Method and system for realizing full-process dynamic tracking and abnormity early warning based on Internet of Things

Abstract

The invention discloses a method and a system for realizing full-process dynamic tracking and abnormity early warning based on the Internet of Things. Luggage information parameters, luggage stage state parameters and operator inspection and sorting error features which are obtained through processing are obtained, dynamic tracking and early warning features are obtained through chain type grading processing, a luggage abnormal state recognition method is obtained, and an abnormal behavior recognition model is constructed through the dynamic tracking and early warning features and the luggage abnormal state recognition method. According to the invention, a multi-fusion discrimination classification model is constructed by considering luggage parameters, states and personnel operation conditions, and training correction of a classification strategy distribution coefficient-based multi-fusion classification model of an SVM model and a Fisher criterion classifier is carried out according to the same data set trained by the model. According to the method, the luggage dynamic state can be tracked in real time, the luggage query time is greatly shortened, the luggage error rate is reduced, and the intelligent level of an airport is improved.

Description

Technical Field

[0001] The present invention relates to a method for realizing full-process dynamic tracking and abnormal warning based on the Internet of Things, which is used in the technical field of airport intelligent transportation Internet of Things services, and in particular to a method and system for realizing full-process dynamic tracking and abnormal warning based on the Internet of Things. Background Art

[0002] With the rapid development of the air transport industry, airport baggage check-in volume continues to grow. To improve service efficiency, major airports have launched self-service baggage check-in services, allowing passengers to complete the baggage check-in process independently, alleviating the pressure on manual check-in counters.

[0003] Currently, airport self-service baggage drop-off systems typically utilize a combination of manual verification, barcode labeling, and RFID tagging. After confirming the passenger's identity, staff print a barcode label and attach it to the baggage. The system then records the baggage information and assigns a transportation route. During the baggage transport process, baggage location information is updated through regular manual inspections and scanning at fixed checkpoints. Baggage handover is completed at the destination airport through manual tag verification. However, manual inspections result in delayed updates of baggage status information, failing to reflect the baggage's actual location. Furthermore, when baggage volume surges, manual verification becomes inefficient, easily leading to baggage backlogs and impacting airport operational efficiency.

[0004] In existing scenarios, machine learning models are used to match big data to predict baggage status to facilitate customers' full-process dynamic tracking of airport baggage and abnormal warnings. However, the data processing and model utilization are still too simple, and the prediction error rate is still relatively high.

[0005] Therefore, there are still challenges in acquiring and processing baggage information, baggage status parameters, and manual operation data, as well as in building models for identifying abnormal behavior. Existing technologies still have room for improvement in how to comprehensively consider baggage parameters, status, and human operation conditions to build more accurate and diverse machine learning models, and in which models to select for training and refining multi-fusion discriminant classification models, enabling dynamic prediction and tracking of baggage throughout the entire process, as well as accurate anomaly warnings. Furthermore, these models can track baggage movements in real time, significantly reducing baggage query time, lowering baggage error rates, and improving airport intelligence. Summary of the Invention

[0006] In order to solve the above technical problems, the present invention proposes a method and system for realizing full-process dynamic tracking and abnormal warning based on the Internet of Things.

[0007] In a first aspect of the present invention, a method for implementing full-process dynamic tracking and abnormality early warning based on the Internet of Things is provided, the method comprising: Baggage information parameters are obtained by collecting and processing RFID tags on the baggage, and the baggage phase status parameters are obtained through the Internet of Things. The operator's inspection and sorting error characteristics are obtained using the local deployment server. Based on the baggage information parameters, the baggage stage status parameters, and the operator inspection and sorting error characteristics, a dynamic tracking warning feature is obtained through chain hierarchical processing, and a method for identifying the abnormal state of the baggage at that time is obtained; Based on the dynamic tracking warning features and the baggage abnormal state identification method, an abnormal behavior identification model based on classification strategy allocation improvement is constructed; Based on the abnormal behavior recognition model, a real-time recognition method for abnormal luggage status is output for luggage status in different scenarios.

