Volleyball spiking training detection method and system

Abstract

The invention provides a volleyball spiking training detection method and system. The volleyball spiking training detection method comprises the following steps: extracting an action image frame sequence from a spiking action video of a target athlete; based on the static background edge of the target athlete for volleyball training and the visual edge in each action image frame in the action image frame sequence, constructing the recognition constraint of the action contour during volleyball spiking training of the target athlete; according to the recognition constraint and a change curve of the motion contour of the target athlete corresponding to adjacent motion image frames in the motion image frame sequence, performing verification analysis on the arm motion trajectory of the target athlete during volleyball spiking training to obtain a plurality of distortion verification points of the arm motion of the target athlete; and repairing the position of each distortion check point on the arm action track to obtain a confidence action contour sequence when the target athlete carries out volleyball spiking training. According to the scheme, accurate capture of volleyball spiking actions of the target athlete can be realized under multi-dimensional accompanying interference on the basis of the action contour sequence.

Description

Technical Field

[0001] The present application relates to the field of sports training detection technology, and more specifically, to a volleyball spiking training detection method and system. Background Art

[0002] As people pay more and more attention to health and sports performance, sports training testing has emerged and developed rapidly. In traditional sports training, the evaluation of athletes' training effects often relies on the coach's experience and subjective judgment, and lacks accurate quantitative data support. However, modern sports training has put forward higher requirements for scientificity and precision. With the help of advanced sensors, biomechanical analysis systems and big data analysis technologies, it is possible to monitor athletes' physical condition, sports technical movements and other multi-dimensional information in real time, thereby providing a strong basis for the formulation of personalized training plans and helping athletes improve their athletic ability more efficiently.

[0003] In existing sports training tests, especially in motion capture, motion capture is mainly based on computer vision principles, using cameras to capture human movements, extracting key information through image processing and analysis algorithms, and extracting the target athlete's motion posture based on multi-view geometry principles. However, in volleyball spiking training tests, when capturing the target athlete's high-speed spiking motion, due to the multi-dimensional accompanying interference caused by the rapid movement of the target athlete's limbs (such as static background edge interference and contour occlusion between consecutive frames), the target athlete's arm movement trajectory has a positioning deviation, resulting in misidentification of the target athlete's volleyball spiking motion capture. Therefore, how to achieve accurate capture of the target athlete's volleyball spiking motion under multi-dimensional accompanying interference has become a difficult problem faced by the industry. Summary of the Invention

[0004] The present application provides a volleyball spiking training detection method and system, which can accurately capture the volleyball spiking movements of a target athlete under multi-dimensional interference.

[0005] In a first aspect, the present application provides a volleyball spiking training detection method, comprising the following steps: Acquire a spiking action video of a target athlete during volleyball spiking training, and extract an action image frame sequence of the target athlete from the spiking action video; extracting static background edges from a picture taken when a target athlete is performing volleyball training, and constructing recognition constraints for a motion contour of the target athlete during the volleyball spiking training based on the static background edges and visible edges within each action image frame in the action image frame sequence; determining a change curve of a target athlete's motion profile corresponding to adjacent motion image frames in the motion image frame sequence, and performing verification analysis on an arm motion trajectory of the target athlete during volleyball spiking training based on the recognition constraint and the change curves corresponding to each adjacent motion image frame, thereby obtaining a plurality of distortion verification points of the target athlete's arm motion during the spiking training; The positions of the distorted check points on the arm motion trajectory are repaired, thereby obtaining a reliable motion contour sequence of the target athlete during volleyball spiking training.

[0006] In some embodiments, extracting static background edges from a picture taken during volleyball training of a target athlete specifically includes: Obtain a fixed background image during volleyball spiking training; The edges of all edge areas in the fixed background image are extracted as static background edges in a picture when photographing a target athlete during volleyball training.

[0007] In some embodiments, the recognition constraints for constructing the action contour of the target athlete during volleyball spiking training based on the static background edge and the visible edges in each action image frame in the action image frame sequence specifically include: extracting visible edges of each action image frame in the action image frame sequence; Determining a target edge set of an action contour of a target athlete performing volleyball spiking training according to a difference relationship between a visible edge of each action image frame and an edge of the static background; Marking each target edge in the target edge set in the corresponding action image frame to obtain all action marked images; The recognition constraints of the target player's action contour during volleyball spiking training are determined according to the motion vectors of the target edge points in each adjacent action-labeled image.

