A method and system for tracking, timing and photographing events

Abstract

This application relates to the field of sports event tracking and filming technology, and discloses a method and system for sports event tracking and timing filming. The method includes: analyzing the real-time status of participants based on real-time data collected from at least two mobile high-speed cameras and real-time timing data of the event to obtain real-time status information; calculating the probability of participants at each position based on the real-time status information to construct a set of position distribution probability pipelines; analyzing the coverage value of cameras at different positions on each position distribution probability pipeline in the set of position distribution probability pipelines, and selecting the camera combination with the optimal coverage value; analyzing the real-time status information of participants, assigning at least one tracking role and tracking target to each camera in the camera combination to obtain a tracking and filming strategy; and controlling the corresponding camera to perform target tracking according to the tracking and filming strategy. This application can improve the success rate of capturing moments such as overtaking and sprinting of participants, and improve the quality of tracking and filming.

Description

Technical Field

[0001] This application relates to the field of sports event tracking and shooting technology, and more specifically to a sports event tracking and timing shooting method and system. Background Technology

[0002] Currently, with the trend of professionalization and intelligent broadcasting in sports events, circular track events place extremely high demands on the accuracy and coordination of tracking, timing, and filming. Circular track events include, but are not limited to, middle- and long-distance running in track and field, and track cycling. Traditional event filming relies on fixed-position cameras and manual operation, resulting in numerous blind spots and resource redundancy during filming. It is difficult to accurately and completely capture key moments in high-speed, multi-target circular track events. Furthermore, the separation of timing data and video data leads to an overly simplistic filming strategy, making it impossible to dynamically adjust and coordinate cameras in real time according to the event's dynamics.

[0003] The existing technology has the following problems: fixed camera positions result in incomplete coverage and low resource utilization efficiency; the motion trajectory is calculated by linear motion without considering the curvature constraint of the circular track, resulting in low accuracy of the predicted motion trajectory; each camera is responsible for a single fixed area, ignoring the coverage value of each camera position, resulting in fixed tracking process, low tracking accuracy and low camera utilization; each camera is responsible for a single tracking task, lacking dynamic tracking role division, and in the process of multi-target tracking, lacking a performance-oriented dynamic allocation method, it is easy to experience tracking interruption; in order to solve at least one of the above problems, this application proposes a sports event tracking timing shooting method and system. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the purpose of this application is to provide a method and system for tracking and timing events, effectively solving the problems described in the background section. The specific technical solution of this application is as follows:

[0005] A method for tracking and timing events, comprising:

[0006] Based on real-time data collected from at least two mobile high-speed cameras and real-time timing data of the event, the real-time status of the participants is analyzed through a preset status recognition model to obtain real-time status information.

[0007] Based on the real-time status information, the motion trajectory is predicted and analyzed according to the preset circular track model to obtain the target motion trajectory of the participants and predict the overtaking behavior and overtaking probability between the participants. The dynamic adjustment factor of different areas in the track is calculated and the target motion model is constructed. The probability of the participants at each position is calculated and the position distribution probability pipeline set is constructed.

[0008] In the set of location distribution probability pipelines, the coverage value of cameras at different locations to each location distribution probability pipeline is analyzed, the coverage value game process of different cameras is simulated, and the camera combination with the optimal coverage value is selected.

[0009] Analyze the real-time status information of the participants, assign at least one tracking role and tracking target to each camera in the camera array, and obtain a tracking shooting strategy. The tracking roles include an anchoring role responsible for global area coverage, an auxiliary role responsible for camera position supplementation, and a locking role responsible for precise tracking.

[0010] According to the tracking and shooting strategy, the corresponding camera is controlled to track the target in order to conduct real-time tracking and timing shooting of the event.

[0011] Specifically, the real-time status information of the participants is obtained by analyzing real-time data collected from at least two mobile high-speed cameras and real-time timing data of the event through a preset status recognition model, including:

[0012] Based on real-time data acquired from at least two mobile high-speed cameras and real-time timing data of the event, target bounding boxes including the positions of participants are identified.

[0013] Within the target bounding box, the real-time status of the participants is analyzed using a preset status recognition model to obtain real-time status information.

[0014] Specifically, based on the real-time status information, the motion trajectory is predicted and analyzed according to the preset circular track model to obtain the target motion trajectory of the participants and predict the overtaking behavior and overtaking probability among the participants. Dynamic adjustment factors for different areas of the track are calculated and a target motion model is constructed. The probability of each participant at each position is calculated, and a set of position distribution probability pipelines is constructed, including:

[0015] Based on real-time status information and a preset circular track model, the target motion trajectory of each participant on the track is analyzed and predicted. Combining the target motion trajectory and the relative positions between participants, the overtaking behavior and overtaking probability between participants are predicted, and the status prediction results are obtained.

[0016] Calculate the dynamic adjustment factor for different regions of the runway based on the state prediction results, and construct the target motion model according to the dynamic adjustment factor.

[0017] By combining the target motion model and target motion trajectory, the predicted running trajectory of the participants is simulated and predicted. The position probability of each participant in different areas of the track is statistically analyzed, and a set of position distribution probability pipelines is constructed.

[0018] Specifically, based on real-time status information and a preset circular track model, the system analyzes and predicts the target trajectory of each participant on the track. Combining the target trajectory with the relative positions of the participants, it predicts the overtaking behavior and probability among the participants, obtaining the status prediction results, including:

[0019] Based on the instantaneous speed and real-time position of the participants in the real-time status information, the linear motion of the participants is converted into the curvilinear motion of the tangential and normal vectors through the preset circular track model. The track curvature constraint is calculated and the target motion trajectory of each participant on the track is predicted.

[0020] Based on the target motion trajectory and the relative positions of the participants, analyze the interaction forces between the participants and construct an interactive motion model;

[0021] The interactive motion model is used to calculate the attractive and repulsive forces exerted on participants behind by participants in front, and to predict the overtaking behavior and probability of participants behind.

[0022] By combining the target's trajectory, overtaking behavior, and overtaking probability, the state prediction result is obtained.

[0023] Specifically, within the set of location distribution probability pipelines, the coverage value of cameras at different locations to each location distribution probability pipeline is analyzed, the game process of coverage value competition among different cameras is simulated, and the camera combination with the optimal coverage value is selected, including:

[0024] Based on the location distribution of participants in each pipe of the location distribution probability pipe set, analyze the coverage of each camera on the pipe and the overlap coverage between different cameras, and calculate the area coverage value of each camera.

[0025] Based on the coverage value of the area, the cooperative game process between each camera is simulated to select the camera combination with the optimal coverage value.

