Early machine inspection target tracking method and tracking system

Abstract

The invention provides an early machine inspection target tracking method and tracking system. The tracking method comprises the steps that partition detection is carried out on a target, multiple pieces of initial detection data of multiple partitions are obtained, and the target comprises to-be-detected luggage and / or a to-be-detected person; the partition processing units of the multiple partitions are used for processing the multiple pieces of corresponding initial detection data, and first processing data of the multiple partitions are obtained; sending the first processing data of the plurality of partitions to a central server; processing the first processing data of the plurality of partitions by using the central server to obtain second processing data of the plurality of partitions; and tracking and detecting the targets in the plurality of partitions according to the second processing data of the plurality of partitions.

Description

Technical Field

[0001] The present disclosure relates to the technical field of article security inspection, and more specifically, to a method and system for tracking a target for early machine inspection. Background Art

[0002] There are two main means of tracking and supervising targets for preliminary machine inspection. The first is to manually check whether there is target luggage marked by customs in the surveillance video. If it is manually determined that the target luggage appears, the on-site customs officers will be notified to handle it. The second is to install an electronic fence lock on the target luggage, so that the on-site customs officers can handle the target luggage within the supervision area.

[0003] However, the current pre-machine inspection system does not track pedestrians or luggage throughout the entire process, and lacks real-time tracking, accurate positioning, and abnormal behavior identification after pedestrians pick up their checked luggage. Among them, the first method of video surveillance requires staff to identify and judge the monitoring screen with the naked eye, which greatly increases the demand for human resources and is prone to misjudgment. The second method of installing electronic fence locks has relatively high installation and maintenance costs, and requires the use of more complex electronic equipment and technology. If a large area needs to be covered, it will cost more. If there is interference from electronic equipment around, it will affect the accuracy of the fence. At the same time, there is also the possibility that the electronic lock is damaged or broken. After the electronic lock fails, it is impossible to continue to track the target person, which will increase the risk of missed interception. In addition, the above two methods use manual screening of surveillance videos to review the events that have occurred, which makes it difficult to link the target luggage with the monitoring perspective in real time, and it is impossible to achieve pre-warning and post-fast positioning, which is not conducive to improving the efficiency of customs disposal. Summary of the invention

[0004] In view of this, an embodiment of the present disclosure provides a method and a system for tracking a target with machine inspection in advance.

[0005] One aspect of the present disclosure provides a method for tracking a target detected by a machine in advance, which includes:

[0006] Performing partition detection on a target to obtain a plurality of initial detection data of a plurality of partitions, wherein the target includes luggage to be detected and / or a person to be detected;

[0007] Using the partition processing units of the plurality of partitions to process the corresponding plurality of initial detection data respectively to obtain first processed data of the plurality of partitions;

[0008] sending the first processed data of the plurality of partitions to a central server;

[0009] The first processed data of the plurality of partitions are processed by the central server to obtain second processed data of the plurality of partitions; and the targets in the plurality of partitions are tracked and detected according to the second processed data of the plurality of partitions.

[0010] According to an embodiment of the present disclosure, the plurality of zones include a baggage identification zone, a baggage claim and lobby zone, a concealed zone, and an exit zone;

[0011] The step of performing partition detection on the target to obtain a plurality of initial detection data of the plurality of partitions includes:

[0012] In the baggage identification area, the baggage to be detected is detected to obtain first detection data of the baggage identification area;

[0013] In the baggage claim and lobby area, inspect the baggage to be inspected and the person to be inspected to obtain second inspection data of the baggage claim and lobby area; and

[0014] In the hidden area and the exit area, the luggage to be detected and the person to be detected are detected to obtain third detection data of the hidden area and the exit area.

[0015] According to an embodiment of the present disclosure, the partition processing units of the plurality of partitions include a first processing unit, a second processing unit, and a third processing unit;

[0016] The using the partition processing unit of the plurality of partitions to process the corresponding plurality of initial detection data respectively to obtain the first processed data of the plurality of partitions comprises:

[0017] Processing the first detection data using the first processing unit to obtain first processed data of the baggage identification area;

[0018] Processing the second detection data using the second processing unit to obtain first processing data of the baggage claim and lobby area; and

[0019] The third processing unit is used to process the third detection data to obtain first processed data of the hidden area and the exit area.

[0020] According to an embodiment of the present disclosure, the step of processing the first detection data by the first processing unit to obtain first processed data of the baggage identification area includes:

[0021] Using the first processing unit to perform identification processing on the first detection data to determine target baggage that needs attention;

[0022] adding a luggage tag to the target luggage, wherein the luggage tag includes a first tracking address;

[0023] Tracking and detecting the target luggage based on the first tracking address to obtain luggage coordinates and video frames of the target luggage; and

[0024] The luggage coordinates and the video frame of the target luggage are encoded together with the first tracking address to obtain first processing data of the luggage identification area.

[0025] According to an embodiment of the present disclosure, the using the central server to process the first processed data of the plurality of partitions to obtain the second processed data of the plurality of partitions comprises: processing the first processed data of the baggage identification area to obtain the second processed data of the baggage identification area,

[0026] The step of processing the first processed data of the baggage identification area to obtain the second processed data of the baggage identification area includes:

[0027] extracting the baggage coordinates, the first tracking address and the video frame data from the first processed data of the baggage identification area;

[0028] Acquire the image of the target luggage from the video frame data according to the luggage coordinates; and

[0029] The image of the target luggage is analyzed and extracted to obtain a first feature of the target luggage.

[0030] According to an embodiment of the present disclosure, the processing of the first processed data of the baggage identification area to obtain the second processed data of the baggage identification area further includes: saving the first feature of the target baggage to the baggage global cache of the central server; and

[0031] According to the current state of the target luggage, the first feature of the target luggage in the luggage global cache is updated.

[0032] According to an embodiment of the present disclosure, the using the second processing unit to process the second detection data to obtain the first processed data of the baggage claim and lobby area includes:

[0033] Obtaining the second position coordinates of all luggage to be detected, and adding a second tracking address to the luggage to be detected;

[0034] Acquire the third position coordinates of the person to be detected near the luggage to be detected, and add a third tracking address to the person to be detected; and

[0035] The data including the second position coordinates, the second tracking address and the video frame of the luggage to be detected and the data including the third position coordinates, the third tracking address and the video frame of the person to be detected are encoded to obtain the first processing data of the luggage claim and hall area.

[0036] According to an embodiment of the present disclosure, the processing of the first processed data of the plurality of partitions by the central server to obtain the second processed data of the plurality of partitions includes: processing the first processed data of the baggage claim and lobby area to obtain the second processed data of the baggage claim and lobby area,

[0037] The step of processing the first processed data of the baggage claim and lobby area to obtain the second processed data of the baggage claim and lobby area comprises:

[0038] extracting data including the second position coordinates, the second tracking address and the video frame of the luggage to be detected and data including the third position coordinates, the third tracking address and the video frame of the person to be detected based on the first processed data of the baggage claim and hall area;

[0039] Acquire an image of the luggage to be detected from a video frame containing the luggage to be detected according to the second position coordinates, and perform feature extraction on the image of the luggage to be detected to obtain a second feature of the luggage to be detected;

[0040] Acquire an image of the person to be detected from a video frame containing the person to be detected according to the third position coordinates, and perform feature extraction on the image of the person to be detected to obtain a third feature of the person to be detected;

[0041] Comparing the second feature of the luggage to be detected with the first feature of the target luggage in the luggage global cache to determine whether the luggage to be detected is the target luggage; and

[0042] The target person associated with the target luggage is verified by calculating and rechecking the target person associated with the target luggage using continuous frames to obtain the features of the target person associated with the target luggage.

[0043] According to an embodiment of the present disclosure, the use of continuous frame calculation to verify the target person associated with the target luggage includes: using continuous frame calculation to verify the person whose coordinates overlap with the target luggage within a unit time.

