Badminton motion data analysis methods and systems combining big data

Abstract

This invention relates to the field of motion analysis technology, specifically disclosing a badminton motion data analysis method and system that combines big data. It captures athletes' joint angles, speeds, and accelerations in real time, and uses continuous wavelet transform and phase synchronization coefficients to calculate the consistency coefficient on the time series, ensuring data accuracy and reliability. Based on accurate data, it extracts short-term fine-tuning features and long-term trend features to construct comprehensive time-dimensional features. It also extracts joint angle change features and body posture features to form comprehensive spatial-dimensional features. These features are integrated into a comprehensive feature representation, input into a clustering model to generate an athlete's posture recognition score. Based on the score results, the athlete's posture is classified. For severely abnormal postures, it provides real-time feedback instructions based on spatiotemporal location to guide athletes in adjusting their movements, and establishes a closed-loop feedback system to monitor the correction effect.

Description

Technical Field

[0001] This invention relates to the field of motion analysis technology, specifically to a method and system for analyzing badminton motion data using big data. Background Technology

[0002] Badminton is a widely popular sport, and the mastery and training of its techniques significantly impact athletes' overall performance. The development of big data technology has opened up new possibilities for the collection, analysis, and application of badminton data. One study proposed a "wireless intelligent motion correction system" based on a "three-friction nanogenerator" to monitor athletes' "pedal frequency" and "pedal force" data in real time, providing feedback to coaches through data analysis to help improve athletes' techniques. Other studies have explored using "big data stream" analytics to obtain train braking performance information for optimizing railway operations. These researches indicate that "big data" technology has broad application prospects in sports training monitoring and optimization.

[0003] The existing technology has the following shortcomings:

[0004] Athlete motion data contains information across multiple temporal and spatial scales, such as short-term minor adjustments and long-term overall trends. Effectively fusing these multi-scale features to ensure they complement each other rather than interfere with each other during feature extraction is a highly challenging problem. Failure to properly address the fusion of multi-scale features may result in the omission of important motion details or overfitting, causing the system to perform poorly on new data and unable to provide effective personalized guidance. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for badminton motion data analysis that combines big data, in order to solve the problems mentioned above.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] The badminton motion data analysis method that combines big data includes the following steps:

[0008] S1: Captures athletes' motion data in real time and performs accuracy analysis on the motion data to ensure data accuracy;

[0009] The motion data includes: the athlete's joint angles, speed, and acceleration information;

[0010] S2: Based on accurate action data, extract short-term fine-tuning features for the data in each small window, merge multiple consecutive small windows into a larger window, and extract long-term trend features; construct a comprehensive time-dimensional feature from the short-term fine-tuning features and the long-term trend features.

[0011] S3: Based on accurate motion data, through geometric transformation and topological analysis of the motion data, joint angle change features and body posture features are extracted, and the joint angle change features and body posture features are built into comprehensive spatial dimension features.

[0012] S4: Construct a comprehensive feature representation by integrating the time dimension features and the spatial dimension features, and use it as input to the action model. Output the athlete's posture recognition score based on the action model.

[0013] S5: Based on the model output, classify the athlete's posture recognition into normal posture, slightly abnormal posture, and severely abnormal posture.

[0014] S6: Provide early warnings and corrections based on severely abnormal postures.

[0015] As a further aspect of the present invention: the real-time capture of the athlete's motion data and the accuracy analysis of the motion data are performed to ensure the accuracy of the data;

[0016] The system collects athletes' motion data in real time, including joint angles, speed, and acceleration information. It analyzes the consistency of the motion data over time and calculates the consistency coefficient of the motion data over time based on the interrelationships and error range of the motion data. It then determines whether the consistency coefficient of the motion data is greater than or equal to a preset threshold. If it is, the data is recorded as accurate; otherwise, it is recorded as inaccurate.

[0017] As a further aspect of the present invention: the logic for obtaining the consistency coefficient is as follows:

[0018] Real-time acquisition of athlete's motion data, including joint angles, velocity, and acceleration information, organized into a time series and recording the value at each moment;

[0019] Multi-scale analysis of time series is performed using continuous wavelet transform to generate wavelet transform coefficients that reflect signal characteristics, including amplitude and phase information of the signal.

[0020] Phase information is extracted from the wavelet transform results, the phase difference between joint angle, velocity and acceleration is calculated, and the synchronization degree between signals is analyzed based on the phase difference to obtain the phase synchronization coefficient.

[0021] By combining the signal amplitude and phase synchronization coefficient of different data dimensions, the consistency of the time series at various scales and time points is weighted and calculated to obtain the time series consistency coefficient of the motion data.