[0008] In some embodiments, determining a change curve of a target athlete's motion profile corresponding to adjacent motion image frames in the motion image frame sequence specifically includes: Extracting an action contour from each action image frame in the action image frame sequence to obtain an action contour for each action image frame; Performing point-to-point matching on the motion contours of each adjacent motion image frame to obtain the temporal mapping relationship between the contour points of each adjacent motion image frame; For each adjacent image frame, a contour point movement description is performed based on a time sequence mapping relationship corresponding to the adjacent image frames, thereby obtaining a change curve corresponding to the adjacent motion image frames in the motion image frame sequence.

[0009] In some embodiments, based on the recognition constraints and the change curves corresponding to each adjacent motion image frame, the arm movement trajectory of the target athlete during volleyball spiking training is verified and analyzed, thereby obtaining multiple distortion verification points of the target athlete's arm movement during the spiking training process, specifically including: Perform time-series fitting on the corresponding change curves of all adjacent motion image frames to obtain the arm movement trajectory of the target athlete during volleyball spiking training; extracting multiple abnormal trajectory segments during the target athlete's spiking training from the arm motion trajectory based on a preset time window; The identification constraint is used to extract trajectory mutation position points from each trajectory abnormal segment as distortion check points of the target athlete's arm movement during the spiking training process, thereby obtaining multiple distortion check points of the target athlete's arm movement during the spiking training process.

[0010] In some embodiments, a high-definition camera is used to obtain a video of a target athlete performing volleyball spiking training.

[0011] In some embodiments, an open source computer vision library of video processing software is used to extract a sequence of action image frames containing complete action information of the athlete from the spiking action video frame by frame at set time intervals.

[0012] In a second aspect, the present application provides a volleyball spiking training and detection system, comprising: An acquisition module is used to acquire a spiking action video of a target athlete during volleyball spiking training, and extract an action image frame sequence of the target athlete from the spiking action video; a processing module for extracting static background edges from a picture taken when a target athlete is performing volleyball training, and constructing recognition constraints for a motion contour of the target athlete during the volleyball spiking training based on the static background edges and visible edges within each action image frame in the action image frame sequence; The processing module is further configured to determine a change curve of a target athlete's motion profile corresponding to adjacent motion image frames in the motion image frame sequence, and to perform verification analysis on an arm motion trajectory of the target athlete during volleyball spiking training based on the recognition constraint and the change curves corresponding to each adjacent motion image frame, thereby obtaining a plurality of distortion verification points of the target athlete's arm motion during the spiking training. The execution module is used to repair the position of each distortion check point on the arm movement trajectory, thereby obtaining a reliable movement contour sequence of the target athlete when performing volleyball spiking training.

[0013] In a third aspect, the present application provides a computer device, comprising a memory and a processor, wherein the memory stores a code, and the processor is configured to obtain the code and execute the above-mentioned volleyball spiking training detection method.

[0014] In a fourth aspect, the present application provides a computer-readable storage medium storing a computer program, which implements the above-mentioned volleyball spiking training detection method when executed by a processor.

[0015] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects: In the volleyball spiking training detection method and system provided by the present application, first, a spiking action video of a target athlete during volleyball spiking training is obtained, and a motion image frame sequence of the target athlete is extracted from the spiking action video; secondly, static background edges in the shot of the target athlete during volleyball training are extracted, and recognition constraints of the motion contour of the target athlete during volleyball spiking training are constructed based on the static background edges and the visible edges in each motion image frame in the motion image frame sequence; then, a change curve of the motion contour of the target athlete corresponding to adjacent motion image frames in the motion image frame sequence is determined, and based on the recognition constraints and the change curves corresponding to each adjacent motion image frame, the arm motion trajectory of the target athlete during volleyball spiking training is verified and analyzed, thereby obtaining a plurality of distortion verification points of the arm motion during the target athlete's spiking training; finally, the position of each distortion verification point on the arm motion trajectory is repaired, thereby obtaining a reliable motion contour sequence of the target athlete during volleyball spiking training.

