A swimming underwater physical training effect analysis system based on training data

Abstract

The invention relates to the field of swimming physical training analysis, and specifically discloses a swimming underwater physical training effect analysis system based on training data. The invention collects video data, motion data, and physiological data of a swimmer by integrating an underwater camera, an inertial measurement unit, and electromyography, heart rate, and blood oxygen sensors to realize spatiotemporal synchronous analysis of multimodal data; based on the consistency of joint angle changes and the synchronization of muscle activation, it is judged whether the swimmer's movements are standard and AR real-time feedback is provided, which is beneficial for the swimmer to correct deviations in real time during training; based on underwater speed, stroke length, stroke frequency, swimming distance, and swimming time, the swimmer's training load is evaluated and an over-limit warning is issued, which is beneficial for formulating a personalized training plan and avoiding overtraining at the same time; by monitoring the swimmer's heart rate, blood oxygen saturation, muscle activation duration, and joint activity, it is judged whether the swimmer has an injury risk and triggers an early warning to reduce the injury incidence rate.

Description

Technical Field

[0001] The invention relates to the field of swimming physical fitness training analysis, and in particular to a swimming underwater physical fitness training effect analysis system based on training data. Background Art

[0002] In the field of modern competitive sports, swimming has extremely high physical requirements for athletes, and the effectiveness of underwater physical training is directly related to the performance of athletes in competitions. With the continuous development of swimming, improving training efficiency and accurately evaluating training results have become key issues.

[0003] Traditional underwater swimming physical training mostly relies on the coach's experience and observation and the athlete's subjective feelings to judge the training effect. This method has significant limitations. On the one hand, it is difficult for the coach to comprehensively and accurately capture and analyze every detail of the athlete's movements and physical energy consumption in the complex underwater environment; on the other hand, the athlete's subjective feelings may be biased due to fatigue, psychological factors, etc., and cannot accurately reflect the true physical state. In addition, the lack of systematic and quantitative data support means that the adjustment of the training plan often lacks scientific basis, making it difficult to achieve personalized and precise training.

[0004] Therefore, developing a system that can comprehensively and accurately analyze the effects of underwater swimming physical training based on training data is of great practical significance for improving the scientific nature of swimming training, optimizing training plans, and improving the competitive level of athletes.

[0005] At present, there are obvious defects in the research and application of analyzing the effect of underwater swimming physical training.

[0006] For example, the existing Chinese patent with publication number CN119149973A discloses a real-time swimming style recognition method and system, which can use a head-mounted six-axis sensor to obtain speed data in real time during swimming; use a trained deep neural network to perform feature recognition on the real-time speed data to obtain a recognition result carrying the swimming style type and swimming style data; use a client to wirelessly receive the recognition result, and display the swimming style type and swimming style data carried by the recognition result in real time. The technical solution shown in this invention enables the coach to monitor the swimmer's swimming style type and swimming style data in real time, and then guide the swimmer to adjust and optimize the movement in time during training. The provided swimming style data enables the coach to accurately control the swimmer's swimming movement and achieve better training results.

[0007] The shortcomings of the above patents are: 1. Insufficient integration of multi-dimensional data, and failure to achieve spatiotemporal synchronous analysis of multimodal data such as video posture, underwater dynamic parameters, electromyographic signals, and physiological parameters, resulting in blind spots in the analysis of the complete training cycle.

[0008] 2. The real-time training load over-limit risk warning mechanism and injury risk warning mechanism are missing, and it is impossible to warn of over-training and injury risks, resulting in a high incidence of injuries.

[0009] 3. Real-time feedback is delayed, and training-related information cannot be fed back to swimmers in time so that they can make real-time corrections during training, which cannot meet the urgent need for real-time adjustment of training plans during high-intensity training. Summary of the invention

[0010] In view of the above problems, the present invention proposes a swimming underwater physical training effect analysis system based on training data to realize the function of swimming physical training analysis.

[0011] The technical solution adopted by the present invention to solve its technical problem is: the present invention provides a swimming underwater physical training effect analysis system based on training data, including: a multimodal data acquisition module: through an underwater camera, an inertial measurement unit, an electromyography sensor, a heart rate sensor, and a blood oxygen sensor, real-time video data, motion data, and physiological data of the swimmer during the training cycle are collected.

