Target tracking method, computer-readable storage medium and computer device

Abstract

This application relates to the field of video processing and provides a target tracking method, a computer-readable storage medium, and a computer device. The method includes: matching a target tracking frame with a target candidate frame to determine a target candidate frame that best matches the target tracking frame; matching one or more remaining target candidate frames, excluding the best-matching target candidate frame, with a second target candidate frame detected in a previous video frame to determine a corresponding matching relationship; obtaining a distance and overlap relationship between the best-matching target candidate frame and each of the one or more remaining target candidate frames and the second target candidate frame according to the corresponding matching relationship between the best-matching target candidate frame and the second target candidate frame, thereby determining an occlusion relationship between the target and other objects in the current video frame; and determining whether to resume target tracking according to the occlusion relationship. The present invention can improve the robustness of long-term tracking of a single target, avoid erroneous results caused by interference of the tracked target with other objects, ensure the accuracy of the tracked target, and prevent false detection from occurring.

Description

[Technical field]

[0001] The present application relates to the field of video processing, and in particular to a target tracking method, a computer-readable storage medium and a computer device. [Background technology]

[0002] The occlusion problem is a very serious problem in target tracking, and currently many target tracking algorithms have proposed several methods to deal with the occlusion problem. For example, the TLD algorithm (Tracking-Learning-Detection) is an architecture for long-term tracking of unknown objects in video, and the TLD algorithm consists of three parts: a tracker, a detector, and a learner. The tracker is for observing the moving direction of the target between video frames; the detector regards each video frame as independent and performs positioning; the learner evaluates the error of the detector according to the tracker's result, generates training samples to update the target model of the detector, and prevents similar errors from appearing in the future. Summary of the Invention [Problem to be solved by the invention]

[0003] However, prior art target tracking algorithms cannot accurately track targets when applied to scenes where the target is occluded by other objects (e.g., people, animals, vehicles, etc.) during single target tracking. [Means for solving the problem]

[0004] To solve one of the above problems, embodiments of the present application provide a target tracking method, an apparatus, a computer-readable storage medium, a computer device, and a camera.

[0005] In a first aspect, an embodiment of the present application provides a method for tracking an object, the method comprising:

[0006] S101. Obtaining a target tracking frame tracked in a current video frame and a target candidate frame detected in the current video frame;

[0007] S102. Matching the target tracking frame with the target candidate frame, and determining a target candidate frame that best matches the target tracking frame;

[0008] S103. Matching one or more remaining target candidate frames, excluding the most matching target candidate frame, among the target candidate frames with a second target candidate frame detected in a previously detected video frame, to determine a corresponding matching relationship;

[0009] S104. According to the best-matching target candidate frame and the corresponding matching relationship, obtain a distance and overlap relationship between the best-matching target candidate frame and each of the one or more remaining target candidate frames and the second target candidate frame, and determine an occlusion relationship between the target and other objects in the current video frame;

[0010] S105. The method includes determining whether to resume target tracking according to an occlusion relationship between the target and other objects.

[0011] In a second aspect, the present application provides a target tracking device, the device comprising:

[0012] an acquisition module for acquiring a target tracking frame tracked in the current video frame and a target candidate frame detected in the current video frame;

[0013] a target candidate frame determination module for matching the target tracking frame with the target candidate frames and determining a target candidate frame that best matches the target tracking frame;

[0014] a matching relationship determination module for matching one or more remaining target candidate frames, excluding the most matching target candidate frame, among the target candidate frames and a second target candidate frame detected in a previously detected video frame, and determining a corresponding matching relationship;

[0015] an occlusion relationship determination module for obtaining a distance and an overlap relationship between the best-matching target candidate frame and each of the one or more remaining target candidate frames, the second target candidate frame according to the best-matching target candidate frame and the corresponding matching relationship, and determining an occlusion relationship between the target and other objects in the current video frame;

[0016] A decision module is provided for deciding whether to resume target tracking depending on an occlusion relationship between the target and other objects.

