Mtmct system, information exchange method, device and readable medium

Abstract

The embodiment of the application relates to the technical field of edge computing and image processing, and particularly relates to an MTMCT system, an information exchange method and device and a computer readable medium, the MTMCT system (100) comprising at least two cameras (10) covering non-overlapping areas, configured to collect image frames, and to perform target identification and feature extraction on the collected image frames, and to maintain the identification ID and features of each target tracked by the camera; a central node (20) configured to maintain the ID and features of each target lost in the MTMCT system (100). Target detection and feature extraction are respectively performed by each camera in the MTMCT system, the central node realizes information sharing between the cameras, reduces bandwidth occupation and communication cost, shortens delay, improves efficiency and has higher computing capacity and reliability. Information can be effectively exchanged between different cameras, and the cameras are coordinated with each other.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the technical field of edge computing and image processing, and particularly relate to an MTMCT system, an information exchange method and device, and a computer readable medium. BACKGROUND

[0002] Multi-target Multi-camera tracking (MTMCT) automatically tracks multiple targets through a network of multiple cameras. Due to the multiple changes in the camera's view angle, lighting, and target posture, it is relatively difficult to implement multi-camera tracking.

[0003] In current multi-camera tracking solutions, target re-identification (ReID) technology is used to extract target features through a neural network to distinguish different targets. In implementation, images captured by each camera are usually sent to a central node for target detection and feature extraction. The disadvantage of this approach is that a large amount of image information needs to be transmitted between the cameras and the central node, resulting in high bandwidth occupation and low efficiency. SUMMARY

[0004] Current multi-camera tracking solutions can be considered as using a centralized processing method. Embodiments of the present application propose a new distributed processing method, in which target detection and feature extraction are performed by each camera in the MTMCT system, and the central node implements information sharing between the cameras, thereby reducing bandwidth occupation and communication cost, shortening delay, improving efficiency, and having higher computing power and reliability. In addition, for effective exchange of information between different cameras and mutual coordination, a feasible and efficient solution is also proposed.

[0005] The solution provided by embodiments of the present application can be used in a non-overlapping MTMCT system (the areas monitored by each camera do not overlap), in which the central node implements information sharing and exchange between the cameras, which can not only guarantee the performance of cross-camera target tracking, but also has small communication overhead.

[0006] In a first aspect, a non-overlapping distributed MTMCT system is provided, which can include a plurality of cameras and a central node in communication with each of the cameras. Each camera performs image acquisition, target identification and feature extraction. Each camera maintains an identification (ID) and features of tracked targets, and the central node maintains IDs and features of lost targets. The camera performs target identification and feature extraction on acquired image frames, and first matches the extracted features with tracked targets. If a match is found, the extracted features are merged with the original features. If no match is found, the camera further obtains IDs and features of lost targets from the central node and matches the extracted features with the IDs and features of lost targets one by one. Once a match is found, the ID and features of the matched lost target are maintained, and the newly extracted features are merged with the features of the matched lost target obtained from the central node. If the newly identified target cannot be matched with tracked targets or lost targets maintained by the central node, the camera can request the central node to assign a new ID to the newly identified target, and the camera maintains features of the newly identified target. In addition, the camera can periodically check whether a tracked target has been lost for a long time, and if so, the camera can send the ID and features of the lost target to the central node for maintenance.

[0007] In a second aspect, a method for information interaction is provided, which can be performed by a camera in the MTMCT system of the first aspect.

[0008] In a third aspect, a method for information interaction is provided, which can be performed by a central node in the MTMCT system of the first aspect.

[0009] In a fourth aspect, a camera is provided, which includes modules for performing the steps of the method performed by the camera in the MTMCT system of the first aspect.

[0010] In a fifth aspect, a central node is provided, which includes modules for performing the steps of the method performed by the central node in the MTMCT system of the first aspect.

[0011] In a sixth aspect, a camera is provided, which includes at least one memory configured to store computer readable code, and at least one processor configured to invoke the computer readable code to perform the steps of the method of the second aspect.

[0012] In a seventh aspect, a central node is provided, which includes at least one memory configured to store computer readable code, and at least one processor configured to invoke the computer readable code to perform the steps of the method of the third aspect.

