An airport monitoring and early warning method and system based on automation

Abstract

This application discloses an automated airport monitoring and early warning method and system, relating to the field of security monitoring technology. The method includes calculating the risk assessment value of luggage to be delivered to the airport based on a preset risk indicator set, and marking luggage that exceeds the assessment threshold as risky luggage. The preset risk indicator set includes similarity risk, and the similarity risk of luggage is determined by comparing appearance similarities, with the highest similarity being taken as the similarity risk. When risky luggage is delivered to the luggage carousel, a corresponding risk level marker is projected on the luggage surface, and a comparison of luggage difference features is pushed to associated passengers. Warning prompts are then issued based on different risk levels. Only risky luggage is issued with different risk warnings, thereby reducing energy consumption in luggage handling. This method also avoids passive, mechanical, and repetitive prompts that can cause passengers to become auditory fatigued and ignore prompts. By providing differentiated warnings to the owners of risky luggage, the rate of mis-collection of luggage is effectively reduced.

Description

Technical Field

[0001] The present application relates to the field of security monitoring technology, and in particular to an automated airport monitoring and early warning method and system. Background Art

[0002] Passengers need to check in large luggage at the airport. After checking in, the passengers and luggage will be separated during the journey until they arrive at the destination and pick it up at the luggage carousel. With the development of the aviation industry, in the current airport's checked luggage transportation safety management, although luggage labels are printed when checking in, in order to save time on luggage collection, luggage collection at the destination is usually done by passengers after subjectively identifying the luggage. Therefore, there is a risk of wrong luggage collection or accidental collection. When luggage is taken by mistake, passengers usually need to spend a lot of energy to retrieve it, which is extremely energy-consuming and time-consuming.

[0003] At present, in order to reduce the rate of wrong baggage collection, airports usually arrange manual reminders at the baggage carousel to remind passengers to carefully check the luggage tags before picking up the luggage. This method consumes manpower and is prone to omissions when there is a large flow of people. Airports also usually remind passengers through broadcasting, but this repetitive and passive warning method can easily cause passengers to become auditory fatigue and ignore the prompts. Both warning methods are difficult to truly and effectively reduce the rate of wrong baggage collection.

[0004] Therefore, there is an urgent need for an automated airport monitoring and early warning method and system. By establishing a risk indicator set and taking the similarity of luggage appearance as an important indicator, the risk assessment of the luggage to be delivered is carried out and the luggage with the risk of being mis-picked is automatically marked. When passengers pick up their luggage, different warning prompts are given for luggage of different risk levels, effectively reducing the rate of mis-picking of luggage. Summary of the Invention

[0005] This application addresses the problem that the current airport method of arranging manual or broadcast reminders at the baggage carousel to remind passengers to carefully check the baggage tags before picking up the bag is difficult to truly and effectively reduce the rate of wrong baggage collection. It provides an automated airport monitoring and early warning method and system. By establishing a risk indicator set, using the similarity of baggage appearance as an important indicator, the risk assessment of the baggage to be dropped is performed, and the baggage with the risk of wrong baggage is automatically marked. When the passenger picks up the baggage, different warnings are issued for baggage of different risk levels, effectively reducing the rate of wrong baggage collection. The specific technical solution is as follows:

[0006] In a first aspect of the present application, an automated airport monitoring and early warning method is provided, comprising:

[0007] Based on the preset risk indicator set, the risk assessment value of the baggage to be delivered to the airport is calculated, and baggage with a risk assessment value greater than the assessment threshold is marked as risky baggage. Baggage with a higher assessment value has a higher risk level, which is divided into high-risk baggage, medium-risk baggage and low-risk baggage.

[0008] The preset risk indicator set includes similarity risk. The similarity risk of the baggage is determined by comparing its appearance with that of all other bags placed in the same baggage carousel area during the same time period, and the highest similarity is taken as the similarity risk.

[0009] When risky luggage is placed on the luggage carousel, the corresponding risk level logo is projected on the surface of the luggage through a projection device, and the luggage difference feature comparison is pushed to the mobile terminal of the associated passenger. When the risky luggage is picked up, different degrees of early warning prompts are given according to different risk levels.

[0010] Furthermore, before taking the highest similarity as the similarity risk of the luggage, the method further includes:

[0011] Establish a baggage appearance database to store and collect image data of historical baggage at the airport, classify historical baggage based on appearance features including size, color, and texture, and count the number of different classifications;

[0012] Classifying the appearance features of the current luggage based on the image data of the current luggage, and calculating a rarity value of the current luggage based on a ratio of the number of categories to which the luggage belongs to and the maximum number of categories in the current classification, wherein the smaller the ratio, the greater the rarity value;

[0013] The rarity value of the luggage is used as a first weight, and the first weight is assigned to the similarities of all other luggage.

[0014] Furthermore, after the risk luggage is placed on the luggage carousel, the method further includes obtaining the position of the risk luggage on the luggage carousel, deriving a second weight based on the distance between the risk luggage, wherein the closer the distance, the smaller the second weight; and assigning the second weight to the similarity.

[0015] Furthermore, the preset risk indicator set also includes human flow risk. The human flow risk is determined based on the real-time update of the human flow in the area where the luggage is located by a multi-source data fusion device after the luggage enters the luggage carousel. The higher the human flow, the higher the human flow risk. The multi-source data fusion device includes a video surveillance camera array, an infrared thermal imaging sensor, and a geomagnetic sensor.

[0016] Furthermore, a spatiotemporal prediction model is constructed based on the LSTM neural network for the flow of people in the area where the luggage is located, integrating historical passenger flow data, flight schedules, and mobile device signaling data to predict the distribution trend of people flow in the future.

[0017] When the forecast shows that the luggage area is about to exceed the standard, the early warning mechanism is activated, including activating the spare luggage carousel, issuing diversion suggestions through the display screen, pushing the optimal route navigation to passengers, setting up temporary diversion isolation belts and playing temporary navigation evacuation prompts.

[0018] Furthermore, after the luggage enters the luggage carousel, the luggage location is updated in real time, and the luggage carousel rotation speed is obtained at the same time. According to the current distance between the passenger and the luggage and the walking speed, the time required for different paths to pick up the luggage is calculated, and the path with the shortest time is pushed to the passenger as the optimal path.

[0019] Furthermore, before determining the optimal path, a security verification stage is also included. Specifically, the multi-source data fusion device is used to update the passenger positions in the baggage carousel area in real time. If there are other passengers staying on the optimal path, the positions of other passengers are regarded as inaccessible locations, and the baggage collection path is regenerated and the time taken is calculated until there are no other passengers on the final optimal path.

[0020] Furthermore, it also includes:

[0021] For passengers claiming risky luggage, a Bluetooth beacon network and multi-camera trajectory tracking algorithm are used to generate a heat map of their movement paths until all luggage in the same risky luggage group has been retrieved.

