Scenic spot personnel distribution monitoring and early warning system and method

Abstract

The invention discloses a scenic spot personnel distribution monitoring and early warning system and method, and relates to the technical field of scenic spot monitoring. The system comprises a memory for storing different security level early warning areas divided based on monitoring area boundaries and allowed moving paths; the real-time monitoring system is used for comparing the actual position of the personnel with the allowed path and generating an alarm signal when the path exceeds; the trajectory tracking system continuously tracks the subsequent path of the personnel until the personnel enter the allowed path again; and the behavior learning system updates the security rules according to feedback of the management terminal. The method comprises the steps of obtaining the real-time position of a person and carrying out trajectory compliance analysis, selecting an alarm level when deviation is detected, executing dynamic resynchronization through a trajectory tracking system, and updating safety rules of a geographic information model according to feedback. According to the system and the method, dynamic monitoring and early warning of the scenic spot personnel are realized, and the efficiency and the accuracy of scenic spot safety management are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of scenic area monitoring, and in particular to a scenic area personnel distribution monitoring and early warning system and method. Background Art

[0002] Traditional surveillance systems often use fixed cameras or single sensors for localized monitoring. This limited detection range prevents them from fully covering every corner of a scenic area. This is especially true for large scenic areas with complex terrain, where blind spots can easily form, preventing timely detection of unusual personnel movements or dangerous behavior. For example, in mountainous scenic areas, traditional cameras may be unable to monitor tourist activity in remote areas due to terrain obstructions.

[0003] Secondly, traditional systems lack effective management of human movement paths, often using static safety rules that are unable to adapt to the dynamic changes in scenic environments. For example, during the rainy season, the risk level in certain stream areas increases, but traditional systems cannot automatically adjust safety levels and permitted paths, and still use conventional rules, which can easily lead to false positives or missed alerts.

[0004] Furthermore, traditional surveillance systems are inadequate for detecting unusual human behavior. When a person strays from their normal path, they often only issue an alarm signal, but lack the ability to continuously track and dynamically manage their subsequent movements, making it impossible to achieve a closed-loop risk management system. For example, if a tourist strays into a restricted area, traditional systems cannot continuously track their movements after issuing an alarm, making it difficult to ensure they return to a safe area in a timely manner.

[0005] Furthermore, traditional systems don't fully utilize historical data, failing to optimize and update safety rules based on actual operational conditions and historical accident data. This results in a gradual decline in the system's adaptability and accuracy over time. For example, if visitors frequently stray from the designated path in a certain area, the traditional system won't automatically include that path in the permitted movement path and will still consider it a violation, leading to an increased rate of false alarms.

[0006] While some scenic spots have introduced intelligent monitoring technology with the advancement of technology, they still face challenges in managing complex scenic environments and multi-target monitoring tasks, such as low collaborative efficiency, low detection accuracy, and irrational task allocation. For example, the lack of effective collaboration between monitoring subsystems leads to information gaps during personnel tracking. Monitoring strategies cannot be adjusted promptly to meet real-time monitoring needs as tourist behavior patterns change with factors such as season and activity. Furthermore, a lack of scientific and rational methods for optimizing safety rules prevents the full utilization of historical data, resulting in wasted detection resources and low efficiency.

[0007] Therefore, there is an urgent need for a scenic area personnel distribution monitoring and early warning system and method that can fully cover, dynamically adjust, continuously track and have self-learning capabilities to solve the above-mentioned problems existing in traditional monitoring systems and improve the efficiency and accuracy of scenic area personnel safety management. Summary of the Invention

[0008] The purpose of the present invention is to provide a scenic area personnel distribution monitoring and early warning system and method to solve the technical problems raised in the technical background.

[0009] The main purpose of this application is to provide a scenic area personnel distribution monitoring and early warning system, including: The memory stores a scenic area geographic information model, wherein the scenic area geographic information is divided into multiple warning areas with different security levels based on multiple predefined monitoring area boundaries, and stores a security level value and at least one allowed personnel movement path for each warning area.

[0010] The real-time monitoring system compares the actual position of a person with at least one allowed movement path in the current warning area during the operation of the scenic area, and generates an alarm signal and a corresponding safety level when the actual position of a person exceeds the allowed movement path.

[0011] The trajectory tracking system, in response to the alarm signal, completes synchronization by continuously tracking the subsequent movement path of the person and comparing it with the safety rules of the warning area until the person's position re-enters the allowed path of any warning area.

[0012] The behavior learning system receives a feedback instruction from the scenic area management terminal in response to the alarm signal, and updates the safety rules of the geographic information model according to the feedback instruction.

[0013] Furthermore, the behavior learning system dynamically adjusts the safety level value associated with the current warning area according to the feedback instruction.

[0014] The behavior learning system includes a feedback parsing engine, which is configured to: When the feedback instruction includes a safety level adjustment, the level value associated with the current warning area is automatically updated.

[0015] When the feedback is confirmed to be a false alarm, the sensitivity parameters of similar alarms in the area are reduced.

[0016] When feedback requires a path correction, the path generator is activated to create a temporary permissible path.