[0022] The expression for calculating the synchronization coefficient is as follows:

[0023]

[0024] In the formula, R(a,b) represents the phase synchronization coefficient, N represents the total length of the time series, t represents the time acquisition point, j represents the imaginary unit, and Δφ xy (a,b) represents the phase difference between the joint angle signal and the velocity signal;

[0025] The formula for calculating the consistency coefficient is as follows:

[0026]

[0027] In the formula, C represents the consistency coefficient, A represents the range of wavelet scale values, a represents the scale parameter, b represents the translation parameter, and W... x (a,b) represent the wavelet coefficients of the joint angle, W y (a,b) represent the wavelet coefficients of velocity, W z (a,b) represents the wavelet coefficients of acceleration, |W x (a,b)| represents the amplitude of the wavelet coefficients at the joint angle, |W y (a,b) represents the amplitude of the wavelet coefficients for velocity, |W z (a,b)| represents the amplitude of the wavelet coefficients of acceleration.

[0028] As a further aspect of the present invention: the process of obtaining the comprehensive time dimension features is as follows:

[0029] The real-time collected motion data is divided into multiple small windows according to fixed time intervals. Each small window contains time series data of joint angle, velocity and acceleration.

[0030] For each small window, calculate the short-term fine-tuning characteristics of the action data, including the range of change, the first-order difference average, and the statistical characteristics of volatility;

[0031] Multiple consecutive small windows are combined into a larger window. Within the larger window, the most significant short-term features are extracted based on the time series trend through a recursive aggregation method to construct long-term trend features.

[0032] Short-term fine-tuning features and long-term trend features are directly spliced ​​together in chronological order to form a comprehensive time-dimensional feature.

[0033] As a further aspect of the present invention: the process of obtaining the comprehensive spatial dimension features is as follows:

[0034] The joint angle change data is embedded into a three-dimensional Euclidean space. By constructing a local coordinate system, the local curvature, first derivative curvature, second derivative curvature, and third derivative curvature of the trajectory are calculated to extract higher-order geometric change features.

[0035] A topological graph model is constructed based on joint trajectory point clouds, and topological invariants are extracted through persistent homology analysis to characterize the global topological properties of body posture.

[0036] By combining the results of higher-order curvature analysis and topological invariant features, joint angle variation features and body posture features are extracted, and comprehensive spatial dimension features are constructed by directly splicing time series data.

[0037] As a further aspect of the present invention: the construction of a comprehensive feature representation by integrating time-dimensional features and spatial-dimensional features, which is then used as input to the action model, and the output of the athlete's posture recognition score based on the action model, specifically includes:

[0038] Based on comprehensive time and spatial features, an action model is constructed. The action model is a clustering model. The model determines the local density distribution and global connectivity of the feature points by analyzing the distribution characteristics of each feature point in the multi-dimensional feature space and the relative relationship between the points. This determines the cluster center point and the corresponding set of feature points.

[0039] For each cluster center, a dynamic influence factor is calculated based on the local distribution of its surrounding feature points. The dynamic influence factor reflects the strength of the influence of the feature points on the cluster center.

[0040] By combining the dynamic influence factor of each feature point with the local density information of the cluster center, the overall distribution characteristics of the cluster are comprehensively evaluated, and the characteristic relationships between different clusters are further analyzed.

[0041] The comprehensive evaluation results are then converted into a posture recognition score for the athlete.

[0042] As a further aspect of the present invention: the specific calculation process for the pose recognition score is as follows:

[0043] The distance between feature points is calculated using the following expression:

[0044]

[0045] In the formula, w k D represents the local distribution weights of the k-th feature, where k represents the number of dimensions of the feature. cd This represents the distance between the c-th sample point and the d-th sample point. This represents the value of feature k at the c-th sample point. The value of feature k at the d-th sample point, where α represents the nonlinear expansion coefficient, and c and d represent sample points;

[0046] Based on the distance between feature points, and considering the dynamic changes in features across both time and space dimensions, the dynamic influence factor of each cluster center is calculated. The calculation expression is as follows:

[0047]

[0048] Where, γ c The local density at the c-th center point is expressed as: γ c = σ represents the kernel function parameter;

[0049] Based on dynamic influencing factors, the label of each cluster is determined and the classification is completed using the model adaptive optimization strategy, forming the basic structure of the action model;

[0050] Based on the clustering results, the pose recognition score is calculated using the following expression:

[0051]

[0052] In the formula, N c Represents the number of cluster centers, |C c | represents the number of samples in the c-th cluster, δ d D represents the dynamic influence factor of the d-th sample. df S represents the distance between feature points. p The expression represents the pose recognition score, where d and f represent sample points, and e represents the logarithm of the natural number base.

