshop area, the statistical results of the "operator" functional sub-density directly reflect the personnel configuration of core positions in that area, providing basic data for subsequent functional compliance assessments.

[0137] The behavioral state sub-density is obtained by classifying and statistically analyzing personnel behavioral states identified through visual posture analysis or radio frequency signal features. Behavioral state recognition relies on an optimized visual recognition strategy and radio frequency feature analysis: visual devices analyze personnel limb postures and movement trajectories through image analysis to distinguish states such as "stationary," "slowly moving," "gathering," and "emergency running"; the dynamic rate of change characteristics of the radio frequency signal assists in verification. Frequent signal fluctuations correspond to moving states, while stable signals correspond to stationary states. During statistical analysis, the system groups personnel in the target area according to their behavioral states and calculates the distribution density of each state. This sub-density can intuitively reflect the activity characteristics of personnel within the area; for example, a sudden increase in the emergency running sub-density indicates potential risk.

[0138] The spatiotemporal compliance sub-density is obtained by classifying and statistically analyzing the matching results of personnel's scheduling information, access history, and real-time location information. The backend database pre-stores information such as personnel's scheduling time slots and authorized access areas. Real-time location information acquired through visual and radio frequency collaboration provides the core basis for compliance verification. The system matches personnel's real-time location with authorized spatiotemporal ranges, categorizing the results into "spatiotemporal compliance," "time-period violation," "area unauthorized access," and "double violation," and then groups and counts the number of personnel in the target area by category and calculates the density. For example, in the core computer room area, if the area unauthorized access sub-density is greater than zero, it indicates that unauthorized personnel have entered, providing a direct clue for anomaly detection.

[0139] The process of comparing the real-time calculated overall population density and attribute sub-density with a preset dynamic threshold rule base in step S3 to determine whether there is an abnormal density situation is described. The abnormal density situation may include, but is not limited to, the following four types:

[0140] The first type is regional cluster density anomaly:

[0141] The overall population density is compared with the maximum carrying capacity threshold calculated based on the regional three-dimensional terrain structure, passage capacity, and real-time event status. When the overall population density exceeds the threshold and the duration exceeds the preset time, it is determined to be an abnormal regional population density.

[0142] The process of generating management operation instructions for the evacuation guidance triggered by the abnormal regional agglomeration density based on multi-dimensional attribute collaboration further includes:

[0143] ① The movement status of people in the area is classified based on the identified behavioral state sub-density, and the people are grouped by function based on the identified functional sub-density;

[0144] ② By calculating the optimal path load for people to reach each safety exit within the area, and combining it with the abnormal dwelling areas identified by the spatiotemporal compliance sub-density, a differentiated evacuation route plan is generated;

[0145] ③ Based on the differentiated evacuation route plan, send evacuation instructions with different priorities and different route guidance to the smart terminals held by different functional groups.

[0146] Specifically, the core criterion for identifying abnormal regional population density is the imbalance between the actual carrying capacity and the maximum safe carrying capacity of the area. This judgment process requires comprehensive consideration of the area's physical characteristics and real-time dynamics. Specifically, the system first compares the real-time calculated overall population density with a maximum carrying capacity threshold generated based on the area's three-dimensional terrain structure, passageway capacity, and real-time event status. The three-dimensional terrain structure determines the upper limit of the physical space the area can accommodate, passageway capacity limits the efficiency boundary of personnel evacuation, and real-time event statuses such as exhibitions, gatherings, and shift changes dynamically adjust the carrying capacity threshold to adapt to temporary personnel flow needs. When the overall population density exceeds this dynamic threshold, and the duration of this threshold exceedance reaches a preset time, the system determines that the regional population density is abnormal. This effectively eliminates false alarms caused by instantaneous personnel movement, ensuring the reliability of the judgment.

