Target tracking method, device, equipment and medium

Abstract

The present invention relates to a target tracking method, apparatus, device, and medium. The method comprises: acquiring a first video frame and a second video frame of a long jump area at the same time, the first video frame being captured by a first camera device deployed above the long jump area, and the second video frame being captured by a second camera device deployed to the side of the long jump area; when the first video frame satisfies a first preset condition for target tracking of a test subject in the long jump area, target tracking is performed on the test subject in the acquired second video frame to obtain a tracking result corresponding to the second video frame. Thus, the present invention can promptly detect abnormal situations in a standing long jump scenario, effectively avoid ID switching, improve the real-time and robustness of analysis and processing, and enhance the user experience.

Description

Technical Field

[0001] Embodiments of the present invention relate to the technical field of physical measurement equipment, and in particular to a target tracking method, apparatus, device, and medium. Background Art

[0002] In the standing long jump scenario, in order to provide scientific and objective operation analysis data, a depth camera (such as ZED) is usually used to track the test object.

[0003] Depth cameras typically use 3D object detection boxes for target tracking, which strongly relies on the depth information captured by the camera. However, in practical applications, the resulting depth map often contains noise and holes, influenced by environmental factors such as lighting conditions, object occlusion, and the camera's effective range. While filling these holes with noise and holes through a series of filtering techniques creates a temporally stable and dense depth map, this often changes the actual distance of the target in the image, resulting in inaccurate depth estimation. Furthermore, in standing long jump scenarios, when tracking a test subject, if the test subject overlaps with unrelated objects in the image, the estimated depth value of the test subject can be subject to significant error. This sudden depth change often causes the target tracking algorithm to fail, forcing it to track unrelated objects outside the long jump area, resulting in ID switching and inaccurate tracking of the test subject. Furthermore, incorporating human features into the target tracking algorithm and using a person re-identification model for target matching can alleviate the ID switching issue to some extent, but this significantly increases the time overhead of the target tracking algorithm and makes it difficult to guarantee real-time tracking. Summary of the Invention

[0004] In view of this, in order to solve the above technical problems, embodiments of the present invention provide a target tracking method, apparatus, device and medium.

[0005] In a first aspect, an embodiment of the present invention provides a target tracking method, comprising:

[0006] Acquire a first video frame and a second video frame of the long jump area at the same time, wherein the first video frame is captured by a first camera device deployed above the long jump area, and the second video frame is captured by a second camera device deployed to the side of the long jump area;

[0007] When the first video frame satisfies a first preset condition for target tracking of the test object in the long jump area, target tracking is performed on the test object in the acquired second video frame to obtain a tracking result corresponding to the second video frame.

[0008] In an optional embodiment, the method further includes:

[0009] capturing an initial video frame using the first camera device;

[0010] When the initial video frame satisfies a second preset condition for target tracking of the test object in the long jump area, the step of acquiring the first video frame and the second video frame of the long jump area at the same moment is performed.

[0011] In an optional embodiment, the long jump area includes a take-off area and a test area, and the initial video frame includes a first image corresponding to the take-off area and a second image corresponding to the test area;

[0012] The second preset condition includes:

[0013] A head detection frame exists in the first image and the head detection frame does not exist in the second image.

[0014] In an optional embodiment, the first video frame includes a third image corresponding to the test area;

[0015] The first preset condition includes:

[0016] The number of the head detection frames in the third image is less than a first preset threshold; or,

[0017] The head detection frame does not exist in the third image and the head detection frame does not exist in the third image corresponding to each first video frame in the target video frame set; wherein, the target video frame set is composed of a plurality of consecutive first video frames that are preceding the first video frame at the current moment.

[0018] In an optional embodiment, the second video frame includes a fourth image corresponding to the test area;

[0019] The performing target tracking on the test object in the acquired second video frame to obtain a tracking result corresponding to the second video frame includes:

[0020] Determine whether the head detection frame exists in the third image corresponding to the first video frame at the same time as the second video frame;

[0021] When the head detection frame exists, obtaining a human body detection frame in the fourth image corresponding to the second video frame according to the head detection frame, and using the human body detection frame as a tracking result corresponding to the second video frame;

[0022] When the head detection frame does not exist, the human body detection frame in the fourth image corresponding to the second video frame is determined based on the head prediction frame in the third image corresponding to the first video frame, and the determined human body detection frame is used as the tracking result corresponding to the second video frame. The head prediction frame is determined by a first preset tracker.

[0023] In an optional embodiment, the method further includes:

[0024] When there are multiple human body detection frames determined based on the head detection frame or the head prediction frame, matching the multiple human body detection frames with the human body prediction frame in the fourth image corresponding to the second video frame using a Hungarian algorithm to obtain a target matching result, wherein the human body prediction frame is determined by a second preset tracker;

[0025] According to the target matching result, a tracking result corresponding to the second video frame is obtained.