[0022] If there is a luggage retrieval error, or if not all luggage in the same risk luggage group has been collected, but the passenger who collected the risk luggage is about to leave the airport area, staff will go and confirm whether the luggage has been collected correctly based on the passenger's predicted trajectory.

[0023] Furthermore, it also includes real-time updating of the passenger positions in the baggage claim area. When the number of passengers decreases or increases to a preset value within a predetermined time, an alert is given to the staff, and the real-time surveillance video of the corresponding position is switched. Corresponding preset measures are taken according to the type of emergency, which includes falling objects from high altitudes, floor collapse, luggage or facility fire, as well as conflicts and injuries.

[0024] In a second aspect of the present application, an automated airport monitoring and early warning system is provided, comprising:

[0025] The risk assessment module calculates the risk assessment value of baggage to be delivered to the airport based on a preset risk indicator set and marks baggage with a risk assessment value greater than the assessment threshold as risky baggage. Baggage with a higher assessment value has a higher risk level, which is divided into high-risk baggage, medium-risk baggage and low-risk baggage.

[0026] An appearance comparison module, wherein the preset risk indicator set includes a similarity risk. The similarity risk of the baggage is determined by comparing its appearance with that of all other bags placed in the same baggage carousel area during the same time period, and the highest similarity is taken as the similarity risk;

[0027] The retrieval warning module projects the corresponding risk level logo on the surface of the luggage through a projection device when the risky luggage is placed on the luggage carousel, pushes the luggage difference feature comparison to the mobile terminal of the associated passenger, and issues different degrees of warning prompts according to different risk levels when the risky luggage is retrieved.

[0028] Furthermore, the appearance comparison module further includes a first weighting submodule before taking the highest similarity as the similarity risk of the luggage:

[0029] Establish a baggage appearance database to store and collect image data of historical baggage at the airport, classify historical baggage based on appearance features including size, color, and texture, and count the number of different classifications;

[0030] Classifying the appearance features of the current luggage based on the image data of the current luggage, and calculating a rarity value of the current luggage based on a ratio of the number of categories to which the luggage belongs to and the maximum number of categories in the current classification, wherein the smaller the ratio, the greater the rarity value;

[0031] The rarity value of the luggage is used as a first weight, and the first weight is assigned to the similarities of all other luggage.

[0032] Furthermore, a second weight submodule is included. After the risk luggage is placed on the luggage carousel, the second weight submodule is further included. The second weight submodule is used to obtain the position of the risk luggage on the luggage carousel, and to derive a second weight based on the distance between the risk luggage. The closer the distance, the smaller the second weight. The second weight is also assigned to the similarity.

[0033] Furthermore, it also includes a traffic risk module, and the preset risk indicator set also includes human traffic risk. The human traffic risk is determined based on the real-time update of the human traffic in the area where the luggage is located by the multi-source data fusion device after the luggage enters the luggage carousel. The higher the human traffic, the higher the human traffic risk. The multi-source data fusion device includes a video surveillance camera array, an infrared thermal imaging sensor and a geomagnetic sensor.

[0034] Furthermore, the traffic risk module also includes a traffic prediction submodule, which constructs a spatiotemporal prediction model for the flow of people in the area where the luggage is located based on an LSTM neural network, integrating historical passenger flow data, flight schedules, and mobile device signaling data to predict the distribution trend of people flow in the future;

[0035] When the forecast shows that the luggage area is about to exceed the standard, the early warning mechanism is activated, including activating the spare luggage carousel, issuing diversion suggestions through the display screen, pushing the optimal route navigation to passengers, setting up temporary diversion isolation belts and playing temporary navigation evacuation prompts.

[0036] Furthermore, the traffic risk module also includes a path push submodule. After the luggage enters the luggage carousel, the luggage location is updated in real time, and the luggage carousel rotation speed is obtained. Based on the current distance between the passenger and the luggage and the walking speed, the time required for different paths to pick up the luggage is calculated. The path with the shortest time is pushed to the passenger as the optimal path.

[0037] Furthermore, the path push submodule also includes a security verification submodule. Before the optimal path is determined, a security verification stage is also included. Specifically, the passenger positions in the baggage carousel area are updated in real time through the multi-source data fusion device. If there are other passengers staying on the optimal path, the positions of other passengers are regarded as inaccessible locations, and the baggage collection path is regenerated and the time taken is calculated until there are no other passengers on the final optimal path.

[0038] Furthermore, it also includes a baggage tracking module:

[0039] For passengers claiming risky luggage, a Bluetooth beacon network and multi-camera trajectory tracking algorithm are used to generate a heat map of their movement paths until all luggage in the same risky luggage group has been retrieved.

[0040] If there is a luggage retrieval error, or if not all luggage in the same risk luggage group has been collected, but the passenger who collected the risk luggage is about to leave the airport area, staff will go and confirm whether the luggage has been collected correctly based on the passenger's predicted trajectory.

[0041] Furthermore, it also includes an emergency module, which updates the passenger positions in the baggage claim area in real time. When the number of passengers decreases or increases to a preset value within a predetermined time, it alerts the staff and switches to the real-time surveillance video of the corresponding location. According to the type of emergency, corresponding preset measures are taken. The types of emergencies include falling objects from high altitudes, floor collapse, luggage or facility fires, as well as conflicts and injuries.

[0042] This application has the following beneficial effects:

[0043] 1. By building a dynamically optimized risk assessment model and a multi-dimensional early warning system, we achieve precise control of the risk of mishandling luggage throughout the entire process. Based on a preset risk indicator set and an adaptive assessment threshold setting based on the airport's scale, the system can dynamically adjust risk assessment criteria based on the passenger flow characteristics of different airports. For example, at large hub airports, stricter assessment thresholds are adopted, such as a threshold exceeding 55 points, which indicates risky luggage. Combined with a normalized weight distribution algorithm trained on historical mishandling data, multiple factors such as appearance similarity are integrated into the calculation. By collecting luggage appearance feature data during baggage check-in and completing a similarity comparison of multiple target luggage before dropping off luggage at the destination, an improved Siamese neural network algorithm is used to extract feature vectors such as luggage texture, brand logos, and sticker patterns. When risky luggage enters the carousel, an intelligent positioning system with integrated laser projection uses indoor positioning technology to track the luggage's movement in real time, projecting a dynamic warning halo and simultaneously sending a comparison map of differences to the passenger's mobile device. For example, a message like "Your black suitcase is 92% similar to an adjacent black suitcase. Please note the red side strap." This system incorporates a differentiated warning strategy, projecting different light patterns for different risk levels (white for low-risk, yellow for medium-risk, and red for high-risk), along with projecting a corresponding font within the halo. Low-risk luggage also triggers a vibration alert in the app, while medium-risk luggage uses audio and visual alerts in the carousel area. High-risk luggage also uses biometric verification at the security checkpoint for secondary verification, effectively reducing the incidence of misplaced luggage. This technical solution, leveraging a three-tiered architecture encompassing the IoT perception layer, edge computing layer, and cloud platform decision-making layer, eliminates the need to issue risk alerts for every piece of luggage, reducing baggage handling energy consumption and avoiding the risk of passengers ignoring passive, repetitive notifications that can lead to auditory fatigue. By providing differentiated warnings to the owners of risky luggage, the system effectively reduces the incidence of misplaced luggage, providing a scalable risk management paradigm for smart airports.

