CN115624735B - Auxiliary training system for ball games and working method - Google Patents
Auxiliary training system for ball games and working method Download PDFInfo
- Publication number
- CN115624735B CN115624735B CN202211247470.9A CN202211247470A CN115624735B CN 115624735 B CN115624735 B CN 115624735B CN 202211247470 A CN202211247470 A CN 202211247470A CN 115624735 B CN115624735 B CN 115624735B
- Authority
- CN
- China
- Prior art keywords
- training
- ball
- motion
- data
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63B—APPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
- A63B71/00—Games or sports accessories not covered in groups A63B1/00 - A63B69/00
- A63B71/06—Indicating or scoring devices for games or players, or for other sports activities
- A63B71/0619—Displays, user interfaces and indicating devices, specially adapted for sport equipment, e.g. display mounted on treadmills
- A63B71/0622—Visual, audio or audio-visual systems for entertaining, instructing or motivating the user
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/20—Scenes; Scene-specific elements in augmented reality scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
- G06V20/42—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items of sport video content
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63B—APPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
- A63B71/00—Games or sports accessories not covered in groups A63B1/00 - A63B69/00
- A63B71/06—Indicating or scoring devices for games or players, or for other sports activities
- A63B71/0619—Displays, user interfaces and indicating devices, specially adapted for sport equipment, e.g. display mounted on treadmills
- A63B71/0622—Visual, audio or audio-visual systems for entertaining, instructing or motivating the user
- A63B2071/0636—3D visualisation
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63B—APPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
- A63B71/00—Games or sports accessories not covered in groups A63B1/00 - A63B69/00
- A63B71/06—Indicating or scoring devices for games or players, or for other sports activities
- A63B71/0619—Displays, user interfaces and indicating devices, specially adapted for sport equipment, e.g. display mounted on treadmills
- A63B2071/0647—Visualisation of executed movements
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63B—APPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
- A63B71/00—Games or sports accessories not covered in groups A63B1/00 - A63B69/00
- A63B71/06—Indicating or scoring devices for games or players, or for other sports activities
- A63B71/0619—Displays, user interfaces and indicating devices, specially adapted for sport equipment, e.g. display mounted on treadmills
- A63B2071/0658—Position or arrangement of display
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30221—Sports video; Sports image
- G06T2207/30224—Ball; Puck
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30241—Trajectory
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Abstract
The invention discloses an auxiliary training system and a working method for ball games, which are characterized in that training videos are subjected to target detection, movement detail data such as a movement track, a batting drop point position, a ball discharging speed and the like are obtained through calculation, and visual data analysis and model display are performed on the movement details, so that a player can be helped to intuitively know the defects existing in training and the progress obtained by the player, meanwhile, the player can perform incremental learning based on training parameters to obtain an incremental learning result, the network parameters of a return ball route network model are continuously corrected based on the incremental learning result, the training system is applied to virtual training of the player, the player can enter a training mode after wearing a VR helmet, the weakness of the player can be overcome, and better training experience and training effects are brought to the player.
Description
Technical Field
The invention belongs to the technical field of sports training, and particularly relates to an auxiliary training system for ball games and a working method.
Background
With the development of economy and society and the arousal of health consciousness of the whole people, the sports population of China which frequently participates in sports exercise is nearly 5 hundred million. In all the construction and sport modes, ball games are always the most important and popular components of modern sports, and with the development of ball games becoming increasingly widespread and deep, professional athletes and ball game fans want to improve their own technical and tactical level.
The table tennis ball is used as a national ball in China, and has ultrahigh popularity. However, the current table tennis training can only rely on a coach to teach through experience basically, the mode is single, the term is boring and difficult to understand, the intuitiveness is poor, and because the number of students is large, the coach is generally one-to-many teaching, and the different situations of each student are difficult to care. Therefore, in the current situation that technology is advancing continuously and has penetrated into life, there is an urgent need to provide an auxiliary training system for ball games to improve training effect.
Disclosure of Invention
The invention aims to provide an auxiliary training system and a working method for ball games, which are used for solving the problems that the conventional sports training at present only depends on training experience teaching, the intuitiveness is poor, one-to-many teaching cannot be carried out on each athlete, and the like.
