Big data driven intelligent traffic monitoring system

Abstract

The invention discloses an intelligent traffic monitoring system driven by big data, and an operation method of the system comprises the following steps: 1, after passing identity verification, binding passenger identity information with face information; 2, deep article information binding is carried out, and passengers and articles are tracked in real time; step 3, realizing object abnormal condition marking through distance calculation; 4, the guarding condition of the carried article is judged through motion analysis and machine learning, an identity verification and association establishment module is used for verifying the identity of the passenger, binding face information with identity card information and preliminarily establishing the association between the passenger and the carried article, and a real-time tracking and article list updating module is used for updating the real-time tracking and article list. The passenger identification module is used for identifying passengers and articles in real time, extracting detailed information, updating a list of carried articles, monitoring time association and providing accurate tracking and monitoring services; the method has the advantages of being accurate in monitoring and rapid in abnormal condition response.

Description

Technical Field

[0001] The present invention relates to the technical field of computer vision, and particularly to a big data-driven intelligent transportation monitoring system. Background Art

[0002] In modern urban transportation, as a place with concentrated passenger flow, the problem of passengers losing their items is becoming increasingly prominent, bringing certain troubles to the travel safety and experience of passengers. The current station monitoring system mainly relies on traditional camera technology, but it has certain limitations in real-time tracking and item association.

[0003] One of the main problems in current stations is that after a passenger loses an item, they often board the vehicle by the time they notice, which leads to the item not being retrieved in time, causing certain economic and psychological losses. The limited capabilities of traditional monitoring systems make it difficult to accurately track passengers and items, especially in the case of high passenger flow in stations, where the monitoring system is difficult to provide sufficient refined information.

[0004] In addition, traditional systems often cannot distinguish different types of carried items, nor can they easily determine the association between the owner of the item and it, restricting the intelligence and practicality of the monitoring system. Therefore, it is necessary to design a big data-driven intelligent transportation monitoring system with accurate monitoring and rapid response to abnormal situations. Summary of the Invention

[0005] The purpose of the present invention is to provide a big data-driven intelligent transportation monitoring system. All user data collected in this application is collected with the consent and authorization of the users, and the collection, use, and processing of relevant users need to comply with the laws, regulations, and standards of the relevant regions to solve the problems raised in the above background art.

[0006] To solve the above technical problems, the present invention provides the following technical solution: A big data-driven intelligent transportation monitoring system, and the operation method of the system includes the following steps:

[0007] Step 1: After passing the identity verification, bind the passenger identity information with the face information;

[0008] Step 2: Perform in-depth item information binding and real-time track passengers and items;

[0009] Step 3: Mark abnormal item situations through distance calculation;

[0010] Step 4: Determine the custody situation of the carried item through motion analysis and machine learning.

[0011] According to the above technical solution, the step of binding the passenger identity information with the face information after passing the identity verification includes:

[0012] Use deep learning algorithms to extract passengers' facial features and ID card information;

[0013] Establish a preliminary association between passengers and items using feature point matching;

[0014] Provide a reliable basis for subsequent monitoring and form a preliminary list of carried items.

[0015] According to the above technical solution, the step of using deep learning algorithms to extract passengers' facial features and ID card information includes:

[0016] Deploy a high-resolution camera device directly in front of the entrance gate. This camera device can capture clear frontal face images of passengers. When a passenger swipes their ID card to enter the station, the system starts the face recognition process. During the face recognition process, the camera captures the passenger's facial image, and uses deep learning algorithms to accurately extract facial features. At the same time, optical character recognition technology is used to extract the ID card information swiped in, including the name and ID number, for establishing a temporary identity information file.

[0017] According to the above technical solution, the step of establishing a preliminary association between passengers and items using feature point matching includes:

[0018] While extracting the facial image of the passenger, also extract the image features of the items carried by the passenger. First, use the classic feature point extraction algorithm SURF to obtain the feature point sets of both. Then, use the feature point matching algorithm to match the facial feature points of the passenger and the feature points of the carried item image. Through the matching of feature points, establish a preliminary association between the passenger and the carried item. On the basis of the matching, further use region of interest matching to determine the specific position of the carried item on the passenger, which helps to reduce unnecessary matching interference and improve the accuracy of the matching. Through the method based on feature point matching, the system can effectively establish a preliminary association between the passenger and the items they carry using image features after the identity verification is passed.