[0044] 2. By introducing a weighted allocation mechanism for luggage appearance rarity, quantitative modeling and precise intervention are achieved to address passengers' mistaking behavior. A luggage appearance feature classification system, built based on historical data, encompasses multiple dimensions, including size, color, and texture. A clustering algorithm is used to deconstruct the features of historical checked baggage data, forming a benchmark database containing multiple representative categories. A real-time update mechanism ensures that the classification model is synchronized with luggage trends. When processing a current piece of luggage, the system uses an object detection model to extract various appearance features and determine the category to which each feature belongs. For example, if the current piece of luggage is pink, it is classified as pink. The system then determines the ratio of the number of pieces in the pink category to the most numerous category. The lower the ratio, the higher the rarity. For example, if the pink category has 2,000 pieces, the black category has 10,000 pieces, and the most numerous category is silver, with 20,000 pieces. Ultimately, pink luggage is more rare than black, which is higher than silver. By cumulatively calculating multiple features, the primary weight of regular luggage is reduced to below 1, reducing the similarity risk when calculating similarity and the corresponding risk assessment value. However, for uniquely rare luggage, the primary weight is greater than 1, increasing the similarity risk when calculating similarity and the corresponding risk assessment value, potentially leading to a decrease or increase in the risk level of the risky luggage. By integrating the rarity weight with the appearance similarity matrix, the system successfully captures the "rare luggage mismatch" scenario in misplaced possession incidents. For example, the similarity of two purple, irregularly shaped fabric backpacks is amplified from 87% to 112%, moving them from low-risk to high-risk.

[0045] 3. By introducing a weighting mechanism based on the distance between risky luggage, a quantitative modeling and precise intervention are implemented to mitigate the risk of passengers mistakenly picking up their luggage. Specifically, when risky luggage is close together on the baggage carousel, for example, two risky luggage pieces are placed side by side, passengers are more likely to find luggage similar to their own when they collect their luggage. This makes it easier for them to check the luggage information when retrieving their luggage to confirm whether they have taken the wrong luggage. However, when risky luggage is farther apart, for example, two risky luggage pieces are placed on opposite sides of the baggage carousel, passengers are less likely to find luggage similar to their own when they collect their luggage. This makes it relatively difficult for them to check the luggage information when retrieving their luggage, potentially leading to mistaken luggage. Therefore, this embodiment quantifies the second weight based on the distance between risky luggage pieces. When the distance between the risky luggage pieces is relatively close, the second weight of the risky luggage piece is reduced to below 1, reducing the similarity risk when calculating the similarity and correspondingly lowering the risk assessment value. When the distance between the risky luggage pieces is relatively far, the second weight of the risky luggage piece is greater than 1, increasing the similarity risk when calculating the similarity and correspondingly increasing the risk assessment value, which may result in a decrease or increase in the risk level of the risky luggage piece. By integrating distance weights with appearance similarity into a matrix, we successfully captured the scenario of "no similar boxes were found to be mismatched" in the incident of mistaking. For example, the similarity of two risky bags separated on both sides of the baggage carousel was amplified from 85% to 95% through weighting, upgrading the baggage from low-risk to medium-risk.

[0046] 4. By building a multimodal dynamic passenger flow risk assessment system and an intelligent early warning linkage mechanism, we achieve spatiotemporal coordinated prevention and control of the risk of mishandling luggage. Using multi-source data fusion technology, the system updates the passenger flow density index of the baggage carousel area every second. Based on a neural network architecture, it integrates historical passenger flow fluctuation data with real-time mobile signaling features to construct a spatiotemporal prediction model. When the passenger flow in a certain area is predicted to exceed a threshold, for example, more than 1.2 people per square meter, the early warning mechanism is activated, including guiding the activation of backup carousels, using LED curved screens to publish dynamic diversion routes, and pushing AR real-time navigation instructions to passengers' mobile phones. Simultaneously, the intelligent diversion system creates isolation zones using guardrails and uses directional sound field devices to play evacuation instructions, providing a comprehensive risk management solution for smart airport operations.

[0047] 5. By integrating real-time dynamic path planning and multimodal sensing technology, precise spatiotemporal coordinated control of the baggage retrieval process is achieved. Based on the synchronization of the positioning system and the baggage carousel encoder data, a three-dimensional spatiotemporal model of the baggage movement trajectory is constructed. Combined with the real-time position of the passengers, a path planning algorithm is used to integrate the carousel tangential speed, the current walking speed of the passengers or the average walking speed of the passengers and the distribution of obstacles. For different passengers, such as elderly people with slower walking speeds and adults in a hurry, multiple paths to the baggage carousel where they can directly pick up their luggage are calculated based on the current walking speed of the passengers. For passengers who are currently stopping for a rest, multiple paths to the baggage carousel where they can directly pick up their luggage are generated based on the average walking speed of the passengers. For example Figure 3In the example, passenger 1 is located on the right side of the baggage carousel, and their baggage is currently located on the left side of the carousel, that is, point A. Based on their current walking speed, the passenger will pick up their baggage along path 1 at point B; along path 2, they will pick up their baggage at point C; and along path 3, they will pick up their baggage at point D. The time taken for all paths is calculated, and path 1, which takes the shortest time, is selected as the optimal path. Furthermore, when the passenger is detected to be more than 15 meters away from the target baggage, the system uses a reinforcement learning model to dynamically evaluate the efficiency of multiple possible paths, including detouring around the isolation zone and using an empty carousel area. Based on a time window prediction algorithm and the real-time position updates of the baggage driven by the baggage carousel, the system prioritizes the shortest path, such as "walk 20 meters along channel C3 toward the 11 o'clock position of baggage carousel 1, and expect to pick up the baggage at the 11 o'clock position on baggage carousel 1 in 38 seconds." This shortens the average time it takes for passengers to retrieve their baggage, thereby reducing congestion at the baggage carousel and the rate of misplaced baggage. The system also uses a dynamic obstacle map constructed from LiDAR point cloud data to update its path feasibility assessment every 200ms. When it detects sudden congestion (using heat maps to identify areas with a density exceeding 1.8 people per square meter), it automatically switches to an alternative path and triggers AR navigation guidance. This technical solution, through a parallel optimization algorithm deployed on edge computing nodes, ensures the effectiveness of recommended routes, reducing average baggage claim time for passengers and improving baggage carousel turnover efficiency, providing a real-time dynamic guidance solution for smart airport operations.