In order to achieve the above object, the present invention provides an auxiliary training system for ball games, which is characterized in that the auxiliary training system comprises a hardware part and a software part, the hardware part comprises a data acquisition device 110, a data transmission device 120, a computing platform 130, a data storage device 140, a display device 150 and a VR helmet 160, and the software part comprises a data communication transmission module 131, an image preprocessing module 132, a target detection module 133, a three-dimensional space positioning module 134, a motion analysis module 135, a motion training module 136, a data storage 137 and a three-dimensional display module 138.
And the data acquisition equipment is used for transmitting the motion video data to the computing platform through the data transmission equipment after shooting and acquiring the motion video of the athlete during training. The computing platform firstly preprocesses video images; then, carrying out target detection on each frame of image, identifying a target (ball) and positioning the position of the target (ball) in each frame of image; then combining the camera parameters and the depth information to calculate the position of the target (ball) in the three-dimensional space; and finally, tracking the position information of the target (ball) in the three-dimensional space at each moment, recording the complete movement track of the target, marking the position of the ball striking drop point at each time, and calculating the ball discharging speed of the sporter. In addition, the computing platform can continuously perform incremental learning based on the target complete motion track, the batting drop point position, the batting speed and the like to obtain an incremental learning result, continuously correct network parameters of the ball return route network based on the incremental learning result, and apply the network parameters to virtual training of athletes. The athlete can enter the training mode after wearing the VR headset.
The computing platform 130 stores the motion analysis data, the motion training data, and the raw video data obtained by the processing in the data storage device 140, and three-dimensionally displays the data on the display device 150 through three-dimensional modeling, so as to assist the coach in guiding the athlete training.
Further, the camera of the data acquisition equipment for shooting the motion video is a binocular stereo camera or an RGB-D camera, and the sampling frequency of the camera can be set according to the requirement;
further, the image preprocessing module comprises algorithms such as image difference, edge extraction, binarization processing, filtering denoising processing and the like;
further, the target detection module includes, but is not limited to, a conventional method-based target detection algorithm and a deep learning-based target detection algorithm; the target detection algorithm based on the traditional method is mainly a circle detection algorithm, including but not limited to a Hough transform algorithm or various improved Hough transform algorithms; the target detection algorithm based on deep learning comprises, but is not limited to, R-CNN, SPP-net, fast R-CNN, R-FCN, YOLO, SSD and the like;
further, the camera parameters comprise an internal parameter and an external parameter of the camera, and are obtained through camera calibration;
further, the acquisition mode of the depth information is determined by camera type selection, and can be obtained by calculation of a binocular stereo camera according to a triangulation method, and can also be obtained by a depth map of an RGB-D camera;
further, the data storage and three-dimensional display module automatically reads the type of the data storage device, sets a storage scheme corresponding to the data storage device according to different device types, and can establish a special training file for each person trained by the system so as to store personalized training data;
further, the transmission mode of the data transmission device can be wireless or wired to realize data transmission;
further, the target (ball) complete motion trail information can be used for analyzing the ball return habit of athletes and performing tactical training; the ball striking drop point position information can be used for analyzing the accuracy and stability of the ball striking of the athlete; the ball-out speed information can be used for analyzing whether the force of the athlete is correct or not and measuring whether physical training needs to be increased or not;
further, the athletic training module may set a dedicated virtual training opponent for each player, where the return route of the virtual training opponent is computationally generated by a return route network deployed on the computing platform. The ball return route network model can continuously perform incremental learning according to the motion data such as the complete motion trail of the target, the batting drop point position, the ball outlet speed and the like obtained by analyzing the real training video of the athlete.
Further, the ball return route network model can calculate and output the ball hitting point position, the ball hitting angle and the force of the virtual training opponent when returning the ball;
further, the incremental learning adopts a regularized incremental learning method, including but not limited to LwF algorithm and related improvement strategies;
further, the ball games include, but are not limited to, table tennis, badminton, billiards, snooker, and the like.