[0044] According to an embodiment of the present disclosure, the method further includes: forming a pairing group between the target luggage and the associated target person, and saving features of the pairing group to a pairing group global cache of the central server.

[0045] According to an embodiment of the present disclosure, the method further includes: simultaneously tracking and detecting the target luggage and the target person according to the second tracking address and the third tracking address, and updating the features of the optimal viewing angles of the target luggage and the target person to the pairing group global cache.

[0046] According to an embodiment of the present disclosure, the method further includes: in the process of tracking the target luggage, sending the latest image and position of the target luggage to a display terminal.

[0047] According to an embodiment of the present disclosure, the method further includes: when an abnormal situation occurs in the pairing group, sending an alarm message to the display terminal, the abnormal situation includes the target luggage being lost; and / or,

[0048] The method further includes: when the luggage mark is broken, sending an alarm message to the display terminal.

[0049] According to an embodiment of the present disclosure, the using the central server to process the first processed data of the plurality of partitions to obtain the second processed data of the plurality of partitions comprises: processing the first processed data of the hidden area and the exit area to obtain the second processed data of the hidden area and the exit area,

[0050] The step of processing the first processed data of the hidden area and the exit area to obtain the second processed data of the hidden area and the exit area comprises:

[0051] Determine the target luggage and the target person based on the pairing group global cache of the central server;

[0052] Detect whether the target luggage and / or the target person enters a hidden area or an exit area, and if so, send a reminder message to the display terminal.

[0053] According to an embodiment of the present disclosure, the method includes: detecting whether the target luggage enters the hidden area;

[0054] If the target luggage enters the hidden area, an abnormal reminder message is sent to the display terminal.

[0055] According to an embodiment of the present disclosure, the method further comprises: detecting whether the target luggage and the target person enter the exit area;

[0056] If any one of the target luggage and the target person enters the exit area, an alarm message is sent to the display terminal.

[0057] Another aspect of the present disclosure provides a pre-machine inspection target tracking system, which includes:

[0058] A plurality of partition cameras, wherein the plurality of partition cameras are used to perform partition detection on a target to obtain a plurality of initial detection data of the plurality of partitions, wherein the target includes luggage to be detected and a person to be detected;

[0059] a plurality of partition processing units, the plurality of partition processing units being used to process the corresponding plurality of initial detection data respectively to obtain first processed data of the plurality of partitions, and to send the first processed data of the plurality of partitions to a central server; and

[0060] A central server is used to process the first processed data of the multiple partitions to obtain the second processed data of the multiple partitions, and to track and detect targets in the multiple partitions according to the second processed data of the multiple partitions.

[0061] According to an embodiment of the present disclosure, the plurality of zone cameras include a baggage identification camera and a baggage tracking camera, the baggage identification camera is located in the baggage identification zone, and the baggage tracking camera is located in the baggage claim and hall zone and the exit zone.

[0062] According to an embodiment of the present disclosure, the plurality of partition processing units include a first processing unit, a second processing unit and a third processing unit.

[0063] The first processing unit is used to process the first detection data detected by the baggage identification camera in the baggage identification area to obtain first processed data of the baggage identification area;

[0064] The second processing unit is used for processing the second detection data detected by the baggage tracking camera in the baggage claim and lobby area to obtain the first processing data of the baggage claim and lobby area; and

[0065] The third processing unit is used for processing the third detection data detected by the baggage tracking camera in the hidden area and the exit area to obtain the first processed data of the hidden area and the exit area.

[0066] According to an embodiment of the present disclosure, the central server includes a first feature extraction module, a second feature extraction module, a comparison module and an association module, wherein:

[0067] The first feature extraction module is used to extract the first feature of the target luggage;

[0068] The second feature extraction module is used to extract the second feature of the luggage to be detected and the third feature of the person to be detected;

[0069] The comparison module is used to compare the second feature of the luggage to be detected with the first feature of the target luggage to determine whether the luggage to be detected is the target luggage;

[0070] The association module is used to associate the target luggage with the target person.

[0071] According to an embodiment of the present disclosure, the central server further includes:

[0072] a pairing group generating module, the pairing group generating module being used to generate a pairing group of the target luggage and the target person according to the results of the comparison module and the association module; and

[0073] The pairing group global cache module is used to store and update the characteristics of the target luggage and the characteristics of the target person.

[0074] According to an embodiment of the present disclosure, the central server further includes: a luggage global cache module, and the luggage global cache module is used to store and update the first feature of the target luggage.

[0075] According to an embodiment of the present disclosure, the tracking system further includes a display terminal, which is used to receive the image and position of the target luggage, the image and position of the target person, an abnormality reminder message and an alarm information.

[0076] One or more of the above embodiments have the following beneficial effects:

[0077] (1) It solves the problem that the existing pre-inspection target luggage and target persons rely too much on human subjective judgment, especially the high risk of missed inspection during peak hours. At the same time, compared with the electronic fence lock method, it effectively reduces the installation and maintenance costs and solves the problem that the target person cannot be tracked separately after the target luggage is lost.

[0078] (2) By refining the preliminary machine inspection scenarios and adopting cloud-edge collaborative deployment, single-view target detection and tracking are placed on the edge camera side, improving the real-time detection performance. The central server performs differentiated computing and analysis based on the camera's detection results, effectively reducing the central computing power consumption and solving the computing power bottleneck problem during peak traffic hours.

[0079] (3) The target baggage and the target person are tracked in association. When the physical identification of the target baggage is found to be damaged, an alarm message will be issued in real time, and the associated target person will continue to be tracked. When the target baggage is blocked in the concealed area and then reappears in the lobby area, the system will issue an alarm message to the target baggage that has entered the concealed area within the specified time and has not reappeared, alerting the customs officers, and continue to track the target person associated with the target baggage. When the lost or destroyed target baggage and the target person arrive at the exit area, the system will push a message that the target person has arrived at the exit.

[0080] (4) Intelligent detection and tracking of target luggage and passengers, real-time push of suspicious information, and early warning before reaching the exit, so that customs officers can prepare in advance, improve customs handling efficiency, and reduce missed inspections. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] The above and other objects, features and advantages of the present disclosure will become more apparent through the following description of the embodiments of the present disclosure with reference to the accompanying drawings, in which:

[0082] Figure 1 is a schematic diagram of a scenario of a pre-machine inspection target tracking system according to an embodiment of the present disclosure;

[0083] Figure 2 is a flow chart of a method for tracking a target according to an embodiment of the present disclosure;

[0084] Figure 3 is a flow chart of multiple partition detection according to an embodiment of the present disclosure;

[0085] Figure 4 is a flowchart of processing of partition processing unit data of multiple partitions according to an embodiment of the present disclosure;

[0086] Figure 5 is a flowchart of processing of first detection data by a first processing unit according to an embodiment of the present disclosure;

[0087] Figure 6 is a flowchart of processing of first processing data of a baggage identification zone by a central server according to an embodiment of the present disclosure;

[0088] Figure 7 is a flowchart of processing of second detection data by a second processing unit according to an embodiment of the present disclosure;

[0089] Figure 8 is a flowchart of processing of first processing data of baggage claim and lobby area by a central server according to an embodiment of the present disclosure;

[0090] Fig. 9 is a flowchart of processing of first processing data of a hidden area and an exit area by a central server according to an embodiment of the present disclosure;

[0091] Fig.10 is a structural block diagram of a pre-machine inspection target tracking system according to an embodiment of the present disclosure; and

[0092] Fig.11 is a structural block diagram of a central server according to an embodiment of the present disclosure.