[0026] In an optional embodiment, the head detection frame in the first image is determined by:

[0027] Determine the head detection frame in the first image using a head detection model;

[0028] Determining an intersection-over-union ratio between the head detection frame and the first image;

[0029] When the intersection-over-union ratio is greater than a second preset threshold, it is determined that the head detection frame exists in the first image.

[0030] In a second aspect, an embodiment of the present invention provides a target tracking device, comprising:

[0031] an acquisition module, configured to acquire a first video frame and a second video frame of a long jump area at the same moment, wherein the first video frame is acquired by a first camera device deployed above the long jump area, and the second video frame is acquired by a second camera device deployed to the side of the long jump area;

[0032] The tracking module is used to track the test object in the acquired second video frame when the first video frame meets the first preset condition for tracking the test object in the long jump area, and obtain a tracking result corresponding to the second video frame.

[0033] In a third aspect, an embodiment of the present invention provides an electronic device, characterized in that it includes: a processor and a memory, wherein the processor is used to execute a target tracking program stored in the memory to implement the target tracking method as described above.

[0034] In a fourth aspect, an embodiment of the present invention provides a storage medium, which stores one or more programs, and the one or more programs can be executed by one or more processors to implement the target tracking method as described above.

[0035] An embodiment of the present invention provides a target tracking method, comprising: acquiring a first video frame and a second video frame of a long jump area at the same moment, the first video frame being captured by a first camera device deployed above the long jump area, and the second video frame being captured by a second camera device deployed to the side of the long jump area; when the first video frame satisfies a first preset condition for target tracking of a test subject in the long jump area, target tracking is performed on the test subject in the acquired second video frame to obtain a tracking result corresponding to the second video frame. When tracking a test subject in a standing long jump scenario, the embodiment of the present invention uses a video stream of the long jump area captured by cameras located at different positions in the long jump area, uses one video stream to judge tracking anomalies, and uses another video stream to track the test subject. This allows for timely detection of anomalies in the standing long jump scenario, effectively avoids ID switching, improves the real-time and robustness of analysis and processing, and enhances user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A schematic diagram of an application scenario of a target tracking method provided by an embodiment of the present invention;

[0037] Figure 2 A flowchart of a target tracking method provided by an embodiment of the present invention;

[0038] Figure 3 A flowchart of another target tracking method provided by an embodiment of the present invention;

[0039] Figure 4 A flowchart of another target tracking method provided by an embodiment of the present invention;

[0040] Figure 5 A schematic diagram of an application scenario of another target tracking method provided by an embodiment of the present invention;

[0041] Figure 6 A schematic structural diagram of a target tracking device provided by an embodiment of the present invention;

[0042] Figure 7 A schematic structural diagram of an electronic device provided in an embodiment of the present invention;

[0043] In the above attached figures:

[0044] 10. Acquisition module; 20. Tracking module;

[0045] 100, first camera device; 200, second camera device; 300, edge computing server;

[0046] 500. Electronic device; 501. Processor; 502. Memory; 5021. Operating system; 5022. Application; 503. User interface; 504. Network interface; 505. Bus system. DETAILED DESCRIPTION

[0047] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0048] To facilitate understanding of the embodiments of the present invention, specific embodiments will be further explained below with reference to the accompanying drawings. The embodiments do not limit the embodiments of the present invention.

[0049] refer to Figure 1 , Figure 1 A schematic diagram of an application scenario of a target tracking method provided by an embodiment of the present invention. The application scenario provided by an embodiment of the present invention includes: two camera devices and an edge computing server 300, the two camera devices are a first camera device 100 and a second camera device 200, the first camera device 100 is a spherical camera, the second camera device 200 is an ordinary camera, and the first camera device 100 and the second camera device 200 are both communicatively connected to the edge computing server 300. The first camera device 100 is deployed above the long jump area, and the second camera device is deployed to the side of the long jump area. The field of view of the first camera device 100 and the second camera device 200 can cover the entire long jump area. The first camera device 100 and the second camera device 200 are used to collect video streams for the long jump area, and send the collected video streams to the edge computing server 300 for target tracking of the test object in the long jump area.

[0050] refer to Figure 2 , Figure 2 A flow chart of a target tracking method provided by an embodiment of the present invention. A target tracking method provided by an embodiment of the present invention includes the following steps:

[0051] S101: Acquire a first video frame and a second video frame of a long jump area at the same moment.

[0052] The first video frame is captured by a first camera positioned above the long jump area, and the second video frame is captured by a second camera positioned to the side of the long jump area. In this embodiment, the first video frame is used to determine if there are any anomalies in the target tracking process, and the second video frame is used to track the test subject in the second video frame based on the determination result of the first video frame. To accurately track the test subject in the long jump area, the first and second video frames must be captured at the same time.

