Motion monitoring management method and system based on artificial intelligence

Abstract

The invention is suitable for the technical field of intelligent exercise monitoring, and provides an artificial intelligence-based exercise monitoring management method and system, and the method comprises the steps: obtaining a training reference video and a training process video of a trainer, analyzing whether the trainer has a target exercise action with the deviation degree exceeding a preset threshold value or not through combining the training reference video and the training process video, and if the target exercise action exists, executing the step 2; a training intensity index for the target athletic action is determined. According to the method, the training reference video and the training process video of the trainee are obtained, the deviation between the movement action and the standard action of the trainee is analyzed in combination with the artificial intelligence technology, the physical performance and heart rate change of the trainee are monitored in real time, and therefore the personalized training intensity index is generated.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent motion monitoring, and in particular relates to a motion monitoring management method and system based on artificial intelligence. Background Art

[0002] At present, with the rapid development of intelligent technology, intelligent motion monitoring technology is increasingly used in the fields of fitness, sports training, and rehabilitation therapy. Traditional motion monitoring methods usually rely on wearable devices or sensors to monitor exercise intensity and evaluate training effects by collecting physiological data of athletes or trainers (such as heart rate, gait, etc.) in real time. However, these methods are often limited to data collection and basic statistical analysis, lacking in-depth analysis and personalized adjustment of movement movements and physiological feedback. Existing technologies make it difficult to comprehensively and accurately evaluate the movement performance and physiological reactions of trainers, and provide targeted training suggestions for trainers of different body shapes and health conditions, especially in the precise adjustment between movement deviation and training intensity.

[0003] In addition, although the existing technology can collect the trainee's physiological data (such as heart rate fluctuations, acceleration, etc.), these data are mostly single-dimensional and lack monitoring of the quality of the trainee's action execution. Traditional training evaluation systems often focus on the monitoring of physiological data, such as heart rate, calorie consumption, etc., but lack a comprehensive evaluation system based on the combination of kinematics and physiological feedback. Although some advanced training systems can use motion capture technology or sensors to analyze the standardization of movements, they often only provide static or single-point feedback, making it difficult to comprehensively consider the dynamic changes and individual differences of the trainee throughout the training process. In addition, the existing technology lacks an effective correction mechanism for the relationship between the accuracy of sports movements and physiological responses, and therefore fails to achieve precise adjustments for different trainees. Summary of the invention

[0004] The purpose of the present invention is to provide a motion monitoring management method and system based on artificial intelligence, aiming to solve the problems raised in the background technology.

[0005] The present invention is implemented as follows: a motion monitoring and management method based on artificial intelligence, the method comprising: Obtaining a training reference video and a training process video of the trainee, and combining the two to analyze whether the trainee has a target movement action whose deviation exceeds a preset threshold, and if so, determining a training intensity index for the target movement action; According to the training reference video, the main limbs involved in the implementation of the target movement are clearly identified, and the standard human body model and the trainer human body model are extracted from the training reference video and the training process video respectively; The first correction factor is calculated by combining the differences between the standard human model and the trainer human model in the main limb parts; Acquire the heart rate fluctuation data of the trainee during the exercise, and obtain a second correction factor according to the difference between the target exercise action and the previous and subsequent non-target exercise actions; The training intensity index is adjusted in sequence by the first correction factor and the second correction factor to obtain a final training intensity index.

[0006] As a further limitation of the technical solution of the embodiment of the present invention, the steps of clarifying the main limb parts involved in implementing the target movement action according to the training reference video, and extracting the standard human body model and the trainer human body model from the training reference video and the training process video respectively include: Extract the standard movement corresponding to the target movement in the training reference video, interpret the standard movement and determine the main limb parts involved; Analyze the training reference video and the training process video, and select the standard frontal full-body images of the reference and the trainer; Based on artificial intelligence technology, a standard human body model and a trainer human body model are extracted from the standard frontal full-body image of the reference person and the standard frontal full-body image of the trainer respectively.

[0007] As a further limitation of the technical solution of the embodiment of the present invention, the step of calculating the first correction factor based on the difference between the standard human body model and the trainer human body model in the main limb parts includes: Extract the main limb parts data of the standard human body model and the trainer human body model, and mark them using posture estimation technology; The difference between the data of the main limb parts of the standard human body model and the trainer human body model is quantified based on artificial intelligence, and the difference is set as a first correction factor in numerical form.

[0008] As a further limitation of the technical solution of the embodiment of the present invention, the main limb part data refers to the geometric features of each limb part in the standard human body model or the trainer human body model, including length, circumference and joint angle.

[0009] As a further limitation of the technical solution of the embodiment of the present invention, the steps of obtaining the heart rate fluctuation data of the trainee during the exercise and deriving the second correction factor according to the difference between the target motion and the previous and subsequent non-target motions include: Acquire the heart rate fluctuation data of the trainee during the exercise, and extract the first local heart rate fluctuation data corresponding to the target exercise action time point; Extracting the second local heart rate fluctuation data within a preset time period before and after the target motion action time point; The difference between the first local heart rate fluctuation data and the second local heart rate fluctuation data is analyzed, and the difference value is set as the second correction factor.

[0010] As a further limitation of the technical solution of the embodiment of the present invention, the step of adjusting the training intensity index in sequence by the first correction factor and the second correction factor to obtain the final training intensity index includes: Obtaining a first correction factor, and applying it to the training intensity index, reducing the corresponding amplitude, so as to obtain a preliminary corrected training intensity index; A second correction factor is obtained and applied to the initially corrected training intensity index to increase the corresponding amplitude to obtain a final training intensity index.