[0016] It can be seen that the present application can realize the accurate capture of the target athlete's volleyball spiking action under multi-dimensional accompanying interference; first, the target athlete's action image frame sequence is extracted from the target athlete's spiking action video during volleyball spiking training, thereby providing a data basis for identifying the action standardization of the target athlete during volleyball spiking training in the video; secondly, based on the static background edge in the shooting picture when the target athlete is performing volleyball training and the visible edge of each action image frame in the action image frame sequence, the recognition constraint of the target athlete's action contour during volleyball spiking training is constructed to limit and optimize the recognition process of the target athlete's action contour, prevent the introduction of static background edges and the loss of key action parts, and thus avoid the arm movement trajectory caused by the multi-dimensional accompanying interference generated by the rapid movement of the target athlete's limbs. Current positioning deviation; then, based on the recognition constraint and the change curve corresponding to each adjacent motion image frame, the arm movement trajectory of the target athlete in the spiking action is verified and analyzed, and then a plurality of distortion verification points of the arm movement of the target athlete in the spiking training process are obtained, thereby providing an accurate intervention entry for the correction and dynamic reconstruction of the subsequent movement trajectory, and using the adjacent trajectory information for fitting correction, thereby improving the authenticity and coherence of the movement data, so as to accurately capture the movement of the target athlete when performing volleyball spiking; finally, the position of each distortion verification point on the arm movement trajectory is repaired, and then a reliable movement contour sequence of the target athlete when performing volleyball spiking training is obtained; in summary, the technical solution provided by the present application can realize the accurate capture of the movement of the target athlete in volleyball spiking under multi-dimensional accompanying interference. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is an exemplary flow chart of a volleyball spiking training detection method according to some embodiments of the present application; Figure 2 is an exemplary flow chart of determining a static background edge according to some embodiments of the present application; Figure 3 is an exemplary flow chart of determining a change curve according to some embodiments of the present application; Figure 4 is a structural diagram of a volleyball spiking training and detection system according to some embodiments of the present application; Figure 5 It is a structural diagram of a computer device for implementing a volleyball spiking training detection method according to some embodiments of the present application. DETAILED DESCRIPTION

[0018] In order to better understand the technical solution of the present application, the technical solution of the present application will be described in detail below with reference to the accompanying drawings and specific implementation methods.

[0019] refer to Figure 1, which is an exemplary flow chart of a volleyball spike training detection method according to some embodiments of the present application. The volleyball spike training detection method 100 mainly includes the following steps: In step 101, a spiking action video of a target athlete during volleyball spiking training is obtained, and an action image frame sequence of the target athlete is extracted from the spiking action video.

[0020] In specific implementation, first, a high-definition camera is used to obtain a volleyball spiking action video of a target athlete during volleyball spiking training. Then, the video processing software OpenCV (OpenCV) is used to extract a sequence of action image frames containing the athlete's complete action information from the spiking action video frame by frame at a set time interval. The time interval can be set according to actual needs and is not limited here.

[0021] It should be noted that the action image frame sequence in the present application includes multiple action image frames, and the action image frame represents a single static image extracted from the video sequence. Each action image frame represents the picture content of the spiking action video at a specific time point, and the action image frame is used to analyze and process the dynamic content in the video.

[0022] In step 102, static background edges are extracted from the footage captured when the target athlete is performing volleyball training, and recognition constraints for the action contour of the target athlete during volleyball spiking training are constructed based on the static background edges and the visible edges within each action image frame in the action image frame sequence.

[0023] It should be noted that the static background edge in this application refers to the edge feature information from the background part in the volleyball spiking training shooting picture, which is used to distinguish the real action edge of the target athlete from the background edge in the background that may interfere with the identification. These background edges may be mistaken for being part of the athlete's action in the motion image of the target athlete, but in fact have nothing to do with the athlete's action.

[0024] In some embodiments, reference Figure 2 As shown in FIG. 1 , this figure is an exemplary flow chart of determining a static background edge according to some embodiments of the present application. In this embodiment, extracting a static background edge in a shot of a target athlete undergoing volleyball training can be achieved by using the following steps: In step 1021, a fixed background image during volleyball spiking training is acquired; In step 1022, edges of all edge regions in the fixed background image are extracted as static background edges in a picture taken when photographing the target athlete during volleyball training.

[0025] In specific implementation, first, a fixed background image during volleyball spiking training can be obtained through a high-definition camera; secondly, edge features of all edge areas in the fixed background image are extracted as static background edges in the picture when the target athlete is performing volleyball training. That is, the edges of all edge areas can be extracted from the fixed background image through the Canny edge detection in the existing edge detection algorithm as the static background edges in the picture when the target athlete is performing volleyball training. The static background edges in this application are the edges in the image taken when there are no trainees in the volleyball spiking training venue. In addition, the edge detection algorithm can also use the Sobel operator and the Laplacian operator, which are not limited here.

[0026] It should be noted that, in this embodiment, the fixed background image is an image taken when there are no trainers at the fixed volleyball training ground. The fixed background image can effectively identify the external interference to the training status of the target athlete.