[0093] It should be noted that, for the sake of clarity, in the drawings used to describe the embodiments of the present disclosure, the sizes of the overall / local structures or the overall / local areas may be enlarged or reduced, that is, these drawings are not drawn according to the actual scale. DETAILED DESCRIPTION

[0094] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0095] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0096] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0097] Baggage inspection is an important part of airports and customs. Checked baggage first passes through the pre-inspection system, which uses perspective image analysis to mark problematic baggage. When these bags pass through the detection gate, the tags on the baggage are read to intercept or release them. In order to further strengthen supervision, when the baggage leaves the baggage carousel, the problematic target baggage or the target person associated with the target baggage is marked. When the target baggage or the target person is identified at the exit, the alarm system is triggered, and the staff intercepts and inspects it.

[0098] However, in the prior art, the discovery of target luggage and target persons for machine inspection relies too much on subjective human judgment, resulting in a high risk of missed inspections during peak traffic hours. In addition, the prior art does not track the target luggage or target persons throughout the entire process. If pedestrians tear off the labels after picking up their luggage but before clearing customs, the problematic luggage may be missed, bringing the risk of missed inspections. In addition, the problematic luggage may change hands or be abandoned during the journey, which may easily lead to missed inspections or the inability to find one or more target persons associated with the problematic luggage. This may result in the inability to track and update the status of the associated target luggage and target persons in real time, and it is impossible to provide advance warnings and rapid positioning afterwards, which is not conducive to improving the efficiency of security inspections.

[0099] Based on this, the embodiments of the present disclosure provide a preliminary machine inspection target tracking method and tracking system. The preliminary machine inspection target tracking method includes: performing partition detection on the target to obtain multiple initial detection data of multiple partitions, and the target includes luggage to be detected and / or people to be detected; using the partition processing units of multiple partitions to process the corresponding multiple initial detection data respectively to obtain first processed data of multiple partitions; sending the first processed data of multiple partitions to the central server; using the central server to process the first processed data of multiple partitions to obtain second processed data of multiple partitions; and tracking and detecting the targets in multiple partitions according to the second processed data of multiple partitions. By refining the preliminary machine inspection scenario and adopting the cloud-edge collaborative deployment method, the single-view target detection and tracking are placed on the edge camera side, which improves the real-time detection. The central server performs differentiated calculation and analysis based on the detection results of the camera, effectively reducing the central computing power consumption and solving the computing power bottleneck problem during peak traffic periods.

[0100] The following explanation will be given using the example of luggage arriving at the airport. It should be noted that the pre-machine inspection target tracking method and tracking system disclosed in the present invention can be used in places such as ports, airports, and stations. Of course, the above-mentioned places cannot be used as a limitation of the present invention. In the place, pedestrians have a certain trajectory displacement and time period after picking up their luggage and before the luggage undergoes special inspection (such as customs inspection), which is within the scope of the present invention.

[0101] Figure 1 is a schematic diagram of a scene of a pre-machine inspection target tracking system according to an embodiment of the present disclosure, Figure 2 It is a flowchart of a method for tracking a target according to an embodiment of the present disclosure.

[0102] For example, refer to Figure 1 , the preliminary machine inspection target tracking system provided by the embodiment of the present disclosure can perform partition detection on the target and obtain multiple initial detection data of multiple partitions. For example, the target may include luggage to be detected 140 and a person to be detected 170. The multiple partitions include a baggage identification area N1, a baggage claim and hall area N2, and a hidden area and an exit area N3. The luggage to be detected 140 can be sent from the luggage identification area N1 to the baggage carousel 200 of the baggage claim and hall area N2 through the security inspection conveyor belt 130. The person to be detected 170 can pick up the luggage to be detected 140 in the baggage claim and hall area N2, and carry the luggage to be detected 140 to move in the baggage claim and hall area N2 or the hidden area and exit area N3. Among them, the hidden area and exit area N3 include a hidden area 180 and an exit area 190.

[0103] Exemplarily, the hidden area 180 may be an area not monitored by the baggage tracking camera 160, for example, the hidden area 180 may include a toilet. The baggage claim and lobby area N2 and the exit area 190 may be provided with a plurality of baggage tracking cameras 160 for tracking and monitoring. For example, a plurality of baggage tracking cameras 160 located in different areas may perform cross-view tracking of the baggage to be inspected 140 and the person to be inspected 170, for example, may perform cross-view tracking of the target baggage and the target person associated with the target baggage.

[0104] Exemplarily, the baggage identification area N1 may be located at the entrance of the area to be inspected, and the baggage to be inspected 140 may first enter the baggage identification area N1. In the baggage identification area N1, the baggage identification camera 120 may be used to perform security inspection on the baggage to be inspected 140 to obtain initial inspection data in the baggage identification area N1. The first processing unit M1 may process the initial inspection data in the baggage identification area N1 to obtain first processing data of the baggage identification area N1. For example, the first processing unit M1 may screen out target baggage from a plurality of baggage to be inspected 140. The target baggage may be a suspicious item, such as suspicious baggage. The first processing unit M1 may also be used to encode data such as the coordinates and video frames of the target baggage to obtain first processing data.

[0105] Exemplarily, the first processing unit M1 may process initial detection data collected by one or more luggage identification cameras 120 .

[0106] Exemplarily, the first processing unit M1 may include a processor or a processing module embedded in one or more baggage identification cameras 120 of the baggage identification area S1. That is, the first processing unit M1 and the baggage identification camera 120 may be integrated together, for example, the first processing unit M1 and the baggage identification camera 120 may be integrated together to form a smart camera with data processing function. Alternatively, the first processing unit M1 may be an independent processor or a processing module capable of communicating with the baggage identification camera 120 and the central server.

[0107] After the preliminary inspection of the luggage 140 to be inspected is completed in the luggage identification area N1, it can be transferred to the luggage claim and lobby area N2, for example, it can be transferred to the luggage carousel 200 of the luggage claim and lobby area N2 through the luggage conveyor belt 130. The person to be inspected 170 can pick up the luggage 140 to be inspected in the luggage claim and lobby area N2, for example, pick up the luggage 140 to be inspected on the luggage carousel 200. The person to be inspected 170 can carry the luggage 140 to be inspected and move in various different areas of the luggage claim and lobby area N2 and / or the hidden area and exit area N3 until the person to be inspected 170 and / or the luggage 140 to be inspected moves to the exit area 190, and completes the exit security inspection in the exit area, thereby ending the security inspection process.

[0108] Exemplarily, the pre-machine inspection target tracking system may further include a second processing unit M2 and a third processing unit M3. The second processing unit M2 may be used to process the initial detection data detected by the baggage tracking camera 160 in the baggage claim and lobby area N2 to obtain first processed data of the baggage claim and lobby area N2. The third processing unit M3 may be used to process the initial detection data detected by the baggage tracking camera 160 in the hidden area and exit area N3 to obtain first processed data of the hidden area and exit area N3.

[0109] Exemplarily, the second processing unit M2 may include a processor or processing module embedded in one or more baggage tracking cameras 160 in the baggage claim and lobby area N2. That is, the second processing unit M2 and the baggage tracking camera 160 may be integrated together, for example, the second processing unit M2 and the baggage tracking camera 160 are integrated together to form a smart camera with data processing functions. Alternatively, the second processing unit M2 may be an independent processor or processing module capable of communicating with the baggage tracking camera 160 in the baggage claim and lobby area N2 and the central server 110.

[0110] Exemplarily, the third processing unit M3 may include a processor or processing module embedded in one or more baggage tracking cameras 160 in the hidden area and the exit area N3. Alternatively, the third processing unit M3 may be an independent processor or processing module capable of communicating with the baggage tracking cameras 160 in the hidden area and the exit area N3 and the central server 110.

[0111] Exemplarily, the pre-machine inspection target tracking system may include a central server 110. The central server 110 may be a cloud server for processing security inspection information. For example, the central server 110 may process images of luggage to be inspected and / or persons to be inspected, extract features of luggage to be inspected and extract features of persons to be inspected. The central server 110 may also be used to send information to security inspectors, such as sending images and location information of target luggage, or sending warning information, etc.