[0011] A sports monitoring and management system based on artificial intelligence, the system comprises: a target sports action determination module, a human body model extraction module, a first correction factor generation module, a second correction factor generation module and a training intensity index correction module, wherein: The target motion action determination module is used to obtain the training reference video and the training process video of the trainee, and combine the two to analyze whether the trainee has a target motion action whose deviation exceeds a preset threshold. If so, determine the training intensity index for the target motion action; A human body model extraction module is used to identify the main limbs involved in implementing the target movement action according to the training reference video, and to extract a standard human body model and a trainer human body model from the training reference video and the training process video respectively; A first correction factor generation module is used to calculate the first correction factor by combining the differences between the standard human body model and the trainer human body model in the main limb parts; A second correction factor generating module is used to obtain the heart rate fluctuation data of the trainee during the exercise process, and to obtain a second correction factor according to the difference between the target exercise action and the preceding and following non-target exercise actions; The training intensity index correction module is used to adjust the training intensity index in sequence through a first correction factor and a second correction factor to obtain a final training intensity index.

[0012] As a further limitation of the technical solution of the embodiment of the present invention, the human body model extraction module specifically includes: A video interpretation unit, used for extracting the standard movement corresponding to the target movement in the training reference video, interpreting the standard movement and determining the main limb parts involved; An image selection unit, used for analyzing the training reference video and the training process video, and selecting standard frontal full-body images of the reference and the trainee; The human body model extraction unit is used to extract a standard human body model and a trainer human body model respectively from the standard frontal full-body image of the referenced person and the standard frontal full-body image of the trainer based on artificial intelligence technology.

[0013] As a further limitation of the technical solution of the embodiment of the present invention, the first correction factor generation module specifically includes: A data extraction unit, used to extract data of main limb parts of the standard human body model and the trainer human body model, and mark them using posture estimation technology; A first correction factor generating unit, for quantifying the difference between the data of the main limb parts of the standard human body model and the trainer human body model based on artificial intelligence, and setting the difference in numerical form as a first correction factor; The main limb part data refers to the geometric features of each limb part in the standard human body model or the trainer human body model, including length, circumference and joint angle.

[0014] As a further limitation of the technical solution of the embodiment of the present invention, the second correction factor generating module specifically includes: A first heart rate data extraction unit is used to obtain the heart rate fluctuation data of the trainee during the exercise, and extract the first local heart rate fluctuation data corresponding to the target exercise action time point; A second heart rate data extraction unit, used to extract second local heart rate fluctuation data within a preset time period before and after a target motion action time point; The second correction factor generating unit is used to analyze the difference between the first local heart rate fluctuation data and the second local heart rate fluctuation data, and set the difference value as the second correction factor.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The present invention obtains the trainee's training reference video and training process video, combines artificial intelligence technology to analyze the deviation between the trainee's movement and standard movement, and monitors the trainee's physical performance and heart rate changes in real time, thereby generating a personalized training intensity index. Specifically, the present invention calculates the first correction factor based on the difference between the trainee and the standard human body model in key limb parts, and at the same time obtains the second correction factor by analyzing the heart rate fluctuation data, doubly correcting the training intensity index to ensure that the training intensity is adapted to the trainee's actual physiological state, maximize the training effect, and reduce the risk of sports injuries. This method realizes the scientific adjustment of training intensity, has strong adaptability, and can provide accurate and personalized training plans for trainees of different body shapes and physical fitness levels. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A flowchart of a method provided by an embodiment of the present invention; Figure 2 A flowchart of obtaining a standard human body model and a trainer human body model in the method provided in an embodiment of the present invention; Figure 3 A flow chart of calculating a first correction factor in a method provided in an embodiment of the present invention; Figure 4 A flow chart of calculating a second correction factor in a method provided in an embodiment of the present invention; Figure 5 A flowchart of calculating the final training intensity index in the method provided in an embodiment of the present invention; Figure 6 An application architecture diagram of a system provided by an embodiment of the present invention; Figure 7 A structural block diagram of a human body model extraction module in a system provided by an embodiment of the present invention; Figure 8 A structural block diagram of a first correction factor generating module in a system provided by an embodiment of the present invention; Fig. 9 This is a structural block diagram of a second correction factor generating module in a system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0018] It is understood that the terms "first", "second", etc. used in this application may be used herein to describe various elements, but unless otherwise specified, these elements are not limited by these terms. These terms are only used to distinguish a first element from another element. For example, without departing from the scope of this application, a first xx script may be referred to as a second xx script, and similarly, a second xx script may be referred to as a first xx script.

[0019] Figure 1 A flow chart of a method provided by an embodiment of the present invention is shown.

[0020] Specifically, a training reference video and a training process video of a trainee are obtained, and the two are combined to analyze whether the trainee has a target motion action whose deviation exceeds a preset threshold, and if so, determine a training intensity index for the target motion action. The method specifically includes the following steps: Step S100, obtaining the training reference video and the training process video of the trainee, and combining the two to analyze whether the trainee has a target movement action whose deviation exceeds a preset threshold. If so, determining the training intensity index for the target movement action.

[0021] In the embodiment of the present invention, the trainees may be fitness enthusiasts who learn by themselves at home, beginners in the gym, and patients who need rehabilitation training at home. These people may lack professional guidance when exercising, and the main purpose of the technical solution of the present invention is to help them improve the quality and safety of exercise.

[0022] The specific methods for obtaining the trainer's training reference video and training process video include: first, users can record their own training process through a smartphone or camera, making sure to shoot from multiple angles to capture complete movement details. Second, demonstration videos can be provided, which are produced by professional coaches and show standard target movement movements for trainers to compare and refer to.

[0023] Preset thresholds can be based on sports science research or expert advice, usually derived from analyzing large amounts of sports data. These thresholds can be set to specific posture deviation ratios, such as angle or length deviations of key limb parts not exceeding 10%. Combining these two, it is analyzed whether the trainee has a target movement action that deviates from the preset threshold. The algorithm can compare the standard action and the trainee's movement data, use computer vision technology to detect deviations in real time, and generate corresponding feedback.

[0024] The formulation and calculation of the training intensity index can be based on multiple factors, including the trainee's physical fitness level, exercise type, difficulty of the target exercise, and physiological indicators such as heart rate measured during the exercise. The specific calculation formula may involve a scoring system for standard exercises, comparing the ideal performance of the target exercise with the trainee's actual performance, and then deriving the corresponding training intensity index.