[0027] In some embodiments, constructing recognition constraints for the target athlete's action contour during volleyball spiking training based on the static background edge and the visible edges within each action image frame in the action image frame sequence can be achieved by the following steps, namely: extracting visible edges of each action image frame in the action image frame sequence; Determining a target edge set of an action contour of a target athlete performing volleyball spiking training according to a difference relationship between a visible edge of each action image frame and an edge of the static background; Marking each target edge in the target edge set in the corresponding action image frame to obtain all action marked images; The recognition constraints of the target player's action contour during volleyball spiking training are determined according to the motion vectors of the target edge points in each adjacent action-labeled image.

[0028] It should be noted that the recognition constraints in this application represent discriminative parameters used to assist in identifying the target athlete's action contour. The recognition constraints can limit and optimize the recognition process of the target athlete's action contour to prevent the introduction of static background edges and the loss of key action parts.

[0029] In specific implementation, first, the action edge of the target athlete can be extracted from each motion image frame as a visible edge through the Canny edge detection in the existing edge detection algorithm, and then the visible edge of each action image frame in the action image frame sequence is obtained. In addition, the edge detection algorithm can also use the Sobel operator and the Laplacian operator, which are not limited here; secondly, for each action image frame, the Hausdorff distance between the visible edge of the action image frame and the static background edge can be used as the difference relationship between the visible edge of the action image frame and the static background edge, and the visible edge with a Hausdorff distance less than a set threshold is classified as the target edge of the action contour of the target athlete when performing volleyball spiking training, and then the target edge set of the action contour of the target athlete when performing volleyball spiking training is obtained. The Hausdorff distance is a measure of the similarity between two point sets. Indicators are commonly used in image processing and edge matching, wherein the threshold can be set according to actual needs or set by machine learning based on the Hausdorff distance of historically identified target edges, which will not be elaborated here; then, the image processing tool OpenCV can be used to mark each target edge in the target edge set in the corresponding action image frame to obtain all action mark images, and the action mark image represents the image after the action image frame is marked with target edges; finally, the recognition constraints of the action contour of the target athlete during volleyball spiking training are determined according to the motion vectors of the target edge points in each adjacent action mark image, that is: the existing optical flow method is used to calculate the motion vectors of the target edge points in the adjacent action mark images, the motion vectors include displacement direction and displacement speed, and the standard deviation of all displacement speeds is used as the recognition constraint of the action contour of the target athlete during volleyball spiking training.

[0030] It should be noted that, in this embodiment, the visible edge represents edge information directly extracted from the action image frame, and the visible edge includes the edge of the target athlete's motion contour and the background edge. The visible edge is the contour line corresponding to the area where the pixel grayscale changes significantly in the image; in this embodiment, the target edge set represents a combination of multiple target edges, wherein the target edge is the edge of the motion contour of the target athlete when moving.

[0031] In step 103, a change curve of the target athlete's motion profile corresponding to adjacent motion image frames in the motion image frame sequence is determined, and based on the recognition constraint and the change curves corresponding to each adjacent motion image frame, the arm motion trajectory of the target athlete during volleyball spiking training is verified and analyzed, thereby obtaining multiple distortion verification points of the target athlete's arm motion during the spiking training.

[0032] It should be noted that the change curve in this application represents the evolution trajectory formed by the change of the target athlete's movement contour over time in adjacent motion image frames. The change curve reflects the spatial change trend of the target athlete's morphological boundary during the movement process. By determining the change curve, the structural dynamics of the target athlete during the movement process can be quantified and expressed.

[0033] In some embodiments, reference Figure 3 As shown in FIG. 1 , this figure is an exemplary flow chart of determining a change curve according to some embodiments of the present application. In this embodiment, determining a change curve of a target athlete's action profile corresponding to adjacent motion image frames in the motion image frame sequence can be achieved by using the following steps: First, in step 1031, action contour extraction is performed on each action image frame in the action image frame sequence to obtain the action contour of each action image frame; Then, in step 1032, point-to-point matching is performed on the motion contours of each adjacent motion image frame to obtain a temporal mapping relationship between the contour points of each adjacent motion image frame; Finally, in step 1033, for each adjacent image frame, a contour point movement description is performed based on the time sequence mapping relationship corresponding to the adjacent image frames, thereby obtaining a change curve corresponding to the adjacent motion image frames in the motion image frame sequence.