[0019] According to the above technical solution, the step of performing deep item information binding and real-time tracking of passengers and items includes:

[0020] Use the YOLO model to detect passengers' items;

[0021] Introduce deep learning object recognition algorithms to identify carried items;

[0022] Establish an item category mapping table and extract detailed information;

[0023] Form a comprehensive list of carried items to provide a basis for real-time tracking and anomaly detection.

[0024] According to the above technical solution, the step of introducing a deep learning object recognition algorithm to recognize carried items includes:

[0025] Furthermore, the system uses a deep learning object recognition algorithm to recognize specific objects within the item contour. The ResNet image classification model is selected, and the model is trained by constructing a labeled data set. In practical applications, the item contour images obtained from the camera are preprocessed, including size adjustment and other enhancement techniques. Subsequently, the trained image classification model is used to infer the preprocessed item contour images, and the class probability distribution of the items is output. To ensure that high-confidence item classes are accurately recognized, the system sets a probability threshold to filter low-confidence class predictions. Then, the item classes output by the model are mapped to actual item types. In addition, an item class mapping table needs to be established. According to the item classes, the system further extracts relevant detailed information to form a final list of carried items. Finally, the recognized item information is associated with the previously established identity information to ensure that each item corresponds to the identity of a specific passenger. This integration process forms a more comprehensive and accurate list of carried items.

[0026] According to the above technical solution, the step of marking abnormal item situations through distance calculation includes:

[0027] Deploy a depth camera network to calculate the relative positions of passengers and items;

[0028] Use depth information to calculate the distance and set a distance threshold.

[0029] According to the above technical solution, the step of using depth information to calculate the distance and set a distance threshold includes:

[0030] The coordinates of a pixel point in the depth map are (U, V), and the corresponding depth value is D. Through the camera internal parameter matrix K, the depth information can be converted into coordinates (X, Y, Z) in the camera coordinate system. The formula: Z = D, where K cx , K cy is the optical center coordinate of the depth camera, K fx , K fy is the focal length. Finally, using the depth information, calculate the distance from the coordinate point to Similarly, the distance between the object and the passenger can also be calculated. Subsequently, using the calculation result, monitoring and determination are carried out. A distance threshold is set in the system, and this threshold represents the maximum distance allowed between the carried item and the passenger. Subsequently, the relative positions of the carried item and the passenger are monitored in real time using a depth camera network. Moreover, the real-time distance between the monitored item and the passenger is determined. If the distance exceeds the set threshold, the system determines it as an abnormal situation. Secondly, a time threshold is set. A time threshold is set in the system to monitor in real time whether there is a disconnection phenomenon between the item and the passenger, that is, whether the passenger item identified and recorded in step two disappears. When the disappearance time is greater than the threshold, it is marked as an abnormal situation. Once the system determines that there is an abnormal situation, a corresponding safety alarm or notification is triggered, and further measures can be taken according to the preset safety strategy, such as notifying security personnel or activating surveillance cameras for real-time monitoring.

[0031] According to the above technical solution, the steps for determining the custody situation of the carried item through motion analysis and machine learning include:

[0032] First of all, through the analysis of motion postures, the system pays attention to the passenger's gesture and head movement characteristics, and measures the distance between the passenger and the carried item in real time, correlating the distance information with the motion postures. In addition, the system introduces a machine learning algorithm for modeling. The deep learning model is trained to learn normal custody situations and potential abnormal behaviors. This model is trained on motion postures and distance characteristics to adapt to various complex scenarios. At the same time, in order to enhance the robustness of the system, an error correction mechanism is implemented. The machine learning model is updated in real time through a feedback mechanism to adapt to the changing motion environment. Such real-time updates help to reduce the misjudgment rate and ensure the accuracy of the system in different scenarios. Once the system determines the custody situation, real-time decisions can be triggered. At the same time, the system can trigger an alarm mechanism to notify relevant personnel or execute other safety measures.

