A method for correcting the athletic posture of a Frisbee competitive player

Abstract

The present invention proposes a method for correcting the athletic posture of a Frisbee competitive player. The method comprises the following steps: obtaining acceleration and angular velocity data of the Frisbee throwing posture by installing an inertial measurement unit, performing information screening on the data, and classifying the screened data to generate quantitative data and non-quantitative data; converting the non-quantitative data into sub-quantitative posture data through a model; fusing the quantitative data and the sub-quantitative posture data to generate a comprehensive evaluation result of the player's posture behavior, and assigning a quantitative label value to the player's posture behavior, thereby effectively correcting the player's posture, reducing the risk of injury caused by incorrect posture, improving the athlete's overall performance level, and ensuring the safety and quality of Frisbee competitions.

Description

Technical Field

[0001] The present invention relates to the field of sports and athletics, and in particular to a method for correcting the athletic posture of a Frisbee athlete. Background Art

[0002] Ultimate Frisbee, also known as Frisbee or disc sports, is a sport played with a flying disc as its primary weapon. Participants demonstrate skill and strategy by throwing and catching the disc to score points or win team games. Ultimate Frisbee sports typically include various events, such as disc golf, mixed events, and rapid-fire events, attracting numerous enthusiasts.

[0003] However, although Frisbee competitions are widely welcomed and participated in, in actual sports, players may affect the trajectory and flight distance of the Frisbee due to inaccurate posture during the throwing and catching process, thereby reducing the effectiveness of the game; secondly, players need to maintain good movement stability in Frisbee competitions to ensure that the throwing and catching of the Frisbee can be carried out smoothly. In Frisbee competitions, players need to throw and catch the Frisbee. If the posture is unstable or the movements are uncoordinated, it may cause joint sprains, muscle strains or other physical injuries. Therefore, the angle and rotation control in Frisbee competitions have an important impact on the flight trajectory and distance of the Frisbee. Players need to accurately control the rotation speed and flight angle of the Frisbee, but there are problems that are difficult to master in practice, making it difficult to find problems with posture and movement in actual training. Therefore, long-term training with incorrect posture leads to various problems. Summary of the Invention

[0004] To address the aforementioned technical issues, the present invention provides a method for correcting a Frisbee player's posture. This method addresses issues such as inaccurate posture, poor stability, and difficulty controlling angles during Frisbee play. This method allows for accurate control of the disc's rotational speed and flight angle, while also addressing joint sprains, muscle strains, and other injuries caused by incorrect posture control during play.

[0005] In order to solve the above problems, the present invention adopts the following technical solutions:

[0006] A method for correcting the athletic posture of a Frisbee competitive player is characterized in that the method is implemented through a posture model and comprises the following steps: obtaining acceleration and angular velocity data of the player's Frisbee throwing posture, wherein the data is captured by an inertial measurement unit installed on the Frisbee; performing information screening on the captured posture data and classifying the screened data to generate quantitative data and non-quantitative data; converting the non-quantitative data into sub-quantitative posture data through a model; fusing the quantitative data and the sub-quantitative posture data to generate a comprehensive evaluation result of the player's posture behavior, and assigning a quantitative label value to the player's posture behavior, thereby accurately evaluating the player's posture behavior.

[0007] Furthermore, the non-quantitative data is converted into sub-quantitative posture data through a model, and the conversion is performed through a flying disc rotation model, a tilt angle model, and a flight direction model.

[0008] Furthermore, the captured posture data is screened and the screened data is classified to generate quantitative data and non-quantitative data by calculating the tilt angle and the rotation angular velocity to classify the quantitative data and the non-quantitative data.

[0009] Furthermore, the classification of the quantitative data and the non-quantified data by calculating the tilt angle and the rotation angular velocity includes the following steps:

[0010] S1, obtains characteristic values ​​such as rotation angular velocity and tilt angle from the information screening process;

[0011] S2, set the corresponding quantitative label and threshold for each feature;

[0012] S3, classify the data of each frame by comparing the feature value with the threshold.

[0013] Furthermore, the S3, performing the following classification operation on the data of each frame includes the following steps:

[0014] S31, calculating the absolute value of the rotation angular velocity and comparing it with the threshold value to classify the rotation angular velocity features;

[0015] S32, calculate the absolute value of the tilt angle and compare it with the threshold to classify the tilt angle feature.

[0016] Furthermore, the fusion of the quantitative data and the sub-quantified posture data to generate a comprehensive evaluation result of the player's posture behavior and assigning a quantitative label value to the player's posture behavior includes the following steps:

[0017] 1051. Fusing quantized data and sub-quantized posture data;

[0018] 1052. Generate comprehensive evaluation results of the players’ posture and behavior;

[0019] 106. Assign quantitative label values ​​to the players' gesture behaviors.