[0053] As a further aspect of the present invention: the athlete's posture recognition is divided according to the model output, including normal posture, slightly abnormal posture, and severely abnormal posture, specifically including:

[0054] The athlete's posture recognition score is compared with a first threshold. If the athlete's posture recognition score is greater than or equal to the first threshold, it is recorded as a normal posture. If the athlete's posture recognition score is less than the first threshold, the athlete's posture recognition score is compared with a second threshold. If the athlete's posture recognition score is greater than or equal to the second threshold, it is recorded as a slightly abnormal posture. If the athlete's posture recognition score is less than the second threshold, it is recorded as a severely abnormal posture.

[0055] As a further aspect of the present invention: the early warning and correction based on severe abnormal posture specifically includes:

[0056] For athletes with severe abnormal postures, real-time feedback instructions based on spatiotemporal location are provided. These instructions are displayed in real time through wearable devices and a virtual reality training environment to guide athletes in making adjustments.

[0057] By monitoring the effect of corrective actions in real time, the athlete's posture recognition score is re-output through the action model and compared with a first threshold. If it is greater than or equal to the first threshold, it means that the correction is successful. The correction intensity is adjusted according to the athlete's adaptability to form a self-optimizing correction process.

[0058] A badminton motion data analysis system that integrates big data includes:

[0059] The data acquisition module is used to capture the athlete's motion data in real time;

[0060] A data accuracy analysis module is used to perform accuracy analysis on motion data to ensure data accuracy.

[0061] The data feature extraction module, based on accurate motion data, extracts short-term fine-tuning features for the data within each small window, merges multiple consecutive small windows into a larger window, and extracts long-term trend features; constructs comprehensive time-dimensional features from the short-term fine-tuning features and long-term trend features; and extracts joint angle change features and body posture features by performing geometric transformation and topological analysis on the motion data, and constructs comprehensive spatial-dimensional features from the joint angle change features and body posture features.

[0062] The motion model construction and recognition module constructs a comprehensive feature representation by integrating time-dimensional features and spatial-dimensional features, which serves as input to the motion model, and outputs a posture recognition score for the athlete based on the motion model.

[0063] The athlete posture segmentation module classifies the athlete's posture based on the output of the model, including normal posture, slightly abnormal posture, and severely abnormal posture.

[0064] The correction and early warning module performs early warning and correction based on severe abnormal posture.

[0065] The beneficial effects of this invention are:

[0066] (1) This invention captures joint angle, velocity, and acceleration information of athletes in real time and with high precision during exercise. By combining continuous wavelet transform with phase synchronization coefficient calculation methods, consistency analysis is performed on time series data, ensuring the high accuracy and reliability of the collected data. Furthermore, by performing geometric transformation and topological analysis on these raw data, detailed joint angle change features and body posture features are extracted to construct comprehensive spatial dimension features. This multi-level data processing method can not only accurately describe the athlete's movement patterns but also identify potential technical defects or non-standard movements. Based on these detailed and accurate analysis results, coaches and athletes can scientifically evaluate movement posture, promptly identify and correct subtle movement deviations, thereby optimizing training effects and significantly improving athletic performance. More importantly, this method helps prevent chronic injuries caused by maintaining improper postures for a long time, ensuring the health and sustainable development of athletes' careers. By organically combining modern sensing technology and complex data analysis methods, this invention provides unprecedented technical support and scientific basis for badminton training, ushering in a new era of personalized and refined training.

[0067] (2) This invention not only encompasses advanced data acquisition and analysis technologies but also innovatively introduces a closed-loop feedback system, enabling real-time monitoring and dynamic adjustment of athletes' movements. When a severely abnormal posture is detected, the system can instantly generate precise feedback instructions based on spatiotemporal location and intuitively convey these adjustments to the athlete through wearable devices or a virtual reality training environment. Furthermore, the system possesses intelligent adjustment capabilities, automatically optimizing the intensity and frequency of corrections based on each athlete's specific adaptation and training progress, forming a self-iteratory and continuously improving process. This highly personalized training guidance method ensures that each athlete receives a training plan best suited to their individual needs, improving training efficiency and effectiveness, and promoting continuous technological progress and development. Simultaneously, coaches can utilize the detailed data analysis and scientific basis provided by the system to formulate more precise and effective training plans, thereby achieving highly customized and scientific management of the training process. Attached Figure Description

[0068] The invention will now be further described with reference to the accompanying drawings.

[0069] Figure 1 This is a flowchart illustrating the specific steps of the badminton motion data analysis method combining big data according to the present invention.