[0013] In an eighth aspect, a computer readable medium is provided, and the computer readable medium has stored thereon computer readable instructions which, when executed by a processor, cause the processor to perform each step of the method provided in the second aspect or the third aspect.

[0014] The scheme provided by the embodiment of the application mainly differs from the current MTMCT system in that target identification and feature extraction are performed by each camera rather than a central node, the camera maintains the ID and features of the tracked target, and the central node maintains the ID and features of the lost target. In this way, a large amount of image information does not need to be transmitted between the camera and the central node, and only a small amount of information needs to be transmitted when the target is lost or a new target is found. Information sharing between the cameras is achieved through the central node, and cross-camera target tracking is achieved simply and with a small amount of overhead.

[0015] The camera can only transmit the features of the lost target under limited conditions. Moreover, the central node only sends the relevant features to different cameras. In this way, the central node only receives necessary information, and only sends necessary information to all the cameras, thereby reducing the communication overhead.

[0016] Since the calculation is distributed to different cameras, the camera can maintain tracking when the central node is closed. Moreover, since the amount of information exchanged is minimal, it is easier to set up multiple central nodes to prevent single-point failure (when there are multiple central nodes, each central node can be a backup for each other, or each is associated with part of the cameras, and information sharing of the MTMCT system can be achieved through information exchange between the central nodes).

[0017] If the target identified by the camera is not a tracked target, the features of the lost target can be obtained from the central node to achieve uninterrupted tracking of the target in a scenario in which the target moves from the coverage area of one camera to the coverage area of another camera. Since the ID and features of the lost target are uniformly maintained by the central node, it is not necessary to store and maintain the relevant information at each camera, which reduces the storage overhead of the camera and also achieves cross-camera information sharing.

[0018] If the target identified by the camera is a tracked target, the newly extracted features are merged with the features of the target maintained by the camera, which enriches the features of the target and provides more abundant data for subsequent cross-camera feature matching, thereby improving the accuracy of matching.

[0019] If the target identified by the camera is neither a tracked target nor a lost target, a new ID is uniformly assigned by the central node and sent to the camera that identified the target, thereby preventing confusion that can be caused by different cameras identifying and maintaining the target.

[0020] The camera periodically checks whether the tracked target has been lost, and if so, sends the ID and features of the lost target to the central node for unified maintenance, so that the target can be tracked in time when it moves to the next camera, and it can be determined that different cameras track the same target, thereby improving the accuracy of target tracking across cameras. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 A structure diagram of an MTMCT system provided by an embodiment of the application is shown.

[0022] Figure 2 Information maintained by each camera and the central node in the MTMCT system provided by an embodiment of the application is shown.

[0023] Figure 3 A structure diagram of a camera provided by an embodiment of the application is shown.

[0024] Figure 4 A structure diagram of a central node provided by an embodiment of the application is shown.

[0025] Figure 5A And Figure 5B A flowchart of information exchange between the camera and the central node in an embodiment of the application is shown.

[0026] Figure 6 A flowchart of the information exchange method on the camera side provided by an embodiment of the application is shown.

[0027] Figure 7 A flowchart of the information exchange method on the central node side provided by an embodiment of the application is shown.

[0028] LIST OF REFERENCE NUMERALS

[0029] 100: MTMCT system 10: camera 20: central node

[0030] 31: ID of the target tracked by the camera 10 32: features of the target tracked by the camera 10

[0031] 41: ID of the lost target 42: features of the lost target

[0032] 101: memory 102: processor 103: communication interface 104: camera

[0033] 30: image processing program 301: image processing module 302: feature matching module

[0034] 303: information exchange module

[0035] 201: memory 202: processor 203: communication interface

[0036] 40: information processing program 401: information interaction module 402: information processing module

[0037] 501, 502, 600, 700: information exchange method

[0038] S5011-S5017, S5021-S5026, S601-S612, S701-S709: method steps DETAILED DESCRIPTION

[0039] The subject matter described herein will now be discussed with reference to example implementations. It should be understood that these implementations are discussed solely for the purpose of illustrating aspects of the subject matter described herein and are not a limitation of the scope, applicability, or examples set forth in the claims. Changes in the function and arrangement of elements discussed can be made without departing from the scope of the subject matter's embodiments. Various examples can omit, substitute, or add various procedures or components as appropriate. For instance, the methods described can be performed in an order different than that described, and in an example, various steps can be added, omitted, or combined. Also, features described with respect to some examples can be combined in other examples.

