Intelligent physical education real-time feedback system and method based on motion perception

Abstract

The invention relates to the technical field of motion perception, in particular to an intelligent physical education real-time feedback system and method based on motion perception. The device comprises a joint capturing unit, a deviation coordination unit and an adjustment feedback unit. An angle judgment module calculates an upper arm vector and a front arm vector of an arm part, the dot product of the upper arm vector and the front arm vector is calculated according to the upper arm vector included component of the arm part and the front arm vector included component of the arm part, and the component of the upper arm vector of the arm part is used for calculating the modulus of the upper arm vector; calculating a model of a forearm vector by using a component of the forearm vector of the arm part, calculating an action angle of a joint by combining an upper arm vector and a forearm vector dot product with the model of the upper arm vector and the model of the forearm vector, judging whether the captured whole-body action of the student is a common activity or a movement, and calculating the whole-body action of the student through vector calculation and angle derivation. The action amplitude and speed of the joints can be accurately quantified, the error of subjective judgment is avoided, and the accuracy of analyzing the actions of students is improved.

Description

Technical Field

[0001] The present invention relates to the field of motion perception technology, and in particular to an intelligent physical education teaching real-time feedback system and method based on motion perception. Background Art

[0002] Intelligent physical education real-time feedback is the product of the combination of traditional physical education and modern technology. It solves the problems of feedback lag and lack of personalization in traditional teaching through sensors, artificial intelligence and big data technology. In online physical education, when students move too fast or with too large an amplitude, the quantification results are inaccurate. At the same time, when multiple joints move simultaneously and affect each other, it is difficult to accurately separate the motion data of each joint, affecting the quantification accuracy and reducing the accuracy of the quantified joint movement amplitude and speed, thereby reducing the accuracy of analyzing students' movements. Therefore, we provide a real-time feedback system and method for intelligent physical education based on motion perception. Summary of the Invention

[0003] The purpose of the present invention is to provide an intelligent physical education real-time feedback system and method based on motion perception to solve the problems raised in the above background technology, including:

[0004] 1. Since the quantification results may be inaccurate when the student's movement speed is too fast or the amplitude is too large, this case uses a data splitting method to split the movement data of different parts into n joint points, record the position coordinates of the joint points to obtain the movement amplitude deviation, joint movement speed and joint movement angle. Through vector calculation and angle deduction, the movement amplitude and speed of the joint can be accurately quantified, avoiding the inaccurate quantification results caused by the student's movement speed being too fast or the amplitude being too large;

[0005] 2. When multiple joints move simultaneously and influence each other, the accuracy of quantifying the joint's movement amplitude and speed is reduced, resulting in a decrease in the accuracy of analyzing student movements. Therefore, this case calculates the joint's movement angle by combining the dot product of the upper arm vector and the forearm vector with the modulus of the upper arm vector and the forearm vector. This can accurately quantify the joint's movement amplitude and speed, avoid errors in subjective judgment, and improve the accuracy of analyzing student movements.

[0006] To achieve the above objectives, one of the objectives of the present invention is to provide an intelligent physical education real-time feedback system based on motion perception, comprising a joint capture unit, a deviation coordination unit and an adjustment feedback unit;

[0007] As a further improvement to this technical solution, the capture and splitting module captures the student's whole-body movements and splits the whole-body movement data into movement data of different parts. After splitting the whole-body movement data, the movement details of each part of the student's body can be clearly observed, such as the arm extension angle and leg bending degree. Precise guidance can be given to specific parts, allowing students to clearly know which movements need improvement, avoiding the ambiguity of general guidance and improving the pertinence of teaching.

[0008] The data splitting method is used to split the motion data of different parts into n joints, record the position coordinates of the joints, and then analyze the maximum motion amplitude value and the minimum motion amplitude value of the joints to obtain the motion amplitude deviation. By calculating the motion amplitude deviation, it can be clearly judged whether the student or athlete's movements meet the standards;

[0009] As a further improvement of the present technical solution, the joint velocity module collects the position coordinates of joint i at time t through the capture and splitting module to analyze the motion trajectory of the part. By recording the joint position coordinates, the displacement distance of the joint can be accurately calculated, the motion amplitude can be quantified, and the movement speed of the joint can be accurately quantified.