[0017] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored therein, the computer program implementing the steps of the target tracking method as described above when executed by a processor.

[0018] In a fourth aspect, the present application provides a computing device, the computing device comprising:

[0019] one or more processors;

[0020] Memory and;

[0021] The method includes one or more computer programs, the processor and the memory are connected by a bus, the one or more computer programs are stored in the memory and arranged to be executed by the one or more processors, and when the processor executes the computer programs, the steps of the target tracking method as described above are realized.

[0022] In a fifth aspect, the present application provides a camera, the camera comprising:

[0023] one or more processors;

[0024] Memory and;

[0025] The method includes one or more computer programs, the processor and the memory are connected by a bus, the one or more computer programs are stored in the memory and arranged to be executed by the one or more processors, and when the processor executes the computer programs, the steps of the target tracking method as described above are realized. Effect of the Invention

[0026] In the embodiment of the present application, one or more remaining target candidate frames, excluding the most matching target candidate frame among the target candidate frames, are matched with a second target candidate frame detected in a previously detected video frame, and corresponding matching relationships are determined; according to the most matching target candidate frame and the corresponding matching relationships, a distance and overlap relationship is obtained between the most matching target candidate frame and each of the one or more remaining target candidate frames and the second target candidate frame, and an occlusion relationship between the target and other objects in the current video frame is determined; whether to resume target tracking is determined according to the occlusion relationship between the target and other objects, so as to improve the robustness of long-term tracking of a single target and avoid erroneous results caused by the interference of the tracked target with other objects; if the target is occluded by other objects, tracking is resumed and tracking of the tracked target is restored, so as to ensure the accuracy of the tracked target and prevent erroneous detection from occurring. [Brief description of the drawings]

[0027] [Figure 1] FIG. 2 is a schematic diagram of an application scenario of the target tracking method provided in an embodiment of the present application;

[0028] [Diagram 2]1 is a flowchart of a target tracking method provided in an embodiment of the present application.

[0029] [Diagram 3] FIG. 2 is a schematic diagram of a target tracking frame tracked in a current video frame;

[0030] [Figure 4] FIG. 2 is a schematic diagram of target candidate frames detected in a current video frame;

[0031] [Diagram 5] FIG. 2 is a schematic diagram of an area of ​​interest.

[0032] [Figure 6] FIG. 2 is a schematic diagram showing another object approaching a target.

[0033] [Figure 7] FIG. 1 is a schematic diagram showing a target being occluded by another object.

[0034] [Figure 8] 4 is a schematic diagram of a target candidate frame that best matches the target tracking frame; FIG.

[0035] [Figure 9] 2 is a schematic diagram of a video frame following a current video frame; FIG.

[0036] [Figure 10] 1 is a schematic diagram of a target tracking device provided in one embodiment of the present application;

[0037] [Figure 11] FIG. 2 is a block diagram of a specific configuration of a computer device provided in an embodiment of the present application.

[0038] [Figure 12] FIG. 2 is a block diagram showing a specific configuration of a camera provided in an embodiment of the present application. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0039] In order to clarify the purpose, technical solution and effects of the present invention, the present application will be described in more detail below with reference to the drawings and examples. It should be understood that the specific examples described herein are only for interpreting the present application, and are not intended to limit the present application.

[0040] In order to illustrate the technical solutions described in this application, specific examples are given below.

[0041] An application scenario of the target tracking method provided in an embodiment of the present application may be a computer device or a camera, and the computer device or the camera executes the target tracking method provided in an embodiment of the present application to track a target in a video. The application scenario of the target tracking method provided in an embodiment of the present application may include a computer device 100 and a camera 200 connected to each other (as shown in FIG. 1). At least one application program may be executed in the computer device 100 and the camera 200. The computer device 100 may be a server, a desktop computer, a mobile terminal, etc., and the mobile terminal includes a mobile phone, a tablet computer, a notebook computer, a personal digital assistant, etc. The camera 200 may be a normal camera, a panoramic camera, etc. An ordinary camera is a photographing device for taking planar images or planar videos. The computer device 100 or the camera 200 executes the target tracking method provided in an embodiment of the present application to track a target in a video.