[0025] The above technical solutions are based on the latest advances in the fields of computer vision, posture estimation and artificial intelligence analysis in existing technologies. These technologies are used for real-time data collection and analysis, making sports monitoring more accurate and personalized. These technologies have been widely used in smart fitness equipment, sports analysis software and wearable devices, and can provide scientific sports guidance for trainers.

[0026] Furthermore, the artificial intelligence-based motion monitoring management method further includes the following steps: Step S200, clarifying the main body parts involved in implementing the target movement action according to the training reference video, and extracting the standard human body model and the trainer human body model from the training reference video and the training process video respectively.

[0027] Specifically, Figure 2 A flow chart for obtaining a standard human body model and a trainer human body model is shown.

[0028] Specifically, the main limb parts involved in implementing the target movement action are clarified according to the training reference video, and the standard human body model and the trainer human body model are extracted from the training reference video and the training process video respectively, including the following steps: Step S201, extracting the standard movement corresponding to the target movement in the training reference video, interpreting the standard movement and determining the main limb parts involved; Step S202, analyzing the training reference video and the training process video, and selecting standard frontal full-body images of the reference subject and the trainer; Step S203: Based on artificial intelligence technology, a standard human body model and a trainer human body model are respectively extracted from the standard frontal full-body image of the referenced person and the standard frontal full-body image of the trainer.

[0029] In an embodiment of the present invention, when implementing the above technical solution, it is first necessary to extract standard sports movements from the training reference video, interpret these movements and determine the main limb parts involved. Standard sports movements are usually demonstrated by professional coaches or athletes, and biomechanical analysis is performed to ensure the standardization of the movements. Through computer vision technology (such as OpenCV), image data of each frame can be extracted from the video, and the key limb parts in the movement (such as shoulders, arms, legs, etc.) can be analyzed to provide a basis for subsequent analysis.

[0030] Next, the training reference video and the training process video are analyzed to select the standard frontal full-body images of the reference and trainer. This process relies on posture estimation technology, and common frameworks include OpenPose and MediaPipe. Through these technologies, the key points of the human body in the video (such as joints, skeletons, angles, etc.) can be automatically detected, and the standard frontal full-body images of the trainer and reference can be generated based on these key point information. Posture estimation technology can identify the position of each joint in real time, thereby accurately establishing the trainer's motion posture model.

[0031] Then, based on artificial intelligence technology, the system extracts the standard human body model and the trainer human body model from the training reference video and the training process video. This process usually uses deep learning algorithms (such as convolutional neural networks) to process the joint coordinates extracted from the video, and reconstruct the standard human body model and the trainer human body model through these coordinate data. The deep learning model automatically generates these models based on the pre-trained data set, thereby obtaining high-precision human motion capture results.

[0032] Finally, the system will analyze the differences between the standard human model and the trainer's human model, focusing on the geometric differences of the main limb parts (such as angles, lengths, widths, etc.). These differences will be quantified through biomechanical and computational geometry methods. For example, if the arm length in the standard human model is 20% longer than the trainer's arm, the system will calculate this deviation and provide feedback for subsequent training.

[0033] Furthermore, the artificial intelligence-based motion monitoring management method further includes the following steps: Step S300, calculating a first correction factor based on the differences between the standard human body model and the trainee human body model in the main limbs.

[0034] Specifically, Figure 3 A flow chart for calculating the first correction factor is shown.

[0035] The calculation of the first correction factor based on the difference between the standard human body model and the trainer human body model in the main limbs includes the following steps: Step S301, extracting data of main limb parts of the standard human body model and the trainer human body model, and marking them using posture estimation technology; Step S302, quantifying the difference between the data of the main limb parts of the standard human body model and the trainer human body model based on artificial intelligence, and setting the difference in numerical form as a first correction factor.

[0036] The main limb part data refers to the geometric features of each limb part in the standard human body model or the trainer human body model, including length, circumference and joint angle.

[0037] In the embodiment of the present invention, the implementation of the technical solution mainly relies on computer vision, artificial intelligence, posture estimation and deep learning technology. The following is a detailed description of the specific implementation steps, as well as the specific definition and calculation method of the degree of difference.

[0038] First, extract the main limb parts data from the standard human body model and the trainer human body model. Specifically, the standard human body model and the trainer human body model include the geometric features of each limb part, which usually refer to the length, circumference and joint angle of the limb part. For example, the length can be the length of the arm or leg, the circumference refers to the circumference of the thigh, upper arm, etc., and the joint angle refers to the angle of the knee joint, elbow joint, etc. during the action. These data provide the basis for the subsequent difference analysis.

[0039] During the implementation process, the coordinates of each key limb part are first marked through posture estimation technology. This process relies on advanced posture estimation frameworks, such as OpenPose, MediaPipe or HRNet, which can identify the joint positions of the human body in real time and mark the two-dimensional or three-dimensional coordinates of each part. Taking OpenPose as an example, this technology can identify the 18 main joints of the human body in the video (such as shoulders, knees, ankles, etc.). Through these coordinates, the system can construct the geometric features of each limb part and further generate parameter data for the standard human body model and the trainer human body model.

[0040] Next, the degree of difference between the standard human model and the trainer human model is quantified based on artificial intelligence (AI) technology. The degree of difference here refers to the specific deviation in geometric features such as length, circumference and joint angle. For example, the upper arm length in the standard human model is 60 cm, while the upper arm length of the trainer is 55 cm, so the difference in upper arm length is 5 cm. The calculated difference ratio is the ratio of the length of the standard model to the length of the trainer model, that is: 60 divided by 55, which is equal to 1.09, indicating that the trainer's upper arm is about 9% shorter than the standard model.

[0041] The degree of difference can also include the deviation of the joint angle. For example, in the standard action, the knee angle of a certain action is 150 degrees, while the knee angle of the trainer is 160 degrees when performing the same action. The difference is 10 degrees. This difference also needs to be quantified, usually expressed as the ratio of the standard angle to the trainer's angle. The calculation of the degree of difference can be obtained by this ratio or other appropriate formulas (such as Euclidean distance).