[0010] The position coordinates of joint i at time t-1 are extracted from the historical data of the capture and splitting module, and the time interval is calculated using time t and time t-1. The movement speed of the joint is then calculated based on the position coordinates of joint i at time t and the position coordinates of joint i at time t-1. By calculating this speed, the student's performance in a specific movement can be accurately understood. For example, in a sprint event, the movement speeds of the hip, knee, and ankle joints directly affect the running frequency and stride length. Accurately calculating the speeds of these joints helps to evaluate the athlete's starting, mid-run, and sprinting abilities.

[0011] As a further improvement of this technical solution, the capture joint unit further includes an angle judgment module;

[0012] The angle judgment module calculates the upper arm vector of the arm part and the forearm vector of the arm part through the arm motion data in the capture and splitting module, calculates the dot product of the upper arm vector and the forearm vector through the components of the upper arm vector of the arm part and the components of the forearm vector of the arm part, calculates the modulus of the upper arm vector using the components of the upper arm vector of the arm part, and then calculates the modulus of the forearm vector using the components of the forearm vector of the arm part, calculates the motion angle of the joint through the dot product of the upper arm vector and the forearm vector and the modulus of the upper arm vector and the modulus |v| of the forearm vector, and through vector calculation and angle deduction, the motion amplitude and speed of the joint can be accurately quantified, thereby avoiding errors in subjective judgment and improving the accuracy of analyzing students' movements.

[0013] A second object of the present invention is to provide a method for operating the intelligent physical education real-time feedback system based on motion perception described above, comprising the following steps:

[0014] S1. The joint capture unit perceives students' movements in real time during online physical education, captures students' whole-body movements to split them into n joints, and then calculates the movement amplitude deviation, joint movement speed, and joint movement angle to determine whether the captured student's whole-body movements are normal activities or sports;

[0015] S2, the deviation coordination unit obtains the joint action angle and joint action speed from the capture joint unit to analyze the joint action angle deviation, joint action speed peak value and student part action coordination value to determine whether the student is doing gymnastics;

[0016] S3. The adjustment feedback unit determines whether the gymnastics movements of different parts of the student need to be adjusted by comparing the movement angles of the joints. When adjustment is required, feedback information is sent to the student end.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] 1. In the intelligent physical education real-time feedback system and method based on motion perception, the angle judgment module calculates the upper arm vector of the arm part and the forearm vector of the arm part, calculates the dot product of the upper arm vector and the forearm vector by using the components of the upper arm vector of the arm part and the components of the forearm vector of the arm part, calculates the modulus of the upper arm vector by using the components of the upper arm vector of the arm part, and then calculates the modulus of the forearm vector by using the components of the forearm vector of the arm part, calculates the movement angle of the joint by combining the dot product of the upper arm vector and the forearm vector with the modulus of the upper arm vector and the modulus of the forearm vector, judges whether the captured whole-body movement of the student is a normal activity or exercise, and can accurately quantify the movement amplitude and speed of the joint through vector calculation and angle deduction, avoids errors in subjective judgment, and improves the accuracy of analyzing student movements.

[0019] 2. In the intelligent physical education real-time feedback system and method based on motion perception, the motion coordination module obtains the average value of the forearm joint motion angle according to the motion angle of the forearm joint, and then calculates the standard deviation of the forearm joint motion angle. The motion angle of the upper arm joint is used to obtain the average value of the upper arm joint motion angle, and then calculates the standard deviation of the upper arm joint motion angle. The motion angle of the forearm joint, the motion angle of the upper arm joint, the standard deviation of the forearm joint motion angle and the standard deviation of the upper arm joint motion angle are combined to calculate the student's part motion coordination value to determine whether the student is doing gymnastics. By monitoring the motion coordination, it is ensured that the student's movements meet the standards, avoid sports injuries caused by uncoordinated movements, reduce sports risks, and improve the safety of students during exercise. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is the overall structural diagram of the present invention;

[0021] Figure 2 This is a block diagram of the module unit of the present invention.