[0034] In the specific implementation, first, the Canny operator in the image edge detection algorithm is used to extract the action contour of each action image frame in the action image frame sequence to obtain the action contour of each action image frame. In addition, the Sobel operator and the Laplacian operator can also be used to extract the action contour of each action image frame in the action image frame sequence, which is not limited here; secondly, the action contours of each adjacent action image frame are point-to-point matched to obtain the temporal mapping relationship between the contour points of each adjacent motion image frame, that is: the point set matching algorithm is used to match the contour points of any two adjacent action image frames to the corresponding relationship, and then the temporal mapping relationship between the contour points of each adjacent motion image frame is obtained. For example, for any two adjacent action image frames, the iterative nearest point algorithm in the point set matching algorithm is used to match the first action image frame to the second action image frame. A one-to-one mapping relationship is established between each contour point in the image frame and the contour point in the second motion image frame that is closest to the contour point and has a similar local structure, to obtain a temporal mapping relationship. The temporal mapping relationship includes multiple mapping point pairs, and the mapping point pairs are corresponding contour points of two adjacent motion image frames. The point set matching algorithm can adopt an iterative nearest point matching algorithm or a matching method based on an edge distance graph, which is not limited here. Finally, for each adjacent image frame, a contour point movement description is performed based on the temporal mapping relationship corresponding to the adjacent image frames, and then a change curve corresponding to the adjacent motion image frames in the motion image frame sequence is obtained, that is: for each adjacent image frame, a B-spline curve is used to perform path fitting on each mapping point pair in the temporal mapping relationship corresponding to the adjacent image frames, and all path sets obtained by fitting are used as the change curve corresponding to the adjacent motion image frames.

[0035] It should be noted that, in this embodiment, the action contour set represents a combination of multiple action contours, and the action contour represents a set of curves that describe the action boundaries when the target athlete moves; in this embodiment, the temporal mapping relationship represents the relationship between corresponding elements in two adjacent motion image frames, which is used to characterize the changing trajectory of the elements in the time dimension. In volleyball spiking training detection, the temporal mapping relationship usually refers to the spatial position mapping relationship between the contour points of the target athlete in the action image frame and the adjacent image frames. The essence of this temporal mapping relationship is to find how the target structure moves or deforms continuously in time-continuous frames, providing a data basis for subsequent action trajectory modeling, distortion point detection and dynamic consistency analysis.

[0036] In some embodiments, based on the recognition constraints and the change curves corresponding to each adjacent motion image frame, the arm movement trajectory of the target athlete during volleyball spiking training is verified and analyzed, thereby obtaining multiple distortion verification points of the target athlete's arm movement during the spiking training. This can be achieved by the following steps, namely: Perform time-series fitting on the corresponding change curves of all adjacent motion image frames to obtain the arm movement trajectory of the target athlete during volleyball spiking training; extracting multiple abnormal trajectory segments during the target athlete's spiking training from the arm motion trajectory based on a preset time window; The identification constraint is used to extract trajectory mutation position points from each trajectory abnormal segment as distortion check points of the target athlete's arm movement during the spiking training process, thereby obtaining multiple distortion check points of the target athlete's arm movement during the spiking training process.

[0037] In the present application, the distortion check point represents a key position point with abnormal characteristics in the arm motion trajectory. Specifically, the distortion check point is a key position point with abnormal characteristics in the arm motion trajectory when the motion contour of the target athlete is identified during volleyball spiking training. The distortion check point reflects the position where the arm motion trajectory suddenly changes and deviates from the normal motion pattern. By determining the distortion check point, the motion recognition error existing in the motion process can be effectively identified and quantified; in volleyball spiking training detection, when capturing the high-speed spiking motion of the target athlete, due to the multi-dimensional accompanying interference generated by the rapid movement of the target athlete's limbs (such as static background edge interference and contour occlusion between consecutive frames), the arm motion trajectory of the target athlete has a positioning deviation, resulting in incoherent or abrupt trajectory points (i.e., distortion check points) when identifying the motion contour. By determining the distortion check point, an accurate intervention entry can be provided for the correction and dynamic reconstruction of the subsequent motion trajectory, and then the adjacent frame information can be used for fitting correction, thereby improving the authenticity and consistency of the motion data.

