A visual detection-based ball hitting posture adaptive feedback method

Abstract

This invention relates to the field of image recognition technology, and more particularly to a visual detection-based adaptive feedback method for hitting posture. The method includes: acquiring a multi-view image sequence of the user's hitting process; preprocessing the received image sequence to generate an image coordinate sequence containing key point information; constructing three-dimensional coordinates based on the image coordinate sequence to calculate evaluation parameters such as the instantaneous power output efficiency of the hitting motion and comparing them with preset thresholds in a database to generate corresponding processing decisions; and adjusting the parameters of corresponding modules based on the received processing decisions. This invention improves the user's instantaneous power output efficiency of the hitting motion and the flexibility of adjustment.

Description

Technical Field

[0001] This invention relates to the field of image recognition technology, and in particular to an adaptive feedback method for hitting posture based on visual detection. Background Technology

[0002] In the fields of competitive sports training and mass fitness, scientifically quantifying and promptly correcting technical movements is crucial for improving athletic performance and preventing sports injuries. Most existing technologies assess hitting posture based on wearing measurement unit sensors and monocular cameras, but they do not consider the impact of timely feedback and spatial relative relationships on hitting posture. In the case of hitting posture involving a swing motion, reasonable adjustments to the user's posture from macroscopic to microscopic muscle coordination can significantly improve the precision analysis and output efficiency of hitting posture. Therefore, how to improve the instantaneous power output efficiency of hitting and overcome single adjustment are issues of great concern to those skilled in the art.

[0003] Chinese Patent Publication No. CN119896847A discloses a vision-based billiards auxiliary training system and method. The method includes: when the cue direction and the shot line are detected to coincide, recording a standard movement and extracting and storing key points; the trainee aligning key points with the standard position projected on the table; repeating the cue stroke after posture correction; and extracting the cue trajectory through image analysis and comparing it with the shot line to determine the positional deviation of the shot. It is evident that while the proposed solution can achieve positional correction, it still suffers from the following problems: it cannot flexibly adjust the trainee based on actual positional deviations and micro-muscle control, easily leading to movement assistance deviations and poor instantaneous power output efficiency during the shot. Summary of the Invention

[0004] To address this issue, the present invention provides a visual detection-based adaptive feedback method for striking posture, which overcomes the problem in existing technologies where the inability to flexibly adjust the posture according to the user's actual situation easily leads to a reduction in the instantaneous power output efficiency during striking.

[0005] To achieve the above objectives, the present invention provides a visual detection-based adaptive feedback method for hitting posture, comprising: Extract the three-dimensional coordinate sequence of key points in the user's hitting action; The three-dimensional coordinate sequence is preprocessed and dynamic data is constructed; The instantaneous power output efficiency at the moment of impact is calculated based on the preprocessed dynamic data. The power efficiency for a single impact action is then determined based on the calculated instantaneous power output efficiency. Whether to generate a corresponding processing decision is determined based on the degree of power transmission coupling. The decision to determine the trunk force line preparation angle is based on the shoulder-hip distance on the non-racket side and the hip axis leading angular velocity. Treatment decisions for determining elbow extension synchronization rate based on the relative lateral deviation angle of the humerus and the elbow flexion-extension stiffness coefficient; After adjustment, the accuracy of the visual acquisition module is verified based on the average confidence level of the key points; Continuous testing is conducted after confirming that the dynamic performance of each individual shot meets the standards.

[0006] Furthermore, for operating conditions where the instantaneous power output efficiency at impact is less than or equal to the preset instantaneous power output efficiency at impact, a decision is made on the preparatory angle of the torso force line based on the power chain transmission coupling degree as the judgment criterion. For operating conditions where the instantaneous power output efficiency at the moment of impact exceeds the preset instantaneous power output efficiency at the moment of impact, the power efficiency for a single impact action is determined to meet the standard and continuously monitored.

[0007] Furthermore, in cases where the power chain transmission coupling degree is greater than the preset power chain transmission coupling degree, it is determined that the contribution of torsional energy is insufficient. When the kinetic chain transmission coupling degree is less than or equal to the preset kinetic chain transmission coupling degree, it is determined that the kinetic energy conversion efficiency of the upper limb swing is low.

[0008] Furthermore, when the contribution of trunk torsional energy is insufficient, the trunk force line preparation angle is adjusted based on the shoulder-hip distance on the non-racket side, and the shoulder-hip distance on the non-racket side is positively correlated with the trunk force line preparation angle.

[0009] Furthermore, a feedback correction instruction is generated to determine that the preparatory angle of the trunk force line should be increased based on the hip axis leading angular velocity, and that the hip axis leading angular velocity and the preparatory angle of the trunk force line are positively correlated.

[0010] Furthermore, in response to the first repeated detection condition, it was determined that the energy conversion efficiency of the upper limb swing was low.