[0026] Specifically, based on the participant location distribution in each pipe of the location distribution probability pipe set, the coverage of each camera on the pipe and the overlap coverage between different cameras are analyzed, and the area coverage value of each camera is calculated, including:

[0027] Based on the participant position distribution in each pipe of the position distribution probability pipe set, the coverage of each camera to the pipe is analyzed through a preset simulated charge value assessment model. Areas with charge density greater than a preset charge threshold are simulated as high charge density areas. The potential energy value of each camera affected by charge at the corresponding position is calculated to obtain the corresponding first coverage value.

[0028] Analyze the spatial relationships between different cameras, identify the overlapping coverage areas between different cameras, and reduce the first coverage value of the overlapping coverage area according to the area to obtain the second coverage value.

[0029] By combining the first coverage value and the second coverage value, the regional coverage value of each camera is obtained.

[0030] Specifically, the analysis of the participants' real-time status information, the assignment of at least one tracking role and tracking target to each camera in the camera array, and the resulting tracking and shooting strategy include:

[0031] Analyze the real-time status information of the participants and calculate the comprehensive performance of each camera in the camera combination under different tracking role configurations through a preset role evaluation model;

[0032] Based on the overall performance, the tracking target conflict between cameras is analyzed, and at least one tracking role and tracking target are assigned to each camera to obtain a tracking shooting strategy.

[0033] Specifically, based on the overall performance, the tracking target conflict between cameras is analyzed, and at least one tracking role and tracking target are assigned to each camera to obtain a tracking shooting strategy, including:

[0034] When multiple cameras compete for the same tracking target, the camera with the best overall performance is selected to match the corresponding tracking target.

[0035] By combining the tracking target matched to each camera with the camera position, at least one tracking role is assigned to each camera;

[0036] According to the preset time window, the tracking role and tracking target of each camera are dynamically adjusted to obtain the tracking shooting strategy.

[0037] Specifically, according to the aforementioned tracking and shooting strategy, the corresponding camera is controlled to track the target in order to conduct real-time tracking and timing of the event, including:

[0038] Generate corresponding tracking shooting instructions based on the tracking role and target information assigned to each camera in the tracking shooting strategy;

[0039] According to the tracking and shooting instructions, the corresponding camera is controlled to track the target in order to conduct real-time tracking and timing shooting of the event.

[0040] A sports event tracking and timing shooting system, used to implement the aforementioned sports event tracking and timing shooting method, includes:

[0041] The status analysis module analyzes the real-time status of participants based on real-time data collected from at least two mobile high-speed cameras and real-time timing data of the event, and obtains real-time status information through a preset status recognition model.

[0042] The position analysis module, based on the real-time status information, predicts and analyzes the movement trajectory according to the preset circular track model, obtains the target movement trajectory of the participants, predicts the overtaking behavior and overtaking probability between the participants, calculates the dynamic adjustment factor of different areas in the track and constructs the target movement model, calculates the probability of the participants at each position, and constructs a set of position distribution probability pipelines.

[0043] The camera value game module analyzes the coverage value of cameras at different locations to each location distribution probability pipeline in the set of location distribution probability pipelines, simulates the coverage value game process of different cameras, and selects the camera combination with the optimal coverage value.

[0044] The tracking and shooting strategy formulation module analyzes the real-time status information of the participants and assigns at least one tracking role and tracking target to each camera in the camera combination to obtain the tracking and shooting strategy.

[0045] The tracking and shooting module controls the corresponding camera to track the target according to the tracking and shooting strategy, so as to carry out real-time tracking and timing shooting of the event process.

[0046] The beneficial effects of this application are as follows: By fusing data from at least two mobile high-speed cameras and timing data, the real-time status information of the participants is analyzed. Combined with the curvature constraints of the circular track to transform motion vectors and predict overtaking behavior through interactive motion between participants, a set of position distribution probability pipelines is constructed. The optimal camera combination is selected by simulating the cooperative game process of each camera, and the tracking target and tracking role of each camera are dynamically allocated according to comprehensive effectiveness, achieving real-time event tracking and filming. By constructing probability pipelines and conducting coverage value analysis, high-value areas with a high probability of participant appearance in the race can be monitored and filmed in real time. Real-time scheduling of cameras based on the real-time trajectory prediction results of participants can improve the success rate of capturing moments such as overtaking and sprinting. Simulating and analyzing the cooperative game process between cameras and dynamically allocating the tracking role and tracking target of each camera allows for optimal filming results with fewer devices, avoiding ineffective overlapping work between cameras and improving the tracking quality and filming effectiveness of the event process. Attached Figure Description

[0047] Figure 1 This is a flowchart illustrating the process of a sports event tracking and timing photography method as described in this application.

[0048] Figure 2This is a schematic diagram of the target border in an embodiment of this application;

[0049] Figure 3 This is a schematic diagram of the overlapping coverage area in the embodiments of this application;

[0050] Figure 4 This is a schematic diagram of the structure of a race tracking and timing shooting system according to an embodiment of this application. Detailed Implementation

[0051] The present application will be further described in detail below with reference to the accompanying drawings and embodiments. Identical components are indicated by the same reference numerals. In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0052] Hereinafter, the terms "first," "second," and other generic terms are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0053] refer to Figure 1 The image shows a specific implementation of a sports event tracking and timing photography method according to this application, comprising:

[0054] S101. Based on real-time data collected from at least two mobile high-speed cameras and real-time timing data of the event, analyze the real-time status of the participants through a preset status recognition model to obtain real-time status information.

[0055] S102. Based on the real-time status information, predict and analyze the movement trajectory according to the preset circular track model to obtain the target movement trajectory of the participants and predict the overtaking behavior and overtaking probability between the participants. Calculate the dynamic adjustment factor of different areas in the track and construct the target movement model. Calculate the probability of the participants at each position and construct the position distribution probability pipeline set.

[0056] S103. In the set of location distribution probability pipelines, analyze the coverage value of cameras at different locations to each location distribution probability pipeline, simulate the coverage value game process of different cameras, and select the camera combination with the optimal coverage value.

[0057] S104. Analyze the real-time status information of the participants, assign at least one tracking role and tracking target to each camera in the camera combination, and obtain a tracking shooting strategy. The tracking role includes an anchoring role responsible for global area coverage, an auxiliary role responsible for camera position supplementation, and a locking role responsible for precise tracking.

[0058] S105. According to the tracking and shooting strategy, control the corresponding camera to track the target in order to track and shoot the event in real time.