[0147] The evacuation guidance mechanism triggered by abnormal regional density is not a traditional unified command issuance, but rather a collaborative decision-making process based on multi-dimensional attribute sub-densities. The corresponding management operation command generation process exhibits refined and differentiated characteristics: the system first relies on behavioral state sub-densities to classify personnel within the area into categories such as "stationary," "slowly moving," and "fast moving" based on their movement status. Simultaneously, it combines functional sub-densities to categorize personnel into functional groups such as "emergency guides," "general personnel," "special groups such as the elderly, infirm, and disabled," and "area management personnel," to achieve differentiated guidance subsequently. The system calculates the optimal path load for each person within the area to reach each safety exit through path planning algorithms, avoiding congestion at single exits. Furthermore, it incorporates abnormally loitering areas identified by spatiotemporal compliance sub-densities, such as areas of personnel illegally blocking passageways, proactively avoiding such obstacles in path planning, ultimately generating differentiated evacuation paths for different groups. Based on the route planning results, personalized evacuation instructions are sent to the smart terminals (such as mobile phones, work bracelets, walkie-talkies, etc.) held by personnel in different functional groups. For example, priority instructions for designated locations are issued to emergency guides, guidance on the optimal route and real-time congestion alerts are sent to ordinary personnel, and exclusive instructions for special groups to receive personal assistance and accessibility are provided. Through the dual differences in priority and route, efficient and orderly evacuation is achieved.

[0148] The second type is characterized by abnormal functional compliance density:

[0149] Based on the functional sub-density and the spatiotemporal compliance sub-density, a dual compliance detection method, including job absence detection and unauthorized trajectory analysis, is used to determine whether there is an abnormal functional compliance density.

[0150] The execution process of the dual compliance testing further includes:

[0151] Job Deficiency Detection: The functional sub-density of each key area is compared with the expected functional density distribution map dynamically generated based on scheduling information and workflow dependencies. The job deficiency detection result is determined by detecting whether the functional sub-density corresponding to the required function is lower than the minimum configuration threshold and whether the duration exceeds the business process tolerance window.

[0152] Unauthorized Trajectory Analysis: The spatiotemporal compliance sub-density is matched with a predefined three-dimensional permission matrix. When an unauthorized target is detected whose real-time location information does not match the authorized spatiotemporal range, the historical trajectory and current activity pattern of the unauthorized target are back-analyzed to construct an abnormal behavior sequence to determine the unauthorized trajectory analysis results.

[0153] Specifically, the functional compliance density anomaly is used to assess two main indicators: the matching degree between regional functional needs and staffing, and the adaptability of personnel permissions to spatiotemporal scope. Accurate judgment is achieved through dual compliance checks of functional sub-density and spatiotemporal compliance sub-density, ensuring the functional integrity and permission security of regional operations. The dual compliance checks are not conducted independently, but rather through data correlation to form complementary verification. The specific execution process is as follows:

[0154] Job shortage detection aims to ensure that job functions meet business needs. This is achieved by comparing the density of functional sub-functions with a dynamically generated expected functional density distribution map. The expected functional density distribution map is not a fixed template but is dynamically generated by the system based on personnel scheduling information and the dependencies between job workflows. For example, a production line requires the collaboration of operators, quality inspectors, and material handlers. The system clearly defines the necessary functions and corresponding personnel density for each key area at different times. The system compares the actual functional sub-density of each key area with this expected distribution map, focusing on whether the sub-density corresponding to necessary functions is below the minimum configuration threshold (e.g., nurses in the operating room or maintenance personnel in a substation), and whether the duration of this deficiency exceeds the business process tolerance window. If the density of a key functional sub-function is consistently insufficient and affects business operations, the job shortage detection result is considered abnormal.