[0042] Once these differences are quantified, they need to be converted into a first correction factor. The first correction factor is a numerical parameter that adjusts the training intensity index by reflecting the difference between the standard mannequin and the trainer mannequin in the main body parts. Assuming that there are differences between the standard model and the trainer model in certain body parts (such as arms, legs, torso, etc.), the correction factor will be adjusted according to the size of the difference. Generally speaking, the greater the difference, the greater the correction factor. For example, if the arm length difference between the standard model and the trainer is large, the first correction factor will be appropriately increased, otherwise, the correction factor will be reduced.

[0043] When applying this correction factor, the system adjusts the training intensity index, aiming to adjust the intensity of the training according to the trainee's body shape and movement norms. For example, if the trainee's legs are short or the joint angle deviation is large, the adjustment of the training intensity index may cause the trainee to perform low-intensity training to avoid sports injuries.

[0044] Furthermore, the artificial intelligence-based motion monitoring management method further includes the following steps: Step S400, obtaining the heart rate fluctuation data of the trainee during the exercise, and deriving a second correction factor according to the difference between the target exercise action and the preceding and following non-target exercise actions.

[0045] Specifically, Figure 4 A flow chart for calculating the second correction factor is shown.

[0046] The process of obtaining the heart rate fluctuation data of the trainee during the exercise and obtaining the second correction factor according to the difference between the target exercise action and the preceding and following non-target exercise actions specifically includes the following steps: Step S401, obtaining the heart rate fluctuation data of the trainee during the exercise, and extracting the first local heart rate fluctuation data corresponding to the target exercise action time point; Step S402, extracting second local heart rate fluctuation data within a preset time period before and after the target exercise action time point; Step S403: Analyze the difference between the first local heart rate fluctuation data and the second local heart rate fluctuation data, and set the difference value as the second correction factor.

[0047] In an embodiment of the present invention, when implementing this technical solution, it is first necessary to use appropriate equipment to obtain the heart rate fluctuation data of the trainee during exercise. Common devices include heart rate belts, smart watches, fitness bracelets, etc., which can monitor the heart rate of the trainee in real time and record the data. Through optical sensors (such as PPG technology) or electrode sensors (such as ECG technology), the device can measure the number of heart beats per minute (BPM) and continuously record these data. The device will transmit the heart rate data to the relevant application or platform in real time to generate a heart rate fluctuation curve in order to analyze the changes in the heart rate of the trainee at different time points.

[0048] The preset time period refers to a selected period of time before and after the target movement, usually 5 to 10 seconds before and after the trainee performs a specific movement. This time period is set based on the principles of physiology and sports science, that is, during training, the body will have a certain preparation phase before the start of the movement, and will go through a recovery process after the movement. By analyzing the heart rate fluctuations during this period, the training intensity, body adaptability and recovery speed can be more accurately evaluated.

[0049] In order to analyze the difference between the first local heart rate fluctuation data and the second local heart rate fluctuation data, it is first necessary to extract data of two key time periods from the overall heart rate fluctuation curve. The first local data refers to the fluctuation data of the trainee's heart rate at the beginning of the target exercise action; the second local data refers to the heart rate fluctuation data in a preset time period before and after the target exercise action occurs. The data of these two time periods reflect the changes in the trainee's heart rate before and after performing a specific action.

[0050] During the specific analysis, the system will compare the heart rate data of these two time periods. First, extract the heart rate fluctuation curves before and after the target action time point, and calculate the difference between the two. For example, if the trainer's heart rate rises from 110BPM to 160BPM after performing a high-intensity action, the heart rate may gradually drop to 140BPM during the recovery period after the exercise. By calculating the change in heart rate before and after the target action (such as an increase from 110 to 160, or a decrease from 160 to 140), the system can get the difference between the two segments of data.

[0051] This difference value reflects the fluctuation range of the trainee's heart rate when performing the action, and can be used to quantify the training intensity. The larger the difference value, the higher the trainee's physical load and exercise intensity in the action, and the second correction factor will also increase accordingly, indicating that the training intensity index needs to be improved accordingly. If the difference value is small, it means that the trainee's load is low when performing the action, and accordingly, the correction factor may be smaller, and the training intensity index will also decrease accordingly.

[0052] For example, if the heart rate before the target exercise is 110BPM, it reaches 150BPM during exercise, and recovers to 120BPM after exercise, the heart rate change range is 40BPM. By analyzing this change, the system can determine the fluctuation range of the trainee's heart rate during this action, and then calculate the difference value (such as the ratio of the fluctuation range to the average heart rate or other appropriate algorithms), and use the difference value as the second correction factor to adjust the training intensity index.

[0053] Furthermore, the artificial intelligence-based motion monitoring management method further includes the following steps: Step S500: adjusting the training intensity index in sequence by using the first correction factor and the second correction factor to obtain a final training intensity index.

[0054] Specifically, Figure 5 A flow chart for calculating the final training intensity index is shown.

[0055] The training intensity index is adjusted in sequence by the first correction factor and the second correction factor to obtain the final training intensity index, which specifically includes the following steps: Step S501, obtaining a first correction factor, and applying it to the training intensity index, reducing the corresponding amplitude, so as to obtain a preliminary corrected training intensity index; Step S502, obtaining a second correction factor, and applying it to the initially corrected training intensity index, increasing the corresponding amplitude, so as to obtain a final training intensity index.

[0056] In the embodiment of the present invention, it is assumed that the trainee is doing high-intensity jumping training, wearing a heart rate monitoring device, and the system obtains its motion data through video analysis. The system first extracts the data of standard sports movements from the training reference video, and compares it with the real-time training process video of the trainee to analyze whether there are any irregular movements. If it is found that the trainee's movements deviate from the standard movements, the system will adjust the training intensity according to the difference in body models.