[0008] Furthermore, the baggage information parameters are constructed using baggage weight, size, and baggage process node time characteristics.

[0009] Furthermore, the baggage process node time features are obtained by taking the time when the baggage enters the airport intelligent baggage check-in conveyor belt during check-in. The baggage process node time features include check-in time, security check time, sorting and loading time, and baggage retrieval time.

[0010] Furthermore, the luggage stage status parameter indicates the stage status of the luggage, which is located and monitored according to the process node where the luggage is located, and is set according to the status of the luggage during check-in, security inspection, sorting, loading and collection.

[0011] Furthermore, the operator inspection and sorting error characteristics are composed of the operator number, the manual security inspection error correction rate and the sorting error rate.

[0012] Furthermore, the chain-type hierarchical processing is specifically to set weights for the baggage information parameters, the baggage stage status parameters, and the operator inspection and sorting error features according to the needs of the improved abnormal behavior recognition model based on classification strategy allocation, and then chain-fuse them into the dynamic tracking warning features.

[0013] Furthermore, the improved abnormal behavior recognition model based on classification strategy allocation adopts a threshold fusion multi-analysis discrimination model.

[0014] Furthermore, the threshold fusion multi-analysis discriminant model is an improved SVM discriminant classifier fused with the Fisher criterion classifier based on classification strategy allocation, and the calculation formula is:

[0015] Where, The output baggage abnormal status recognition method is Assign coefficients to the classification strategy of the SVM discriminant classifier, is the classification strategy allocation coefficient of the Fisher criterion classifier, is the normal vector of the hyperplane for training the SVM discriminant classifier, b is the intercept for training the SVM discriminant classifier, is the normal vector of the Fisher criterion classifier perpendicular to the hyperplane. The above normal vector and intercept are obtained through training of each model. To dynamically track early warning features.

[0016] It also provides a full-process dynamic tracking and abnormality warning system based on the Internet of Things. The system implements a full-process dynamic tracking and abnormality warning method based on the Internet of Things, including an Internet of Things baggage information parameter processing module, a baggage stage status parameter retrieval module, a client operator sorting error feature inspection module, an abnormal behavior recognition model construction module, and a full-process dynamic tracking and abnormality warning module: The Internet of Things luggage information parameter processing module is used to obtain luggage information parameters by collecting and processing the RFID tags on the luggage; The luggage phase status parameter retrieval module is configured to obtain the luggage phase status parameters through the Internet of Things; The client operator inspection and sorting error feature module uses the local deployment server to obtain the operator inspection and sorting error feature; The abnormal behavior recognition model construction module: Based on the baggage information parameters, the baggage stage status parameters, and the operator inspection and sorting error characteristics, a chain-type hierarchical process is performed to obtain dynamic tracking and warning characteristics, and a method for identifying the abnormal baggage status at that time; and based on the dynamic tracking and warning characteristics and the abnormal baggage status identification method, an improved abnormal behavior recognition model based on classification strategy allocation is constructed; The full-process dynamic tracking and abnormal warning module: based on the abnormal behavior recognition model, outputs a real-time baggage abnormal status recognition method for baggage status in different scenarios.

[0017] Furthermore, the improved abnormal behavior recognition model based on classification strategy allocation adopts a threshold fusion multi-analysis discrimination model; The threshold fusion multi-analysis discrimination model is based on the improved SVM discrimination classifier fused with the Fisher criterion classifier based on classification strategy allocation. The calculation formula is:

[0018] Where, The output baggage abnormal status recognition method is Assign coefficients to the classification strategy of the SVM discriminant classifier, is the classification strategy allocation coefficient of the Fisher criterion classifier, is the normal vector of the hyperplane for training the SVM discriminant classifier, b is the intercept for training the SVM discriminant classifier, is the normal vector of the Fisher criterion classifier perpendicular to the hyperplane. The above normal vector and intercept are obtained through training of each model. To dynamically track early warning features.