[0012] Movement Standardization Assessment Module: Identifies the swimmer's swimming posture and obtains the angle of each joint and the activation time and RMS value of each muscle group in each movement. It assesses the consistency of joint angle changes and muscle activation synchronization of each movement, determines whether the swimmer's movements are standard, and provides real-time AR feedback.

[0013] Training load monitoring module: Based on the actual measured values ​​of the swimmer's underwater speed, stroke length, stroke frequency, swimming distance and swimming time during the training cycle, the swimmer's training load is evaluated and an over-limit warning is issued.

[0014] Injury risk warning module: real-time monitoring of the swimmer's heart rate, blood oxygen saturation, duration of muscle activation, and angular velocity and angle of each joint movement during the training cycle, to determine whether the swimmer is at risk of injury and trigger a warning.

[0015] Database: Stores sample image libraries of various swimming styles and standard angles of each joint in each action, as well as information on precursors to swimming injuries.

[0016] Based on the above embodiments, the specific working process of the action normativeness evaluation module includes: S1: detecting the turning points of the stroke phase of the swimmer's video data during the training cycle through the optical flow method to extract key frames, obtain the key images of the swimmer during the training cycle, and compare them with the sample image library of various swimming styles stored in the database to identify the swimmer's swimming style.

[0017] S2: Obtain the shooting time points corresponding to each key image of the swimmer during the training cycle, and record them as the action switching time points during the training cycle.

[0018] The inertial measurement units installed at each joint of the swimmer are used to obtain the three-dimensional acceleration data, angular velocity data and posture data of each joint of the swimmer at each action switching time point during the training cycle. Combined with the key images of the swimmer during the training cycle, the three-dimensional posture is reconstructed using a graph convolutional network to generate a three-dimensional skeletal model of each action of the swimmer.

[0019] A baseline is selected in the swimmer's three-dimensional skeleton model according to a preset principle. According to the three-dimensional skeleton model of each action of the swimmer, the angle between each joint and the baseline in each action of the swimmer is obtained, and the angle is recorded as the angle of each joint in each action of the swimmer.

[0020] S3: The activation time and RMS value of each muscle group of the swimmer at each action switching time point during the training cycle are obtained through the electromyographic sensors deployed on each muscle group of the swimmer, and are recorded as the activation time and RMS value of each muscle group under each action of the swimmer.

[0021] Based on the above embodiment, the specific working process of the movement standardization evaluation module also includes: extracting the standard angles of each joint in each action of various swimming styles stored in the database, and screening the standard angles of each joint in each action corresponding to the swimmer's swimming style.

[0022] The angles of each joint in each action of the swimmer are compared with the corresponding standard angles to obtain the angle deviations of each joint in each action of the swimmer, and the maximum angle deviations of the joints in each action of the swimmer are further obtained and substituted into the calculation formula Analyze the consistency coefficient of joint angle changes in each action of swimmers ,in , Respectively represent the swimmer The consistency coefficient of joint angle change and the maximum joint angle deviation of each action, , Indicates The number of the action, , Represents a natural constant.

[0023] On the basis of the above-mentioned embodiment, the specific working process of the action normative evaluation module also includes: comparing the activation time and RMS value of each muscle group under each action of the swimmer, obtaining the maximum deviation of the muscle group activation time and the maximum deviation of the activation RMS value under each action of the swimmer, and recording them as , and substitute it into the calculation formula Analyze the muscle activation synchronization coefficient of each swimmer's action ,in They represent the preset thresholds for the muscle group activation time and the maximum deviation of the activation RMS value, respectively.

[0024] Based on the above embodiments, the specific working process of the action normativeness evaluation module also includes: calculating the weighted average of the joint angle change consistency coefficient and the muscle activation synchronization coefficient of each action of the swimmer, analyzing the normativeness evaluation index of each action of the swimmer, and comparing it with the preset normativeness evaluation index threshold. If the normativeness evaluation index of an action is less than the threshold, the action is not normative, and the swimmer's non-normative actions are counted, projected in real time through the AR goggles worn by the swimmer, and transmitted to the remote mobile terminal device.

[0025] Compared with the prior art, the swimming underwater physical training effect analysis system based on training data described in the present invention has the following beneficial effects: 1. The present invention integrates underwater cameras, inertial measurement units, electromyography sensors, heart rate sensors, and blood oxygen sensors to collect swimmers' video data, motion data, and physiological data in real time, realizes spatiotemporal synchronous analysis of multimodal data, and can capture swimmers' underwater motion characteristics in all dimensions, solving the problem of one-sided data in traditional single-dimensional monitoring.

