Intelligent technical and tactical style evaluation method and system for boxing sports

Abstract

The invention belongs to the technical field of intelligent evaluation, and discloses an intelligent evaluation method and system for technique and tactical styles for boxing sports, and the method comprises the steps: collecting motion video data of red and blue boxing hands, obtaining time sequence data of the red and blue boxing hands through an openpose deep learning algorithm and a pinhole imaging model principle in combination with openpose confidence, and carrying out the recognition of the time sequence data of the red and blue boxing hands; the method comprises the following steps: performing anomaly detection, elimination, filling and smoothing processing on time sequence data, then endowing each timestamp of the preprocessed time sequence data with an action label, and realizing technical and tactical analysis and evaluation on a boxing hand through a technical and tactical style and characteristic analysis and evaluation module. The problems that in the prior art, a real-time feedback and quantification technique and tactics means cannot be provided, defense, counterattack and other techniques and tactics are not comprehensively considered, and comprehensive assessment of the techniques and tactics is not facilitated are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent evaluation, and in particular relates to a method and system for intelligent evaluation of technical and tactical styles of boxing. Background Art

[0002] Boxers need to develop technical and tactical abilities through intensive training and a great deal of practical experience. The traditional way of evaluating boxers relies on human guidance and on-site sparring, which is inefficient and subjective, and cannot provide real-time feedback and personalized training plans.

[0003] The existing technology uses video action recognition but cannot provide real-time feedback and quantify technical and tactical means. It improves the offensive technical and tactical capabilities of boxers by calculating indicators such as the angle and speed of the boxer's punch, but does not fully consider defensive and counterattack techniques and tactics, which is not conducive to a comprehensive evaluation of techniques and tactics. The existing technology cannot provide real-time feedback to boxers, which makes it difficult for athletes to adjust their techniques and tactics in a timely manner. At the same time, the traditional evaluation method relies on human guidance and on-site sparring, which is easily affected by personal bias, resulting in subjective evaluation results. Summary of the invention

[0004] The present invention aims at the problems existing in the prior art and provides a method and system for intelligent evaluation of technical and tactical styles for boxing.

[0005] The objective of the present invention is achieved through the following technical solutions:

[0006] In a first aspect, the present invention provides a method for intelligently evaluating the technical and tactical styles of boxing, the method comprising the following steps:

[0007] Step S1, deploy four cameras around the boxing ring to collect action video data of the red and blue boxers; calibrate the cameras to obtain the intrinsic parameter matrix and extrinsic parameter matrix of each camera;

[0008] Step S2: First, the OpenPose deep learning algorithm is used to detect the human joints in the acquired action video data, that is, the two-dimensional plane coordinates of the joints are obtained. Then, according to the principle of the pinhole imaging model and combined with the OpenPose confidence, the three-dimensional coordinates of the joints are obtained, that is, the time series data of the red and blue sides are obtained respectively.

[0009] Step S3, preprocessing the time series data: analyzing the time series data, identifying and marking the timestamps where the outliers are located in the time series data, then setting the data values ​​of the timestamps where the outliers are located to blank and filling the blank values, and then performing smoothing filtering on the filled time series data to obtain the preprocessed time series data;

[0010] Step S4, using the time series classification task to realize action recognition, and then assigning an action label to each timestamp of the preprocessed time series data;

[0011] Step S5, annotating the time series data with action labels, including sprinting steps, circling steps, sliding steps and passive marking, punching hand marking and punching position marking, and marking effective punches and ineffective punches in attack and counterattack tactics;

[0012] Step S6, counting the boxer's use of skills and tactics: according to the data obtained in step S5, respectively counting the number and proportion of offensive, defensive, counterattack, and feint skills and tactics, statistics on the dominant hand of the skills and tactics, statistics on the hitting parts, and calculation of the hit rate and effective punch ratio and important parameters for evaluating the skills and tactics;

[0013] Step S7, labeling the style and characteristics of the time series preprocessed in step S3, integrating the labels with the important technical and tactical parameters obtained in step S6 to build a classification model based on machine learning, and inputting the important technical and tactical parameters into the trained classification model based on machine learning to evaluate the technical and tactical style and characteristics of the boxer in a round.

[0014] Preferably, the specific steps of step S2 are as follows:

[0015] Step S2-1, using the OpenPose deep learning algorithm to detect the human joints in the acquired action video data, and obtain the joint points P i The two-dimensional plane coordinates (u i , v i ), i=1, 2, 3 or 4;

[0016] Step S2-2, using the pinhole imaging model principle to obtain the three-dimensional coordinates P (X w ,Y w ,Z w ):

[0017]

[0018] Among them, f x and f y are the horizontal focal length and vertical focal length of the camera respectively; u0 and v0 are the pixel coordinates of the intersection of the camera optical axis and the imaging plane, M i is the external parameter matrix from the world coordinate system to the camera i coordinate system, x w ,y w 、z w is the three-dimensional coordinate of the joint point P under ideal conditions, z c is the proportionality coefficient;

[0019] Step S2-3, in order to avoid noise, it is necessary to estimate the spatial position of the real joint point R, consider the distance between the rays emitted by the camera, set a spherical area, and the radius of the sphere is 10 times the shortest distance between the rays. The three-dimensional coordinate formula of the real joint point R is as follows:

[0020]

[0021] Among them, l i is the ray emitted by the i-th camera, |P,l i | represents the ideal state of the joint point P to the ray l i The shortest distance, ρ i is the Openpose confidence parameter, n=4;

[0022] Step S2-4, the action video data covers multiple joints of the human body, and time series data is obtained according to the real three-dimensional coordinates of the multiple joints.