[0017] Furthermore, the boundaries of the monitoring area are determined by running a safety risk analysis algorithm on the terrain data of the scenic area through a risk modeling engine, and identifying the boundaries of high-risk areas in combination with historical accident data.

[0018] The risk modeling engine is configured to perform the following operations: Input scenic area terrain point cloud data, hydrogeological report, and historical accident database.

[0019] Run the Monte Carlo security risk simulation algorithm.

[0020] Combined with crowd flow dynamics model to identify the boundaries of high-risk areas.

[0021] Outputs a set of graded boundary coordinates with weight coefficients.

[0022] Furthermore, the geographic information model is generated by simulating the flow paths of personnel, recording the safety status of each warning area and the path conversion rules.

[0023] The geographic information model is constructed by a behavior simulator: Load the scenic area building BIM model and vegetation coverage data.

[0024] Inject typical tourist behavior patterns based on seasonal changes.

[0025] Run a multi-agent path planning simulation.

[0026] A state transition matrix of each area is recorded, wherein the transition matrix includes a transition relationship between a normal path / an emergency path / a restricted area.

[0027] Generate a path rule base with temporal and spatial constraints.

[0028] Furthermore, the scenic area geographical information includes: The area rule table stores the security level value and path index corresponding to each warning area.

[0029] A path mapping table stores a set of allowed movement paths for each warning area, and the path mapping table is accessed through the path index.

[0030] The path mapping table is organized by index cluster and contains: The master path set contains the GPS coordinate sequence of the allowed movement paths.

[0031] Emergency path set, including evacuation paths and triggering conditions.

[0032] The transfer rule set records the topological relationship of legal cross-region transfers.

[0033] Furthermore, the trajectory tracking system enters a verification mode after completing synchronization. If the position of the person is detected to meet the warning area rules for N consecutive times, the synchronization is determined to be valid.

[0034] The trajectory tracking system includes a verification controller: After synchronization is complete, it enters verification mode with adjustable duration.

[0035] Perform position compliance checks N times consecutively.

[0036] When the pass rate of N verifications is ≥95%, the synchronization is determined to be valid.

[0037] Activate the regional rule adaptation module to adjust the current path confidence weight.

[0038] Furthermore, the real-time monitoring system suspends alarming in verification mode, and if any subsequent position violates the area rules, the trajectory tracking system is triggered to re-execute synchronization.

[0039] Wherein, the real-time monitoring system performs the following operations in verification mode: Enable the relaxed alert policy.

[0040] If the single position deviation standard deviation is greater than the preset tolerance value, the trajectory tracking system is triggered to restart synchronization.

[0041] If there are two consecutive deviations, the verification mode will be forced to exit and a regional abnormality report will be generated.

[0042] Furthermore, it also includes: The abnormal behavior record library and the spatiotemporal indexed deviation trajectory data set are stored in the memory.

[0043] The trajectory tracking system stores the trajectory data of the person deviating from the path into the abnormal behavior record library, wherein: The trajectory tracking system performs: Compress and store the GPS coordinate sequence that deviates from the path.

[0044] Associate the corresponding timestamp and environment parameters.

[0045] High-frequency deviation areas are marked to form a risk thermal layer.

[0046] Furthermore, the behavior learning system copies the abnormal trajectory data to a learning database and counts the number of repetitions of the same path deviation; when the number of repetitions reaches a threshold, the geographic information model is updated to include the path as a new allowed movement path.

[0047] The present invention also provides a method for monitoring and early warning of personnel in a scenic area, comprising the following steps: The real-time location of personnel is obtained through a multi-source perception layer, and the real-time monitoring system compares the position of the personnel to be measured with the allowed movement path of the current warning area; and performs trajectory compliance analysis: projecting the personnel coordinates onto a three-dimensional geographic model; comparing with the current warning area path rule library; when a path deviation is detected, selecting an alarm level based on the deviation type.

[0048] Dynamic resynchronization is performed through the trajectory tracking system: trajectory prediction based on the motion model is initiated, the adjacent area transfer rules are continuously matched, a synchronization signal is triggered when a person enters a legal path, and subsequent positions are continuously tracked until any warning area rules are rematched.

[0049] The behavioral learning system updates the security rules of the geographic information model based on feedback from the management terminal. The behavioral learning system performs rule optimization: parsing the structured feedback instructions from the management terminal; reducing the monitoring sensitivity of the area when the feedback confirms a false alarm; starting the regional rule reconstruction engine when the feedback requires a rule update; and outputting an incrementally updated version of the geographic information model.