[0155] Unauthorized access trajectory analysis aims to ensure that personnel activities comply with authorized scope, and is accomplished by matching spatiotemporal compliance sub-densities with a predefined three-dimensional permission matrix. The three-dimensional permission matrix integrates authorization information across three dimensions: personnel, time, and region, clearly defining the area a specific type of personnel can access during a specific time period. The system compares the personnel location and permission matching results corresponding to the spatiotemporal compliance sub-densities with this matrix. When an unauthorized target is detected whose real-time location does not match the authorized spatiotemporal range, it is not directly judged as abnormal. Instead, it retrospectively analyzes the historical trajectory of the unauthorized target (e.g., whether there have been multiple unauthorized attempts recently, whether the unauthorized path deliberately evades monitoring) and current activity patterns (e.g., the time spent in the unauthorized area, whether it has contacted sensitive equipment), constructing a complete abnormal behavior sequence. If the sequence shows that the unauthorized behavior is proactive and purposeful, the unauthorized access trajectory analysis result is abnormal. Combined with the job absence detection results, it is ultimately determined whether it constitutes a functional compliance density anomaly. For example, if a quality inspector in a production workshop has insufficient functional sub-density (job absence), or if an unauthorized visitor appears in the quality inspection area (unauthorized access trajectory), it is comprehensively judged as a functional compliance density anomaly.

[0156] The third type is abnormal functional clustering density:

[0157] The instantaneous change rate and spatial distribution entropy value of the specific functional sub-density are monitored. When the instantaneous change rate generated by the concentration of personnel in a specific function exceeds the dynamic risk threshold in a short period of time, and the spatial distribution entropy value changes abruptly from a uniform distribution to a highly clustered distribution, it is determined that the functional cluster density is abnormal, and cross-validation based on the behavioral state sub-density within the region is triggered.

[0158] The cross-validation process based on the density of behavioral state sub-states within the region further includes:

[0159] ① When the functional aggregation density is determined to be abnormal, the behavioral state sub-density is obtained within the same time period and spatial range as the abnormal functional aggregation, and the time series change curve of the functional sub-density is aligned and the correlation is calculated with the concurrent change curve of the behavioral state sub-density.

[0160] ② If the calculation results show that the occurrence time of functional aggregation is highly correlated with the outbreak time of abnormal behavior and the spatial range highly overlaps, then the confidence level of the functional aggregation density anomaly is increased and it is determined to be a high-probability emergency response event; otherwise, the confidence level of the functional aggregation density anomaly is decreased.

[0161] Specifically, the anomaly in functional cluster density addresses the potential risk of unplanned concentrations of personnel in specific functions. Its judgment logic considers both quantitative changes and spatial distribution, while also incorporating behavioral state sub-density for cross-validation to eliminate interference from routine work-related clustering. Specifically, the system monitors the instantaneous change rate and spatial distribution entropy value of specific functional sub-densities in real time. The instantaneous change rate reflects the fluctuation in the number of personnel in that function over a short period, while the spatial distribution entropy value characterizes the degree of dispersion and concentration of personnel within a region. When the instantaneous change rate of specific functional personnel exceeds the dynamic risk threshold within a short period, and the spatial distribution entropy value abruptly changes from a uniform distribution (personnel dispersed in their respective work areas) to a highly concentrated distribution (personnel concentrated in a certain area), the system initially determines that the functional cluster density is abnormal, and then triggers the cross-validation process.

[0162] Cross-validation based on behavioral state sub-density is key to improving the accuracy of anomaly detection. This involves analyzing the correlation between functional clustering and behavioral anomalies to eliminate false positives. Specifically, after initially determining an abnormal functional clustering density, the system immediately extracts behavioral state sub-density data within the same time period and spatial range as the abnormal clustering. The time-series change curves of functional sub-densities are aligned with the contemporaneous change curves of behavioral state sub-densities, and their correlation is calculated. If the correlation analysis shows that the occurrence time of the functional clustering highly overlaps with the outbreak time of the abnormal behavioral state (such as "running rapidly," "emotional agitation," or "blocking equipment"), and their spatial coverage is completely identical, it indicates that the functional clustering is not routine work collaboration but rather triggered by an emergency event. In this case, the system increases the confidence level of the abnormal functional clustering density, classifying it as a high-probability emergency response event and immediately triggering the corresponding contingency plan. Conversely, if the functional clustering is associated with routine behavioral states such as "calm conversation" or "collaborative work," the anomaly confidence level is reduced, classifying it as a routine work clustering, and no emergency response is required.