[0020] Furthermore, the fusion of the quantized data and the sub-quantized posture data includes the following steps:

[0021] 10511, standardize quantized and subquantized posture data;

[0022] 10512, computing the mutual information between the quantized data and the sub-quantized posture data;

[0023] 10513, calculating the phase relationship between posture data based on mutual information;

[0024] 10514, performing weighted averaging on the normalized quantized data and the sub-quantized posture data according to the mutual information correlation to obtain the fused data.

[0025] Furthermore, the standardization of the quantized and sub-quantized posture data includes the following steps:

[0026] i. Calculate the mean and standard deviation of the quantitative data and sub-quantized posture data;

[0027] ii, perform standardized calculations;

[0028] iii. Repeat steps i and ii for each point.

[0029] Furthermore, the standardized calculation is performed using the following formula: Where z is the standardized data point, μ is the mean, and σ is the standard deviation.

[0030] The present invention also discloses a method for evaluating the motion posture of a Frisbee player taking into account wind influence, which specifically comprises the following steps:

[0031] S21, obtaining wind speed and direction data, wherein the data is obtained by installing wind speed and direction sensors around the site;

[0032] S22, processing the captured data using a wind impact model to obtain trajectory prediction data and attitude stability data;

[0033] S23, fusing the trajectory prediction data and the posture stability data with the quantified data and the sub-quantified posture data to generate a comprehensive player posture behavior assessment result;

[0034] S24, assigning a quantitative label value to the player's posture behavior based on the comprehensive evaluation results, thereby accurately evaluating the quality of the player's posture behavior.

[0035] The present invention provides accurate assessment and guidance, aiming to improve players' competitive performance, reduce the risk of injury, and provide strong support for the development and training of Frisbee competitions, while preventing joint sprains, muscle strains or other physical injuries during exercise. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A flowchart of a method for correcting the athletic posture of a Frisbee competitive player proposed by the present invention;

[0037] Figure 2 A flow chart of a method for data fusion and evaluation of a method for correcting the motion posture of a Frisbee competitive player proposed by the present invention;

[0038] Figure 3 A flowchart of the standardization of quantified data and sub-quantified data for a method of correcting the athletic posture of a Frisbee competitive player proposed by the present invention;

[0039] Figure 4 This is a flow chart of information screening, feature extraction and classification for a method proposed by the present invention for correcting the athletic posture of disc players. DETAILED DESCRIPTION

[0040] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings, but the protection scope of the present invention is not limited to the following.

[0041] In order to make the objectives, technical solutions, and advantages of the present invention more clearly understood, the present invention is further described in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present invention and are not intended to limit the present invention. That is, the embodiments described herein are only some embodiments of the present invention, not all embodiments. Generally, the components of the embodiments of the present invention described and illustrated in the drawings herein may be arranged and designed in various different configurations.

[0042] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative work shall fall within the scope of protection of the present invention. It should be noted that relational terms such as "first" and "second" are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations.

[0043] Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not preclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0044] The features and performance of the present invention are further described in detail below with reference to the embodiments.

[0045] like Figure 1 As shown, a method for correcting the athletic posture of a Frisbee competitive player is characterized in that the method is implemented through a posture model and includes the following steps:

[0046] Step 101, obtaining acceleration and angular velocity data of a player's disc throwing posture, wherein the data is captured by an inertial measurement unit installed on the disc;

[0047] First, the inertial measurement unit (IMU) is installed, in one or more embodiments, at the center of the disc, typically at its center of gravity. This reduces the IMU's impact on flight stability because the sensor doesn't significantly alter the disc's balance. This allows for relatively stable data, suitable for measuring the disc's overall motion. In competitive disc racing, the IMU is a sensor system comprised of accelerometers and gyroscopes. These sensors can sense changes in the disc's acceleration and angular velocity in three dimensions, respectively. This device is designed to accurately track the disc's posture changes, revealing the details of the player's movements during the throw.

[0048] When a player prepares to throw a Frisbee, the inertial measurement unit (IMU) comes into play. The accelerometer first senses the disc's acceleration changes along each axis. It applies Newton's second law to measure acceleration based on the change in mass displacement. For example, when a player throws a Frisbee, the accelerometer detects the disc's acceleration increasing during the throw and then gradually decreasing during flight. This process provides the necessary data for subsequent attitude calculations. The gyroscope also plays a key role, sensing the disc's angular velocity, or rotational speed. This sensor uses the law of conservation of angular momentum to measure the disc's angular velocity by sensing the disc's rotational changes about each axis. For example, when a Frisbee is thrown and begins to spin, the gyroscope accurately senses the disc's rotational velocity, including not only roll and yaw but also pitch motion. This provides comprehensive data on the disc's throwing attitude.