[0059] In this embodiment, based on real-time data collected from at least two mobile high-speed cameras and real-time timing data of the event, the real-time status of the participants is analyzed through a preset status recognition model. The corresponding participants are identified and selected from each frame of the image, and the instantaneous velocity and acceleration of the participants in the frame are analyzed to obtain real-time status information. By fusing the real-time data collected from the cameras and the real-time timing data of the event, the status of the athletes can be analyzed and judged from multiple dimensions, improving the accuracy of the participant status analysis, providing accurate data support for the prediction of the participants' movement trajectory, and improving the accuracy of the participants' path prediction results.

[0060] Specifically, based on real-time status information, the motion trajectory is predicted and analyzed according to a preset circular track model. The athlete's linear motion is decomposed into tangential and normal motion vectors according to the track shape. The tangential vector is along the track direction, and the normal vector points towards the center of the circle. Combined with the track curvature analysis and centripetal acceleration constraints, the target motion trajectory of the participants is predicted. The overtaking behavior and overtaking probability are analyzed by combining the influence forces between the participants, and the probability of the participants at each position is calculated to construct a set of position distribution probability pipelines. By constructing probability pipelines and analyzing the influence forces between participants, the overtaking situation of the participants can be predicted in advance, improving the accuracy of the target motion trajectory and effectively avoiding the problem of low tracking accuracy caused by trajectory prediction errors. This provides accurate trajectory references for optimizing the allocation of photography resources and improves the accuracy and effectiveness of camera tracking and shooting.

[0061] Specifically, within the set of location distribution probability channels, areas with a high probability of participants appearing are simulated as regions with high charge density in physics. The coverage value of cameras at different locations for each location distribution probability channel is analyzed and calculated. Each camera is treated as a game participant, simulating the game process of coverage value among different cameras, and selecting the camera combination with the optimal coverage value. By simulating charge distribution, the shooting value of each camera at its corresponding location can be accurately quantified, ensuring that cameras can capture the most exciting areas of the event in a timely manner. By simulating the cooperative game process of cameras, the problems of overlapping coverage and blind spots in multi-camera shooting can be effectively avoided. The cameras are automatically coordinated to form a complementary relationship in the shooting coverage area, achieving comprehensive coverage of the track probability channels with a limited number of cameras, improving camera utilization efficiency and effectiveness.

[0062] Specifically, the system analyzes the real-time status information of the participants, and calculates the overall performance score for each camera by analyzing the camera's position, focal length, gimbal performance, and the status, position, and importance of the tracked target. Based on this overall performance score, each camera in the camera group is assigned at least one tracking role and tracking target, resulting in a tracking and shooting strategy. The tracking roles include an anchoring role responsible for overall area coverage, an auxiliary role responsible for supplementing camera positions, and a locking role responsible for precise tracking. Dynamically adjusting the camera's tracking roles and tracking targets ensures a comprehensive grasp of the overall situation of the event, accurate tracking and shooting of key participants and key time points, and timely filling of camera gaps and handling of unexpected situations through auxiliary roles. Dynamic tracking adjustments ensure the integrity of overall information and the accuracy of local information. The dynamic adjustment mechanism ensures that the strategy can adapt to real-time changes in the event, allowing for rapid deployment of cameras for target tracking in unexpected situations, maintaining shooting continuity and quality.

[0063] According to the tracking shooting strategy, each camera is assigned a corresponding tracking shooting command to control the corresponding camera to track the target and shoot the event in real time. By generating commands and controlling the cameras in real time, the accuracy and real-time performance of the camera shooting actions are improved. When the event is in progress at high speed, the corresponding competition process can be captured in real time, improving the accuracy and efficiency of the event shooting process.

[0064] This application analyzes the real-time status information of participants by fusing data from at least two mobile high-speed cameras and timing data. It combines the curvature constraints of the circular track with the transformation of motion vectors and the interactive motion prediction between participants to construct a set of position distribution probability pipelines. By simulating the cooperative game process of each camera, the optimal camera combination is selected, and the tracking target and tracking role of each camera are dynamically assigned according to comprehensive effectiveness, achieving real-time event tracking and filming. Through the construction of probability pipelines and coverage value analysis, high-value areas with a high probability of participant appearance in the race are ensured to be monitored and filmed in real time. Real-time scheduling of cameras based on the real-time trajectory prediction results of participants can improve the success rate of capturing moments such as overtaking and sprinting. Simulating and analyzing the cooperative game process between cameras and dynamically assigning the tracking role and tracking target of each camera allows for optimal filming results with fewer devices, avoiding ineffective overlapping work between cameras and improving the tracking quality and filming effectiveness of the event.

[0065] Furthermore, based on real-time data collected from at least two mobile high-speed cameras and real-time timing data of the event, the real-time status of the participants is analyzed through a preset status recognition model to obtain real-time status information, including:

[0066] S201. Based on real-time acquisition data from at least two mobile high-speed cameras and real-time timing data of the event, identify the target bounding box including the location of the participants.

[0067] S202. Within the target bounding box, the real-time status of the participants is analyzed using a preset status recognition model to obtain real-time status information.

[0068] In this embodiment, based on real-time acquisition data from at least two mobile high-speed cameras and real-time timing data of the event, the mobile high-speed cameras are deployed at different locations on the circular track, including but not limited to the outside of the curves and the starting point of the straightaways, to acquire real-time image stream data. The real-time timing data includes but is not limited to the timing data output by the starting blocks and the finish line timers. The video data from different mobile high-speed cameras are synchronously decoded and time-aligned to ensure the synchronization and temporal consistency between the data acquired from different camera positions. A convolutional neural network model pre-trained with a large amount of sports event data is used to perform in-depth analysis on each video frame.

[0069] like Figure 2As shown, the model identifies regions with human morphological features and generates a rectangular target bounding box that completely surrounds the participant. Based on real-time timing data, the model identifies the target bounding box at the corresponding position in real time. By combining real-time acquired data and real-time timing data, positional errors can be effectively corrected, improving the accuracy of the identified target bounding box. Through collaborative processing of multi-camera data and fusion of multi-view information, the model continuously tracks the participant, improving the accuracy and continuity of the tracking process.

[0070] Specifically, within the target bounding box, a pre-defined state recognition model analyzes the real-time state of the participants to obtain corresponding real-time state information. This state recognition model includes, but is not limited to, a neural network model. The neural network model is trained using a large amount of real-time state data from the participants to obtain a pre-trained model. The model performs posture change analysis and position migration analysis on the participants within each target bounding box, outputting the participants' real-time state information, including but not limited to angular velocity, acceleration, and body center of gravity offset angle. By analyzing the participants' state information in real time, the model can reveal changes in their state, providing data support for trajectory analysis and prediction, improving the accuracy of trajectory prediction results, and providing accurate target trajectory references for real-time camera tracking.