[0163] The fourth type is anomaly in state aggregation density:

[0164] The specific density of the behavior state is compared in real time with the state alarm threshold preset based on the scene safety rules. At the same time, the spatial propagation speed of the behavior state and the direction of personnel flow are monitored. When the state alarm threshold is exceeded and a trend of spreading to key areas is shown, it is determined that the state aggregation density is abnormal.

[0165] Specifically, the abnormal density of state clusters targets the concentration of risk-spreading behavioral states. Its assessment not only focuses on the density threshold of the behavioral state but also emphasizes its propagation trend and potential hazards, preventing local risks from escalating into regional crises. Specifically, the system compares the density of specific behavioral states (such as "pushing and shoving," "emergency calls for help," and "disorderly running") with pre-set state alarm thresholds based on scenario safety rules in real time; this is the fundamental condition for judgment. Simultaneously, the system monitors the spatial propagation speed of the behavioral state (such as the rate of diffusion from the edge of the area to the center) and the direction of personnel flow, which is a crucial supplement to the risk level assessment. If the density of a risky behavior only slightly exceeds the threshold, but the propagation speed is extremely fast and the direction of personnel flow is directly towards critical areas, such as safety exits in shopping malls, emergency passages in hospitals, or warehouses storing flammable and explosive materials in factories, then its potential hazard is far greater than high-density clustering within a fixed area.

[0166] When the density of a specific behavioral state exceeds the state alarm threshold, and a clear trend of this behavioral state spreading towards key areas is detected, the system determines that the state aggregation density is abnormal. For example, in a train station waiting hall, if the density of the "running and crowding" behavioral state exceeds the threshold, and the flow of people is all pointing towards a single ticket gate, the system will immediately determine the anomaly and trigger a multi-level response: on the one hand, it sends alarm information to the terminals of area management personnel, marking the location of the anomaly and the direction of spread; on the other hand, it issues guidance instructions through broadcasts and electronic screens in the area, prompting people to move in an orderly manner; if the spread trend intensifies, it will also link security equipment to focus on the abnormal area, providing real-time video support for subsequent handling. This judgment mechanism breaks through the limitations of judging anomalies solely by density, and achieves proactive early warning and handling of risks by combining propagation characteristics.

[0167] The regional personnel intelligent management method provided in this application embodiment can be applied to regional personnel intelligent management equipment. Figure 3 The hardware structure block diagram of the regional personnel intelligent management device is shown. Figure 3 The hardware structure of the regional personnel intelligent management device may include: at least one processor 1, at least one communication interface 2, at least one memory 3, and at least one communication bus 4;

[0168] In this embodiment of the application, the number of processor 1, communication interface 2, memory 3, and communication bus 4 is at least one, and processor 1, communication interface 2, and memory 3 communicate with each other through communication bus 4;

[0169] Processor 1 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention.

[0170] Memory 3 may include high-speed RAM, and may also include non-volatile memory, such as at least one disk storage device;

[0171] The memory stores a program, which the processor can call. The program is used for:

[0172] The radio frequency positioning network is used to obtain the approximate location information of the personnel, and the corresponding visual recognition device is dispatched based on the approximate location information. The attribute category and location information of the personnel in the area are determined by recognizing the dynamic coded tag. The coded information of the dynamic coded tag includes a periodically dynamically updated visible light part and / or a fluorescent hidden part, and is associated with the personnel attribute category information in the background database.

[0173] Based on the identified attribute categories and location information of the personnel, the overall personnel density of the target area is calculated in real time, and at the same time, at least one attribute sub-density based on the personnel attribute classification is calculated. The attribute sub-density is the density obtained by statistically analyzing the personnel with the same attribute category in the target area.

[0174] The overall personnel density and the attribute sub-density calculated in real time are compared with a preset dynamic threshold rule library to determine whether there is an abnormal density situation. The dynamic threshold rule library includes attribute sub-density thresholds set based on different attribute category combinations and / or different spatiotemporal scenarios.

[0175] If an abnormal density situation is detected, corresponding management operation instructions will be generated and executed based on the anomaly type and the preset management contingency plan library.

[0176] Optionally, the refined and extended functions of the program can be referred to the above description.