[0023] Preferably, the specific steps of step S3 are as follows:

[0024] Step S3-1, obtaining the boxer's motion data in an unobstructed environment, constructing a training data set that can reflect the real human body movement law, and using the athlete training data collected in the boxing ring environment as the test data set;

[0025] Step S3-2, using a normalization method to map the value range of the time series data under the training data set and the test data set to a real number space from 0 to 1;

[0026] Step S3-3, inputting the normalized training data set into the Transformer model, so that the Transformer model learns the normal pattern of boxing movement;

[0027] Step S3-4, input the normalized test data set into the Transformer model obtained in step S3-3, construct anomaly score using reconstruction error MSE, set a threshold, and select the top 5% of data with anomaly scores greater than the threshold as abnormal data segments, in timestamp units, otherwise, they are normal data segments;

[0028] Step S3-5, empty the abnormal data segment and fill it with normal values ​​through interpolation or Transformer prediction;

[0029] Step S3-6, using adaptive filtering technology to perform smoothing filtering on the filled time series data.

[0030] Preferably, the abnormal data segments in step S3-5 are set to empty and filled with normal values, specifically: when there are few abnormal data segments, an interpolation method is used, and adjacent normal data points are used to estimate the values ​​of the abnormal data segments through linear interpolation and spline interpolation; when there are many abnormal data segments, a trained Transformer model or regression model is used to predict the abnormal data segments and fill them with predicted values.

[0031] Preferably, the specific steps of step S4 are as follows:

[0032] Step S4-1, using the time series classification task to realize action recognition, classifying actions into offensive actions, defensive actions, counter-attack actions and feints, annotating the pre-processed time series data, and dividing it into a training set and a test set;

[0033] Step S4-2, using the training set to train a multivariate time series classification model based on the LSTM architecture;

[0034] Step S4-3, input the preprocessed time series data into the trained multivariate time series classification model based on the LSTM architecture, and assign a corresponding action label to each timestamp of the preprocessed time series data.

[0035] Preferably, the specific steps of step S5 are as follows:

[0036] Step S5-1, filter out the timestamps corresponding to the movement categories in the action tags, divide the steps into sprinting, circling, sliding and passive steps according to the displacement, speed, rotation angle and acceleration of the joint points, and perform fine-grained annotation of the steps for the time series data of the movement categories in the action tags;

[0037] Step S5-2, filtering out the timestamps corresponding to the attack and counterattack in the action tags, marking the puncher according to the coordinate values, movement speed, and acceleration of both hands, and marking the hitting parts according to the attack and counterattack time series data in the action tags according to the distance from the opponent's joint points;

[0038] Step S5-3, taking each interaction between the two parties as a unit, combining the puncher and the hitting part in each interaction, identifying whether there is a hit or not, and at the same time, marking the hitting effect of the time series with action labels as attack and defense, that is, marking effective punches and invalid punches.

[0039] Preferably, the important parameters for evaluating techniques and tactics in step S6 include: pressure coefficient, distance coefficient, balance coefficient and movement coefficient.

[0040] Preferably, the specific steps of step S7 are as follows:

[0041] Step S7-1, labeling the style and characteristics of the time series data obtained after preprocessing in step S3, and dividing it into a training set and a test set, the style labels include active offensive type, defensive counterattack type, offensive round-trip type, and wandering type, and the characteristic labels include fake moves, direct attacks, tight pressure, and attack first and then defense;

[0042] Step S7-2, integrating the number statistics and proportion calculation of the offensive, defensive and counterattack techniques and tactics described in step S6, the statistics of the dominant hand of the techniques and tactics, the statistics of the hitting parts, the calculation of the hit rate and the effective punch ratio, and the important parameters of the techniques and tactics as well as the style and characteristic labels, to build and train a classification model based on machine learning;

[0043] Step S7-3, inputting important technical and tactical parameters into the trained classification model based on machine learning to evaluate the technical and tactical style and characteristics of the boxer in a round.

[0044] In a second aspect, the present invention provides an intelligent evaluation system for boxing techniques and tactics, which is used to implement the above method, and includes: an acquisition module, a motion capture module, a time series preprocessing module, a motion recognition module, a motion analysis module, a technique and tactics statistics module, and a technique and tactics style and characteristics analysis and evaluation module;

[0045] The acquisition module is connected to the motion capture model and acquires the motion video data of the red and blue sides around the boxing ring;

[0046] The motion capture module obtains the action video data of the red and blue parties through the input channel, and obtains the time series data of the red and blue parties through the OpenPose deep learning algorithm and the pinhole imaging model principle, combined with the OpenPose confidence parameter;

[0047] The time series preprocessing module includes a time series anomaly detection submodule, an anomaly removal and filling submodule, and a smoothing filter submodule. The time series anomaly detection submodule analyzes the time series data, identifies and marks the timestamps where the outliers are located in the time series data, the anomaly removal and filling submodule sets the data value of the timestamp where the outliers are located to empty, and fills the empty value by interpolation or Transformer prediction, and the smoothing filter submodule performs smoothing filtering on the filled time series data to obtain the smoothed filtered time series data as the output of the time series preprocessing module;

[0048] The action recognition module realizes action recognition by using the time series classification task, and then assigns an action label to each timestamp of the preprocessed time series data;

[0049] The action analysis labels the time series data with action tags, including sprint steps, circling steps, sliding steps and passive labels, punching hand labels and punching parts labels, and labels of effective punches and ineffective punches in attack and counterattack tactics;

[0050] The technical and tactical statistics module counts the boxer's use of technical and tactical skills, including the number and proportion of offensive, defensive and counterattack technical and tactical skills; the boxer's dominant hand usage pattern of technical and tactical punching dominant hand statistics; the statistics of striking parts; the calculation of hit rate and effective punch ratio; and the important parameters for evaluating technical and tactical skills;

[0051] The technical and tactical style and characteristic analysis and evaluation module constructs a classification model based on machine learning according to the output features of the technical and tactical statistics module combined with style and characteristic labels, and uses the classification model based on machine learning to implement technical and tactical style evaluation of boxing.