[0040] As used herein, the terms "includes," "including," "has," "having," "contains," "containing," "comprises," "comprising," "including," "contains," "containing" or variants thereof do not exclude additional elements. For example, the expression "A includes B" means that A includes B or other elements not specified, as well as B. The term "based on" means "based, at least in part, on" and not "based solely on" unless specifically specified. The terms "one embodiment" and "an embodiment" mean "at least one embodiment." The term "another embodiment" means "at least one additional embodiment." The terms "a first," "a second," etc. do not require that there be only one of the indicated elements. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of this disclosure. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. The terms "comprises", "comprising", "includes", "including" and "contains", "containing" as used herein, are specifically intended to be open-ended. These terms encompass the elements or steps set forth herein, but not excluding additional elements or steps. It will be understood that when an element or step is referred to as being "connected" or "coupled" to another element or step, it can be directly connected or coupled or it can be indirectly connected or coupled, as appropriate.

[0041] The embodiments of the present application will be described below in detail with reference to the accompanying drawings.

[0042] Figure 1An MTMCT system 100 provided by an embodiment of the present application is shown, which includes a plurality of cameras 10 and one or more central nodes 20. Each camera 10 performs image acquisition, target recognition and feature extraction. For the case where there is only one central node 20, each camera 10 performs necessary information exchange with the central node 20 to realize tracking of a target when the target moves between areas covered by the plurality of cameras 10. For the case where there are a plurality of central nodes 20, each central node 20 can form a group with part of the cameras 10, information sharing between the cameras 10 in the group is realized through the central node 20 in the group, and information sharing between the cameras 10 in different groups is realized through the central nodes 20 of the groups. Alternatively, for the case where there are a plurality of central nodes 20, all the central nodes 20 can form a cluster of central nodes, which is equivalent to one central node, and perform information exchange with each camera 10.

[0043] As shown in Figure 2 , each camera 10 maintains an identification ID 31 and a feature 32 of each target tracked by itself, and the central node 20 maintains an ID and a feature of each target lost in the MTMCT system 100. The areas covered by the cameras 10 in the MTMCT system 100 are mutually non-overlapping, which means that the IDs of the targets tracked by different cameras are different, and the ID 41 and the feature 42 of the lost target maintained by the central node 20 are common to all the cameras 10.

[0044] On one hand, the camera 10 performs target recognition and feature extraction on the image frame collected by itself, first matches with the target tracked by itself, if matched, merges the extracted feature with the original saved feature, if not matched, further acquires the ID and the feature of the lost target from the central node 20 and matches one by one, once matched, maintains the ID and the feature of the matched lost target, and merges the newly extracted feature with the feature of the matched lost target acquired from the central node, for details and implementation, refer to steps S5011-S5013, S5015-S5017 and related descriptions in the following Figure 5A .

[0045] On the other hand, if the newly recognized target cannot be matched with the target tracked by the camera 10 or the lost target maintained by the central node 20, the camera 10 can request the central node 20 to allocate a new ID for the newly recognized target, and the camera 10 maintains the feature of the newly recognized target, for details and implementation, refer to steps S5014 and S5019 and related descriptions in the following Figure 5A .

[0046] In still another aspect, the camera 10 can also periodically check whether a target it has tracked has been lost for a long time, i.e., has been lost, and if so, send the ID and features of the lost target to the central node 20 for maintenance. The details of the process and implementation can refer to the subsequent description of steps S5021-S5026 and related description. Figure 5B In still another aspect, the camera 10 can also periodically check whether a target it has tracked has been lost for a long time, i.e., has been lost, and if so, send the ID and features of the lost target to the central node 20 for maintenance. The details of the process and implementation can refer to the subsequent description of steps S5021-S5026 and related description.

[0047] In the following, the camera 10 and the central node 20 will be described in detail with reference to the accompanying drawings. Figure 5A and the accompanying drawings. Figure 5B The information interaction between the camera 10 and the central node 20 will be described in detail.

[0048] Figure 5A The method 501 of information interaction between the camera 10 and the central node 20 is shown, which specifically includes the following steps:

[0049] S5011: The camera 10 performs target recognition and feature extraction on a frame of image collected by itself. The method of target recognition and feature extraction can use the existing method, which will not be described here.