A method of operating an auxiliary training system for ball games, comprising:
step S01, shooting and transmitting a motion video; after the auxiliary training system is started, the camera shoots the whole training process of the athlete and sends the shot motion video to the computing platform through the data transmission equipment;
step S02: preprocessing an image; after receiving a video image, the computing platform carries out preprocessing operation on the video image through an image preprocessing module;
step S03: detecting a table tennis target; the target detection module carries out target detection on each preprocessed frame of image, identifies the table tennis ball and positions the table tennis ball in each frame of image;
step S04: positioning the table tennis ball in a three-dimensional space; after the position information of the table tennis ball in each frame of image is obtained, calculating the corresponding moment of each frame of image and the position of the table tennis ball in a three-dimensional space by combining camera parameters and depth information;
step S05: motion analysis; the method comprises the steps of recording the complete movement track of the table tennis by tracking the position information of the table tennis in a three-dimensional space at each moment, marking the position of a ball striking drop point at each time, and calculating the ball discharging speed of a sporter after each swing;
step S06: virtual training; the training module in the computing platform can set a dedicated virtual training opponent for each athlete, a return route of the virtual training opponent is calculated and generated by a return route network model deployed on the computing platform, the return route network model can calculate and output the position, the angle and the force of a ball hitting point when the virtual training opponent returns a ball, the return route network model can perform incremental learning according to target complete motion track, the position of the ball hitting point and the movement data of the ball discharging speed obtained by analyzing real training video of the athlete, an incremental learning result is obtained, and based on the incremental learning result, network parameters of the return route network are corrected and applied to virtual training of the athlete, and when the athlete wants to train, the training mode can be entered after wearing a VR helmet;
step S07: storing data and displaying three-dimensionally; the computing platform stores the motion data, the motion training data and the original video which are obtained through processing in the data storage equipment, and can be displayed on the display device in real time and three-dimensionally through three-dimensional modeling, so as to assist a coach to guide the athlete to train;
further, the motion data comprises complete motion track information of the table tennis, batting drop point position information and ball outlet speed information; the athletic training data includes a modified return route network.
The invention has the beneficial effects that:
the invention discloses an auxiliary training system and a working method for ball sports, which are characterized in that a training video is subjected to target detection to identify a target (ball) and position the target (ball) in each frame of image; then combining the camera parameters and the depth information to calculate the position of the target (ball) in the three-dimensional space; finally, tracking the position information of the target (ball) in a three-dimensional space at each moment, recording the complete movement track of the target, marking the ball striking drop point position of each time, calculating the ball discharging speed of the sportsman, analyzing and displaying the data, and visually analyzing and displaying the data of the movement details to help the sportsman to visually know the insufficient progress and the progress in training, and also assist a coach to formulate a more scientific, reasonable and personalized training scheme aiming at the situation of each sportsman;
the invention also obtains an increment learning result by continuously performing increment learning based on the target complete motion track, the batting drop point position, the ball outlet speed and the like, continuously corrects the network parameters of the ball return route network model based on the increment learning result, and is applied to the virtual training of athletes. The athlete can enter a training mode after wearing the VR helmet; through the virtual training with exclusive training object, can constantly challenge "oneself", overcome own weakness, bring better training experience and training effect for the sportsman.
Drawings
In order to more clearly illustrate the technical solutions of the present invention, the following brief description will be given to the drawings required for the invention, and it is obvious that the drawings in the following description are some embodiments of the present invention, and that other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic diagram of a system according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a system according to an embodiment of the present invention;
FIG. 3 is a flow chart of a system operation method of the present invention for one embodiment.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention. In addition, some technical features that are known in the art have not been described in order to avoid obscuring the present invention.
According to one embodiment of the present application, the present invention relates to a training aid system for ball games, comprising a hardware part and a software part. As shown in fig. 1, the hardware components include a data acquisition device 110, a data transmission device 120, a computing platform 130, a data storage device 140, a display 150, and a VR headset 160. The software part mainly comprises a data communication transmission module 131, an image preprocessing module 132, a target detection module 133, a three-dimensional space positioning module 134, a motion analysis module 135, a motion training module 136 and a data storage and three-dimensional display module 137.
The data acquisition device 110 is used for shooting a motion video of an athlete during training, and the video transmission device 120 is used for transmitting the motion video shot by the data acquisition device 110 to the motion platform 130 to process the motion video.
Specifically, the computing platform 130 first pre-processes the video image; then, carrying out target detection on each frame of image, identifying a target (ball) and positioning the position of the target (ball) in each frame of image; then combining the camera parameters and the depth information to calculate the position of the target (ball) in the three-dimensional space; and finally, tracking the position information of the target (ball) in the three-dimensional space at each moment, recording the complete movement track of the target, marking the position of the ball striking drop point at each time, and calculating the ball discharging speed of the sporter. In addition, the computing platform 130 may perform incremental learning based on the target complete motion track, the hitting drop point position, the ball output speed, etc., to obtain an incremental learning result, and continuously correct the network parameters of the return route network based on the incremental learning result, and apply the network parameters to the virtual training of the athlete. The athlete may enter the training mode after wearing VR headset 160.
