An intelligent online correction method and system for pull-ups

Abstract

The application provides a kind of intelligent online correction method and system for pull-up, it is related to data processing technical field, method includes: obtaining real-time pull-up video data to be corrected;Through gesture recognition algorithm, the skeletal point coordinate data in pull-up video data is extracted;Skeletal point coordinate data is handled by missing value complementation;According to the skeletal point coordinate data after complementation processing, joint state space model is established;According to joint state space model, the target angle is predicted in combination with Kalman filter algorithm and curve fitting algorithm;In the case where there is predicted target angle that does not conform to pull-up movement standard index, output correction prompt through preset correction rule.The application can have a good identification and correction effect on non-standard action in pull-up movement, and can give user correction feedback in real time according to correction rule, assist user to better independently complete pull-up movement training.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to an intelligent online correction method and system for pull-ups. Background Technology

[0002] Pull-ups are a common form of physical exercise that primarily works the muscles in the back, shoulders, and arms. Intelligent online correction uses intelligent technologies (such as computer vision and artificial intelligence algorithms) to analyze and correct a movement or behavior in real time. The intelligent online correction method for pull-ups refers to a technical solution that uses intelligent technology to analyze pull-up movements in real time and automatically provide correction suggestions.

[0003] Intelligent online correction based on pull-up movements can provide personalized guidance, suitable for all types of fitness enthusiasts, especially those who cannot obtain professional coaching. It can effectively improve the training quality and efficiency of athletes and is of great significance in effectively reducing sports injuries caused by incorrect posture.

[0004] Currently, research combining artificial intelligence with sports science to assist exercise has made significant progress, but most studies have only completed tasks such as classification, recognition, and counting. While existing intelligent online correction methods for pull-ups can provide real-time feedback through video analysis and posture recognition, they lack adaptability when dealing with athletes of different body types and skill levels, resulting in inaccurate correction suggestions. Furthermore, devices relying on cameras or sensors are easily affected by factors such as ambient light and device angle, leading to decreased recognition accuracy. Moreover, existing methods often focus on technical implementation, neglecting athletes' psychological feedback and personalized needs, and may not fully consider athletes' physiological characteristics and training goals. Summary of the Invention

[0005] To address the shortcomings of existing intelligent online correction methods for pull-ups, which suffer from poor adaptability to individuals with different body types and skill levels, resulting in inaccurate correction suggestions, reliance on cameras or sensors to collect motion images leading to decreased recognition accuracy, and inability to fully consider the physiological characteristics and training goals of athletes, this invention provides an intelligent online correction method and system for pull-ups.

[0006] The technical solutions provided by the embodiments of the present invention are as follows:

[0007] First aspect:

[0008] This invention provides an intelligent online correction method for pull-ups, comprising:

[0009] S1: Acquire real-time pull-up video data to be corrected;

[0010] S2: Extract the skeletal point coordinate data from pull-up video data using a posture recognition algorithm;

[0011] S3: Perform missing value imputation on the skeletal point coordinate data;

[0012] S4: Based on the coordinate data of the completed skeletal points, establish a joint state space model;

[0013] S5: Predict the target angle based on the joint state space model, combined with the Kalman filter algorithm and curve fitting algorithm;

[0014] S6: If the predicted target angle does not meet the standard indicators for pull-up exercise, output correction prompts through preset correction rules.

[0015] The second aspect:

[0016] This invention provides an intelligent online correction system for pull-ups, comprising:

[0017] processor;

[0018] The memory stores computer-readable instructions, which, when executed by a processor, implement the intelligent online correction method for pull-ups as described in the first aspect.

[0019] Third aspect:

[0020] The present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the intelligent online correction method for pull-ups as described in the first aspect.

[0021] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0022] In this embodiment of the invention, real-time pull-up video data to be corrected is acquired. Then, a posture recognition algorithm is used to extract the skeletal point coordinates from the pull-up video data, ensuring accurate capture of the athlete's movements. Further, missing values ​​in the skeletal point coordinate data are filled in, and a joint state space model is established based on the skeletal point coordinate data. A Kalman filter algorithm and a curve fitting algorithm are combined to predict the target angle, providing a scientific basis for movement correction. Finally, based on a comparison between the predicted target angle and the standard movement, timely correction prompts are given. Through precise posture analysis and intelligent algorithms, non-standard movements are monitored and corrected in real time, improving training effectiveness and reducing sports injuries. Real-time correction feedback is provided to the user according to correction rules, assisting the user in better independently completing pull-up training and meeting the personalized requirements of different athletes. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 A flowchart illustrating an intelligent online correction method for pull-ups provided in an embodiment of the present invention;

[0025] Figure 2 This is a schematic diagram of the structure of an intelligent online correction system for pull-ups provided in an embodiment of the present invention. Detailed Implementation

[0026] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0027] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0028] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, their intended meanings are consistent. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, their intended meanings are consistent.