[0050] S5012: The camera 10 matches the extracted features with the features of the targets it has tracked (herein referred to as "first targets", which can be all the targets tracked by the camera 10 or part of them), and if matched, step S5018 is executed, and if not matched, step S5013 is executed.

[0051] S5013: The camera 10 acquires the features of the lost targets (herein referred to as "second targets", which can be all the lost targets or part of them) from the central node 20.

[0052] S5014: The camera 10 matches the extracted features with the features of the second targets, and if matched successfully (i.e., one of the second targets can be matched), steps S5015-S5017 are executed, and if not matched successfully (i.e., none of the features of the second targets can be matched with the features extracted by the camera 10), step S5019 is executed.

[0053] S5015: The camera 10 sets the ID of the recognized target as the ID of the second target matched successfully.

[0054] S5016: The camera 10 notifies the central node 20 that the second target matched successfully has been tracked by the camera 10, and the central node 20 can delete the ID and features of the second target matched successfully from the list of maintained lost targets.

[0055] S5017: The camera 10 merges the extracted features with the features of the second target matched successfully acquired from the central node 20, and maintains the ID and features of the second target.

[0056] - S5018: The camera 10 merges the extracted features with the features of the matched first target, and continues to maintain the ID and features of the first target.

[0057] The merging of features in step S5017 and step S5018 includes adding the information of the newly acquired features to the originally saved feature information of the target, so as to enrich the features of the target, which can make the result of subsequent feature matching more accurate.

[0058] - S5019: The camera 10 requests the central node 20 to assign a new ID to the recognized target, and the central node 20 assigns a new ID to the camera 10 after receiving the request and returns it to the camera 10. The camera 10 will maintain the ID and features (i.e. the features extracted in step S5011) of the newly recognized target.

[0059] Figure 5B The method 502 of information interaction between the camera 10 and the central node 20 is shown, which specifically includes the following steps:

[0060] - S5021: The camera 10 calculates the time length from the time when the last time the third target was recognized to the current time for each third target tracked by itself.

[0061] - S5022: The camera 10 compares the calculated time length with a preset threshold value, and if the time length of a third target is greater than the preset threshold value (at this time, the third target can have moved out of the coverage area of the camera 10), steps S5023-S5026 are executed, otherwise, step S5021 is returned. The threshold value can be set according to the area of the coverage area of the camera 10 and the moving speed of the tracked target under normal circumstances. The larger the area and the slower the moving speed, the longer the threshold value can be set; otherwise, the smaller the area and the faster the moving speed, the shorter the threshold value can be set.

[0062] - S5023: The camera 10 sends the ID of the third target that exceeds the preset threshold value to the central node 20.

[0063] - S5024: The central node 20 marks the third target identified by the ID as lost.

[0064] - S5025: The camera 10 sends the features of the third target to the central node 20.

[0065] - S5026: The central node 20 maintains the ID and features of the third target.

[0066] Through the above steps, when a target moves out of the coverage area of the camera 10, the ID and features of the target are maintained by the unified central node 20, and other cameras 10 can obtain the features of the target and perform matching, thereby achieving cross-camera tracking.

[0067] Figure 3 The structure of the camera 10 provided by the embodiment of the application is shown. In addition to the camera device 104, the camera can further include at least one memory 101 including a computer readable medium, such as a random access memory (RAM); in addition, the camera 10 can further include at least one processor 102 coupled with the at least one memory 101. Computer executable instructions are stored in the at least one memory 101, and when executed by the at least one processor 102, can cause the at least one processor 102 to perform the steps described herein.

[0068] Figure 3 The at least one memory 101 shown in the figure can include an image processing program 30, so that the at least one processor 102 performs the information exchange method 600 described in the embodiment of the application. As shown in the figure, the image processing program 30 can include: Figure 3

[0069] - an image processing module 301 configured to perform target recognition and feature extraction on a frame of image collected by the camera 10;

[0070] - a feature matching module 302 configured to match the features extracted by the image processing module 301 with the features of at least one first target tracked by the camera 10;

[0071] - an information interaction module 303 configured to, if the features extracted by the image processing module 301 do not match the features of the at least one first target tracked by the camera 10, obtain the features of at least one second target lost from the central node 20 of the MTMCT system 100;

[0072] - the feature matching module 302 is further configured to match the extracted features with the features of the at least one second target;

[0073] - the information interaction module 303 is further configured to, if the matching of the features extracted by the image processing module 301 with the features of the at least one second target is successful,

[0074] - set the identification ID of the recognized target as the ID of the second target whose matching is successful,

[0075] - inform the central node 20 that the second target whose matching is successful has been tracked by the camera 10, and

[0076] ​- merge the extracted features with the features of the second target that the matching was successful with.