[0011] Furthermore, when the energy conversion efficiency of upper limb swing is low, the elbow joint extension synchronization rate is adjusted based on the relative lateral deviation angle of the humerus.

[0012] Furthermore, a feedback correction instruction is generated to determine that the adjusted elbow joint extension synchronization rate should be increased based on the elbow joint flexion-extension stiffness coefficient, and that the elbow joint flexion-extension stiffness coefficient and the elbow joint extension synchronization rate are negatively correlated.

[0013] Furthermore, in response to the second repeated detection condition, the accuracy verification result of the visual acquisition module is determined based on the average confidence level of the key points.

[0014] In addition, the present invention also provides a system for the aforementioned vision-based adaptive feedback method for hitting posture, comprising: The visual acquisition module is used to acquire a multi-frame image sequence containing the user's hitting action; The image processing module is used to preprocess the image sequence, including the identification and extraction of user key points and racket feature points, and outputs a pixel coordinate sequence containing key point information. The data processing module is used to calculate the hitting posture evaluation parameters based on the pixel coordinate sequence and output the three-dimensional motion trajectory data of the user's joints and racket; The data analysis module determines the processing decision for the power efficiency of a single hitting action based on the instantaneous power output efficiency at the moment of impact. The execution module is used to adjust the parameters of the corresponding module to the corresponding values ​​based on the received processing decisions.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention reflects the comprehensive efficiency of the kinetic chain of the hitting action through the instantaneous power output efficiency of the hitting action, reflects the degree of difference in energy transfer efficiency between trunk force generation and upper limb swing through the kinetic chain transmission coupling degree, and when it is determined that the instantaneous power output efficiency of the hitting action is lower than a preset threshold, it performs two-level progressive diagnosis and adjustment through the trunk force line preparation angle and elbow joint extension synchronization rate, so that the action feedback instructions are more in line with the user's technical stage, avoiding invalid or erroneous guidance caused by the use of fixed parameters, and thus enabling the training process to adaptively approach the user's current optimal action mode.

[0016] Furthermore, by setting the evaluation parameter of instantaneous power output efficiency upon hitting the ball, this invention comprehensively quantifies the torso torsional energy, upper limb swing energy, and hitting accuracy, avoiding the limitations of a single fixed threshold that leads to poor instantaneous power output efficiency upon hitting the ball. This makes the cause analysis more consistent with the user's actual scenario and further improves the flexibility of the data analysis module.

[0017] Furthermore, by calculating the coupling degree of the kinetic chain transmission and setting a preset value, this invention can quickly distinguish whether the root cause of the problem is insufficient contribution of torso torso twisting energy or low conversion efficiency of upper limb swing kinetic energy when the instantaneous power output efficiency of the ball does not meet the standard. This provides direction for subsequent precise adjustment and makes the diagnosis more accurate.

[0018] Furthermore, this invention addresses the core power generation issue from a spatial geometric perspective by adjusting the preparatory angle of the trunk force line based on the shoulder-hip distance on the non-racket side. When the shoulder-hip distance on the non-racket side is less than a preset threshold, the adjustment is increased, making the preparatory angle of the trunk force line more consistent with the actual situation and further improving the instantaneous power output efficiency of the shot.

[0019] Furthermore, after completing the initial adjustment of the trunk force line preparation angle, the present invention corrects the adjusted trunk force line preparation angle according to the hip axis leading angular velocity to ensure that the stored elastic potential energy is released efficiently, thereby increasing the starting efficiency of the kinetic chain and further improving the instantaneous power output efficiency of the ball.

[0020] Furthermore, in response to the first repeated detection condition, the present invention switches the cause, ensuring the rigor and adaptability of the diagnostic logic, avoiding the system from getting stuck in a loop on a single attribution path, thereby further improving the instantaneous power output efficiency of the ball.

[0021] Furthermore, this invention adjusts the elbow joint extension synchronization rate by adjusting the relative lateral deviation angle of the humerus, avoiding the problem of shoulder and elbow dislocation during power transmission in the upper limb, improving the joint coordination efficiency, and selecting an appropriate adjustment coefficient within the threshold range of the relative lateral deviation angle of the humerus to avoid the irrationality of linear adjustment, thereby further improving the elbow joint extension synchronization rate.

[0022] Furthermore, after completing the initial adjustment of the elbow joint extension synchronization rate, the present invention corrects the adjusted elbow joint extension synchronization rate according to the elbow joint flexion-extension stiffness coefficient, and intervenes at the muscle level, thereby making the elbow joint extension synchronization rate that conforms to the user's actual situation more in line with the standard, avoiding the internal braking effect caused by the dyssynergy of flexor and extensor muscles, and thus further improving the elbow joint extension synchronization rate.