[0048] 6. Introduce a safety verification mechanism. When there are other passengers on the optimal path, especially those waiting to pick up their luggage at the baggage carousel, that is, when there are other passengers at the end of the optimal path, they will block the passengers from picking up their luggage. Forcibly taking the luggage may cause accidents, and temporary changes may cause the luggage to be missed. For example Figure 4 In the example, the original optimal path for passenger 1 was Path 1. However, the end point of Path 1, Point B, and nearby passengers 2-5 blocked Passenger 1 from picking up their luggage. Therefore, Path 2 was changed to pick up their luggage at Point C, which became the optimal path. By further optimizing the path calculation method, the path was not blocked by other passengers, which reduced the time wasted and the risk of accidents.

[0049] 7. By building a multimodal perception network and a dynamic risk management closed-loop mechanism, we achieve comprehensive intelligent control of the risk of mishandled luggage. Only passengers with risky luggage are subject to heterogeneous data fusion based on a Bluetooth beacon network and an intelligent camera array. This generates a heat map of passenger movement trajectories and enables targeted trajectory tracking. If mishandled luggage is subsequently discovered, it can be quickly retrieved based on trajectory tracking. When all luggage within the same risky luggage group has been retrieved, tracking of the previous passenger who previously retrieved the group's risky luggage is canceled, avoiding wasted computing resources. Furthermore, if there is still luggage to be retrieved within the same risky luggage group and the passenger who previously retrieved the risky luggage is about to leave the airport area, staff are notified to confirm whether the luggage has been mishandled. This avoids significant increases in patrol time and costs caused by the mishandled luggage leaving the airport.

[0050] 8. By building a multimodal perception network and intelligent emergency response system, full-dimensional dynamic prevention and control of security risks in the baggage claim area is achieved. Based on heterogeneous data fusion from a positioning system and a multi-camera array, the system generates real-time passenger density heat maps. Combining target detection algorithms with optical flow motion trajectory prediction models, the system detects abnormal fluctuations in passenger numbers within a predetermined timeframe. When a sudden drop in passenger flow (such as a floor collapse) or a gathering (such as a conflict zone) is detected, the system triggers an early warning mechanism through edge computing nodes. This involves automatically switching the camera's view angle to the target area through the intelligent video dispatching platform and initiating video structured analysis. The system also categorizes emergency events, such as accurately identifying a luggage fire through smoke detection. It then activates corresponding response procedures based on pre-set emergency protocols. For falling objects, an audible and visual alarm is triggered, directing passenger evacuation; for floor collapses, a laser projection warning boundary is triggered; and for conflicts, security personnel are dispatched to implement physical isolation. The system implements differentiated response strategies for different emergency types. For luggage fires, the system locates the fire point upon appearance, activates the fire sprinkler and smoke exhaust systems for initial fire control, and simultaneously issues evacuation orders through a directional sound field device. Through automated monitoring and early warning, the system prevents staff from delaying emergency detection, which could lead to further deterioration. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0052] Figure 1 A schematic diagram of the electronic device structure of the hardware operating environment involved in the embodiments of the present application.

[0053] Figure 2 It is a step flow chart and functional module diagram of an automated airport monitoring and early warning method provided in an embodiment of the present application.

[0054] Figure 3 This is a planar schematic diagram of path push provided in an embodiment of the present application.

[0055] Figure 4 This is a planar schematic diagram of an optimized path push provided in an embodiment of the present application.

[0056] Figure 5 This is a functional module diagram of an automated airport monitoring and early warning system provided in an embodiment of the present application.

[0057] Figure 6 This is a module construction diagram of an automated airport monitoring and early warning system provided in an embodiment of the present application.

[0058] Symbols in the figure: 1001 - processor, 1002 - communication bus, 1003 - user interface, 1004 - network interface, 1005 - memory. DETAILED DESCRIPTION

[0059] To make the above-mentioned purposes, features, and advantages of this application more clearly understood, the present application is further described in detail below with reference to the accompanying drawings and specific embodiments. It is apparent that the embodiments described are only a portion of the embodiments of this application, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments in this application without inventive effort are also within the scope of protection of this application.

[0060] The solution of this application is further described below with reference to the accompanying drawings.

[0061] like Figure 1 As shown, the electronic device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display and an input unit such as a keyboard. Optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a wireless fidelity (WI-FI) interface). The memory 1005 may be a high-speed random access memory (RAM) or a stable non-volatile memory (NVM), such as a disk storage device. The memory 1005 may also be a storage device independent of the processor 1001.

[0062] Those skilled in the art will understand that Figure 1 The structure shown in the figure does not constitute a limitation to the electronic device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0063] like Figure 1 As shown, the memory 1005 as a storage medium may include an operating system, a network communication module, a user interface module and a data storage module.

[0064] exist Figure 1 In the electronic device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the electronic device of the present invention can be set in the electronic device, and the electronic device calls the data storage module in the memory 1005 through the processor 1001 to store an automated airport monitoring and early warning method, and executes an automated airport monitoring and early warning system provided in an embodiment of the present application.

[0065] Based on the aforementioned hardware operating environment and system architecture, in the first aspect of this application, refer to Figure 2 As shown, an automated airport monitoring and early warning method is provided, comprising:

[0066] Based on the preset risk indicator set, the risk assessment value of the baggage to be delivered to the airport is calculated, and baggage with a risk assessment value greater than the assessment threshold is marked as risky baggage. Baggage with a higher assessment value has a higher risk level, which is divided into high-risk baggage, medium-risk baggage and low-risk baggage.

[0067] It should be noted that by establishing a preset risk indicator set, various factors that affect the baggage mis-collection rate are used as indicators to calculate the baggage risk assessment value, and the influence of various factors is normalized. Since larger airports have greater peak passenger flow and more frequent accidents, the assessment threshold can be formulated according to the airport size and needs. Baggage above the assessment threshold is considered risky baggage. At the same time, risky baggage is further refined based on the value exceeding the assessment threshold. The higher the risk assessment value, the higher the risk of mis-collection of the baggage. For example, a risk assessment value exceeding 60 is considered risky baggage, 60-75 is considered low-risk baggage, 76-90 is considered medium-risk baggage, and 91-100 is considered high-risk baggage.

[0068] The preset risk indicator set includes similarity risk. The similarity risk of the baggage is determined by comparing its appearance with that of all other bags placed in the same baggage carousel area during the same time period, and the highest similarity is taken as the similarity risk.

[0069] It should be noted that historical statistics on the causes of mishandled baggage reveal that the most common cause is similar baggage appearance. Therefore, when checking in at the departure airport, we capture images of the baggage's appearance. Before dropping the baggage off at the destination airport, we calculate the appearance similarity of all bags dropped off at the same baggage carousel area during the same time period. The similarity risk of any piece of baggage is calculated by taking the highest similarity value after calculating similarities with all other bags.

[0070] When risky luggage is placed on the luggage carousel, the corresponding risk level logo is projected on the surface of the luggage through a projection device, and the luggage difference feature comparison is pushed to the mobile terminal of the associated passenger. When the risky luggage is picked up, different degrees of early warning prompts are given according to different risk levels.