[0053] In this embodiment, in order to ensure the real-time performance of target tracking, the edge computing server establishes two threads, and the edge computing server obtains the first video frame and the second video frame respectively according to the two threads. The two threads are the first thread T A and the second thread T B , the first thread T A The second thread T is used to obtain the first video frame captured by the first camera device. B Used to obtain a second video frame captured by a second device. Specifically, the first camera device uses a bird's-eye view, and there is no target occlusion in the XZ dimension of the spatial coordinate system, but it can detect possible occlusion in the XY dimension. Therefore, the first video frame captured by the first camera device is used to determine abnormalities in the target tracking process, and the second video frame in the XY dimension captured by the second camera device is used to track the test subject in the long jump area.

[0054] S102: Determine whether the first video frame meets a first preset condition for target tracking of the test subject in the long jump area.

[0055] During target tracking for the test subject, the first preset condition is used to determine whether an abnormality occurs in the long jump area. An abnormality in the long jump area can be considered as the detection of an object unrelated to the test subject. Alternatively, an abnormality in the long jump area can be considered as the continued failure of the test subject to be detected in the long jump area. The first preset condition is described below and is not further elaborated in this embodiment.

[0056] In this embodiment, in order to further ensure the real-time performance of target tracking, target tracking and data processing are processed in separate threads. A third thread T is also established in the edge computing server, and the third thread T is used to A The first video frame collected and the second thread T B The collected second video frame is subjected to data processing, that is, the third thread T is used to determine whether the first preset condition for target tracking of the test object in the long jump area is currently met based on the first video frame, and to execute subsequent processing steps according to the determination result.

[0057] S103: When the first preset condition is met, target tracking is performed on the test object in the acquired second video frame to obtain a tracking result corresponding to the second video frame.

[0058] Among them, step S103 is executed by the third thread T. In this embodiment, for the first video frame and the second video frame at the same time, target tracking is performed on the test object in the second video frame only when the first video frame meets the first preset condition. By setting the first preset condition, when an irrelevant person appears in the detected long jump area (i.e., an abnormal situation), target tracking of the test object is stopped, thereby avoiding the problem of target tracking errors caused by ID switching when there are multiple people in the test area in the prior art, ensuring the real-time and accuracy of target tracking, and eliminating the need to limit the tracking scene (such as there cannot be too many irrelevant objects and object occlusion in the long jump area), thereby improving the robustness of the analysis and processing.

[0059] It should be noted that after step S103 is completed, the process proceeds to step S101 from the next moment. The determination of the tracking result corresponding to the second video frame will be described below and will not be elaborated on in detail in this embodiment.

[0060] S104: When the first preset condition is not met, target tracking of the test object in the acquired second video frame is stopped, and starting from the next moment, the steps of acquiring the first video frame and the second video frame of the long jump area at the same moment are continued.

[0061] Among them, step S104 is executed by the third thread T. In this embodiment, for the first video frame and the second video frame at the same time, when it is determined that the first video frame does not meet the first preset condition, target tracking of the test object in the second video frame is stopped, and the process proceeds to step S101 to obtain the first video frame and the second video frame at the next time, so as to target track the test object in the second video frame at the next time. When the test object performs a long jump, only the test object is allowed to be in the long jump area. When an unrelated person other than the test object is detected in the long jump area, it may be that the unrelated person has mistakenly entered the long jump area. Since it is impossible for an unrelated person to exist in the long jump area for a long time, when a first video frame does not meet the first preset condition, target tracking of the test object in the long jump area at that time can be stopped, and the tracking result of the test object in the entire long jump process will not be affected.

[0062] It should be noted that the frame represents a moment, and different frames represent different moments. For the first video frame at the current moment, the next moment represents the next first video frame adjacent to the current first video frame. Similarly, for the second video frame at the current moment, the next moment represents the next second video frame adjacent to the current second video frame.

[0063] This embodiment provides a target tracking method. When tracking a test subject in a standing long jump scenario, video streams of the long jump area are captured by cameras located at different positions in the long jump area. One video stream is used to judge tracking anomalies, and another video stream is used to track the test subject. This method can promptly detect anomalies in the standing long jump scenario, effectively avoid ID switching, improve the real-time and robustness of analysis and processing, and enhance user experience.

[0064] refer to Figure 3 , Figure 3 A flow chart of a target tracking method provided by an embodiment of the present invention. A target tracking method provided by an embodiment of the present invention includes the following steps:

[0065] S201: Capturing an initial video frame using a first camera device.

[0066] In this embodiment, the video frames captured by the first camera device include an initial video frame and a first video frame. The initial video frame is a video frame in which the test object is not tracked.