[0023] Furthermore, in response to the second repeated detection condition, the present invention determines the cause of the error by the visual acquisition module rather than the user action by using the average confidence level of the visual key points. This avoids continuously giving incorrect processing decisions when the system itself is unreliable, ensuring the safety and effectiveness of training, further improving the integrity of the system's self-testing process, and thus further improving the instantaneous power output efficiency of the ball. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of the visual detection-based adaptive feedback method for hitting posture according to the present invention; Figure 2 This is a flowchart illustrating the present invention for determining whether the power efficiency of a single hitting action meets the standard based on the instantaneous power output efficiency at the moment of impact; Figure 3 This is a flowchart illustrating the process of determining the adjustment direction and adjustment procedure based on the coupling degree of the power chain in this invention. Figure 4 This is a module connection diagram of the visual detection-based adaptive feedback system for hitting posture according to the present invention. Detailed Implementation

[0025] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0026] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0027] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0028] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0029] It should be noted that the data in this embodiment are all derived from a comprehensive analysis and evaluation of historical test data and corresponding historical test results from the system described in this invention over the three months prior to this test. Before this test, the system described in this invention comprehensively determines the preset values ​​stored in the database based on the analysis results of 25,863 cumulative tests over the previous three months and the processing results after handling 19,584 specific cases. Those skilled in the art will understand that the system described in this invention can determine the above-mentioned parameters for a single item by selecting the value with the highest proportion based on the data distribution as the preset standard parameter, using weighted summation to obtain the value as the preset standard parameter, substituting each historical data point into a specific formula and using the value obtained by the formula as the preset standard parameter, or other selection methods, as long as the system described in this invention can clearly define different specific situations in the single-item judgment process through the obtained values.

[0030] Please see Figure 1 As shown, it is a module link diagram of the visual detection-based adaptive feedback system for hitting posture of the present invention, including: The visual acquisition module is used to acquire a multi-frame image sequence containing the user's hitting action; The image processing module is used to preprocess the image sequence, including the identification and extraction of user key points and racket feature points, and outputs a pixel coordinate sequence containing key point information. The data processing module is used to calculate the hitting posture evaluation parameters based on the pixel coordinate sequence and output the three-dimensional motion trajectory data of the user's joints and racket; The data analysis module determines the processing decision for the power efficiency of a single hitting action based on the instantaneous power output efficiency at the moment of impact. The execution module is used to adjust the parameters of the corresponding module to the corresponding values ​​based on the received processing decisions.

[0031] Specifically, the visual acquisition module described in this embodiment of the invention controls multiple cameras to ensure spatiotemporal consistency and transmits the acquired raw image sequence to the image processing module in real time.

[0032] Specifically, the image processing module of this embodiment of the invention performs the following preprocessing process on the received image sequence: lens distortion correction based on camera intrinsic parameters; scaling the image resolution to the model input size and normalizing the pixel values; outputting a heatmap containing user key points and a heatmap of the racket center point and orientation; determining the approximate location of a single key point by finding local maxima in the heatmap, and accurately locating it by combining offset prediction; and finally outputting a two-dimensional point pixel coordinate sequence.

[0033] Specifically, in this embodiment of the invention, the data processing module receives and identifies the pixel coordinate sequence output by the image processing module to reconstruct three-dimensional spatial coordinates, and performs reasonableness detection on the reconstruction results to remove and mark outliers; and calculates hitting posture evaluation parameters, including the instantaneous output efficiency of the hitting moment, based on key moments such as the lowest point of the backswing, the instant of hitting the ball, and the starting point of the swing.

[0034] Specifically, the data analysis module described in this embodiment of the invention compares the real-time calculated ball-hitting posture evaluation parameters with a preset threshold to generate a diagnostic result and selects the corresponding processing decision to perform a diagnosis on whether the power efficiency of a single ball-hitting action meets the standard.

[0035] Specifically, the system described in this invention stores and manages the received processing results and constructs a database that can be continuously updated.

[0036] Specifically, in this embodiment of the invention, the visual acquisition module acquires multi-view image sequences of the user's hitting process using multiple synchronous cameras; the image processing module preprocesses the received image sequences to generate image coordinate sequences containing key point information; the data analysis module constructs three-dimensional coordinates based on the image coordinate sequences to calculate evaluation parameters such as the instantaneous power output efficiency of the hitting process and compares them with preset thresholds in the database to generate corresponding processing decisions; and the execution module adjusts the parameters of the corresponding module based on the received processing decisions.

[0037] Please see Figure 2 As shown, this is a flowchart illustrating the process of determining whether the power efficiency of a single hitting action meets the standard based on the instantaneous power output efficiency of the ball, as described in this invention. The process includes: Extract the three-dimensional coordinate sequence of key points in the user's hitting action; The three-dimensional coordinate sequence is preprocessed and dynamic data is constructed; The instantaneous power output efficiency at the moment of impact is calculated based on the preprocessed dynamic data. The power efficiency for a single impact action is then determined based on the calculated instantaneous power output efficiency. Whether to generate a corresponding processing decision is determined based on the degree of power transmission coupling. The decision to determine the trunk force line preparation angle is based on the shoulder-hip distance on the non-racket side and the hip axis leading angular velocity. Treatment decisions for determining elbow extension synchronization rate based on the relative lateral deviation angle of the humerus and the elbow flexion-extension stiffness coefficient; After adjustment, the accuracy of the visual acquisition module is verified based on the average confidence level of the key points.