[0070] Figure 2 This is a flowchart of the badminton motion data analysis system that combines big data in this invention. Detailed Implementation

[0071] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0072] Please see Figure 1 As shown, this invention is a badminton motion data analysis method combining big data, including the following steps:

[0073] S1: Captures athletes' motion data in real time and performs accuracy analysis on the motion data to ensure data accuracy;

[0074] The motion data includes: the athlete's joint angles, speed, and acceleration information;

[0075] S2: Based on accurate action data, extract short-term fine-tuning features for the data in each small window, merge multiple consecutive small windows into a larger window, and extract long-term trend features; construct a comprehensive time-dimensional feature from the short-term fine-tuning features and the long-term trend features.

[0076] S3: Based on accurate motion data, through geometric transformation and topological analysis of the motion data, joint angle change features and body posture features are extracted, and the joint angle change features and body posture features are built into comprehensive spatial dimension features.

[0077] S4: Construct a comprehensive feature representation by integrating the time dimension features and the spatial dimension features, and use it as input to the action model. Output the athlete's posture recognition score based on the action model.

[0078] S5: Based on the model output, classify the athlete's posture recognition into normal posture, slightly abnormal posture, and severely abnormal posture.

[0079] S6: Provide early warnings and corrections based on severely abnormal postures.

[0080] In S1, the athlete's motion data is captured in real time, and the data is analyzed for accuracy to ensure its accuracy. Specifically, this includes:

[0081] This invention employs a self-powered sensor based on a "triple-friction nanogenerator," integrated into sports equipment to collect athletes' motion data in real time, including joint angles, speeds, and accelerations. The sensor requires no external power supply, making it easy for athletes to wear.

[0082] The system collects athletes' motion data in real time, including joint angles, speed, and acceleration information. It analyzes the consistency of the motion data over time and calculates the consistency coefficient of the motion data over time based on the interrelationships and error range of the motion data. It then determines whether the consistency coefficient of the motion data is greater than or equal to a preset threshold. If it is, the data is recorded as accurate; otherwise, it is recorded as inaccurate.

[0083] The logic for obtaining the consistency coefficient is as follows:

[0084] Real-time acquisition of athletes' joint angles, velocities, and accelerations is used to construct a time series: S = {x} t ,y t ,z t}, where t represents the time acquisition point, x t y represents the value of the joint angle at time t. t z represents the velocity at time t. t This represents the value of acceleration at time t;

[0085] For the time series S={x t ,y t ,z t Perform a continuous wavelet transform to obtain the wavelet coefficients W. x (a,b),W y (a,b),W z (a, b), the calculation expression is:

[0086]

[0087] In the formula, a represents the scaling parameter, b represents the translation parameter, ψ represents the mother wavelet function, and W x (a,b) represent the wavelet coefficients of the joint angle, W y (a,b) represent the wavelet coefficients of velocity, W z (a,b) represents the wavelet coefficients of acceleration;

[0088] Phase information in each dimension is extracted using wavelet transform, and the calculation expression is as follows:

[0089] φ x (a,b)=arg(W x (a,b));

[0090] φ y (a,b)=arg(W y (a,b));

[0091] φ z (a,b)=arg(W z (a,b));

[0092] In the formula, φ x (a,b) represents the phase information of the joint angle, φ y (a,b) represents the phase information of the velocity, φ z (a,b) represents the bit information of acceleration, and arg represents the phase angle operation of taking a complex number;

[0093] The phase difference is calculated using the following expression:

[0094] Δφ xy (a,b)=φ x (a,b)-φ y (a,b);

[0095] Δφ yz (a,b)=φ y (a,b)-φ z (a,b);

[0096] In the formula, Δφ xy (a,b) represents the phase difference Δφ between the joint angle signal and the velocity signal. yz (a,b) represents the phase difference between the velocity signal and the acceleration signal;

[0097] The phase synchronization coefficient is constructed, and its calculation expression is as follows:

[0098]

[0099] In the formula, R(a,b) represents the phase synchronization coefficient, N represents the total length of the time series, t represents the time acquisition point, j represents the imaginary unit, and Δφ xy (a,b) represents the phase difference between the joint angle signal and the velocity signal;

[0100] Combining wavelet coefficients and phase synchronization coefficients, the consistency coefficient of the time series is calculated using the following expression:

[0101]

[0102] In the formula, C represents the consistency coefficient, A represents the range of wavelet scale values, a represents the scale parameter, b represents the translation parameter, and W... x (a,b) represent the wavelet coefficients of the joint angle, W y (a,b) represent the wavelet coefficients of velocity, W z (a,b) represents the wavelet coefficients of acceleration, |W x (a,b)| represents the amplitude of the wavelet coefficients at the joint angle, |W y (a,b) represents the amplitude of the wavelet coefficients for velocity, |W z (a,b)| represents the amplitude of the wavelet coefficients of acceleration.