[0057] First, the system will calculate the first correction factor based on the limb data of the trainee and the standard human model (such as the length, angle and other geometric features of the arms, legs, torso, etc.). If the upper arm length of the standard human model is 60 cm, and the upper arm length of the trainee is 50 cm, the difference is 10 cm. This difference will be quantified as a correction factor, reflecting that the trainee's upper arm is 20% shorter than the standard model. This first correction factor will be used to reduce the training intensity index because the trainee's body shape is different from the standard model, and the exercise intensity needs to be appropriately reduced to avoid overtraining. For example, assuming the initial training intensity index is 10, after adjustment by the first correction factor, the intensity index is reduced to 8.

[0058] Next, the system will obtain the real-time heart rate fluctuation data of the trainee during the exercise and analyze the fluctuation range of the heart rate when the target exercise action occurs. Suppose that when performing a high-intensity jump, the trainee's heart rate rises from 110BPM to 150BPM, and then returns to 120BPM. The system compares the heart rate fluctuation data before and after the target action and calculates the difference in heart rate fluctuations. Assuming that this difference is 40BPM, this indicates that the trainee's heart rate fluctuates greatly when performing the action, indicating that the exercise intensity is high, so the system will use the second correction factor to adjust the training intensity index. Since the trainee's heart rate fluctuates greatly, indicating that the action is more difficult, the system will ensure the training effect by increasing the training intensity index. For example, after adjustment by the second correction factor, the training intensity index is increased from the initial correction of 8 to 10.

[0059] Ultimately, the system obtains a reasonable and personalized training intensity index through the interaction of these two correction factors. The first correction factor adapts to the physical differences of the trainees by reducing the training intensity, while the second correction factor further optimizes the intensity based on the feedback of heart rate fluctuations to ensure that the training is in line with the trainee's physical fitness level and can achieve effective training results.

[0060] Through this dual correction mechanism, the system can more accurately customize the training intensity for different types of trainees. On the one hand, the first correction factor takes into account the difference in body shape between the trainee and the standard model, avoiding the risk of injury caused by excessively high-intensity training; on the other hand, the second correction factor reflects the actual training intensity based on heart rate fluctuations, ensuring that the trainee can achieve the best training effect within a safe range. Combining these two correction factors, the training intensity can be dynamically adjusted according to the trainee's actual physiological response, making the training more scientific and reasonable, and maximizing the training effect.

[0061] Furthermore, Figure 6 The application architecture diagram of the system provided by the embodiment of the present invention is shown.

[0062] Among them, in another preferred embodiment provided by the present invention, a sports monitoring and management system based on artificial intelligence includes: The target motion action determination module 100 is used to obtain the trainee's training reference video and training process video, and combine the two to analyze whether the trainee has a target motion action whose deviation exceeds a preset threshold. If so, determine the training intensity index for the target motion action.

[0063] In the embodiment of the present invention, the trainees may be fitness enthusiasts who learn by themselves at home, beginners in the gym, and patients who need rehabilitation training at home. These people may lack professional guidance when exercising, and the main purpose of the technical solution of the present invention is to help them improve the quality and safety of exercise.

[0064] The specific methods for obtaining the trainer's training reference video and training process video include: first, users can record their own training process through a smartphone or camera, making sure to shoot from multiple angles to capture complete movement details. Second, demonstration videos can be provided, which are produced by professional coaches and show standard target movement movements for trainers to compare and refer to.

[0065] Preset thresholds can be based on sports science research or expert advice, usually derived from analyzing large amounts of sports data. These thresholds can be set to specific posture deviation ratios, such as angle or length deviations of key limb parts not exceeding 10%. Combining these two, it is analyzed whether the trainee has a target movement action that deviates from the preset threshold. The algorithm can compare the standard action and the trainee's movement data, use computer vision technology to detect deviations in real time, and generate corresponding feedback.

[0066] The formulation and calculation of the training intensity index can be based on multiple factors, including the trainee's physical fitness level, exercise type, difficulty of the target exercise, and physiological indicators such as heart rate measured during the exercise. The specific calculation formula may involve a scoring system for standard exercises, comparing the ideal performance of the target exercise with the trainee's actual performance, and then deriving the corresponding training intensity index.

[0067] The above technical solutions are based on the latest advances in the fields of computer vision, posture estimation and artificial intelligence analysis in existing technologies. These technologies are used for real-time data collection and analysis, making sports monitoring more accurate and personalized. These technologies have been widely used in smart fitness equipment, sports analysis software and wearable devices, and can provide scientific sports guidance for trainers.

[0068] Furthermore, the artificial intelligence-based motion monitoring and management system also includes: The human body model extraction module 200 is used to identify the main body parts involved in implementing the target movement action according to the training reference video, and to extract the standard human body model and the trainer human body model from the training reference video and the training process video respectively.

[0069] Specifically, Figure 7 It shows a structural block diagram of the human body model extraction module 200 in the system provided by the embodiment of the present invention.

[0070] Among them, in the preferred embodiment provided by the present invention, the human body model extraction module 200 specifically includes: The video interpretation unit 201 is used to extract the standard motion corresponding to the target motion in the training reference video, interpret the standard motion and determine the main body parts involved; An image selection unit 202 is used to analyze the training reference video and the training process video, and select standard frontal full-body images of the reference and the trainee; The human body model extraction unit 203 is used to extract a standard human body model and a trainer human body model from the standard frontal full-body image of the referenced person and the standard frontal full-body image of the trainer respectively based on artificial intelligence technology.

[0071] In an embodiment of the present invention, when implementing the above technical solution, it is first necessary to extract standard sports movements from the training reference video, interpret these movements and determine the main limb parts involved. Standard sports movements are usually demonstrated by professional coaches or athletes, and biomechanical analysis is performed to ensure the standardization of the movements. Through computer vision technology (such as OpenCV), image data of each frame can be extracted from the video, and the key limb parts in the movement (such as shoulders, arms, legs, etc.) can be analyzed to provide a basis for subsequent analysis.