Still further, the computing platform 130 stores the motion analysis data, the motion training data and the original video data obtained by the processing in the data storage device 140, and three-dimensionally displays the data on the display device 150 through three-dimensional modeling, so as to assist a coach to guide the athlete to train;
specifically, the motion analysis data comprise complete motion track information of the table tennis, batting drop point position information and ball outlet speed information; the athletic training data includes a modified return route network.
Fig. 2 is a schematic diagram of a system according to an embodiment of the present invention. In order to assist the training of table tennis players, the cameras are arranged above the table tennis table, and the field of view covers the whole movement area.
According to another embodiment of the present invention, a working method of the training aid system for ball games is shown in fig. 3:
step S01, shooting and transmitting a motion video;
specifically, after the auxiliary training system is started, the camera shoots the whole training process of the athlete, and the shot sports video is sent to the computing platform through the data transmission equipment.
It should be noted that, the camera of the data acquisition device for capturing the motion video is a binocular stereo camera or an RGB-D camera, the sampling frequency of the camera can be set according to the requirement, and considering the hardware cost, the data acquisition device 110 in this embodiment adopts a binocular stereo camera with relatively low price.
In addition, the computing platform may be a local computer, a terminal device, a remote computing center, or a cloud platform, which is not particularly limited herein.
The transmission mode of the data transmission equipment can be wireless or wired mode to realize data transmission;
step S02: preprocessing an image;
and after receiving the video image, the computing platform performs preprocessing operation on the video image through an image preprocessing module.
Specifically, the image preprocessing module comprises algorithms such as image difference, edge extraction, binarization processing, filtering denoising processing and the like.
Step S03: detecting a table tennis target;
the target detection module carries out target detection on each preprocessed frame of image, identifies the table tennis ball and positions the table tennis ball in each frame of image;
the target detection module comprises, but is not limited to, a target detection algorithm based on a traditional method and a target detection algorithm based on deep learning; the target detection algorithm based on the traditional method is mainly a circle detection algorithm, including but not limited to a Hough transform algorithm or various improved Hough transform algorithms; the target detection algorithm based on deep learning comprises, but is not limited to, R-CNN, SPP-net, fast R-CNN, R-FCN, YOLO, SSD and the like;
step S04: positioning the table tennis ball in a three-dimensional space;
after the position information of the table tennis ball in each frame of image is obtained, calculating the corresponding moment of each frame of image and the position of the table tennis ball in a three-dimensional space by combining camera parameters and depth information;
in this embodiment, the camera parameters are internal parameters and external parameters of the binocular stereo camera, and mainly include focal length, optical center position, distortion coefficient, base line distance, rotation matrix R and translation vector T, and further, the internal and external parameters of the camera can be obtained through camera calibration.
It should be noted that, the obtaining mode of the depth information is determined by the camera model selection, and may be obtained by calculating the depth information by a binocular stereo camera according to a triangulation method, or may be obtained by a depth map provided by an RGB-D camera. In this embodiment, the depth information of each point on the table tennis is calculated by a triangulation method.
Exemplary, it is shown in a specific case how after the position of the table tennis ball in each frame of image is acquired, the table tennis ball is positioned in combination with camera parameters and depth information, i.e. the coordinates (X W ,Y W ,Z W ) Is a process of (1);
suppose that when a certain frame of image corresponds toThe coordinates of the table tennis ball position in the world coordinate system are (X W ,Y W ,Z W ) The coordinates in the camera coordinate system are (X C ,Y C ,Z C ) Said coordinates (X C ,Y C ,Z C ) Coordinates of the projection points in the image pixel coordinate system are (u, v), and the conversion relation is as follows:
f in the above x Is the scale factor on the u-axis, also called normalized focal length on the u-axis, and similarly, F y Is a scale factor on the v-axis, (u) 0 ,v 0 ) Representing the coordinates of the camera's optical center in the image pixel coordinate system, R is a 3×3 orthogonal rotation matrix, and T is a 3×1 translation matrix.