[0038] Continuous testing is conducted after confirming that the dynamic performance of each individual shot meets the standards.

[0039] In this embodiment of the invention, the user's hitting action is synchronously acquired through a vision system. Key points are extracted, a three-dimensional coordinate sequence is calculated, and temporal smoothing is performed. Based on the pre-processed dynamic data, the instantaneous power output efficiency of the hitting action is calculated and compared with a preset threshold to determine if the power efficiency of the hitting action meets the standard. If it does not meet the standard, the cause is distinguished by calculating the coupling degree of the kinetic chain transmission. The trunk force line preparation angle is adjusted by adjusting the shoulder-hip distance and hip axis lead angular velocity on the non-racket side, or the elbow extension synchronization rate is adjusted by adjusting the relative lateral deviation angle of the humerus and the elbow flexion-extension stiffness coefficient. When the action adjustment remains ineffective, the system automatically detects the average confidence level of the visual key points to determine if it is a problem with the accuracy of the vision acquisition module. For actions that meet the standard, their data will be continuously incorporated into the user profile for dynamic updating of evaluation values ​​and completion of database construction.

[0040] Please see Figure 3As shown, this is a flowchart illustrating the process of determining whether the power efficiency of a single hitting action meets the standard based on the instantaneous power output efficiency of the ball, as described in this invention. The process includes: Specifically, for operating conditions where the instantaneous power output efficiency at the moment of impact is less than or equal to the preset instantaneous power output efficiency at the moment of impact, the processing decision is made based on the coupling degree of the power chain transmission and the preparatory angle of the torso force line. For operating conditions where the instantaneous power output efficiency at the moment of impact exceeds the preset instantaneous power output efficiency at the moment of impact, the power efficiency for a single impact action is determined to meet the standard and continuously monitored.

[0041] Specifically, the data processing module described in this embodiment of the invention includes: The power determination submodule calculates the instantaneous power output efficiency at the moment of impact by using torso torsional energy, upper limb swing output energy, and the precision penalty coefficient at the point of impact. This efficiency is used to determine the energy conversion efficiency transferred from the body to the racket at the moment of impact. Higher power output efficiency results in higher energy transfer efficiency, better coordinated force application, and less energy loss; conversely, lower power output efficiency leads to lower energy transfer efficiency, resulting in issues such as disjointed force application, muscle compensation, and greater energy loss.

[0042] In this embodiment, the instantaneous power output efficiency upon impact is a comprehensive evaluation value calculated based on the torso torso torso rotation energy, upper limb swing output energy, and the accuracy penalty coefficient at the point of impact, wherein: The calculation process for the torsional energy Et of the torso includes: The time period from the lowest point of the backswing to the instant of impact is recorded as the detection period; Calculate the change in the relative shoulder-hip torsion angle during the testing period; The ratio of the change in the relative shoulder-hip torsional angle to the detection period is calculated and denoted as the average angular velocity of the trunk.

[0043] The calculation process for the upper limb swing output energy Es includes: Position the coordinates of the center point of the wrist joint at the moment of impact; The wrist linear velocity is calculated based on the coordinate difference of the center point; Calculate the product of half the user's body mass and the square of the wrist's linear velocity, and denote it as the upper limb swing output energy.

[0044] The calculation process for the accuracy penalty coefficient at the point of impact includes: Calculate the projected distance between the sweet spot of the racket and the center of the ball on the horizontal plane at the moment of impact, and record it as the horizontal offset distance; Take the absolute value of the horizontal offset distance, add 1, and take the reciprocal of the sum as the penalty coefficient; After logarithmically converting the torso torsional energy and the upper limb swing output energy, the sum is weighted and summed with the hitting point precision penalty coefficient, and recorded as the instantaneous dynamic output energy at the moment of impact.

[0045] It is understood that the instantaneous power output efficiency of the ball strike described in the embodiments of the present invention is only used to illustrate the technical principles and feasibility of the present invention. When implementing the present invention, those skilled in the art can make adaptive adjustments to the above parameters or adopt other equivalent digital transformations.

[0046] Specifically, when the power chain transmission coupling degree is greater than the preset power chain transmission coupling degree, it is determined that the trunk torsional energy contribution is insufficient. When the kinetic chain transmission coupling degree is less than or equal to the preset kinetic chain transmission coupling degree, it is determined that the kinetic energy conversion efficiency of the upper limb swing is low.