[0050] This application has the following beneficial effects: 1) This application features dynamic safety level adjustment. The behavioral learning system dynamically adjusts the safety level of a warning area based on feedback from the management terminal. When feedback includes a safety level adjustment, the feedback parsing engine automatically updates the current warning area's level. For example, administrators can adjust the risk level of a zone based on seasonal changes (e.g., raising the risk level of a stream area from 2 to 4 during the rainy season) or activity patterns. This dynamic adjustment mechanism avoids the response lag caused by fixed levels and adapts to the dynamic changes in the scenic area environment. Furthermore, when feedback is confirmed as a false alarm, the system reduces the sensitivity parameters for similar alarms in that area, including positioning error tolerance, path deviation threshold, and alarm delay time. This reduces false alarms and prevents excessive alarms from disrupting scenic area operations. When feedback requests a path correction, the system activates the path generator to create temporary permitted paths. For example, it calculates feasible paths based on the BIM model and incorporates spatiotemporal constraints, which are then updated in the path mapping table. This dynamic adjustment of safety levels and paths enables the system to accurately respond to actual conditions, improving both the accuracy and effectiveness of warnings.

[0051] 2) Precisely demarcate monitoring area boundaries based on a risk modeling engine: Monitoring area boundaries are determined by applying a safety risk analysis algorithm to scenic area terrain data using the risk modeling engine, combined with historical accident data to identify high-risk area boundaries. The risk modeling engine inputs scenic area terrain point cloud data (such as 3D topography generated by LiDAR scans, identifying steep slopes and rockfall areas), hydrogeological reports (identifying flood / landslide risk areas), and a historical accident database (statistically analyzing accident hotspots). The engine then runs a Monte Carlo safety risk simulation algorithm, randomly generating a large number of visitor trajectories, simulating the environmental risk associated with each trajectory, calculating accident probabilities, identifying high-risk area boundaries, and outputting a set of graded boundary coordinates with weighted coefficients. This regional demarcation approach, based on multi-source data and scientific algorithms, accurately identifies high-risk areas within the scenic area, providing an accurate geographic information foundation for subsequent personnel monitoring and early warning. This allows the monitoring system to focus on high-risk areas, improving the accuracy of scenic area safety management.

[0052] 3) Self-Optimization of the Behavioral Learning System: The behavioral learning system responds to alarm signals and receives feedback from the management terminal. Based on this feedback, it updates the safety rules of the geographic information model. The system copies abnormal trajectory data to a learning database and counts the number of recurrences of similar path deviations. When the number of recurrences reaches a threshold, the geographic information model is updated to incorporate that path as a new permitted path. For example, if a deviated path exceeds the threshold for three consecutive months with no accidents and meets safety standards (slope < 25°, no geological hazards), the system automatically adds it as a temporarily permitted path and sets spatiotemporal constraints. Furthermore, the behavioral learning system includes a feedback parsing engine. When feedback is confirmed as a false alarm, the monitoring sensitivity of that area is reduced. When feedback requires a rule update, the regional rule reconstruction engine is activated. This self-learning and self-optimization mechanism enables the system to continuously iterate the geographic information model based on actual operational conditions and adapt to changes in tourist behavior patterns within the scenic area.

[0053] 4) Tracking System Verification Mechanism Ensures Closed-Loop Risk Management: After synchronization, the tracking system enters verification mode. If the person's location is detected N times in a row as conforming to the warning zone rules, synchronization is considered valid. After synchronization is complete, the verification controller enters verification mode with adjustable duration, performing N consecutive position compliance checks. When the pass rate is ≥95%, synchronization is considered valid and the regional rule adaptive module is activated to adjust the current path confidence weight. For example, the initial weight is 0.6, and it increases according to the learning rate with each verification pass. Monitoring frequency is reduced when the weight is > 0.8, and focused drone tracking is initiated when the weight is < 0.4. In verification mode, the real-time monitoring system uses a relaxed alarm policy. If a single position deviation exceeds a preset tolerance, the tracking system will be triggered to restart synchronization. If there are two consecutive deviations, verification mode is forced to exit and a regional anomaly report is generated. This verification mechanism solves the problem of "missing tracking after an alarm" in traditional monitoring. It ensures that people who accidentally enter a restricted area can be continuously tracked until they return to a safe path and pass verification, forming a closed-loop risk management system and improving the reliability of scenic area safety management.

[0054] 5) Abnormal Behavior Data Analysis: The system stores a database of abnormal behavior records and a spatiotemporally indexed dataset of deviation trajectories. The trajectory tracking system compresses and stores trajectory data of individuals deviating from their paths (using the Douglas-Peucker algorithm, with an accuracy loss of < 0.5 meters), associates corresponding timestamps with environmental parameters (such as weather and foot traffic), and marks high-frequency deviation areas to form a risk thermal layer. The system counts deviations in 1km×1km grids, calculates regional risk scores (risk score = number of deviations per month × regional safety level), and outputs a visualization layer (red, yellow, and green to indicate high, medium, and low risk), which is updated in real time. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 Flow chart of the method of the present invention.