[0049] As the disc is thrown, the inertial measurement unit (IMU) continuously collects acceleration and angular velocity data. This data is stored as digital signals in the sensor's internal memory and can be transmitted to an external device via Bluetooth or other wireless communication methods. This real-time data transmission ensures real-time monitoring and analysis of the player's throwing process. As the player throws the disc, the IMU continuously collects acceleration and angular velocity data. This raw data is stored as digital signals and can be transmitted to an external device for subsequent processing and analysis. However, raw data is not immediately usable as it may be affected by noise, vibration, and calibration errors. Therefore, data processing and filtering are required to improve its accuracy and reliability. This processing includes using digital filters to remove noise and performing calibration to correct for sensor errors. The processed data provides the foundation for subsequent attitude calculations. Ultimately, after data processing and fusion, in one or more embodiments, the acceleration and angular velocity data are converted into quaternions that accurately describe the disc's attitude in space, including its tilt angle and rotation. This data representation offers improved mathematical properties and computational efficiency, providing a more powerful tool for subsequent attitude assessment and analysis.

[0050] In summary, an inertial measurement unit (IMU) mounted on a Frisbee can fundamentally capture acceleration and angular velocity data during the throwing process. This data, after processing and conversion, is ultimately stored in the form of quaternions, providing a precise foundation for understanding the player's posture and behavior, further promoting the development and advancement of Frisbee competitions.

[0051] Step 102, filtering the captured posture data and classifying the filtered data to generate quantitative data and non-quantitative data, as shown in step 103;

[0052] In many cases, existing posture analysis technologies may directly evaluate sensor data without further processing, classification, or feature extraction. Sensors generate data at a high frequency, containing a significant amount of redundant information. Directly using this data leads to redundant analysis and computation, reducing efficiency. Furthermore, for non-quantitative posture features, such as the fluidity and coordination of movements, existing technologies may lack subjective analysis and judgment, thus failing to provide a more comprehensive assessment. To address these issues, a process has been introduced to filter and classify the captured posture data, generating both quantitative and non-quantitative data. It should be noted that quantitative data refers to data that, after feature extraction and threshold processing, can be used for relatively accurate classification and analysis. This data contains certain patterns and rules, such as the angular velocity characteristics of a particular movement falling within a certain range. Non-quantitative data, on the other hand, refers to data that is not easily categorized and analyzed directly using specific rules. It involves elusive characteristics due to its diversity and randomness. Initially, this data may be difficult to accurately classify and evaluate.

[0053] First, the captured data information is screened and feature extraction is performed. Feature extraction extracts information related to posture behavior assessment from the quaternion data. In one or more embodiments, rotation angle and rotation speed data information are extracted. The difference between quaternions can provide information about the rotation change of the posture. By calculating the difference of the quaternion and the corresponding angular velocity, the required features can be obtained. Specifically, assuming two adjacent quaternions, q prev Represents the quaternion of the previous frame, q curr The quaternion representing the current frame. To calculate the difference between two quaternions, quaternion multiplication and inverse operations are used in this embodiment.

[0054] Specifically, calculate the difference between the two quaternions and get the quaternion q diff : Then, through q diff Calculate the rotation angle and axis. This can be done by converting the quaternion to Euler angles or a rotation matrix, and then extracting the required angle information from it. The specific steps are as follows:

[0055] (a) Obtaining quaternion data: Obtaining quaternion data of the player's disc throwing posture from the inertial measurement unit;

[0056] (b) For each frame of data, calculate q diff , where q prev is the quaternion of the previous frame, q curr is the quaternion of the current frame.

[0057] (c) From q diff Calculate the angular velocity ω. This is done by taking qdiff Convert it into a rotation matrix and calculate the angular velocity of the rotation matrix to achieve it.

[0058] Specifically, use q diff Calculate the corresponding rotation matrix R diff The conversion from quaternion to rotation matrix can be done using the following formula:

[0059] where q x ,q y ,q z ,q w are the components of the quaternion. By comparing the rotation matrices between consecutive frames, the angular velocity can be calculated. The difference between the rotation matrices can be used to calculate the angular velocity ω. Specifically, the difference in the rotation matrices between consecutive frames can be used to estimate the angular velocity. Suppose there are two rotation matrices R1 and R2 for consecutive frames. The rotation matrix describes the rotation transformation from one coordinate system to another. The difference between the rotation matrices is calculated using the following formula: Where T represents the rotation operation of the matrix. Next, use calculus to estimate the angular velocity: Here Δθ is the rotation angle change calculated from ΔR, and Δt is the time interval between two consecutive frames.

[0060] Spin velocity is the rate of change of angle over time and can reflect the speed, compactness, and stability of a player's throwing motion. A higher spin velocity may mean the disc spins faster during a throw, which is associated with higher levels of skill and precision.

[0061] (d) From the quaternion difference q diff Calculate the tilt angle.