[0026] 2. The present invention identifies the swimmer's swimming posture and evaluates the consistency of joint angle changes and muscle activation synchronization of each action to determine whether the swimmer's movements are standard and provide AR real-time feedback, which is beneficial for the swimmer to correct deviations in real time during training and improve training effects.

[0027] 3. The present invention evaluates the swimmer's training load based on underwater speed, stroke length, stroke frequency, swimming distance and swimming time, and issues an over-limit warning, which is conducive to formulating a personalized training plan while avoiding overtraining.

[0028] 4. The present invention determines whether the swimmer is at risk of injury and triggers an early warning by monitoring the swimmer's heart rate, blood oxygen saturation, duration of muscle activation, and joint activity, thereby providing a timely early warning of the risk of injury and reducing the incidence of injury.

[0029] 5. The present invention realizes zero-delay training feedback through AR visual channel and augmented reality human-computer interaction, meeting the needs of real-time correction and timely adjustment of training plans during training. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0031] Figure 1 This is a system module connection diagram of the present invention.

[0032] Figure 2 It is a workflow diagram of the action normativeness evaluation module of the present invention.

[0033] Figure 3 This is a workflow diagram of the training load monitoring module of the present invention. DETAILED DESCRIPTION

[0034] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] See also Figure 1 As shown, the present invention provides a swimming underwater physical training effect analysis system based on training data, including a multimodal data acquisition module, an action normative evaluation module, a training load monitoring module, an injury risk warning module, and a database.

[0036] The multimodal data acquisition module is respectively connected to the action normative evaluation module, the training load monitoring module, and the injury risk warning module, and the database is respectively connected to the action normative evaluation module and the injury risk warning module.

[0037] The multimodal data acquisition module collects video data, motion data, and physiological data of the swimmer during the training cycle in real time through an underwater camera, an inertial measurement unit, an electromyography sensor, a heart rate sensor, and a blood oxygen sensor.

[0038] Furthermore, the specific working process of the multimodal data acquisition module is: according to preset principles, a number of underwater high-frame rate cameras are deployed at set positions in the swimming pool, and a camera network is built in conjunction with a head-mounted miniature camera integrated in the swimmer's swimming cap to collect video data of the swimmer during the training cycle in real time.

[0039] According to the preset principles, inertial measurement units are fixedly installed at each joint of the swimmer, and electromyography sensors are deployed at each muscle group of the swimmer to collect the swimmer's motion data during the training cycle in real time.

[0040] Through the heart rate sensor and blood oxygen sensor integrated in the swimmer's wearable device, the swimmer's physiological data during the training cycle is collected in real time.

[0041] It should be noted that the joints of the swimmer include but are not limited to wrists, waist, head, legs, etc.

[0042] It should be noted that the muscle groups of the swimmer are power-generating muscle groups, including but not limited to the pectoralis major, latissimus dorsi, triceps brachii, etc.

[0043] It should be noted that the heart rate sensor can be integrated into a waterproof wristband, a waterproof chest strap, a swimming goggles nose pad, etc., to continuously monitor the heart rate through photoplethysmography.

[0044] It should be noted that the blood oxygen sensor can be integrated into a waterproof wristband, etc., and detects blood oxygen saturation based on the photoplethysmography method in spectrophotometry.

[0045] As a preferred solution, the underwater high frame rate camera adopts a wide-angle polarized lens and cooperates with a flicker suppression algorithm to achieve an image clarity of more than 98% in a turbulent environment.

[0046] It should be noted that by deploying a camera network, the swimmer's motion trajectory can be captured from multiple perspectives.

[0047] As a preferred solution, the video data collected by each camera is spliced ​​to obtain the video data of the swimmer during the training period.

[0048] As a preferred solution, the inertial measurement unit (IMU) is installed through a magnetic waterproof buckle, which can collect three-dimensional acceleration data, angular velocity data, attitude data, etc. When the inertial measurement unit is started, the sensor calibration is automatically performed: a coordinate system is established in the swimming pool according to the set principles, and the swimming pool coordinate system is used as the reference coordinate system to perform zero bias compensation and coordinate system alignment on the inertial measurement unit.

[0049] As a preferred solution, the electromyographic sensor uses a flexible electrode patch. The electromyographic sensor is calibrated by a resting state baseline when it is started.