[0071] It should be noted that the projection device adjusts the illumination angle according to the location of the risky luggage. When a passenger picks up the risky luggage, they will first see a special sign projected by the projection device, reminding them to check the information after picking up the luggage. Furthermore, according to the risk level, the mobile terminal can push information warnings, light warnings, and voice warnings;

[0072] In this implementation, a dynamically optimized risk assessment model and a multi-dimensional early warning system are constructed to achieve precise, full-process control of the risk of mishandled luggage. Based on a preset risk indicator set and adaptive assessment thresholds tailored to the airport's passenger flow characteristics, the system dynamically adjusts risk assessment criteria based on the passenger flow characteristics of different airports. For example, stricter assessment thresholds are adopted at large hub airports, such as those exceeding 55 points, to identify a bag as risky. A normalized weight allocation algorithm trained on historical mishandled luggage data is used to integrate multiple factors, such as appearance similarity, into the calculation. By collecting baggage appearance feature data during check-in and performing a similarity comparison of multiple target bags before dropping them off at the destination, an improved Siamese neural network algorithm is used to extract feature vectors, including texture, brand logos, and sticker patterns. When risky luggage enters the carousel, an intelligent positioning system with integrated laser projection uses indoor positioning technology to track the luggage's movement in real time, projecting a dynamic warning halo and simultaneously sending a comparison map of differences to the passenger's mobile device. For example, a message like "Your black suitcase is 92% similar to an adjacent black suitcase. Please note the red side strap." This system incorporates a tiered warning strategy, projecting different light patterns for different risk levels (white for low-risk, yellow for medium-risk, and red for high-risk), along with projecting a corresponding font within the halo. Low-risk luggage also triggers a vibration alert in the app, while medium-risk luggage uses audio and visual alerts in the carousel area. High-risk luggage also uses biometric verification at the security checkpoint for secondary verification, effectively reducing the incidence of mis-claimed luggage. This technical solution, leveraging a three-tiered architecture encompassing the IoT perception layer, edge computing layer, and cloud platform decision-making layer, eliminates the need for risk warnings for every 10,000 pieces of luggage processed, reducing energy consumption per 10,000 pieces of luggage processed. It also avoids the risk of passengers ignoring warnings due to passive, repetitive notifications. By providing differentiated warnings to the owners of risky luggage, the system effectively reduces the rate of mis-claimed luggage, providing a new risk management paradigm for smart airports.

[0073] In one embodiment of the present application, before taking the highest similarity as the similarity risk of the luggage, the following steps are further included:

[0074] Establish a baggage appearance database to store and collect image data of historical baggage at the airport, classify historical baggage based on appearance features including size, color, and texture, and count the number of different classifications;

[0075] Classifying the appearance features of the current luggage based on the image data of the current luggage, and calculating a rarity value of the current luggage based on a ratio of the number of categories to which the luggage belongs to and the maximum number of categories in the current classification, wherein the smaller the ratio, the greater the rarity value;

[0076] The rarity value of the luggage is used as a first weight, and the first weight is assigned to the similarities of all other luggage.

[0077] It should be noted that owners of relatively rare luggage with unusual sizes and shapes are more likely to assume that no similar luggage will appear at their designated carousel area. Therefore, when retrieving luggage, if a similar luggage appears, they are more likely to pick up the wrong luggage. On the other hand, owners of regular luggage, such as a 24-inch silver plastic suitcase, are more likely to suspect that it is someone else's luggage when they pick it up, as they often see similar luggage. This makes them less likely to pick up the wrong luggage.

[0078] This implementation achieves quantitative modeling and precise intervention for passengers' mistaking luggage by introducing a weighted allocation mechanism for luggage appearance rarity. A luggage appearance classification system, built based on historical data, encompasses multiple feature dimensions, including size, color, and texture. A clustering algorithm is used to deconstruct the features of historical checked baggage data, forming a benchmark database containing multiple representative categories. A real-time update mechanism ensures that the classification model remains synchronized with luggage trends. When processing a current piece of luggage, the system uses an object detection model to extract various appearance features and determine the category to which each feature belongs. For example, if the current piece of luggage is pink, it will be classified into the pink category. The system then determines the ratio of the number of pieces in the pink category to the most numerous category. The lower the ratio, the higher the rarity. For example, if the pink category has 2,000 pieces, the black category has 10,000 pieces, and the silver category, with 20,000 pieces, is the most numerous. Consequently, the rarity of pink luggage is greater than that of black, which in turn is greater than that of silver. By cumulatively calculating multiple features, the primary weight of regular luggage is reduced to below 1, reducing the similarity risk when calculating similarity and the corresponding risk assessment value. However, for uniquely rare luggage, the primary weight is greater than 1, increasing the similarity risk when calculating similarity and the corresponding risk assessment value, potentially leading to a decrease or increase in the risk level of the risky luggage. By integrating the rarity weight with the appearance similarity matrix, the system successfully captures the "rare luggage mismatch" scenario in misplaced possession incidents. For example, the similarity of two purple, irregularly shaped fabric backpacks is amplified from 87% to 112%, moving them from low-risk to high-risk.

[0079] In one embodiment of the present application, after the risk luggage is placed on the luggage carousel, the method further includes obtaining the position of the risk luggage on the luggage carousel, deriving a second weight based on the distance between the risk luggage, wherein the closer the distance, the smaller the second weight; and assigning the second weight to the similarity.

[0080] In this embodiment, a weighting mechanism based on the distance between risky luggage is introduced to quantitatively model and precisely intervene in the risk of passengers mistakenly picking up their luggage. Specifically, when risky luggage is close together on the baggage carousel, for example, two risky luggage pieces are placed side by side, passengers are more likely to find luggage similar to their own when they collect their luggage. This makes it easier for them to check the luggage information and confirm whether they have picked up the wrong bag. However, when risky luggage is farther apart, for example, two risky luggage pieces are placed on opposite sides of the baggage carousel, passengers are less likely to find similar luggage when they collect their luggage. This makes it harder for them to check the luggage information when they collect their luggage, potentially leading to mistaken luggage. Therefore, this embodiment quantifies the second weight based on the distance between risky luggage pieces. When the distance between the risky luggage pieces is relatively close, the second weight of the risky luggage piece is reduced to below 1, reducing the similarity risk when calculating the similarity and correspondingly lowering the risk assessment value. When the distance between the risky luggage pieces is relatively far, the second weight of the risky luggage piece is greater than 1, increasing the similarity risk when calculating the similarity and correspondingly increasing the risk assessment value, which may result in a decrease or increase in the risk level of the risky luggage piece. By integrating distance weights with appearance similarity into a matrix, we successfully captured the scenario of "no similar boxes were found to be mismatched" in the incident of mistaking. For example, the similarity of two risky bags separated on both sides of the baggage carousel was amplified from 85% to 95% through weighting, upgrading the baggage from low-risk to medium-risk.