[0029] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0030] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0031] Reference manual attached Figure 1 The diagram shows a flowchart of an intelligent online correction method for pull-ups provided by an embodiment of the present invention.

[0032] This invention provides an intelligent online correction method for pull-ups. This method can be implemented by an intelligent online correction device for pull-ups, which can be a terminal or a server. The processing flow of the intelligent online correction method for pull-ups may include the following steps:

[0033] S1: Acquire real-time pull-up video data to be corrected.

[0034] Among them, "to be corrected" refers to the athlete's pull-up movements that need to be analyzed and corrected through intelligent methods, and "pull-up video data" refers to video records of athletes performing pull-ups captured by cameras and other devices, which includes real-time images and movement information of the athletes when performing the movements.

[0035] It should be noted that acquiring video data of pull-ups provides a comprehensive foundation for motion analysis. Video data is crucial for subsequent corrective processes, accurately recording the athlete's movement details and capturing every moment of posture. Compared to traditional manual supervision or equipment assistance, video data provides a wealth of information that is easy to analyze and can be easily replayed and processed multiple times.

[0036] S2: Extract the coordinate data of skeletal points from pull-up video data using a posture recognition algorithm.

[0037] Among them, posture recognition algorithm is an algorithm that analyzes human posture through computer vision technology. It is used to extract the position of key skeletal points (such as joints, limb extremities, etc.) of the human body from images or videos. Skeletal point coordinate data refers to the two-dimensional or three-dimensional coordinate data of key human body points (such as head, shoulders, wrists, etc.) extracted by posture recognition algorithm.

[0038] It's important to note that posture recognition algorithms precisely extract the coordinates of skeletal points during pull-ups. This process enables accurate capture of the positions of each joint in the athlete's body, providing detailed information for subsequent motion analysis. Unlike traditional manual recording or the use of complex sensors, posture recognition algorithms can monitor athletes in real time using ordinary camera equipment, reducing equipment costs and operational complexity.

[0039] In one possible implementation, the pose recognition algorithm is specifically the lightweight OpenPose algorithm.

[0040] Extracting skeletal point coordinate data using the lightweight OpenPose algorithm specifically includes:

[0041] S201: Extracts image frames from pull-up video data based on the lightweight OpenPose algorithm.

[0042] The lightweight OpenPose algorithm is an optimized version of the original OpenPose algorithm, aiming to reduce computational resource consumption and improve real-time performance. The OpenPose algorithm uses a deep learning model to extract key skeletal points of the human body from images or videos, such as the head, shoulders, elbows, wrists, knees, and ankles, for pose recognition. An image frame refers to each still image in video data.

[0043] S202: Extracting image features from image frames using a lightweight convolutional network.

[0044] F(t) = MobileNet(I(t))

[0045] Where F(t) represents the image features at time t, MobileNet represents a lightweight convolutional neural network, and I(t) represents the image frame at time t.

[0046] S203: Based on image features, a target point regression algorithm is used to output a heatmap of each skeletal point:

[0047] H k (t)=Heatmap(F(t),k),k∈{1,2,...,N}

[0048] Among them, H k (t) represents the heatmap of the k-th skeletal point at time t, and Heatmap(F(t),k) represents the heatmap generation function, which is used to extract the heatmap of the k-th skeletal point from the input image features F(t), where k = 1, 2, ..., N, and N represents the total number of skeletal points.

[0049] S204: Extract the maximum value locations from each heatmap and generate skeletal point coordinate data:

[0050] (x k (t),y k (t))=arg max(H k (t))

[0051] Where, x k (t) represents the x-coordinate of the k-th skeletal point at time t, y k (t) represents the ordinate of the k-th skeletal point at time t, and argmax represents the independent variable that takes the maximum value.

[0052] It should be noted that lightweight OpenPose optimizes the algorithm by reducing the complexity of the network model, reducing computation, or using more efficient network architectures (such as MobileNet, EfficientNet, etc.), enabling the algorithm to run more efficiently on resource-constrained devices (such as mobile devices, embedded devices, etc.).

[0053] S3: Perform missing value imputation on the skeletal point coordinate data.

[0054] Missing value imputation refers to the process where, during data processing, some skeletal points may not be correctly identified or data may not be available. Missing value imputation uses algorithms to fill in these missing data points, ensuring data integrity and consistency. Specifically, it ensures the integrity of skeletal point coordinate data. When pose recognition algorithms cannot fully extract the coordinates of certain skeletal points, missing value imputation can use interpolation, prediction, or other techniques to fill in the missing data, thereby avoiding analytical errors caused by incomplete data.