[0047] Specifically, the data processing module described in this embodiment of the invention includes: The kinetic chain transmission determination submodule calculates the kinetic chain transmission coupling degree using the torso torsional energy and upper limb swing output energy, thereby determining the energy transfer efficiency between the torso and upper limbs. A higher kinetic chain transmission coupling degree results in more consistent force application, less energy loss, and higher power transfer efficiency; conversely, a lower coupling degree makes it easier for force to disengage and leads to greater energy loss.

[0048] Specifically, the calculation process of the power chain transmission coupling degree in the embodiments of the present invention includes: The trunk torsional energy and upper limb swing output energy obtained through real-time calculation are input into the database for synchronous storage; The data processing module calculates the ratio of the upper limb swing output energy to the trunk torsional energy of a single stroke in a single absorbing action, and obtains the actual kinetic chain transmission coupling degree. Specifically, the baseline value of the kinetic chain transmission coupling degree is based on the user's historical performance data. The system continuously records the kinetic chain transmission coupling degree values ​​corresponding to all shots that are judged to meet the standard within the same statistical period. After excluding obvious outliers from the set of values, the statistical median is calculated and set as the baseline value of the kinetic chain transmission coupling degree for the user's current stage.

[0049] It is understood that the power chain transmission coupling degree described in the embodiments of the present invention does not participate in the intermediate judgment of whether it meets the standard, but is only used to trace the reasons for not meeting the power efficiency standard of a single shot.

[0050] It is understood that the power chain transmission coupling degree described in the embodiments of the present invention is only used to illustrate the technical principles and feasibility of the present invention. When implementing the present invention, those skilled in the art can make adaptive adjustments to the above parameters or adopt other equivalent digital transformations.

[0051] Specifically, when the contribution of trunk torsional energy is insufficient, an adjustment command for the trunk force line preparation angle is generated based on the shoulder-hip distance on the non-racket side, and the shoulder-hip distance on the non-racket side is positively correlated with the trunk force line preparation angle.

[0052] Specifically, the data processing module described in this embodiment of the invention includes: The shoulder-hip distance determination submodule calculates the shoulder-hip distance on the non-racket side using the three-dimensional coordinates of the shoulder joint on the non-racket side and the hip joint on the same side, in order to determine the preparatory range for trunk lateral flexion and rotation.

[0053] Specifically, the process of obtaining the shoulder-hip distance on the non-racket side as described in this embodiment of the invention includes: The minimum velocity point of the elbow joint of the holding hand in space is detected to determine the moment when the backswing amplitude reaches its maximum. This process is called the positioning of the frame with the maximum backswing amplitude. The maximum amplitude frame is input to the image processing module to generate the maximum amplitude frame image. In this frame image, the three-dimensional coordinates of the shoulder joint on the non-holding side and the hip joint on the same side are obtained and projected onto the horizontal ground. The projection length of the line connecting the two points of the three-dimensional coordinate system in the direction of the hitting target is measured. This length is the shoulder-hip distance L on the non-racket side.

[0054] The process of obtaining the preparatory angle of the trunk force line according to the embodiments of the present invention includes: The transition point from the backswing to the forward swing is determined by the instant when the torso twist angular velocity changes from negative to positive. This process is the positioning of the swing initiation frame. The startup frame is input to the image processing module to generate a startup frame image. In this frame image, the projection line of the line connecting the two shoulder joints on the horizontal plane is calculated, which is the trunk force line. The acute angle between the projection line of the line connecting the two hip joints on the horizontal plane is recorded as the trunk force line preparation angle.

[0055] Understandably, the trunk force line preparatory angle represents the degree of torsional separation between the shoulder axis and the hip axis before the start of force exertion. The larger the trunk force line preparatory angle, the more efficient the transmission of power from the lower limbs to the upper limbs, which is a key mechanical posture. If the trunk force line preparatory angle is too small, it indicates that the hips are thrusting or the shoulders are rotated excessively, resulting in power leakage.

[0056] It is understood that in the embodiments of the present invention, during the backswing phase, increasing the shoulder-hip distance on the non-holding side forces the torso to perform a greater lateral twist and stretch, thereby increasing the torso force line preparation angle accordingly. Therefore, the increase in the torso force line preparation angle is proportional to the shoulder-hip distance on the non-holding side.

[0057] Specifically, the system generates voice adjustment commands based on the actual shoulder-hip distance on the non-racket side to prompt the user's hitting posture.

[0058] It is understood that those skilled in the art can select the threshold corresponding to the non-holding side shoulder-hip distance and trunk force line preparation angle based on the correlation coefficient of the non-holding side shoulder-hip distance and trunk force line preparation angle in the historical data of different monitoring periods. Any method can be used to adjust and optimize the correlation threshold, all of which are equivalent embodiments of the present invention.

[0059] Specifically, a feedback correction instruction is generated to increase the preparatory angle of the trunk force line based on the hip axis leading angular velocity, and the hip axis leading angular velocity and the preparatory angle of the trunk force line are positively correlated.