[0022] The meaning of each number in the figure is:

[0023] 10. Capture joint unit; 101. Capture splitting module; 102. Joint speed module; 103. Angle judgment module;

[0024] 20. Deviation coordination unit; 201. Deviation peak module; 202. Action coordination module; 30. Adjustment feedback unit. DETAILED DESCRIPTION

[0025] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making any creative efforts shall fall within the scope of protection of the present invention.

[0026] Example 1

[0027] One of the purposes of this embodiment is to provide an intelligent physical education real-time feedback system based on motion perception, including a joint capture unit 10, a deviation coordination unit 20 and an adjustment feedback unit 30;

[0028] The following is a refinement of the above units, see Figure 1-Figure 2 As shown;

[0029] The joint capture unit 10 senses the student's movements in real time during online physical education, collects historical data, and captures the student's whole body movements to split them into n joints. The position coordinates of the joints are recorded to derive the movement amplitude deviation, joint movement speed, and joint movement angle, and determine whether the captured student's whole body movements are normal activities or sports.

[0030] The capture joint unit 10 includes a capture splitting module 101 and a joint velocity module 102;

[0031] The capture and splitting module 101 uses a camera to perceive the student's movements in online physical education in real time, and captures the student's whole-body movements and collects historical data. When the camera fails to capture the student's whole-body movement data within the set time range threshold, the "failure to capture the student's whole-body movement data" message is fed back to the teacher in real time, and then the "please stand within the camera range as soon as possible" message is fed back to the student. When the camera successfully captures the student's whole-body movement data, since the student's whole-body movement data usually contains the position information of the whole-body joints (such as three-dimensional coordinates), in order to analyze the movements of various parts of the body in the student's whole-body movement data, the student's whole-body movement data needs to be split into movement data of different parts (such as arms);

[0032] The historical data includes the position coordinates P of joint i at time t-1 i (t-1), joint angle θ of standard gymnastics movement g (θ g The subscript “g” in has no special physical meaning and is used for distinction);

[0033] We can group joints into parts based on their anatomical relationships:

[0034] Arm: shoulder joint, elbow joint, wrist joint;

[0035] Use the data splitting method to split the action data of different parts into n joint points, and record the position coordinates P of each joint point at time t i =(x i ,y i ,z i ), where i = 1, 2, ..., n, the unit is meter (m), according to the position coordinates P of the joint point i =(x i ,y i ,z i )Analyze the maximum motion amplitude of the joint point max(P i (t)) and the minimum motion amplitude value min(P i (t)), the arm motion data can be expressed as: Arm = {P shoulder ,P elbow ,P wrist}, P shoulder is the position coordinate of the shoulder joint, usually expressed as (x shoulder ,y shoulder ,z shoulder ) indicates that P elbow is the position coordinate of the elbow joint point, usually expressed as (x elbow ,y elbow ,z elbow ) indicates that P wrist is the position coordinate of the wrist joint point, usually expressed as (x wrist,y wrist ,z wrist )express;

[0036] Analyze the implementation process of the maximum motion amplitude value and the minimum motion amplitude value of the joint point:

[0037] First, record the position coordinates P of the joint point i =(x i ,y i ,z i ) The amplitude of the movement at time t; for example, for the position coordinate P of a joint point i =(x i ,y i ,z i ), at t1, t2, ..., t k Record the corresponding P at each time i (t1), P i (t2), ..., P i (t k );

[0038] Secondly, the position coordinates P of the joint point i =(x i ,y i ,z i ) The action amplitude value at time t1 is given the maximum and minimum values at the same time, that is, max temp =P i (t1), min temp =P i (t1), then record the position coordinates P of the joint point i =(x i ,y i ,z i ) the motion amplitude value at time t2, and comparing each motion amplitude value with the current maximum motion amplitude value and minimum motion amplitude value in turn;

[0039] If the current action amplitude value is greater than max temp , then the updated max temp is the current maximum movement amplitude value;

[0040] If the current action amplitude value is less than min temp , then update min temp is the current minimum action amplitude value;