[0071] Furthermore, based on the real-time status information, the motion trajectory is predicted and analyzed according to the preset circular track model to obtain the target motion trajectory of the participants and predict the overtaking behavior and overtaking probability among the participants. Dynamic adjustment factors for different areas of the track are calculated and a target motion model is constructed. The probability of each participant at each position is calculated, and a set of position distribution probability pipelines is constructed, including:

[0072] S301. Based on real-time status information and a preset circular track model, analyze and predict the target motion trajectory of each participant on the track. Combine the target motion trajectory with the relative positions between participants to predict the overtaking behavior and overtaking probability between participants, and obtain the status prediction result.

[0073] S302. Calculate the dynamic adjustment factor for different areas in the runway based on the state prediction results, and construct the target motion model according to the dynamic adjustment factor.

[0074] S303. Combining the target motion model and target motion trajectory, simulate and predict the predicted running trajectory of the participants, calculate the position probability of each participant in different areas of the track, and construct a set of position distribution probability pipelines.

[0075] In this embodiment, based on real-time status information and a preset circular track model, the target trajectory of each participant on the track is analyzed and predicted. Combining the target trajectory with the relative positions between participants, the overtaking behavior and overtaking probability between participants are predicted, resulting in a status prediction result. By combining the circular track model with the relative interaction between participants, the accuracy of the trajectory prediction result can be improved, the probability of overtaking events can be quantified, and accurate scheduling decision support can be provided for camera scheduling. For participants whose overtaking probability is in a rapidly increasing phase, shooting resources can be deployed in advance to accurately capture the corresponding overtaking moment image.

[0076] Specifically, the runway is divided into regional grids based on the state prediction results. The grid size is set according to the tracking and shooting accuracy requirements. For each grid region, a corresponding dynamic adjustment factor is calculated. The dynamic adjustment factor is evaluated and calculated in real time in conjunction with the real-time runway environment. The overtaking probability and the participant density within the grid region are calculated. The overtaking probability and the participant density are multiplied to obtain the corresponding dynamic adjustment factor. The participant density is based on the number of participants expected to pass through per unit time. According to the dynamic adjustment factor, for tangential velocity, the acceleration weight is increased in high adjustment factor regions. For normal components, the curvature constraint weight is enhanced in high adjustment factor regions of curves. Target motion models corresponding to different regional characteristics are constructed. By introducing the dynamic adjustment factor to construct the corresponding motion model, it is possible to identify and focus on key areas with high dynamic characteristics, improving the adaptability of the motion model in complex track environments. It can reduce the prediction trajectory error in complex scenarios such as curves, sprint sections, and areas of intense competition among athletes, improving the accuracy of the prediction trajectory analysis results.

[0077] Specifically, combining the target motion model and target motion trajectory, the Monte Carlo simulation method is used to simulate and predict the predicted running trajectory of the participants. A large number of particles are generated, each corresponding to a virtual athlete. Each particle follows the rules of the corresponding target motion model. The initial state of the particles is determined based on the predicted target motion trajectory value. The corresponding overtaking event is simulated based on the predicted overtaking probability. The trajectory simulation process is analyzed throughout the overall simulation, and the predicted running trajectory is simulated. Statistical analysis is performed on the predicted running trajectory. In each region of the track space, the frequency of occurrence of trajectory particles for each participant is counted, and the probability of that participant appearing in that region at a future time point is calculated. The position probabilities of each participant in different regions of the track are combined to construct a set of position distribution probability pipelines. By constructing the probability pipeline set, the activity range and probability of the participants are analyzed in the form of a probability field. When the actual movement path of a participant deviates from the predicted optimal trajectory, corresponding camera resources can be scheduled according to the high-probability pipeline range corresponding to that path, improving the success rate of the camera capturing exciting moments.

[0078] Furthermore, based on real-time status information and a pre-set circular track model, the target trajectory of each participant on the track is analyzed and predicted. Combining the target trajectory with the relative positions of the participants, the overtaking behavior and probability among the participants are predicted, resulting in a status prediction outcome, including:

[0079] S401. Based on the instantaneous speed and real-time position of the participants in the real-time status information, the linear motion of the participants is converted into the curvilinear motion of the tangential and normal vectors through the preset circular track model, the track curvature constraint is calculated, and the target motion trajectory of each participant on the track is predicted.

[0080] S402. Based on the target motion trajectory and the relative positions of the participants, analyze the interaction forces between the participants and construct an interactive motion model;

[0081] S403. Calculate the attractive and repulsive forces exerted on the participants in front by the participants behind using the interactive motion model, and predict the overtaking behavior and overtaking probability of the participants behind.

[0082] S404. Combining the target's trajectory, overtaking behavior, and overtaking probability, the state prediction result is obtained.

[0083] In this embodiment, based on the instantaneous velocity and real-time position of the participants in the real-time status information, the linear motion of the participants is converted into curvilinear motion with tangential and normal vectors through a preset circular track model. The circular track model is used to describe the geometry of the track, including but not limited to physical parameters such as the radius of curvature of each segment of the track, the center of the curve, the track width, and the slope. At the position of the participants, the velocity vector is decomposed into two orthogonal components: a tangential vector along the tangent direction of the track and a normal vector pointing towards the center of the track curvature. The tangential vector reflects the forward velocity of the participants, and the normal vector reflects the centripetal acceleration of the participants. Combined with the real-time curvature of the track at that position, the centripetal acceleration required for the participants to maintain curvilinear motion is calculated to obtain the track curvature constraint. According to the track curvature constraint, the tangential velocity of the participants is analyzed to calculate the velocity decay range or velocity maintenance range of the participants. The tangential velocity and normal acceleration are inferred to predict the target motion trajectory of each participant on the track.

[0084] It should be noted that by constructing runway curvature constraints and motion vector decomposition, the obtained trajectory prediction results conform to the pattern of a circular track, which can reduce trajectory prediction errors in complex sections such as curves, improve the accuracy of trajectory prediction results, and enhance the environmental adaptability of trajectory prediction results.