[0112] Exemplarily, the first processing unit M1 may transmit the first processing data of the baggage identification area N1 to the central server 110. The central server 110 processes the first processing data of the baggage identification area N1 to obtain the second processing data of the baggage identification area N1. The central server 110 may also feed back the second processing data of the baggage identification area N1 to the first processing unit M1, so as to further track and detect the baggage to be detected in the baggage identification area N1.

[0113] Exemplarily, the second processing unit M2 may transmit the first processing data of the baggage claim and lobby area N2 to the central server 110. The central server 110 processes the first processing data of the baggage claim and lobby area N2 to obtain the second processing data of the baggage claim and lobby area N2.

[0114] Exemplarily, the third processing unit M3 may transmit the first processed data of the hidden area and the exit area N3 to the central server 110. The central server 110 then processes the first processed data of the hidden area and the exit area N3 to obtain the second processed data of the hidden area and the exit area N3.

[0115] By refining the preliminary machine inspection scenario, the first processing unit M1, the second processing unit M2 and the third processing unit M3 located on the camera side and the central server 110 located on the center side adopt a cloud-edge collaborative deployment method, and the single-view target detection and tracking are placed on the edge camera side, which improves the real-time detection. The central server performs differentiated computational analysis based on the camera's detection results (i.e., the detection results after processing by the first processing unit M1, the second processing unit M2 and the third processing unit M3), which can effectively reduce the center's computing power consumption and solve the computing power bottleneck problem during peak traffic periods.

[0116] Exemplarily, cameras located in different areas can exchange information with the same central server 110 through the partition processing unit. For example, the central server 110 can extract and analyze the first processed data of the baggage identification area N1 processed by the first processing unit M1, so as to determine the target baggage in the baggage identification area N1 and extract the first feature of the target baggage. The central server 110 can also extract and analyze the first processed data of the baggage claim and lobby area N2 processed by the second processing unit M2 and the first processed data of the hidden area and exit area N3 processed by the third processing unit M3, and obtain the second feature of the baggage to be detected 140 and the third feature of the person to be detected 170 located in the baggage claim and lobby area N2 and / or the hidden area and exit area N3. The central server 110 can also associate the baggage to be detected 140 with the person to be detected 170, for example, by analyzing the coordinate coincidence relationship between the baggage to be detected 140 and the person to be detected 170, and associate the baggage to be detected 140 with the person to be detected 170.

[0117] Exemplarily, the central server 110 may also compare the second feature of the luggage to be detected 140 retrieved from the baggage claim and lobby area N2 and / or the hidden area and exit area N3 with the first feature of the target luggage previously retrieved from the baggage identification area N1, thereby determining the target luggage located in the baggage claim and lobby area N2 and / or the hidden area and exit area N3, and determining the target person located in the baggage claim and lobby area N2 and / or the hidden area and exit area N3 according to the association between the luggage to be detected 140 and the person to be detected 170 in the baggage claim and lobby area N2 and / or the hidden area and exit area N3. The central server 110 may also be used to perform cross-view tracking detection on the target luggage and the associated target person.

[0118] This method can avoid the discovery of target luggage and target people relying too much on human subjective judgment, which is conducive to reducing the risk of missed detection and effectively reducing installation and maintenance costs. In addition, by tracking the target luggage and the target person in association, the problem of being unable to track the target person alone after the target luggage is lost can be solved.

[0119] For example, in conjunction with reference Figure 1 and Figure 2 The method for tracking a target by machine inspection in advance provided by an embodiment of the present disclosure may include the following steps S1-S5.

[0120] In step S1, a target is subjected to partition detection to obtain a plurality of initial detection data of a plurality of partitions, wherein the target includes luggage to be detected and / or a person to be detected.

[0121] In step S2, the partition processing units of the plurality of partitions are used to process the corresponding plurality of initial detection data respectively to obtain the first processed data of the plurality of partitions.

[0122] In step S3, the first processed data of the plurality of partitions are sent to the central server.

[0123] In step S4, the first processed data of the plurality of partitions are processed by the central server to obtain the second processed data of the plurality of partitions.

[0124] In step S5, the targets in the multiple partitions are tracked and detected according to the second processed data of the multiple partitions.

[0125] By refining the preliminary machine inspection scenario, the first processing unit M1, the second processing unit M2 and the third processing unit M3 located on the camera side and the central server 110 located on the center side adopt a cloud-edge collaborative deployment method, and place single-view target detection and tracking on the edge camera side, thereby improving the real-time detection. The central server performs differentiated computational analysis based on the camera's detection results (i.e., the detection results after processing by the first processing unit M1, the second processing unit M2 and the third processing unit M3), effectively reducing the center's computing power consumption and solving the computing power bottleneck problem during peak traffic periods.

[0126] Figure 3 is a flowchart of multiple partition detection according to an embodiment of the present disclosure.

[0127] For example, in the embodiments of the present disclosure, in combination with reference Figure 1 and Figure 3 In step S1, performing partition detection on the target to obtain multiple initial detection data of multiple partitions may specifically include the following steps S11-S13.

[0128] In step S11 , in the baggage identification area N1 , the baggage identification camera 120 is used to perform single-view detection on the baggage to be detected 140 , and obtain first detection data of the baggage identification area N1 .

[0129] In step S12, in the baggage claim and lobby area N2, the baggage tracking camera 160 is used to perform cross-view detection on the baggage 140 and the person 170 to be detected, and obtain second detection data of the baggage claim and lobby area N2.

[0130] In step S13, in the hidden area and the exit area N3, the baggage tracking camera 160 is used to perform cross-viewing angle detection on the baggage 140 and the person 170 to be detected, and the third detection data of the hidden area and the exit area N3 is obtained.

[0131] Figure 4 is a data processing flow chart of a partition processing unit of multiple partitions according to an embodiment of the present disclosure.

[0132] For example, in the embodiments of the present disclosure, referring to Figure 4 In step S2, the corresponding multiple initial detection data are processed respectively by using the partition processing units of multiple partitions to obtain the first processed data of multiple partitions, which can specifically include the following steps S21-S23.

[0133] In step S21, the first processing unit M1 is used to process the first detection data to obtain first processed data of the baggage identification area N1.

[0134] In step S22, the second processing unit M2 is used to process the second detection data to obtain first processed data of the baggage claim and lobby area N2.

[0135] In step S23, the third detection data is processed by the third processing unit M3 to obtain first processed data of the hidden area and the exit area N3.

[0136] Figure 5 is a flowchart of processing of first detection data by a first processing unit according to an embodiment of the present disclosure.

[0137] For example, in the embodiments of the present disclosure, referring to Figure 5 In step S21, the first processing unit M1 is used to process the first detection data to obtain the first processed data of the baggage identification area N1, which may specifically include the following steps S211-S214.

[0138] In step S211, the first processing unit M1 is used to perform identification processing on the first detection data to determine the target baggage that needs attention. For example, the target baggage that needs attention can be screened out from a plurality of baggage to be detected.

[0139] In step S212, a luggage tag is added to the target luggage, where the luggage tag includes a first tracking address.

[0140] In step S213, the target luggage is tracked and detected based on the first tracking address to obtain luggage coordinates and video frames of the target luggage.

[0141] In step S214, the luggage coordinates and the video frame of the target luggage are encoded together with the first tracking address to obtain the first processing data of the luggage identification area.

[0142] Exemplarily, in step S4, using the central server to process the first processed data of multiple partitions to obtain the second processed data of multiple partitions may specifically include the following step S41.

[0143] In step S41, the first processed data of the baggage identification area is processed to obtain the second processed data of the baggage identification area.

[0144] Figure 6 The figure is a flowchart of processing first processing data of a baggage identification zone by a central server according to an embodiment of the present disclosure.

[0145] For example, in the embodiments of the present disclosure, referring to Figure 6 In step S41, the first processed data of the baggage identification area is processed to obtain the second processed data of the baggage identification area, which may specifically include the following steps S411-S413.

[0146] In step S411, the luggage coordinates, the first tracking address and the video frame data in the first processed data of the luggage identification area N1 are extracted.