[0103] In S2: Based on accurate action data, short-term fine-tuning features are extracted for the data within each small window. Multiple consecutive small windows are merged into a larger window to extract long-term trend features. The short-term fine-tuning features and long-term trend features are then combined to construct a comprehensive time-dimensional feature set, specifically including:

[0104] The real-time collected motion data is divided into multiple small windows according to fixed time intervals. Each small window contains time series data of joint angle, velocity and acceleration.

[0105] For each small window, calculate the short-term variation characteristics of joint angle, velocity and acceleration, including the range of variation (maximum value minus minimum value), the first difference mean and volatility (the ratio of standard deviation to mean), and construct a short-term feature vector;

[0106] By recursively aggregating, multiple consecutive small windows are combined into a larger window. Within the larger window, the most significant short-term feature (the feature with the largest change) is selected based on the changing trend of the time series data and extracted as the long-term trend feature.

[0107] Within a larger window, short-term fine-tuning features and long-term trend features are directly concatenated in chronological order to form a comprehensive time-dimensional feature vector, so as to fully preserve short-term and long-term feature information and ensure the structural integrity of the data.

[0108] The sliding window approach is used to gradually move the boundaries of a larger window, introducing a new smaller window each time, recursively recalculating short-term features and long-term trend features, and updating the overall time dimension features.

[0109] In S3, based on accurate motion data, geometric transformations and topological analyses are performed on the motion data to extract joint angle variation features and body posture features. These features are then used to construct a comprehensive spatial dimension feature set, specifically including:

[0110] The system collects athletes' joint angles, velocities, and accelerations in real time, and performs interpolation and noise reduction on the collected data to ensure accuracy and continuity.

[0111] The data is normalized to conform to the same dimension for subsequent geometric analysis and topological modeling.

[0112] Joint angle change data are embedded into a three-dimensional Euclidean space to generate joint trajectory point clouds for each time step, denoted as .

[0113] In the formula, i represents the number of data points, and M represents the total number of data points;

[0114] By constructing a local coordinate system (based on the normal and tangent vectors of neighboring points), the local curvature, second derivative curvature, and third derivative curvature of the joint trajectory are calculated to extract higher-order geometric transformation features.

[0115] Construct a topological graph model G = (V, E) of the motion data based on the trajectory point cloud, where V represents the set of points in the trajectory point cloud and E represents the connection relationship between the points;

[0116] Persistent cohomology analysis is performed on the topology graph to extract the topological invariants of the trajectory (0-dimensional and 1-dimensional persistent barcodes) to describe the global topological features of body pose.

[0117] For each time step, the curvature change trajectory of the feature points is calculated based on the local curvature, generating a higher-order curvature vector. Among them, κ (i) Represents the curvature vector. Indicates first-order curvature. Indicates second-order curvature. Indicates third-order curvature;

[0118] In the topological graph model, key turning points and stable segments in the trajectory are extracted by utilizing the curvature differences and connection relationships between nodes to construct body posture features;

[0119] The joint angle change features and body posture features are directly spliced ​​into comprehensive spatial dimension features based on time series. A dynamic recursive approach is used to analyze newly added trajectory data points step by step. By recalculating higher-order curvature and topological invariants, the comprehensive spatial dimension features are updated in real time.

[0120] In S4, a comprehensive feature representation is constructed by integrating temporal and spatial features, which serves as input to the action model. Based on the action model, the athlete's posture recognition score is output, specifically including:

[0121] Based on comprehensive time and spatial features, an action model is constructed. The action model is a clustering model. The model determines the local density distribution and global connectivity of the feature points by analyzing the distribution characteristics of each feature point in the multi-dimensional feature space and the relative relationship between the points. This determines the cluster center point and the corresponding set of feature points.

[0122] For each cluster center, a dynamic influence factor is calculated based on the local distribution of its surrounding feature points. The dynamic influence factor reflects the strength of the influence of the feature points on the cluster center.

[0123] By combining the dynamic influence factor of each feature point with the local density information of the cluster center, the overall distribution characteristics of the cluster are comprehensively evaluated, and the characteristic relationships between different clusters are further analyzed.

[0124] The comprehensive evaluation results are converted into a posture recognition score for the athlete;

[0125] The specific calculation process for the pose recognition score is as follows:

[0126] The integrated time dimension features and integrated spatial dimension features are standardized separately to ensure that all feature dimensions have the same scale.