[0041] Finally, after traversing all the action amplitude values, max temp That is max(P i (t)), min temp That is min(P i(t)), using the maximum motion amplitude value max(P i (t)) and the minimum motion amplitude value min(P i (t)) to obtain the action amplitude deviation ΔP i =max(P i (t))-min(P i (t));

[0042] Since the position of the joints will change over time, the joint velocity module 102 collects the position coordinates P of the joint i at time t through the capture and split module 101. i (t)=(x i (t),y i (t),z i (t)), according to the position coordinates P of joint i at time t i (t)=(x i (t),y i (t),z i (t)) analyzes the motion trajectory of the parts, and then extracts the position coordinates P of joint i at time t-1 by capturing the historical data of the splitting module 101 i (t-1), using time t and time t-1 to calculate the time interval Δt=t-(t-1);

[0043] Since the movement speed of the joint is the rate of change of the joint position over time, according to the position coordinate P of joint i at time t i (t) and the position coordinates P of joint i at time t-1 i (t-1) Calculate the joint movement speed v i (t), its algorithm formula is Among them, the unit of time t and t-1 is second (s), the movement speed v i The unit of (t) is meter per second (m / s);

[0044] The capture joint unit 10 further includes an angle determination module 103;

[0045] The angle determination module 103 receives the shoulder joint coordinates P of the arm part from the capture and separation module 101. shoulder =(x shoulder ,y shoulder ,z shoulder ), coordinates of the elbow joint P elbow =(x elbow ,y elbow ,z elbow ) and the coordinates of the wrist joint P wrist =(x wrist ,y wrist ,zwrist ), using the coordinates P of the elbow joint elbow =(x elbow ,y elbow ,z elbow ) and shoulder joint coordinates P shoulder =(x shoulder ,y shoulder ,z shoulder ) Calculate the upper arm vector u of the arm part, and the specific algorithm formula is:

[0046] u=P elbow -P shoulder =(x elbow -x shoulder ,y elbow -y shoulder ,z elbow -z shoulder ); where the unit of the upper arm vector u of the arm part is meter (m);

[0047] Using the coordinates P of the wrist joint wrist =(x wrist ,y wrist ,z wrist ) and the coordinates of the elbow joint P elbow =(x elbow ,y elbow ,z elbow ) Calculate the forearm vector v of the arm part, and the specific algorithm formula is:

[0048] v=P wrist -P elbow =(x wrist -x elbow ,y wrist -y elbow ,z wrist -z elbow ); where the unit of the forearm vector v of the arm part is meter (m);

[0049] Since the upper arm vector u of the arm part includes components u1, u2, ..., u m , the forearm vector v of the arm part includes components v1, v2, ..., v m , calculate the dot product of the upper arm vector and the forearm vector by the components of the upper arm vector of the arm part and the components of the forearm vector of the arm part u·v=u1v1+u2v2+、...、+u m v m , where u·v is in meters (m), and the modulus of the upper arm vector is calculated using the components of the upper arm vector at the arm site. The unit of |u| is meter (m), and the modulus of the forearm vector is calculated using the components of the forearm vector at the arm site. The unit of |v| is meter (m). The joint motion angle θ is calculated by combining the dot product u·v of the upper arm vector and the forearm vector with the modulus |u| of the upper arm vector and the modulus |v| of the forearm vector.

[0050] Steps to calculate the joint action angle:

[0051] Step 1: Collect the dot product u·v of the upper arm vector and the forearm vector, the modulus |u| of the upper arm vector, and the modulus |v| of the forearm vector, multiply the modulus |u| of the upper arm vector and the modulus |v| of the forearm vector to obtain the modulus product |u|·|v|, and use the dot product u·v of the upper arm vector and the forearm vector and the modulus product |u|·|v| to calculate the cosine value cosθ. The specific algorithm formula is:

[0052] Step 2: Calculate the joint motion angle θ using the arccosine function (arccos) based on the product of the upper arm vector and forearm vector dot product u·v and the modulus |u|·|v|. The specific algorithm formula is:

[0053] The set action amplitude deviation threshold α and the set action speed threshold β are combined with the action amplitude deviation ΔP i , joint movement speed v i (t) is used to determine whether the captured student's whole-body movements are "normal activities" or "sports", which is used to distinguish sports from normal activities, while avoiding the accidental use of a large number of capture resources and saving computing resources;

[0054] Specific judgment situations include:

[0055] Case ①: When ΔP i >α, and v i When (t)>β, the captured student's whole body motion is judged to be in motion;

[0056] Case ②: When ΔP i <α, and v i When (t)>β, it is determined that the captured student's whole body movement is a normal activity, and feedback is sent to the capture and splitting module 101 to stop capturing the student's movement;

[0057] Case ③, when ΔP i >α, and v i When (t)<β, it is determined that the captured student's whole body movement is a normal activity, and feedback is sent to the capture and splitting module 101 to stop capturing the student's movement;

[0058] Case ④: When ΔP i <α, and v i When (t)<β, it is determined that the captured student's whole body movement is a normal activity, and feedback is sent to the capture and splitting module 101 to stop capturing the student's movement;

[0059] The deviation coordination unit 20 receives a command from the capture joint unit 10 to determine whether the captured student's whole body movement is in motion. The deviation coordination unit 20 obtains the joint movement angle and joint movement speed from the capture joint unit 10 to analyze the joint movement angle deviation, the joint movement speed peak value and the student's part movement coordination value to determine whether the student is doing gymnastics.

[0060] The deviation coordination unit 20 includes a deviation peak module 201 and an action coordination module 202;

[0061] The deviation peak module 201 receives the command that the angle judgment module 103 determines that the captured student's whole body movement is a movement command, and the deviation peak module 201 obtains the joint movement angle θ from the angle judgment module 103 to analyze the maximum movement angle max (θ j (t)) and the minimum movement angle min(θ j (t)),

[0062] Analyze the process of achieving the maximum and minimum movement angles of the parts:

[0063] First, obtain the joint action angle θ at time t; for example, for a joint action angle θ, at t1, t2, ..., t k Record the corresponding θ at each moment j (t1),θ j (t2), ..., θ j (t k ), where θ j (t1) refers to the action angle θ of the j-th joint at time t1, θ j (t2) refers to the action angle θ of the j-th joint at time t2, θ j (t k ) refers to t k The action angle θ of the jth joint at the moment,

[0064] Secondly, the joint action angle θ at time t1 is given the maximum and minimum values at the same time, that is, max temp =θ j (t1), min temp =θ j (t1), then obtain the joint action angle θ at time t2, and compare the action angle θ of each joint with the current maximum action angle and minimum action angle;

[0065] If the current joint's action angle is greater than max temp , then the updated max temp is the current maximum action angle;

[0066] If the current joint's action angle is less than min temp , then update min temp is the current minimum action angle;

[0067] Finally, after traversing the action angles of all joints, max temp That is max(θ i (t)), min temp That is min(θ i (t)), the maximum motion angle max(θ j (t)) and the minimum movement angle min(θ j (t)) to obtain the joint action angle deviation Δθ j =max(θ j (t))-min(θ j (t));

[0068] When the action angle deviation Δθ is known j When the deviation peak module 201 obtains the joint motion speed v from the joint speed module 102 i (t) to analyze the peak velocity v of the joint peak,j ;

[0069] Analyze the process of achieving the peak value of the maximum movement speed of the joint:

[0070] First, obtain the joint motion velocity v at time t i (t); For example, for the movement speed v of a joint i (t), at t1, t2, ..., t k Record the corresponding v at each moment i (t1), v i (t2),...,v i (t k );

[0071] Secondly, the joint movement speed v at time t1 i (t) is given the maximum value, that is, max temp =v i (t1), the movement speed v of each joint is i (t) is compared with the current maximum movement speed; if the current joint movement speed is greater than max temp , then the updated max temp is the current maximum movement speed;

[0072] Finally, after traversing the movement speeds of all joints, max temp That is max(v i(t)), and then calculate the absolute value of the current maximum movement speed to analyze the joint movement speed peak value v peak,i =max(|v i (t)|);