[0056] Figure 2 This is a schematic diagram of the system framework principle in the present invention. DETAILED DESCRIPTION

[0057] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0058] Reference Figure 1 and Figure 2The main purpose of this application is to provide a scenic area personnel distribution monitoring and early warning system, including: a memory, storing a scenic area geographic information model, the scenic area geographic information is divided into multiple early warning areas with different security levels based on multiple predefined monitoring area boundaries, and a security level value and at least one allowed personnel movement path are stored for each early warning area; a real-time monitoring system, comparing the actual position of the personnel with the at least one allowed movement path of the current early warning area during the operation of the scenic area, and generating an alarm signal and a corresponding safety level when the actual position of the personnel exceeds the allowed movement path; a trajectory tracking system, in response to the alarm signal, continuously tracking the subsequent movement path of the personnel and comparing it with the safety rules of the early warning area until the personnel position re-enters the allowed path of any early warning area to complete synchronization; a behavior learning system, in response to the alarm signal, receiving feedback instructions from the scenic area management terminal, and updating the safety rules of the geographic information model according to the feedback instructions.

[0059] In the above embodiment, the logic for constructing the geographic information of the scenic area is as follows: the warning area is divided based on the principle of differentiated safety levels. For example, the high-risk area (cliff, deep water area) has a high safety level value (such as level 5), and very few paths are allowed; the medium-risk area (steep slope, construction area) has a medium level value (level 3), and the paths are restricted.

[0060] Low-risk areas (visitor centers, flat land) have low level values (level 1) and loose paths.

[0061] Each of these paths must meet spatiotemporal constraints. For example, movement along the plank road is restricted to the guardrail path, and deviation from the viewing platform path is prohibited in wildlife sanctuaries. By digitally mapping the rules of physical space, a structured data foundation is provided for real-time monitoring, avoiding misjudgments caused by ambiguous rules in traditional monitoring.

[0062] In the above-mentioned embodiment, the real-time monitoring system utilizes dynamic compliance checking. This involves continuously matching GPS / Beidou positioning data with the current regional path database and using a spatial topology algorithm to determine whether the location is within the path buffer (e.g., deviations >5 meters trigger an alarm). The alarm classification mechanism is as follows: low-level alarms (deviating from the safe path) notify patrol personnel; high-level alarms (entering a restricted area) trigger evacuation via the public address system. Position deviations are converted into safety events, and the level of the alarm is combined to enable precise emergency response, preventing excessive alarms from disrupting operations.

[0063] In the above-mentioned embodiment, the trajectory tracking system employs a resynchronization mechanism. This mechanism means that when a person strays into a restricted area, the system activates a motion trajectory prediction model to predict their direction of movement. It continuously compares the system's trajectory with the transition rules of adjacent areas (for example, returning to the trail from a cliff requires a designated safe exit). Only when the trajectory overlaps with a legal path in a specific area is synchronization determined and the alarm is released. This solves the problem of "missing tracking after an alarm" in traditional monitoring, ensuring closed-loop risk management.

[0064] In the above embodiment, the behavioral learning system utilizes dynamic rule evolution. This dynamic rule evolution involves receiving administrator feedback (e.g., "Yesterday's cliff was falsely reported due to a tourist taking photos") and interpreting the following instructions: false alarm confirmation, reducing sensor sensitivity in that area; route change request, generating a temporary viewing path. By continuously iterating the geographic information model, the rule base adapts to actual operational needs, reducing manual configuration costs.

[0065] In the above embodiment, the four subsystems of real-time monitoring to detect anomalies, trajectory tracking for continuous intervention, behavior learning to optimize rules, and memory updating of geographic information models form a "monitoring-tracking-learning" closed loop.

[0066] In some embodiments, the behavioral learning system dynamically adjusts the security level value associated with the current warning area based on feedback instructions; and the behavioral learning system includes a feedback parsing engine, which is configured to: automatically update the level value associated with the current warning area when the feedback instruction includes a security level adjustment; reduce the sensitivity parameters of similar alarms in the area when the feedback is confirmed to be a false alarm; and activate the path generator to create a temporary allowed path when the feedback requires path correction.

[0067] In the above embodiment, dynamic security level adjustment is triggered by an administrator adjusting the regional risk level based on seasons or activities (e.g., raising the stream risk level from 2 to 4 during the rainy season). The implementation logic is as follows: the engine calls the geographic information model's API to modify the regional level; this simultaneously updates the alarm thresholds of the real-time monitoring system (e.g., reducing the deviation tolerance for a level 4 area from 10 meters to 2 meters).

[0068] This application avoids the response lag caused by fixed levels through dynamic adjustment and adapts to dynamic changes in the environment.

[0069] In the above embodiment, the parameter adjustment mechanism is also adjusted through false alarm processing and sensitivity optimization. The "sensitivity parameters" include: positioning error tolerance, path deviation threshold, and alarm delay time.

[0070] In the above embodiment, the temporary path is generated according to the following method: receiving administrator instructions; calculating feasible paths based on the BIM model; injecting spatiotemporal constraints; and updating the path mapping table.

[0071] In some embodiments, the boundaries of the monitoring area are determined by running a safety risk analysis algorithm on the scenic area terrain data through a risk modeling engine, and identifying the boundaries of high-risk areas in combination with historical accident data; wherein, the risk modeling engine is configured to perform the following operations: input scenic area terrain point cloud data, hydrogeological reports, and historical accident databases; run a Monte Carlo safety risk simulation algorithm to identify the boundaries of high-risk areas; and output a graded boundary coordinate set with weight coefficients.