[0055] S4: Based on the completed skeletal point coordinate data, establish a joint state space model.

[0056] The joint state-space model is a mathematical model used to describe and predict the motion state of a joint. It considers physical quantities such as joint angles, angular velocities, and angular accelerations, and describes their changes over time through mathematical equations. This model is commonly used in dynamic system analysis to simulate the motion trajectory and behavior of joints. By establishing a joint state-space model, joint motion behavior can be accurately described and predicted. The joint state-space model can not only predict the dynamic changes of movements but also help determine whether movements are performed correctly, thus providing timely corrective suggestions.

[0057] In one possible implementation, S4 specifically includes:

[0058] S401: Calculate the distance between the coordinates of each bone point:

[0059]

[0060] Where a represents the distance between bone point B and bone point C, b represents the distance between bone point A and bone point C, c represents the distance between bone point A and bone point B, (x a ,y a (x) represents the coordinates of skeletal point A. b ,y b (x) represents the coordinates of skeletal point B. c ,y c ) represents the coordinates of the skeletal point C.

[0061] S402: Calculate the target angle using the law of cosines based on the distance.

[0062]

[0063] Where cosα represents the cosine value of the target angle α, and arccos represents the inverse cosine operation.

[0064] S403: Based on the target angle and combined with knowledge of physics, establish the equation of motion for the target angle:

[0065]

[0066] ω(k)=ω(k-1)+β(k-1)·Δt

[0067] (x k (t),y k (t))=arg max(H k (t))

[0068] Where θ(k) represents the angle value of the athlete at time k, ω(k-1) represents the angular velocity of the athlete at time k-1, Δ represents the increment operator, β(k-1) represents the angular acceleration of the athlete at time k-1, and ω(k) represents the angular velocity of the athlete at time k-1.

[0069] S404: Establish a joint state space model based on the equation of motion.

[0070] It should be noted that by calculating the distances between skeletal points and applying the law of cosines to calculate the target angle, the system can accurately measure changes in joint angles, thus providing a reliable data foundation for subsequent motion analysis. This precise dynamic analysis helps improve the effectiveness of movement correction, reduce sports injuries, and optimize the training process.

[0071] In one possible implementation, when the target angle is a periodically changing angle, the state-space model is specifically as follows:

[0072]

[0073] Where, Θ i (k) represents the state vector of the i-th target angle at time k, θ i (k) represents the position of the i-th target angle at time k, ω i (k) represents the angular velocity of the i-th target angle at time k, Ai represents the state transition matrix of the i-th target angle, Bi represents the control input matrix of the i-th target angle, and U i (k-1) represents the control input for the i-th target angle at time k-1, ni represents the noise of the i-th target angle, T represents the time step, and β i (k-1) represents the angular acceleration of the i-th target angle at time k-1.

[0074] The output equation is as follows:

[0075]

[0076] Among them, Y i (k) and y i (k) represents the predicted value of the i-th target angle, H i This represents the observation matrix for the i-th target angle.

[0077] When the target angle is a stable angle, the state-space model is as follows:

[0078]

[0079] The output equation is as follows:

[0080] Y i (k)=y i (k)=θ i (k).

[0081] It should be noted that during a pull-up, some angles (such as the angles of the elbow and shoulder) change periodically with the movement, while other angles (such as the relative position of the chin to the bar) may stabilize at certain stages. Modeling these two types of angles separately allows for a more accurate capture of the characteristics and patterns of change of each angle, thereby improving the model's adaptability to different types of movements. This effectively enhances the model's accuracy, computational efficiency, and adaptability, ultimately improving the effect of motion correction.

[0082] S5: Based on the joint state space model, the target angle is predicted by combining the Kalman filter algorithm and the curve fitting algorithm.

[0083] Kalman filtering is a commonly used algorithm for estimating and predicting dynamic systems, especially suitable for noisy real-time data. In motion analysis, Kalman filtering can accurately predict joint states based on current and historical data, suppressing noise and improving the stability and accuracy of predictions. Curve fitting algorithms use mathematical methods to fit data points into a curve, often used to predict the future state of a system. Target angle refers to the angle formed between joints during motion. For example, the angle between the elbow, shoulder, and wrist is a target angle.

[0084] It's important to note that the Kalman filter and curve fitting algorithms are combined to accurately predict the target angle. The Kalman filter removes noise from real-time data, providing stable angle predictions, while the curve fitting algorithm predicts angle changes based on motion trends. This combination enables the model to provide accurate and reliable feedback in dynamic environments, especially during rapid movement, reducing errors and improving prediction accuracy.