[0073] The motion coordination module 202 obtains the motion angle θ of the forearm joint from the motion angle θ of the joint in the angle judgment module 103 q (θ q The subscript "q" is used to distinguish and has no physical meaning) and the action angle θ of the upper arm joint s (θ s The subscript "s" is used for distinction and has no physical meaning), record the number of forearm joints N and the number of upper arm joints M, and calculate the angle of the forearm joint according to the action angle θ q The average angle of the forearm joint action μθ is obtained by combining the number of forearm joint points N. q , and then according to the action angle θ of the forearm joint q , the average value of the forearm joint action angle μθ q and the number of forearm joint points N to calculate the standard deviation of the forearm joint action angle σθ q ;

[0074] Using the upper arm joint movement angle θ s The average angle of the upper arm joint is obtained by adding the number of upper arm joint points M. s , and then use the upper arm joint movement angle θ s , average upper arm joint angle μθ s and the average angle of the upper arm joint action μθ s Calculate the standard deviation of the upper arm joint movement angle σθ s ;

[0075] The action angle θ of the forearm joint q 、The angle of the upper arm joint θ s and the standard deviation of the forearm joint angle σθ q and the standard deviation of the upper arm joint angle σθ s Combined calculation of student body movement coordination value Corr q,s , its specific algorithm formula is:

[0076] Among them, Cov represents covariance, which is a dimensionless quantity;

[0077] When the student's body movement coordination is known q,s When the motion coordination module 202 obtains the maximum motion velocity peak value v of the joint from the deviation peak module 201 peak,i and joint motion angle deviation Δθ j, through the set joint action velocity peak threshold γ, the set joint action angle deviation threshold δ and the set part action coordination threshold ε and the joint maximum action velocity peak v peak,i , joint movement angle deviation Δθ j And the student's body movement coordination value Corr q,s Determine whether students are doing gymnastics;

[0078] If v peak,i >γ, and Δθ j >δ, and Corr q,s When >ε, it is determined that the student is doing gymnastics; otherwise, it is determined that the student is not doing gymnastics;

[0079] The adjustment feedback unit 30 receives the command from the deviation coordination unit 20 that the student is performing gymnastics exercises. The adjustment feedback unit 30 obtains historical data from the joint capture unit 10 and compares the data with the joint movement angles to determine whether the gymnastics exercises performed by different parts of the student need to be adjusted. If adjustment is required, feedback information is sent to the student end.

[0080] When the adjustment feedback unit 30 receives the command from the action coordination module 202 that the student is doing gymnastics, the adjustment feedback unit 30 obtains the historical data and the joint action angle θ from the angle judgment module 103 from the capture and separation module 101, and collects the joint angle θ of the standard gymnastics action from the historical data. g , through the joint action angle θ and the joint angle θ of the standard gymnastics movement g Compare and calculate the joint angle difference Δθ = θ g -θ, uses the joint angle difference Δθ and the set joint action angle deviation threshold δ to determine whether the gymnastics movements of different parts of the student need to be adjusted. When the joint angle difference Δθ is greater than the set joint action angle deviation threshold δ, it is determined that the gymnastics movements of different parts of the student need to be adjusted. The current joint action angle and the joint angle of the standard gymnastics movement are fed back to the student end in real time, and the feedback information "Please raise your arms" or "Please bend your arms" is played to the student end;

[0081] A second object of the present invention is to provide a method for operating the above-mentioned intelligent physical education teaching real-time feedback system based on motion perception, comprising the following method steps:

[0082] S1. The joint capture unit 10 senses the student's movements in real time during online physical education, captures the student's whole body movements to split them into n joint points, and then calculates the movement amplitude deviation, joint movement speed, and joint movement angle to determine whether the captured student's whole body movement is a normal activity or exercise;

[0083] S2, the deviation coordination unit 20 obtains the joint action angle and joint action speed from the joint capture unit 10 to analyze the joint action angle deviation, the joint action speed peak value and the student's part action coordination value to determine whether the student is doing gymnastics;

[0084] S3. The adjustment feedback unit 30 determines whether the gymnastics movements of different parts of the student need to be adjusted by comparing the movement angles of the joints. If adjustment is required, feedback information is sent to the student end.