[0052] The present invention designs a method and system for intelligent evaluation of technical and tactical styles for boxing. In the present invention, four cameras are used to collect video data of training or competition of red and blue boxers, and the time series data of red and blue are obtained by combining the openpose and pinhole imaging model principles in the motion capture module according to four perspective video streams, and the openpose confidence. The time series is detected, eliminated, and filled by the time series preprocessing module, and the time series is smoothed. The LSTM-based action recognition module uses the time series classification task to obtain rough classification labels such as actions. In addition, the present invention designs an action analysis module to fine-grainedly identify the single-fist classification and hitting parts of the red and blue footwork, effective punches, and combination punches, and constructs a standard action knowledge base. Through the technical and tactical statistics module, the offensive, defensive, and counterattack techniques and tactics in one round of the game are statistically analyzed, and the hit rate, effective punch ratio, pressure system, balance coefficient, distance coefficient, and movement coefficient are calculated. The boxing technical and tactical style and characteristic analysis and evaluation module merges the features, thereby realizing the technical and tactical analysis and evaluation of the boxer.

[0053] The present invention has the following beneficial effects: (1) The present invention provides a method and system for intelligently evaluating the technical and tactical style of boxing, which solves the problem that the prior art cannot provide real-time feedback and means for quantifying technical and tactical styles, does not comprehensively consider defensive and counterattack-like technical and tactical styles, and is not conducive to comprehensive evaluation of technical and tactical styles.

[0054] (2) The present invention collects video data through four cameras and uses openpose and the pinhole imaging model principle to capture motion, so as to detect the boxer's movements in all directions, which is convenient for the comprehensive analysis of subsequent technical and tactical styles and characteristics.

[0055] (3) Based on the time series data output by the motion capture module, the present invention uses the LSTM-based time series classification task to complete the motion recognition task, avoiding the use of complex computer vision models. Through the inherent memory mechanism of LSTM, the key motion features are captured, greatly improving the efficiency of motion recognition.

[0056] (4) The present invention uses fine-grained recognition of the footwork of both red and blue players, effective punches, and the sub-classification of single punches in combination punches, and the striking parts, and constructs a standard action knowledge base to count the offensive, defensive, and counterattack techniques and tactics in a round of the game, calculate the hit rate, effective punch ratio, pressure system, balance coefficient, distance coefficient, and movement coefficient, and integrate the features through a boxing technique and tactical style and characteristic analysis and evaluation module, so as to comprehensively analyze the boxer's techniques and tactics. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 A flowchart of an intelligent evaluation method for technical and tactical styles in boxing;

[0058] Figure 2 This is a framework diagram of an intelligent evaluation system for technical and tactical styles in boxing; DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0060] A method for intelligently evaluating the technical and tactical style of boxing, the method comprising the following steps:

[0061] Step S1, obtain the action video data of the boxer; set up four cameras around the boxing ring to achieve multi-angle coverage of shooting. Use Zhang Zhengyou binocular camera calibration method to determine the relative position relationship between two adjacent cameras. By calibrating the camera, the internal and external parameter matrices of each camera are obtained. The internal parameter matrix contains information such as the focal length of the camera, and the external parameter matrix contains two different coordinate system conversion information, that is, the transformation matrix. The four cameras are connected to the input channel of the motion capture module through the synchronization interface. The motion capture module supports one-click synchronous shooting. To ensure that the generated motion capture data is stable and there is no confusion in the timestamp of the data obtained by the motion capture module, a master camera is selected from the four cameras to be connected to the motion capture module. Multiple slaves and the host are synchronized through a dedicated synchronization line. The trigger signal is sent to the main camera by pressing a button, and the signal is transmitted to the slave through the synchronization line, so that the start time and end time of the shooting of the four cameras are completely synchronized, and the synchronization time error must be less than one frame.

[0062] Furthermore, by calibrating the camera, we can obtain the transformation matrix from the world coordinate system to each camera and the transformation matrix between any two cameras. The transformation matrix can reflect the conversion method from one perspective coordinate system to another perspective coordinate system. Through the transitivity of the transformation matrix, we can obtain the transformation matrix between any camera and the other three cameras, and then we can get the transformation matrix between multiple cameras.

[0063] Step S2, the motion capture module captures four real-time video data streams through the input channel, and obtains the time series of the red and blue sides through the OpenPose deep learning algorithm and the pinhole imaging model principle, combined with the OpenPose confidence. The specific implementation steps include:

[0064] By using the OpenPose deep learning algorithm, the key joints of the human body in four real-time video data streams are detected. The algorithm identifies and locates the key points in the video that reflect the athlete's body posture and movement trajectory, such as the head, shoulders, elbows, wrists, hips, knees and ankles, and obtains the joint point P i The two-dimensional plane coordinates (u i , v i ). The algorithm supports real-time video stream input and accurately locates the key joints of each frame in the video.

[0065] The imaging of a three-dimensional space point on a camera can be represented by a pinhole imaging model, which is expressed as follows:

[0066]

[0067] where f x and f yare the horizontal focal length and vertical focal length of the camera respectively; u0, v0 are the pixel coordinates of the intersection of the camera optical axis and the imaging plane, f x , f y , u0, v0 are the internal parameters of the camera, M i is the external parameter matrix from the world coordinate system to the camera i coordinate system, x w ,y w , z w is the three-dimensional coordinate of the joint point P in the ideal state in space, z c is the proportional coefficient, which can be 1 by default.

[0068] For the i-th camera, the three-dimensional coordinate point P(X w ,Y w ,Z w ) and the two-dimensional plane coordinates of the point imaged by a certain camera.