[0033] According to the above technical solution, the system includes:

[0034] An authentication and association establishment module, which is used to verify the passenger's identity, bind the face information with the ID card information, and initially establish the association between the passenger and the carried item;

[0035] A real-time tracking and item list update module, which is used to identify the passenger and the item in real time, extract detailed information, update the carried item list, and monitor the time association;

[0036] A custody situation determination and decision trigger module, which is used to analyze the video stream captured by the depth camera, judge whether the passenger is in custody of the item, and once the custody situation is determined, trigger a real-time decision.

[0037] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: First, in the present invention, identity verification is performed through a high-resolution camera at the entrance of the station. The identity information of the passenger is bound to the face image, and the items carried by the passenger are initially associated. Subsequently, after security inspection, a deep learning object detection algorithm is started. The items carried by the passenger are identified through the trained YOLO model, and the specific types of the items are identified by using the deep learning object recognition algorithm. Second, a depth camera network is deployed for real-time three-dimensional perception to monitor the relative positions of the passengers and the items, and distance and time thresholds are set to determine abnormal situations. Finally, the monitoring situation of the carried items is determined through intelligent motion analysis and machine learning, improving the accuracy and real-time responsiveness of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The accompanying drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention, and do not constitute a limitation to the present invention. In the drawings:

[0039] Figure 1 is a flowchart of a big data-driven intelligent traffic monitoring method provided by Embodiment 1 of the present invention;

[0040] Figure 2 is a schematic diagram of the module composition of a big data-driven intelligent traffic monitoring system provided by Embodiment 2 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0041] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0042] Embodiment 1: Figure 1 is a flowchart of a big data-driven intelligent traffic monitoring method provided by Embodiment 1 of the present invention. This embodiment can be applied to the scenario of passenger item safety monitoring. This method can be executed by the big data-driven intelligent traffic monitoring system provided by this embodiment, as Figure 1 shown. The method specifically includes the following steps:

[0043] Step 1: After identity verification, bind the passenger identity information with the face information;

[0044] In the embodiment of the present invention, a camera is installed directly in front of the station entrance gate to achieve accurate face shooting and identity verification;

[0045] Exemplarily, directly in front of the inbound gate, a high-resolution camera device is deployed. This camera device can capture clear frontal face images of passengers. When a passenger swipes their ID card to enter the station, the system starts the face recognition process. During the face recognition process, the camera collects the passenger's facial images and uses deep learning algorithms to accurately extract facial features. At the same time, optical character recognition technology is used to extract the ID card information swiped in, including the name, ID number, etc., for establishing a temporary identity information file.

[0046] Exemplarily, while extracting the facial images of passengers, image features of the items carried by passengers are also extracted. First, the classic feature point extraction algorithm SURF is used to obtain the feature point sets of both. Then, a feature point matching algorithm is used to match the facial feature points of the passenger and the feature points of the image of the carried item. Through the matching of feature points, a preliminary association between the passenger and the carried item is established. On the basis of the matching, region of interest matching is further used to determine the specific position of the carried item on the passenger, which helps to reduce unnecessary matching interference and improve the accuracy of matching. Through the method based on feature point matching, the system can effectively establish a preliminary association between the passenger and the items they carry using image features after the identity verification is passed. The establishment of this association provides a reliable basis for subsequent item tracking and monitoring.

[0047] Step two: Perform in-depth item information binding and real-time tracking of passengers and items;

[0048] In the embodiment of the present invention, after the passenger passes the security check, the system starts the deep learning object detection algorithm to identify the items carried by the passenger.

[0049] Exemplarily, a pre-trained YOLO model is deployed in the system to achieve the detection of passenger items. This algorithm can achieve the simultaneous detection of multiple objects and provide the bounding boxes and confidence scores of the objects. When the passenger items are detected, the system enters the in-depth item binding stage. In this stage, the system binds all the items of the passenger. The system uses a convolutional neural network to perform in-depth binding of the outlines of the items. Next, the system associates the in-depth binding information of all the items carried by the passenger with the passenger's identity information.