[0147] In step S412, an image of the target luggage is obtained from the video frame data according to the luggage coordinates.

[0148] In step S413, the image of the target luggage is analyzed and extracted to obtain the first feature of the target luggage.

[0149] The detection data of the baggage identification area is processed by the first processing unit located on the camera side to obtain the first processed data of the baggage identification area, and then the first processed data of the baggage identification area is transmitted to the central server for subsequent calculation and processing. This method can reduce the computing power consumption of the central server and solve the computing power bottleneck problem during peak traffic periods. In addition, the first processing unit located on the camera side has higher timeliness in data processing, which is conducive to improving the real-time performance of detection.

[0150] In some embodiments, the step S41 of processing the first processed data of the baggage identification area to obtain the second processed data of the baggage identification area may also include the following steps S414-S415.

[0151] In step S414, the first feature of the target baggage is saved in the baggage global cache of the central server.

[0152] By saving the first feature of the target baggage to the global baggage cache of the central server, in the subsequent detection process, the first feature of the target baggage can be used for comparison, and the target baggage can be screened out from the numerous baggage to be detected in the baggage claim and hall area N2 and / or the hidden area and exit area N3. There is no need to perform perspective detection again, and the target baggage can be screened out only by obtaining the external image of the baggage to be detected by the camera, which is conducive to improving the detection efficiency.

[0153] In step S415, the first feature of the target baggage in the baggage global cache is updated according to the current state of the target baggage.

[0154] This method can improve the accuracy and efficiency of detection, and can help staff better understand the current status of target luggage and target persons, thereby improving security inspection efficiency and reducing the risk of missed inspections.

[0155] According to an embodiment of the present disclosure, the early machine inspection target tracking method may further include: pushing a reminder message to a display terminal, for example, the reminder message may include that the target luggage has been found; and using the display terminal to display an image of the target luggage and a monitoring screen of the current viewing angle. For example, the display terminal may be a mobile display terminal carried by a staff member.

[0156] This method can help staff to timely understand the information of target luggage, so that they can conduct relevant inspections and processes more promptly and accurately, which is conducive to improving the efficiency of security inspections.

[0157] For example, in the baggage identification area N1, different types of image detection devices can be used to inspect the baggage to be inspected. For example, a perspective image detection device can be used to obtain a perspective image of the baggage to be inspected, and a baggage identification camera can be used to obtain an external image of the baggage to be inspected, so as to better screen out the target baggage and track and detect the target baggage.

[0158] Figure 7 is a flowchart of processing second detection data by the second processing unit according to an embodiment of the present disclosure.

[0159] For example, in the embodiments of the present disclosure, referring to Figure 7 In step S22, the second processing unit M2 is used to process the second detection data to obtain the first processing data of the baggage claim and lobby area N2, which can specifically include the following steps S221-S223.

[0160] In step S221, the second position coordinates of all the luggage to be detected are obtained, and the second tracking address is added to the luggage to be detected. For example, after all the luggage to be detected is detected by the luggage tracking camera 160 in the baggage claim and lobby area N2, the second server M2 can process the initial detection data of all the luggage to be detected, obtain the second position coordinates of all the luggage to be detected, and add the second tracking address to the luggage to be detected.

[0161] In step S222, the third position coordinates of the person to be detected near the luggage to be detected are obtained, and a third tracking address is added to the person to be detected. For example, the luggage tracking camera 160 in the luggage claim and lobby area N2 can also detect the person to be detected near each luggage to be detected, and the second server M2 can obtain the third position coordinates of the person to be detected near the luggage to be detected, and add the third tracking address to the person to be detected.

[0162] In step S223, the data of the video frame containing the second position coordinates, the second tracking address and the luggage to be detected and the data of the video frame containing the third position coordinates, the third tracking address and the person to be detected are encoded to obtain the first processing data of the luggage claim and the hall area.

[0163] This method can improve the real-time performance of baggage collection and lobby area detection, while also reducing the computing power consumption of the central server.

[0164] Exemplarily, in step S4, using the central server to process the first processed data of multiple partitions to obtain the second processed data of multiple partitions may also include the following step S42.

[0165] In step S42, the first processed data of the baggage claim and lobby area N2 is processed to obtain the second processed data of the baggage claim and lobby area N2.

[0166] Figure 8 4 is a flowchart of processing of first processing data of baggage claim and lobby area by a central server according to an embodiment of the present disclosure.

[0167] For example, in the embodiments of the present disclosure, referring to Figure 8 In step S42, the first processed data of the baggage claim and lobby area N2 is processed to obtain the second processed data of the baggage claim and lobby area N2, which may specifically include the following steps S421-S425.

[0168] In step S421, based on the first processed data of the baggage claim and lobby area N2, data of the video frame containing the second position coordinates, the second tracking address and the luggage to be detected and data of the video frame containing the third position coordinates, the third tracking address and the person to be detected are extracted.

[0169] In step S422, an image of the luggage to be detected is obtained from the video frame containing the luggage to be detected according to the second position coordinates, and feature extraction is performed on the image of the luggage to be detected to obtain a second feature of the luggage to be detected.

[0170] In step S423, an image of the person to be detected is obtained from a video frame containing the person to be detected according to the third position coordinates, and feature extraction is performed on the image of the person to be detected to obtain a third feature of the person to be detected.

[0171] In some embodiments, the central server may also filter out some video frames according to the second tracking address and / or the third tracking address, so as to obtain images of luggage to be detected and images of persons to be detected more quickly, which is beneficial to improving the processing efficiency of the central server.

[0172] In step S424, the second feature of the luggage to be detected is compared with the first feature of the target luggage in the luggage global cache to determine whether the luggage to be detected is the target luggage.

[0173] For example, the second feature of the luggage to be detected obtained in the luggage claim and lobby area N2 can be compared with the first feature of the target luggage obtained in the luggage identification area N1. If the overlap between the first feature and the second feature is higher than a preset threshold, the luggage to be detected can be considered as the target luggage. Subsequently, the tracking and monitoring of the target luggage can be strengthened in the luggage claim and lobby area N2 and the hidden area and exit area N3, so as to better grasp the real-time information of the target luggage.

[0174] In step S425, the target person associated with the target luggage is verified by using continuous frame calculation to obtain the features of the target person associated with the target luggage.

[0175] Exemplarily, using continuous frame calculation to verify the target person associated with the target luggage may include: using continuous frame calculation to verify the person whose coordinates overlap with the target luggage within a unit time. For example, the unit time may be less than or equal to 10 seconds. That is, the person to be detected who is at the same position as the target luggage within 10 seconds may be considered as the target person associated with the target luggage.

[0176] Through continuous video frames, one or more persons to be detected whose coordinates overlap with the target luggage can be screened out, and the one or more persons to be detected are considered as the target person associated with the target luggage. Then, according to the third feature of the person to be detected extracted in step S423, the feature of the target person is determined.

[0177] According to an embodiment of the present disclosure, the characteristics of the target person may include at least one of a body image and a face image.

[0178] In some embodiments, the baggage tracking camera can obtain a clear facial image of the target person, and can use the facial image of the target person to extract features of the target person.

[0179] In some embodiments, the baggage tracking camera may not be able to obtain a clear facial image of the target person. For example, the camera in the airport hall is located at a high position and the facial image obtained is not clear enough. Alternatively, the target person's features may be extracted based on the target person's body image.

[0180] In some embodiments, features of the target person's facial image and body image may also be extracted simultaneously, which may improve the reliability of tracking and monitoring.

[0181] According to an embodiment of the present disclosure, the preliminary machine inspection target tracking method may further include: forming a pairing group of target luggage and an associated target person, and saving features of the pairing group to a pairing group global cache of a central server.

[0182] Pairing the target luggage and the target person can facilitate cross-perspective tracking of the target luggage and the target person. Even if the target luggage and the target person are separated, the real-time status of the target luggage and the target person can be grasped at the same time, solving the problem of being unable to track the target person alone after the target luggage is lost.