[0069]

[0070] in

[0071] Ideally, the above mapping relationship is applicable to the four cameras, and the rays emitted by the four cameras can intersect at one point, that is, P(x w ,y w ,z w ). However, due to the inevitable noise in the data, the ray li deviates, resulting in multiple rays not intersecting at one point, so it is necessary to estimate the spatial position of the joint point P in space. The present invention considers the distance between rays, sets a spherical area, and the radius of the sphere is 10 times the shortest distance between rays. The moving point P' moves in the spherical area. Combined with the OpenPose confidence parameter, the spatial point with the minimum distance to multiple rays in space is taken as the intersection of multiple rays.

[0072] If the rays are l1, l2, l3, l4, where |P,l i | represents the ideal state from the joint point P to the ray l i The shortest distance, the three-dimensional coordinates of the real joint point R are expressed as follows:

[0073]

[0074] Among them, the above formula indicates that the final joint point P in space is the point closest to the four rays. Since the rays have confidence parameters, when solving the position of the real joint point R, the spatial point should be close to the ray with higher confidence. In order to reduce the measurement error, the Openpose confidence parameter ρ is introduced i ; Then the three-dimensional coordinate formula of the real joint point R is as follows:

[0075]

[0076] Furthermore, the weighted distance and expression of the joint point to multiple rays in space and the three-dimensional coordinate expression of the spatial point are:

[0077]

[0078] where d i For ray l i The direction vector, c i The coordinates of the origin of the camera coordinate system in the world coordinate system.

[0079] With the input of real-time video data stream, the OpenPose deep learning algorithm outputs a two-dimensional coordinate sequence. Through the above process, a three-dimensional multivariate time series covering the motion information of multiple joints of the human body can be further obtained, that is, time series data. The variable of the multivariate time series is a certain dimension of a certain joint point.

[0080] Step S3, the time series preprocessing module detects abnormal timestamps in the time series data through the time series anomaly detection submodule, and the anomaly removal and filling submodule removes them and fills them with null values. The smoothing filter submodule uses Kalman filtering to adaptively adjust the filter parameters under the MSE error constraint to smooth the filled time series data.

[0081] The present invention recognizes boxing actions in time series, rather than action recognition under video, because the present invention takes into account that the deep learning model under video data cannot provide real-time functions for the scenarios faced by the present invention, and the model is generally large. The present invention recognizes boxing actions through time series classification tasks, not only identifying the type of boxing, but also identifying the defensive actions of the opponent under the punch, providing a basis for the subsequent analysis of boxing skills, tactics, styles and characteristics.

[0082] Due to the occlusion of the boxing ring during the video shooting process, the motion capture module cannot capture the movement trajectory of the athlete within a specific time period, resulting in abnormalities in some timestamps in the time series data, which affects the accuracy of data processing and motion analysis. For this reason, the present invention designs a time series preprocessing module before motion recognition, including a time series anomaly detection submodule, an anomaly removal and filling submodule, and a smoothing filter submodule. The time series anomaly detection submodule analyzes the time series data, identifies and marks the timestamps where the abnormal values ​​are located; the anomaly removal submodule sets the data value of the abnormal timestamp to zero, and fills the empty value through interpolation or Transformer prediction; the smoothing filter submodule smoothes the filled time series data through Kalman filtering, adaptively adjusts the filter parameters under the MSE error constraint, and outputs the time series data after smoothing filtering. The specific implementation steps include:

[0083] It is known that the training data set should not have any anomalies. To this end, we first collect boxing athlete training data in an unobstructed environment to construct a training data set that can reflect the real laws of human movement and ensure that the training data has no anomalies. Subsequently, we collect athlete training data in an actual boxing ring environment as the test data set to simulate data anomalies in real scenarios.

[0084] The normalization method is used to map the value range of the time series under the training data and the test data to the real number space from 0 to 1, thereby eliminating the impact of different feature dimensions. Specifically, the min-max method is used to normalize the time series data of the training set and the test set. Considering that the time series is multivariate, this method needs to be used for each time series sample and each variable. Therefore, x is the time series data under a certain variable, and the dimension is the length of the time series. The specific implementation is shown in the following formula.

[0085]

[0086] The time series anomaly detection submodule has a built-in Transformer reconstruction model. For normal actions, the model can accurately reconstruct, but for data segments with anomalies, the reconstruction results will deviate greatly from the original input. Based on this principle, the location of the abnormal timestamp is identified. The Transformer deep learning model architecture is used to learn the normal patterns of the data in the training set, including but not limited to: 1. The normal movement rules of each joint point when punching and retracting; 2. The reasonable range of speed and acceleration in human movement; 3. The reasonable range of joint angle changes.

[0087] Furthermore, the training data set is input into the Transformer model. The encoder part of the Transformer model uses the self-attention mechanism to learn the relationship between different time steps in the time series, and encodes the entire sequence into a feature vector containing global information. The decoder reconstructs the original time series data from the feature vector, and by minimizing the reconstruction error, the Transformer model learns the normal pattern of boxing movement. The reconstruction error MSE in the training phase is shown as follows, where the dimension is the time series dimension, where x∈R L×M is the original data, To reconstruct the data, L is the length of the time series and m is the number of joint points.

[0088]

[0089] The reconstruction error is used to construct anomaly scores to measure the anomalies of the data. A larger anomaly score means that it is difficult for the model to accurately reconstruct the data segment, which means that there may be anomalies caused by occlusion during this time period. By setting a suitable threshold, the abnormal fragments in the boxing ring collected data are identified. Specifically, the model training is implemented by minimizing the reconstruction error, but during the test, due to the occurrence of anomalies, the abnormal position cannot be reconstructed normally, resulting in a larger reconstruction error at the abnormal position. Based on this principle, a large reconstruction error means a large anomaly score, which can measure the anomaly of the data and then output the abnormal position. For threshold setting, if the threshold is set larger, normal data may be marked as abnormal, which is detrimental to data mining; if the threshold is set smaller, the data at the abnormal position will be marked as normal. The embodiment of the present invention does not tend to damage the normal movement pattern of the data. In order to filter out the abnormal position, the anomaly scores are sorted, and the timestamp position corresponding to the top 5%, that is, the abnormal data segment, is the output of the time series anomaly module.