[0062] Tilt angle represents the angle of the disc relative to a reference plane and can be used to assess the disc's tilt. Changes in tilt angle can indicate changes in the disc's posture during a throw, such as a tumble or tilt. diff Converted to rotation matrix R diff , extract the tilt angle from the rotation matrix. Specifically, the tilt angle information is obtained by calculating the Euler angle of the rotation matrix (the rotation angle around the fixed coordinate axis). In the embodiment, it is obtained by the ZYX Euler angle (the rotation angle around the Z axis, Y axis and X axis). Specifically, assuming that the rotation matrix is ​​R, its Euler angle is expressed as (α, β, γ), where α is the rotation angle around the Z axis, β is the rotation angle around the Y axis, and γ is the rotation angle around the X axis. For the extraction of the ZYX Euler angle, the following formula can be used: β = arcsin(R 31 ),α=arctan2(-R 32 ,R 33),γ=arctan2(-R 21 ,R 11 ), where R ij Represents an element in the rotation matrix. In practical applications, the resulting angular velocity and tilt angle are used to determine the speed, stability, and postural changes of movements. Quantitative and non-quantitative data can be classified based on these feature values, leading to a more accurate assessment of the athlete's postural behavior.

[0063] Through the above steps, the captured quaternion data is converted into information with actual posture characteristics. In one or more embodiments, this information is converted into rotational angular velocity and tilt angle data. These characteristics are used in the subsequent classification step to determine quantitative and non-quantitative data. This feature extraction process helps to better understand the player's posture behavior and provides a meaningful data foundation for subsequent evaluation.

[0064] Next, we use the pre-set quantitative labels and thresholds to classify the data into quantitative data and non-quantitative data based on the numerical attributes of the features. The quantitative data are those that can be divided by clear numerical standards, while the non-quantitative data require subjective judgment and analysis. I will explain the classification steps in detail below, such as Figure 4 As shown,

[0065] S1. Obtaining characteristic values ​​such as rotation angular velocity and tilt angle from the feature extraction (information screening) process;

[0066] S2. Set the corresponding quantitative label and threshold for each feature, such as T rotation and T tilt ;

[0067] S3. For each frame of data, perform the following classification operations:

[0068] S31. For the rotational angular velocity feature: calculate the absolute value of the rotational angular velocity and compare it with T rotation If the absolute value is greater than the threshold, it is classified as quantitative data; otherwise, it is classified as non-quantitative data.

[0069] S32. For the tilt angle feature: calculate the absolute value of the tilt angle and compare it with T tilt If the absolute value is greater than the threshold, it is classified as non-quantitative data; otherwise, it is classified as quantitative data.

[0070] The data after classification operation is divided into two categories, obtaining a quantitative data set and a non-quantitative data set.

[0071] For example, for a frame with an angular velocity of 120° / s, according to the set threshold T rotation , because 120° / s>Trotation , this data will be classified as quantitative data. For another frame, the tilt angle is 25°, because 25° <T tilt , so this data will be classified as quantitative data.

[0072] It should be noted that angular velocity represents the speed of the player's rotation during the throw, while tilt angle represents the tilt angle of the disc in the air. These feature values ​​can help quantitatively describe the speed, stability, and variability of the disc's throwing motion. Quantitative data refers to data that can be categorized using clear numerical criteria or thresholds. The angular velocity and tilt angle obtained after feature extraction can be directly used to classify quantitative data because they are numerical attributes and can be compared with thresholds for classification. For example, if the angular velocity exceeds a certain threshold, the corresponding data can be classified as quantitative data, and threshold classification can clearly categorize the data into different action types. Both angular velocity and tilt angle fall within a certain range, making it possible to clearly identify the action type. Non-quantitative data refers to data that cannot be simply categorized using numerical values ​​or thresholds. In other words, non-quantitative data is data that cannot be clearly classified into a specific action type using threshold classification. In this case, non-quantitative data requires more subjective judgment and may be influenced by various factors such as context and disc state. Angular velocity and tilt angle are the results of feature extraction and serve as the basis for classifying data into quantitative and non-quantitative data. During classification, the data is divided into quantitative and non-quantitative data by comparing the specific values ​​of angular velocity and tilt angle with the set threshold. The angular velocity and tilt angle values ​​used in feature extraction are themselves quantitative descriptions of the player's posture and behavior. These features can be compared with the threshold to achieve classification. As for determining the threshold, there are many practical methods. In this implementation, the threshold was determined based on expert advice and experience, and adjusted through practice. I will not elaborate on this here.

[0073] Step 104, converting the non-quantized data into sub-quantized attitude data based on the flying disc rotation model, the tilt angle model, and the flight direction model;

[0074] Model mapping converts non-quantitative data into sub-quantitative data. This transforms raw, abstract non-quantitative data into more concrete, easier-to-understand sub-quantitative data that provides greater meaning when describing the disc's posture. For example, by converting angular velocity into rotation and tilt angle into pitch, we can better capture the disc's spin, tilt, and flight direction. This helps more accurately assess a player's posture and provides a more useful data foundation for further analysis and evaluation.