[0085] Specifically, based on the target trajectory and the relative positions of the participants, the interaction forces between them are simulated and analyzed. Each participant is simulated as a potential energy field centered on them, and the attractive and repulsive forces between different potential energy fields are analyzed to construct an interactive motion model. The model can reflect the attractive forces that participants exert to reduce air resistance by following each other, the repulsive forces that prevent collisions and maintain personal tactical space, and the forces that drive overtaking at curves. By constructing an interactive motion model, the analysis is no longer limited to individual participants; instead, it incorporates correlational analysis between participants. By combining the overtaking and following intentions of the participants, the accuracy of the participant state analysis results can be improved.

[0086] Specifically, based on the interactive motion model, the analysis is performed on each pair of participants with a front-to-back positional relationship. For the participant in the rear, the cooperative force exerted by one or more athletes in front is calculated. The attraction force is calculated by superimposing the speed difference between the participant and the athletes in front and behind, and the reciprocal of the relative distance between the participant and the athletes in front and behind is calculated to obtain the corresponding repulsive force. The cooperative force is calculated by superimposing the attraction and repulsive forces. Based on the cooperative force, the overtaking behavior of the participant in the rear is predicted. The magnitude of the cooperative force and the currently available overtaking space are analyzed, and the corresponding overtaking probability is calculated. By simulating the attraction and repulsion forces between participants, the overtaking situation between athletes can be analyzed, and the corresponding overtaking probability value can be calculated. This provides accurate data support for camera scheduling and improves the effectiveness of camera tracking and filming.

[0087] Specifically, by combining the target trajectory, overtaking behavior, and overtaking probability, the predicted target trajectory, overtaking behavior, and overtaking probability are spatiotemporally aligned and correlated. The predicted overtaking behavior and overtaking probability are marked as key event points on the corresponding athlete's target trajectory. The athlete's trajectory is evaluated based on the overtaking probability to obtain the state prediction result. The state prediction result includes each athlete's predicted position, speed, posture tendency, and interaction events with other athletes and their confidence levels. Combining each athlete's target trajectory with the corresponding overtaking behavior allows for the analysis of the corresponding motion state by integrating the athlete's dynamic motion information, improving the accuracy of the state prediction result, providing accurate data support for camera scheduling decisions, and improving camera shooting efficiency and the effectiveness of field tracking.

[0088] Furthermore, within the set of location distribution probability pipelines, the coverage value of cameras at different locations to each location distribution probability pipeline is analyzed, the game process of coverage value competition among different cameras is simulated, and the camera combination with the optimal coverage value is selected, including:

[0089] S501. Based on the location distribution of participants in each pipe of the location distribution probability pipe set, analyze the coverage of each camera on the pipe and the overlapping coverage between different cameras, and calculate the area coverage value of each camera.

[0090] S502. Based on the coverage value of the area, simulate the cooperative game process between each camera and select the camera combination with the optimal coverage value.

[0091] In this embodiment, based on the participant location distribution in each pipeline of the location distribution probability pipeline set, the existence probability of each participant in each spatiotemporal unit is simulated as the corresponding charge density. The coverage of each camera on the pipeline and the overlapping coverage between different cameras are analyzed, and the regional coverage value of each camera is calculated. By simulating the charge density, the spatiotemporal regions where the most exciting events and overtaking events occur can be analyzed and identified. The overlapping coverage areas can be identified and their values ​​dynamically adjusted, which can avoid ineffective scheduling of camera resources, provide accurate data support for the collaborative optimization of the camera scheduling process, and improve the effectiveness and efficiency of the camera scheduling process.

[0092] Specifically, based on regional coverage value, each camera is treated as a game participant, simulating a cooperative game process between them. The strategy space is set to selected or not selected, and the payoff function defines the sum of the regional coverage values ​​of the selected cameras. Considering that different cameras may have complementary or overlapping views, different cameras are combined, and the global coverage value of the combined cameras is calculated. A greedy algorithm is used to traverse and search all camera combination subsets, evaluating the global coverage value of each subset, and finally selecting the camera combination with the optimal global coverage value. By simulating the camera cooperative game process, the various cameras can be dynamically coordinated, avoiding coverage blind spots caused by cameras focusing on a single hotspot, and preventing cameras from being too scattered to capture key areas in real time, thereby improving the overall quality and effectiveness of the shooting during event tracking.

[0093] Furthermore, based on the participant location distribution in each pipe of the location distribution probability pipe set, the coverage of each camera on the pipe and the overlap coverage between different cameras are analyzed, and the area coverage value of each camera is calculated, including:

[0094] S601. Based on the participant position distribution in each pipe of the position distribution probability pipe set, analyze the coverage of each camera on the pipe through the preset simulated charge value assessment model, simulate the area with charge density greater than the preset charge threshold as a high charge density area, calculate the potential energy value of each camera affected by charge at the corresponding position, and obtain the corresponding first coverage value.

[0095] S602. Analyze the spatial relationship between different cameras, identify the overlapping coverage area between different cameras, and reduce the first coverage value in the overlapping coverage area according to the area to obtain the second coverage value.

[0096] S603. Combining the first coverage value and the second coverage value, the area coverage value of each camera is obtained.

[0097] In this embodiment, based on the participant position distribution in each pipe of the position distribution probability pipe set, the coverage of each camera to the pipe is analyzed using a preset simulated charge value assessment model. Areas with charge density greater than a preset charge threshold are simulated as high charge density areas. The potential energy value of each camera affected by charge at its corresponding position is calculated to obtain the corresponding first coverage value. The position distribution probability pipe set is discretized, and the track is divided into corresponding grid cells. In each grid cell, the probability value of the participant position included in the area is calculated, and the corresponding charge density is obtained by superposition calculation. The charge threshold is set according to the accuracy of event tracking and shooting. Areas with charge density greater than the charge threshold are regarded as high charge density areas. The camera position of each camera is regarded as a detection point through the simulated charge value assessment model. The potential energy value of the point in the electrostatic field composed of the entire probability pipe set is calculated. The corresponding potential energy value is obtained by integrating all grid cells within the camera's field of view. This potential energy value is used as the corresponding first coverage value.

[0098] It is important to emphasize that by using a simulated charge value assessment model to quantify the value of camera footage, the discrete probability distribution can be transformed into a continuous potential energy field. This allows for the quantification of the coverage value of each camera position for dynamic events, enabling the selection of camera positions that can provide the best imaging quality for high-value areas and improving the quality of event tracking and filming.