[0050] It should be noted that the present invention integrates underwater cameras, inertial measurement units, electromyography sensors, heart rate sensors, and blood oxygen sensors to collect swimmers' video data, motion data, and physiological data in real time, realizes spatiotemporal synchronous analysis of multimodal data, and can capture swimmers' underwater motion characteristics in all dimensions, solving the problem of one-sided data in traditional single-dimensional monitoring.

[0051] The movement standardization evaluation module identifies the swimmer's swimming posture and obtains the angle of each joint and the activation time and RMS value of each muscle group in each movement, evaluates the consistency of joint angle changes and muscle activation synchronization of each movement, determines whether the swimmer's movements are standard, and provides AR real-time feedback.

[0052] Further, see Figure 2As shown, the specific working process of the action normativeness evaluation module includes: S1: detecting the turning points of the stroke phase of the swimmer's video data during the training cycle through the optical flow method to extract key frames, obtain the key images of the swimmer during the training cycle, and compare them with the sample image library of various swimming styles stored in the database to identify the swimmer's swimming style.

[0053] As a preferred solution, the specific process of identifying a swimmer's swimming style is as follows: comparing each key image of the swimmer during the training cycle with a sample image library of various swimming styles, obtaining the maximum similarity between the key image of the swimmer during the training cycle and the sample images of various swimming styles, and comparing them with each other; if the maximum similarity between the key image of the swimmer during the training cycle and a certain swimming style sample image is higher than that of other swimming style sample images, then this swimming style is recorded as the swimmer's swimming style.

[0054] S2: Obtain the shooting time points corresponding to each key image of the swimmer during the training cycle, and record them as the action switching time points during the training cycle.

[0055] The inertial measurement units installed at each joint of the swimmer are used to obtain the three-dimensional acceleration data, angular velocity data and posture data of each joint of the swimmer at each action switching time point during the training cycle. Combined with the key images of the swimmer during the training cycle, the three-dimensional posture is reconstructed using a graph convolutional network to generate a three-dimensional skeletal model of each action of the swimmer.

[0056] A baseline is selected in the swimmer's three-dimensional skeleton model according to a preset principle. According to the three-dimensional skeleton model of each action of the swimmer, the angle between each joint and the baseline in each action of the swimmer is obtained, and the angle is recorded as the angle of each joint in each action of the swimmer.

[0057] S3: The activation time and RMS value of each muscle group of the swimmer at each action switching time point during the training cycle are obtained through the electromyographic sensors deployed on each muscle group of the swimmer, and are recorded as the activation time and RMS value of each muscle group under each action of the swimmer.

[0058] It should be noted that the optical flow method is a computer vision technology used to estimate the movement of pixels in an image sequence. It is based on the brightness consistency assumption, that is, the brightness of an object remains unchanged in a short period of time, and combines the changes in the grayscale value of the image to infer the movement direction and speed of the pixel.

[0059] It should be noted that the paddling action can generally be divided into different stages such as entering the water, holding the water, paddling, and exiting the water. The movement characteristics and limb forms of each stage are different. The transition points of these stages are the turning points of the paddling phase.

[0060] It should be noted that when using the optical flow method to detect the turning point of the stroke phase, the video sequence of the stroke action is first processed. The motion vectors of the pixels of the human limbs between adjacent frames are calculated by the optical flow algorithm, and the changes of these motion vectors are analyzed. When the stroke action switches from one phase to another, the direction and speed of the arm's movement will change significantly, which will be reflected in the optical flow vector. There will also be corresponding mutations or characteristic changes. By setting appropriate thresholds and analysis algorithms, the turning point of the stroke phase can be determined based on these changes in the optical flow vector.

[0061] It should be noted that the optical flow method for detecting the turning points of the stroke phase and the key frame extraction are both relatively mature technical means available and will not be elaborated here.

[0062] It should be noted that the use of graph convolutional networks for 3D posture reconstruction is a relatively mature technical means available, which will not be elaborated here.

[0063] In a specific embodiment, the electromyographic sensor is a surface electromyograph. When in use, the electrode sheet is attached to the relevant muscle groups of the swimmer to collect the electrical signals of the muscles during exercise, and the muscle activation time and RMS value are calculated through signal processing and analysis software. It should be noted that the electromyographic signal is a bioelectric signal generated by the muscle during contraction or relaxation. The RMS value can reflect the intensity or power level of the electromyographic signal. The stronger the muscle contraction and the more muscle fibers involved in the activity, the greater the amplitude of the electromyographic signal generated, and the higher the corresponding RMS value.