[0019] Therefore, the present invention has the beneficial effect of chaining and hierarchically processing baggage information parameters, baggage stage status parameters, and operator inspection and sorting error rate characteristics to generate dynamic tracking and warning features, thereby obtaining a baggage abnormality identification method. This method utilizes the dynamic tracking and warning features and the baggage abnormality identification method to construct an abnormal behavior recognition model. The present invention considers baggage parameters, status, and operator operation conditions to construct a multi-fusion discriminant classification model. Based on the same training data set, the multi-fusion discriminant classification model based on the classification strategy allocation coefficient is trained and modified for both the SVM model and the Fisher criterion classifier. This model enables dynamic prediction and tracking of baggage throughout the entire process, as well as accurate abnormality warnings. Furthermore, this method enables real-time tracking of baggage movements, significantly reducing baggage query time, lowering baggage error rates, and improving the intelligence level of airports.

[0020] More embodiments and improved effects of the present invention will be further introduced in conjunction with the drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a flow chart of the method for realizing full-process dynamic tracking and abnormal warning based on the Internet of Things of the present invention; Figure 2 This is a schematic diagram of the system for realizing full-process dynamic tracking and abnormal warning based on the Internet of Things of the present invention; Figure 3 This is a schematic diagram of an airport baggage check-in system in an embodiment of the present invention; Figure 4 This is a diagram showing the basic principle of threshold fusion multi-analysis discrimination in an embodiment of the present invention; Figure 5 This is a schematic diagram of the structure of an electronic device for implementing the method of the present invention in an embodiment of the present invention. DETAILED DESCRIPTION

[0022] The invention is further described below in conjunction with the accompanying drawings and specific implementation methods. The discriminant analysis model used in the present invention is an improvement of the multi-fusion classification discriminant model in the application scenario of airport baggage status tracking and abnormal warning.

[0023] like Figure 1 and Figure 2As shown, the method and system for realizing full-process dynamic tracking and abnormal warning based on the Internet of Things of the present invention belong to the airport transportation Internet of Things application service, and therefore belong to other Internet services such as agriculture, intelligent transportation Internet of Things application services, and therefore belong to the emerging software and new information technology service industry.

[0024] In a first aspect of the present invention, a method for implementing full-process dynamic tracking and abnormality early warning based on the Internet of Things is provided, the method comprising: Baggage information parameters are obtained by collecting and processing RFID tags on the baggage, and the baggage phase status parameters are obtained through the Internet of Things. The operator's inspection and sorting error characteristics are obtained using the local deployment server. Based on the baggage information parameters, the baggage stage status parameters, and the operator inspection and sorting error characteristics, a dynamic tracking warning feature is obtained through chain hierarchical processing, and a method for identifying the abnormal state of the baggage at that time is obtained; Based on the dynamic tracking warning features and the baggage abnormal state identification method, an abnormal behavior identification model based on classification strategy allocation improvement is constructed; Based on the abnormal behavior recognition model, a real-time recognition method for abnormal luggage status is output for luggage status in different scenarios.

[0025] Support Vector Machine (SVM) is a type of generalized linear classifier that performs binary classification on data using supervised learning. Its decision boundary is the maximum margin hyperplane solved for the learning samples.

[0026] If we project the points in a sample's multidimensional feature space onto a straight line, we can compress the feature space into one dimension. The key, then, is to find the direction of this line. Therefore, the goal of the Fisher criterion classifier is to find the optimal direction of this line and how to achieve the transformation that projects it in that optimal direction. This projection is precisely the solution vector we seek.

[0027] By fusing the data of the two in existing scenarios, the advantages of the two classifiers can be effectively combined to reduce the classification generalization, obtain more accurate baggage status prediction results, and improve airport customer satisfaction.

[0028] Furthermore, the baggage information parameters are constructed using baggage weight, size, and baggage process node time characteristics.