[0085] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely preferred examples of the present invention and are not intended to limit the present invention. Various changes and improvements may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and improvements fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. Intelligent physical education real-time feedback system based on motion perception, characterized by: It includes a capture joint unit (10), a deviation coordination unit (20) and an adjustment feedback unit (30); The joint capture unit (10) senses the student's movements in online physical education in real time, collects historical data, and captures the student's whole body movements to split n joint points, records the position coordinates of the joint points to obtain the movement amplitude deviation, the movement speed of the joints and the movement angle of the joints, and determines whether the captured student's whole body movements are normal activities or sports; The deviation coordination unit (20) receives a command from the capture joint unit (10) to determine that the captured student's whole body movement is in motion, and the deviation coordination unit (20) obtains the joint movement angle and joint movement speed from the capture joint unit (10) to analyze the joint movement angle deviation, the joint movement speed peak value and the student's part movement coordination value to determine whether the student is doing gymnastics; The adjustment feedback unit (30) receives a command from the deviation coordination unit (20) to determine that the student is doing gymnastics. The adjustment feedback unit (30) obtains historical data from the joint capture unit (10) and compares the data with the joint movement angles to determine whether the gymnastics movements performed by different parts of the student need to be adjusted. When adjustment is required, feedback information is sent to the student end.

2. The intelligent physical education teaching real-time feedback system based on motion perception according to claim 1 is characterized by: The capture joint unit (10) includes a capture splitting module (101) and a joint speed module (102); The capture and splitting module (101) uses a camera to perceive the student's movements in online physical education in real time, and captures the student's whole body movements, collects historical data, and when the camera fails to capture, it feeds back to the teacher in real time. When the camera succeeds, it splits the student's whole body movement data into movement data of different parts. The data splitting method is used to split the motion data of different parts into n joints, record the position coordinates of the joints, and then analyze the maximum motion amplitude value max(P i (t)) and the minimum motion amplitude value min(P i (t)) to obtain the action amplitude deviation ΔP i , where the arm motion data can be expressed as: Arm = {P shoulder ,P elbow ,P wrist }.

3. The intelligent physical education teaching real-time feedback system based on motion perception according to claim 2 is characterized in that: The joint speed module (102) collects the position coordinates of the joint i at time t through the capture and splitting module (101) to analyze the motion trajectory of the part; The position coordinates of the joint i at time t-1 are extracted from the historical data of the capture and splitting module (101), the time interval is calculated using time t and time t-1, and the movement speed of the joint is calculated based on the position coordinates of the joint i at time t and the position coordinates of the joint i at time t-1.

4. The intelligent physical education teaching real-time feedback system based on motion perception according to claim 2 is characterized by: The capture joint unit (10) further includes an angle judgment module (103); The angle judgment module (103) calculates the upper arm vector and the forearm vector of the arm part through the arm motion data in the capture and separation module (101); Since the upper arm vector of the arm part includes components u1, u2, ..., u m , the forearm vector of the arm part includes components v1, v2, ..., v m , calculate the dot product of the upper arm vector and the forearm vector by the components of the upper arm vector at the arm site and the components of the forearm vector at the arm site, calculate the modulus of the upper arm vector by using the components of the upper arm vector at the arm site, and then calculate the modulus of the forearm vector by using the components of the forearm vector at the arm site, and calculate the action angle of the joint by combining the dot product of the upper arm vector and the forearm vector with the modulus of the upper arm vector and the modulus of the forearm vector |v|; The set action amplitude deviation threshold α and the set action speed threshold β are combined with the action amplitude deviation ΔP i , joint movement speed v i (t) To determine whether the captured student's whole body movements are normal activities or sports.