[0075] On this basis, the non-quantized data is finally converted into sub-quantized data through three model mappings. Specifically, the first is the Frisbee rotation model, in which quaternions are used to describe the rotation state of the Frisbee. For the generation of sub-quantized data, it is updated according to the angular velocity data. Specifically, assuming that the initial quaternion is q old =[q w ,q x ,q y ,q z ], the angular velocity is ω=[ω x ,ω y ,ω z ], the time step is Δt, and the quaternion update can use the numerical integration method. The specific formula is as follows: This will get the updated quaternion q new , which represents the rotation state of the disc. Through this mapping, the angular velocity data is converted to the rotation state, realizing the conversion from non-quantized data to sub-quantized data. Next is the tilt angle model, which is used to describe the tilt angle of the disc. The tilt angle is obtained by calculating the Euler angle of the rotation matrix. Specifically, from the quaternion q new The rotation matrix R is calculated, and the tilt angle information is obtained by calculating the Euler angle of the rotation matrix. In one or more embodiments, the conversion formula from the rotation matrix to the Euler angle is used:

[0076] roll=atan2(R 21 ,R 22 ), pitch=asin(-R 20 ),yaw=atan2(R 10 ,R 00 ), where R ij Represents the elements of the rotation matrix. Through these calculations, the information of the tilt angle is obtained. This will map the tilt angle data into sub-quantized attitude data. Finally, there is the flight direction model. In the flight direction model, angles or vectors are used to describe the flight direction. In some embodiments, angles are used, so the angle information of the flight direction is already included in the rotation model, and the angle of the flight direction is directly used as sub-quantized data; in some embodiments, the flight direction is described by a vector, and the mapping can be achieved through the following steps to obtain sub-quantized tilt angle data (roll, pitch, yaw) according to the tilt angle model. Construct the rotation matrix R based on the sub-quantized tilt angle data. In some embodiments, the Euler angle conversion formula R=R Z (yaw)·R y (pitch) R x (roll), where Rx, Ry, and Rz are the rotation matrices around the X, Y, and Z axes, respectively. Define an initial flight direction vector, such as v initial=[1,0,0] means the flight direction points to the X axis. The initial flight direction vector is transformed by the rotation matrix R to obtain the subquantized flight direction vector v new , the specific formula is as follows: new =R·v initila Through these steps, the non-quantized flight direction data is mapped to sub-quantized attitude data, which includes the information of the flight direction vector. This mapping process utilizes the information of the tilt angle model and transforms the initial vector through the rotation matrix to obtain the sub-quantized flight direction vector.

[0077] In summary, since non-quantitative data represents characteristics such as the fluidity and coordination of a player's movements, the goal of model mapping is to effectively convert these characteristics into sub-quantitative posture data for subsequent analysis. The disc rotation model generates the disc's rotational state and represents it using quaternions. By observing changes in the rotational state, we can understand the player's disc rotation at different stages. If the movement is smooth and coordinated, the rotational state changes are likely to be relatively smooth, while unsmooth movements may lead to unstable changes in the rotational state. The tilt angle model provides information on the disc's tilt angle, including roll, pitch, and yaw. This angle information reveals the disc's tilt relative to the reference direction. For smooth and coordinated movements, the tilt angle changes are likely to be smoother, while uncoordinated movements may result in drastic changes in the tilt angle. The flight direction model provides information on the disc's flight direction. This can be expressed as an angle or a vector. The fluidity and coordination of the movement may affect the changes in flight direction. Uncoordinated movements may result in frequent changes in flight direction, while smooth movements may result in a more stable flight direction. After feature extraction, classification and model mapping, the data can better reflect characteristics such as fluency and coordination, reduce the complexity of the data, improve the accuracy and effectiveness of the analysis, and enable a deeper understanding of the players' movement characteristics, thereby making more valuable evaluations and judgments.

[0078] like Figure 2 As described in steps 105 and 106, the quantitative data and the sub-quantified posture data are integrated to generate a comprehensive evaluation result of the player's posture behavior, and a quantitative label value is assigned to the player's posture behavior, thereby achieving an accurate evaluation of the player's posture behavior.

[0079] Step 1051: Fuse the quantized data and the sub-quantized posture data.

[0080] Step 10511, normalize the quantized data and sub-quantized posture data to ensure that they have the same scale, so that subsequent calculations are more accurate and consistent. The normalization process can be completed through the following steps, such as Figure 3 As shown:

[0081] I. For the quantized data and sub-quantized posture data, calculate their mean μ and standard deviation σ respectively, where, n is the total number of data points, x i is the i-th data point;

[0082] II. Perform normalization calculations. For each data point, use the following formula to perform normalization calculations: z is the normalized data point.

[0083] III. Repeat steps 1.1 and 1.2 for all quantized data and subquantized pose data to calculate and normalize each data point separately.

[0084] Step 10512. Calculate the mutual information between the quantized data and the sub-quantized posture data. The mutual information measures the mutual dependence between two random variables. The calculation formula for the mutual information is as follows: Where X and Y represent the quantized data and subquantized pose data, respectively. p(x,y) is the joint probability distribution, and p(x) and p(y) are the marginal probability distributions.