[0072] Next, the training reference video and the training process video are analyzed to select the standard frontal full-body images of the reference and trainer. This process relies on posture estimation technology, and common frameworks include OpenPose and MediaPipe. Through these technologies, the key points of the human body in the video (such as joints, skeletons, angles, etc.) can be automatically detected, and the standard frontal full-body images of the trainer and reference can be generated based on these key point information. Posture estimation technology can identify the position of each joint in real time, thereby accurately establishing the trainer's motion posture model.

[0073] Then, based on artificial intelligence technology, the system extracts the standard human body model and the trainer human body model from the training reference video and the training process video. This process usually uses deep learning algorithms (such as convolutional neural networks) to process the joint coordinates extracted from the video, and reconstruct the standard human body model and the trainer human body model through these coordinate data. The deep learning model automatically generates these models based on the pre-trained data set, thereby obtaining high-precision human motion capture results.

[0074] Finally, the system will analyze the differences between the standard human model and the trainer's human model, focusing on the geometric differences of the main limb parts (such as angles, lengths, widths, etc.). These differences will be quantified through biomechanical and computational geometry methods. For example, if the arm length in the standard human model is 20% longer than the trainer's arm, the system will calculate this deviation and provide feedback for subsequent training.

[0075] Furthermore, the artificial intelligence-based motion monitoring and management system also includes: The first correction factor generation module 300 is used to calculate the first correction factor by combining the differences between the standard human body model and the trainer human body model in the main limb parts.

[0076] Specifically, Figure 8 It shows a structural block diagram of the first correction factor generating module 300 in the system provided by an embodiment of the present invention.

[0077] In a preferred embodiment of the present invention, the first correction factor generating module 300 specifically includes: A data extraction unit 301 is used to extract data of main limb parts of the standard human body model and the trainer human body model, and mark them using posture estimation technology; A first correction factor generating unit 302 is used to quantify the difference between the data of the main limb parts of the standard human body model and the trainer human body model based on artificial intelligence, and set the difference in numerical form as a first correction factor; The main limb part data refers to the geometric features of each limb part in the standard human body model or the trainer human body model, including length, circumference and joint angle.

[0078] In the embodiment of the present invention, the implementation of the technical solution mainly relies on computer vision, artificial intelligence, posture estimation and deep learning technology. The following is a detailed description of the specific implementation steps, as well as the specific definition and calculation method of the degree of difference.

[0079] First, extract the main limb parts data from the standard human body model and the trainer human body model. Specifically, the standard human body model and the trainer human body model include the geometric features of each limb part, which usually refer to the length, circumference and joint angle of the limb part. For example, the length can be the length of the arm or leg, the circumference refers to the circumference of the thigh, upper arm, etc., and the joint angle refers to the angle of the knee joint, elbow joint, etc. during the action. These data provide the basis for the subsequent difference analysis.

[0080] During the implementation process, the coordinates of each key limb part are first marked through posture estimation technology. This process relies on advanced posture estimation frameworks, such as OpenPose, MediaPipe or HRNet, which can identify the joint positions of the human body in real time and mark the two-dimensional or three-dimensional coordinates of each part. Taking OpenPose as an example, this technology can identify the 18 main joints of the human body in the video (such as shoulders, knees, ankles, etc.). Through these coordinates, the system can construct the geometric features of each limb part and further generate parameter data for the standard human body model and the trainer human body model.

[0081] Next, the degree of difference between the standard human model and the trainer human model is quantified based on artificial intelligence (AI) technology. The degree of difference here refers to the specific deviation in geometric features such as length, circumference and joint angle. For example, the upper arm length in the standard human model is 60 cm, while the upper arm length of the trainer is 55 cm, so the difference in upper arm length is 5 cm. The calculated difference ratio is the ratio of the length of the standard model to the length of the trainer model, that is: 60 divided by 55, which is equal to 1.09, indicating that the trainer's upper arm is about 9% shorter than the standard model.

[0082] The degree of difference can also include the deviation of the joint angle. For example, in the standard action, the knee angle of a certain action is 150 degrees, while the knee angle of the trainer is 160 degrees when performing the same action. The difference is 10 degrees. This difference also needs to be quantified, usually expressed as the ratio of the standard angle to the trainer's angle. The calculation of the degree of difference can be obtained by this ratio or other appropriate formulas (such as Euclidean distance).

[0083] Once these differences are quantified, they need to be converted into a first correction factor. The first correction factor is a numerical parameter that adjusts the training intensity index by reflecting the difference between the standard mannequin and the trainer mannequin in the main body parts. Assuming that there are differences between the standard model and the trainer model in certain body parts (such as arms, legs, torso, etc.), the correction factor will be adjusted according to the size of the difference. Generally speaking, the greater the difference, the greater the correction factor. For example, if the arm length difference between the standard model and the trainer is large, the first correction factor will be appropriately increased, otherwise, the correction factor will be reduced.

[0084] When applying this correction factor, the system adjusts the training intensity index, aiming to adjust the intensity of the training according to the trainee's body shape and movement norms. For example, if the trainee's legs are short or the joint angle deviation is large, the adjustment of the training intensity index may cause the trainee to perform low-intensity training to avoid sports injuries.

[0085] Furthermore, the artificial intelligence-based motion monitoring and management system also includes: The second correction factor generation module 400 is used to obtain the heart rate fluctuation data of the trainee during the exercise process, and to obtain the second correction factor according to the difference between the target motion and the preceding and following non-target motions.

[0086] Specifically, Fig. 9 It shows a structural block diagram of the second correction factor generating module 400 in the system provided by an embodiment of the present invention.

[0087] In a preferred embodiment of the present invention, the second correction factor generating module 400 specifically includes: The first heart rate data extraction unit 401 is used to obtain the heart rate fluctuation data of the trainee during the exercise process, and extract the first local heart rate fluctuation data corresponding to the target exercise action time point; The second heart rate data extraction unit 402 is used to extract the second local heart rate fluctuation data within a preset time period before and after the target motion action time point; The second correction factor generating unit 403 is used to analyze the difference between the first local heart rate fluctuation data and the second local heart rate fluctuation data, and set the difference value as the second correction factor.