[0067] S202: Determine whether the initial video frame meets a second preset condition for target tracking of the test subject in the long jump area.

[0068] To accurately track a test subject in the long jump area, the test subject to be tracked must be determined before tracking the test subject. Once the test subject is determined, it is indicated that target tracking of the test subject is possible. The test subject to be tracked is determined based on the initial video frame and a second preset condition.

[0069] In this embodiment, the long jump area includes a take-off area and a test area, and the initial video frame includes a first image corresponding to the take-off area and a second image corresponding to the test area.

[0070] Specifically, the second precondition includes:

[0071] There is a head detection frame in the first image and there is no head detection frame in the second image.

[0072] Among them, the head detection frame in the first image is a head detection frame, which meets the second preset condition and indicates that the test object is ready and needs to be tested. At this time, target tracking of the test object is required.

[0073] In this embodiment, the head detection frame in the first image corresponding to the first video frame can be determined in the following manner:

[0074] Determine a head detection frame in the first image using a head detection model;

[0075] Determine an intersection-over-union ratio between the head detection frame and the first image;

[0076] When the intersection-over-union ratio is greater than a second preset threshold, it is determined that a head detection frame exists in the first image.

[0077] Among them, all detection boxes in the first image corresponding to the first video frame can be generated by the first thread T A The intersection-over-union ratio refers to the ratio of the intersection and union of the area of ​​the head detection frame and the first image (i.e., the area of ​​the take-off region). The second preset threshold can be set according to actual needs and is not specifically limited in this embodiment. The head detection model can be obtained using existing technologies and is not described in detail in this embodiment.

[0078] S203: When the second preset condition is met, a first video frame and a second video frame of the long jump area at the same moment are acquired.

[0079] In this embodiment, obtaining the first video frame and the second video frame of the long jump area at the same time is consistent with the above step S101, which is not described in detail in this embodiment. For details, please refer to the above description.

[0080] When it is determined that the initial video frame satisfies the second preset condition, in order to achieve target tracking of the test subject within the long jump area, a human body detection frame (i.e., a human body detection frame of the test subject) needs to be determined from the initial video frame captured by the second camera device at the same time as the initial video frame, and the video frame includes a target image corresponding to the take-off area. When the initial video captured by the first camera device satisfies the second preset condition, the human body detection frame can be determined from the target image in the following manner:

[0081] Determine all detection boxes in the target image;

[0082] Determine whether there is a human body detection frame in all detection frames based on the relative position of the head detection frame and the take-off area in the first image;

[0083] If it exists, it is determined that there is a human detection frame in the target image, and the human detection frame is used as the target tracking object. Starting from the next moment, the process returns to step S202 to obtain the first video frame and the second video frame of the long jump area at the same moment;

[0084] If it does not exist, the process returns to step S201 from the next moment.

[0085] In this embodiment, a human body detection frame can be determined from all detection frames based on the position of the head detection frame relative to the take-off area. It should be noted that when the initial video frame captured by the first camera device meets the second preset condition, if the human body detection frame does not exist in the target image corresponding to the initial video frame captured by the second camera device, it may be due to inaccurate detection. At this time, in order to determine the human body detection frame in the target image, starting from the next moment, return to execute step S201 until the human body detection frame is determined from the initial video frame captured by the second camera device. When the initial video frame captured by the first camera device meets the second preset condition and there is a human body detection frame in the target image corresponding to the initial video frame captured by the second camera device, starting from the next moment, return to execute step S203 to obtain the first video frame of the long jump area at the same moment, and the step of obtaining the second video frame.

[0086] In this embodiment, when the initial video frame captured by the first camera device meets the second preset condition and a human body detection frame exists in the target image corresponding to the initial video frame captured by the second camera device, the first preset tracker A corresponding to the first camera device and the second preset tracker B corresponding to the second camera device can be initialized. Based on the first preset tracker A and the head detection frame in the initial video frame, the head prediction frame in the first video frame at the next moment is predicted. Similarly, based on the second preset tracker B and the determined human body detection frame in the initial video frame, the human body prediction frame in the second video frame at the next moment can be predicted. It should be noted that both the first preset tracker A and the second preset tracker B use Kalman filter trackers.

[0087] S204: Determine whether the first video frame meets a first preset condition for target tracking of the test subject in the long jump area.

[0088] In this embodiment, the first video frame includes a third image corresponding to the test area, and the first preset condition includes: the number of head detection frames in the third image is less than a first preset threshold; or,

[0089] There is no head detection frame in the third image and there is no head detection frame in the third image corresponding to each first video frame in the target video frame set; wherein, the target video frame set is composed of a plurality of consecutive first video frames preceding the first video frame at the current moment, and the number of first video frames in the target video frame set is less than a third preset threshold.