[0060] Specifically, the data processing module described in this embodiment of the invention includes: The hip axis lead angle determination submodule calculates the hip axis lead angular velocity by using the coordinates of the left and right hip joints in three-dimensional space, in order to determine the force sequence of the lower limb driving the upper limb.

[0061] Specifically, the process of obtaining the hip axis leading angular velocity according to the embodiments of the present invention includes: Based on the user key points identified by the image processing module, the coordinates of the left and right hip joints in three-dimensional space are obtained, and the hip axis is the virtual axis connecting the left and right hip joints. The moment when the preparatory angle of the trunk force line reaches its maximum value is recorded as the starting time t0; Starting from the initial time t0, a fixed time interval is extended backward, and the end time of this interval is denoted as t1. The interval [t0, t1] is the time window for measuring the hip axis initiation speed. Within the time window, multiple frames of image data are extracted, and the pointing angle of the hip axis on the horizontal plane in each frame is calculated. The difference between the hip axis pointing angle at the start time t0 and the hip axis pointing angle at the end time t1 is calculated. This difference is the angular displacement of the hip axis around the vertical axis of the body during the window period. The ratio of the angular displacement to the time window is calculated and denoted as the hip axis leading angular velocity; the hip axis leading angular velocity characterizes the initial speed at which the hip begins to drive the body to rotate forward.

[0062] Understandably, the greater the hip axis leading angular velocity, the faster the start-up and the more in line with the kinetic chain sequence. The smaller the hip axis leading angular velocity, the slower the start-up after accumulating power, resulting in loss of power in the early stage of power transmission. Therefore, it is necessary to increase the hip axis leading angular velocity to ensure that it matches the adjusted trunk force line preparatory angle.

[0063] It is understood that the assignment of the hip axis leading angular velocity in the embodiments of the present invention is only used to illustrate the technical principle and feasibility of the present invention. When implementing the present invention, those skilled in the art can make adaptive adjustments to the above parameters or use other equivalent digital transformations.

[0064] Specifically, the system generates voice adjustment commands based on the actual hip axis leading angular velocity to prompt the user's hitting posture.

[0065] Specifically, in response to the first repeated detection condition, it was determined that the energy conversion efficiency of the upper limb swing was low.

[0066] It is understandable that the first repeated detection condition is that after the data analysis module completes the adjustment for insufficient contribution of torso torso torso energy, the recalculated instantaneous power output efficiency of the ball still does not meet the standard.

[0067] Specifically, if the power transmission coupling degree is greater than a preset value, the above-mentioned processing decision for repeated adjustment in the case of insufficient torsional energy contribution is issued. If the power transmission coupling degree is less than or equal to the preset value, and the preset angle of the trunk force line and related parameters are all within the normal range after multiple tests, it is determined that there is no abnormality in the trunk link, and the problem may be due to the low energy conversion efficiency of the upper limb swing.

[0068] Specifically, to rule out misjudgments, the dynamic parameters of the upper limb joints were further calculated: The ratio of the shoulder joint rotational angular velocity to the elbow joint extension angular velocity during the acceleration phase is denoted as the shoulder-elbow velocity ratio. Calculate the angle between the line connecting the center point of the elbow joint and the sweet spot of the racket at the moment of impact and the direction of the racket velocity, and record it as the impact hysteresis angle; If the shoulder-elbow speed ratio is lower than the preset shoulder-elbow speed ratio, or if the ball-hitting lag angle is greater than the preset ball-hitting lag angle, the reason is determined to be low kinetic energy conversion efficiency of the upper limb swing.

[0069] It is understood that those skilled in the art can select the threshold corresponding to the shoulder-elbow speed ratio and the hitting lag angle based on the historical data of different monitoring users. Any method can be used to adjust and optimize the correlation threshold, and all of these are equivalent embodiments of the present invention.

[0070] Please see Figure 4 As shown, it is a flowchart of the process of determining the adjustment direction and adjustment process based on the coupling degree of the power chain according to the present invention, and the process includes: Specifically, when the energy conversion efficiency of the upper limb swing is low, the elbow joint extension synchronization rate is adjusted based on the relative lateral deviation angle of the humerus.

[0071] Specifically, the data processing module described in this embodiment of the invention includes: The humeral angle determination submodule calculates the relative lateral deviation angle of the humerus using the local coordinate system of the torso to determine the lateral lifting range of the upper arm. The larger the relative lateral deviation angle of the humerus, the higher the elbow is raised, which makes it easier for the force line to deviate and cause shoulder injury; the smaller the relative lateral deviation angle of the humerus, the smaller the angle between the upper arm and the body, and the more limited the swing trajectory.