[0088] In an embodiment of the present invention, when implementing this technical solution, it is first necessary to use appropriate equipment to obtain the heart rate fluctuation data of the trainee during exercise. Common devices include heart rate belts, smart watches, fitness bracelets, etc., which can monitor the heart rate of the trainee in real time and record the data. Through optical sensors (such as PPG technology) or electrode sensors (such as ECG technology), the device can measure the number of heart beats per minute (BPM) and continuously record these data. The device will transmit the heart rate data to the relevant application or platform in real time to generate a heart rate fluctuation curve in order to analyze the changes in the heart rate of the trainee at different time points.

[0089] The preset time period refers to a selected period of time before and after the target movement, usually 5 to 10 seconds before and after the trainee performs a specific movement. This time period is set based on the principles of physiology and sports science, that is, during training, the body will have a certain preparation phase before the start of the movement, and will go through a recovery process after the movement. By analyzing the heart rate fluctuations during this period, the training intensity, body adaptability and recovery speed can be more accurately evaluated.

[0090] In order to analyze the difference between the first local heart rate fluctuation data and the second local heart rate fluctuation data, it is first necessary to extract data of two key time periods from the overall heart rate fluctuation curve. The first local data refers to the fluctuation data of the trainee's heart rate at the beginning of the target exercise action; the second local data refers to the heart rate fluctuation data in a preset time period before and after the target exercise action occurs. The data of these two time periods reflect the changes in the trainee's heart rate before and after performing a specific action.

[0091] During the specific analysis, the system will compare the heart rate data of these two time periods. First, extract the heart rate fluctuation curves before and after the target action time point, and calculate the difference between the two. For example, if the trainer's heart rate rises from 110BPM to 160BPM after performing a high-intensity action, the heart rate may gradually drop to 140BPM during the recovery period after the exercise. By calculating the change in heart rate before and after the target action (such as an increase from 110 to 160, or a decrease from 160 to 140), the system can get the difference between the two segments of data.

[0092] This difference value reflects the fluctuation range of the trainee's heart rate when performing the action, and can be used to quantify the training intensity. The larger the difference value, the higher the trainee's physical load and exercise intensity in the action, and the second correction factor will also increase accordingly, indicating that the training intensity index needs to be improved accordingly. If the difference value is small, it means that the trainee's load is low when performing the action, and accordingly, the correction factor may be smaller, and the training intensity index will also decrease accordingly.

[0093] For example, if the heart rate before the target exercise is 110BPM, it reaches 150BPM during exercise, and recovers to 120BPM after exercise, the heart rate change range is 40BPM. By analyzing this change, the system can determine the fluctuation range of the trainee's heart rate during this action, and then calculate the difference value (such as the ratio of the fluctuation range to the average heart rate or other appropriate algorithms), and use the difference value as the second correction factor to adjust the training intensity index.

[0094] Furthermore, the artificial intelligence-based motion monitoring and management system also includes: The training intensity index correction module 500 is used to adjust the training intensity index in sequence by using a first correction factor and a second correction factor to obtain a final training intensity index.

[0095] In the embodiment of the present invention, it is assumed that the trainee is doing high-intensity jumping training, wearing a heart rate monitoring device, and the system obtains its motion data through video analysis. The system first extracts the data of standard sports movements from the training reference video, and compares it with the real-time training process video of the trainee to analyze whether there are any irregular movements. If it is found that the trainee's movements deviate from the standard movements, the system will adjust the training intensity according to the difference in body models.

[0096] First, the system will calculate the first correction factor based on the limb data of the trainee and the standard human model (such as the length, angle and other geometric features of the arms, legs, torso, etc.). If the upper arm length of the standard human model is 60 cm, and the upper arm length of the trainee is 50 cm, the difference is 10 cm. This difference will be quantified as a correction factor, reflecting that the trainee's upper arm is 20% shorter than the standard model. This first correction factor will be used to reduce the training intensity index because the trainee's body shape is different from the standard model, and the exercise intensity needs to be appropriately reduced to avoid overtraining. For example, assuming the initial training intensity index is 10, after adjustment by the first correction factor, the intensity index is reduced to 8.

[0097] Next, the system will obtain the real-time heart rate fluctuation data of the trainee during the exercise and analyze the fluctuation range of the heart rate when the target exercise action occurs. Suppose that when performing a high-intensity jump, the trainee's heart rate rises from 110BPM to 150BPM, and then returns to 120BPM. The system compares the heart rate fluctuation data before and after the target action and calculates the difference in heart rate fluctuations. Assuming that this difference is 40BPM, this indicates that the trainee's heart rate fluctuates greatly when performing the action, indicating that the exercise intensity is high, so the system will use the second correction factor to adjust the training intensity index. Since the trainee's heart rate fluctuates greatly, indicating that the action is more difficult, the system will ensure the training effect by increasing the training intensity index. For example, after adjustment by the second correction factor, the training intensity index is increased from the initial correction of 8 to 10.

[0098] Ultimately, the system obtains a reasonable and personalized training intensity index through the interaction of these two correction factors. The first correction factor adapts to the physical differences of the trainees by reducing the training intensity, while the second correction factor further optimizes the intensity based on the feedback of heart rate fluctuations to ensure that the training is in line with the trainee's physical fitness level and can achieve effective training results.

[0099] Through this dual correction mechanism, the system can more accurately customize the training intensity for different types of trainees. On the one hand, the first correction factor takes into account the difference in body shape between the trainee and the standard model, avoiding the risk of injury caused by excessively high-intensity training; on the other hand, the second correction factor reflects the actual training intensity based on heart rate fluctuations, ensuring that the trainee can achieve the best training effect within a safe range. Combining these two correction factors, the training intensity can be dynamically adjusted according to the trainee's actual physiological response, making the training more scientific and reasonable, and maximizing the training effect.