[0038] In the specific implementation, first, the change curves corresponding to all adjacent motion image frames are time-series fitted to obtain the arm movement trajectory of the target athlete during volleyball spiking training, that is, the B-spline curve fitting algorithm can be used to time-series fit the change curves corresponding to all adjacent motion image frames, and the curve trajectory obtained by fitting is used as the arm movement trajectory of the target athlete during volleyball spiking training. The time-series fitting represents the process of fitting in time sequence. In addition, polynomial curve fitting can also be used to time-series fit the change curves corresponding to all adjacent motion image frames, which will not be repeated here; then, based on a preset time window, multiple trajectory abnormal segments in the target athlete's spiking training process are extracted from the arm movement trajectory, that is, the arm movement trajectory is divided into multiple arm movement trajectory segments in time sequence by the preset time window, the first-order velocity change value of each arm movement trajectory segment is compared with a set threshold, and the arm movement trajectory segment with a first-order velocity change value greater than the set threshold is used as the trajectory abnormal segment in the target athlete's spiking training process, thereby obtaining the target athlete's spiking training process. There are multiple trajectory abnormality segments during the training process, and the preset time window can be set according to actual needs, wherein the first-order velocity variation mean value during the normal movement of the historical athlete can be set as a threshold value, and the trajectory abnormal segment represents a local trajectory segment in the arm movement trajectory that is inconsistent with the normal movement law; finally, the identification constraint is used to extract the trajectory mutation position point from each trajectory abnormal segment as the distortion check point of the arm movement during the target athlete's spiking training, thereby obtaining multiple distortion check points of the arm movement during the target athlete's spiking training, that is: the trajectory velocity variation mean value within each trajectory abnormal segment is calculated, for each trajectory abnormal segment, the identification constraint and the trajectory velocity variation mean value within the trajectory abnormal segment are used as the standard deviation parameter and mean parameter of the Laida criterion respectively, thereby obtaining the abnormal discrimination interval corresponding to each trajectory abnormal segment, and extracting the trajectory points in each trajectory abnormal segment whose trajectory velocity value is not in the corresponding abnormal discrimination interval as the distortion check point of the arm movement during the target athlete's spiking training, and the trajectory velocity value represents the rate at which the spatial position of the arm changes over time during the movement.

[0039] It should be noted that, in this embodiment, the arm motion trajectory represents the spatial motion path formed by the target athlete's arm in continuous image frames during volleyball spiking training; in this embodiment, the abnormality discrimination interval represents the range of judging whether the target athlete's arm swing is abnormal during spiking.

[0040] It should also be noted that the verification analysis in the present application represents the process of identifying abnormal points in the arm swing of the target athlete during the spiking process, wherein the arm movement trajectory of the target athlete during volleyball spiking training is verified and analyzed based on the identification constraints and the change curves corresponding to each adjacent motion image frame, that is: the change curves corresponding to all adjacent motion image frames are time-series fitted to obtain the arm movement trajectory of the target athlete during volleyball spiking training; based on a preset time window, multiple trajectory abnormal segments during the target athlete's spiking training are extracted from the arm movement trajectory; the identification constraints extract trajectory mutation position points from each trajectory abnormal segment as distortion verification points of the target athlete's arm movement during the spiking training, and then multiple distortion verification points of the target athlete's arm movement during the spiking training are obtained, that is, the distortion verification points are used as the results of the verification analysis, and then the verification analysis of the arm movement trajectory in the target athlete's spiking action is completed.

[0041] In step 104, the positions of the distortion check points on the arm motion trajectory are repaired, thereby obtaining a reliable motion contour sequence of the target athlete during volleyball spiking training.

[0042] It should be noted that the trusted action profile sequence in this application represents a set of temporal variation data of the trustworthy motion profile of the target athlete in the volleyball spiking action, specifically, the profile of the athlete's arm swing in the volleyball spiking action. By determining the trusted action profile sequence, the subjective action observation of the target athlete during the volleyball spiking can be converted into objective data, providing measurable technical evaluation indicators for identifying the training effect of the target athlete.

[0043] In some embodiments, the positions of the distortion check points on the arm motion trajectory are repaired to obtain a reliable motion contour sequence for the target athlete during volleyball spiking training, which can be achieved by: For each distortion check point, extract the two trajectory points before and after the distortion check point on the arm movement trajectory, and correct the position of the distortion check point by using the two extracted trajectory points, and use the corrected position as the new position of the distortion check point, thereby obtaining the distortion check points after correction of each position; The distortion check points after correction of each position are re-embedded into the arm motion trajectory, and the motion contour of the corresponding motion image frame is updated; The motion contours of each corrected motion image frame are integrated in time sequence to generate a reliable motion contour sequence of the target athlete when performing volleyball spiking training.