[0077] Other functions of the above modules can refer to the camera 10 side processing in the aforementioned method 501 and method 502, which will not be repeated here. Each module can also be considered as a functional module implemented by hardware, for implementing various functions involved in the information exchange method performed by the camera 10, such as pre-burning the control logic of each flow involved in the method into a chip such as a Field-Programmable Gate Array (FPGA) chip or a Complex Programmable Logic Device (CPLD), and executing the functions of the above modules by the chip or device. The specific implementation manner can be determined according to engineering practice.

[0078] In addition, the camera 10 can further include a communication interface 103 for communication between the camera 10 and other devices, such as the central node 20.

[0079] It should be mentioned that the embodiments of the application can include devices having architectures different from Figure 3 The above architecture is only exemplary and is used to explain the information exchange method 600 provided by the embodiments of the application.

[0080] The at least one processor 102 can include a microprocessor, an application-specific integrated circuit (ASIC), a digital signal processor (DSP), a central processing unit (CPU), a graphics processing unit (GPU), a state machine, or the like. Embodiments of the computer-readable medium include, but are not limited to, a floppy disk, a CD-ROM, a magnetic disk, a memory chip, a ROM, a RAM, an ASIC, a configured processor, an all-optical medium, all magnetic tapes or other magnetic media, or any other medium from which a computer processor can read instructions. In addition, various other forms of computer-readable media can send or carry instructions to a computer, including routers, special-purpose or public networks, or other wired and wireless transmission devices or channels. The instructions can include code in any computer programming language, including C, C++, C language, Visual Basic, java, and JavaScript.

[0081] Figure 4The structure of the central node 20 provided by the embodiment of the present application is shown. At least one memory 201 can be included, which comprises a computer readable medium, such as a random access memory (RAM); in addition, the central node 20 can also include at least one processor 202 coupled with the at least one memory 201. Computer executable instructions are stored in the at least one memory 201, and when executed by the at least one processor 202, can cause the at least one processor 202 to perform the steps described herein.

[0082] Figure 4 The at least one memory 201 shown in the middle can contain an information interaction program 40, so that the at least one processor 202 executes the information exchange method 700 described in the embodiment of the present application. As shown, the information interaction program 40 can include: Figure 4

[0083] - an information interaction module 401 configured to send at least one of the features of part or all of the second target to the first camera 10 in response to a request of the first camera 10 in the MTMCT system 100, and receive a notification from the first camera 10 that a second target has been tracked by the first camera 10;

[0084] - an information processing module 402 configured to mark the notified second target as having been tracked by the first camera 10.

[0085] The other functions of the above-mentioned modules can refer to the processing of the central node 20 side in the aforementioned method 501 and method 502, which will not be repeated here. Each module can also be considered as a functional module realized by hardware, for realizing various functions involved in the execution of the information exchange method by the central node 20, such as pre-burning the control logic of each flow involved in the method into a chip such as a Field-Programmable Gate Array (FPGA) or a Complex Programmable Logic Device (CPLD), and executing the functions of the above-mentioned modules by the chip or device. The specific implementation manner can be determined according to engineering practice.

[0086] In addition, the central node 20 can also include a communication interface 203 for communication between the central node 20 and other devices, such as each camera 10.

[0087] It should be mentioned that the embodiment of the present application can include devices with architectures different from Figure 4 The above-mentioned architecture is only exemplary and is used to explain the information exchange method 700 provided by the embodiment of the present application.

[0088] ​The at least one processor 202 can include a microprocessor, an application-specific integrated circuit (ASIC), a digital signal processor (DSP), a central processing unit (CPU), a graphics processing unit (GPU), a state machine, or the like. Embodiments of the computer-readable medium include, but are not limited to, a floppy disk, a CD-ROM, a magnetic disk, a memory chip, a ROM, a RAM, an ASIC, a configured processor, an all-optical medium, all magnetic tapes or other magnetic media, or any other medium from which a computer processor can read instructions. In addition, various other forms of computer-readable media can send or carry instructions to a computer, including routers, special-purpose or public networks, or other wired and wireless transmission devices or channels. Instructions can include code in any computer programming language, including C, C++, C language, Visual Basic, java, and JavaScript.