5. The intelligent physical education teaching real-time feedback system based on motion perception according to claim 4 is characterized in that: The specific judgment situations in the angle judgment module (103) include: Case ①: When ΔP i >α, and v i When (t)>β, the captured student's whole body motion is judged to be in motion; Case ②, when ΔP i <α, and v i When (t)>β, it is determined that the captured student's whole body movement is a normal activity, and feedback is sent to the capture and splitting module (101) to stop capturing the student's movement; Case ③, when ΔP i >α, and v i When (t)<β, it is determined that the captured student's whole body movement is a normal activity, and feedback is sent to the capture and splitting module (101) to stop capturing the student's movement; Case ④: When ΔP i <α, and v i When (t)<β, it is determined that the captured student's whole body movement is a normal activity, and feedback is sent to the capture and separation module (101) to stop capturing the student's movement.

6. The intelligent physical education teaching real-time feedback system based on motion perception according to claim 4 is characterized by: The deviation coordination unit (20) includes a deviation peak module (201) and an action coordination module (202); The deviation peak module (201) receives a command from the angle judgment module (103) that the captured student's whole body movement is in motion, and the deviation peak module (201) obtains the joint movement angle from the angle judgment module (103) to analyze the maximum movement angle and the minimum movement angle of the part; The joint movement angle deviation is obtained by using the maximum movement angle and the minimum movement angle of the part; When the movement angle deviation is known, the deviation peak module (201) obtains the movement speed of the joint from the joint speed module (102) to analyze the movement speed peak of the joint.

7. The intelligent physical education teaching real-time feedback system based on motion perception according to claim 6 is characterized by: The action coordination module (202) obtains the action angle of the forearm joint and the action angle of the upper arm joint from the action angle of the joint in the angle judgment module (103), obtains the average action angle of the forearm joint according to the action angle of the forearm joint, and then calculates the standard deviation of the forearm joint action angle; The upper arm joint movement angle is used to obtain the average upper arm joint movement angle, and then the upper arm joint movement angle standard deviation is calculated; The student's body movement coordination value Corr is calculated by combining the action angle of the forearm joint, the action angle of the upper arm joint, the standard deviation of the forearm joint action angle, and the standard deviation of the upper arm joint action angle. q,s , its specific algorithm formula is: Among them, Cov represents covariance, which is a dimensionless quantity; When the coordination of the student's body movements is known, the movement coordination module (202) obtains the joint's maximum movement speed peak value and the joint movement angle deviation from the deviation peak module (201), and calculates the joint's maximum movement speed peak value v by using the set joint movement speed peak threshold value γ, the set joint movement angle deviation threshold value δ and the set body movement coordination threshold value ε. peak,i , joint movement angle deviation Δθ j And the student's body movement coordination value Corr q,s Determine whether students are doing gymnastics; If v peak,i >γ, and Δθ j >δ, and Corr q,s >ε, it is determined that the students are doing gymnastics.

8. The intelligent physical education teaching real-time feedback system based on motion perception according to claim 6 is characterized by: When the adjustment feedback unit (30) receives a command from the action coordination module (202) that determines that the student is doing gymnastics, the adjustment feedback unit (30) obtains historical data and the action angle of the joint in the angle judgment module (103) from the capture and separation module (101), and compares the action angle θ of the joint with the historical data to determine whether the gymnastics movements performed by different parts of the student need to be adjusted. If adjustment is required, feedback information such as "please raise your arms" or "please bend your arms" is played to the student end.

9. A method for operating the intelligent sports teaching real-time feedback system based on motion perception according to any one of claims 1 to 8, characterized in that: The method comprises the following steps: S1, the capture joint unit (10) senses the student's movements in real time during online physical education, and captures the student's whole body movements to split n joint points, and then calculates the movement amplitude deviation, joint movement speed and joint movement angle, and determines whether the captured student's whole body movement is a normal activity or exercise; S2, the deviation coordination unit (20) obtains the joint movement angle and joint movement speed from the capture joint unit (10) to analyze the joint movement angle deviation, the joint movement speed peak value and the student's part movement coordination value to determine whether the student is doing gymnastics; S3. The adjustment feedback unit (30) determines whether the gymnastics movements performed by different parts of the student need to be adjusted by comparing the movement angles of the joints. If adjustment is required, feedback information is sent to the student end.

Images