[0183] According to an embodiment of the present disclosure, the preliminary machine inspection target tracking method may further include: simultaneously tracking and detecting the target luggage and the target person according to the second tracking address and the third tracking address, and updating the features of the optimal viewing angles of the target luggage and the target person to the pairing group global cache.

[0184] Through this method, the latest status of the target luggage and the target person can be timely understood. Even if the target luggage and / or the target person have undergone a makeover (for example, the target person changes clothes, shoes or hats, or the target luggage changes color or casing), the target luggage and the target person can still be accurately detected, thereby reducing the risk of missed inspection.

[0185] Exemplarily, in step S4, using the central server to process the first processed data of multiple partitions to obtain the second processed data of multiple partitions may also include the following step S43.

[0186] In step S43, the first processed data of the hidden area and the exit area N3 are processed to obtain the second processed data of the hidden area and the exit area N3.

[0187] Fig. 9 It is a flowchart of processing of first processing data of a hidden area and an exit area by a central server according to an embodiment of the present disclosure.

[0188] For example, in the embodiments of the present disclosure, referring to Fig. 9 In step S43, the first processed data of the hidden area and the exit area N3 are processed to obtain the second processed data of the hidden area and the exit area N3, which can specifically include the following steps S431-S432.

[0189] In step S431, the target baggage and the target person are determined based on the global cache of the pairing groups of the central server. For example, the target baggage and the target person located in or about to enter the hidden area and the exit area N3 can be determined based on the global cache of the pairing groups of the target baggage and the target person previously obtained in the baggage claim and lobby area N2, combined with the processing result of the detection data of the hidden area and the exit area N3 by the third server M3.

[0190] In step S432, it is detected whether the target luggage and / or the target person enters the hidden area or the exit area. If so, a reminder message is sent to the display terminal.

[0191] According to an embodiment of the present disclosure, the pre-machine inspection target tracking method may further include: in the process of tracking the target luggage, sending the latest image and position of the target luggage to the display terminal.

[0192] According to an embodiment of the present disclosure, the preliminary machine-detected target tracking method may further include: in the process of tracking the target person, sending the latest image and position of the target person to the display terminal.

[0193] Through the information linkage of cameras, partition processing units, central servers and display terminals, staff can timely understand the latest images and locations of target luggage and target persons, so that staff can quickly reach the locations of target luggage and target persons and inspect and process them.

[0194] According to an embodiment of the present disclosure, the pre-machine inspection target tracking method may further include: when an abnormal situation occurs in the target luggage and the target person pairing group, sending an alarm message to the display terminal. For example, the abnormal situation may include the target luggage being lost.

[0195] According to an embodiment of the present disclosure, the pre-machine inspection target tracking method may further include: when the baggage mark is broken, sending an alarm message to the display terminal. For example, it can be determined whether the baggage mark is broken by identifying the target person's action (detecting the target person's action of breaking the mark).

[0196] Through this method, the target luggage and the target person can be tracked in association. When the physical identification of the target luggage is found to be damaged, an alarm message will be issued in real time. At the same time, the associated target person will continue to be tracked, which can reduce the chance of missed detection.

[0197] In the embodiments of the present disclosure, the early machine inspection target tracking method may also include sending a reminder message or an alarm message to the display terminal in a timely manner when the target luggage and / or the target person enters or is about to enter a special area, so that the staff can notice the relevant situation and deal with it in time. For example, the special area may include a hidden area and an exit area.

[0198] According to an embodiment of the present disclosure, the preliminary machine-inspected target tracking method may further include the following steps S4321-S4322.

[0199] In step S4321, it is detected whether the target luggage enters the hidden area.

[0200] In step S4322, if the target luggage enters the hidden area, an abnormal reminder message is sent to the display terminal.

[0201] For example, when the target luggage reappears in the hall area after being blocked in the hidden area, the system will send an alarm message to the target luggage that has entered the hidden area within the specified time and has not reappeared, alerting the customs officer and continuing to track the target person associated with the target luggage.

[0202] According to an embodiment of the present disclosure, the preliminary machine-inspected target tracking method may further include the following steps S4323-S4324.

[0203] In step S4323, it is detected whether the target luggage and the target person enter the exit area.

[0204] In step S4324, if any one of the target luggage and the target person enters the exit area, an alarm message is sent to the display terminal.

[0205] When the target luggage and / or target person arrives at the exit area, the system will push a message to the display terminal that the target luggage and / or target person has arrived at the exit, which can help the staff to timely understand the information of the target luggage and target person, which is conducive to improving the efficiency of security inspection and reducing the probability of missed inspection.

[0206] Through this method, target luggage and passengers can be intelligently detected and tracked, suspicious information can be pushed in real time, and early warnings can be issued before reaching the exit, so that customs officers can prepare in advance, improve customs handling efficiency, and reduce missed inspections.

[0207] Fig.10 It is a structural block diagram of a preliminary machine inspection target tracking system according to an embodiment of the present disclosure.

[0208] Another aspect of the present disclosure provides a pre-machine inspection target tracking system 1000. Figure 1 and Fig.10 The preliminary machine inspection target tracking system 1000 may include: multiple partition cameras, multiple partition processing units and a central server.

[0209] Exemplarily, the multiple zone cameras include, for example, a baggage identification camera 120 located in the baggage identification area N1, a baggage tracking camera 1601 located in the baggage claim and lobby area N2, and a baggage tracking camera 1602 located in the hidden area and exit area N3. The multiple zone cameras can be used to perform zone detection on the target to obtain multiple initial detection data of the multiple zones, and the target includes the baggage to be detected and the person to be detected. For example, the baggage identification camera 120 located in the baggage identification area N1 can be used to detect the baggage to be detected located in the baggage identification area N1. For another example, the baggage tracking camera 1601 located in the baggage claim and lobby area N2 can detect the baggage to be detected and the person to be detected located in the baggage claim and lobby area N2. For another example, the baggage tracking camera 1602 located in the hidden area and exit area N3 can detect the baggage to be detected and the person to be detected located in the hidden area and exit area N3.

[0210] Exemplarily, the plurality of partition processing units include, for example, a first processing unit M1, a second processing unit M2, and a third processing unit M3. The plurality of partition processing units can be used to process the corresponding plurality of initial detection data respectively, obtain the first processed data of the plurality of partitions, and send the first processed data of the plurality of partitions to the central server.

[0211] Exemplarily, the partition processing unit may be designed to be integrated with the corresponding partition camera.

[0212] Exemplarily, the partition processing unit can be designed independently from the corresponding partition camera.

[0213] Exemplarily, the first processing unit M1 is used to process the first detection data detected by the baggage identification camera 120 in the baggage identification area N1 to obtain the first processed data of the baggage identification area. For example, the first processing unit M1 can process the initial detection data detected by the baggage identification camera 120 in the baggage identification area N1, such as screening out the target baggage, marking the address of the target baggage, and encoding the video frame and coordinate address of the target baggage to form a data packet.

[0214] Exemplarily, the second processing unit M2 can be used to process the second detection data detected by the baggage tracking camera 160 in the baggage claim and lobby area N2 to obtain the first processing data of the baggage claim and lobby area N2. For example, the second processing unit M2 can be used to detect all the baggage to be detected in the baggage claim and lobby area N2 and the persons to be detected near the baggage to be detected, and can also add tracking addresses to the baggage to be detected and the persons to be detected, respectively, so as to facilitate continuous tracking and detection. The second processing unit M2 can also package and encode the address coordinates and video frames of the detected baggage to be detected and the persons to be detected into a data packet, and send it to the central server 110.

[0215] Exemplarily, the third processing unit M3 may be used to process the third detection data detected by the baggage tracking camera 160 in the hidden area and the exit area N3 to obtain the first processed data of the hidden area and the exit area N3.