[0085] Step 10513. Based on the calculated mutual information value, calculate the weight to reflect the correlation between the quantized data and the sub-quantized posture data. The weight is calculated using the following formula: Among them, I max is the maximum possible value of mutual information.

[0086] Step 10514: Perform weighted averaging on the normalized quantized data and sub-quantized posture data based on mutual information correlation to obtain fused data. The weighted average calculation formula is as follows: used =w·X+(1-w)·Y, where F use is the fused data, X and Y represent the standardized quantitative data and sub-quantized posture data respectively.

[0087] Step 1052, generating a comprehensive evaluation result of the player's posture and behavior;

[0088] Specifically, in this step, the fused data is used as features to build an evaluation model. In one or more embodiments, a linear regression model is used: in, are evaluation scores, x1,x2,…,x n are the dimensions of the fused data, β0,β1,…,β n is a parameter of the model. By training the parameter β of the model, the model can better fit the training data set. In one or more embodiments, the mean square error is used, specifically, Where N is the number of training samples, y i is the actual assessment score, is the evaluation score predicted by the model. For each player's gesture, the corresponding fused data is extracted and fed into the trained model. The model then predicts an evaluation score. This score reflects the player's gesture performance across different dimensions, as the fused data contains information about all aspects of the gesture.

[0089] Step 106 assigns a quantitative label value to the player's posture behavior, so as to provide targeted training to the player based on the evaluation results, provide targeted guidance, and help the player continuously improve his or her movement skills and enhance the performance of the Frisbee competition.

[0090] Based on the generated comprehensive evaluation results, a quantitative label value is assigned to the athlete's posture behavior. This quantitative label value can be a digital score, grade, index, or other representation that clearly expresses the quality of the athlete's posture behavior. When designing the label value, a multidimensional label space is defined for the label. In one or more embodiments, the labels stability, angle rationality, and movement smoothness are defined. Mapping rules are established for each label dimension to map the evaluation score to the corresponding label value. These rules can be based on expert knowledge, experimental data, or other reliable sources. For example, the evaluation score range can be divided into different ranges, each corresponding to a label, or a curve function can be used for mapping. For each posture behavior evaluation score, the score is mapped to the corresponding label value according to the mapping rules for each dimension. For example, if the evaluation score is mapped to 0.7 for the stability dimension, 0.6 for the angle rationality dimension, and 0.8 for the movement smoothness dimension, according to the rules, these scores can be mapped to the labels "slightly unstable, angle not rational, movement smooth." The labels obtained from the multidimensional mapping are combined to form the final quantitative label. In this embodiment, the quantitative label for the posture behavior can be "slightly unstable, angle not rational, movement smooth." Specifically, three, five, or seven levels of mapping values ​​can be specified for the following labels: "Rotational State Label, Tilt Angle Label, Movement Fluency Label, Angle Reasonableness Label, and Stability Label." This effectively distinguishes movement levels, such as angle ranges. This multi-dimensional label mapping method allows assessment scores to be mapped to labels across multiple dimensions simultaneously, providing a richer and more accurate assessment of a player's posture. These quantitative labels not only help players comprehensively understand their performance but also provide more specific guidance to improve various aspects of posture, effectively identifying and addressing issues. Assigning quantitative labels to player posture is intended to assist and optimize the posture of disc players. By assigning quantitative labels to each posture behavior, the system can more accurately assess and analyze the quality of a player's movements. These quantitative labels not only provide an objective, quantitative evaluation of posture, but also provide players with targeted feedback and guidance, helping them better understand and improve their technique. Therefore, assigning quantitative labels to player posture is intended to assist and optimize the posture of disc players. By assigning quantitative labels to each gesture, the system can more accurately assess and analyze the quality of a player's movements. These labels not only provide an objective and quantitative evaluation of gestures, but also provide players with targeted feedback and guidance, helping them better understand and improve their movement techniques.

[0091] In actual training, the stability and consistency of movement are crucial for an athlete's development and improvement. A stable and consistent movement can help athletes maintain better performance during competition and help reduce the risk of injury. Therefore, a step will be introduced to assess movement consistency to better evaluate the quality of an athlete's movement. The specific steps are as follows:

[0092] The fused comprehensive evaluation results are extracted from each action. These evaluation results include multiple aspects of the action, such as stability, angle rationality, and movement smoothness.

[0093] The fused comprehensive evaluation results of each action are regarded as multidimensional feature vectors. The comprehensive evaluation results of two actions are compared using the Euclidean distance method to calculate the similarity between the two sets of data.

[0094] Set a threshold for similarity measurement, which is adjusted according to the actual situation. A similarity measurement above the threshold means that the two actions have a certain consistency in comprehensive evaluation.