[0090] The first preset threshold is two, and the third preset threshold can be set according to actual needs and is not specifically limited in this embodiment. For example, the third preset threshold can be 20, 30, etc.

[0091] S205: When the first preset condition is met, target tracking is performed on the test object in the acquired second video frame to obtain a tracking result corresponding to the second video frame.

[0092] The second video frame includes a fourth image corresponding to the test area, and target tracking is performed on the test object in the acquired second video frame to obtain a tracking result corresponding to the second video frame, including:

[0093] Determine whether there is a head detection frame in a third image corresponding to the first video frame at the same time as the second video frame;

[0094] When a head detection frame exists, determining a human body detection frame in the fourth image corresponding to the second video frame according to the head detection frame, and using the human body detection frame as a tracking result corresponding to the second video frame;

[0095] When there is no head detection frame, the human body detection frame in the fourth image corresponding to the second video frame is determined based on the head prediction frame in the third image corresponding to the first video frame, and the determined human body detection frame is used as the tracking result corresponding to the second video frame. The head prediction frame is determined by the first preset tracker.

[0096] In this embodiment, to accurately determine the human detection frame, when a head detection frame exists in the third image corresponding to the first video frame, the human detection frame in the fourth image corresponding to the second video frame can be determined based on the position of the head detection frame and the test area. When no head detection frame exists in the third image corresponding to the first video frame, the human detection frame in the fourth image corresponding to the second video frame can be determined based on the position of the head prediction frame in the third image corresponding to the first video frame and the test area.

[0097] It should be noted that the first preset tracker is a Kalman filter tracker. For the first video frame at the current moment, when there is no head detection frame in the third image corresponding to the first video frame at the previous moment, the head prediction frame determined by the first preset tracker at the previous moment is used as the head detection frame, and the head prediction frame at the current moment is determined based on the head detection frame determined at the previous moment and the first preset tracker; when there is a head detection frame in the third image corresponding to the first video frame at the previous moment, the head prediction frame at the current moment is determined based on the head detection frame at the previous moment and the first preset tracker.

[0098] In this embodiment, it can be determined whether there is a head detection frame in the third image corresponding to the first video frame by the following method:

[0099] Using a head detection model to determine a head detection frame in a third image corresponding to the first video frame;

[0100] Determine an intersection-over-union ratio between the head detection frame and the third image;

[0101] When the intersection-over-union ratio is greater than a fourth preset threshold, it is determined that a head detection frame exists in the third image corresponding to the first video frame.

[0102] Among them, all detection frames in the third image corresponding to the first video frame can be generated by the first thread T A The intersection-over-union ratio refers to the ratio of the intersection and union of the area of ​​the head detection frame and the area of ​​the third image (i.e., the area of ​​the test area). The fourth preset threshold can be set according to actual needs and is not specifically limited in this embodiment. The head detection model can be obtained using existing technologies and is not described in detail in this embodiment.

[0103] In this embodiment, when there are multiple human body detection frames determined in step S205, the tracking result corresponding to the second video frame can be obtained in the following manner:

[0104] When there are multiple human detection frames determined based on the head detection frame or the head prediction frame, the Hungarian algorithm is used to match the multiple human detection frames with the human prediction frame in the fourth image corresponding to the second video frame to obtain a target matching result, and the human prediction frame is determined by the second preset tracker;

[0105] According to the target matching result, the tracking result corresponding to the second video frame is obtained.

[0106] Among them, multiple qualified human body detection frames may be determined based on the head detection frame. In order to ensure accurate target tracking of the test subject, it is necessary to determine the human body detection frame for target tracking from multiple qualified human body detection frames. In this embodiment, the human body detection frame for target tracking is obtained by the Hungarian algorithm. After the human body detection frame for target tracking is determined, the second preset tracker B is updated based on the determined human body detection frame, and the human body prediction frame corresponding to the second video frame at the next moment is predicted based on the updated second preset tracker B. Similarly, the first preset tracker A is updated based on the head detection frame to predict the head prediction frame corresponding to the first video frame at the next moment.

[0107] S206: When the first preset condition is not met, target tracking of the test object in the acquired second video frame is stopped, and starting from the next moment, the process returns to step S203 to continue acquiring the first video frame and the second video frame of the long jump area at the same moment.

[0108] Among them, step S206 is consistent with step S104, and is not described in detail in this embodiment. Please refer to the above description for details.

[0109] S207: When the second preset condition is not met, the process returns to step S201 from the next moment.

[0110] In this embodiment, when the first video frame does not meet the second preset condition, an initial video frame of the long jump area at the next moment is acquired until an initial video frame meeting the second preset condition is determined.