[0177] This application embodiment also provides a readable storage medium that can store a program suitable for execution by a processor, the program being used for:

[0178] The radio frequency positioning network is used to obtain the approximate location information of the personnel, and the corresponding visual recognition device is dispatched based on the approximate location information. The attribute category and location information of the personnel in the area are determined by recognizing the dynamic coded tag. The coded information of the dynamic coded tag includes a periodically dynamically updated visible light part and / or a fluorescent hidden part, and is associated with the personnel attribute category information in the background database.

[0179] Based on the identified attribute categories and location information of the personnel, the overall personnel density of the target area is calculated in real time, and at the same time, at least one attribute sub-density based on the personnel attribute classification is calculated. The attribute sub-density is the density obtained by statistically analyzing the personnel with the same attribute category in the target area.

[0180] The overall personnel density and the attribute sub-density calculated in real time are compared with a preset dynamic threshold rule library to determine whether there is an abnormal density situation. The dynamic threshold rule library includes attribute sub-density thresholds set based on different attribute category combinations and / or different spatiotemporal scenarios.

[0181] If an abnormal density situation is detected, corresponding management operation instructions will be generated and executed based on the anomaly type and the preset management contingency plan library.

[0182] Optionally, the refined and extended functions of the program can be referred to the above description.

[0183] This application also provides a computer program product, including a computer program, wherein the computer program is executed by a processor using the following method:

[0184] The radio frequency positioning network is used to obtain the approximate location information of the personnel, and the corresponding visual recognition device is dispatched based on the approximate location information. The attribute category and location information of the personnel in the area are determined by recognizing the dynamic coded tag. The coded information of the dynamic coded tag includes a periodically dynamically updated visible light part and / or a fluorescent hidden part, and is associated with the personnel attribute category information in the background database.

[0185] Based on the identified attribute categories and location information of the personnel, the overall personnel density of the target area is calculated in real time, and at the same time, at least one attribute sub-density based on the personnel attribute classification is calculated. The attribute sub-density is the density obtained by statistically analyzing the personnel with the same attribute category in the target area.

[0186] The overall personnel density and the attribute sub-density calculated in real time are compared with a preset dynamic threshold rule library to determine whether there is an abnormal density situation. The dynamic threshold rule library includes attribute sub-density thresholds set based on different attribute category combinations and / or different spatiotemporal scenarios.

[0187] If an abnormal density situation is detected, corresponding management operation instructions will be generated and executed based on the anomaly type and the preset management contingency plan library.

[0188] Optionally, the refined and extended functions of the program can be referred to the above description.

[0189] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0190] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0191] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for intelligent management of regional personnel, characterized in that, This is achieved based on dynamically coded tags configured for each person within the area, a radio frequency positioning network deployed throughout the area, and multimodal visual recognition devices. The method includes: The radio frequency positioning network is used to obtain the approximate location information of the personnel, and the corresponding visual recognition device is dispatched based on the approximate location information. The attribute category and location information of the personnel in the area are determined by recognizing the dynamic coded tag. The coded information of the dynamic coded tag includes a periodically dynamically updated visible light part and / or a fluorescent hidden part, and is associated with the personnel attribute category information in the background database. Based on the identified attribute categories and location information of the personnel, the overall personnel density of the target area is calculated in real time, and at the same time, at least one attribute sub-density based on the personnel attribute classification is calculated. The attribute sub-density is the density obtained by statistically analyzing the personnel with the same attribute category in the target area. The overall personnel density and the attribute sub-density calculated in real time are compared with a preset dynamic threshold rule library to determine whether there is an abnormal density situation. The dynamic threshold rule library includes attribute sub-density thresholds set based on different attribute category combinations and / or different spatiotemporal scenarios. If an abnormal density situation is detected, corresponding management operation instructions will be generated and executed based on the anomaly type and the preset management contingency plan library.