[0064] Furthermore, the specific working process of the movement standardization evaluation module also includes: extracting the standard angles of each joint in each movement of various swimming styles stored in the database, and screening the standard angles of each joint in each movement corresponding to the swimmer's swimming style.

[0065] The angles of each joint in each action of the swimmer are compared with the corresponding standard angles to obtain the angle deviations of each joint in each action of the swimmer, and the maximum angle deviations of the joints in each action of the swimmer are further obtained and substituted into the calculation formula Analyze the consistency coefficient of joint angle changes in each action of swimmers ,in , Respectively represent the swimmer The consistency coefficient of joint angle change and the maximum joint angle deviation of each action, , Indicates The number of the action, , Represents a natural constant.

[0066] It should be noted that the lower the deviation between the angle of each joint of the swimmer in each action and its corresponding standard angle, the more standardized the action. In a specific embodiment, the abduction angle of the swimmer's shoulder during freestyle swimming should be 90°-120°, and the deviation of the swimmer's shoulder angle from this range in actual training is recorded. The average deviation within 10° is considered good.

[0067] Furthermore, the specific working process of the action normative evaluation module also includes: comparing the activation time and RMS value of each muscle group under each action of the swimmer, obtaining the maximum deviation of the muscle group activation time and the maximum deviation of the activation RMS value under each action of the swimmer, and recording them as , and substitute it into the calculation formula Analyze the muscle activation synchronization coefficient of each swimmer's action ,in They represent the preset thresholds for the muscle group activation time and the maximum deviation of the activation RMS value, respectively.

[0068] It should be noted that by measuring the activation time and activation RMS value of different muscle groups in each action and calculating the synchronization coefficient of muscle activation, the higher the synchronization coefficient, the better the muscle cooperation and the higher the action efficiency. In a specific embodiment, in the breaststroke kicking action, the front thigh muscles and the back thigh muscles should be activated synchronously at a specific time point.

[0069] Furthermore, the specific working process of the action normativeness evaluation module also includes: calculating the weighted average of the joint angle change consistency coefficient and the muscle activation synchronization coefficient of each action of the swimmer, analyzing the normativeness evaluation index of each action of the swimmer, and comparing it with the preset normativeness evaluation index threshold. If the normativeness evaluation index of an action is less than the threshold, the action is non-standard, and the non-standard actions of the swimmer are counted, projected in real time through the AR goggles worn by the swimmer, and transmitted to the remote mobile terminal device.

[0070] It should be noted that the weights of the joint angle change consistency coefficient and the muscle activation synchronization coefficient are set values, and the cumulative sum is 1.

[0071] It should be noted that when AR goggles project the swimmer's irregular movements in real time, the images can be projected onto the bottom or sides of the swimming pool and other areas within the swimmer's field of view. The content of the images can be signs, symbols, shapes or names of irregular movements.

[0072] It should be noted that the present invention determines whether the swimmer's movements are standard and provides real-time AR feedback by identifying the swimmer's swimming posture and evaluating the consistency of joint angle changes and muscle activation synchronization of each action, which is beneficial for the swimmer to correct deviations in real time during training and improve training effects.

[0073] It should be noted that the present invention realizes zero-delay training feedback through AR visual channels and augmented reality human-computer interaction, meeting the needs of real-time correction and timely adjustment of training plans during training.

[0074] The training load monitoring module evaluates the swimmer's training load and issues an over-limit warning based on the actual measured values ​​of the swimmer's underwater speed, stroke length, stroke frequency, swimming distance and swimming time during the training cycle.

[0075] Further, see Figure 3 As shown, the specific working process of the training load monitoring module includes: through the inertial measurement unit installed at the joints of the swimmer, the speed, single stroke propulsion distance, and number of strokes per unit time of the swimmer during the training cycle are collected in real time, and the actual measured values ​​of the swimmer's underwater speed, stroke length, and stroke frequency during the training cycle are obtained, and the actual measured values ​​of the swimmer's training intensity coefficient during the training cycle are matched in combination with a preset relationship comparison table between the underwater speed, stroke length, stroke frequency and the training intensity coefficient.