[0111] This embodiment provides a target tracking method. When tracking a test subject in a standing long jump scenario, video streams of the long jump area are captured by cameras located at different positions in the long jump area. One video stream is used to judge tracking anomalies, and another video stream is used to track the test subject. This method can promptly detect anomalies in the standing long jump scenario, effectively avoid ID switching, improve the real-time and robustness of analysis and processing, and enhance user experience.

[0112] As an example, see Figure 4 and Figure 5 ,The following is a detailed introduction to the target tracking process, as follows:

[0113] To track the target, computing device C simultaneously parses and processes the two video streams as follows:

[0114] 1. Algorithm initialization, computing device C starts sub-thread T A and T B Get the video frames F captured by the camera device A respectively A and the video frame F captured by the camera device B B , and get the video frame F A Test area J in A The corresponding first image and the jump area R A The corresponding second image, video frame F B Test area J in B The corresponding third image and the take-off area R B The corresponding fourth image.

[0115] 2. Thread T responsible for camera device A A Start executing head detection, thread T responsible for camera device B B Start pedestrian detection and start another sub-thread T for the video frame F of the two threads A and F B With the detection box D A and D B synchronization and processing.

[0116] 3. If thread T recognizes video frame F A Corresponding take-off area R A There is a head in the video frame F A Corresponding test area J A If there is no head in the video frame F BCorresponding take-off area R B D in B Find the human detection frame. If not found, skip it. If found, start the Kalman filter tracker K. A and Kalman filter tracker K B .

[0117] Fourth, use the detection information to update the tracker for the video frame F A Once more than two heads are detected in the test area J A or N consecutive frames are not in the test area J A If a head is detected within the range, tracking stops. Otherwise, the tracker K A Get the predicted box P A , using tracker K B Get the predicted box P B , then according to P A Relative to test area J A The position of the detection box D B Filter out the test area J B The detection box of the corresponding position in the B Perform Hungarian matching and finally use the matched detection box to update the tracker K B , using the test area J A Update the tracker K within the head detection box A .

[0118] refer to Figure 6 , Figure 6 A schematic diagram of the structure of a target tracking device provided by an embodiment of the present invention. The target tracking device provided by an embodiment of the present invention includes an acquisition module 10 and a tracking module 20. The acquisition module 10 is used to acquire a first video frame and a second video frame of a long jump area at the same time, the first video frame being captured by a first camera device deployed above the long jump area, and the second video frame being captured by a second camera device deployed to the side of the long jump area. The tracking module 20 is used to track a test subject in the acquired second video frame when the first video frame meets a first preset condition for target tracking of the test subject in the long jump area, thereby obtaining a tracking result corresponding to the second video frame.

[0119] In this embodiment, the acquisition module 10 is further configured to:

[0120] An initial video frame is captured using a first camera device.

[0121] The target tracking device provided in this embodiment further includes a judgment module, which is configured to:

[0122] When the initial video frame meets the second preset condition for target tracking of the test object in the long jump area, the acquisition module 10 is executed.

[0123] In this embodiment, the long jump area includes a take-off area and a test area, and the initial video frame includes a first image corresponding to the take-off area and a second image corresponding to the test area;

[0124] The second precondition includes:

[0125] There is a head detection frame in the first image and there is no head detection frame in the second image.

[0126] In this embodiment, the first video frame includes a third image corresponding to the test area;

[0127] The first precondition includes:

[0128] The number of head detection frames in the third image is less than a first preset threshold; or,

[0129] There is no head detection frame in the third image and there is no head detection frame in the third image corresponding to each first video frame in the target video frame set; wherein the target video frame set is composed of multiple consecutive first video frames before the first video frame at the current moment.

[0130] In this embodiment, the second video frame includes a fourth image corresponding to the test area. The tracking module 20 is further configured to:

[0131] Determine whether there is a head detection frame in a third image corresponding to the first video frame at the same time as the second video frame;

[0132] When a head detection frame exists, obtaining a human body detection frame in the fourth image corresponding to the second video frame according to the head detection frame, and using the human body detection frame as the tracking result corresponding to the second video frame;

[0133] When there is no head detection frame, the human body detection frame in the fourth image corresponding to the second video frame is determined based on the head prediction frame in the third image corresponding to the first video frame, and the determined human body detection frame is used as the tracking result corresponding to the second video frame. The head prediction frame is determined by the first preset tracker.

[0134] In this embodiment, the tracking module 20 is further configured to:

[0135] When there are multiple human detection frames determined based on the head detection frame or the head prediction frame, the Hungarian algorithm is used to match the multiple human detection frames with the human prediction frame in the fourth image corresponding to the second video frame to obtain a target matching result, and the human prediction frame is determined by the second preset tracker;

[0136] According to the target matching result, the tracking result corresponding to the second video frame is obtained.