[0072] In the above embodiment, the multi-source data includes topographic point cloud data, hydrogeological reports and historical accident databases. The topographic point cloud data includes three-dimensional landforms generated by lidar scanning, identifying steep slopes and rockfall areas; the hydrogeological reports include the determination of flood / landslide risk areas (such as the expansion range of river channels during the rainy season).

[0073] The historical accident database includes statistical analysis of accident hotspots, etc.

[0074] In the above-mentioned embodiment, the Monte Carlo safety risk simulation algorithm involves randomly generating 100,000 sets of tourist trajectories to construct a diverse sample of tourist movement paths. For each step of each trajectory, the corresponding environmental risk is simulated, taking into account factors such as terrain, facilities, and weather. For example, complex terrain has higher risk, and severe weather has increased risk. Based on the environmental risk and trajectory characteristics, a probabilistic model (such as an algorithm based on historical accident data) is used to calculate the probability of an accident at each point on the trajectory. All trajectory points are clustered by location, and high-accident-probability areas are counted and visualized using a heat map, where darker colors represent higher risks. The darkest areas in the heat map, representing peak probability areas, are identified and their boundaries extracted as high-risk areas for use in safety warnings and control planning.

[0075] In some embodiments, the geographic information model is generated by simulating the flow of personnel, recording the safety status of each warning area and the path conversion rules; wherein, the geographic information model is constructed through a behavior simulator: loading the scenic area building BIM model and vegetation coverage data; injecting typical tourist behavior patterns based on seasonal changes; running multi-agent path planning simulation; recording the state transition matrix of each area, the transition matrix includes the conversion relationship between normal path / emergency path / restricted area; generating a path rule library with time and space constraints.

[0076] In the above, the static defects of the traditional scenic spot rule base are solved by dynamically generating a geographic information model through a behavioral simulator. It is implemented in five steps: 1) Multi-source data loading: importing the scenic spot building BIM model (including load-bearing structures, safety exit locations) and vegetation coverage data (distinguishing between accessible lawns and ecological protection areas); 2) Tourist behavior pattern injection: presetting typical behavior rule base based on seasons / events: Peak season mode: high tourist density, path selection tends to be shortcuts; Snow season mode: automatically activate anti-slip paths, close ice viewing platforms; Large-scale event mode: bind temporary evacuation channel coordinates. 3) Multi-agent path simulation: simulate the movement trajectories of tens of thousands of tourists (including elderly, children and other agents with different speeds), and discover bottleneck areas through stress testing. 4) Output key indicators: risk coefficient of plank road detention during peak hours; emergency evacuation path traffic efficiency (such as the number of people passing through per minute). 5) State transition matrix generation: record the state transition relationship of each area under abnormal conditions: Normal path → Emergency path: In the event of a fire, the main passage of the observation deck will be closed and the west side backup exit will be automatically switched; Emergency path → Restricted area: When the backup exit is overcrowded, the adjacent dangerous area will be blocked.

[0077] In some embodiments, the scenic area geographic information includes: an area rule table, which stores the safety level value and path index corresponding to each warning area; a path mapping table, which stores the set of allowed movement paths for each warning area, and the path mapping table is accessed through the path index; wherein the path mapping table is organized by index cluster, including: a main path set, which contains a GPS coordinate sequence of allowed movement paths; an emergency path set, which contains evacuation paths and trigger conditions; and a transfer rule set, which records the topological relationship of legal cross-regional transfers.

[0078] The geographic information model defined above features a three-tiered structured storage scheme, enabling millisecond-level path compliance assessment. The regional rule table (core index) contains a record containing the region ID, security level, and path index pointer. The security level is numerically categorized (1-5) and directly linked to the alarm response level. The path mapping table (organized by index clusters) includes: The primary path set stores the GPS coordinate sequence of all compliant movement paths, each with a travel time window; the emergency path set contains evacuation paths bound to trigger conditions; and the transfer rule set describes the topological relationships between inter-regional movements.

[0079] The above data association mechanism: the real-time monitoring system calls the regional rule table to obtain the security level, locates the coordinate sequence through the path index, and verifies the cross-region legitimacy with the topological relationship.

[0080] In some embodiments, the trajectory tracking system enters a verification mode after completing synchronization. If the personnel position is detected to be in compliance with the warning area rules for N consecutive times, the synchronization is determined to be valid; wherein, the trajectory tracking system includes a verification controller: after the synchronization is completed, it enters a verification mode with adjustable duration; performs N consecutive position compliance checks; when the pass rate of N checks is ≥95%, the synchronization is determined to be valid; and activates the area rule adaptive module to adjust the current path confidence weight.

[0081] In the above, the verification controller workflow includes entering verification mode when the trajectory tracking system confirms that the person has returned to the allowed path (such as a tourist returning to the observation deck from a restricted area); the adjustable parameter N (default 5 times) is used to continuously perform position compliance verification.

[0082] In the above, the dynamic judgment logic includes: if the position is within the path buffer (such as ±5 meters) for 5 consecutive times, the synchronization is determined to be valid; if the pass rate is ≥95%, the adaptive module is activated to increase the confidence weight of the path.