[0127] A multimodal feature clustering model is constructed. This model is based on the local density distribution and global connectivity of the feature subspace. It utilizes an improved density peak clustering algorithm and calculates the distance between feature points by defining the expression:

[0128]

[0129] In the formula, w k D represents the local distribution weights of the k-th feature, where k represents the number of dimensions of the feature. cd This represents the distance between the c-th sample point and the d-th sample point. This represents the value of feature k at the c-th sample point. The value of feature k at the d-th sample point, where α represents the nonlinear expansion coefficient, and c and d represent sample points;

[0130] Based on the distance between feature points, and considering the dynamic changes in features across both time and space dimensions, the dynamic influence factor of each cluster center is calculated. The calculation expression is as follows:

[0131]

[0132] Where, γ c The local density at the c-th center point is expressed as follows: σ represents the kernel function parameter;

[0133] Based on dynamic influencing factors, the label of each cluster is determined and the classification is completed using the model adaptive optimization strategy, forming the basic structure of the action model;

[0134] Based on the clustering results, the pose recognition score is calculated using the following expression:

[0135]

[0136] In the formula, N c Represents the number of cluster centers, |C c | represents the number of samples in the c-th cluster, δ d D represents the dynamic influence factor of the d-th sample. df S represents the distance between feature points. pThe pose recognition score is represented by d and f, where d and f represent sample points, and e represents the logarithm of the natural number base.

[0137] By inputting the comprehensive feature representation into the clustering algorithm model, the athlete's posture recognition score is obtained.

[0138] In S5, based on the model's output, the athlete's posture is categorized into normal posture, slightly abnormal posture, and severely abnormal posture, specifically including:

[0139] The athlete's posture recognition score is compared with a first threshold. If the athlete's posture recognition score is greater than or equal to the first threshold, it means that the athlete's posture is normal and is recorded as a normal posture. If the athlete's posture recognition score is less than the first threshold, the athlete's posture recognition score is compared with a second threshold. If the athlete's posture recognition score is greater than or equal to the second threshold, it means that the athlete's posture is slightly abnormal and is recorded as a slightly abnormal posture. If the athlete's posture recognition score is less than the second threshold, it means that the athlete's posture is seriously abnormal and needs to be adjusted in time and is recorded as a seriously abnormal posture.

[0140] In S6, early warnings and corrections are performed based on severe abnormal postures, specifically including:

[0141] For athletes with severe abnormal postures, real-time feedback instructions based on spatiotemporal location are provided. These instructions are displayed in real time through wearable devices or virtual reality training environments to guide athletes in making adjustments. This feedback is provided not only through visual or auditory means but also by combining tactile feedback systems to remind athletes to make timely adjustments.

[0142] Establish a closed-loop feedback system to monitor the effect of corrective actions in real time, re-output the athlete's posture recognition score through the action model, and compare it with a first threshold. If it is greater than or equal to the first threshold, it means that the correction is successful. Adjust the correction intensity according to the athlete's adaptability to form a self-optimizing correction process.

[0143] Please see Figure 2 As shown, the badminton motion data analysis system, which combines big data, includes:

[0144] The data acquisition module is used to capture the athlete's motion data in real time;

[0145] A data accuracy analysis module is used to perform accuracy analysis on motion data to ensure data accuracy.

[0146] The data feature extraction module, based on accurate motion data, extracts short-term fine-tuning features for the data within each small window, merges multiple consecutive small windows into a larger window, and extracts long-term trend features; constructs comprehensive time-dimensional features from the short-term fine-tuning features and long-term trend features; and extracts joint angle change features and body posture features by performing geometric transformation and topological analysis on the motion data, and constructs comprehensive spatial-dimensional features from the joint angle change features and body posture features.

[0147] The motion model construction and recognition module constructs a comprehensive feature representation by integrating time-dimensional features and spatial-dimensional features, which serves as input to the motion model, and outputs a posture recognition score for the athlete based on the motion model.

[0148] The athlete posture segmentation module classifies the athlete's posture based on the output of the model, including normal posture, slightly abnormal posture, and severely abnormal posture.

[0149] The correction and early warning module performs early warning and correction based on severe abnormal posture.

[0150] The working principle of this invention is as follows: By capturing the athlete's joint angles, velocities, and accelerations in real time, and using continuous wavelet transform and phase synchronization coefficients to calculate the consistency coefficient on the time series, data accuracy is ensured. Then, based on accurate data, a combination of short-window and long-window methods is used to extract short-term fine-tuning features and long-term trend features, constructing a comprehensive time-dimensional feature set. Simultaneously, through geometric transformation and topological analysis of the motion data, joint angle change features and body posture features are extracted, forming a comprehensive spatial-dimensional feature set. These two types of features are then integrated into a comprehensive feature representation, which is input into a clustering model to generate an athlete's posture recognition score. Based on the score results, the athlete's posture is classified into normal posture, slightly abnormal posture, and severely abnormal posture. For severely abnormal postures, real-time feedback instructions based on spatiotemporal location are provided, guiding the athlete to adjust their movements through wearable devices or virtual reality training environments. A closed-loop feedback system is established to monitor the correction effect, achieving a self-optimizing correction process. This technical solution not only improves the accuracy and efficiency of motion data analysis but also provides athletes with scientific basis and technical support, helping to improve athletic performance and reduce the risk of injury. Throughout the process, the fusion of multi-source data and the application of intelligent algorithms enabled a deep understanding and precise optimization of the details of badminton players' movements, demonstrating the important role of modern technology in sports training.