[0029] In this embodiment, one form of constructing baggage information parameters is a feature vector representation (baggage weight, size, and baggage process node time feature construction). It can also be represented as a matrix vector:

[0030] The airport's intelligent baggage control system, leveraging the Internet of Things (IoT) to achieve full-process dynamic tracking and anomaly warnings, first establishes a stable and efficient data transmission channel with the check-in counter and applies RFID tags to checked-in baggage. The system utilizes a combination of 5G communication technology and a wired network. When a passenger places their baggage at the check-in counter, the high-precision weighing and distance sensors installed there activate and read the RFID tags.

[0031] The weighing sensor uses the pressure sensing principle to convert the luggage weight into an electrical signal and transmit it to the control chip at the check-in counter. The distance sensor measures the time it takes for light waves to be emitted and reflected back to the receiver, and combines this with the distance sensor's distance calculation method to obtain the length, width, and height data of each side of the luggage and thereby obtain the relevant luggage volume parameters as the luggage size.

[0032] Furthermore, the baggage process node time features are obtained by taking the time when the baggage enters the airport intelligent baggage check-in conveyor belt during check-in. The baggage process node time features include check-in time, security check time, sorting and loading time, and baggage retrieval time.

[0033] Check-in time is the time from the moment a user checks in to the moment their baggage begins to be transported from the conveyor belt to the security checkpoint. Security check time is the time it takes for baggage to pass security. Sorting and loading time is the time it takes for baggage to be sorted and loaded after passing security. Baggage claim time is the time it takes from the moment the aircraft docks at the airport, the moment the cargo hold is opened, the baggage is unloaded, and the baggage is placed on the check-in conveyor belt, until the baggage is claimed by the customer. All of these times are integer values.

[0034] In this embodiment, the time characteristics of the baggage process node are composed of the check-in time, security check time, sorting and loading time, and baggage claim time to form a feature vector. A specific embodiment is represented as (check-in time, security check time, sorting and loading time, baggage claim time).

[0035] Typically, the time it takes for baggage to be checked in on an airplane, from placement on the check-in conveyor belt to the security checkpoint, is approximately 20-40 minutes. After being transported from the check-in conveyor belt to the security check area, the baggage typically undergoes an X-ray scan and other steps. Traditional security systems process approximately one to three minutes per piece. Baggage must complete security screening, sorting, and transfer to the apron for loading. According to typical airport reports, the entire sorting and loading process typically takes 5-20 minutes. Similarly, baggage claim time is slightly earlier than the sum of the security check and sorting and loading times, typically taking 10-15 minutes. This differs from the user's wait time for baggage pickup because it excludes the time spent waiting for the aircraft to open its hold and the time it takes to get onto the conveyor belt. The specific time characteristics for each baggage process node are represented as (0.5, 0.05, 0.25, 0.15), representing 0.5 hour for check-in, 0.05 hour for security screening, 0.25 hour for sorting and loading, and 0.15 hour for baggage claim.

[0036] In one representation of this embodiment, the baggage information parameter is expressed as a complete matrix form by supplementing the baggage information matrix vector according to the general form of data processing. In this embodiment, 1 is added. Then, a baggage information parameter in this embodiment is represented as: , where 1 represents the supplementary value of the eigenvector, 15 represents the baggage weight of 15 kg, and 45 represents the checked baggage volume of 45 inches.

[0037] Furthermore, the luggage stage status parameter indicates the stage status of the luggage, and the luggage is positioned and its status is monitored according to the process node where the luggage is located.