[0085] In this invention, the process first determines whether there is any input data. If there is no input data, the process directly terminates. When the input data is θ5, prediction is performed directly using the Kalman filter algorithm. When the input data is θ3, θ4, and θ6, if the angle change Δθ is greater than 5°, a linear fitting algorithm is used for prediction; otherwise, the Kalman filter algorithm is used, and the process checks again whether Δθ is greater than 5°. If it is, the linear fitting algorithm is used again; otherwise, the Kalman filter algorithm is used. When the input data is θ1 and θ2, if the angle change Δθ is less than 5°, a sine curve fitting algorithm is used for prediction; otherwise, the Kalman filter algorithm is used, and the process checks again whether Δθ is less than 5°. If it is, the sine curve fitting algorithm is used again; otherwise, the Kalman filter algorithm is used.

[0086] By combining algorithms such as linear fitting, sine curve fitting, and Kalman filtering, the system can select the best prediction method according to different situations, avoiding the limitations of a single algorithm and improving the ability to model motion data. Then, by judging the conditions, the prediction method used in different situations is clarified, making the entire correction and prediction process logical, easy to understand and implement, thus providing good technical support for intelligent pull-up correction.

[0087] In one possible implementation, the target angle specifically includes: angle θ1 with the left elbow as the apex and the left wrist and left shoulder as the two endpoints; angle θ2 with the right elbow as the apex and the right wrist and right shoulder as the two endpoints; angle θ3 with the left knee as the apex and the left ankle and left hip as the two endpoints; angle θ4 with the right knee as the apex and the right ankle and right hip as the two endpoints; angle θ5 with the chin as the apex and the left and right wrists as the two endpoints; and angle θ6 with the midpoint of the left and right hips as the apex and the left and right ankles as the two endpoints.

[0088] In one possible implementation, the curve fitting algorithm includes a linear fitting algorithm and a sine curve fitting algorithm.

[0089] Linear fitting algorithms are a type of algorithm that uses methods such as least squares to fit a set of data points into a linear model. Specifically, linear fitting finds a straight line that minimizes the sum of the squares of the perpendicular distances from each data point to that line. This algorithm is typically used when there is an approximately linear relationship between data points, and this line is used to predict or describe the trend of the data. Sine curve fitting algorithms use a sine function to fit a dataset, and are particularly suitable for data exhibiting periodic fluctuations. A sine function has a specific period, amplitude, and phase, and can accurately describe periodically changing data. Sine curve fitting algorithms use the least squares method to fit data points, ensuring that the fitted sine function best represents the fluctuation trend of the data.

[0090] For the prediction of θ1 and θ2, the Kalman filter algorithm and the sine curve fitting algorithm are combined for prediction.

[0091] Specifically, angles θ1 and θ2 exhibit a sinusoidal periodic change during a standard pull-up, and Kalman filtering may introduce some errors at the inflection points of the curve. Based on this sinusoidal periodic characteristic of the curve, an algorithm combining Kalman filtering and sine curve fitting is used. When the curve slope is small, i.e., near the inflection point, a sine curve is fitted to obtain the angle value at the next moment.

[0092] For the prediction of θ3, θ4, and θ6, the Kalman filter algorithm and the line fitting algorithm are combined for prediction.

[0093] It should be noted that angles θ3, θ4, and θ6 remain relatively constant during a standard pull-up. When the slope of the angle curve suddenly increases, it is likely due to false positives, and the values ​​predicted directly using the Kalman filter algorithm may deviate significantly from the actual values. Therefore, using a Kalman filter combined with linear fitting, when the curve slope changes significantly, allows for a more accurate prediction of the angle values ​​at subsequent moments.

[0094] The prediction of θ5 is performed using the Kalman filter algorithm.

[0095] For angle θ5, there are many inflection points, and the effect is poor after two-step prediction. Therefore, the Kalman filter algorithm is adopted to perform only one-step prediction, thereby improving the stability and accuracy of the prediction.

[0096] In one possible implementation, prediction is performed by combining the Kalman filter algorithm and the sine curve fitting algorithm, specifically including:

[0097] When the change in angle θ1 or angle θ2 exceeds a preset threshold, prediction is performed using a sine curve fitting algorithm.

[0098] θ k =Asin(ωt) k +φ)+D

[0099] Where, θ k Let A represent the target angle value at time k, ω represent the amplitude, and t represent the frequency. k Let k represent time k, φ represent phase shift, and D represent overall angle shift.

[0100] Otherwise, use the Kalman filter algorithm for prediction.