[0044] In a specific implementation, first, for each distortion check point, two trajectory points before and after the distortion check point on the arm motion trajectory are extracted, and the position of the distortion check point is corrected using the two extracted trajectory points, and the corrected position is used as the new position of the distortion check point, thereby obtaining the distortion check points with corrected positions. That is, for each distortion check point, two trajectory points before and after the distortion check point on the arm motion trajectory are extracted, and based on the two extracted trajectory points, a local smooth curve corresponding to the distortion check point is constructed using the existing cubic spline interpolation method, and the position of the distortion check point is optimized with the goal of minimizing the second-order derivative energy function of the local smooth curve, and the corrected position is used as the new position of the distortion check point, thereby obtaining the distortion check points with corrected positions; then, each corrected distortion check point is re-embedded into the arm motion trajectory, and the motion contour of the corresponding motion image frame is updated using the contour update algorithm of the image processing tool OpenCV (such as chain code-based contour reconstruction); finally, the motion contours of each corrected motion image frame are integrated in chronological order using the image processing tool OpenCV to generate a reliable motion contour sequence of the target athlete when performing volleyball spiking training.

[0045] It should be noted that the local smooth curve in this embodiment represents a curve constructed within a local range, a mathematically continuous and second-order differentiable curve segment (local smooth curve) constructed by the cubic spline interpolation algorithm. The local smooth curve retains the acceleration change and joint curvature, and can be used to constrain and correct the position of the intermediate distortion check point so that its motion trajectory conforms to the physical laws of human body movements.

[0046] In addition, in another aspect of the present application, in some embodiments, the present application provides a volleyball spike training detection system, referring to Figure 4 , which is a schematic diagram of the structure of a volleyball spike training and detection system according to some embodiments of the present application. The volleyball spike training and detection system 200 includes: an acquisition module 201, a processing module 202, and an execution module 203, which are described as follows: Acquisition module 201, in this application, acquisition module 201 is mainly used to acquire a spiking action video of a target athlete during volleyball spiking training, and extract an action image frame sequence of the target athlete from the spiking action video; Processing module 202, in this application, is mainly used to extract static background edges in the image captured when the target athlete is performing volleyball training, and to construct recognition constraints for the target athlete's action contour when performing volleyball spiking training based on the static background edges and the visible edges in each action image frame in the action image frame sequence; The processing module 202 is further configured to determine a change curve of a target athlete's motion profile corresponding to adjacent motion image frames in the motion image frame sequence, and perform verification analysis on the target athlete's arm motion trajectory during volleyball spiking training based on the recognition constraint and the change curves corresponding to each adjacent motion image frame, thereby obtaining a plurality of distortion verification points of the target athlete's arm motion during the spiking training. The execution module 203 in this application is mainly used to repair the position of each distortion check point on the arm movement trajectory, thereby obtaining a reliable movement contour sequence for the target athlete when performing volleyball spiking training.

[0047] In addition, the present application also provides a computer device, which includes a memory and a processor, wherein the memory stores code, and the processor is configured to obtain the code and execute the above-mentioned volleyball spiking training detection method.

[0048] In some embodiments, reference Figure 5 , which is a schematic diagram of the structure of a computer device for implementing a volleyball spike training detection method according to some embodiments of the present application. The volleyball spike training detection method in the above embodiment can be Figure 5 The computer device 300 shown in FIG. 1 is implemented as shown in FIG. 1 , and the computer device 300 includes at least one processor 301 , a communication bus 302 , a memory 303 , and at least one communication interface 304 .

[0049] The processor 301 may be a general-purpose central processing unit (CPU), or an application-specific integrated circuit (ASIC) or one or more processors for controlling the execution of the volleyball spiking training detection method of the present application.

[0050] The communication bus 302 may be used to transmit information between the aforementioned components.

[0051] Memory 303 may be, but is not limited to, a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, a random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, an optical disc storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), a magnetic disk or other magnetic storage device, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer. Memory 303 may be independent and connected to processor 301 via communication bus 302. Memory 303 may also be integrated with processor 301.

[0052] Memory 303 is used to store program code for executing the solution of the present application, and is controlled by processor 301 for execution. Processor 301 is used to execute the program code stored in memory 303. The program code may include one or more software modules. The determination of the volleyball spiking training detection method in the above embodiment can be implemented by processor 301 and one or more software modules in the program code in memory 303.

[0053] The communication interface 304 uses any device such as a transceiver to communicate with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area network (WLAN), etc.

[0054] In a specific implementation, as an example, a computer device may include multiple processors, each of which may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. A processor herein may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0055] The aforementioned computer device can be a general-purpose computer device or a dedicated computer device. In a specific implementation, the computer device can be a desktop computer, a portable computer, a network server, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. The embodiments of this application do not limit the type of computer device.

[0056] In addition, the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the above-mentioned volleyball spiking training detection method.

[0057] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present application.