[0050] Exemplarily, further, the system utilizes a deep learning object recognition algorithm to identify specific objects within the item contour. The ResNet image classification model is selected, and the model is trained by constructing a labeled dataset. In practical applications, the item contour images obtained from the camera are preprocessed, including size adjustment and other enhancement techniques. Subsequently, the trained image classification model is used to infer the preprocessed item contour images, and the class probability distribution of the items is output. To ensure that items with high confidence are accurately identified, the system sets a probability threshold to filter out low-confidence class predictions. Then, the item classes output by the model are mapped to actual item types, such as handbags, luggage, etc. In addition, an item class mapping table needs to be established. According to the item classes, the system further extracts relevant detailed information, such as brand, color, etc., to form the final list of carried items. Finally, the identified item information is associated with the previously established identity information to ensure that each item corresponds to the identity of a specific passenger. This integrated process forms a more comprehensive and accurate list of carried items, which helps to achieve more intelligent real-time tracking and anomaly detection.

[0051] Step 3: Mark abnormal item situations through distance calculation;

[0052] In the embodiment of the present invention, a depth camera based on stereo vision is deployed in the station waiting hall, and this camera can provide accurate depth information at each pixel point;

[0053] Exemplarily, multiple depth cameras based on stereo vision are deployed in the waiting hall. These cameras are arranged in different areas to monitor various parts of the waiting hall. When a passenger enters the waiting hall and enters the monitoring range of the depth camera, these cameras generate real-time depth maps. Using the depth maps, the system adopts an object detection algorithm to real-time identify the passengers and carried items in the depth maps. Through the results of the object detection algorithm, the system can accurately identify the position of each target in the depth maps. Subsequently, the system utilizes the distance information in the depth maps. By calculating the distance from each coordinate point to the camera, the relative positions of the passengers and items in the physical space are determined. This enables the system to real-time understand the spatial relationships between each target and the surrounding environment, including the specific positions of the passengers and items in the waiting hall and the distances between them;

[0054] Exemplarily, the method for calculating the distance from each coordinate point to the camera is as follows: Assume that the coordinate of a pixel point in the depth map is (U, V), and the corresponding depth value is D. Through the camera internal parameter matrix K, the depth information can be converted into the coordinates (X, Y, Z) in the camera coordinate system. The formula: Z = D, where, K cx , K cy is the optical center coordinate of the depth camera, K fx, K fy is the focal length. Finally, using the depth information, calculate the distance from the coordinate point to Similarly, the distance between the object and the passenger can also be calculated. Subsequently, using the calculation result, carry out monitoring and determination;

[0055] Exemplarily, set a distance threshold in the system. This threshold represents the maximum distance allowed between the carried item and the passenger. Subsequently, use the depth camera network to monitor the relative positions of the carried item and the passenger in real time, and determine the real-time distance between the monitored item and the passenger. If the distance exceeds the set threshold, the system determines it as an abnormal situation. Secondly, set a time threshold. Set a time threshold in the system to monitor in real time whether there is a disconnection phenomenon between the item and the passenger, that is, whether the passenger item identified and recorded in step two disappears. When the disappearance time is greater than the threshold, it is marked as an abnormal situation. Once the system determines that there is an abnormal situation, trigger the corresponding security alarm or notification, and further measures can be taken according to the preset security strategy, such as notifying the security personnel or activating the surveillance camera for real-time monitoring.

[0056] Step Four: Determine the custody situation of the carried item through motion analysis and machine learning.

[0057] In the embodiment of the present invention, the system uses the video stream captured by the depth camera for motion posture analysis to determine whether the passenger is currently taking care of the carried item;

[0058] Exemplarily, first, through the analysis of the motion posture, the system focuses on features such as the passenger's gestures and head movements, and measures the distance between the passenger and the carried item in real time, correlating the distance information with the motion posture. In addition, the system introduces a machine learning algorithm for modeling. The deep learning model is trained to learn normal custody situations and potential abnormal behaviors. This model is trained on features such as motion posture and distance to adapt to various complex scenarios. At the same time, in order to enhance the robustness of the system, an error correction mechanism is implemented. The machine learning model is updated in real time through a feedback mechanism to adapt to the changing motion environment. Such real-time updates help reduce the misjudgment rate and ensure the accuracy of the system in different scenarios. Once the system determines the custody situation, real-time decisions can be triggered, such as updating the associated timestamp to maintain the association with the item. At the same time, the system can trigger an alarm mechanism to notify relevant personnel or execute other security measures. Through this intelligent custody situation determination mechanism, the system can more effectively adapt to the changing real-world scenarios, improving the accuracy and real-time responsiveness of the monitoring system.