[0216] Exemplarily, the central server 110 can be used to process the first processed data of multiple partitions to obtain the second processed data of multiple partitions, and to track and detect the targets in the multiple partitions according to the second processed data of the multiple partitions. For example, the central server 110 can obtain the first feature of the target baggage in the baggage identification area N1 according to the first processed data of the baggage identification area N1 processed by the first processing unit M1. For another example, the central server 110 can also obtain the second feature of the baggage to be detected and the third feature of the person to be detected in the baggage claim and lobby area N2 according to the first processed data of the baggage claim and lobby area N2 processed by the second processing unit M2, and screen out the target baggage and target person in the baggage claim and lobby area N2 through comparison.

[0217] This design solves the problem that the existing pre-machine inspection target luggage and target person detection relies too much on human subjective judgment, especially the high risk of missed detection during peak hours. At the same time, compared with the electronic fence lock method, it effectively reduces the installation and maintenance costs and solves the problem of being unable to track the target person alone after the target luggage is lost. In addition, by refining the pre-machine inspection scenario and adopting a cloud-edge collaborative deployment method, single-view target detection and tracking are placed on the edge camera side, improving the real-time detection. The central server performs differentiated calculation and analysis based on the camera's detection results, effectively reducing the center's computing power consumption and solving the computing power bottleneck problem during peak hours.

[0218] Exemplarily, the central server 110 may also associate and pair the target luggage and the target person, and then continuously track and detect the target luggage and the target person in combination with the partition processing units and cameras in each zone, for example, the target luggage and the target person may be continuously monitored across viewing angles.

[0219] By tracking the target luggage and the target person in association, the problem of not being able to track the target person alone after the target luggage is lost can be solved.

[0220] For example, continue to refer to Fig.10 The pre-machine inspection target tracking system 1000 may also include a display terminal 500. The display terminal 500 may be used to receive various information such as the image and position of the target luggage, the image and position of the target person, abnormal reminder messages and alarm information. For example, the display terminal may display the screens that need attention in each partition, and may also receive reminder messages of each partition.

[0221] This design can help staff members to timely understand the information of target luggage and target persons, so that they can conduct relevant inspections and processes more promptly and accurately, which is conducive to improving the efficiency of security inspections.

[0222] Fig.11 is a structural block diagram of a central server according to an embodiment of the present disclosure.

[0223] According to the embodiments of the present disclosure, referring to Fig.11 The central server 110 may include a first feature extraction module 1101 , a second feature extraction module 1102 , a comparison module 1103 and an association module 1104 .

[0224] Exemplarily, the first feature extraction module 1101 may be used to extract the first feature of the target luggage, for example, extract the first feature of the target luggage detected in the luggage identification area N1.

[0225] Exemplarily, the second feature extraction module 1102 can be used to extract the second feature of the luggage to be detected and the third feature of the person to be detected. For example, the second feature of the luggage to be detected and the third feature of the person to be detected in the baggage claim and lobby area N2 and / or the hidden area and exit area N3 is extracted.

[0226] Exemplarily, the comparison module 1103 may be used to compare the second feature of the luggage to be detected with the first feature of the target luggage to determine whether the luggage to be detected is the target luggage.

[0227] Exemplarily, the association module 1104 may be used to associate the target luggage with the target person.

[0228] The first feature extraction module is used to extract features from the target baggage in the baggage identification area N1, which reduces the amount of calculation and improves the real-time detection. The second feature extraction module targets and extracts features from the baggage and people to be detected in the baggage claim and lobby area N2 and / or the hidden area and exit area N3, ensuring the accuracy and timeliness of the detection. By refining the preliminary machine inspection scene, the central server processes information according to the refined scene, which is conducive to reducing the central computing power consumption and solving the computing power bottleneck problem during peak traffic hours.

[0229] According to the embodiments of the present disclosure, continue to refer to Fig.11 The central server 110 may further include: a pairing group generating module 1105 and a pairing group global caching module 1106 .

[0230] Exemplarily, the pairing group generating module 1105 may be used to generate a pairing group of target luggage and target person according to the results of the comparing module 1103 and the associating module 1104 .

[0231] Exemplarily, the pairing group global cache module 1106 may be used to store and update the characteristics of the target luggage and the characteristics of the target person.

[0232] Through such a design, the target luggage and the target person can be paired together, so that they can be easily tracked across perspectives. Even if the target luggage and the target person are separated, the real-time status of the target luggage and the target person can be grasped at the same time, solving the problem of being unable to track the target person alone after the target luggage is lost.

[0233] According to the embodiments of the present disclosure, continue to refer to Fig.11 The central server 110 may further include: a luggage global cache module 1107. The luggage global cache module 1107 may be used to store and update the first feature of the target luggage.

[0234] Through such a design, clearer and more accurate characteristics of the target luggage can be obtained, which is conducive to improving the accuracy of tracking detection and reducing the probability of missed detection.

[0235] In some embodiments, the pre-machine inspection target tracking system 1000 may further include a baggage tagging device located in the baggage identification area, which may be used to add a baggage tag to the target baggage, wherein the baggage tag includes a first tracking address.

[0236] This design can facilitate subsequent tracking and monitoring of target luggage, quickly lock the target luggage and the associated target person, and improve tracking and monitoring efficiency.

[0237] According to an embodiment of the present disclosure, the baggage tracking camera 160 can be set in multiple areas such as the baggage claim area, the lobby area, and the exit area to continuously track and update the image and location of the target baggage and the image and location of the target person across perspectives, thereby facilitating staff to perform relevant processing in a timely manner and improving security inspection efficiency.

[0238] Exemplarily, the target persons are all pedestrians whose coordinates overlap with the target luggage within a unit time. For example, the unit time is less than or equal to 10 seconds. For example, one target luggage can be associated with one or more target persons.

[0239] By setting up multiple tracking cameras in multiple areas such as the baggage claim area, lobby area, and exit area, cross-perspective tracking of target luggage and target persons can be achieved, which can solve the problem of being unable to track the target person alone after the target luggage is lost.

[0240] Those skilled in the art can understand that the embodiments described above are exemplary and can be improved by those skilled in the art. The structures described in various embodiments can be freely combined without causing conflicts in structure or principle.

[0241] After describing the preferred embodiments of the present disclosure in detail, those skilled in the art can clearly understand that various changes and modifications can be made without departing from the scope and spirit of the appended claims, and the present disclosure is not limited to the implementation methods of the exemplary embodiments listed in the specification.

Claims

1. A method for tracking a target by early machine inspection, characterized in that: include: Performing partition detection on a target to obtain a plurality of initial detection data of a plurality of partitions, wherein the target includes luggage to be detected and / or a person to be detected; Using the partition processing units of the plurality of partitions to process the corresponding plurality of initial detection data respectively to obtain first processed data of the plurality of partitions; sending the first processed data of the plurality of partitions to a central server; Using the central server to process the first processed data of the plurality of partitions to obtain second processed data of the plurality of partitions; as well as The targets in the multiple partitions are tracked and detected according to the second processed data of the multiple partitions.

2. The method according to claim 1, wherein: The plurality of zones include a baggage identification zone, a baggage claim and lobby zone, a concealed zone, and an exit zone; The step of performing partition detection on the target to obtain a plurality of initial detection data of the plurality of partitions includes: In the baggage identification area, the baggage to be detected is detected to obtain first detection data of the baggage identification area; In the baggage claim and lobby area, inspect the baggage to be inspected and the person to be inspected to obtain second inspection data of the baggage claim and lobby area; and In the hidden area and the exit area, the luggage to be detected and the person to be detected are detected to obtain third detection data of the hidden area and the exit area.

3. The method according to claim 2, wherein: The partition processing units of the plurality of partitions include a first processing unit, a second processing unit and a third processing unit; The using the partition processing unit of the plurality of partitions to process the corresponding plurality of initial detection data respectively to obtain the first processed data of the plurality of partitions comprises: Processing the first detection data using the first processing unit to obtain first processed data of the baggage identification area; Processing the second detection data using the second processing unit to obtain first processing data of the baggage claim and lobby area; and The third processing unit is used to process the third detection data to obtain first processed data of the hidden area and the exit area.