[0099] like Figure 3As shown, the spatial relationships between different cameras are analyzed to identify overlapping coverage areas. A weighted reduction of the first coverage value within each overlapping coverage area yields a second coverage value. Furthermore, overlapping coverage areas are identified by calculating the spatial intersection of the coverage areas of any two cameras. A weight is calculated by combining the area of ​​the overlapping coverage area with its proportion within the original camera's coverage area. This weighted reduction of the first coverage value within the overlapping coverage area yields a second coverage value. By identifying overlapping coverage areas and applying value reduction, ineffective competition between cameras in these areas can be avoided. This results in a final camera combination that is functionally complementary rather than functionally redundant. This allows for wider coverage shooting combinations with limited camera resources, improving camera utilization efficiency during tracking and shooting, and ultimately leading to an optimal camera collaborative shooting strategy.

[0100] Specifically, by combining the first and second coverage values, the regional coverage value of each camera is obtained. The first coverage value of non-overlapping areas and the second coverage value of overlapping areas are then combined to obtain the regional coverage value for each camera. This value combination allows for the integration of each camera's overall shooting capabilities and collaborative shooting capabilities, enhancing the collaborative shooting effect of the camera combination and improving the overall performance of the entire event tracking and shooting process.

[0101] Furthermore, by analyzing the real-time status information of the participants, at least one tracking role and tracking target are assigned to each camera in the camera array to obtain a tracking shooting strategy, including:

[0102] S701. Analyze the real-time status information of the participants and calculate the comprehensive performance of each camera in the camera combination under different tracking role configurations through the preset role evaluation model;

[0103] S702. Based on the overall performance, analyze the tracking target conflict between cameras, assign at least one tracking role and tracking target to each camera, and obtain a tracking shooting strategy.

[0104] In this embodiment, the real-time status information of the participants is analyzed, and the comprehensive performance of each camera in the camera combination under different tracking role configurations is calculated through a preset role evaluation model. The role evaluation model includes, but is not limited to, a multi-factor weighted decision model. The evaluation dimensions include static and dynamic attributes. Static attributes include, but are not limited to, the focal length range of the camera lens, the rotation speed and accuracy of the gimbal, and the resolution of the sensor. Dynamic attributes include, but are not limited to, the geometric relationship between the current spatial position and orientation of the camera and the target area, and whether there is redundant coverage provided by other cameras in the current field of view. The multi-factor weighted decision model formulates corresponding performance evaluation criteria for anchoring roles, auxiliary roles, and locking roles. For example, for anchoring roles, the model prioritizes evaluating the camera's field of view and its basic coverage capability of key areas. For locking roles, it prioritizes evaluating the optical zoom capability, the response speed of the tracked target, and the smoothness of the gimbal mechanism's motion on the predicted trajectory. The model calculates the evaluation value of each camera for each evaluation dimension when playing different tracking roles and sums them to obtain the corresponding comprehensive performance.

[0105] It should be noted that by quantitatively evaluating the performance of each camera in its respective role, accurate data support can be provided for the allocation of camera roles. By conducting multi-role and multi-dimensional performance evaluations of each camera, cameras with advantages in their respective roles can be identified, thereby improving the synergy and effectiveness of the selected camera combinations.

[0106] Specifically, based on comprehensive performance analysis, the tracking target conflicts between cameras are analyzed. Each camera is assigned at least one tracking role and target. The tracking role and target for each camera are dynamically adjusted according to the latest developments in the event, resulting in a tracking shooting strategy. This comprehensive performance-based conflict resolution mechanism ensures that the most suitable camera performs the most critical tasks in competitive scenarios, improving the quality of key shots. Dynamically adjusting the tracking role and target ensures both global coverage and local detail during the tracking process, enhancing the completeness of the content. Dynamically adjusting the tracking role and target throughout the event maintains accurate tracking of the focus of the competition, effectively responding to various unforeseen circumstances and improving the effectiveness of the event tracking shooting process.

[0107] Furthermore, based on the aforementioned overall performance, the tracking target conflict between cameras is analyzed, and at least one tracking role and tracking target are assigned to each camera to obtain a tracking shooting strategy, including:

[0108] S801. When multiple cameras compete for the same tracking target, select the optimal camera to match the corresponding tracking target based on overall performance.

[0109] S802. Based on the tracking target matched by each camera and the camera position, assign at least one tracking role to each camera;

[0110] S803. According to the preset time window, the tracking role and tracking target of each camera are dynamically adjusted to obtain the tracking shooting strategy.

[0111] In this embodiment, when multiple cameras compete for the same tracking target, the camera with the best overall performance is selected to be matched with the corresponding tracking target. Tracking target conflict is detected; when two or more cameras calculate a high overall performance for the same tracking target, a resource competition is determined. The overall performance calculated by the competing cameras for that tracking target is analyzed, compared, and the camera with the best overall performance is selected. The tracking target is then assigned to that camera, completing the matching process. By selecting the appropriate camera, the waste of multiple cameras repeatedly tracking and filming the same target can be effectively avoided, ensuring that each tracking target receives a corresponding matched camera. This effectively prevents resource waste caused by camera competition and improves the success rate and image quality of capturing the most critical moments of the event.

[0112] Specifically, based on the tracking target matched to each camera and the camera's position, at least one tracking role is assigned to each camera. Tracking roles are assigned according to each camera's physical location, field of view characteristics, and the nature of the task it has undertaken. The assignment rules include: for cameras not specifically assigned to an athlete target but located at a strategic high point with a wide-angle view, an anchoring role is assigned; for cameras that have successfully matched a key athlete target, a locking role is assigned; for the remaining cameras, auxiliary roles are assigned based on the camera's position and field of view to fill the coverage gaps between the anchoring and locking roles. By assigning corresponding roles to each camera, both the overall situation and local details can be taken into account, ensuring complete monitoring of the overall event process while ensuring focused capture of key figures and crucial events. This effectively addresses unexpected situations such as single-device malfunctions or obstructed views, improving the shooting quality of the event tracking process.

[0113] Specifically, according to a preset time window, the tracking role and target of each camera are dynamically adjusted to obtain a tracking and shooting strategy. A time window is set based on the required tracking accuracy for the event. The latest participant status information is obtained within the time window, and the latest competition situation is analyzed. The tracking role and target of each camera are then dynamically adjusted. For example, when an athlete who was previously locked onto a position loses value due to a lower ranking, the corresponding locked camera is reassigned to track the new leader. When a new high-probability overtaking event is predicted in another area, a nearby auxiliary camera is promoted to the locked role to track that target. By dynamically adjusting the tracking target and tracking role, the tracking and shooting strategy can be kept synchronized with the dynamic progress of the event, solving the problem that traditional static strategies cannot adapt to long-duration, highly dynamic events. This effectively avoids the risk of missing key shots during the event using a fixed strategy, thereby improving the coverage and shooting quality throughout the entire event.