[0059] Embodiment Two: Embodiment Two of the present invention provides a big data-driven intelligent transportation monitoring system. Figure 2 It is a schematic diagram of the module composition of the big data-driven intelligent transportation monitoring system provided by Embodiment Two of the present invention, as Figure 2As shown, the system includes:

[0060] An authentication and association establishment module, which is used to verify the identity of passengers, bind the face information with the ID card information, initially establish the association between passengers and the carried items, and ensure that the legally entering passengers can be accurately associated with the items they carry;

[0061] A real-time tracking and item list update module, which is used to identify passengers and items in real time, extract detailed information, update the carried item list, and monitor the time association, so as to ensure that the system accurately records the association time between the carried items and passengers, and provides accurate tracking and monitoring services;

[0062] A custody situation determination and decision trigger module, which is used to analyze the video stream captured by the depth camera, determine whether the passenger is in custody of the item, and trigger a real-time decision once the custody situation is determined;

[0063] In some embodiments of the present invention, the authentication and association establishment module includes:

[0064] An authentication and face information binding module, which is used to establish passenger identity authentication, associate the face information with the ID card information, and initially establish the association between passengers and the carried items;

[0065] A depth item information binding module, which is used to update the carried item list in real time through object detection and deep learning, and establish a deep association between passengers and items;

[0066] A distance calculation and abnormal situation marking module, which is used to calculate the relative position through the depth map, mark abnormal situations exceeding the specified distance or time, and trigger safety measures;

[0067] In some embodiments of the present invention, the real-time tracking and item list update module includes:

[0068] A deep learning object recognition module, which is used to identify passengers and items in real time, extract detailed information, and update the carried item list;

[0069] A time association monitoring module, which is used to monitor the time association between the carried items and passengers in real time, and ensure that the system accurately records the time stamp of the association;

[0070] An intelligent motion analysis and machine learning module, which is used to analyze the motion posture of passengers through the video stream, determine whether they are in custody of the item, and trigger a real-time decision and alarm mechanism;

[0071] In some embodiments of the present invention, the custody situation determination and decision trigger module includes:

[0072] A custody situation determination module, which is used to analyze the motion posture through the video stream captured by the depth camera, measure the distance, and determine whether the passenger is in custody of the item;

[0073] A real-time decision-making and alert mechanism module, which is used to trigger real-time decision-making when the system determines the supervision situation.

[0074] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device.