[0081] In one embodiment of the present application, the preset risk indicator set also includes human flow risk. The human flow risk is determined based on the real-time update of the human flow in the area where the luggage is located by a multi-source data fusion device after the luggage enters the luggage carousel. The higher the human flow, the higher the human flow risk. The multi-source data fusion device includes a video surveillance camera array, an infrared thermal imaging sensor, and a geomagnetic sensor.

[0082] It should be noted that, as another indicator in the preset risk indicator set, crowd flow risk is different from similarity risk. When the crowd flow is large, it is easy for people to push and shove, block the baggage, and let the passengers behind them take it in a hurry, which may lead to the wrong baggage being taken.

[0083] In one embodiment of the present application, a spatiotemporal prediction model is constructed based on an LSTM neural network for the flow of people in the area where the luggage is located, integrating historical passenger flow data, flight schedules, and mobile device signaling data to predict the distribution trend of people flow in the future.

[0084] When the forecast shows that the luggage area is about to exceed the standard, the early warning mechanism is activated, including activating the spare luggage carousel, issuing diversion suggestions through the display screen, pushing the optimal route navigation to passengers, setting up temporary diversion isolation belts and playing temporary navigation evacuation prompts.

[0085] In this implementation, a multimodal dynamic passenger flow risk assessment system and an intelligent early warning linkage mechanism are constructed to achieve spatiotemporal coordinated control of the risk of mishandled luggage. Using multi-source data fusion technology, the system updates the passenger flow density index of the baggage carousel area every second. Based on a neural network architecture, the spatiotemporal prediction model is constructed by integrating historical passenger flow fluctuation data with real-time mobile signaling features. When the passenger flow in a certain area is predicted to exceed a threshold, for example, exceeding 1.2 people per square meter, an early warning mechanism is activated, including guiding the activation of backup carousels, using LED curved screens to display dynamic diversion routes, and pushing AR real-time navigation instructions to passengers' mobile phones. Simultaneously, an intelligent diversion system is activated, using guardrails to create isolation zones and using a directional sound field device to play evacuation instructions, providing a comprehensive risk management solution for smart airport operations.

[0086] In one embodiment of the present application, see Figure 3 As shown, after the luggage enters the luggage carousel, the luggage location is updated in real time, and the luggage carousel rotation speed is obtained at the same time. According to the current distance between the passenger and the luggage and the walking speed, the time required for different paths to pick up the luggage is calculated, and the path with the shortest time is pushed to the passenger as the optimal path.

[0087] In this embodiment, by integrating real-time dynamic path planning and multimodal sensing technology, precise spatiotemporal coordinated control of the baggage retrieval process is achieved. Based on the synchronization of the positioning system and the luggage carousel encoder data, a three-dimensional spatiotemporal model of the luggage movement trajectory is constructed. Combined with the real-time position of the passengers, a path planning algorithm is used to integrate the carousel tangential speed, the passenger's current walking speed or the passenger's average walking speed and the distribution of obstacles. For different passengers, such as elderly people with slower walking speeds and adults in a hurry to travel, multiple paths to the luggage carousel where they can directly pick up their luggage are calculated based on the passenger's current walking speed. For passengers who are currently stopping for a rest, multiple paths to the luggage carousel where they can directly pick up their luggage are generated based on the passenger's average walking speed. For example Figure 3In the example, passenger 1 is located on the right side of the baggage carousel, and their baggage is currently located on the left side of the carousel, that is, point A. Based on their current walking speed, the passenger will pick up their baggage along path 1 at point B; along path 2, they will pick up their baggage at point C; and along path 3, they will pick up their baggage at point D. The time taken for all paths is calculated, and path 1, which takes the shortest time, is selected as the optimal path. Furthermore, when the passenger is detected to be more than 15 meters away from the target baggage, the system uses a reinforcement learning model to dynamically evaluate the efficiency of multiple possible paths, including detouring around the isolation zone and using an empty carousel area. Based on a time window prediction algorithm and the real-time position updates of the baggage driven by the baggage carousel, the system prioritizes the shortest path, such as "walk 20 meters along channel C3 toward the 11 o'clock position of baggage carousel 1, and expect to pick up the baggage at the 11 o'clock position on baggage carousel 1 in 38 seconds." This shortens the average time it takes for passengers to retrieve their baggage, thereby reducing congestion at the baggage carousel and the rate of misplaced baggage. The system also uses a dynamic obstacle map constructed from LiDAR point cloud data to update its path feasibility assessment every 200ms. When it detects sudden congestion (using heat maps to identify areas with a density exceeding 1.8 people per square meter), it automatically switches to an alternative path and triggers AR navigation guidance. This technical solution, through a parallel optimization algorithm deployed on edge computing nodes, ensures the effectiveness of recommended routes, reducing average baggage claim time for passengers and improving baggage carousel turnover efficiency, providing a real-time dynamic guidance solution for smart airport operations.

[0088] In one embodiment of the present application, see Figure 4 As shown, the determination of the optimal path also includes a security verification stage. Specifically, the multi-source data fusion device is used to update the passenger positions in the baggage carousel area in real time. If there are other passengers staying on the optimal path, the positions of other passengers are regarded as inaccessible locations, and the baggage collection path is regenerated and the time taken is calculated until there are no other passengers on the final optimal path.

[0089] In this implementation, a safety verification mechanism is introduced. When there are other passengers on the optimal path, mainly passengers waiting to pick up their luggage at the baggage carousel, that is, when there are other passengers at the end of the optimal path, they will block the passengers from picking up their luggage. Forcibly picking up the luggage may cause accidents, and a sudden change of direction may result in missing the luggage. For example, Figure 4 In the example, the original optimal path for passenger 1 was Path 1. However, the end point of Path 1, Point B, and nearby passengers 2-5 blocked Passenger 1 from picking up their luggage. Therefore, Path 2 was changed to pick up their luggage at Point C, which became the optimal path. By further optimizing the path calculation method, the path was not blocked by other passengers, which reduced the time wasted and the risk of accidents.

[0090] In one embodiment of the present application, it further includes:

[0091] For passengers claiming risky luggage, a Bluetooth beacon network and multi-camera trajectory tracking algorithm are used to generate a heat map of their movement paths until all luggage in the same risky luggage group has been retrieved.

[0092] If there is a luggage retrieval error, or if not all luggage in the same risk luggage group has been collected, but the passenger who collected the risk luggage is about to leave the airport area, staff will go and confirm whether the luggage has been collected correctly based on the passenger's predicted trajectory.

[0093] It should be noted that risk luggage consists of at least two similar pieces of luggage, and there may be more than one, forming a risk luggage group. There may be multiple risk luggage groups in the same baggage carousel area at the same time period. Therefore, two pieces of luggage may both be risk luggage, but they are not similar to each other and there is no risk of mis-collection. When all the luggage in the same risk luggage group has been collected and no mis-collection has been reported, it means that the luggage in the risk luggage group has been collected normally.