[0042] Referring to FIG. 2, FIG. 2 is a flowchart of a target tracking method provided in one embodiment of the present application. In this embodiment, the target tracking method is mainly described by taking the application of the target tracking method to a computer device or a camera as an example. The target tracking method provided in one embodiment of the present application includes the following steps:

[0043] S101. Obtain a target tracking frame tracked in a current video frame and a target candidate frame detected in the current video frame.

[0044] The target tracking frame looks like the rectangular frame in FIG. 3; the target candidate frame looks like the rectangular frame in FIG.

[0045] In one embodiment of the present application, the target tracking frame tracked in the current video frame may be specifically obtained by tracking a target in the video using a tracker.

[0046] The target candidate frame detected in the current video frame may specifically be a target candidate frame detected in the current video frame by invoking a detector.

[0047] In one embodiment of the present application, the target candidate frame detected in the current video frame is specifically:

[0048] The target candidate frame may be detected in the current video frame by performing one detection every predetermined frame interval. By performing one detection every predetermined frame interval, the processing time of the detection algorithm can be saved. Of course, in order to improve the accuracy of detection, it is also possible to set one detection for each frame of the video frame, and the present application does not specifically limit it.

[0049] In one embodiment of the present application, the video may be a panoramic video or a planar video, or may be other video formats, but the present application does not specifically limit the same.

[0050] As shown in FIG. 3, the target tracking frame tracked in the current video frame is the target tracking frame 31 in FIG. 3, and the target candidate frames detected in the current video frame are five target candidate frames, namely, target candidate frames 41, 42, 43, 44, and 45 in FIG. 4, and the number of the target candidate frames may be multiple.

[0051] S102, matching the target tracking frame with the target candidate frame, and determining a target candidate frame that is most matched with the target tracking frame.

[0052] In one embodiment of the present application, S102 may specifically be as follows:

[0053] The target tracking frame and the target candidate frame are respectively matched, specifically, matched using IoU (Intersection over Union), and an IoU value corresponding to each target candidate frame can be obtained; IoU is a criterion for measuring the accuracy of detecting a corresponding object in a specific data set, and IoU is a simple measurement criterion, and any task that can obtain a prediction range in the output can be measured by IoU;

[0054] Specifically, the target candidate frame that has the largest corresponding IoU value and satisfies a preset threshold can be determined as the target candidate frame that best matches the target tracking frame.

[0055] S103. Match one or more remaining target candidate frames (e.g., all or part of the remaining target candidate frames) among the target candidate frames, excluding the most matching target candidate frame, with a second target candidate frame detected in the previously detected video frame, and determine a corresponding matching relationship.

[0056] In one embodiment of the present application, S103 may specifically be as follows:

[0057] Among the target candidate frames, one or more remaining target candidate frames, excluding the most matching target candidate frame, are matched with a second target candidate frame detected in a previously detected video frame, and specifically, the matching is performed using IoU (Intersection over Union), and a corresponding matching relationship between the one or more remaining target candidate frames and the second target candidate frame can be determined using a Hungarian matching algorithm (Munkres Assignment Algorithm).

[0058] For example, the multiple remaining target candidate frames in the current video frame (the tth video frame) are A1, A2, A3, and A4.

[0059] The target candidate frames detected in the previously detected video frame (for example, the t-1th video frame or the t-3rd video frame, and more specifically, determined according to a specified frame interval at the time of detection, if the specified frame interval is 1, the previous video frame is the t-1th video frame, and if the specified frame interval is 3, the previous video frame is the t-3rd video frame) are B1, B2, B3, and B4.

[0060] The corresponding matching relations are A1-B1, A2-B2, A3-B3, and A4-B4. Four targets appear simultaneously in the two video frames, and the target candidate frames in the two video frames are related.