[0089] The information exchange method 600 provided by the embodiments of the application will be described below in combination with Figure 6 The information exchange method 600 provided by the embodiments of the application will be described below in combination with

[0090] S601: A camera 10 of the MTMCT system 100 performs target recognition and feature extraction on a frame of image collected by itself.

[0091] S602: The camera 10 matches the extracted feature with the feature of at least one first target tracked by itself, if not matched, step S603 is performed, if matched, step S608 is performed.

[0092] S603: Obtain the feature of at least one second target lost from the central node 20 of the MTMCT system 100.

[0093] S604: Match the extracted feature with the feature of the at least one second target, if matched successfully, steps S605-S607 are performed, otherwise, step S609 is performed.

[0094] S605: Set the identification ID of the recognized target as the ID of the second target matched successfully.

[0095] S606: Inform the central node 20 that the second target matched successfully has been tracked by the camera 10.

[0096] S607: Merge the extracted feature with the feature of the second target matched successfully.

[0097] S608: Merge the extracted feature with the feature of the matched first target.

[0098] S609: Request the central node 20 to allocate a new ID for the recognized target.

[0099] In addition, for each third target tracked by the camera 10, the camera 10 can further calculate, by step S610, a time length from a time point when the third target was last identified to a current time point for each third target tracked by the camera 10. If the time length of a third target is greater than a preset threshold, the following steps S611-S612 are performed.

[0100] S611: sending the ID of the third target exceeding the preset threshold to the central node 20 to mark as lost.

[0101] S612: sending the features of the third target exceeding the preset threshold to the central node 20.

[0102] Other steps and implementation manners performed by the camera 10 can refer to Figure 5A and Figure 5B and corresponding descriptions, which will not be repeated here.

[0103] The information exchange method 700 provided by the embodiments of the present application will be described below in combination with Figure 7 The method can be performed by the central node 20 described above, and the method can include the following steps:

[0104] S701: the central node 20 sends the features of part or all of the at least one second target to the first camera 10 in response to a request of the first camera 10 in the MTMCT system 100;

[0105] S702: receiving a notification from the first camera 10 that a second target has been tracked by the first camera 10;

[0106] S703: marking the notified second target as having been tracked by the first camera 10.

[0107] The method 700 can further include the following steps:

[0108] S704: receiving a request from the second camera 10 in the MTMCT system 100 to assign an ID to a newly identified target;

[0109] S705: assigning an ID to the newly identified target;

[0110] S706: returning the assigned ID to the second camera 10.

[0111] The method 700 can further include the following steps:

[0112] S707: receiving a notification from the third camera 10 in the MTMCT system 100 indicating that a third target has been lost;

[0113] S708: recording the ID of the third target;

[0114] S709: receiving and recording the features of the third target from the third camera 10.

[0115] Other steps and implementation manners performed by the central node 20 can refer to the descriptions of the steps and implementation manners performed by the central node 20 in the above-mentioned embodiments, which will not be repeated here. Figure 5A and Figure 5B and the corresponding descriptions, which will not be repeated here.

[0116] In addition, the embodiments of the present application also provide a computer readable medium, which stores computer readable instructions. When the computer readable instructions are executed by a processor, the processor executes the information exchange method described above. Embodiments of the computer readable medium include floppy disks, hard disks, magneto-optical disks, optical disks (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), magnetic tapes, non-volatile memory cards and ROM. Alternatively, the computer readable instructions can be downloaded from a server computer or the cloud through a communication network.

[0117] It should be noted that not all steps and modules in the above-mentioned flowcharts and system structure diagrams are necessary, and some steps or modules can be ignored according to actual needs. The execution order of each step is not fixed and can be adjusted according to needs. The system structure described in each of the above-mentioned embodiments can be a physical structure or a logical structure, that is, some modules can be implemented by the same physical entity, or some modules can be implemented by multiple physical entities, or some modules can be implemented by some components in multiple independent devices.