[0076] Get the measured values ​​of the swimmer's swimming distance and swimming time.

[0077] As a preferred solution, the swimmer's underwater speed can also be analyzed through video data of the swimmer during the training cycle, or the swimmer's underwater speed can be recorded through a pressure sensor array at the bottom of the swimming pool or a wearable waterproof GPS.

[0078] As a preferred solution, the swimmer's underwater stroke distance can also be analyzed through video markers.

[0079] As a preferred solution, the video data of the swimmer during the training cycle can be analyzed frame by frame by video analysis software to obtain the swimmer's stroke frequency.

[0080] In another specific embodiment, an infrared camera and waterproof reflective markers are used to capture the swimmer's underwater motion.

[0081] It should be noted that the greater the underwater speed, the longer the stroke distance, and the higher the stroke frequency, the greater the training intensity coefficient.

[0082] As a preferred solution, the swimmer's swimming distance can be obtained through distance markings at the side of the swimming pool or using a device such as a sports watch with a distance recording function.

[0083] As a preferred solution, a stopwatch or training management software can be used to obtain the swimmer's swimming time.

[0084] Furthermore, the specific working process of the training load monitoring module also includes: substituting the measured values ​​of the swimmer's training intensity coefficient, swimming distance and swimming time during the training cycle into the calculation formula Analyze swimmers' real-time training load ,in They represent the swimmer’s swimming distance, swimming time and training intensity coefficient during the training cycle. They represent the influencing factors corresponding to the preset unit swimming distance and unit swimming time respectively. Indicates the preset training intensity coefficient threshold, and provides AR real-time feedback and remote transmission of the swimmer's training load.

[0085] The swimmer's real-time training load is compared with the preset training load limit. If the swimmer's real-time training load exceeds the limit, a corresponding warning is issued.

[0086] It should be noted that the calculation formula for the swimmer's real-time training load uses a linear combination and a square root function. The linear combination is used to combine the effects of swimming distance and swimming time, and the square root function is used to adjust the influence of the training intensity coefficient so that its effect on the training load is smoother when the training intensity coefficient is high.

[0087] It should be noted that, in a feasible simulation process, the pre-set influencing factors corresponding to the unit swimming distance and the unit swimming time and the training intensity coefficient threshold are manually set at the beginning of the system operation. , the simulation results can be obtained. Please refer to Table 1 for details, which lists some representative data.

[0088] Table 1. Some of the collected swimming distances, swimming durations, training intensity coefficients and calculation results

[0089]

[0090] Through computational simulation, it was verified that as the swimming distance, swimming time and training intensity coefficient increased, the training load also increased accordingly, indicating that the calculation formula can reasonably reflect the swimmer's training load, and can effectively solve the problem and provide real-time training feedback and remote transmission.

[0091] As a preferred solution, when the real-time training load of a swimmer exceeds a limit, an early warning is given through a voice prompt device worn by the swimmer or a prompt message is projected through the AR goggles worn by the swimmer.

[0092] It should be noted that the present invention evaluates the swimmer's training load based on underwater speed, stroke length, stroke frequency, swimming distance and swimming time, and issues an over-limit warning, which is conducive to formulating a personalized training plan while avoiding overtraining.

[0093] The injury risk warning module monitors the swimmer's heart rate, blood oxygen saturation, duration of muscle activation, and angular velocity and angle of each joint movement in real time during the training cycle, determines whether the swimmer has an injury risk and triggers a warning.

[0094] Furthermore, the specific working process of the injury risk warning module includes: monitoring the heart rate and blood oxygen saturation of the swimmer in real time during the training period through the heart rate sensor and the blood oxygen sensor, setting the risk factors corresponding to each heart rate range and each blood oxygen saturation range, screening the risk factors corresponding to the swimmer's heart rate and blood oxygen saturation, and recording them as .

[0095] The duration of the swimmer's muscle activation and the angular velocity and angle of each joint activity during the training period are monitored in real time by electromyographic sensors and inertial measurement units, which are recorded as , Indicates The number of joints, .

[0096] Extract the swimming injury precursor information stored in the database, obtain the warning value of muscle activation duration and the limit value of angular velocity and angle of each joint activity, and record them as .

[0097] By calculating the formula Analyzing swimmers’ injury risk factors .

[0098] It should be noted that too high or too low heart rate and blood oxygen saturation may increase the risk of sports injuries.