[0151] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0152] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0153] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0154] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0155] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A badminton action data analysis method combined with big data, characterized in that, The method comprises the following steps: S1: Real-time capture of athlete action data, and accuracy analysis of the action data to ensure data accuracy; The action data includes: joint angle, speed, and acceleration information of the athlete; S2: Based on the accurate action data, for the data in each small window, extract short-term fine-tuning features, combine multiple consecutive small windows into larger windows, and extract long-term trend features; the short-term fine-tuning features and the long-term trend features are constructed into comprehensive time dimension features; S3: Based on the accurate action data, by geometric transformation and topological analysis of the action data, extract joint angle change features and body posture features, and construct the joint angle change features and the body posture features into comprehensive spatial dimension features; S4: The comprehensive time dimension features and the comprehensive spatial dimension features are constructed into comprehensive feature representation as the action model input, and the athlete's posture recognition score is output according to the action model; S5: According to the output of the model output, the athlete's posture recognition is divided into normal posture, slight abnormal posture, and serious abnormal posture; S6: Based on the serious abnormal posture, pre-warning and correction are performed; The acquisition process of the comprehensive time dimension features is: The real-time collected action data is divided into multiple small windows according to a fixed time interval, and each small window contains time series data of joint angle, speed, and acceleration; For each small window, the short-term fine-tuning features of the action data are calculated, including the statistical features of the change amplitude range, the first-order difference average, and the fluctuation rate; Multiple consecutive small windows are combined into larger windows, and the most significant short-term features are extracted according to the time series trend in the larger windows by recursive aggregation method to construct long-term trend features; The short-term fine-tuning features and the long-term trend features are directly spliced in time sequence to form comprehensive time dimension features; The acquisition process of the comprehensive spatial dimension features is: The joint angle change data is embedded into a three-dimensional Euclidean space, the local curvature, the first-order derivative curvature, the second-order derivative curvature, and the third-order derivative curvature of the trajectory are calculated by constructing a local coordinate system, and high-order geometric change features are extracted; A topological graph model is constructed based on the joint trajectory point cloud, and topological invariants are extracted by persistent homology analysis to represent the global topological characteristics of the body posture; The joint angle change features and the body posture features are extracted by combining the high-order curvature analysis results and the topological invariant features, and the comprehensive spatial dimension features are constructed by directly splicing the time series.

2. The badminton motion data analysis method using big data according to claim 1, characterized in that, The real-time capture of athlete action data, and accuracy analysis of the action data to ensure data accuracy; Real-time collection of athlete action data, including joint angle, speed, and acceleration information of the athlete, analysis of the consistency of the action data in time series, calculation of the consistency coefficient of the action data in time series according to the cross-correlation and error range of the action data, and judgment of whether the consistency coefficient of the action data is greater than or equal to a preset threshold, if yes, it is recorded as accurate action data, if not, it is recorded as inaccurate action data.

3. The badminton motion data analysis method using big data according to claim 2, characterized in that, The acquisition logic of the consistency coefficient is: Real-time acquisition of athlete action data, including joint angle, speed and acceleration information, organized as a time series and recording the numerical value at each time; Through continuous wavelet transform, the time series is analyzed at multiple scales to generate wavelet transform coefficients reflecting the characteristics of the signal, which contains amplitude and phase information of the signal; From the wavelet transform results, the phase information is extracted, the phase difference between joint angle, speed and acceleration is calculated, and the synchronization degree between signals is analyzed according to the phase difference to obtain the phase synchronization coefficient; The consistency of the time series at each scale and time point is calculated by weighting the signal amplitude and phase synchronization coefficient of different data dimensions to obtain the time series consistency coefficient of the action data; The calculation expression of the synchronization coefficient is: ； wherein denotes a phase synchronization coefficient, denotes the total length of the time series, denotes a time acquisition point, denotes the imaginary unit, denotes the phase difference between the joint angle signal and the velocity signal; The calculation expression of the consistency coefficient is: ； wherein denotes the conformity coefficient, denotes the range of values of the wavelet scale, denotes the scale parameter, denotes the translation parameter, denotes the wavelet coefficients of the joint angle, denotes the wavelet coefficients of the velocity, denotes the wavelet coefficients of the acceleration, denotes the wavelet coefficients amplitude of the joint angle, denotes the wavelet coefficients amplitude of the velocity, denotes the wavelet coefficients amplitude of the acceleration.