[0114] Furthermore, according to the aforementioned tracking and shooting strategy, the corresponding camera is controlled to track the target in order to perform real-time tracking and timing shooting of the event process, including:

[0115] S901. Generate corresponding tracking shooting instructions according to the tracking role and target information assigned to each camera in the tracking shooting strategy;

[0116] S902. According to the tracking and shooting instruction, control the corresponding camera to track the target in order to perform real-time tracking and timing shooting of the event process.

[0117] In this embodiment, corresponding tracking and shooting instructions are generated according to the tracking role and target information assigned to each camera in the tracking and shooting strategy. The tracking and shooting strategy determines the tracking role and corresponding target information assigned to each camera. Based on the strategy information and the camera parameters, corresponding tracking and shooting instructions are generated using a random forest model pre-trained with a large amount of historical tracking data. These instructions include, but are not limited to, position sequences, gimbal yaw and pitch angle control values, and optical zoom parameters. By generating specific control instructions, execution deviations caused by ambiguous instructions can be avoided, enabling each camera to work according to its corresponding tracking role. This improves the accuracy of tracking lock, the stability of anchor coverage, and the flexibility of auxiliary positioning, thereby enhancing the quality of the tracking and shooting process.

[0118] Specifically, based on the tracking and shooting instructions, the corresponding camera is controlled to track the target for real-time tracking and timing of the event. The generated tracking and shooting instructions are sent to the underlying control unit of the corresponding camera, which executes the instructions in real time to track and time the event. By executing the corresponding control instructions in real time, it is ensured that the moving target is continuously and clearly locked in the frame, improving the reliability and effectiveness of the event tracking and shooting process.

[0119] like Figure 4 As shown, a sports event tracking and timing shooting system is used to implement a sports event tracking and timing shooting method, including:

[0120] The status analysis module analyzes the real-time status of participants based on real-time data collected from at least two mobile high-speed cameras and real-time timing data of the event, and obtains real-time status information through a preset status recognition model.

[0121] The position analysis module, based on the real-time status information, predicts and analyzes the movement trajectory according to the preset circular track model, obtains the target movement trajectory of the participants, predicts the overtaking behavior and overtaking probability between the participants, calculates the dynamic adjustment factor of different areas in the track and constructs the target movement model, calculates the probability of the participants at each position, and constructs a set of position distribution probability pipelines.

[0122] The camera value game module analyzes the coverage value of cameras at different locations to each location distribution probability pipeline in the set of location distribution probability pipelines, simulates the coverage value game process of different cameras, and selects the camera combination with the optimal coverage value.

[0123] The tracking and shooting strategy formulation module analyzes the real-time status information of the participants and assigns at least one tracking role and tracking target to each camera in the camera combination to obtain the tracking and shooting strategy.

[0124] The tracking and shooting module controls the corresponding camera to track the target according to the tracking and shooting strategy, so as to carry out real-time tracking and timing shooting of the event process.

[0125] In this embodiment, the state analysis module integrates real-time data from at least two mobile high-speed cameras with real-time event timing data. By analyzing the real-time state of participants through a pre-defined state recognition model, it avoids the limitations and errors of single-device data. Through multi-source data fusion analysis, it obtains high-precision, multi-dimensional real-time state information, providing accurate data support for participant trajectory prediction and camera scheduling, thus improving tracking and shooting accuracy. The position analysis module, based on the participants' real-time state information, predicts their movement trajectories using a pre-defined circular track model and calculates the probability of occurrence at different positions, constructing a set of position distribution probability pipelines. It transforms trajectory prediction into a probabilistic spatial distribution description, quantifies the uncertainty of participant positions, identifies high-probability areas and time-series changes, and provides data support for camera scheduling to cover the entire competition area, ensuring accurate capture of participant footage.

[0126] Specifically, the camera value game module analyzes the coverage value of cameras at different locations within the probability pipeline of location distribution. By simulating the coverage value game process, it selects the camera combination with the optimal coverage value. Through quantitative evaluation and game analysis, it analyzes the coverage efficiency and resource redundancy of each camera, avoiding the one-sidedness of single-device evaluation and the resource waste of multiple cameras covering the same area. This improves the coverage value and shooting efficiency of the selected camera combination, enhancing the utilization efficiency of equipment resources and the reliability of area coverage. The tracking and shooting strategy formulation module, based on the real-time status information of the participants, assigns at least one tracking role and a corresponding tracking target to each device in the selected optimal camera combination, thus obtaining a tracking and shooting strategy. By matching camera roles with targets, it can fully utilize the characteristics of each camera, avoid tracking failures caused by mismatch between camera tracking roles and targets, and improve tracking accuracy and coordination.

[0127] Specifically, the tracking and shooting module generates and executes specific control commands according to the established tracking and shooting strategy, driving the corresponding camera to complete target tracking, realizing real-time tracking and timing linkage shooting of the event process; real-time camera control ensures the stability of the tracking process, and combined with timing data, accurately corresponds to the participants' participation process, ensuring the quality of capturing details of key moments, and providing reliable data support for event timing and referee judgment, thereby improving the integrity and quality of event records.

[0128] The above description is merely a preferred embodiment of this application. The scope of protection of this application is not limited to the above embodiments. All technical solutions falling within the scope of this application's concept are within the scope of protection of this application. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of this application should also be considered within the scope of protection of this application.

Claims

1. A method for tracking and timing a sporting event, characterized in that, include: Based on real-time data collected from at least two mobile high-speed cameras and real-time timing data of the event, the real-time status of the participants is analyzed through a preset status recognition model to obtain real-time status information. Based on the real-time status information, the motion trajectory is predicted and analyzed according to the preset circular track model to obtain the target motion trajectory of the participants and predict the overtaking behavior and overtaking probability between the participants. The dynamic adjustment factor of different areas in the track is calculated and the target motion model is constructed. The probability of the participants at each position is calculated and the position distribution probability pipeline set is constructed. In the set of location distribution probability pipelines, the coverage value of cameras at different locations to each location distribution probability pipeline is analyzed, the coverage value game process of different cameras is simulated, and the camera combination with the optimal coverage value is selected. Analyze the real-time status information of the participants, assign at least one tracking role and tracking target to each camera in the camera array, and obtain a tracking shooting strategy. The tracking roles include an anchoring role responsible for global area coverage, an auxiliary role responsible for camera position supplementation, and a locking role responsible for precise tracking. According to the tracking and shooting strategy, the corresponding camera is controlled to track the target in order to conduct real-time tracking and timing shooting of the event.