4. The method according to claim 3, wherein: The using the first processing unit to process the first detection data to obtain the first processed data of the baggage identification area includes: Using the first processing unit to perform identification processing on the first detection data to determine target baggage that needs attention; adding a luggage tag to the target luggage, wherein the luggage tag includes a first tracking address; Tracking and detecting the target luggage based on the first tracking address to obtain luggage coordinates and video frames of the target luggage; and The luggage coordinates and the video frame of the target luggage are encoded together with the first tracking address to obtain first processing data of the luggage identification area.

5. The method according to claim 4, wherein: The using the central server to process the first processed data of the plurality of partitions to obtain the second processed data of the plurality of partitions comprises: processing the first processed data of the baggage identification area to obtain the second processed data of the baggage identification area, The step of processing the first processed data of the baggage identification area to obtain the second processed data of the baggage identification area includes: extracting the baggage coordinates, the first tracking address and the video frame data from the first processed data of the baggage identification area; Acquire the image of the target luggage from the video frame data according to the luggage coordinates; and The image of the target luggage is analyzed and extracted to obtain a first feature of the target luggage.

6. The method according to claim 5, wherein: The processing of the first processed data of the baggage identification area to obtain the second processed data of the baggage identification area further includes: saving the first feature of the target baggage to the baggage global cache of the central server; and According to the current state of the target luggage, the first feature of the target luggage in the luggage global cache is updated.

7. The method according to claim 3, wherein: The using the second processing unit to process the second detection data to obtain the first processing data of the baggage claim and lobby area comprises: Obtaining the second position coordinates of all luggage to be detected, and adding a second tracking address to the luggage to be detected; Acquire the third position coordinates of the person to be detected near the luggage to be detected, and add a third tracking address to the person to be detected; and The data including the second position coordinates, the second tracking address and the video frame of the luggage to be detected and the data including the third position coordinates, the third tracking address and the video frame of the person to be detected are encoded to obtain the first processing data of the luggage claim and hall area.

8. The method according to claim 7, wherein: The using the central server to process the first processed data of the plurality of partitions to obtain the second processed data of the plurality of partitions comprises: processing the first processed data of the baggage claim and lobby area to obtain the second processed data of the baggage claim and lobby area, The step of processing the first processed data of the baggage claim and lobby area to obtain the second processed data of the baggage claim and lobby area comprises: extracting data including the second position coordinates, the second tracking address and the video frame of the luggage to be detected and data including the third position coordinates, the third tracking address and the video frame of the person to be detected based on the first processed data of the baggage claim and hall area; Acquire an image of the luggage to be detected from a video frame containing the luggage to be detected according to the second position coordinates, and perform feature extraction on the image of the luggage to be detected to obtain a second feature of the luggage to be detected; Acquire an image of the person to be detected from a video frame containing the person to be detected according to the third position coordinates, and perform feature extraction on the image of the person to be detected to obtain a third feature of the person to be detected; Comparing the second feature of the luggage to be detected with the first feature of the target luggage in the luggage global cache to determine whether the luggage to be detected is the target luggage; and The target person associated with the target luggage is verified by calculating and rechecking the target person associated with the target luggage using continuous frames to obtain the features of the target person associated with the target luggage.

9. The method according to claim 8, wherein: The method of verifying the target person associated with the target luggage by using continuous frame calculation includes: verifying the person whose coordinates overlap with the target luggage within a unit time by using continuous frame calculation.

10. The method according to claim 9, wherein: The method further includes: forming a pairing group with the target luggage and the associated target person, and saving the features of the pairing group to a pairing group global cache of the central server.

11. The method according to claim 10, wherein: The method further includes: simultaneously tracking and detecting the target luggage and the target person according to the second tracking address and the third tracking address, and updating the features of the optimal viewing angles of the target luggage and the target person to the pairing group global cache.

12. The method according to claim 11, wherein: The method further includes: in the process of tracking the target luggage, sending the latest image and position of the target luggage to a display terminal.

13. The method according to claim 12, wherein: The method further includes: when an abnormal situation occurs in the pairing group, sending an alarm message to the display terminal, the abnormal situation including the target luggage being lost; and / or, The method further includes: when the luggage mark is broken, sending an alarm message to the display terminal.

14. The method according to claim 10, wherein: The using the central server to process the first processed data of the plurality of partitions to obtain the second processed data of the plurality of partitions comprises: processing the first processed data of the hidden area and the exit area to obtain the second processed data of the hidden area and the exit area, The step of processing the first processed data of the hidden area and the exit area to obtain the second processed data of the hidden area and the exit area comprises: Determine the target luggage and the target person based on the pairing group global cache of the central server; Detect whether the target luggage and / or the target person enters a hidden area or an exit area, and if so, send a reminder message to the display terminal.

15. The method according to claim 14, wherein: The method comprises: detecting whether the target luggage enters the hidden area; If the target luggage enters the hidden area, an abnormal reminder message is sent to the display terminal.

16. The method according to claim 14 or 15, wherein: The method further comprises: detecting whether the target luggage and the target person enter the exit area; If any one of the target luggage and the target person enters the exit area, an alarm message is sent to the display terminal.

17. A pre-machine inspection target tracking system, characterized in that: include: A plurality of partition cameras, wherein the plurality of partition cameras are used to perform partition detection on a target to obtain a plurality of initial detection data of the plurality of partitions, wherein the target includes luggage to be detected and a person to be detected; A plurality of partition processing units, the plurality of partition processing units being used to process the corresponding plurality of initial detection data respectively to obtain first processed data of the plurality of partitions, and send the first processed data of the plurality of partitions to a central server; and A central server is used to process the first processed data of the multiple partitions to obtain the second processed data of the multiple partitions, and to track and detect targets in the multiple partitions according to the second processed data of the multiple partitions.

18. The system of claim 17, wherein: The plurality of zone cameras include baggage identification cameras and baggage tracking cameras, wherein the baggage identification cameras are located in the baggage identification zone and the baggage tracking cameras are located in the baggage claim and lobby zone and the exit zone.

19. The system of claim 17, wherein: The plurality of partition processing units include a first processing unit, a second processing unit and a third processing unit, The first processing unit is used to process the first detection data detected by the baggage identification camera in the baggage identification area to obtain the first processed data of the baggage identification area; The second processing unit is used to process the second detection data detected by the baggage tracking camera in the baggage claim and lobby area to obtain the first processing data of the baggage claim and lobby area; and The third processing unit is used to process the third detection data detected by the baggage tracking camera in the hidden area and the exit area to obtain the first processed data of the hidden area and the exit area.

20. The system of claim 17, wherein: The central server includes a first feature extraction module, a second feature extraction module, a comparison module and an association module, wherein: The first feature extraction module is used to extract the first feature of the target luggage; The second feature extraction module is used to extract the second feature of the luggage to be detected and the third feature of the person to be detected; The comparison module is used to compare the second feature of the luggage to be detected with the first feature of the target luggage to determine whether the luggage to be detected is the target luggage; The association module is used to associate the target luggage with the target person.

21. The system of claim 20, wherein: The central server also includes: a pairing group generating module, the pairing group generating module being used to generate a pairing group of the target luggage and the target person according to the results of the comparison module and the association module; and The pairing group global cache module is used to store and update the characteristics of the target luggage and the characteristics of the target person.

22. The system of claim 20, wherein: The central server further comprises: a luggage global cache module, and the luggage global cache module is used to store and update the first feature of the target luggage.

23. The system according to any one of claims 17 to 22, wherein: The tracking system further comprises a display terminal, which is used to receive the image and position of the target luggage, the image and position of the target person, an abnormality reminder message and an alarm message.

Images