[0090] The abnormal data segments are emptied and filled with normal values ​​using a variety of methods, including but not limited to: 1. When there is less missing data, the interpolation method is used to use adjacent normal data points to estimate the values ​​of the abnormal data points through linear interpolation and spline interpolation; 2. When there is more missing data, the trained Transformer model is used to predict the abnormal data segments and fill them with predicted values.

[0091] Due to the high intensity and rapid change characteristics of boxing, the data stream contains high dynamic changes and rapid changes, and the spatial information reflected in the time series cannot be destroyed. For this reason, the present invention uses adaptive filtering technology to smooth the filled time series data, and uses the most suitable filter for each time series sample to ensure that the coordinate value will not change while the data is smooth. This process uses MSE error measurement and adaptively adjusts the filtering parameters.

[0092] Step S4: The action recognition module uses the time series classification task to assign an action label to each timestamp of the time series data. The specific implementation steps include:

[0093] The present invention realizes action recognition through time series classification tasks. This scenario involves: offensive actions, defensive actions, counterattack actions and feints, including but not limited to the following specific actions:

[0094] Offensive moves: straight punch, swing punch, hook punch, two-punch combination, three-punch combination, multi-punch combination;

[0095] Defensive actions: holding the head, blocking, sticking to each other, dodging, slapping, moving, and swinging arms;

[0096] Counterattack: counter punch, counter hook, counter hook, counter combination punch;

[0097] Feint: A move by a boxer to mislead his opponent into reacting.

[0098] Professional sports science practitioners and computer-related practitioners annotated the time series data output by the time preprocessing module and split it into training and test sets in a ratio of 8:2.

[0099] The action recognition module uses a multivariate time series classification model based on the LSTM architecture. To facilitate model training, data enhancement operations are used to cut the time series using overlapping time windows and feed them into the model for training. There are multiple LSTM layers in the multivariate time series classification model based on the LSTM architecture to capture the time and variable dependencies in the preprocessed time series data. The fully connected layer maps the output of the LSTM layer to the probability distribution of the target category. The output layer uses the softmax activation function to output the probability of each category.

[0100] The cross entropy loss function is used to measure the difference between the model prediction results and the true labels. By calculating the difference between the predicted probability distribution and the true label distribution, the prediction error of the model is quantified, thereby guiding the model to continuously optimize parameters during the training process.

[0101] The preprocessed time series data is input into the trained multivariate time series classification model based on LSTM architecture, and a corresponding action label is assigned to each timestamp of the preprocessed time series data.

[0102] Step S5, the action analysis module further analyzes the time series with action tags: analyzes the footwork of the movement category; identifies each single punch in the combination punch; identifies the hitting part of the punch; and identifies effective and invalid punches.

[0103] The movement category analyzes the footwork and classifies the movement labels of boxers in fine granularity, and divides the movement footwork into sprinting, circling, sliding, and passive. When a boxer moves at high speed and large angle to quickly shorten the distance with the opponent and launch an attack, it is classified as a sprinting step; when a boxer moves around the opponent at a medium speed and angle to control the venue and find offensive opportunities, it is classified as a circling step; when a boxer moves at a smooth speed and small angle, which is more used for defense, it is classified as a sliding step; when a boxer is forced to move under the attack of the opponent, it is classified as a passive step.

[0104] The combination punch recognizes each single punch and classifies it into combination punches and counter-combination punches in a fine-grained manner to ensure that the combination punch label remains unchanged. It has a secondary label to distinguish between straight punches, swinging punches, and hooks.

[0105] Single-punch puncher identification, judging the puncher based on the coordinate values, speed, and acceleration fluctuations of both hands during attack and counterattack techniques and tactics.

[0106] Identify the striking part of a punch, and determine whether it hits the opponent based on the distance from the punch to the opponent's joint point during offensive and counterattack techniques and tactics, and output the striking part.

[0107] The recognition of effective and invalid punches, in a competition scenario, takes each interaction between the two parties as a unit, combines the punching parameters and the hitting parts of the attacker during each interaction, and identifies whether the attacker's punch is effective or invalid.

[0108] The specific implementation steps are as follows:

[0109] The time series data with action labels contains information on 21 joints, including important joints for measuring boxing action information, such as hand joints, wrist joints, elbow joints, knee joints, ankle joints and foot joints. The displacement, speed, rotation angle and acceleration of these important joints are calculated and summarized into a data set. Professional sports science practitioners and computer-related practitioners perform fine-grained annotation for the time series with action labels, including sprint steps, circle steps, sliding steps and passive annotations in footwork; single punch annotations in combination punches and counterattack combination punches; and annotations of effective punches and invalid punches in offensive and counterattack techniques and tactics.

[0110] In particular, taking the red team’s perspective as an example, the timestamps corresponding to the red team’s movement categories are filtered out, and the blue team’s data at the same timestamps are retrieved. A new multivariate time series dataset is constructed, which contains the xyz coordinate values ​​of the knee, ankle, and foot joints, as well as the speed, rotation angle, and acceleration of these joints, and further derives the distance information between the knee, ankle, and foot joints of the red and blue teams. This new multivariate time series dataset with derived features is used for fine-grained classification of the steps of the movement categories.

[0111] Furthermore, the dataset is split into training and test sets in a ratio of 8:2. Based on these data, a LSTM-based fine-grained step classification model is trained to provide corresponding fine-grained step labels for each timestamp of the movement category, including sprint, circle, slide, and passive recognition.

[0112] In particular, the timestamps of combination punches and counter-combination punches are screened out, and the xyz coordinate values ​​of the hand joints, wrist joints, and elbow joints, as well as the speed, rotation angle, and acceleration of these joints are combined to construct a new multivariate time series dataset for the classification of single punches in combination punches.