[0075] Finally, it should be noted that the above are only the preferred embodiments of the present invention and are not used to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A big data-driven intelligent traffic monitoring method, characterized in that, The method includes the following steps: Step 1: After authentication, bind the passenger identity information with the face information; Step 2: Perform in-depth item information binding and track passengers and items in real time; Step 3: Mark abnormal situations of items through distance calculation; Step 4: Determine the custody situation of carried items through motion analysis and machine learning; The step of binding the passenger identity information with the face information after authentication includes: Use a deep learning algorithm to extract the passenger's facial features and ID card information; Establish a preliminary association between the passenger and the item using feature point matching; Provide a reliable basis for subsequent monitoring and form a preliminary list of carried items; The step of using a deep learning algorithm to extract the passenger's facial features and ID card information includes: Deploy a high-resolution camera device directly in front of the inbound gate. This camera device captures a clear frontal face image of the passenger. When the passenger swipes the ID card to enter the station, the system starts the face recognition process. During the face recognition process, the camera collects the passenger's facial image and uses a deep learning algorithm to accurately extract the facial features. At the same time, optical character recognition technology is used to extract the ID card information swiped in, including the name and ID number, for establishing a temporary identity information file; The step of establishing a preliminary association between the passenger and the item using feature point matching includes: While extracting the facial image of the passenger, also extract the image features of the items carried by the passenger. First, use the classic feature point extraction algorithm SURF to obtain the feature point sets of both. Then, use the feature point matching algorithm to match the passenger's facial feature points with the feature points of the carried item image. Through the matching of feature points, establish a preliminary association between the passenger and the carried item. On the basis of the matching, further use region of interest matching to determine the specific position of the carried item on the passenger, which helps to reduce unnecessary matching interference and improve the accuracy of the matching. Through the method based on feature point matching, the system can effectively establish a preliminary association between the passenger and the items carried by the passenger after authentication through image features; The step of performing in-depth item information binding and tracking passengers and items in real time includes: Use the YOLO model to detect passenger items; Introduce a deep learning object recognition algorithm to identify carried items; Establish an item category mapping table and extract detailed information; Form a comprehensive list of carried items to provide a basis for real-time tracking and anomaly detection; The step of introducing a deep learning object recognition algorithm to identify carried items includes: Furthermore, the system utilizes a deep learning object recognition algorithm to identify specific objects within the item contour. The ResNet image classification model is selected, and the model is trained by constructing a labeled dataset. In practical applications, the item contour images obtained from the camera are preprocessed, including size adjustment and other enhancement techniques. Subsequently, the trained image classification model is used to infer the preprocessed item contour images, and the class probability distribution of the items is output. To ensure that items with high confidence are accurately identified, the system sets a probability threshold to filter out low-confidence class predictions. Then, the item classes output by the model are mapped to actual item types. In addition, an item class mapping table needs to be established. According to the item classes, the system further extracts relevant detailed information to form the final list of carried items. Finally, the identified item information is associated with the previously established identity information to ensure that each item corresponds to the identity of a specific passenger. This integrated process forms a more comprehensive and accurate list of carried items; The steps for determining the custody status of carried items through motion analysis and machine learning include: First, through the analysis of motion postures, the system focuses on the passenger's gestures and head movement characteristics, and measures the distance between the passenger and the carried items in real time. The distance information is correlated with the motion postures. In addition, the system introduces a machine learning algorithm for modeling. The deep learning model is trained to learn normal custody situations and potential abnormal behaviors. This model is trained on motion postures and distance features to adapt to various complex scenarios. At the same time, to enhance the robustness of the system, an error correction mechanism is implemented, and the machine learning model is updated in real time through a feedback mechanism. Such real-time updates help reduce the misjudgment rate and ensure the accuracy of the system in different scenarios. Once the system determines a custody situation, a real-time decision is triggered. At the same time, the system triggers an alarm mechanism to notify relevant personnel or execute other security measures; The steps for marking abnormal item situations through distance calculation include: Multiple stereo vision-based depth cameras are deployed inside the waiting hall. These cameras are arranged in different areas to monitor various parts of the waiting hall. When a passenger enters the waiting hall and comes within the monitoring range of the depth cameras, these cameras generate real-time depth maps. Using the depth maps, the system adopts an object detection algorithm to real-time identify the passengers and carried items in the depth maps. Through the results of the object detection algorithm, the system can accurately identify the position of each target in the depth maps. Subsequently, the system utilizes the distance information in the depth maps to determine the relative positions of the passengers and items in physical space by calculating the distance from each coordinate point to the camera. This enables the system to understand the spatial relationships between each target and the surrounding environment in real time, including the specific positions of the passengers and items in the waiting hall and the distances between them.

2. The big data-driven intelligent traffic monitoring system according to claim 1, wherein: The system includes: An authentication and association establishment module for verifying the passenger's identity, binding the face information with the ID card information, and initially establishing the association between the passenger and the carried items; Real-time tracking and item list update module, which is used to identify passengers and items in real time, extract detailed information, update the list of carried items, and monitor time correlation; Custody situation determination and decision trigger module, which is used to analyze the video stream captured by the depth camera, determine whether the passenger is in custody of the item, and trigger a real-time decision once the custody situation is determined.

Images