[0099] It should be noted that excessively high angular velocity of joint movement may lead to joint injury, and the angular velocity of joint movement directly reflects the intensity of exercise and potential risk of injury.

[0100] It should be noted that abnormal angles of joint movement may lead to improper posture and injury. The angle of joint movement is an important indicator for evaluating sports posture and potential injury.

[0101] It should be noted that the activation durations of the swimmer's muscle groups are compared with each other to obtain the maximum activation duration of the muscle group, which is used as the muscle activation duration of the swimmer during the training cycle.

[0102] Furthermore, the specific working process of the injury risk warning module also includes: comparing the swimmer's injury risk coefficient with a preset injury risk coefficient threshold. If the swimmer's injury risk coefficient is greater than the corresponding threshold, the swimmer is at risk of injury and a warning is triggered.

[0103] As a preferred solution, when a swimmer is at risk of injury, a warning is given through a voice prompt device worn by the swimmer or a warning message is projected through AR goggles worn by the swimmer.

[0104] It should be noted that the present invention determines whether the swimmer is at risk of injury and triggers an early warning by monitoring the swimmer's heart rate, blood oxygen saturation, duration of muscle activation, and joint movement, thereby providing a timely early warning of the risk of injury and reducing the incidence of injury.

[0105] The database stores sample image libraries of various swimming styles and standard angles of various joints in various movements, and stores swimming injury precursor information.

[0106] The above contents are merely examples and explanations of the concept of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, they should all fall within the protection scope of the present invention.

Claims

1. A swimming underwater physical training effect analysis system based on training data, characterized in that: include: Multimodal data acquisition module: collects video data, motion data, and physiological data of swimmers in real time during the training cycle through underwater cameras, inertial measurement units, electromyography sensors, heart rate sensors, and blood oxygen sensors; Movement Standardization Evaluation Module: Identify the swimmer's swimming posture and obtain the angle of each joint and the activation time and RMS value of each muscle group in each movement, evaluate the consistency of joint angle changes and muscle activation synchronization of each movement, determine whether the swimmer's movement is standard and provide AR real-time feedback; Extract the standard angles of each joint under each action of various swimming styles stored in the database, and select the standard angles of each joint under each action corresponding to the swimming style of the swimmer; The angles of each joint in each action of the swimmer are compared with the corresponding standard angles to obtain the angle deviations of each joint in each action of the swimmer, and the maximum angle deviations of the joints in each action of the swimmer are further obtained and substituted into the calculation formula Analyze the consistency coefficient of joint angle changes in each action of swimmers , Indicates swimmer's The maximum angle deviation of the joints for each action, , Indicates The number of the action, , represents a natural constant; The activation time and RMS value of each muscle group under each action of the swimmer are compared with each other to obtain the maximum deviation of the muscle group activation time and the maximum deviation of the activation RMS value under each action of the swimmer, which are recorded as , and substitute it into the calculation formula Analyze the muscle activation synchronization coefficient of each swimmer's movements ,in The thresholds represent the preset muscle group activation time and the maximum deviation of the activation RMS value respectively; right and The weighted average value is calculated to analyze the normative evaluation index of each action of the swimmer, and the index is compared with the preset normative evaluation index threshold. If the normative evaluation index of a certain action is less than the threshold, the action is not normative, and the swimmer's non-normative actions are counted, projected in real time through the AR goggles worn by the swimmer, and transmitted to the remote mobile terminal device; Training load monitoring module: Based on the actual measured values ​​of the swimmer's underwater speed, stroke length, stroke frequency, swimming distance and swimming time during the training cycle, the module evaluates the swimmer's training load and issues an over-limit warning. Injury risk warning module: real-time monitoring of the swimmer's heart rate, blood oxygen saturation, duration of muscle activation, and angular velocity and angle of each joint movement during the training cycle, to determine whether the swimmer is at risk of injury and trigger a warning; Database: Stores sample image libraries of various swimming styles and standard angles of each joint in each action, as well as information on precursors to swimming injuries.

2. A swimming underwater physical training effect analysis system based on training data according to claim 1, characterized in that: The specific working process of the multimodal data acquisition module is as follows: According to the preset principles, several underwater high-frame rate cameras are deployed at set positions in the swimming pool, and a camera network is built with a head-mounted miniature camera integrated into the swimmer's swimming cap to collect real-time video data of the swimmer during the training cycle; Inertial measurement units are fixedly installed at each joint of the swimmer according to the preset principles, and electromyographic sensors are deployed at each muscle group of the swimmer to collect the swimmer's motion data during the training cycle in real time; Through the heart rate sensor and blood oxygen sensor integrated in the swimmer's wearable device, the swimmer's physiological data during the training cycle is collected in real time.