[0101] Prediction is performed by combining the Kalman filter algorithm and the line fitting algorithm, specifically including:

[0102] When the angular changes of θ3, θ4, or θ6 are less than a preset threshold, prediction is performed using a linear fitting algorithm.

[0103] θ k =at k +b

[0104] Where, θ k Let represent the target angle value at time k, 'a' represent the slope of the angle change, and 't' represent the angle value at time k. k Let k represent time k, and b represent the initial angle.

[0105] Otherwise, use the Kalman filter algorithm for prediction.

[0106] Prediction using the Kalman filter algorithm specifically includes:

[0107] Calculate the target angle value based on the Kalman filter state prediction equation:

[0108]

[0109] in, This represents the predicted state value at time k, i.e., the target angle value. Let A represent the state estimate at time k-1, A represent the state transition matrix, B represent the control input matrix, and u represent the state estimate at time k-1. k-1 This represents the control vector at time k-1. Let P represent the state prediction error covariance matrix at time k. k-1 Let represent the state estimation error covariance matrix at time k-1, and Q represent the process noise covariance matrix.

[0110] Update the target angle value and the state prediction error covariance matrix:

[0111]

[0112] Among them, K k Let H represent the Kalman gain at time k, H represent the measurement matrix, and R represent the measurement noise covariance matrix. Let z represent the state estimate at time k. k Let P represent the measurement vector at time k. k Let represent the state estimation error covariance matrix at time k after measurement and update.

[0113] S6: If the predicted target angle does not meet the standard indicators for pull-up exercise, output correction prompts through preset correction rules.

[0114] The pull-up standard indicators are set for the joint angles and range of motion during the pull-up movement, aiming to ensure that the movement meets correct biomechanical requirements, avoid sports injuries, and maximize training effects. Standard indicators may include elbow angle, knee angle, and whether the chin goes over the bar. Preset correction rules are a series of rules pre-set by the system based on the exercise standards, used to judge whether the athlete's movement is correct and to provide improvement suggestions when the movement does not meet the standards. Correction prompts are adjustment suggestions given by the system based on the athlete's real-time movement data and preset correction rules when the movement does not meet the standards.

[0115] It's worth noting that the system automatically outputs corrective prompts based on a comparison between the target angle and the standard movement, helping athletes adjust improper movements in real time. Through preset correction rules, the system can assess changes in key angles in real time during exercise and provide feedback based on the athlete's performance. This intelligent correction method eliminates the shortcomings of traditional training that relies on manual observation and judgment, thus improving training efficiency.

[0116] In one possible implementation, the sports standard indicators specifically include: Indicator One to Indicator Five.

[0117] The first indicator is: during the exercise, when the lowest point is reached, the angles θ1 and θ2 between the two arms with the elbow as the apex should be close to 180 degrees.

[0118] The second indicator is as follows: During the exercise, the athlete cannot use the momentum of pushing off with their legs, and the angles θ3 and θ4 between the two legs with the knee as the apex should be close to 180 degrees.

[0119] Indicator 3 specifically states that during exercise, one cannot use the left or right leg swings to gain momentum, meaning the angle θ5 between the two legs and the midpoint of the hip is less than the preset angle value.

[0120] The fourth indicator is as follows: During the exercise, at the highest point, the athlete's chin must be higher than the horizontal bar, that is, θ6 is less than 180 degrees, and the vertical coordinate of the chin must be higher than the coordinates of both wrists.

[0121] The fifth indicator is: during the exercise, the time from when the athlete starts to rise from the moment they grab the bar to when they lower it to the lowest point should not exceed the preset time.

[0122] It's important to note that judging whether a pull-up is performed correctly usually relies on certain standards. These include whether the athlete uses body swing to gain momentum during the pull-up, and whether the chin does not extend above the bar at the top. Establishing these standards not only reduces sports injuries caused by improper form but also provides athletes with personalized improvement suggestions, helping them achieve their desired training results more quickly.

[0123] In one possible implementation, the preset correction rules specifically include: Rule 1 to Rule 5.

[0124] Rule 1 specifically states: By assessing the degree of arm bending and the lowering of the body, different levels of reminders are output.

[0125] Rule 2 specifically involves classifying and providing feedback on the degree of leg flexion to help avoid using leverage.

[0126] Rule 3 specifically states: Check the angle between the midpoint of the legs and the hips to ensure proper form.

[0127] Rule four specifically states: Ensure that your chin is above the horizontal bar and adjust the height accordingly.

[0128] Rule 5 specifically states: based on the speed of movement, prevent movements from being too slow or improper.