[0038] Specifically, the status of the baggage during check-in, security check, sorting, loading and collection is set. If the baggage is at check-in, the baggage stage status parameter is defined as (1,1), which means that the baggage is at the check-in state and is on the check-in conveyor belt. If the baggage stage status parameter is (2,2), it means that the baggage is at the security check state and is inside the security check instrument. If the baggage stage status parameter is (3,1), it means that the baggage is at the sorting state and is on the check-in conveyor belt during sorting. If the baggage stage status parameter is (4,0), it means that the baggage is at the loading state and is on the transfer vehicle. If the baggage stage status parameter is (4,-1), it means that the baggage is at the loading state and is in the aircraft cargo hold. If the baggage stage status parameter is (5,1), it means that the baggage is at the claim state and is on the check-in conveyor belt during claiming. If the baggage stage status parameter is (0,0), it means that the baggage has been claimed by the customer. If the status parameter is (X,9), it means that the baggage has an abnormality at stage X. If it is represented by a matrix vector, one of the baggage stage status parameters is .

[0039] Furthermore, the operator inspection and sorting error characteristics are composed of the operator number, the manual security inspection error correction rate and the sorting error rate.

[0040] Because airport security checks and sorting can cause baggage errors, resulting in abnormal baggage status, airports consider the operator's security check error correction rate and sorting error rate to make corresponding status prediction corrections for subsequent full-process dynamic tracking of baggage and abnormality warnings. The manual security check error correction rate, that is, the rate of correction of security check anomalies, can, to a certain extent, reflect the correction index of baggage status prediction, facilitating the construction and use of subsequent models.

[0041] The operator number is generally a general numerical number set by this application and is the general number of the security inspector at the time. The security inspection error rate and the manual correction rate after the security inspection machine error are generally less than 1%, and the sorting error rate is generally less than 2%. Therefore, one expression of the operator inspection and sorting error characteristics can be (1, 3, 0.35%, 1.35%), which means that security inspector No. 1 and sorter No. 3 have a manual security inspection error correction rate of 0.35% and a sorting error rate of 1.35%.

[0042] Furthermore, the chain-type hierarchical processing is specifically to set weights for the baggage information parameters, the baggage stage status parameters, and the operator inspection and sorting error features according to the needs of the improved abnormal behavior recognition model based on classification strategy allocation, and then chain-fuse them into the dynamic tracking warning features.

[0043] The chain-based hierarchical processing in this embodiment is to assign coefficients to the importance of the features of the input model for subsequent baggage status tracking prediction. The coefficient assignment in this application is based on the high correlation between the accuracy of dynamic tracking of the entire baggage process and abnormal warning after training of the subsequent input model, and is an adaptive value for the airport baggage status tracking and warning scenario. The dynamic tracking and warning feature is represented by ( Baggage information parameters, baggage phase status parameters, (1- ) Operator checks for sorting error characteristics). is the feature importance distribution coefficient. If the baggage information parameter is represented by a matrix vector, the dynamic tracking warning feature is Since the time feature dimension of the baggage process node is 4 and the numerical dimension of the baggage stage state parameter is 2, additional matrix feature 1-filling processing is performed when performing matrix-vector representation. This is a processing method for model processing data to facilitate the normalization of input feature processing for subsequent models. It is a data preprocessing step before the model performs data processing.

[0044] Furthermore, the improved abnormal behavior recognition model based on classification strategy allocation adopts a threshold fusion multi-analysis discrimination model.

[0045] like Figure 4 As shown in FIG, the basic principle diagram of the threshold fusion multi-analysis discriminant model is shown. Furthermore, the threshold fusion multi-analysis discriminant model is an improved SVM discriminant classifier fused with the Fisher criterion classifier based on classification strategy allocation. The calculation formula is:

[0046] Where, The output baggage abnormal status recognition method is Assign coefficients to the classification strategy of the SVM discriminant classifier, The Fisher criterion classifier classification strategy allocation coefficient is set according to the model classification threshold obtained after the dynamic tracking warning feature is input after model training. In this embodiment, it is set to prevent the difference between the final classification results of the two models from being too large. In the experiment, it was found that the output value of the SVM discriminant classifier was too large, so this application sets Greater than , is the normal vector of the hyperplane for training the SVM discriminant classifier, b is the intercept for training the SVM discriminant classifier, is the normal vector of the Fisher criterion classifier perpendicular to the hyperplane. The above normal vector and intercept are obtained by training each model using the same training data set. To dynamically track early warning features.