[0061] In one embodiment of the present application, S103 may specifically be as follows:

[0062] The area expanded by a predetermined ratio with the target tracking frame as the center is set as an area of ​​interest (for example, block area 32 in FIG. 5), and one or more remaining target candidate frames, excluding the most matching target candidate frame located in the area of ​​interest among the target candidate frames, are matched with a second target candidate frame detected in a previously detected video frame, specifically, the matching is performed using IoU (Intersection over Union), and the corresponding matching relationship between the one or more remaining target candidate frames and the second target candidate frame can be determined using a Hungarian matching algorithm (Munkres Assignment Algorithm). Since only the remaining target candidate frames located in the area of ​​interest are matched, the amount of calculation of the algorithm can be reduced.

[0063] For example, the multiple remaining target candidate frames in the area of ​​interest in the current video frame (the tth video frame) are A1, A2, A3, A4.

[0064] The second target candidate frames detected in the previously detected video frame (for example, the t-1th video frame or the t-3rd video frame, and more specifically, are determined according to a specified frame interval at the time of detection, and if the specified frame interval is 1, the previous video frame is the t-1th video frame, and if the specified frame interval is 3, the previous video frame is the t-3rd video frame) are B1, B2, B3, and B4.

[0065] The corresponding matching relations are A1-B1, A2-B2, A3-B3, and A4-B4. Four targets appear simultaneously in the two video frames, and the target candidate frames in the two video frames are related.

[0066] S104. According to the best-matching target candidate frame and the corresponding matching relationship, obtain a distance and overlap relationship between the best-matching target candidate frame and each of the one or more remaining target candidate frames and the second target candidate frame, and determine an occlusion relationship between the target and other objects in the current video frame.

[0067] The occlusion relationship may include the target not being occluded by another object, the other object moving away from the target, the other object moving closer to the target, and the target being occluded by the other object. For example, Fig. 6 shows that the other object is moving closer to the target, and Fig. 7 shows that the target is occluded by the other object. The appearance of a state in which the other object is moving closer to the target indicates that there is a high probability that the target will be occluded by the other object at the next detection time point.

[0068] In one embodiment of the present application, if the occlusion relationship between the target and other objects in the current video frame is that the other objects are approaching the target or the target is occluded by the other objects, after S104, the method further includes the steps of:

[0069] Whether the target is occluded by another object is determined according to the reliability of the target tracking. For example, if the reliability of the target tracking is less than a preset threshold (e.g., 30%, 40%, etc.), it is determined that the target is occluded by another object.

[0070] S105, determining whether to resume target tracking according to the occlusion relationship between the target and other objects.

[0071] In one embodiment of the present application, S105 may specifically be as follows:

[0072] If the target is not occluded by another object or if the other object moves away from the target, tracking continues; if the target is occluded by another object, tracking is resumed and tracking of the tracked target is regained.

[0073] In one embodiment of the present application, the above-mentioned resuming of tracking and recovering tracking of the tracking target is specifically performed by using a Person Re-identification (ReID) network or an image similarity matching algorithm to resume tracking and recover tracking of the tracking target.

[0074] The ReID (Person Re-identification) network is a technology that uses computer vision techniques to determine whether a particular pedestrian is present in an image or video sequence.

[0075] In another embodiment of the present application, the above-mentioned resuming tracking and restoring tracking to the tracking target may specifically include the following steps:

[0076] Detecting a target candidate frame in a video frame next to the current video frame, the next video frame being a video frame that will be detected next by the detector;

[0077] A candidate result frame is determined according to the relationship of dimensions and distance between the target candidate frame detected in the next video frame and the best matching target candidate frame; specifically, there may be n or less candidate result frames, n>=2, for example n=5; (for example, if the dimensions of two target candidate frames detected in the next video frame are both 5*4cm, and the dimensions of the best matching target candidate frame are also 5*4cm, the distance between one of the two target candidate frames detected in the next video frame and the best matching target candidate frame is 30cm, and the distance between the other one and the best matching target candidate frame is 20cm, then both of these two target candidate frames can be candidate result frames).