[0113] Furthermore, the dataset is split into a training set and a test set in a ratio of 8:2. Based on these data, a LSTM-based fine-grained classification model for combination punches is trained to provide a corresponding single punch label for each timestamp of the combination punches.

[0114] Filter out the timestamps of offensive and counterattack tactics, and retrieve the opponent's data at the same timestamp. Construct a new multivariate time series data set, including the xyz coordinate values, speed, rotation angle, and acceleration of the attacker's hand joints, wrist joints, and elbow joints, and derive the distance between the attacker's hands and the defender's head joints, hand joints, wrist joints, elbow joints, and hip joints, as well as the displacement, speed, and acceleration of the head joints, hand joints, elbow joints, shoulder joints, knee joints, ankle joints, and foot joints. This new multivariate time series data set is used to identify the puncher of a single punch, the punching part of the punch, and the identification of effective and invalid punches.

[0115] Furthermore, through the coordinate values, speed, rotation angle, and acceleration of a pair of hand joints, wrist joints, and elbow joints, the corresponding hand with a larger fluctuation is determined to be the punching hand, and a punching hand label is given to each time series sample.

[0116] Furthermore, thresholds are set based on the distances between the attacking puncher and the defending puncher's head joint, hand joint, elbow joint, and hip joint to identify the hitting part. If multiple distances reach the threshold, the one with the smallest distance is selected as the hitting part. The hitting parts include: hitting the head, hitting the hand, hitting the elbow, and hitting the abdomen, and the hitting part label and no hitting label are marked for the corresponding timestamp.

[0117] Furthermore, the average speed and acceleration of the hand joints, wrist joints and elbow joints of the offensive puncher for 10 consecutive time stamps are calculated, and the average speed and acceleration of the 10 consecutive time stamps are guaranteed to be the largest, as well as the average displacement, speed and acceleration of the head joints, hand joints, elbow joints, shoulder joints, knee joints, ankle joints and foot joints of the defensive side at the corresponding time stamps. Combined with the offensive punch label, hit label, and valid / invalid punch label, a binary classification model based on machine learning is trained. The alternative models include K nearest neighbor, decision tree, random forest, XGBoost and other algorithms, and the model with the best classification accuracy is selected as the valid punch and invalid punch recognition model.

[0118] Step S6, technical and tactical statistics module, statistics on the use of boxing athletes' technical and tactical skills: statistics on the number of offensive, defensive, and counterattack technical and tactical skills and percentage calculation; statistics on the boxer's dominant hand usage pattern of technical and tactical punching; statistics on hitting parts; calculation of hit rate and effective punch ratio; important parameters for evaluating technical and tactical skills include calculation of pressure system, distance coefficient, balance coefficient, and movement coefficient. The specific implementation steps are as follows:

[0119] The number and percentage of offensive, defensive, and counter-attack tactics are counted. Taking the game round as the unit, the number of offensive, defensive, counter-attack, and feint tactics in a round is counted.

[0120] Furthermore, we calculated the proportion of the number of occurrences of straight punch, swing punch, hook punch, and combination punch labels in offensive techniques and tactics; the proportion of the number of occurrences of various defensive action labels such as head hugging and arm swinging in defensive techniques and tactics; the proportion of the number of occurrences of refined footwork action labels in movements; the proportion of the number of occurrences of counterattack single punch and counterattack combination punch labels in counterattack techniques and tactics; the proportion of offensive techniques and tactics timestamps in the entire game; the proportion of defensive techniques and tactics timestamps in the entire game; the proportion of counterattack techniques and tactics timestamps in the entire game; and the proportion of fake action timestamps in the entire game.

[0121] Statistics of the dominant hand of technical and tactical punches, and analysis of the boxer's dominant hand usage pattern. Statistics of offensive and counterattack technical and tactical puncher tags are collected in rounds, and the proportion of left-handed tags to all puncher tags is calculated, and the proportion of right-handed tags to all puncher tags is calculated.

[0122] Statistics of hitting parts. Taking the game round as the unit, calculate the number of labels of hitting parts in one round of the game scene.

[0123] Hit rate and effective punch ratio calculation. Taking the round as the unit, the hit rate is the ratio of the number of offensive and counter-attack tactics that hit the target to the total number of offensive and counter-attack tactics; the effective punch ratio is the ratio of effective punches to the number of offensive and counter-attack tactics that hit the target.

[0124] Calculation of important parameters for evaluating technical and tactical skills, including:

[0125] Calculate the pressure system of the red and blue teams in one round of the game. Taking the round as the unit, combine the proportion of offensive and counterattack tactics timestamps in the entire game, hit rate, effective punch ratio, and the number of sprint step label appearances in the proportion of moving actions, and calculate the weighted average with weights of 0.5, 0.2, 0.2, and 0.1 to get the pressure system of the red and blue teams.

[0126] Distance coefficient calculation. Calculate the average distance between the red and blue foot joints in a round as the distance coefficient.

[0127] Balance coefficient calculation: Calculate the average area of ​​the triangle formed by the two hand joints and the head joint in one round as the balance coefficient.

[0128] Movement coefficient calculation: calculate the ratio of the timestamp of the movement tag in one round to the whole game as the movement coefficient.

[0129] Step S7, the technical and tactical style and characteristics analysis and evaluation module, uses the machine learning classification module and integrates the features output by the technical and tactical statistics module in S6 to evaluate the technical and tactical style and characteristics of the athlete in a round. The specific implementation steps are as follows:

[0130] Based on the parameters output by the technical and tactical statistics module, the features used as the analysis and evaluation of technical and tactical styles and characteristics include: the number of offensive, defensive, and counterattack technical and tactical techniques and tactics in a round; the proportion of the number of straight punches, swinging punches, hooks, and combination punches in offensive technical and tactical techniques; the proportion of the number of various defensive action labels such as head holding and arm swinging in defensive technical and tactical techniques; the proportion of the number of refined footwork action labels in movements; the proportion of the number of counterattack single punches and counterattack combination punch labels in counterattack technical and tactical techniques, combined with the hit rate, effective punch ratio, pressure system, distance coefficient, balance coefficient, and movement coefficient.