[0058] Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims

1. A volleyball spike training detection method, characterized in that: The steps include: Acquire a spiking action video of a target athlete during volleyball spiking training, and extract an action image frame sequence of the target athlete from the spiking action video; extracting static background edges from a picture taken when a target athlete is performing volleyball training, and constructing recognition constraints for a motion contour of the target athlete during the volleyball spiking training based on the static background edges and visible edges within each action image frame in the action image frame sequence; determining a change curve of a target athlete's motion profile corresponding to adjacent motion image frames in the motion image frame sequence, and performing verification analysis on an arm motion trajectory of the target athlete during volleyball spiking training based on the recognition constraint and the change curves corresponding to each adjacent motion image frame, thereby obtaining a plurality of distortion verification points of the target athlete's arm motion during the spiking training; The positions of the distorted check points on the arm motion trajectory are repaired, thereby obtaining a reliable motion contour sequence of the target athlete during volleyball spiking training.

2. The method according to claim 1, wherein Extracting the static background edge in the picture of the target athlete during volleyball training specifically includes: Obtain a fixed background image during volleyball spiking training; The edges of all edge areas in the fixed background image are extracted as static background edges in a picture when photographing a target athlete during volleyball training.

3. The method according to claim 1, wherein The recognition constraints for constructing the action contour of the target athlete during volleyball spiking training based on the static background edge and the visible edges in each action image frame in the action image frame sequence specifically include: extracting visible edges of each action image frame in the action image frame sequence; Determining a target edge set of an action contour of a target athlete performing volleyball spiking training according to a difference relationship between a visible edge of each action image frame and an edge of the static background; Marking each target edge in the target edge set in the corresponding action image frame to obtain all action marked images; The recognition constraints of the target player's action contour during volleyball spiking training are determined according to the motion vectors of the target edge points in each adjacent action-labeled image.

4. The method according to claim 1, wherein Determining a change curve of the target athlete's motion profile corresponding to adjacent motion image frames in the motion image frame sequence specifically includes: Extracting an action contour from each action image frame in the action image frame sequence to obtain an action contour for each action image frame; Performing point-to-point matching on the motion contours of each adjacent motion image frame to obtain the temporal mapping relationship between the contour points of each adjacent motion image frame; For each adjacent image frame, a contour point movement description is performed based on a time sequence mapping relationship corresponding to the adjacent image frames, thereby obtaining a change curve corresponding to the adjacent motion image frames in the motion image frame sequence.

5. The method according to claim 1, wherein Based on the recognition constraints and the change curves corresponding to each adjacent motion image frame, the arm movement trajectory of the target athlete during volleyball spiking training is verified and analyzed, thereby obtaining multiple distortion verification points of the target athlete's arm movement during the spiking training process, specifically including: Perform time-series fitting on the corresponding change curves of all adjacent motion image frames to obtain the arm movement trajectory of the target athlete during volleyball spiking training; extracting multiple abnormal trajectory segments during the target athlete's spiking training from the arm motion trajectory based on a preset time window; The identification constraint is used to extract trajectory mutation position points from each trajectory abnormal segment as distortion check points of the target athlete's arm movement during the spiking training process, thereby obtaining multiple distortion check points of the target athlete's arm movement during the spiking training process.

6. The method according to claim 1, wherein The spiking action video of the target athlete during volleyball spiking training is obtained through a high-definition camera.

7. The method according to claim 1, wherein An open-source computer vision library of video processing software is used to extract a sequence of action image frames containing complete action information of the athlete from the spiking action video frame by frame at set time intervals.

8. A volleyball spiking training and detection system, characterized in that: include: An acquisition module is used to acquire a spiking action video of a target athlete during volleyball spiking training, and extract an action image frame sequence of the target athlete from the spiking action video; a processing module for extracting static background edges from a picture taken when a target athlete is performing volleyball training, and constructing recognition constraints for a motion contour of the target athlete during the volleyball spiking training based on the static background edges and visible edges within each action image frame in the action image frame sequence; The processing module is further configured to determine a change curve of a target athlete's motion profile corresponding to adjacent motion image frames in the motion image frame sequence, and to perform verification analysis on an arm motion trajectory of the target athlete during volleyball spiking training based on the recognition constraint and the change curves corresponding to each adjacent motion image frame, thereby obtaining a plurality of distortion verification points of the target athlete's arm motion during the spiking training. The execution module is used to repair the position of each distortion check point on the arm movement trajectory, thereby obtaining a reliable movement contour sequence of the target athlete when performing volleyball spiking training.

9. A computer device, characterized in that: The computer device includes a memory and a processor, the memory stores codes, and the processor is configured to obtain the codes and execute the volleyball spiking training detection method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the volleyball spiking training detection method according to any one of claims 1 to 7 is implemented.

Images