4. The badminton motion data analysis method using big data according to claim 1, wherein The comprehensive time dimension feature and the comprehensive space dimension feature are constructed into a comprehensive feature representation, which is input into the action model, and the athlete's posture recognition score is output according to the action model, specifically including: Based on the comprehensive time dimension feature and the comprehensive space dimension feature, an action model is constructed, which is a clustering model. The model determines the local density distribution and global connectivity of feature points by analyzing the distribution characteristics of each feature point in the multi-dimensional feature space and the relative relationship between points, so as to determine the cluster center point and the corresponding feature point set; For each cluster center point, the dynamic influence factor is calculated according to the local distribution of the surrounding feature points, which reflects the influence strength of the feature points on the cluster center; Combined with the dynamic influence factor of each feature point and the local density information of the cluster center, the overall distribution characteristics of the cluster are comprehensively evaluated, and the feature relationship between different clusters is further analyzed; The comprehensive evaluation result is converted into the athlete's posture recognition score.

5. The badminton motion data analysis method using big data according to claim 4, wherein The specific calculation process of the posture recognition score is: The distance between feature points is calculated, and the calculation expression is: ； wherein denotes the local distribution weight of the dimension of the feature, denotes the distance between the sample point and the value of the feature at the sample point, value of the feature at the sample point, denotes the non-linear spreading coefficient, and denotes the sample point; Based on the distance between feature points, the dynamic change of the comprehensive time dimension and the comprehensive space dimension is combined to calculate the dynamic influence factor of each cluster center point, and the calculation expression is: ； wherein, represents the local density of the th center point, and the calculation expression is: , represents the kernel function parameter; Based on the dynamic influence factor, the model self-adaptive optimization strategy is used to determine the label of each cluster and complete the classification to form the basic structure of the action model; According to the clustering result, the posture recognition score is calculated, and the calculation expression is: ； In the formula, represents the number of cluster centers, represents the number of samples of the th cluster, represents the dynamic influence factor of the th sample, represents the distance between feature points, represents the pose recognition score, and represents the sample point, represents the natural number base logarithm.

6. The badminton motion data analysis method using big data according to claim 1, wherein The output of the model is used to divide the athlete's posture recognition, including normal posture, slight abnormal posture and serious abnormal posture, specifically including: The athlete's posture recognition score is compared with the first threshold value. If the athlete's posture recognition score is greater than or equal to the first threshold value, it is recorded as normal posture. If the athlete's posture recognition score is less than the first threshold value, the athlete's posture recognition score is compared with the second threshold value. If the athlete's posture recognition score is greater than or equal to the second threshold value, it is recorded as slight abnormal posture. If the athlete's posture recognition score is less than the second threshold value, it is recorded as serious abnormal posture.

7. The method of claim 1, wherein the method is a method of analyzing badminton motion data using big data. Based on the serious abnormal posture, pre-warning and correction are carried out, specifically including: Based on the serious abnormal posture of the athlete, real-time feedback instructions based on the space-time position of the athlete are provided, the feedback instructions are displayed in real time through a wearable device and a virtual reality training environment, and the athlete is guided to adjust; By monitoring the effect of the correction action in real time, the posture recognition score of the athlete is output again through the action model, and compared with the first threshold value. If it is greater than or equal to the first threshold value, it means that the correction is successful. The correction intensity is adjusted according to the adaptability of the athlete, and a self-optimizing correction process is formed.

8. A badminton action data analysis system combined with big data, characterized in that, The badminton action data analysis method combined with big data according to any one of claims 1-7, comprising: A data acquisition module for capturing action data of the athlete in real time; A data accuracy analysis module for analyzing the accuracy of the action data to ensure the accuracy of the data; A data feature extraction module based on accurate action data, which extracts short-term fine-tuning features for data in each small window, combines multiple consecutive small windows into larger windows, and extracts long-term trend features; The short-term fine-tuning features and long-term trend features are constructed into comprehensive time dimension features, the joint angle change features and body posture features are extracted by geometric transformation and topological analysis of the action data, and the joint angle change features and body posture features are constructed into comprehensive space dimension features; An action model construction and recognition module for constructing comprehensive time dimension features and comprehensive space dimension features into comprehensive feature representation as action model input, and outputting the posture recognition score of the athlete according to the action model; An athlete posture division module for dividing the posture recognition of the athlete according to the output of the model, including normal posture, slight abnormal posture, and serious abnormal posture; A correction warning module based on serious abnormal posture for warning and correction.

Images