[0094] In this implementation, a multimodal sensing network and a dynamic closed-loop risk management mechanism are built to achieve comprehensive intelligent control of the risk of mishandled luggage. Only passengers with risky luggage are subject to heterogeneous data fusion based on a Bluetooth beacon network and an intelligent camera array. This generates a heat map of the passenger's movement trajectory and enables targeted trajectory tracking. If a mishandled bag is subsequently discovered, it can be quickly retrieved based on the trajectory tracking. When all bags within the same risky luggage group have been retrieved, tracking of the previous passenger who previously retrieved the group is canceled, avoiding wasted computing resources. Furthermore, if there are still bags remaining within the same risky luggage group and the passenger who previously retrieved the risky bag is about to leave the airport, staff are notified to verify whether the baggage has been mishandled. This avoids the significant increase in travel time and costs associated with mishandled luggage leaving the airport.

[0095] In one embodiment of the present application, the position of passengers in the baggage claim area is also updated in real time. When the number of passengers decreases or increases to a preset value within a predetermined time, an alert is given to the staff, and the real-time surveillance video of the corresponding position is switched. According to the type of emergency, corresponding preset measures are taken. The types of emergencies include falling objects from high altitudes, floor collapse, fire in luggage or facilities, as well as conflicts and injuries.

[0096] In this embodiment, by building a multimodal perception network and an intelligent emergency response system, full-dimensional dynamic prevention and control of security risks in the baggage claim area is achieved. Based on heterogeneous data fusion from a positioning system and a multi-camera array, the system generates real-time passenger density heat maps. Combining target detection algorithms with optical flow motion trajectory prediction models, the system detects abnormal fluctuations in passenger numbers within a predetermined timeframe. When a sudden drop in passenger flow (such as a floor collapse) or a gathering (such as a conflict zone) is detected, the system triggers an early warning mechanism through edge computing nodes. This involves automatically switching the camera's view angle to the target area through the intelligent video dispatching platform and initiating video structured analysis. The system also categorizes emergency events, such as accurately identifying a luggage fire through smoke detection. It then activates corresponding response procedures based on pre-set emergency protocols. For falling objects, an audible and visual alarm is triggered, directing passenger evacuation; for floor collapses, a laser projection warning boundary is triggered; and for conflicts, security personnel are dispatched to implement physical isolation. The system implements differentiated response strategies for different emergency types. For luggage fires, the system locates the fire point upon appearance, activates the fire sprinkler and smoke exhaust systems for initial fire control, and simultaneously issues evacuation orders through a directional sound field device. Through automated monitoring and early warning, the system prevents staff from delaying emergency detection, which could lead to further deterioration.

[0097] In the second aspect of this application, see Figure 5 and Figure 6 As shown, an automated airport monitoring and early warning system is provided, comprising:

[0098] The risk assessment module calculates the risk assessment value of baggage to be delivered to the airport based on a preset risk indicator set and marks baggage with a risk assessment value greater than the assessment threshold as risky baggage. Baggage with a higher assessment value has a higher risk level, which is divided into high-risk baggage, medium-risk baggage and low-risk baggage.

[0099] An appearance comparison module, wherein the preset risk indicator set includes a similarity risk. The similarity risk of the baggage is determined by comparing its appearance with that of all other bags placed in the same baggage carousel area during the same time period, and the highest similarity is taken as the similarity risk;

[0100] The retrieval warning module projects the corresponding risk level logo on the surface of the luggage through a projection device when the risky luggage is placed on the luggage carousel, pushes the luggage difference feature comparison to the mobile terminal of the associated passenger, and issues different degrees of warning prompts according to different risk levels when the risky luggage is retrieved.

[0101] In one embodiment of the present application, the appearance comparison module further includes a first weighting submodule before taking the highest similarity as the similarity risk of the luggage:

[0102] Establish a baggage appearance database to store and collect image data of historical baggage at the airport, classify historical baggage based on appearance features including size, color, and texture, and count the number of different classifications;

[0103] Classifying the appearance features of the current luggage based on the image data of the current luggage, and calculating a rarity value of the current luggage based on a ratio of the number of categories to which the luggage belongs to and the maximum number of categories in the current classification, wherein the smaller the ratio, the greater the rarity value;

[0104] The rarity value of the luggage is used as a first weight, and the first weight is assigned to the similarities of all other luggage.

[0105] In one embodiment of the present application, a second weight submodule is further included. After the risky luggage is placed on the luggage carousel, the second weight is obtained, and a second weight is derived based on the distance between the risky luggage. The closer the distance, the smaller the second weight. The second weight is then assigned to the similarity.

[0106] In one embodiment of the present application, a traffic risk module is further included, and the preset risk indicator set also includes human traffic risk. The human traffic risk is determined by real-time updates of the human traffic in the area where the luggage is located by a multi-source data fusion device after the luggage enters the luggage carousel. The higher the human traffic, the higher the human traffic risk. The multi-source data fusion device includes a video surveillance camera array, an infrared thermal imaging sensor, and a geomagnetic sensor.

[0107] In one embodiment of the present application, the traffic risk module further includes a traffic prediction submodule, which constructs a spatiotemporal prediction model for the flow of people in the area where the luggage is located based on an LSTM neural network, integrating historical passenger flow data, flight schedules, and mobile device signaling data to predict the distribution trend of people flow in the future;

[0108] When the forecast shows that the luggage area is about to exceed the standard, the early warning mechanism is activated, including activating the spare luggage carousel, issuing diversion suggestions through the display screen, pushing the optimal route navigation to passengers, setting up temporary diversion isolation belts and playing temporary navigation evacuation prompts.

[0109] In one embodiment of the present application, the traffic risk module also includes a path push submodule. After the luggage enters the luggage carousel, the location of the luggage is updated in real time, and the rotation speed of the luggage carousel is obtained. Based on the current distance between the passenger and the luggage and the walking speed, the time required for different paths to pick up the luggage is calculated. The path with the shortest time is pushed to the passenger as the optimal path.

[0110] In one embodiment of the present application, the path push submodule also includes a security verification submodule. Before the optimal path is determined, a security verification stage is also included. Specifically, the passenger positions in the baggage carousel area are updated in real time through the multi-source data fusion device. If there are other passengers staying on the optimal path, the positions of other passengers are regarded as inaccessible locations, and the baggage retrieval path is regenerated and the time taken is calculated until there are no other passengers on the final optimal path.

[0111] In one embodiment of the present application, a baggage tracking module is also included:

[0112] For passengers claiming risky luggage, a Bluetooth beacon network and multi-camera trajectory tracking algorithm are used to generate a heat map of their movement paths until all luggage in the same risky luggage group has been retrieved.

[0113] If there is a luggage retrieval error, or if not all luggage in the same risk luggage group has been collected, but the passenger who collected the risk luggage is about to leave the airport area, staff will go and confirm whether the luggage has been collected correctly based on the passenger's predicted trajectory.