[0100] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0101] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0102] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0103] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

[0104] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A sports monitoring and management method based on artificial intelligence, characterized in that: The method comprises: Obtaining a training reference video and a training process video of the trainee, and combining the two to analyze whether the trainee has a target movement action whose deviation exceeds a preset threshold, and if so, determining a training intensity index for the target movement action; According to the training reference video, the main limbs involved in the implementation of the target movement are clearly identified, and the standard human body model and the trainer human body model are extracted from the training reference video and the training process video respectively; The first correction factor is calculated by combining the differences between the standard human body model and the trainer human body model in the main limb parts; Acquire the heart rate fluctuation data of the trainee during the exercise, and obtain a second correction factor according to the difference between the target exercise action and the previous and subsequent non-target exercise actions; The training intensity index is adjusted in sequence by the first correction factor and the second correction factor to obtain a final training intensity index.

2. The artificial intelligence-based sports monitoring and management method according to claim 1, characterized in that: The steps of clarifying the main limb parts involved in implementing the target movement action according to the training reference video, and extracting the standard human body model and the trainer human body model from the training reference video and the training process video respectively include: Extract the standard movement corresponding to the target movement in the training reference video, interpret the standard movement and determine the main limb parts involved; Analyze the training reference video and the training process video, and select the standard frontal full-body images of the reference and the trainer; Based on artificial intelligence technology, a standard human body model and a trainer human body model are extracted from the standard frontal full-body image of the reference person and the standard frontal full-body image of the trainer respectively.

3. The motion monitoring and management method based on artificial intelligence according to claim 2 is characterized in that: The steps of calculating the first correction factor based on the differences between the standard human body model and the trainer human body model in the main limb parts include: Extract the main limb parts data of the standard human body model and the trainer human body model, and mark them using posture estimation technology; The difference between the data of the main limb parts of the standard human body model and the trainer human body model is quantified based on artificial intelligence, and the difference is set as a first correction factor in numerical form.

4. The method for sports monitoring and management based on artificial intelligence according to claim 3 is characterized in that: The main limb part data refers to the geometric features of each limb part in the standard human body model or the trainer human body model, including length, circumference and joint angle.

5. The motion monitoring and management method based on artificial intelligence according to claim 1 is characterized in that: The steps of obtaining the heart rate fluctuation data of the trainee during the exercise and deriving the second correction factor according to the difference between the target exercise action and the preceding and following non-target exercise actions include: Acquire the heart rate fluctuation data of the trainee during the exercise, and extract the first local heart rate fluctuation data corresponding to the target exercise action time point; Extracting the second local heart rate fluctuation data within a preset time period before and after the target motion action time point; The difference between the first local heart rate fluctuation data and the second local heart rate fluctuation data is analyzed, and the difference value is set as the second correction factor.

6. The method for sports monitoring and management based on artificial intelligence according to claim 1, characterized in that: The step of sequentially adjusting the training intensity index by the first correction factor and the second correction factor to obtain the final training intensity index includes: Obtaining a first correction factor, and applying it to the training intensity index, reducing the corresponding amplitude, so as to obtain a preliminary corrected training intensity index; A second correction factor is obtained and applied to the initially corrected training intensity index to increase the corresponding amplitude to obtain a final training intensity index.

7. A sports monitoring and management system based on artificial intelligence, characterized in that: The system comprises: a target motion determination module, a human body model extraction module, a first correction factor generation module, a second correction factor generation module and a training intensity index correction module, wherein: The target motion action determination module is used to obtain the training reference video and the training process video of the trainee, and combine the two to analyze whether the trainee has a target motion action whose deviation exceeds a preset threshold. If so, determine the training intensity index for the target motion action; A human body model extraction module is used to identify the main limbs involved in implementing the target movement action according to the training reference video, and to extract a standard human body model and a trainer human body model from the training reference video and the training process video respectively; A first correction factor generation module is used to calculate the first correction factor by combining the differences between the standard human body model and the trainer human body model in the main limb parts; A second correction factor generating module is used to obtain the heart rate fluctuation data of the trainee during the exercise process, and to obtain a second correction factor according to the difference between the target exercise action and the preceding and following non-target exercise actions; The training intensity index correction module is used to adjust the training intensity index in sequence through a first correction factor and a second correction factor to obtain a final training intensity index.

8. The artificial intelligence-based sports monitoring and management system according to claim 7, characterized in that: The human body model extraction module specifically includes: A video interpretation unit, used for extracting the standard movement corresponding to the target movement in the training reference video, interpreting the standard movement and determining the main limb parts involved; An image selection unit, used for analyzing the training reference video and the training process video, and selecting standard frontal full-body images of the reference and the trainee; The human body model extraction unit is used to extract a standard human body model and a trainer human body model respectively from the standard frontal full-body image of the referenced person and the standard frontal full-body image of the trainer based on artificial intelligence technology.

9. The artificial intelligence-based sports monitoring and management system according to claim 8, characterized in that: The first correction factor generation module specifically includes: A data extraction unit, used to extract data of main limb parts of the standard human body model and the trainer human body model, and mark them using posture estimation technology; A first correction factor generating unit, for quantifying the difference between the data of the main limb parts of the standard human body model and the trainer human body model based on artificial intelligence, and setting the difference in numerical form as a first correction factor; The main limb part data refers to the geometric features of each limb part in the standard human body model or the trainer human body model, including length, circumference and joint angle.

10. The artificial intelligence-based sports monitoring and management system according to claim 9, characterized in that: The second correction factor generation module specifically includes: A first heart rate data extraction unit is used to obtain the heart rate fluctuation data of the trainee during the exercise, and extract the first local heart rate fluctuation data corresponding to the target exercise action time point; A second heart rate data extraction unit, used to extract second local heart rate fluctuation data within a preset time period before and after a target motion action time point; The second correction factor generating unit is used to analyze the difference between the first local heart rate fluctuation data and the second local heart rate fluctuation data, and set the difference value as the second correction factor.

Images