[0095] For each pair of actions, the degree of consistency can be determined by comparing the similarity metric with the threshold. If the similarity metric is higher than the threshold, the two actions are considered to be consistent in terms of comprehensive evaluation; conversely, if the similarity metric is lower than the threshold, the actions may be somewhat different.

[0096] Similarly, the player's motion data can also be compared with pre-defined standard movements to detect unqualified areas. Specifically, a standard movement data set is prepared, which contains the comprehensive evaluation results of multiple standard movements after integration. The player's motion data is compared with each standard movement, and a similarity metric is used to calculate the similarity between the player's motion and each standard movement. In one or more embodiments, the similarity between the movements is calculated using Euclidean distance. Finally, the results of the similarity metric are compared with a pre-set similarity threshold to identify unqualified areas compared to the standard movement. At the same time, it can also help players learn and improve professional throwing techniques. By evaluating the stability of the movement, the rationality of the angle, the smoothness of the movement, etc., athletes can gain a deeper understanding of their own movement performance, and comparing it with the standard movement can help athletes identify the gaps compared with professional skills. Based on the comprehensive evaluation results, athletes can make self-adjustments and try different postures, angles, and movement processes to find the throwing method that best suits them.

[0097] In actual operation, especially in outdoor competitive sports, there is often the influence of natural wind, which may affect the evaluation results. Therefore, this application also proposes a method for evaluating the sports posture of a Frisbee competitive player taking into account the influence of wind. Specifically, the method includes the following steps:

[0098] S21, obtaining wind speed and direction data, wherein the data is obtained by installing wind speed and direction sensors around the site;

[0099] S22, processing the captured data using a wind impact model to obtain trajectory prediction data and attitude stability data;

[0100] S23, fusing the trajectory prediction data and posture stability data with the previously quantified data and sub-quantified posture data to generate a comprehensive player posture behavior assessment result;

[0101] S24, assigning a quantitative label value to the player's posture behavior based on the comprehensive evaluation results, thereby accurately evaluating the quality of the player's posture behavior.

[0102] Step S21, specifically, select a suitable location to install a wind speed and direction sensor to ensure that wind speed and wind direction can be accurately measured. In one or more embodiments, the sensor is installed at the edge or corner of the arena, and the wind speed and direction sensor is correctly connected to the data acquisition system to ensure that the sensor and the data acquisition device communicate normally. The data acquisition system is set to obtain the wind speed and wind direction data measured by the sensor in real time, including the numerical value of the wind speed (m / s) and the angle of the wind direction (relative to the north), and the collected wind speed and wind direction data is stored in a computer or data server for subsequent processing and analysis.

[0103] Step S22: Process the captured data using a wind impact model to decompose the wind's effect on the disc into vertical (lift) and horizontal (side force) forces. Specifically, gravity F = m*g, lift L = 0.5*ρ*A*Cl*V 2 , ρ is the air density, A is the reference area of ​​the disc, Cl is the lift coefficient, V is the relative wind speed; lateral force D = 0.5*ρ*A*Cd*V 2*sin(θ), where Cd is the side force coefficient, and θ is the angle of the disc relative to the wind direction. Combining these forces yields the disc's acceleration a = (lift - gravity) / m + (side force / m). Numerical integration is then used to determine the disc's velocity and position over time, namely V(t) and x(t). The lift coefficient and side force coefficient are calibrated based on the disc used and the wind conditions. This can be accomplished through experimentation and observation of actual flight data. The disc's motion stability also needs to be considered, specifically pitch stability, roll stability, and roll stability. The pitch angle is the angle of the disc relative to the horizontal plane, which affects its ascent and descent. A stable pitch angle indicates a smooth up-and-down motion during flight, without drastic rises or falls. This implementation calculates the rate of change of the disc's pitch angle; a smaller rate of change indicates better pitch stability. Roll stability involves the disc's spin, which is its rotation around its vertical axis. This spin can be estimated using angular velocity data. Stable spin means the disc's angular velocity remains stable without drastic changes. This implementation evaluates rotational stability by calculating the standard deviation or rate of change of angular velocity. Roll is the rotation of the disc around its axis of flight, which affects the disc's lateral stability. The assessment of roll stability can consider the disc's degree of stability during lateral motion. This implementation calculates the rate of change of the disc's displacement in the roll direction; a smaller rate of change indicates better roll stability. Ultimately, attitude stability assessment = w1 * pitch stability index + w2 * rotation stability index + w3 * roll stability index, where w1, w2, and w3 are the weights of the different indicators and can be adjusted based on actual needs and expert advice to obtain quantitative data.

[0104] Step S23, the trajectory prediction data and posture stability data are merged with the previous quantized data and sub-quantized posture data. In one or more embodiments, mutual information calculation is used to determine the degree of association between the data, and then the weights are set. Then, the trajectory prediction data, posture stability data, quantized data and sub-quantized posture data are weighted averaged according to the weights to generate a comprehensive player posture behavior evaluation result. The specific logic and processing method are the same as step 1051, which has been described in detail above and will not be repeated here. In general, this comprehensive result will take into account the various aspects of wind influence, quantized data and posture data, so as to more accurately evaluate the quality of the player's posture behavior.