[0114] In one embodiment of the present application, an emergency module is also included to update the passenger positions in the baggage claim area in real time. When the number of passengers decreases or increases to a preset value within a predetermined time, an alert is given to the staff, and the real-time surveillance video of the corresponding location is switched. According to the type of emergency, corresponding preset measures are taken. The types of emergencies include falling objects from high altitudes, floor collapse, luggage or facility fires, as well as conflicts and injuries.

[0115] It should be noted that the specific implementation of an automated airport monitoring and early warning system in an embodiment of the present application refers to the specific implementation of an automated airport monitoring and early warning method proposed in the first aspect of the aforementioned embodiment of the present application, and will not be repeated here.

[0116] It should be noted that, in this document, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that an article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such article or device. In the absence of further limitations, the elements defined by the phrase "includes..." do not exclude the presence of other identical elements in the article or device comprising the elements.

[0117] The above is a detailed introduction to the provided automated airport monitoring and early warning method. This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the automated airport monitoring and early warning method and its core idea of ​​this application; at the same time, for general technical personnel in this field, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on this application.

Claims

1. An automated airport monitoring and early warning method, characterized in that: include: Calculate the risk assessment value of the baggage to be delivered to the airport based on the preset risk indicator set, and mark the baggage with a risk assessment value greater than the assessment threshold as risky baggage; The higher the assessment value, the higher the risk level of the risk baggage, including high-risk baggage, medium-risk baggage and low-risk baggage; The preset risk indicator set includes similarity risk. The similarity risk of the baggage is determined by comparing its appearance with that of all other bags placed in the same baggage carousel area during the same time period, and the highest similarity is taken as the similarity risk. When risky luggage is placed on the baggage carousel, a projection device projects a risk level marker onto the luggage surface, pushes the luggage's differential features to the associated passenger's mobile terminal for comparison, and issues warnings of varying degrees based on the risk level when the risky luggage is picked up. Before taking the highest similarity as the similarity risk, the baggage also includes: Establish a baggage appearance database to store and collect image data of historical baggage at the airport, classify historical baggage based on appearance features including size, color, and texture, and count the number of different classifications; Classifying the appearance features of the current luggage based on the image data of the current luggage, and calculating a rarity value of the current luggage based on a ratio of the number of categories to which the luggage belongs to and the maximum number of categories in the current classification, wherein the smaller the ratio, the greater the rarity value; The rarity value of the luggage is used as a first weight, and the first weight is assigned to the similarities of all other luggage.

2. The automated airport monitoring and early warning method according to claim 1, characterized in that: After the risky luggage is placed on the luggage carousel, the method further includes obtaining the position of the risky luggage on the luggage carousel, deriving a second weight based on the distance between the risky luggage, wherein the closer the distance, the smaller the second weight; and assigning the second weight to the similarity.

3. The automated airport monitoring and early warning method according to claim 1, characterized in that: The preset risk indicator set also includes human flow risk. The human flow risk is determined based on the real-time update of the human flow in the area where the luggage is located by a multi-source data fusion device after the luggage enters the luggage carousel. The higher the human flow, the higher the human flow risk. The multi-source data fusion device includes a video surveillance camera array, an infrared thermal imaging sensor, and a geomagnetic sensor.

4. The automated airport monitoring and early warning method according to claim 3 is characterized in that: The flow of people in the area where the luggage is located is predicted based on a spatiotemporal prediction model built on an LSTM neural network. This model integrates historical passenger flow data, flight schedules, and mobile device signaling data to predict the distribution trend of people flow in the future. When the forecast shows that the luggage area is about to exceed the standard, the early warning mechanism is activated, including activating the spare luggage carousel, issuing diversion suggestions through the display screen, pushing the optimal route navigation to passengers, setting up temporary diversion isolation belts and playing temporary navigation evacuation prompts.

5. The automated airport monitoring and early warning method according to claim 4, characterized in that: After the luggage enters the luggage carousel, the luggage location is updated in real time, and the luggage carousel rotation speed is obtained at the same time. Based on the current distance between the passenger and the luggage and the walking speed, the time required for different paths to pick up the luggage is calculated. The path with the shortest time is pushed to the passenger as the optimal path.

6. The automated airport monitoring and early warning method according to claim 5, characterized in that: Before determining the optimal path, a security verification stage is also included. Specifically, the multi-source data fusion device is used to update the passenger positions in the baggage carousel area in real time. If other passengers are staying on the optimal path, the locations of other passengers are marked as inaccessible locations, and the baggage retrieval path is regenerated and the time required is calculated until there are no other passengers on the final optimal path.

7. The automated airport monitoring and early warning method according to any one of claims 1 to 6, characterized in that: Also includes: For passengers claiming risky luggage, a Bluetooth beacon network and multi-camera trajectory tracking algorithm are used to generate a heat map of their movement paths until all luggage in the same risky luggage group has been retrieved. If there is a luggage retrieval error, or if not all luggage in the same risk luggage group has been collected, but the passenger who collected the risk luggage is about to leave the airport area, staff will go and confirm whether the luggage has been collected correctly based on the passenger's predicted trajectory.

8. The automated airport monitoring and early warning method according to any one of claims 1 to 6, characterized in that: It also includes real-time updates of the passenger positions in the baggage claim area. When the number of passengers decreases or increases to a preset value within a scheduled time, the staff will be warned and the real-time surveillance video of the corresponding location will be switched. Corresponding preset measures will be taken according to the type of emergency, which includes falling objects from high altitudes, floor collapse, luggage or facility fires, as well as conflicts and injuries.

9. An automated airport monitoring and early warning system, characterized in that: include: The risk assessment module calculates the risk assessment value of the baggage to be delivered to the airport based on a preset risk indicator set, and marks the baggage with a risk assessment value greater than the assessment threshold as risky baggage; The higher the assessment value, the higher the risk level of the risk baggage, including high-risk baggage, medium-risk baggage and low-risk baggage; An appearance comparison module, wherein the preset risk indicator set includes a similarity risk. The similarity risk of the baggage is determined by comparing its appearance with that of all other bags placed in the same baggage carousel area during the same time period, and the highest similarity is taken as the similarity risk; The retrieval warning module projects a risk level indicator onto the luggage surface when risky luggage is placed on the luggage carousel. This indicator is then sent to the passenger's mobile device for comparison. Different levels of risk warnings are then issued when the risky luggage is retrieved. The appearance comparison module further includes a first weight submodule before taking the highest similarity as the similarity risk of the luggage: Establish a baggage appearance database to store and collect image data of historical baggage at the airport, classify historical baggage based on appearance features including size, color, and texture, and count the number of different classifications; Classifying the appearance features of the current luggage based on the image data of the current luggage, and calculating a rarity value of the current luggage based on a ratio of the number of categories to which the luggage belongs to and the maximum number of categories in the current classification, wherein the smaller the ratio, the greater the rarity value; The rarity value of the luggage is used as a first weight, and the first weight is assigned to the similarities of all other luggage.

Images