[0078] The next video frame of the current video frame and the candidate result frame are input into the ReID network to perform forward inference to reacquire the tracking target, and if the tracking target is acquired, tracking is resumed; otherwise, the above-mentioned steps of resuming tracking and recovering tracking for the tracking target are continuously performed until the recovery fails after exceeding the maximum number of attempts.

[0079] In one embodiment of the present application, the forward inference is specifically as shown in Figures 8 and 9, where Figure 8 is the current video frame, the rectangular frame in Figure 8 is the most matching target candidate frame, Figure 9 is the next video frame of the current video frame, and the rectangular frame in Figure 9 is the candidate result frame; the next video frame of the current video frame and the candidate result frame are input into the ReID network to obtain a similarity matrix with the most matching target candidate frame; from the input candidate result frames, select a candidate result frame with the smallest similarity distance, and judge whether the candidate result frame with the smallest similarity distance satisfies a threshold constraint. If so, it is deemed that the tracking target is obtained, and the candidate result frame with the smallest similarity distance is used as a target tracking frame to resume tracking, thereby completing the process of recovering tracking for the tracking target.

[0080] In the embodiment of the present application, one or more remaining target candidate frames, excluding the most matching target candidate frame among the target candidate frames, are matched with a second target candidate frame detected in a previously detected video frame, and corresponding matching relationships are determined; according to the most matching target candidate frame and the corresponding matching relationships, a distance and overlap relationship is obtained between the most matching target candidate frame and each of the one or more remaining target candidate frames and the second target candidate frame, and an occlusion relationship between the target and other objects in the current video frame is determined; whether to resume target tracking is determined according to the occlusion relationship between the target and other objects, so as to improve the robustness of long-term tracking of a single target and avoid erroneous results caused by the interference of the tracked target with other objects; if the target is occluded by other objects, tracking is resumed and tracking of the tracked target is restored, so as to ensure the accuracy of the tracked target and prevent erroneous detection from occurring.

[0081] 10, the target tracking device provided in an embodiment of the present application may be a computer program or program code operated on a computer device or a camera, for example, the target tracking device is application software; the target tracking device can be used to perform corresponding steps in the target tracking method provided in the embodiment of the present application. The target tracking device provided in an embodiment of the present application includes:

[0082] an acquisition module 11 for acquiring a target tracking frame tracked in a current video frame and a target candidate frame detected in the current video frame;

[0083] a target candidate frame determination module 12 for matching the target tracking frame with the target candidate frames and determining a target candidate frame that best matches the target tracking frame;

[0084] a matching relationship determination module 13 for matching one or more remaining target candidate frames, excluding the most matching target candidate frame, among the target candidate frames and a second target candidate frame detected in a previously detected video frame, and determining a corresponding matching relationship;

[0085] an occlusion relationship determination module 14 for obtaining a distance and overlap relationship between the best-matching target candidate frame and each of the one or more remaining target candidate frames, the second target candidate frame according to the best-matching target candidate frame and the corresponding matching relationship, and determining an occlusion relationship between the target and other objects in the current video frame;

[0086] The apparatus further comprises a decision module 15 for deciding whether to resume target tracking depending on an occlusion relationship between the target and other objects.

[0087] The target tracking device provided in one embodiment of the present application belongs to the same concept as the target tracking method provided in one embodiment of the present application, and the details of the specific implementation process thereof may be referred to the entire specification, so they will not be described in detail here.

[0088] In one embodiment of the present application, a computer-readable storage medium is further provided, the computer-readable storage medium storing a computer program, which, when executed by a processor, realizes the steps of the target tracking method as provided in one embodiment of the present application.