[0131] One round generates two samples, namely the red boxer sample and the blue boxer sample, and each sample contains the above features.

[0132] Professional sports science practitioners annotate each sample with its style and characteristics. The annotation labels include:

[0133] Technical and tactical style: active attack, defensive counterattack, offensive round trip, roaming

[0134] Technical and tactical features: fake moves, direct attack, tight pressure, attack first and then defend

[0135] By integrating the above features and labels, a classification model based on machine learning is constructed. The alternative models include K-nearest neighbor, decision tree, random forest and other algorithms. The model with the best classification performance is selected as the technical and tactical style and characteristics evaluation model.

[0136] Important technical and tactical parameters are input into the trained machine learning-based classification model to evaluate the boxer's technical and tactical style and characteristics within a round.

[0137] Furthermore, a standard action knowledge base is established, a standard execution method of boxing actions is set, the action data of professional boxers are preprocessed and action recognition is performed, and important parameters such as the speed, angle, acceleration, etc. of the athlete when using boxing actions are calculated.

[0138] By using the standard action knowledge base, the actual punching data of boxers can be compared with the standard actions for similarity. By analyzing the deviation of the angle, speed and other parameters of boxing actions of athletes from the corresponding standard actions, the quality of athletes' actions can be analyzed and evaluated.

[0139] Furthermore, the dynamic time warping (DTW) algorithm is used to process the similarity comparison of time series data and calculate the deviations of various parameters between the actual punching data of ordinary boxers and the standard action data.

[0140] The above description is only a preferred embodiment of the present invention, but is not intended to limit the present invention. Those skilled in the art may make possible changes and modifications to the present invention using the technical contents disclosed above without departing from the scope of the technical solution of the present invention, or modify it into an equivalent embodiment of equivalent changes. Any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall fall within the protection scope of the technical solution of the present invention.

Claims

1. An intelligent evaluation method for boxing techniques and tactics, characterized in that: The method comprises the following steps: Step S1, deploy four cameras around the boxing ring to collect action video data of the red and blue boxers; calibrate the cameras to obtain the intrinsic parameter matrix and extrinsic parameter matrix of each camera; Step S2: First, the OpenPose deep learning algorithm is used to detect the human joints in the acquired action video data, that is, the two-dimensional plane coordinates of the joints are obtained. Then, according to the principle of the pinhole imaging model and combined with the OpenPose confidence, the three-dimensional coordinates of the joints are obtained, that is, the time series data of the red and blue sides are obtained respectively. Step S3, preprocessing the time series data: analyzing the time series data, identifying and marking the timestamps where the outliers are located in the time series data, then setting the data values ​​of the timestamps where the outliers are located to blank and filling the blank values, and then performing smoothing filtering on the filled time series data to obtain the preprocessed time series data; Step S4, using the time series classification task to realize action recognition, and then assigning an action label to each timestamp of the preprocessed time series data; Step S5, annotating the time series data with action labels, including sprinting steps, circling steps, sliding steps and passive marking, punching hand marking and punching position marking, and marking effective punches and ineffective punches in attack and counterattack tactics; Step S6, counting the boxer's use of skills and tactics: according to the data obtained in step S5, respectively counting the number and proportion of offensive, defensive, counterattack, and feint skills and tactics, statistics of the dominant hand of the skills and tactics, statistics of the hitting parts, calculation of the hit rate and effective punch ratio, and important parameters for evaluating the skills and tactics; Step S7, labeling the style and characteristics of the time series preprocessed in step S3, integrating the labels with the important technical and tactical parameters obtained in step S6 to build a classification model based on machine learning, and inputting the important technical and tactical parameters into the trained classification model based on machine learning to evaluate the technical and tactical style and characteristics of the boxer in a round.

2. The intelligent evaluation method for boxing techniques and tactics according to claim 1 is characterized in that: The specific steps of step S2 are as follows: Step S2-1, using the OpenPose deep learning algorithm to detect the human joints in the acquired action video data, and obtain the joint points P i The two-dimensional plane coordinates (u i , v i ), i=1, 2, 3 or 4; Step S2-2, using the pinhole imaging model principle to obtain the three-dimensional coordinates P (X w ,Y w ,Z w ): Among them, f x and f y are the horizontal focal length and vertical focal length of the camera respectively; u0 and v0 are the pixel coordinates of the intersection of the camera optical axis and the imaging plane, M i is the external parameter matrix from the world coordinate system to the camera i coordinate system, x w ,y w 、z w is the three-dimensional coordinate of the joint point P under ideal conditions, z c is the proportionality coefficient; Step S2-3, in order to avoid noise, it is necessary to estimate the spatial position of the real joint point R, consider the distance between the rays emitted by the camera, set a spherical area, and the radius of the sphere is 10 times the shortest distance between the rays. The three-dimensional coordinate formula of the real joint point R is as follows: Among them, l i is the ray emitted by the i-th camera, |P,l i | represents the ideal state of the joint point P to the ray l i The shortest distance, ρ i is the Openpose confidence parameter, n=4; Step S2-4, the action video data covers multiple joints of the human body, and time series data is obtained according to the real three-dimensional coordinates of the multiple joints.