[0072] Specifically, the process of obtaining the relative lateral deviation angle of the humerus in the embodiments of the present invention includes: A local coordinate system for the torso is established. During the racket-swing acceleration phase, the data processing module establishes the three-dimensional coordinates of the center of the shoulder joint and the center of the elbow joint on the racket-holding side, and calculates the vector pointing from the center of the shoulder joint to the center of the elbow joint, denoted as the humeral unit vector at that moment. Based on key torso markers, the vertical coordinates of the torso's central plane are constructed, denoted as the torso's sagittal plane normal vector. The spatial angle is calculated using the inverse cosine of the dot product of the humeral unit vector and the torso's sagittal plane normal vector, denoted as the basic spatial angle. The difference between 90° and the basic spatial angle is calculated and denoted as the relative lateral deviation angle of the humerus.

[0073] Specifically, the data processing module described in this embodiment of the invention includes: The elbow joint determination submodule calculates the elbow joint extension synchronization rate using the shoulder joint internal rotation angular velocity sequence and the elbow joint extension angular velocity sequence to determine the temporal coordination accuracy of the extension. A higher elbow joint extension synchronization rate results in higher energy transfer efficiency and more fluid movements; a lower elbow joint extension synchronization rate results in lower energy transfer efficiency and a greater likelihood of temporal misalignment.

[0074] Specifically, the process of obtaining the elbow joint extension synchronization rate according to the embodiments of the present invention includes: Two time series were simultaneously acquired at a frequency of 100Hz during the same acceleration phase: the shoulder joint internal rotation angular velocity sequence and the elbow joint extension angular velocity sequence. The covariance of the two angular velocity sequences was calculated and divided by the product of their respective standard deviations. The range of the calculated result ρ is [-1, 1], which is denoted as the elbow joint extension synchronization rate. The closer this value is to 1, the better the coordination between shoulder internal rotation and elbow extension.

[0075] It is understood that the relative lateral deviation angle of the humerus and the elbow joint extension synchronization rate described in the embodiments of the present invention have a non-linear negative correlation adjustment relationship. If the relative lateral deviation angle of the humerus is smaller, the inner arm of the shoulder joint internal rotation is too short, the adduction state is more obvious, and it cannot provide initial kinetic energy.

[0076] It is understood that those skilled in the art can make adaptive adjustments to the above parameters or use other equivalent digital transformations when implementing the present invention.

[0077] Specifically, a feedback correction instruction is generated to determine that the adjusted elbow joint extension synchronization rate should be increased based on the elbow joint flexion-extension stiffness coefficient, and the elbow joint flexion-extension stiffness coefficient and the elbow joint extension synchronization rate are negatively correlated.

[0078] Specifically, the data processing module described in this embodiment of the invention includes: The elbow joint determination submodule uses electromyographic signals from the flexor and extensor muscles to determine the elbow joint flexion-extension stiffness coefficient, thus identifying whether the elbow joint is excessively locked or relaxed at the moment of impact. A larger elbow joint flexion-extension stiffness coefficient indicates greater impact strength and power; conversely, a smaller coefficient results in less impact strength, weaker power, insufficient support at the moment of impact, and power loss.

[0079] Specifically, the process of obtaining the elbow joint flexion-extension stiffness coefficient according to the embodiments of the present invention includes: Acquire electromyographic signals of flexor and extensor muscles during the acceleration phase of the swing; calculate the root mean square (RMS) values ​​of the electromyographic signals of flexor and extensor muscles respectively, and calculate the ratio of the RMS value of the flexor signal to the RMS value of the extensor signal, which is denoted as the elbow joint flexion-extension stiffness coefficient.

[0080] It is understood that the larger the elbow joint flexion-extension stiffness coefficient described in the embodiments of the present invention, the higher the flexion-extension strength, and the more it restricts the explosive extension of the elbow joint. Therefore, it is necessary to select an appropriate coefficient to reduce the elbow joint flexion-extension stiffness coefficient in order to ensure an increase in the elbow joint extension synchronization rate.

[0081] It is understood that those skilled in the art can make adaptive adjustments to the above parameters or use other equivalent digital transformations when implementing the present invention.

[0082] Specifically, in response to the second repeated detection condition, the accuracy verification of the visual acquisition module is determined based on the average confidence of key points.

[0083] Understandably, after the data analysis module completed the adjustment for the low energy conversion efficiency of the upper limb swing, the recalculated instantaneous power output efficiency of the ball still did not meet the standard.

[0084] Specifically, the data processing module described in this embodiment of the invention includes: The confidence level determination submodule calculates the average confidence level of keypoints identified within an image frame to determine the reliability of the visual algorithm in recognizing body joint coordinates. A higher average confidence level indicates higher recognition accuracy, less illumination and occlusion, and higher data reliability; conversely, a lower average confidence level indicates lower recognition accuracy, the presence of occlusion or insufficient lighting, requiring recalibration, and lower data reliability.