Claims

1. An information exchange method (600) for use in an MTMCT system (100), wherein the areas covered by each camera (10) in the MTMCT system (100) do not overlap, characterized in that, include: - One camera (10) of the MTMCT system (100) performs target recognition and feature extraction on a frame of image it acquires (S601); and maintains the identifier ID and features of each target it has tracked; - The camera (10) will match the extracted features with the features of at least one first target it has been tracking (S602). If there is no match, then - Obtain (S603) the features of at least one lost second target from the central node (20) of the MTMCT system (100); - Match the extracted features with the features of at least one second target (S604); -If the match is successful, then - Set the identifier ID of the identified target (S605) to the ID of the successfully matched second target, and - Notification (S606) that the second target successfully matched by the central node (20) has been tracked by the camera (10); - The extracted features are merged with the features of the successfully matched second target (S607); the central node (20) removes the ID and features of the successfully matched second target from the maintained list of missing targets; the camera (10) maintains the ID and features of the second target.

2. The method as described in claim 1, characterized in that, Also includes: If the features extracted by the camera (10) match the features of a first target that it has already tracked, then -Merge the extracted features with the features of the first matched target (S608).

3. The method as described in claim 1, characterized in that, Also includes: If the features extracted by the camera (10) do not match the features of any of the tracked first targets, and the extracted features cannot match the features of any of the lost second targets, then -Request (S609) the central node (20) to assign a new ID to the identified target.

4. The method as described in claim 1, characterized in that, Also includes: - For each third target that the camera (10) has tracked, the camera (10) calculates (S610) the time elapsed since the last time the third target was identified; -If the duration of a third target exceeds a preset threshold, then - Send (S611) the ID of the third target exceeding the preset threshold to the central node (20) to mark it as lost, and - The features of the third target that exceed the preset threshold are sent (S612) to the central node (20).

5. A camera (10) in an MTMCT system (100), wherein the areas covered by each camera (10) in the MTMCT system (100) do not overlap, characterized in that, include: - The image processing module (301) is configured to perform target recognition and feature extraction on a frame of image captured by the camera (10); and to maintain the identifier ID and features of each target it has tracked; - Feature matching module (302) is configured to match features extracted by image processing module (301) with features of at least one first target tracked by camera (10); - The information interaction module (303) is configured to: if the features extracted by the image processing module (301) do not match the features of at least one first target tracked by the camera (10), then obtain the features of at least one lost second target from the central node (20) of the MTMCT system (100); -The feature matching module (302) is further configured to match the extracted features with features of at least one second target; The information interaction module (303) is further configured to: if the features extracted by the image processing module (301) successfully match the features of at least one second target, then - Set the identifier ID of the identified target to the ID of the successfully matched second target. -Notify the central node (20) that the second target that has been successfully matched has been tracked by the camera (10), and -Merge the extracted features with the features of the successfully matched second target; The central node (20) removes the ID and features of the successfully matched second target from the maintained list of missing targets; the camera (10) maintains the ID and features of the second target.

6. The camera (10) as described in claim 5, characterized in that, The information interaction module (303) is also configured to: If the features extracted by the image processing module (301) match the features of a first target that has been tracked by the camera (10), then the features extracted by the image processing module (301) are merged with the features of the matched first target.

7. The camera (10) as described in claim 5, characterized in that, The information interaction module (303) is further configured to: if the features extracted by the image processing module (301) do not match the features of any first target tracked by the camera (10), and the extracted features cannot match the features of any lost second target, then - Request the central node (20) to assign a new ID to the identified target.

8. The camera (10) as described in claim 5, characterized in that, Also includes: The information interaction module (303) is further configured as follows: - For each third target that the camera (10) has tracked, calculate the time elapsed since the last time the camera (10) identified the third target; -If the duration of a third target exceeds a preset threshold, then - Send the IDs of third targets exceeding the preset threshold to the central node (20) to mark them as lost, and - The features of the third target that exceed the preset threshold are sent to the central node (20).

9. An information exchange device, characterized in that, include: At least one memory (101) is configured to store computer-readable code; At least one processor (102) is configured to invoke the computer-readable code to perform the method of any one of claims 1 to 4.

10. A computer-readable medium, characterized in that, The computer-readable medium stores computer-readable instructions that, when executed by a processor, cause the processor to perform the method as described in any one of claims 1 to 4.

Images