[0083] In the above, the confidence weight optimization includes an initial weight of 0.6 (new path or high-risk area), which is increased by the learning rate each time the verification passes (default +0.1); when the weight is >0.8, the system reduces the monitoring frequency (such as positioning sampling from 1 time / second to 1 time / 10 seconds); when the weight is <0.4, it is automatically marked as a high-risk individual and the drone is started to track it.

[0084] In the above, the exception handling mechanism includes a single deviation found during verification, specifically including: the tolerance radius is expanded to 15 meters in flat terrain areas; and tracking is immediately restarted if it exceeds 5 meters in cliff areas.

[0085] In some embodiments, the real-time monitoring system suspends the alarm in verification mode, and if any subsequent position violates the regional rules, the trajectory tracking system is triggered to re-synchronize; wherein, the real-time monitoring system performs the following operations in verification mode: enabling a relaxed alarm strategy; if the standard deviation of a single position deviation is greater than a preset tolerance value, triggering the trajectory tracking system to restart synchronization; if there are two consecutive deviations, forcibly exiting the verification mode and generating a regional anomaly report.

[0086] During the trajectory tracking verification, a graded relaxation strategy is adopted to balance safety and false alarms. The specific strategy switching rules are as follows: when the trajectory tracking system enters the verification mode, the real-time monitoring system automatically switches to a relaxed strategy, and the path buffer radius is expanded from the conventional 5 meters to 15 meters; the maximum tolerance time is extended from 10 seconds to 30 seconds.

[0087] Key trigger conditions: In normal areas, a single deviation from the tolerance value position standard deviation > 20 meters triggers an alarm; in dangerous areas (cliffs / deep water), a single deviation from the tolerance value position standard deviation > 5 meters restarts trajectory synchronization.

[0088] Forced exit mechanism: When there are two consecutive position deviations or three cumulative deviations, the verification mode will be forced to exit and a report will be generated.

[0089] Report generation logic: includes deviation coordinate sequence, environmental parameters (weather / people flow), and historical statistics of similar events.

[0090] In the above, an abnormal behavior record library and a time-space indexed deviation trajectory data set are stored in the memory; the trajectory tracking system stores the trajectory data of the person deviating from the path in the abnormal behavior record library, wherein the trajectory tracking system performs: compressing and storing the GPS coordinate sequence of the deviation path; associating the corresponding timestamps and environmental parameters; marking the high-frequency deviation areas to form a risk thermal layer.

[0091] In the above, a deviation database with spatiotemporal indexing is constructed to drive precise safety management. Specifically, trajectory compression storage technology is used. For example, the Douglas-Peucker algorithm is used to compress GPS points, retaining key inflection points (accuracy loss < 0.5 meters); multi-dimensional data association is formed, and peak period risks are analyzed by binding timestamps; and the deviation rate of specific plank roads in rainy and foggy weather is marked by associating environmental parameters.

[0092] Risk thermal layer generation: Count the number of deviations in a 1km×1km grid and calculate the regional risk score: Risk score = number of monthly deviations × regional safety level Output visualization layer: red (high risk), yellow (medium risk), green (low risk).

[0093] Set up a dynamic update mechanism to refresh the heat map in real time when new deviation data is received.

[0094] In some embodiments, the behavior learning system copies the abnormal trajectory data to a learning database and counts the number of repetitions of the same path deviation; when the number of repetitions reaches a threshold, the geographic information model is updated to include the path as a new allowed movement path.

[0095] In the above, the behavioral learning system has the function of automatically correcting path rules based on group behavior data. Specifically, it performs statistical analysis of high-frequency deviations: extracting similar path deviation data from the abnormal record library; setting dynamic thresholds (default 100 times / month) to trigger the path optimization process.

[0096] The rules for upgrading the legitimacy of a path include: Condition 1: The deviation from the path meets safety standards (slope <25°, no geological disasters); Condition 2: The deviation exceeds the threshold for three consecutive months with no accident records. If the conditions are met, the path will be automatically added as a temporarily permitted path and injected into the path mapping table.

[0097] Temporal and spatial constraint binding includes: Temporary path setting validity period: Lawn shortcut: only activated during the peak season (May-October) Creekside water point: open every day from 9:00 to 17:00 during the rainy season (June-August).

[0098] The present invention also provides a method for monitoring and early warning of personnel in a scenic area, comprising the following steps: The real-time location of personnel is obtained through a multi-source perception layer, and the real-time monitoring system compares the position of the personnel to be measured with the allowed movement path of the current warning area; and performs trajectory compliance analysis: projecting the personnel coordinates onto a three-dimensional geographic model; comparing with the current warning area path rule library; when a path deviation is detected, selecting an alarm level based on the deviation type.

[0099] Dynamic resynchronization is performed through the trajectory tracking system: trajectory prediction based on the motion model is initiated, the adjacent area transfer rules are continuously matched, a synchronization signal is triggered when a person enters a legal path, and subsequent positions are continuously tracked until any warning area rules are rematched.