3. A swimming underwater physical training effect analysis system based on training data according to claim 2, characterized in that: The specific working process of the action normative evaluation module includes: S1: The optical flow method is used to detect the turning points of the stroke phase of the video data of the swimmer during the training period to extract key frames, obtain the key images of the swimmer during the training period, and compare them with the sample image library of various swimming styles stored in the database to identify the swimming style of the swimmer; S2: Obtain the shooting time points corresponding to each key image of the swimmer during the training cycle, and record them as the action switching time points during the training cycle; The inertial measurement units installed at the joints of the swimmer are used to obtain the three-dimensional acceleration data, angular velocity data, and posture data of each joint of the swimmer at each action switching time point during the training cycle. Combined with the key images of the swimmer during the training cycle, the graph convolutional network is used to reconstruct the three-dimensional posture and generate a three-dimensional skeleton model of each action of the swimmer. A reference line is selected in the three-dimensional skeleton model of the swimmer according to a preset principle, and an angle between each joint of the swimmer and the reference line in each action is obtained according to the three-dimensional skeleton model of each action of the swimmer, and the angle is recorded as the angle of each joint in each action of the swimmer; S3: The activation time and RMS value of each muscle group of the swimmer at each action switching time point during the training cycle are obtained through the electromyographic sensors deployed on each muscle group of the swimmer, and are recorded as the activation time and RMS value of each muscle group under each action of the swimmer.

4. The swimming underwater physical training effect analysis system based on training data according to claim 1, characterized in that: The specific working process of the training load monitoring module includes: Through the inertial measurement unit installed at the joints of the swimmer, the speed, single arm stroke distance, and number of arm strokes per unit time of the swimmer during the training cycle are collected in real time, and the measured values ​​of the swimmer's underwater speed, stroke length, and stroke frequency during the training cycle are obtained. Combined with the preset relationship comparison table between underwater speed, stroke length, stroke frequency and training intensity coefficient, the measured value of the swimmer's training intensity coefficient during the training cycle is matched; Get the measured values ​​of the swimmer's swimming distance and swimming time.

5. A swimming underwater physical training effect analysis system based on training data according to claim 4, characterized in that: The specific working process of the training load monitoring module also includes: Substitute the measured values ​​of the swimmer's training intensity coefficient, swimming distance and swimming time during the training cycle into the calculation formula Analyze swimmers' real-time training load ,in They represent the swimmer’s swimming distance, swimming time and training intensity coefficient during the training cycle. They represent the influencing factors corresponding to the preset unit swimming distance and unit swimming time respectively. Indicates the preset training intensity coefficient threshold, and provides AR real-time feedback and remote transmission of the swimmer's training load; The swimmer's real-time training load is compared with the preset training load limit. If the swimmer's real-time training load exceeds the limit, a corresponding warning is issued.

6. The swimming underwater physical training effect analysis system based on training data according to claim 1, characterized in that: The specific working process of the damage risk warning module includes: The heart rate sensor and blood oxygen saturation of the swimmers during the training period are monitored in real time by using a heart rate sensor and a blood oxygen sensor, and the risk factors corresponding to each heart rate range and each blood oxygen saturation range are set. The risk factors corresponding to the swimmers' heart rate and blood oxygen saturation are screened and recorded as ; The duration of the swimmer's muscle activation and the angular velocity and angle of each joint activity during the training period are monitored in real time by electromyographic sensors and inertial measurement units, which are recorded as , Indicates The number of joints, ; Extract the swimming injury precursor information stored in the database, obtain the warning value of muscle activation duration and the limit value of angular velocity and angle of each joint activity, and record them as ; By calculating the formula Analyzing swimmers’ injury risk factors .

7. A swimming underwater physical training effect analysis system based on training data according to claim 6, characterized in that: The specific working process of the damage risk warning module also includes: The swimmer's injury risk factor is compared with a preset injury risk factor threshold. If the swimmer's injury risk factor is greater than the corresponding threshold, the swimmer is at risk of injury and an early warning is triggered.

Images