[0047] In this embodiment, the classification result is determined by the improved comprehensive threshold value after the threshold classification strategy allocation coefficient of the fusion multi-analysis discrimination model, and finally the baggage abnormal state recognition method is obtained. If the value of is greater than 5, the corresponding baggage abnormal status recognition method is to predict that the baggage is in the check-in state and on the check-in belt. If the value is less than or equal to 5 and greater than 3, it means that the luggage is in the security inspection state and is inside the security inspection instrument. If the value is less than or equal to 3 and greater than 2, it means that the baggage is in the sorting state and is on the consignment conveyor belt where it is sorted. If the value is less than or equal to 2 and greater than 1, it means that the luggage is in the loading state and is on the transfer vehicle. If the value is less than or equal to 1 and greater than 0, it means that the luggage is in the loading state and is in the aircraft cargo hold. If the value of is equal to 0, it means that the baggage is in the claim state and is on the check-in conveyor belt at the time of claim or the customer has claimed it. If the value of is less than 0, it means the baggage is abnormal. The absolute value range of negative numbers is the same as the above state, but it indicates an exception in the same state.

[0048] It also provides a full-process dynamic tracking and abnormality warning system based on the Internet of Things. The system implements a full-process dynamic tracking and abnormality warning method based on the Internet of Things, including an Internet of Things baggage information parameter processing module, a baggage stage status parameter retrieval module, a client operator sorting error feature inspection module, an abnormal behavior recognition model construction module, and a full-process dynamic tracking and abnormality warning module: The Internet of Things luggage information parameter processing module is used to obtain luggage information parameters by collecting and processing the RFID tags on the luggage; The luggage phase status parameter retrieval module is configured to obtain the luggage phase status parameters through the Internet of Things; The client operator inspection and sorting error feature module uses the local deployment server to obtain the operator inspection and sorting error feature; The abnormal behavior recognition model construction module: Based on the baggage information parameters, the baggage stage status parameters, and the operator inspection and sorting error characteristics, a chain-type hierarchical process is performed to obtain dynamic tracking and warning characteristics, and a method for identifying the abnormal baggage status at that time; and based on the dynamic tracking and warning characteristics and the abnormal baggage status identification method, an improved abnormal behavior recognition model based on classification strategy allocation is constructed; The full-process dynamic tracking and abnormal warning module: based on the abnormal behavior recognition model, outputs a real-time baggage abnormal status recognition method for baggage status in different scenarios.

[0049] Furthermore, the improved abnormal behavior recognition model based on classification strategy allocation adopts a threshold fusion multi-analysis discrimination model; The threshold fusion multi-analysis discrimination model is based on the improved SVM discrimination classifier fused with the Fisher criterion classifier based on classification strategy allocation. The calculation formula is:

[0050] Where, The output baggage abnormal status recognition method is Assign coefficients to the classification strategy of the SVM discriminant classifier, The Fisher criterion classifier classification strategy allocation coefficient is set according to the model classification threshold obtained after the dynamic tracking warning feature is input after model training. In this embodiment, it is set to prevent the difference between the final classification results of the two models from being too large. In the experiment, it was found that the output value of the SVM discriminant classifier was too large, so this application sets Greater than , is the normal vector of the hyperplane for training the SVM discriminant classifier, b is the intercept for training the SVM discriminant classifier, is the normal vector of the Fisher criterion classifier perpendicular to the hyperplane. The above normal vector and intercept are obtained by training each model using the same training data set. To dynamically track early warning features.