[0105] In step S24, a quantitative label value is assigned to the player's posture behavior based on the comprehensive evaluation results, thereby accurately evaluating the quality of the player's posture behavior. This step is the same as the processing method of step 106, except that the data dimension will be higher, and multiple aspects such as wind impact, posture stability, quantitative data and sub-quantitative posture data will be comprehensively considered. Therefore, in the process of mapping labels, more attention should be paid to reliable resources such as test data and expert opinions.

[0106] The foregoing description is merely a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the form disclosed herein and should not be construed as excluding other embodiments. Rather, the present invention can be used in various other combinations, modifications, and environments and can be modified within the scope of the concept described herein through the above teachings or techniques or knowledge in the relevant field. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention are intended to be protected by the appended claims.

Claims

1. A method for correcting the athletic posture of a Frisbee athlete, characterized in that: The method is implemented through a posture model and includes the following steps: obtaining acceleration and angular velocity data of a player's flying disc throwing posture, wherein the data is captured by an inertial measurement unit installed on the flying disc; performing information screening on the captured posture data, and classifying the screened data to generate quantitative data and non-quantitative data; converting the non-quantitative data into sub-quantitative posture data through a model; fusing the quantitative data and the sub-quantitative posture data to generate a comprehensive evaluation result of the player's posture behavior, and assigning a quantitative label value to the player's posture behavior, thereby accurately evaluating the player's posture behavior; converting the non-quantitative data into sub-quantitative posture data through the model is performed through a flying disc rotation model, a tilt angle model, and a flight direction model; performing information screening on the captured posture data, and classifying the screened data to generate quantitative data and non-quantitative data is performed by calculating the tilt angle and the rotation angular velocity to classify the quantitative data and the non-quantitative data.

2. A method for correcting the athletic posture of a Frisbee athlete according to claim 1, characterized in that: The method of classifying the quantitative data and the non-quantified data by calculating the tilt angle and the rotation angular velocity comprises the following steps: S1, obtains the rotation angular velocity and tilt angle characteristic values ​​from the information screening process; S2, set the corresponding quantitative label and threshold for each feature; S3, classify the data of each frame by comparing the feature value with the threshold.

3. A method for correcting the athletic posture of a Frisbee competitive player according to claim 2, characterized in that: The S3, performing the following classification operation on the data of each frame, includes the following steps: S31, calculating the absolute value of the rotation angular velocity and comparing it with the threshold value to classify the rotation angular velocity features; S32, calculate the absolute value of the tilt angle and compare it with the threshold to classify the tilt angle feature.

4. The method for correcting the athletic posture of a Frisbee athlete according to claim 1, wherein: The process of fusing the quantitative data and the sub-quantified posture data to generate a comprehensive evaluation result of the player's posture behavior and assigning a quantitative label value to the player's posture behavior includes the following steps: 1051. Fusing quantized data and sub-quantized posture data; 1052. Generate comprehensive evaluation results of the players’ posture and behavior; 106. Assign quantitative label values ​​to the players' gesture behaviors.

5. The method for correcting the athletic posture of a Frisbee athlete according to claim 4, characterized in that: The fusion of the quantized data and the sub-quantized posture data includes the following steps: 10511, standardize quantized and subquantized posture data; 10512 , calculate the mutual information between quantized data and sub-quantized posture data; 10513, calculating the phase relationship between posture data based on mutual information; 10514, performing weighted averaging on the normalized quantized data and the sub-quantized posture data according to the mutual information correlation to obtain the fused data.

6. The method for correcting the athletic posture of a Frisbee athlete according to claim 5, characterized in that: The standardization of the quantized and sub-quantized posture data comprises the following steps: i. Calculate the mean and standard deviation of the quantitative data and sub-quantized posture data; ii, perform standardized calculations; iii. Repeat steps i and ii for each point.

7. A method for correcting the athletic posture of a Frisbee athlete according to claim 6, characterized in that: The standardized calculation is performed using the following formula: , where z is the standardized data point, μ is the mean, and σ is the standard deviation.

8. A method for evaluating the athletic posture of a Frisbee player taking into account wind influence, specifically comprising the following steps: S21, obtaining wind speed and direction data, wherein the data is obtained by installing wind speed and direction sensors around the site; S22, processing the captured data using a wind impact model to obtain trajectory prediction data and attitude stability data; S23, fusing the trajectory prediction data and posture stability data with the quantified data and sub-quantized posture data described in claim 1 to generate a comprehensive player posture behavior assessment result; S24, assigning a quantitative label value to the player's posture behavior based on the comprehensive evaluation results, thereby accurately evaluating the quality of the player's posture behavior.

Images