3. The intelligent evaluation method for boxing techniques and tactics according to claim 1 is characterized in that: The specific steps of step S3 are as follows: Step S3-1, obtaining the boxer's motion data in an unobstructed environment, constructing a training data set that can reflect the real human body movement law, and using the athlete training data collected in the boxing ring environment as the test data set; Step S3-2, using a normalization method to map the value range of the time series data under the training data set and the test data set to a real number space from 0 to 1; Step S3-3, inputting the normalized training data set into the Transformer model, so that the Transformer model learns the normal pattern of boxing movement; Step S3-4, input the normalized test data set into the Transformer model obtained in step S3-3, construct anomaly score using reconstruction error MSE, set a threshold, and select the top 5% of data with anomaly scores greater than the threshold as abnormal data segments, in timestamp units, otherwise, they are normal data segments; Step S3-5, empty the abnormal data segment and fill it with normal values ​​through interpolation or Transformer prediction; Step S3-6, using adaptive filtering technology to perform smoothing filtering on the filled time series data.

4. The intelligent evaluation method for boxing techniques and tactics according to claim 3 is characterized in that: In step S3-5, the abnormal data segments are set to empty and filled with normal values. Specifically, when there are few abnormal data segments, the interpolation method is adopted, and the adjacent normal data points are used to estimate the values ​​of the abnormal data segments through linear interpolation and spline interpolation; when there are many abnormal data segments, the trained Transformer model or regression model is used to predict the abnormal data segments and fill them with predicted values.

5. The intelligent evaluation method for boxing techniques and tactics according to claim 1 is characterized in that: The specific steps of step S4 are as follows: Step S4-1, using the time series classification task to realize action recognition, classifying actions into offensive actions, defensive actions, counter-attack actions and feints, annotating the pre-processed time series data, and dividing it into a training set and a test set; Step S4-2, using the training set to train a multivariate time series classification model based on the LSTM architecture; Step S4-3, input the preprocessed time series data into the trained multivariate time series classification model based on the LSTM architecture, and assign a corresponding action label to each timestamp of the preprocessed time series data.

6. The intelligent evaluation method for boxing techniques and tactics according to claim 1 is characterized in that: The specific steps of step S5 are as follows: Step S5-1, filter out the timestamps corresponding to the movement categories in the action tags, divide the steps into sprinting, circling, sliding and passive steps according to the displacement, speed, rotation angle and acceleration of the joint points, and perform fine-grained annotation of the steps for the time series data of the movement categories in the action tags; Step S5-2, filtering out the timestamps corresponding to the attack and counterattack in the action tags, marking the puncher according to the coordinate values, movement speed, and acceleration of both hands, and marking the hitting parts according to the attack and counterattack time series data in the action tags according to the distance from the opponent's joint points; Step S5-3, taking each interaction between the two parties as a unit, combining the puncher and the hitting part in each interaction, identifying whether there is a hit or not, and at the same time, marking the hitting effect of the time series with action labels as attack and defense, that is, marking effective punches and invalid punches.

7. The intelligent evaluation method for boxing techniques and tactics according to claim 1 is characterized in that: The important parameters for evaluating techniques and tactics described in step S6 include: pressure coefficient, distance coefficient, balance coefficient and movement coefficient.

8. The intelligent evaluation method for boxing techniques and tactics according to claim 1 is characterized in that: The specific steps of step S7 are as follows: Step S7-1, labeling the style and characteristics of the time series data obtained after preprocessing in step S3, and dividing it into a training set and a test set, the style labels include active offensive type, defensive counterattack type, offensive round-trip type, and wandering type, and the characteristic labels include fake moves, direct attacks, tight pressure, and attack first and then defense; Step S7-2, integrating the number statistics and proportion calculation of the offensive, defensive and counterattack techniques and tactics described in step S6, the statistics of the dominant hand of the techniques and tactics, the statistics of the hitting parts, the calculation of the hit rate and the effective punch ratio, and the important parameters of the techniques and tactics as well as the style and characteristic labels, to build and train a classification model based on machine learning; Step S7-3, inputting important technical and tactical parameters into the trained classification model based on machine learning to evaluate the technical and tactical style and characteristics of the boxer in a round.

9. An intelligent evaluation system for boxing techniques and tactics, used to implement the method described in any one of claims 1 to 9, characterized in that: include: Acquisition module, motion capture module, time series preprocessing module, motion recognition module, motion analysis module, technical and tactical statistics module, technical and tactical style and characteristics analysis and evaluation module; The acquisition module is connected to the motion capture model and acquires the motion video data of the red and blue sides around the boxing ring; The motion capture module obtains the action video data of the red and blue parties through the input channel, and obtains the time series data of the red and blue parties through the OpenPose deep learning algorithm and the pinhole imaging model principle, combined with the OpenPose confidence parameter; The time series preprocessing module includes a time series anomaly detection submodule, an anomaly removal and filling submodule, and a smoothing filter submodule. The time series anomaly detection submodule analyzes the time series data, identifies and marks the timestamps where the outliers are located in the time series data, the anomaly removal and filling submodule sets the data value of the timestamp where the outliers are located to empty, and fills the empty value by interpolation or Transformer prediction, and the smoothing filter submodule performs smoothing filtering on the filled time series data to obtain the smoothed filtered time series data as the output of the time series preprocessing module; The action recognition module realizes action recognition by using the time series classification task, and then assigns an action label to each timestamp of the preprocessed time series data; The action analysis labels the time series data with action tags, including sprint steps, circling steps, sliding steps and passive labels, punching hand labels and punching parts labels, and labels of effective punches and ineffective punches in attack and counterattack tactics; The technical and tactical statistics module counts the boxer's use of technical and tactical skills, including the number and percentage of offensive, defensive and counterattack technical and tactical skills; technical and tactical punching dominant hand statistics boxer's dominant hand usage pattern; and statistics of striking parts; Calculation of hit rate and effective punch ratio; important parameters for evaluating skills and tactics; The technical and tactical style and characteristic analysis and evaluation module constructs a classification model based on machine learning according to the output features of the technical and tactical statistics module combined with style and characteristic labels, and uses the classification model based on machine learning to implement technical and tactical style evaluation of boxing.

Images