[0051] In this embodiment, the classification result is determined by the improved comprehensive threshold value after the threshold classification strategy allocation coefficient of the fusion multi-analysis discrimination model, and finally the baggage abnormal state recognition method is obtained. If the value of is greater than 5, the corresponding baggage abnormal status recognition method is to predict that the baggage is in the check-in state and on the check-in belt. If the value is less than or equal to 5 and greater than 3, it means that the luggage is in the security inspection state and is inside the security inspection instrument. If the value is less than or equal to 3 and greater than 2, it means that the baggage is in the sorting state and is on the consignment conveyor belt where it is sorted. If the value is less than or equal to 2 and greater than 1, it means that the luggage is in the loading state and is on the transfer vehicle. If the value is less than or equal to 1 and greater than 0, it means that the luggage is in the loading state and is in the aircraft cargo hold. If the value of is equal to 0, it means that the baggage is in the claim state and is on the check-in conveyor belt at the time of claim or the customer has claimed it. If the value of is less than 0, it means the baggage is abnormal. The absolute value range of negative numbers is the same as the above state, but it indicates an exception in the same state.

[0052] Therefore, the present invention has the beneficial effect of chaining and hierarchically processing baggage information parameters, baggage stage status parameters, and operator inspection and sorting error rate characteristics to generate dynamic tracking and warning features, thereby obtaining a baggage abnormality identification method. This method utilizes the dynamic tracking and warning features and the baggage abnormality identification method to construct an abnormal behavior recognition model. The present invention considers baggage parameters, status, and operator operation conditions to construct a multi-fusion discriminant classification model. Based on the same training data set, the multi-fusion discriminant classification model based on the classification strategy allocation coefficient is trained and modified for both the SVM model and the Fisher criterion classifier. This model enables dynamic prediction and tracking of baggage throughout the entire process, as well as accurate abnormality warnings. Furthermore, this method enables real-time tracking of baggage movements, significantly reducing baggage query time, lowering baggage error rates, and improving the intelligence level of airports.

[0053] Of course, it can be understood that each embodiment of the present invention can achieve one of the effects alone, and a combination of multiple embodiments of the present invention can achieve all of the above effects, but it is not required that each embodiment of the present invention achieve all of the above advantages and effects, because each embodiment of the present invention can constitute a separate technical solution and make one or more contributions to the existing technology. Each embodiment of the present invention is not affected, and the solution can still be implemented even if one of the embodiments is deleted.

[0054] The prior art mentioned in the above background technology section and specific embodiments section of the present invention can be regarded as part of the present invention and used to understand the meaning of some technical features or parameters.

Claims

1. A method for dynamic tracking and abnormal warning of the entire process based on the Internet of Things, characterized by: The method comprises: Baggage information parameters are obtained by collecting and processing RFID tags on the baggage, and the baggage phase status parameters are obtained through the Internet of Things. The operator's inspection and sorting error characteristics are obtained using the local deployment server. Based on the baggage information parameters, the baggage stage status parameters, and the operator inspection and sorting error characteristics, a dynamic tracking warning feature is obtained through chain hierarchical processing, and a method for identifying the abnormal state of the baggage at that time is obtained; Based on the dynamic tracking warning features and the baggage abnormal state identification method, an abnormal behavior identification model based on classification strategy allocation improvement is constructed; Based on the abnormal behavior recognition model, a real-time recognition method for abnormal luggage status is output for luggage status in different scenarios.

2. The method for realizing full-process dynamic tracking and abnormality early warning based on the Internet of Things according to claim 1, characterized in that: The baggage information parameters are constructed using baggage weight, size, and baggage process node time characteristics.

3. The method for realizing full-process dynamic tracking and abnormality early warning based on the Internet of Things according to claim 2, characterized in that: The baggage process node time features are obtained by taking the time when the baggage enters the airport intelligent baggage check-in conveyor belt during check-in. The baggage process node time features include check-in time, security check time, sorting and loading time, and baggage collection time.

4. The method for realizing full-process dynamic tracking and abnormality early warning based on the Internet of Things according to claim 3 is characterized by: The luggage stage status parameter indicates the stage status of the luggage. It locates and monitors the status of the luggage according to the process node where the luggage is located, and sets the status of the luggage according to the check-in, security check, sorting, loading and collection.