[0100] The behavioral learning system updates the security rules of the geographic information model based on feedback from the management terminal. The behavioral learning system performs rule optimization: parsing the structured feedback instructions from the management terminal; reducing the monitoring sensitivity of the area when the feedback confirms a false alarm; starting the regional rule reconstruction engine when the feedback requires a rule update; and outputting an incrementally updated version of the geographic information model.

[0101] In the above, a multi-source perception layer is used to obtain the person's real-time location. The real-time monitoring system then compares the path and analyzes the trajectory for compliance. The multi-source perception layer in this application integrates positioning technologies such as GPS and Beidou, as well as other sensor devices, enabling real-time location information of people within the scenic area. After obtaining the person's real-time location, the real-time monitoring system projects the person's coordinates onto a three-dimensional geographic model stored in memory. This geographic model divides warning zones with different security levels based on predefined monitoring area boundaries and stores a set of permitted movement paths for each zone. Using a spatial topology algorithm, the real-time monitoring system compares the person's actual location with a path rule library for the current warning zone to determine whether the person's location is within the buffer zone of the permitted movement path (e.g., a deviation of >5 meters from the path triggers an alarm). When a path deviation is detected, the system selects an appropriate alarm level based on the type of deviation (e.g., straying from a safe path or entering a restricted area). A low-level alarm notifies patrol personnel, while a high-level alarm triggers a broadcast system for crowd evacuation. This real-time comparison of the person's location with the three-dimensional geographic model and path rule library enables dynamic compliance detection of the person's movement path, ensuring timely detection of abnormal behavior and the issuance of appropriate alarm signals.

[0102] This application implements dynamic resynchronization through a trajectory tracking system. When the real-time monitoring system detects a person's path deviation and generates an alarm signal, the trajectory tracking system responds to the alarm signal and initiates a dynamic resynchronization mechanism. The trajectory tracking system first initiates trajectory prediction based on the motion model to predict the person's movement direction, and then continuously matches the transfer rules of adjacent areas (for example, returning to the trail from a cliff area requires passing through a designated safe exit). The system continuously tracks the person's subsequent movement path and compares it with the safety rules of the warning area until the person's position re-enters the permitted path of any warning area, triggering a synchronization signal and completing synchronization. After synchronization is completed, the trajectory tracking system enters verification mode and performs N consecutive position compliance checks. If the verification pass rate is ≥95%, the synchronization is determined to be valid. This dynamic resynchronization mechanism solves the problem of missing tracking after an alarm in traditional monitoring, ensuring continuous tracking of abnormal personnel until they return to a safe path, forming a closed-loop risk management.

[0103] This application uses a behavioral learning system to update the security rules of the geographic information model based on feedback from management terminals. The behavioral learning system receives feedback instructions from scenic area management terminals in response to alarm signals and updates the security rules of the geographic information model based on these instructions. If the management terminal feedback confirms a false alarm, the behavioral learning system reduces the monitoring sensitivity of that area and adjusts sensitivity parameters (such as positioning error tolerance and path deviation threshold). If the feedback requires a rule update, the regional rule reconstruction engine is activated to parse the structured feedback instructions from the management terminal, such as security level adjustment and path correction. For example, if the feedback instruction includes a security level adjustment, the level value of the current warning area is automatically updated, and the alarm threshold of the real-time monitoring system is also updated simultaneously. If the feedback requires a path correction, the path generator is activated to create a temporary permitted path and update the path mapping table. Furthermore, the behavioral learning system copies abnormal trajectory data to a learning database and counts the number of recurrences of the same path deviation. When the number of recurrences reaches a threshold, the geographic information model is updated to include the path as a new permitted movement path. This mechanism of updating the geographic information model based on management terminal feedback and historical abnormal data enables the system to continuously adapt to changes in the actual operation of the scenic area.

[0104] In the foregoing specification, examples have been described with reference to specific exemplary embodiments. However, it will be apparent that various modifications and changes can be made to the specific examples without departing from the scope as set forth in the appended claims, and the claims are not limited to the specific examples described above.

Claims

1. A scenic area personnel distribution monitoring and early warning system, characterized in that: include: A memory storing a scenic area geographic information model, wherein the scenic area geographic information is divided into a plurality of warning areas with different security levels based on a plurality of predefined monitoring area boundaries, and storing a security level value and at least one allowed personnel movement path for each warning area; A real-time monitoring system that compares the actual position of a person with at least one permitted movement path in the current warning area during the operation of the scenic area, and generates an alarm signal and a corresponding safety level when the actual position of a person exceeds the permitted movement path; A trajectory tracking system, in response to the alarm signal, synchronizes the person's subsequent movement path by continuously tracking the person's subsequent movement path and comparing it with the safety rules of the warning area until the person's position re-enters the allowed path of any warning area; The behavior learning system receives a feedback instruction from the scenic area management terminal in response to the alarm signal, and updates the safety rules of the geographic information model according to the feedback instruction.