[0137] The target tracking device provided by this embodiment further includes a determination module, which is configured to:

[0138] Determine a head detection frame in the first image using a head detection model;

[0139] Determine an intersection-over-union ratio between the head detection frame and the first image;

[0140] When the intersection-over-union ratio is greater than a second preset threshold, it is determined that a head detection frame exists in the first image.

[0141] The target tracking device provided in this embodiment, when tracking a test subject in a standing long jump scenario, collects video streams of the long jump area based on camera devices located at different positions in the long jump area, uses one video stream to judge tracking anomalies, and uses another video stream to track the test subject. This device can promptly detect anomalies in the standing long jump scenario, effectively avoid ID switching, improve the real-time and robustness of analysis and processing, and enhance user experience.

[0142] Figure 7 A schematic structural diagram of an electronic device provided by an embodiment of the present invention. Figure 7 The electronic device 500 shown includes: at least one processor 501, a memory 502, at least one network interface 504 and another user interface 503. The various components in the electronic device 500 are coupled together via a bus system 505. It is understood that the bus system 505 is used to achieve connection and communication between these components. In addition to including a data bus, the bus system 505 also includes a power bus, a control bus and a status signal bus. However, for the sake of clarity, the bus system 505 is not shown in FIG. Figure 7 Various buses are labeled as bus system 505.

[0143] The user interface 503 may include a display, a keyboard, or a pointing device (eg, a mouse, a trackball, a touchpad, or a touch screen).

[0144] It is understood that the memory 502 in the embodiment of the present invention can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link DRAM (SLDRAM), and direct RAM bus random access memory (DRRAM). The memory 502 described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0145] In some embodiments, the memory 502 stores the following elements, executable units, or data structures, or a subset thereof, or an extended set thereof: an operating system 5021 and application programs 5022 .

[0146] The operating system 5021 includes various system programs, such as a framework layer, a core library layer, and a driver layer, for implementing various basic services and handling hardware-based tasks. Application programs 5022 include various application programs, such as a media player and a browser, for implementing various application services. Programs implementing the methods of the embodiments of the present invention may be included in application programs 5022.

[0147] In an embodiment of the present invention, by calling the program or instructions stored in the memory 502, specifically, the program or instructions stored in the application 5022, the processor 501 is used to execute the method steps provided by each method embodiment, for example, including: obtaining a first video frame and a second video frame of the long jump area at the same time, the first video frame being captured by a first camera device deployed above the long jump area, and the second video frame being captured by a second camera device deployed on the side of the long jump area; when the first video frame meets the first preset condition for target tracking of the test object in the long jump area, target tracking is performed on the test object in the acquired second video frame to obtain a tracking result corresponding to the second video frame.

[0148] The methods disclosed in the above embodiments of the present invention can be applied to or implemented by processor 501. Processor 501 may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by hardware integrated logic circuits in processor 501 or by software instructions. The above processor 501 may be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The methods, steps, and logic block diagrams disclosed in the embodiments of the present invention can be implemented or executed. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in conjunction with the embodiments of the present invention can be directly implemented and executed by a hardware decoding processor, or by a combination of hardware and software units in the decoding processor. The software units can be located in storage media well-known in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, etc. The storage medium is located in the memory 502 , and the processor 501 reads the information in the memory 502 and completes the steps of the above method in combination with its hardware.

[0149] It is understood that the embodiments described herein may be implemented using hardware, software, firmware, middleware, microcode, or a combination thereof. For hardware implementation, the processing unit may be implemented in one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions described herein, or a combination thereof.

[0150] For software implementation, the technology described herein can be implemented by a unit that performs the functions described herein. The software code can be stored in a memory and executed by a processor. The memory can be implemented in the processor or outside the processor.

[0151] The electronic device provided in this embodiment may be Figure 7 The electronic device shown in FIG. 1 can perform the following operations: Figure 2 、 Figure 3 and Figure 5 All steps of the target tracking method in the Figure 2 、 Figure 3 and Figure 5 For details on the technical effects of the target tracking method shown, please refer to Figure 2 、 Figure 3 and Figure 5 For the sake of brevity, the relevant description will not be repeated here.

[0152] An embodiment of the present invention further provides a storage medium (computer-readable storage medium). The storage medium stores one or more programs. The storage medium may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as read-only memory, flash memory, hard disk, or solid-state drive; and the memory may also include a combination of the aforementioned types of memory.

[0153] When one or more programs in the storage medium can be executed by one or more processors, the target tracking method executed on the target tracking device side can be implemented.

[0154] The processor is used to execute a target tracking program stored in a memory to implement the following steps of a target tracking method performed on the target tracking device side: obtaining a first video frame and a second video frame of a long jump area at the same time, the first video frame being captured by a first camera device deployed above the long jump area, and the second video frame being captured by a second camera device deployed to the side of the long jump area; when the first video frame meets a first preset condition for target tracking of a test object in the long jump area, target tracking is performed on the test object in the acquired second video frame to obtain a tracking result corresponding to the second video frame.