2. The method according to claim 1, characterized in that, The radio frequency positioning network is used to obtain a rough area location information of a person, and a corresponding visual recognition device is scheduled based on the rough area location information. The attribute category and location information of the person in the area are determined by recognizing the dynamic coded tag, including: The radio frequency positioning network periodically detects signals to obtain rough location information of each person in the area; Based on the rough area location information, visual recognition devices deployed above the corresponding area are scheduled to acquire images of the target area; During the image acquisition process, the images are analyzed to detect areas where people block each other or where dynamically coded tags overlap. If present, at least two visual recognition devices with different spectral characteristics are scheduled to perform collaborative image acquisition and multimodal image fusion processing on the target area to resolve the occluded dynamic coded label. From the acquired images and the fused images, locate and identify the current encoding information of the dynamic coded tag worn by each person; The identified current encoding information is matched with the background database to obtain the corresponding personnel attribute category, and the personnel's location information is determined by combining image and radio frequency signal analysis.

3. The method according to claim 2, characterized in that, Before scheduling the visual recognition device, the method further includes extracting and analyzing the multi-dimensional radio frequency features of the radio frequency positioning network signal to optimize the visual recognition strategy. The multi-dimensional radio frequency features include at least multipath delay difference features, signal strength gradient features, and signal dynamic change rate features. The multipath delay difference features are used to characterize the richness of signal propagation paths, the signal strength gradient features are used to characterize the spatial variation of the received signal strength at different positioning anchor points, and the signal dynamic change rate features are used to characterize the frequency and amplitude of signal strength changes over time. The process of optimizing the visual recognition strategy includes: Based on the multipath delay difference characteristics, it is determined whether there is a densely populated sub-region within the approximate location. If so, the visual recognition device is scheduled to prioritize the multimodal collaborative acquisition and fusion recognition of the region. Based on the signal strength gradient characteristics, the relative distance distribution between different people in the approximate area is calculated to determine the focusing and field of view segmentation parameters when the visual recognition device acquires data. Based on the dynamic change rate characteristics of the signal, the visual acquisition frame rate is dynamically adjusted, with a high frame rate for moving personnel and a low frame rate for stationary personnel.

4. The method according to claim 1, characterized in that, The attribute sub-density includes at least one of the function sub-density, behavior state sub-density, and spatiotemporal compliance sub-density; The functional sub-density is obtained by classifying and statistically analyzing the job functions of personnel; The behavioral state sub-density is obtained by classifying and statistically analyzing the human behavioral states identified through visual posture analysis or radio frequency signal features. The spatiotemporal compliance sub-density is obtained by classifying and statistically analyzing the matching results of personnel's scheduling information, access history, and real-time location information.

5. The method according to claim 4, characterized in that, The real-time calculated overall population density and attribute sub-densities are compared with a preset dynamic threshold rule base to determine whether there is an abnormal density situation, including: The overall population density is compared with the maximum carrying density threshold calculated based on the regional three-dimensional terrain structure, passage capacity and real-time event status. When the overall population density exceeds the threshold and the duration exceeds the preset time, it is determined to be an abnormal regional cluster density. Based on the functional sub-density and the spatiotemporal compliance sub-density, a dual compliance detection method, including job absence detection and unauthorized trajectory analysis, is used to determine whether there is an abnormal functional compliance density. Monitor the instantaneous change rate and spatial distribution entropy value of the specific functional sub-density. When the instantaneous change rate generated by the concentration of specific functional personnel exceeds the dynamic risk threshold in a short period of time, and the spatial distribution entropy value changes abruptly from a uniform distribution to a highly clustered distribution, it is determined that the functional cluster density is abnormal, and cross-validation based on the behavioral state sub-density within the region is triggered. The specific density of the behavior state is compared in real time with the state alarm threshold preset based on the scene safety rules. At the same time, the spatial propagation speed of the behavior state and the direction of personnel flow are monitored. When the state alarm threshold is exceeded and a trend of spreading to key areas is shown, it is determined that the state aggregation density is abnormal.