2. The scenic area personnel distribution monitoring and early warning system according to claim 1 is characterized in that: The behavior learning system dynamically adjusts the safety level value associated with the current warning area according to the feedback instruction; The behavior learning system includes a feedback parsing engine, which is configured to: When the feedback instruction includes a safety level adjustment, the level value associated with the current warning area is automatically updated; When the feedback is confirmed to be a false alarm, the sensitivity parameters of similar alarms in the area will be reduced; When feedback requires a path correction, the path generator is activated to create a temporary permissible path.

3. The scenic area personnel distribution monitoring and early warning system according to claim 1 is characterized in that: The boundaries of the monitoring area are determined by running a safety risk analysis algorithm on the terrain data of the scenic area through a risk modeling engine, and identifying the boundaries of high-risk areas in combination with historical accident data; The risk modeling engine is configured to perform the following operations: Input scenic area terrain point cloud data, hydrogeological reports, and historical accident database; Run the Monte Carlo security risk simulation algorithm; Combined with crowd flow dynamics model to identify high-risk area boundaries; Outputs a set of graded boundary coordinates with weight coefficients.

4. The scenic area personnel distribution monitoring and early warning system according to claim 3 is characterized in that: The geographic information model is generated by simulating the flow paths of personnel, recording the safety status of each warning area and the path conversion rules; The geographic information model is constructed by a behavior simulator: Load the scenic area building BIM model and vegetation coverage data; Inject typical tourist behavior patterns based on seasonal changes; Run multi-agent path planning simulations; Recording a state transition matrix of each area, wherein the transition matrix includes a transition relationship between a normal path / an emergency path / a restricted area; Generate a path rule base with temporal and spatial constraints.

5. The scenic area personnel distribution monitoring and early warning system according to claim 1 is characterized in that: The scenic area geographic information includes: Area rule table, which stores the security level value and path index corresponding to each warning area; A path mapping table storing a set of allowed movement paths for each warning area, wherein the path mapping table is accessed through the path index; The path mapping table is organized by index cluster and contains: The main path set contains the GPS coordinate sequence of the allowed movement paths; Emergency path set, including evacuation paths and triggering conditions; The transfer rule set records the topological relationship of legal cross-region transfers.

6. The scenic area personnel distribution monitoring and early warning system according to claim 1 is characterized in that: The trajectory tracking system enters the verification mode after completing the synchronization. If the person's position is detected to meet the warning area rules for N consecutive times, the synchronization is determined to be valid; The trajectory tracking system includes a verification controller: After synchronization is complete, it enters verification mode with adjustable duration; Perform N consecutive position compliance checks; When the pass rate of N verifications is ≥95%, the synchronization is considered valid; Activate the regional rule adaptation module to adjust the current path confidence weight.

7. The scenic area personnel distribution monitoring and early warning system according to claim 6 is characterized in that: The real-time monitoring system suspends alarming in verification mode, and if any subsequent position violates the zone rules, the trajectory tracking system is triggered to re-execute synchronization; Wherein, the real-time monitoring system performs the following operations in verification mode: Enable relaxed alert policy; If the single position deviation standard deviation is greater than the preset tolerance value, the trajectory tracking system is triggered to restart synchronization; If there are two consecutive deviations, the verification mode will be forced to exit and a regional abnormality report will be generated.

8. The scenic area personnel distribution monitoring and early warning system according to claim 1 is characterized in that: Also includes: a library of abnormal behavior records and a spatiotemporally indexed deviation trajectory dataset stored in the memory; The trajectory tracking system stores the trajectory data of the person deviating from the path into the abnormal behavior record library, wherein: The trajectory tracking system performs: Compress and store the GPS coordinate sequence that deviates from the path; Associate the corresponding timestamps and environment parameters; High-frequency deviation areas are marked to form a risk thermal layer.

9. The scenic area personnel distribution monitoring and early warning system according to claim 8, characterized in that: The behavior learning system copies the abnormal trajectory data to the learning database and counts the number of repetitions of the same path deviation; when the number of repetitions reaches a threshold, the geographic information model is updated to include the path as a new allowed movement path.

10. A method for monitoring and early warning of personnel in a scenic area, characterized in that: The steps include: The system uses a multi-source perception layer to obtain the real-time location of the person, and uses a real-time monitoring system to compare the location of the person to be measured with the allowed movement path in the current warning area. It also performs trajectory compliance analysis: projecting the person's coordinates onto a three-dimensional geographic model. Compare the current warning area path rule library; when a path deviation is detected, select the alarm level based on the deviation type Dynamic resynchronization is performed through the trajectory tracking system: trajectory prediction based on the motion model is initiated, the adjacent area transfer rules are continuously matched, a synchronization signal is triggered when the person enters a legal path, and subsequent positions are continuously tracked until any warning area rules are matched again; The behavioral learning system updates the security rules of the geographic information model based on feedback from the management terminal. The behavioral learning system performs rule optimization: parsing the structured feedback instructions from the management terminal; reducing the monitoring sensitivity of the area when the feedback confirms a false alarm; starting the regional rule reconstruction engine when the feedback requires a rule update; and outputting an incrementally updated version of the geographic information model.

Images