[0129] It should be noted that the pre-set correction rules can detect and provide timely feedback on non-standard movements in real time when athletes perform pull-ups, avoiding the delays caused by relying on manual observation and improving the efficiency of correction. By systematically correcting non-standard movements, the pre-set correction rules can effectively reduce sports injuries caused by incorrect movements and ensure the safety of athletes during training.

[0130] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0131] In this embodiment of the invention, real-time pull-up video data to be corrected is acquired. Then, a posture recognition algorithm is used to extract the skeletal point coordinates from the pull-up video data, ensuring accurate capture of the athlete's movements. Further, missing values ​​in the skeletal point coordinate data are filled in, and a joint state space model is established based on the skeletal point coordinate data. A Kalman filter algorithm and a curve fitting algorithm are combined to predict the target angle, providing a scientific basis for movement correction. Finally, based on a comparison between the predicted target angle and the standard movement, timely correction prompts are given. Through precise posture analysis and intelligent algorithms, non-standard movements are monitored and corrected in real time, improving training effectiveness and reducing sports injuries. Real-time correction feedback is provided to the user according to correction rules, assisting the user in better independently completing pull-up training and meeting the personalized requirements of different athletes.

[0132] Reference manual attached Figure 2 The diagram shows a structural schematic of an intelligent online correction system for pull-ups provided by the present invention.

[0133] The present invention also provides an intelligent online correction system 20 for pull-ups, applied to the above-mentioned intelligent online correction method for pull-ups, comprising:

[0134] Processor 201.

[0135] The memory 202 stores computer-readable instructions that, when executed by the processor 201, implement the intelligent online correction method for pull-ups as described in the method embodiment.

[0136] The intelligent online correction system 20 for pull-ups provided by this invention can perform the above-mentioned intelligent online correction method for pull-ups and achieve the same or similar technical effects. To avoid repetition, this invention will not elaborate further.

[0137] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0138] In this embodiment of the invention, real-time pull-up video data to be corrected is acquired. Then, a posture recognition algorithm is used to extract the skeletal point coordinates from the pull-up video data, ensuring accurate capture of the athlete's movements. Further, missing values ​​in the skeletal point coordinate data are filled in, and a joint state space model is established based on the skeletal point coordinate data. A Kalman filter algorithm and a curve fitting algorithm are combined to predict the target angle, providing a scientific basis for movement correction. Finally, based on a comparison between the predicted target angle and the standard movement, timely correction prompts are given. Through precise posture analysis and intelligent algorithms, non-standard movements are monitored and corrected in real time, improving training effectiveness and reducing sports injuries. Real-time correction feedback is provided to the user according to correction rules, assisting the user in better independently completing pull-up training and meeting the personalized requirements of different athletes.

[0139] It should be understood that the processor in the embodiments of the present invention can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0140] It should also be understood that the memory in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDRSDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0141] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the flow or function according to the embodiments of the present invention is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. A computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. Available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. Semiconductor media can be solid-state drives.

[0142] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0143] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0144] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0145] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0146] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0147] In the embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0148] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0149] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0150] If the functionality is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0151] This invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the intelligent online correction method for pull-ups as described in the method embodiments.

[0152] The present invention provides a computer-readable storage medium that can implement the steps and effects of the intelligent online correction method for pull-ups described in the above method embodiments. To avoid repetition, the present invention will not repeat the details.

[0153] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0154] In this embodiment of the invention, real-time pull-up video data to be corrected is acquired. Then, a posture recognition algorithm is used to extract the skeletal point coordinates from the pull-up video data, ensuring accurate capture of the athlete's movements. Further, missing values ​​in the skeletal point coordinate data are filled in, and a joint state space model is established based on the skeletal point coordinate data. A Kalman filter algorithm and a curve fitting algorithm are combined to predict the target angle, providing a scientific basis for movement correction. Finally, based on a comparison between the predicted target angle and the standard movement, timely correction prompts are given. Through precise posture analysis and intelligent algorithms, non-standard movements are monitored and corrected in real time, improving training effectiveness and reducing sports injuries. Real-time correction feedback is provided to the user according to correction rules, assisting the user in better independently completing pull-up training and meeting the personalized requirements of different athletes.

[0155] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

[0156] The following points need to be explained:

[0157] (1) The accompanying drawings of the embodiments of the present invention only involve the structures involved in the embodiments of the present invention. Other structures can refer to the general design.

[0158] (2) For clarity, the thickness of layers or regions is enlarged or reduced in the drawings used to describe embodiments of the present invention; that is, these drawings are not drawn to actual scale. It is understood that when an element such as a layer, film, region, or substrate is referred to as being “above” or “below” another element, the element may be “directly” located “above” or “below” the other element, or there may be intermediate elements.

[0159] (3) Where there is no conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other to obtain new embodiments.