6. The method according to claim 5, characterized in that, The process of generating management operation instructions for evacuation guidance based on multi-dimensional attribute collaboration triggered by the abnormal regional agglomeration density includes: The movement status of people in the area is classified based on the identified behavioral state sub-density, and the people are grouped by function based on the identified functional sub-density. By calculating the optimal path load for people to reach each safety exit within the area, and combining this with the abnormal dwelling areas identified by the spatiotemporal compliance sub-density, differentiated evacuation route planning is generated. Based on the differentiated evacuation route plan, evacuation instructions with different priorities and different route guidance are sent to the smart terminals held by different functional groups.

7. The method according to claim 5, characterized in that, The execution process of the dual compliance check includes: The functional sub-density of each key area is compared with the expected functional density distribution map dynamically generated based on scheduling information and workflow dependencies. The job shortage detection result is determined by detecting whether the functional sub-density corresponding to the necessary functions is lower than the minimum configuration threshold and whether the duration exceeds the business process tolerance window. The spatiotemporal compliance sub-density is matched with a predefined three-dimensional permission matrix. When an unauthorized target is detected whose real-time location information does not match the authorized spatiotemporal range, the historical trajectory and current activity pattern of the unauthorized target are back-analyzed to construct an abnormal behavior sequence in order to determine the unauthorized trajectory analysis results.

8. The method according to claim 5, characterized in that, The cross-validation process based on the density of behavioral state sub-states within the region includes: When the functional aggregation density is determined to be abnormal, the behavioral state sub-density is obtained within the same time period and spatial range as the abnormal functional aggregation, and the time series change curve of the functional sub-density is aligned and the correlation is calculated with the concurrent change curve of the behavioral state sub-density. If the calculation results show that the occurrence time of functional aggregation is highly correlated with the outbreak time of abnormal behavior and the spatial range highly overlaps, then the confidence level of the functional aggregation density anomaly is increased, and it is determined to be a high-probability emergency response event; otherwise, the confidence level of the functional aggregation density anomaly is decreased.

9. The method according to claim 1, characterized in that, When assessing density anomalies, the following should also be included: Real-time monitoring and comparison of the changing trends of attribute sub-densities of the same attribute category across different regions; When a specific attribute sub-density in the first region is detected to decrease significantly within a certain period of time, and the same attribute sub-density in the second region adjacent to the first region increases significantly within a similar period of time, and the changes are correlated, it is determined that a cross-regional personnel transfer event has occurred, and the root cause is inferred by combining the other sub-density trends of the first region and the second region.

10. The method according to claim 1, characterized in that, After determining the attribute category and location information of the personnel within the area by identifying the dynamic coded tag, the method further includes: Obtain a causal knowledge graph containing prior business rules for the scenario; The attribute categories and location information of the personnel within the area determined by identifying the dynamic coding tags, along with the causal knowledge graph, are input into the causal inference engine for logical consistency verification. If the logical consistency check finds a conflict, an identity verification process is triggered, which includes: Re-identification and verification of dynamic coded tags of relevant personnel based on multimodal image fusion; If the verification is correct, based on the personnel's historical behavior data and shift information, combined with the causal knowledge graph, the conflicting attribute categories are probabilistically corrected until the attribute category matching result that conforms to the prior business logic of the scenario is output.

11. A regional personnel intelligent management device, characterized in that, Including memory and processor; The memory is used to store programs; The processor is used to execute the program to implement the various steps of the regional personnel intelligent management method as described in any one of claims 1-10.

12. A readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements each step of the regional personnel intelligent management method as described in any one of claims 1-10.

13. A computer program product, comprising a computer program, characterized in that, The computer program, when run by a processor, executes the steps of the regional personnel intelligent management method as described in any one of claims 1-10.

14. A dynamic coded tag for use in the regional personnel intelligent management method as described in any one of claims 1-10, characterized in that, include: Label base; A visible light coding area is disposed on the label substrate, the visible light coding area containing a combination of numbers and letters that are periodically and dynamically updated; A fluorescent material layer disposed on or within the label substrate, the fluorescent material layer exhibiting a hidden coded pattern under excitation light of a specific wavelength; A microcontroller is used to control the encoding of the visible light encoding area to be synchronously updated according to the personnel attribute category information in the background database.

Images