5. The method for realizing full-process dynamic tracking and abnormality early warning based on the Internet of Things according to claim 4 is characterized in that: The operator inspection and sorting error characteristics are composed of the operator number, the manual security inspection error correction rate and the sorting error rate.

6. The method for realizing full-process dynamic tracking and abnormality early warning based on the Internet of Things according to claim 5, characterized in that: The chain-type hierarchical processing specifically involves weighting the baggage information parameters, the baggage stage status parameters, and the operator inspection and sorting error features according to the needs of the improved abnormal behavior recognition model based on classification strategy allocation, and then chain-fusing them into the dynamic tracking warning features.

7. The method for realizing full-process dynamic tracking and abnormality early warning based on the Internet of Things according to claim 6, characterized in that: The improved abnormal behavior recognition model based on classification strategy allocation adopts a threshold fusion multi-analysis discrimination model.

8. The method for realizing full-process dynamic tracking and abnormality early warning based on the Internet of Things according to claim 7, characterized in that: The threshold fusion multi-analysis discrimination model is based on the improved SVM discrimination classifier fused with the Fisher criterion classifier based on classification strategy allocation. The calculation formula is: Where, The output baggage abnormal status recognition method is Assign coefficients to the classification strategy of the SVM discriminant classifier, is the classification strategy allocation coefficient of the Fisher criterion classifier, is the normal vector of the hyperplane for training the SVM discriminant classifier, b is the intercept for training the SVM discriminant classifier, is the normal vector of the Fisher criterion classifier perpendicular to the hyperplane. The above normal vector and intercept are obtained through training of each model. To dynamically track early warning features.

9. A full-process dynamic tracking and abnormality warning system based on the Internet of Things, which implements the method according to claim 8, comprises an Internet of Things baggage information parameter processing module, a baggage periodic status parameter retrieval module, a client operator sorting error feature verification module, an abnormal behavior recognition model construction module, and a full-process dynamic tracking and abnormality warning module, characterized by: The Internet of Things luggage information parameter processing module is used to obtain luggage information parameters by collecting and processing the RFID tags on the luggage; The luggage phase status parameter retrieval module is configured to obtain the luggage phase status parameters through the Internet of Things; The client operator inspection and sorting error feature module uses the local deployment server to obtain the operator inspection and sorting error feature; The abnormal behavior recognition model construction module: Based on the baggage information parameters, the baggage stage status parameters, and the operator inspection and sorting error characteristics, a chain-type hierarchical process is performed to obtain dynamic tracking and warning characteristics, and a method for identifying the abnormal baggage status at that time; and based on the dynamic tracking and warning characteristics and the abnormal baggage status identification method, an improved abnormal behavior recognition model based on classification strategy allocation is constructed; The full-process dynamic tracking and abnormal warning module: based on the abnormal behavior recognition model, outputs a real-time baggage abnormal status recognition method for baggage status in different scenarios.

10. The system for realizing full-process dynamic tracking and abnormality early warning based on the Internet of Things as claimed in claim 9, characterized in that: The improved abnormal behavior recognition model based on classification strategy allocation adopts a threshold fusion multi-analysis discrimination model; The threshold fusion multi-analysis discrimination model is based on the improved SVM discrimination classifier fused with the Fisher criterion classifier based on classification strategy allocation. The calculation formula is: Where, The output baggage abnormal status recognition method is Assign coefficients to the classification strategy of the SVM discriminant classifier, is the classification strategy allocation coefficient of the Fisher criterion classifier, is the normal vector of the hyperplane for training the SVM discriminant classifier, b is the intercept for training the SVM discriminant classifier, is the normal vector of the Fisher criterion classifier perpendicular to the hyperplane. The above normal vector and intercept are obtained through training of each model. To dynamically track early warning features.

Images