[0160] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. The scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An intelligent online correction method for pull-ups, characterized in that, The method comprises the following steps: S1: acquiring real-time pull-up video data to be corrected; S2: extracting bone point coordinate data in the pull-up video data through a pose recognition algorithm; S3: performing missing value completion processing on the bone point coordinate data; S4: establishing a joint state space model according to the bone point coordinate data after the completion processing; The S4 specifically comprises: S401: calculating the distances between the bone point coordinates; S402: calculating the target angle by using the cosine theorem according to the distances; S403: establishing a motion equation of the target angle according to the target angle and combining physical knowledge; S404: establishing the joint state space model based on the motion equation; When the target angle is a periodic change angle, the state space model is specifically as follows: wherein Θ i (k) denotes the state vector of the i-th target angle at the k-th time instant, θ i (k) denotes the position of the i-th target angle at the k-th time instant, ω i (k) denotes the angular velocity of the i-th target angle at the k-th time instant, A i denotes the state transition matrix of the i-th target angle, B i denotes the control input matrix of the i-th target angle, U i (k-1) denotes the control input of the i-th target angle at the k-1-th time instant, n denotes the noise of the i-th target angle, T denotes the time step, β i (k-1) denotes the angular acceleration of the i-th target angle at the k-1-th time instant; When the target angle is a stable angle, the state space model is specifically as follows: S5: predicting the target angle according to the joint state space model, combining a Kalman filtering algorithm and a curve fitting algorithm; The target angle specifically comprises: an angle θ1 with the left elbow as a vertex and the left wrist and the left shoulder as two endpoints; an angle θ2 with the right elbow as a vertex and the right wrist and the right shoulder as two endpoints; an angle θ3 with the left knee as a vertex and the left ankle and the left crotch as two endpoints; an angle θ4 with the right knee as a vertex and the right ankle and the right crotch as two endpoints; an angle θ5 with the lower jaw as a vertex and the left and right wrists as two endpoints; and an angle θ6 with the midpoint of the left and right crotches as a vertex and the left and right ankles as two endpoints; The curve fitting algorithm comprises a linear fitting algorithm and a sine curve fitting algorithm; For the prediction of θ1 and θ2, the Kalman filtering algorithm and the sine curve fitting algorithm are combined for prediction; When the angle change of θ1 or the angle change of θ2 is greater than a preset threshold, the sine curve fitting algorithm is used for prediction; otherwise, the Kalman filtering algorithm is used for prediction; For the prediction of θ3, θ4 and θ6, the Kalman filtering algorithm and the linear fitting algorithm are combined for prediction; When the angle change of θ3, θ4 or θ6 is less than a preset threshold, the linear fitting algorithm is used for prediction; otherwise, the Kalman filtering algorithm is used for prediction; For the prediction of θ5, the Kalman filtering algorithm is used for prediction; S6: in the case that there is a predicted target angle that does not meet the pull-up movement standard index, a correction prompt is output through a preset correction rule.

2. The intelligent online correction method for pull-ups according to claim 1, characterized in that, The pose recognition algorithm is specifically a lightweight OpenPose algorithm; The extraction of the bone point coordinate data through the lightweight OpenPose algorithm specifically comprises: S201: extracting an image frame in the pull-up video data based on the OpenPose algorithm; S202: extracting image features of the image frame through a lightweight convolutional network: F(t) = MobileNet(I(t)) Wherein, F(t) represents the image features at time t, MobileNet represents a lightweight convolutional neural network, and I(t) represents the image frame at time t. S203: According to the image features, output the heat map of each skeleton point by a target point regression algorithm; H k (t) = Heatmap(F(t), k), k e {1, 2,..., N} wherein H k (t) represents a heat map of the kth skeleton point at time t, Heatmap(F(t), k) represents a heat map generation function for extracting the heat map of the kth skeleton point from the input image feature F(t), k = 1, 2, …, N, and N represents the total number of skeleton points. S204: Extract the maximum position of each heat map to generate the skeleton point coordinate data; (x k (t),y k (t))=argmax(H k (t)) wherein x k (t) denotes the horizontal coordinate of the kth skeleton point at time t, y k (t) denotes the vertical coordinate of the kth skeleton point at time t, and argmax denotes the argument of the maximum.