2. The event tracking and timing shooting method according to claim 1, characterized in that, The process involves analyzing the real-time status of participants using a preset status recognition model based on real-time data collected from at least two mobile high-speed cameras and real-time timing data of the event, to obtain real-time status information, including: Based on real-time data acquired from at least two mobile high-speed cameras and real-time timing data of the event, target bounding boxes including the positions of participants are identified. Within the target bounding box, the real-time status of the participants is analyzed using a preset status recognition model to obtain real-time status information.

3. The event tracking and timing shooting method according to claim 1, characterized in that, Based on the real-time status information, the motion trajectory is predicted and analyzed according to the preset circular track model to obtain the target motion trajectory of the participants and predict the overtaking behavior and overtaking probability among the participants. Dynamic adjustment factors for different areas of the track are calculated and a target motion model is constructed. The probability of each participant at each position is calculated, and a set of position distribution probability pipelines is constructed, including: Based on real-time status information and a preset circular track model, the target motion trajectory of each participant on the track is analyzed and predicted. Combining the target motion trajectory and the relative positions between participants, the overtaking behavior and overtaking probability between participants are predicted, and the status prediction results are obtained. Calculate the dynamic adjustment factor for different regions of the runway based on the state prediction results, and construct the target motion model according to the dynamic adjustment factor. By combining the target motion model and target motion trajectory, the predicted running trajectory of the participants is simulated and predicted. The position probability of each participant in different areas of the track is statistically analyzed, and a set of position distribution probability pipelines is constructed.

4. The event tracking and timing shooting method according to claim 3, characterized in that, Based on real-time status information and a pre-set circular track model, the system analyzes and predicts the target trajectory of each participant on the track. Combining the target trajectory with the relative positions of the participants, it predicts overtaking behavior and the probability of overtaking, obtaining the status prediction results, including: Based on the instantaneous speed and real-time position of the participants in the real-time status information, the linear motion of the participants is converted into the curvilinear motion of the tangential and normal vectors through the preset circular track model. The track curvature constraint is calculated and the target motion trajectory of each participant on the track is predicted. Based on the target motion trajectory and the relative positions of the participants, analyze the interaction forces between the participants and construct an interactive motion model; The interactive motion model is used to calculate the attractive and repulsive forces exerted on participants behind by participants in front, and to predict the overtaking behavior and probability of participants behind. By combining the target's trajectory, overtaking behavior, and overtaking probability, the state prediction result is obtained.

5. The event tracking and timing shooting method according to claim 1, characterized in that, Within the set of location distribution probability pipelines, the coverage value of cameras at different locations for each location distribution probability pipeline is analyzed. The coverage value game process of different cameras is simulated, and the camera combination with the optimal coverage value is selected, including: Based on the location distribution of participants in each pipe of the location distribution probability pipe set, analyze the coverage of each camera on the pipe and the overlap coverage between different cameras, and calculate the area coverage value of each camera. Based on the coverage value of the area, the cooperative game process between each camera is simulated to select the camera combination with the optimal coverage value.

6. The event tracking and timing shooting method according to claim 5, characterized in that, The process involves analyzing the participant location distribution in each pipe of the probability pipe set, examining the coverage of each camera within the pipe, and the overlap of coverage between different cameras. The regional coverage value of each camera is then calculated, including: Based on the participant position distribution in each pipe of the position distribution probability pipe set, the coverage of each camera to the pipe is analyzed through a preset simulated charge value assessment model. Areas with charge density greater than a preset charge threshold are simulated as high charge density areas. The potential energy value of each camera affected by charge at the corresponding position is calculated to obtain the corresponding first coverage value. Analyze the spatial relationships between different cameras, identify the overlapping coverage areas between different cameras, and reduce the first coverage value of the overlapping coverage area according to the area to obtain the second coverage value. By combining the first coverage value and the second coverage value, the regional coverage value of each camera is obtained.

7. The event tracking and timing shooting method according to claim 1, characterized in that, The analysis of the participants' real-time status information, assigning at least one tracking role and tracking target to each camera in the camera array, yields a tracking and shooting strategy, including: Analyze the real-time status information of the participants and calculate the comprehensive performance of each camera in the camera combination under different tracking role configurations through a preset role evaluation model; Based on the overall performance, the tracking target conflict between cameras is analyzed, and at least one tracking role and tracking target are assigned to each camera to obtain a tracking shooting strategy.

8. The event tracking and timing shooting method according to claim 7, characterized in that, Based on the overall performance, the tracking target conflict between cameras is analyzed, and at least one tracking role and tracking target are assigned to each camera to obtain a tracking shooting strategy, including: When multiple cameras compete for the same tracking target, the camera with the best overall performance is selected to match the corresponding tracking target. By combining the tracking target matched to each camera with the camera position, at least one tracking role is assigned to each camera; According to the preset time window, the tracking role and tracking target of each camera are dynamically adjusted to obtain the tracking shooting strategy.

9. A method for tracking and timing a sporting event according to claim 1, characterized in that, According to the aforementioned tracking and shooting strategy, the corresponding camera is controlled to track the target in order to perform real-time tracking and timing shooting of the event process, including: Generate corresponding tracking shooting instructions based on the tracking role and target information assigned to each camera in the tracking shooting strategy; According to the tracking and shooting instructions, the corresponding camera is controlled to track the target in order to conduct real-time tracking and timing shooting of the event.

10. A sports event tracking and timing shooting system, characterized in that, A method for implementing a sports event tracking and timing shooting method as described in any one of claims 1 to 9, comprising: The status analysis module analyzes the real-time status of participants based on real-time data collected from at least two mobile high-speed cameras and real-time timing data of the event, and obtains real-time status information through a preset status recognition model. The position analysis module, based on the real-time status information, predicts and analyzes the movement trajectory according to the preset circular track model, obtains the target movement trajectory of the participants, predicts the overtaking behavior and overtaking probability between the participants, calculates the dynamic adjustment factor of different areas in the track and constructs the target movement model, calculates the probability of the participants at each position, and constructs a set of position distribution probability pipelines. The camera value game module analyzes the coverage value of cameras at different locations to each location distribution probability pipeline in the set of location distribution probability pipelines, simulates the coverage value game process of different cameras, and selects the camera combination with the optimal coverage value. The tracking and shooting strategy formulation module analyzes the real-time status information of the participants and assigns at least one tracking role and tracking target to each camera in the camera combination to obtain the tracking and shooting strategy. The tracking and shooting module controls the corresponding camera to track the target according to the tracking and shooting strategy, so as to carry out real-time tracking and timing shooting of the event process.

Images