[0089] FIG. 11 shows a specific configuration block diagram of a computer device provided in one embodiment of the present application, which may be the computer device shown in FIG. 1, where the computer device 100 includes one or more processors 101; a memory 102; and one or more computer programs, the processor 101 and the memory 102 are connected by a bus, the one or more computer programs are stored in the memory 102 and arranged to be executed by the one or more processors 101, and when the processor 101 executes the computer programs, it realizes the steps of the target tracking method provided in one embodiment of the present application.

[0090] The computing device may be a desktop computer, a mobile terminal, etc., including a mobile phone, a tablet computer, a laptop, a personal digital assistant, etc.

[0091] FIG. 12 shows a specific configuration block diagram of a camera provided in one embodiment of the present application, which may be the camera shown in FIG. 1. The camera 200 includes one or more processors 201; a memory 202; and one or more computer programs, the processor 201 and the memory 202 are connected by a bus, the one or more computer programs are stored in the memory 202 and arranged to be executed by the one or more processors 201, and when the processor 201 executes the computer programs, it realizes the steps of the target tracking method provided in one embodiment of the present application.

[0092] It should be understood that the steps in each embodiment of the present application do not necessarily have to be performed sequentially according to the order indicated by the step numbers. Unless expressly stated herein, there is no strict order restriction on the execution of these steps, and these steps may be performed in other orders. Furthermore, in each embodiment, at least some of the steps may include multiple sub-steps or multiple stages, and these sub-steps or stages do not necessarily have to be performed or completed at the same time, but may be performed at different times, and the order of execution of these sub-steps or stages does not necessarily have to be sequential, but may be performed in order or alternating with other steps or at least some of the sub-steps or stages of other steps.

[0093] As will be understood by those skilled in the art, all or part of the flow of the method for implementing the above-mentioned embodiment can be achieved by instructing the relevant hardware by a computer program, and the program may be stored in a computer-readable non-volatile storage medium, and when the program is executed, the flow of each of the above-mentioned method embodiments may be included. Here, any reference to memory, storage, database, or other medium used in each of the embodiments provided in this application may include non-volatile and / or volatile memory. The non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. The volatile memory may include random access memory (RAM) or external cache memory. RAM comes in various forms such as, but not limited to, Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Dual Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Memory Bus (Rambus) Direct RAM (RDRAM), Direct Memory Bus Dynamic RAM (DRDRAM), and Memory Bus Dynamic RAM (RDRAM).

[0094] The technical features of the above embodiments may be combined in any manner. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, combinations of these technical features should be considered to be within the scope of the present specification, unless contradictory.

[0095] The above examples merely show some embodiments of the present application, and although they are described relatively specifically and in detail, they cannot be understood as limiting the scope of the claims. It should be noted that a person skilled in the art can make some modifications and improvements without departing from the concept of the present application, all of which are included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be according to the scope of the attached claims.

Claims

1. S101. Using a tracker to track a target in a video, perform target tracking for the target to obtain a tracked target tracking frame in a current video frame, and using a detector to obtain a detected target candidate frame in the current video frame; S102. Matching the target tracking frame with the target candidate frame, and determining a target candidate frame that is most matched with the target tracking frame; S103. Match one or more remaining target candidate frames, excluding the most matching target candidate frame, among the target candidate frames and a second target candidate frame detected in a previously detected video frame, to determine a corresponding matching relationship; S104. According to the best-matching target candidate frame and the corresponding matching relationship, obtain a distance and overlap relationship between the best-matching target candidate frame and each of the one or more remaining target candidate frames, the second target candidate frame, and consider the target to be located in the best-matching target candidate frame and other objects to be located in the one or more remaining target candidate frames, and determine an occlusion relationship between the target and the other objects in the current video frame, including: the target is not occluded by the other objects, the other objects are moving away from the target, the other objects are moving toward the target, and the target is occluded by the other objects; S105. A method for tracking a target, comprising: determining whether to continue tracking the target or resume tracking the target after determining that the target is occluded by the other object according to an occlusion relationship between the target and the other object.