[0085] Specifically, the process for determining the accuracy verification of the visual acquisition module as described in this embodiment of the invention includes: The average confidence score of key points identified in all image frames during a single shot is calculated and denoted as the average confidence score of visual key points. If the average confidence level is less than or equal to the preset average confidence level, the data analysis module determines the deviation of the visual acquisition module, issues a command to pause all current model-based fine adjustment, and generates a prompt on the user interface to guide the user into the corresponding equipment calibration or inspection process. If the average confidence level is less than or equal to the preset average confidence level, the data analysis module determines that the visual acquisition module is operating normally and issues a processing decision to return to the upper layer for repeated adjustment.

[0086] It is understood that those skilled in the art can make adaptive adjustments to the above parameters or use other equivalent digital transformations when implementing the present invention.

[0087] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0088] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A visual detection based swing posture adaptive feedback method, characterized in that, include: Extract the three-dimensional coordinate sequence of key points in the user's hitting action; The three-dimensional coordinate sequence is preprocessed and dynamic data is constructed; The instantaneous power output efficiency at the moment of impact is calculated based on the preprocessed dynamic data. The power efficiency for a single impact action is then determined based on the calculated instantaneous power output efficiency. Whether to generate a corresponding processing decision is determined based on the degree of power transmission coupling. The decision to determine the trunk force line preparation angle is based on the shoulder-hip distance on the non-racket side and the hip axis leading angular velocity. Treatment decisions for determining elbow extension synchronization rate based on the relative lateral deviation angle of the humerus and the elbow flexion-extension stiffness coefficient; After adjustment, the accuracy of the visual acquisition module is verified based on the average confidence level of the key points; Continuous testing is conducted after confirming that the dynamic performance of each individual shot meets the standards.

2. The visual detection based swing pose adaptive feedback method according to claim 1, wherein, For operating conditions where the instantaneous power output efficiency at impact is less than or equal to the preset instantaneous power output efficiency at impact, the processing decision is made based on the power chain transmission coupling degree as the judgment criterion for the preparatory angle of the torso force line. For operating conditions where the instantaneous power output efficiency at the moment of impact exceeds the preset instantaneous power output efficiency at the moment of impact, the power efficiency for a single impact action is determined to meet the standard and continuously monitored.

3. The visual detection based swing pose adaptive feedback method according to claim 1 or 2, characterized in that, When the power chain transmission coupling degree is greater than the preset power chain transmission coupling degree, it is determined that the contribution of torsional energy of the torso is insufficient. When the kinetic chain transmission coupling degree is less than or equal to the preset kinetic chain transmission coupling degree, it is determined that the kinetic energy conversion efficiency of the upper limb swing is low.

4. The visual detection based swing pose adaptive feedback method according to claim 3, wherein, When the contribution of trunk torsional energy is insufficient, an adjustment command for the trunk force line preparation angle is generated based on the shoulder-hip distance on the non-racket side, and the shoulder-hip distance on the non-racket side is positively correlated with the trunk force line preparation angle.

5. The visual detection based swing pose adaptive feedback method according to claim 3 or 4, characterized in that, Generate feedback correction instructions to increase the prepared angle of the trunk force line based on the hip axis leading angular velocity, and ensure that the hip axis leading angular velocity and the prepared angle of the trunk force line are positively correlated.

6. The visual detection-based adaptive feedback method for hitting posture according to claim 5, characterized in that, In response to the first repeated detection condition, it was determined that the energy conversion efficiency of the upper limb swing was low.

7. The visual detection based swing pose adaptive feedback method according to claim 3 or 6, characterized in that, When the energy conversion efficiency of upper limb swing is low, adjustment instructions for elbow joint extension synchronization rate are generated based on the relative lateral deviation angle of the humerus.

8. The visual detection based swing pose adaptive feedback method according to claim 3 or 6, characterized in that, A feedback correction instruction is generated, which determines that the adjusted elbow joint extension synchronization rate should be increased based on the elbow joint flexion-extension stiffness coefficient, and that the elbow joint flexion-extension stiffness coefficient and the elbow joint extension synchronization rate are negatively correlated.

9. The visual detection based swing pose adaptive feedback method according to claim 1, wherein, In response to the second repeated detection condition, the accuracy verification of the visual acquisition module is determined based on the average confidence of key points.

10. A system applying the visual detection based ball striking posture adaptive feedback method according to any one of claims 1 to 9, characterized in that, include, The visual acquisition module is used to acquire a multi-frame image sequence containing the user's hitting action; The image processing module is used to preprocess the image sequence, including the identification and extraction of user key points and racket feature points, and outputs a pixel coordinate sequence containing key point information. The data processing module is used to calculate the hitting posture evaluation parameters based on the pixel coordinate sequence and output the three-dimensional motion trajectory data of the user's joints and racket; The data analysis module determines the processing decision for the power efficiency of a single hitting action based on the instantaneous power output efficiency at the moment of impact. The execution module is used to adjust the parameters of the corresponding module to the corresponding values ​​based on the received processing decisions.

Images