[0155] Professionals should also be further aware that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the above description has generally described the components and steps of each example according to their functions. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0156] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein may be implemented using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0157] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A target tracking method, characterized in that: include: Acquire a first video frame and a second video frame of the long jump area at the same time, wherein the first video frame is captured by a first camera device deployed above the long jump area, and the second video frame is captured by a second camera device deployed to the side of the long jump area; When the first video frame satisfies a first preset condition for target tracking of the test subject in the long jump area, target tracking is performed on the test subject in the acquired second video frame to obtain a tracking result corresponding to the second video frame, wherein the first video frame includes a third image corresponding to the test area, and the first preset condition includes: the number of head detection frames in the third image is less than a first preset threshold; or There is no head detection frame in the third image and there is no head detection frame in the third image corresponding to each first video frame in the target video frame set; wherein, the target video frame set is composed of a plurality of consecutive first video frames before the first video frame at the current moment, and the number of first video frames in the target video frame set is less than a third preset threshold.

2. The method according to claim 1, characterized in that The method further comprises: capturing an initial video frame using the first camera device; When the initial video frame satisfies a second preset condition for target tracking of the test object in the long jump area, the step of acquiring the first video frame and the second video frame of the long jump area at the same moment is performed.

3. The method according to claim 2, characterized in that The long jump area includes a take-off area and a test area, and the initial video frame includes a first image corresponding to the take-off area and a second image corresponding to the test area; The second preset condition includes: A head detection frame exists in the first image and the head detection frame does not exist in the second image.

4. The method according to claim 3, characterized in that The first video frame includes a third image corresponding to the test area; The first preset condition includes: The number of the head detection frames in the third image is less than a first preset threshold; or, The head detection frame does not exist in the third image and the head detection frame does not exist in the third image corresponding to each first video frame in the target video frame set; wherein, the target video frame set is composed of a plurality of consecutive first video frames that are preceding the first video frame at the current moment.

5. The method according to claim 4, characterized in that The second video frame includes a fourth image corresponding to the test area; The performing target tracking on the test object in the acquired second video frame to obtain a tracking result corresponding to the second video frame includes: Determine whether the head detection frame exists in the third image corresponding to the first video frame at the same time as the second video frame; When the head detection frame exists, obtaining a human body detection frame in the fourth image corresponding to the second video frame according to the head detection frame, and using the human body detection frame as a tracking result corresponding to the second video frame; When the head detection frame does not exist, the human body detection frame in the fourth image corresponding to the second video frame is determined based on the head prediction frame in the third image corresponding to the first video frame, and the determined human body detection frame is used as the tracking result corresponding to the second video frame. The head prediction frame is determined by a first preset tracker.

6. The method according to claim 5, characterized in that The method further comprises: When there are multiple human body detection frames determined based on the head detection frame or the head prediction frame, matching the multiple human body detection frames with the human body prediction frame in the fourth image corresponding to the second video frame using a Hungarian algorithm to obtain a target matching result, wherein the human body prediction frame is determined by a second preset tracker; According to the target matching result, a tracking result corresponding to the second video frame is obtained.

7. The method according to claim 3, characterized in that The head detection frame in the first image is determined by: Determine the head detection frame in the first image using a head detection model; Determining an intersection-over-union ratio between the head detection frame and the first image; When the intersection-over-union ratio is greater than a second preset threshold, it is determined that the head detection frame exists in the first image.

8. A target tracking device, characterized in that: include: an acquisition module, configured to acquire a first video frame and a second video frame of a long jump area at the same moment, wherein the first video frame is acquired by a first camera device deployed above the long jump area, and the second video frame is acquired by a second camera device deployed to the side of the long jump area; a tracking module configured to track the test subject in the acquired second video frame when the first video frame satisfies a first preset condition for target tracking of the test subject in the long jump area, and obtain a tracking result corresponding to the second video frame, wherein the first video frame includes a third image corresponding to the test area, and the first preset condition includes: the number of head detection frames in the third image is less than a first preset threshold; or There is no head detection frame in the third image and there is no head detection frame in the third image corresponding to each first video frame in the target video frame set; wherein, the target video frame set is composed of a plurality of consecutive first video frames before the first video frame at the current moment, and the number of first video frames in the target video frame set is less than a third preset threshold.

9. An electronic device, characterized in that: include: A processor and a memory, wherein the processor is configured to execute a target tracking program stored in the memory to implement the target tracking method according to any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by the processor, the target tracking method according to any one of claims 1 to 7 is implemented.

Images