2. The target candidate frame detected in the current video frame is specifically:

2. The method for tracking a target according to claim 1, wherein the detection is performed once every predetermined frame interval, and the target candidate frame is detected in the current video frame.

3. Specifically, S102 is Matching the target tracking frame and the target candidate frame respectively using a cross ratio IoU to obtain an IoU value corresponding to each target candidate frame; The target tracking method according to claim 1 , further comprising: determining a target candidate frame that has a maximum corresponding IoU value and satisfies a preset threshold as a target candidate frame that best matches the target tracking frame.

4. Specifically, S103 is The target tracking method according to claim 1, characterized in that: using an IoU cross ratio to match one or more remaining target candidate frames, excluding the best-matching target candidate frame, among the target candidate frames with a second target candidate frame detected in a previously detected video frame; and using a Hungarian matching algorithm to determine a corresponding matching relationship between the one or more remaining target candidate frames and the second target candidate frame.

5. Specifically, S103 is The target tracking method according to claim 1, characterized in that: an area expanded by a predetermined ratio with the target tracking frame as a center is set as an area of ​​interest; one or more remaining target candidate frames, excluding the best-matching target candidate frame located in the area of ​​interest among the target candidate frames, are matched with a second target candidate frame detected in a previously detected video frame using a cross ratio IoU; and a Hungarian matching algorithm is used to determine a corresponding matching relationship between the one or more remaining target candidate frames and the second target candidate frame.

6. Specifically, S105 is continuing to track the target when the target is no longer occluded by the other object or when the other object moves away from the target; 2. The method of claim 1, further comprising: resuming said target tracking when said target is occluded by said other object, and restoring target tracking to said target.

7. The target tracking method described in claim 1, characterized in that the tracker tracks the target in the video by observing the direction of movement of the target between video frames, and the detector detects candidate target frames by treating each video frame as independent and positioning the target.

8. An acquisition module for performing target tracking for a target in a video using a tracker to acquire a tracked target tracking frame in a current video frame, and for acquiring a detected target candidate frame in the current video frame using a detector; a target candidate frame determination module for matching the target tracking frame with the target candidate frames and determining a target candidate frame that best matches the target tracking frame; a matching relationship determination module for matching one or more remaining target candidate frames, excluding the most matching target candidate frame, among the target candidate frames and a second target candidate frame detected in a previously detected video frame, and determining a corresponding matching relationship; an occlusion relationship determination module for obtaining a distance and an overlap relationship between the best-matching target candidate frame and each of the one or more remaining target candidate frames, the second target candidate frame according to the best-matching target candidate frame and the corresponding matching relationship, regarding the target as being located in the best-matching target candidate frame and the one or more remaining target candidate frames as being located in another object, and determining an occlusion relationship between the target and the other object in the current video frame, including: the target is not occluded by the other object, the other object is moving away from the target, the other object is moving toward the target, and the target is occluded by the other object; and a decision module for deciding whether to continue tracking the target or resume tracking the target after determining that the target is occluded by the other object, depending on an occlusion relationship between the target and the other object.

9. The target tracking device described in claim 8, characterized in that the tracker tracks a target in the video by observing the direction of movement of the target between video frames, and the detector detects candidate target frames by treating each video frame as independent and positioning the target.

10. A computer-readable storage medium storing a computer program, the computer program being characterized in that, when executed by a processor, the computer-readable storage medium implements the steps of the target tracking method according to any one of claims 1 to 7.

11. one or more processors; Memory; A computing device including one or more computer programs, The processor and the memory are connected by a bus, the one or more computer programs are stored in the memory and arranged to be executed by the one or more processors, and when the processor executes the computer program, the computer device realizes the steps of the target tracking method described in any one of claims 1 to 7.

12. one or more processors; Memory; A camera including one or more computer programs, The processor and the memory are connected by a bus, the one or more computer programs are stored in the memory and arranged to be executed by the one or more processors, and when the processor executes the computer program, the camera realizes the steps of the target tracking method described in any one of claims 1 to 7.

Images