3. The intelligent online correction method for pull-ups according to claim 1, characterized in that, The formula for calculating the distance between each skeleton point coordinate is specifically: wherein a represents a distance between the skeletal point B and the skeletal point C, b represents a distance between the skeletal point A and the skeletal point C, c represents a distance between the skeletal point A and the skeletal point B, (x a ,y a ) represents coordinates of the skeletal point A, (x b ,y b ) represents coordinates of the skeletal point B, and (x c ,y c ) represents coordinates of the skeletal point C. The formula for calculating the target angle is specifically: Wherein, cosα represents the cosine value of the target angle α, and arccos represents the inverse cosine operation; The motion equation of the target angle is specifically: ω(k) = ω(k-1) + β(k-1)·Δt Wherein, θ(k) represents the angle value of the mover at the kth moment, ω(k-1) represents the angular velocity of the mover at the (k-1)th moment, Δ represents the increment operator, β(k-1) represents the angular acceleration of the mover at the (k-1)th moment, and ω(k) represents the angular velocity of the mover at the kth moment.

4. The intelligent online correction method for pull-ups of claim 1, wherein, When the target angle is a periodic change angle, the output equation is specifically: where Y i (k) and y i (k) are the predicted values of the ith target angle, H i is the observation matrix of the ith target angle. When the target angle is a stable angle, the output equation is specifically: Y i (k) = y i (k) = θ i (k).

5. The intelligent online correction method for pull-ups of claim 1, wherein, The prediction combining the Kalman filtering algorithm and the sine curve fitting algorithm specifically includes: When the angle change of θ1 or the angle change of θ2 is greater than a preset threshold, the prediction is made by the sine curve fitting algorithm: θ k = Asin(ωt k + φ) + D where θ k represents the target angle value at time k, A represents the amplitude, ω represents the frequency, t k represents the time k, φ represents the phase offset, and D represents the offset of the overall angle. Otherwise, the prediction is made by the Kalman filtering algorithm; The prediction combining the Kalman filtering algorithm and the linear fitting algorithm specifically includes: When the angle change of θ3, θ4 or θ6 is less than a preset threshold, the prediction is made by the linear fitting algorithm: θ k = at k +b wherein θ k represents the target angle value at time k, a represents the slope of the angle change, t k represents the initial angle at time k, and b represents the initial angle. Otherwise, the prediction is made by the Kalman filtering algorithm; The prediction by the Kalman filtering algorithm specifically includes: Based on the Kalman filtering state prediction equation, the target angle value is calculated: wherein, denotes a state prediction value at the kth time, i.e., a target angle value, denotes a state estimation value at the (k-1)th time, A denotes a state transition matrix, B denotes a control input matrix, u k-1 denotes a control vector at the (k-1)th time, denotes a state prediction error covariance matrix at the kth time, P k-1 denotes a state estimation error covariance matrix at the (k-1)th time, Q denotes a process noise covariance matrix; The target angle value and the state prediction error covariance matrix are updated: wherein K k represents the Kalman gain at the kth moment, H represents the measurement matrix, and R represents the measurement noise covariance matrix, represents the state estimation value at the kth moment, z k represents the measurement vector at the kth moment, P k represents the state estimation error covariance matrix at the kth moment after measurement update.

6. The intelligent online correction method for pull-ups of claim 1, wherein, The motion standard indicators specifically include indicators one to five; The indicator one is specifically that during the movement process, the included angle θ1 between the two arms with the elbow as the vertex and the included angle θ2 between the two arms are both close to 180 degrees when reaching the lowest point; The indicator two is specifically that during the movement process, the mover cannot use the front and back leg to borrow force, and the included angle θ3 between the two legs with the knee as the vertex and the included angle θ4 between the two legs are both close to 180 degrees; The indicator three is specifically that during the movement process, the mover cannot use the left and right leg to borrow force, that is, the included angle θ5 between the two legs and the crotch midpoint is less than a preset included angle value; The indicator four is specifically that during the movement process, the jaw of the mover must be higher than the single bar when reaching the highest point, that is, θ6 is less than 180 degrees, and the vertical coordinate of the jaw is higher than the wrist coordinate; The indicator five is specifically that during the movement process, the time from the start of the movement to the lowest point cannot be greater than a preset time.

7. The intelligent online correction method for pull -ups according to claim 1, characterized in that, The preset correction rules specifically include rules one to five; Rule one is specifically that the bending degree of the arm and the lowering of the body are evaluated to output different degrees of reminders; Rule two is specifically that the bending degree of the leg is classified and fed back to help avoid borrowing force; Rule three is specifically that the included angle between the two legs and the crotch midpoint is checked to ensure the action standard; Rule four is specifically: ensure the mandibular position is higher than the single bar, and adjust the rising height; Rule five is specifically: prevent slow or non-standard movements based on movement speed.

8. An intelligent online correction system for pull-ups, characterized in that, Comprise: A processor; A memory, wherein computer readable instructions are stored on the memory, and the computer readable instructions are executed by the processor to implement the intelligent online correction method for pull-ups according to any one of claims 1 to 7.

Images