Since the coordinates (u, v) have been obtained by object detection in said step s03, according to the camera imaging principle, in the case of camera internal and external parameters determination, from one point on one image, the position of the point in three-dimensional space determined with respect to the camera position can be back calculated, i.e. the coordinates (X W ,Y W ,Z W )。
Step S05: motion analysis;
the position information of the table tennis ball in the three-dimensional space at each moment is tracked, the complete movement track of the table tennis ball is recorded, the position of the ball striking drop point at each time is marked, and the ball discharging speed of the mobilizer after each swing is calculated.
In this embodiment, the ball-out speed, i.e. the speed of the table tennis, of the player after each swing is calculated, and the table tennis moving distance can be calculated by changing the coordinates of the center of the table tennis in the two frames of images, divided by the interval time of the two frames of images.
Step S06: virtual training;
specifically, the motion training module in the computing platform sets a dedicated virtual training opponent for each player, the ball return route of the virtual training opponent is calculated and generated by a ball return route network model deployed on the computing platform, the ball return route network model can calculate and output the ball striking point position, the ball striking angle and the force of the virtual training opponent when the ball returns, and the ball return route network model continuously carries out incremental learning according to the target complete motion track, the ball striking point position, the ball outlet speed and other motion data obtained by analyzing the real training video of the player to obtain an incremental learning result, and continuously corrects the network parameters of the ball return route network based on the incremental learning result, so that the method is applied to the virtual training of the player. When the athlete wants to train, the athlete can enter the training mode after wearing the VR helmet.
Wherein the incremental learning adopts a regularized incremental learning method, including but not limited to LwF algorithm and related improvement strategies.
Step S07: storing data and displaying three-dimensionally;
specifically, the computing platform stores the motion data (including complete table tennis motion track information, batting drop point position information, ball-out speed information and the like), motion training data (including corrected ball-return route network) and original video obtained through the processing in the data storage device, and can be displayed on the display device in real time and three-dimensionally through three-dimensional modeling, so as to assist a coach to guide a player to train.
The information of the complete motion trail of the ping ball can be used for analyzing the ball returning habit of athletes and performing tactical training; the ball striking drop point position information can be used for analyzing the accuracy and stability of the ball striking of the athlete; the ball-out speed information can be used for analyzing whether the force of the athlete is correct or not and measuring whether physical training needs to be increased or not;
it should be further noted that the data storage and three-dimensional display module may automatically read the type of the data storage device and set a storage scheme corresponding to the data storage device according to different device types, and may set up a specific training file for each person trained using the system to store personalized training data. In this embodiment, a storage scheme combining local solid state disk storage and cloud storage is adopted, that is, the motion original video and motion data within one month are preferentially stored in the local solid state disk, so that the motion original video and motion data exceeding one month can be conveniently and quickly read and called, and the motion original video and motion data exceeding one month can be automatically uploaded to a cloud server for storage, so that the space of the local solid state disk is released.
In addition, the ball games include, but are not limited to, table tennis, badminton, billiards, snooker, and the like.
The auxiliary training system for the ball games provided by the invention can help athletes to intuitively know the defects existing in training and the progress obtained by the athletes through visual data analysis and model display of the details of the sports, can assist coaches to formulate a more scientific, reasonable and more personalized training scheme aiming at the situation of each athlete, and can bring better training experience and training effect to the athletes.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (9)
1. An auxiliary training system for ball games, which is characterized by comprising a hardware part and a software part, wherein the hardware part comprises a data acquisition device (110), a data transmission device (120), a computing platform (130), a data storage device (140), a display device (150) and a VR helmet (160), and the software part comprises a data communication transmission module (131), an image preprocessing module (132), a target detection module (133), a three-dimensional space positioning module (134), a motion analysis module (135), a motion training module (136), a data storage and three-dimensional display module (137);
the data acquisition device (110) is used for shooting a motion video of an athlete during training, and the data transmission device (120) is used for transmitting the motion video shot by the data acquisition device (110) to the computing platform (130) so as to process the motion video; the processing procedure comprises the following steps: -the computing platform (130) first pre-processes the motion video image; then, carrying out target detection on each frame of image, identifying a ball, and positioning the ball in each frame of image; then combining the camera parameters and the depth information to calculate the position of the ball in the three-dimensional space; finally, tracking the position information of the target ball in the three-dimensional space at each moment, recording the complete movement track of the ball, marking the position of the ball striking drop point at each time, and calculating the ball discharging speed of the mobilizer; the computing platform (130) performs incremental learning based on the target complete motion track, the batting drop point position and the ball outlet speed to obtain an incremental learning result, continuously corrects network parameters of a ball return route network model based on the incremental learning result, and is applied to virtual training of a player, and the player can enter a training mode after wearing a VR helmet;
the motion analysis module (135) is used for obtaining complete motion trail information of the ball, analyzing the ball returning habit of the athlete and performing tactical training; the batting drop point position information is used for analyzing the accuracy and stability of batting of the athlete; the ball-out speed information is used for analyzing whether the force of the athlete is correct or not and measuring whether physical training needs to be increased or not; the sports training module (136) sets a dedicated virtual training opponent for each player, the ball return route of the virtual training opponent is calculated and generated by a ball return route network model deployed on the computing platform, and the ball return route network model performs incremental learning according to the target complete motion track, the batting drop point position and the ball outlet speed motion data obtained by analyzing the real training video of the player.
2. An auxiliary training system for ball games according to claim 1, wherein the image pre-processing comprises image differencing, edge extraction, binarization processing and filtering denoising processing algorithms.
3. The training aid system for ball games according to claim 1, wherein the target detection module comprises a conventional method-based target detection algorithm and a deep learning-based target detection algorithm; the target detection algorithm based on the traditional method is a circle detection algorithm, and the circle detection algorithm is one of a Hough transformation algorithm or an improved Hough transformation algorithm; the target detection algorithm based on deep learning is one of R-CNN, SPP-net, fast R-CNN and R-FCN, YOLO, SSD.
4. The training aid system for ball games according to claim 1, wherein the camera parameters comprise in-camera parameters and out-parameters, in particular focal length, optical center position, distortion coefficients, base line distance, rotation matrix R and translation vector T; the camera parameters are obtained through camera calibration.
5. An auxiliary training system for ball games according to claim 1, characterised in that the way of obtaining depth information is determined by camera selection, calculated from triangulation methods by a binocular stereo camera or by a depth map of an RGB-D camera.
6. An auxiliary training system for ball games according to claim 1, wherein the data storage and three-dimensional display module automatically reads the type of said data storage device and sets the corresponding storage scheme according to different device types and creates a specific training profile for each person trained using the system to store personalized training data.
7. A training aid system for ball games according to claim 1, wherein said data transmission means is wireless or wired.
8. A method of operating a training aid system for ball games according to any of claims 1-7, wherein:
step S01, shooting and transmitting a motion video; after the auxiliary training system is started, the camera shoots the whole training process of the athlete and sends the shot motion video to the computing platform through the data transmission equipment;
step S02: preprocessing an image; after receiving a video image, the computing platform carries out preprocessing operation on the video image through an image preprocessing module;
step S03: detecting a table tennis target; the target detection module carries out target detection on each preprocessed frame of image, identifies the table tennis ball and positions the table tennis ball in each frame of image;
step S04: positioning the table tennis ball in a three-dimensional space; after the position information of the table tennis ball in each frame of image is obtained, calculating the corresponding moment of each frame of image and the position of the table tennis ball in a three-dimensional space by combining camera parameters and depth information;
step S05: motion analysis; the method comprises the steps of recording the complete movement track of the table tennis by tracking the position information of the table tennis in a three-dimensional space at each moment, marking the position of a ball striking drop point at each time, and calculating the ball discharging speed of a sporter after each swing;
step S06: virtual training; the training system comprises a training platform, a training module, a virtual training opponent and a virtual training network model, wherein the training opponent is provided with a dedicated virtual training opponent for each athlete, a ball return route of the virtual training opponent is calculated and generated by the ball return route network model deployed on the computing platform, the ball return route network model calculates and outputs the ball hitting point position, the ball hitting angle and the force of the virtual training opponent when the virtual training opponent returns, the ball return route network model carries out incremental learning according to target complete motion track, ball hitting drop point position and ball outlet speed motion data obtained by analyzing real training videos of the athlete, an incremental learning result is obtained, network parameters of the ball return route network are corrected based on the incremental learning result, the virtual training of the athlete is applied to the virtual training of the athlete, and a training mode is entered after the athlete wears a VR helmet when the training is wanted;
step S07: storing data and displaying three-dimensionally; the computing platform stores the motion data, the motion training data and the original video obtained through processing in the data storage device, and can be displayed on the display device in real time and three-dimensionally through three-dimensional modeling, so as to assist a coach to guide an athlete to train.
9. A method of operating an auxiliary training system for ball games according to claim 8, wherein: the motion data comprises complete motion trail information of the table tennis, batting drop point position information and ball outlet speed information; the athletic training data includes a modified return route network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211247470.9A CN115624735B (en) | 2022-10-12 | 2022-10-12 | Auxiliary training system for ball games and working method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211247470.9A CN115624735B (en) | 2022-10-12 | 2022-10-12 | Auxiliary training system for ball games and working method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115624735A CN115624735A (en) | 2023-01-20 |
| CN115624735B true CN115624735B (en) | 2023-06-27 |
Family
ID=84903848
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211247470.9A Active CN115624735B (en) | 2022-10-12 | 2022-10-12 | Auxiliary training system for ball games and working method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115624735B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116665841B (en) * | 2023-07-28 | 2023-10-31 | 山东大学 | Directional shooting athlete reaction training device and real-time evaluation system |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102243687A (en) * | 2011-04-22 | 2011-11-16 | 安徽寰智信息科技股份有限公司 | Physical education teaching auxiliary system based on motion identification technology and implementation method of physical education teaching auxiliary system |
| WO2016116780A1 (en) * | 2015-01-23 | 2016-07-28 | Playsight Interactive Ltd. | Ball game training |
| US10300362B2 (en) * | 2015-04-23 | 2019-05-28 | Win Reality, Llc | Virtual reality sports training systems and methods |
| CN108465224A (en) * | 2018-04-07 | 2018-08-31 | 华北理工大学 | Table tennis track analysis system |
| CN112702481B (en) * | 2020-11-30 | 2024-04-16 | 杭州电子科技大学 | Table tennis track tracking device and method based on deep learning |
| CN215609219U (en) * | 2021-03-25 | 2022-01-25 | 南方科技大学 | Table tennis training device and training system |
| CN113688741A (en) * | 2021-08-26 | 2021-11-23 | 成都大学 | Motion training evaluation system and method based on cooperation of event camera and visual camera |
-
2022
- 2022-10-12 CN CN202211247470.9A patent/CN115624735B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN115624735A (en) | 2023-01-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20220292882A1 (en) | Tracking of handheld sporting implements using computer vision | |
| AU2017331639B2 (en) | A system and method to analyze and improve sports performance using monitoring devices | |
| US20180357472A1 (en) | Systems and methods for creating target motion, capturing motion, analyzing motion, and improving motion | |
| US9744421B2 (en) | Method of analysing a video of sports motion | |
| CN103916586B (en) | image analysis apparatus and image analysis method | |
| US11798318B2 (en) | Detection of kinetic events and mechanical variables from uncalibrated video | |
| US11850498B2 (en) | Kinematic analysis of user form | |
| CN115624735B (en) | Auxiliary training system for ball games and working method | |
| CN108905182B (en) | Intelligent glasses and billiard shooting positioning and aiming method | |
| Malawski | Depth versus inertial sensors in real-time sports analysis: A case study on fencing | |
| KR20100033205A (en) | System suppporting golf practice and method therefor | |
| CN113505662A (en) | Fitness guidance method, device and storage medium | |
| CN108970091B (en) | Badminton action analysis method and system | |
| Yang et al. | Research on face recognition sports intelligence training platform based on artificial intelligence | |
| US10025986B1 (en) | Method and apparatus for automatically detecting and replaying notable moments of a performance | |
| CN111754549B (en) | Badminton player track extraction method based on deep learning | |
| CN110910489B (en) | Monocular vision-based intelligent court sports information acquisition system and method | |
| JP2006181014A (en) | Image analysis device and movement correction system | |
| CN110929595A (en) | System and method for training or entertainment with or without ball based on artificial intelligence | |
| US20210322852A1 (en) | Determining trajectory of a ball from two-dimensional media-content using computer vision | |
| JP7319540B2 (en) | Shot management system, shot management method and program | |
| Malawski | Real-Time First Person Perspective Tracking and Feedback System for Weapon Practice Support in Fencing. | |
| KR102534463B1 (en) | Golf lesson system using a joint of human body and a golf club based on AI | |
| US20220343649A1 (en) | Machine learning for basketball rule violations and other actions | |
| JP7248353B1 (en) | Hitting analysis system and hitting analysis method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |