Intelligent timing and scoring system for competition

Abstract

The invention discloses an intelligent timing and scoring system for a match, and relates to the field of match systems, the intelligent timing and scoring system for the match comprises a first data acquisition port used for receiving a standard shadowboxing video as a scoring standard, and a second data acquisition port used for acquiring a real-time action video of an athlete in a shadowboxing match; according to the intelligent timing and scoring system for the competition, the difficulty value ND of each standard action is accurately quantified, and each real-time action of an athlete is compared and scored with the standard action according to a unified standard, so that the subjectivity and the uncertainty in the scoring process are greatly reduced, the credibility and the authority of a competition result are ensured, and the competition quality is improved. Therefore, the competition really becomes the comparison of technology and strength, the competitive consciousness and training enthusiasm of athletes are stimulated, and the benign development of the whole project is promoted.

Description

Technical Field

[0001] The present invention relates to the field of competition systems, and in particular to an intelligent timing and scoring system for competitions. Background Art

[0002] In the field of Tai Chi competitions, the traditional timing and scoring method mainly relies on the subjective judgment of the referees, which has many disadvantages. On the one hand, the personal experience, professional level and subjective preferences of the referees will have a great impact on the scoring results, making it difficult to ensure the objectivity and fairness of the scoring. Different referees may have different understandings and judgment standards for the accuracy, fluency, strength and difficulty of the movements, which makes the athletes' scores lack stability and consistency. Summary of the invention

[0003] In order to make up for the shortcomings of the existing technical problems, the purpose of the present invention is to provide an intelligent timing and scoring system for competitions, which determines the difficulty value of movements by quantifying the complexity of the amplitude, speed, balance stability and other complexities of each standard movement, objectively evaluates the performance of athletes, reduces the influence of subjective factors, and provides improvement suggestions for athletes to improve their technical level.

[0004] In order to solve the problems of the prior art, the technical solution of the present invention is as follows:

[0005] An intelligent timing and scoring system for a competition includes a first data acquisition port, a second data acquisition port, a scoring terminal, a display screen, and a processor:

[0006] A standard Tai Chi video used as a scoring standard is obtained through the first data acquisition port, a real-time action video of an athlete performing Tai Chi competition is obtained through the second data acquisition port, and timing data is displayed on a display screen to show the athlete;

[0007] The processor divides the standard Tai Chi video into a plurality of unit videos, each unit video covers a standard action, calculates the difficulty value ND of each standard action, compares the real-time action of the real-time action video with the standard action to determine the standard value BZ of the real-time action, and calculates the total score ZF of the athlete's real-time action video according to the difficulty value ND and the standard value BZ;

[0008] The scoring terminal obtains the referee's scoring results for the athlete's real-time actions, and the processor combines the referee's score and the total score ZF to calculate and output the athlete's final score ZZD.

[0009] Preferably, the processor completes the calculation of the final score ZZD by specifically comprising the following steps:

[0010] Step 1: Divide the standard Tai Chi video into multiple unit videos, each unit video covers a standard action;

[0011] Step 2: Based on the multiple unit videos obtained by division, the difficulty value ND of the standard action covered by the unit videos is calculated;

[0012] Step 3: compare the real-time action video with all the unit videos, and use the standard action in the unit video to determine the standard value BZ of the real-time action corresponding to the standard action in the real-time action video;

[0013] Step 4: Calculate the total score ZF of the athlete's real-time action video according to the standard values ​​BZ of all real-time actions in the obtained real-time action video and the difficulty values ​​ND of the corresponding standard actions;

[0014] Step 5. Calculate the athlete's final score ZZD using the total score ZF and the referee's score.

[0015] Compared with the prior art, the advantages of the present invention are as follows:

[0016] 1. In the present invention, since the difficulty value ND of each standard action is accurately quantified, and each real-time action of the athlete is compared and scored with the standard action according to a unified standard, the subjectivity and uncertainty in the scoring process are greatly reduced. Whether in professional competitions or daily training competitions, all participating athletes are evaluated under the same fair and just evaluation criteria, ensuring the credibility and authority of the competition results, making the competition a real contest of technology and strength, stimulating the competitive consciousness and training enthusiasm of athletes, and promoting the healthy development of the entire project;

[0017] 2. The present invention can evaluate the performance of athletes with high accuracy. This accuracy enables athletes to clearly understand their strengths and weaknesses in each specific action and obtain extremely detailed and targeted feedback. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a schematic diagram of the overall structure of the present invention.

[0019] Figure 2 It is a schematic diagram of the steps of the present invention.

[0020] Figure 3 It is a schematic diagram of the key joint points of the present invention.

[0021] Figure numerals: 1. wrist; 2. elbow; 3. shoulder; 4. hip; 5. knee; 6. ankle. DETAILED DESCRIPTION

[0022] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0023] For example, see Figures 1 to 3 This embodiment provides an intelligent timing and scoring system for a competition, including a first data acquisition port for receiving a standard Tai Chi video as a scoring standard, a second data acquisition port for collecting real-time action videos of athletes in Tai Chi competitions, a scoring terminal for a referee to score the real-time action videos of athletes, a processor for obtaining the referee's scoring results, the standard Tai Chi video, and the real-time action video to calculate the final score ZZD of the athlete, and a display screen for displaying the timing data to the athlete. The processor calculates the final score ZZD of the athlete, specifically including the following steps:

[0024] Step 1: The processor divides the standard Tai Chi video into multiple unit videos, each of which covers a standard action.

[0025] A1. Calculate the sum of pixels S between adjacent frames of the standard Tai Chi video:

[0026] The absolute frame difference image JZT between adjacent frames of the standard Tai Chi video is calculated and converted into a grayscale frame difference image HZT. The obtained standard Tai Chi video is a color image, which needs to be converted into a grayscale image during the calculation process to facilitate the processing and analysis. The process of converting a color image into a grayscale image is a common operation in the field of computer vision and image processing. For converting the absolute frame difference image JZT into the grayscale frame difference image HZT, a weighted average method is usually used to convert the color absolute frame difference image into a grayscale image. This is an existing mature technology, so it will not be described in detail.

[0027] Binarize the grayscale frame difference image HZT to obtain a binary absolute frame difference image EZT, calculate the pixel sum S of the binary absolute frame difference image EZT, and use the pixel sum S of the binary absolute frame difference image EZT to measure the motion intensity between adjacent frames of the standard Tai Chi video;

[0028] A2, start traversing from the beginning of the standard Tai Chi video, in the standard Tai Chi video, there is S> first motion intensity starting threshold KA between the nth frame and the n+1th frame, and before the nth frame, continuous F ka If the S between adjacent frames is less than KA, the nth frame in the standard Tai Chi video is determined to be the starting frame of the standard action, and the motion intensity is judged to be large enough based on the frame difference;

[0029] A3. After finding the start frame of the standard action, continue to traverse backwards. In the standard Tai Chi video, S between the mth frame and the m+1th frame is less than the first motion intensity end threshold JA, and after the mth frame, continuous F jaIf the S between adjacent frames is less than JA, the mth frame in the standard Tai Chi video is determined to be the end frame of the standard action, and the motion intensity is reduced to a sufficiently small value based on the frame difference;

[0030] A4. In the standard Tai Chi video, extract the video from the start frame to the end frame of the standard action to obtain a unit video covering one standard action;

[0031] The first exercise intensity starting threshold KA, F in the above content ka , the first exercise intensity end threshold JA and F ja The four parameters are obtained based on industry experience or experimental tests;

[0032] In addition to the above-mentioned methods, a variety of existing technologies such as clustering technology based on visual features, segmentation technology based on optical flow, or technology based on model fusion can also be used to divide the video into multiple unit videos, each of which covers an action. These are all existing mature technologies, so they will not be described in detail here, and only one solution will be introduced;

[0033] Step 2: The processor calculates the difficulty value ND of the standard action covered by the unit video based on the multiple unit videos obtained by the division.

[0034] The human body posture estimation model MediaPipe Pose is used to detect the key joint positions in each frame of the standard Tai Chi video based on existing technologies such as optical motion capture or optical flow method.

[0035] B1. Calculate the amplitude complexity C of the standard action in the unit video a

[0036] The following formula is used to calculate the range R of the key joint points in the unit video:

[0037]

[0038] Among them, R i represents the activity range of the i-th key joint point in the unit video, and Respectively represent the average horizontal coordinate and average vertical coordinate of the i-th key joint point in all frames of the unit video, x i，u represents the horizontal coordinate of the i-th key joint point in the u-th frame of the unit video, y i，u Represents the vertical coordinate of the i-th key joint point in the u-th frame of the unit video;

[0039] The key joint points are preset manually based on experience. Please refer to the following table 1 for the preset key joint points:

[0040] Table 1: Key joint point preset table

[0041] Wrist(1) Elbow(2) Shoulder (3) Hip(4) Knee(5) Ankle(6)

[0042] Further explanation of the calculation formula for the range of activity R:

[0043] What is calculated is the distance of the i-th key joint point relative to its average position in the u-th frame of the unit video;

[0044] Represents the maximum distance between the i-th key joint point and its average position in all frames of the entire unit video;

[0045] Represents the minimum distance between the i-th key joint point and its average position in all frames of the entire unit video;

[0046] The amplitude complexity C of the standard action is calculated using the following formula a :

[0047]

[0048] Among them, M is the total number of key joints, w i Represents the range of motion R of the i-th key joint point i The weight of Weight w i Preset according to the importance of key joints in standard movements, such as shoulders and hips with higher weights;

[0049] B2. Calculate the speed change complexity of the standard action in the unit video C v

[0050] The following formula is used to calculate the displacement da of the key joint points between adjacent frames in the unit video:

[0051]

[0052] Among them, i，u represents the displacement of the i-th key joint point between the u-th frame and the u+1-th frame in the unit video, x i，u represents the horizontal coordinate of the i-th key joint point in the u-th frame of the unit video, y i，u represents the ordinate of the i-th key joint point in the u-th frame of the unit video, x i，u+1 represents the horizontal coordinate of the i-th key joint point in the u+1-th frame of the unit video, y i，u+1 Represents the ordinate of the i-th key joint point in the u+1-th frame of the unit video;

[0053] The average rate of change of standard actions in the unit video is calculated using the following formula: d:

[0054]

[0055] Among them, N represents the total number of frames of the unit video, and M is the total number of key joint points;

[0056] The average rate of change r d做 Further explanation:

[0057] Numerator: It means that the displacement d of each key joint point between all adjacent frames (from the 1st frame to the N-1th frame, because the adjacent frame comparison is between n and n+1 frames, the last frame cannot participate in this adjacent comparison) is summed up first, and then the sum of these displacements of all M key joint points is summed up, that is, the total displacement of all key joint points between all adjacent frames in the whole standard action process is calculated;

[0058] Denominator: The denominator (N-1) represents the total number of pairs of adjacent frames involved in calculating displacement changes (the total number of frames N minus 1 is the number of adjacent frame pairs), M is the total number of key joints, then (N-1)*M represents the number of combinations of all possible key joints and adjacent frame pairs in the entire standard action process, that is, the total number of "calculation units";

[0059] By comparing the displacement d of all key joints between all adjacent frames in the unit video, the maximum value is determined and recorded as the peak value P. d , using the following formula: The speed change complexity C of the standard action in the unit video v :

[0060] C v =α*r d +β*P d ;

[0061] Where α and β are weight coefficients, and α+β=100%;

[0062] B3. Calculate the balance stability complexity of the standard action in the unit video C b

[0063] According to the human posture estimation model, the approximate coordinates of the human body's center of gravity are obtained by weighted calculation of multiple joint coordinates, that is, the position of the human body's center of gravity. It can also be obtained by using existing technologies such as the OpenPose algorithm and the MediaPipe algorithm. The following formula is used to calculate the moving trajectory length L of the human body's center of gravity in the standard action in the entire unit video:

[0064]

[0065] Among them, x g，u and g，uRespectively represent the horizontal and vertical coordinates of the center of gravity of the human body in the u-th frame of the unit video, y g，u+1 and g，u+1 They respectively represent the horizontal and vertical coordinates of the center of gravity of the human body in the u+1th frame in the unit video, and N represents the total number of frames in the unit video;

[0066] Determine the maximum offset distance D of the center of gravity of the human body relative to the center of gravity of the human body in the starting frame of the unit video in all frames of the unit video g This process is completed in three steps: first, the center of gravity of the human body in the starting frame of the unit video is used as the reference point, and finally the straight-line distance between the center of gravity of the human body and the reference point in each frame of the unit video is calculated. Finally, a maximum value is determined among all the straight-line distances using a comparison algorithm. The maximum value is the maximum offset distance D g ;

[0067] The following formula is used to calculate the complexity C of standard action balance stability b :

[0068] C b =γ*L g +δ*D g ;

[0069] Among them, γ and δ are weight coefficients, and γ+δ=100%;

[0070] B4. Use the following formula to calculate the difficulty value ND of the standard action in the unit video:

[0071] ND=ε*C a +ζ*C v +η*C b ;

[0072] Among them, ε, ζ and η are all weight coefficients, and ε+ζ+η=100%;

[0073] Step 3: The processor compares the real-time action video with all the unit videos, and uses the standard action in the unit video to determine the standard value BZ of the real-time action corresponding to the standard action in the real-time action video. The processor can combine the existing technologies such as action recognition and temporal action localization to achieve the standard value BZ. This step is illustrated as follows;

[0074] C1. Use format conversion technology and frame rate conversion technology to convert the format and frame rate of the real-time action video to be consistent with the parameters of the standard Tai Chi video, or set the device for collecting the real-time action video to ensure that the parameters of the real-time action video collected are consistent with the parameters of the standard Tai Chi video, use the human body posture estimation model to extract the coordinates of the key joint points in each frame of the real-time action video and all the unit videos, and calculate the joint angle and joint position change data based on the coordinates of each key joint point;

[0075] Use deep learning models, OpenPose and HRNet, to estimate human posture. After extensive training, the model can accurately detect human joints. The calculation of action feature joint angles and joint position change data can be achieved through geometric transformation, vector operations and other methods.

[0076] C2. In real-time action videos, use temporal action detection technology to find real-time actions similar to standard action unit videos. This is achieved by comparing the features of real-time video frames with those of standard action unit video frames, and using cosine similarity or Euclidean distance in similarity measurement methods to evaluate the degree of match;

[0077] C3. For each detected real-time action, compare it with the corresponding standard action unit video, and score the real-time action using the regression model or classification model in the machine learning model according to the accuracy, fluency, strength and other standards of the action. Improve the accuracy of the score by training a special scoring model, which predicts the score based on the difference between the action features and the target detection results;

[0078] Step 4: The processor calculates the total score ZF of the athlete's real-time action video according to the obtained standard values ​​BZ of all real-time actions in the real-time action video and the difficulty values ​​ND of the corresponding standard actions using the following formula:

[0079]

[0080] Among them, ND v Indicates the difficulty value of the standard action in the vth unit video, BZ v represents the standard value of the real-time action corresponding to the standard action in the vth unit video, and t is the total number of unit videos split from the standard Tai Chi video, that is, the total number of standard actions;

[0081] Step 5: The scoring terminal obtains the referee's score for the athlete's real-time action video, and the processor calculates the athlete's final score ZZD using the following formula:

[0082]

[0083] in, and φ are weight factors, and CPF k represents the scoring result of the kth referee on the athlete’s real-time action video, and j represents the total number of referees.

[0084] Embodiment 2: The difference between this embodiment and embodiment 1 is that:

[0085] Step 1: The processor divides the standard Tai Chi video into multiple unit videos, each of which covers a standard action:

[0086] A1. Use the human posture estimation model MediaPipe Pose, based on existing technologies such as optical motion capture or optical flow method to detect the position of key joints in each frame of the standard Tai Chi video, obtain the coordinate information of the key joints in each frame, and use the following formula to calculate the displacement db of the key joints between adjacent frames in the standard Tai Chi video:

[0087]

[0088] Among them, db i represents the displacement of the i-th key joint point between adjacent frames in the standard Tai Chi video, x i，n is the horizontal coordinate of the i-th key joint point in the n-th frame of the standard Tai Chi video, y i，n is the ordinate of the i-th key joint point in the n-th frame of the standard Tai Chi video, x i，n+1 is the horizontal coordinate of the i-th key joint point in the n+1th frame of the standard Tai Chi video, y i，n+1 is the ordinate of the i-th key joint point in the n+1-th frame of the standard Tai Chi video;

[0089] The following formula is used to calculate the total displacement D of all key joints between adjacent frames: total , using D total Comprehensive measurement of the degree of change in human posture:

[0090]

[0091] Among them, M is the total number of key joints;

[0092] A2. Start traversing from the beginning of the standard Tai Chi video. D appears between the nth frame and the n+1th frame in the standard Tai Chi video. total >The second motion intensity threshold KB, and before the nth frame in the standard Tai Chi video, continuous F kb D between adjacent frames total If both are less than KB, the nth frame in the standard Tai Chi video is determined to be the starting frame of the standard action:

[0093] A3. After finding the starting frame of the standard action, continue to traverse backwards and find the D between the mth frame and the m+1th frame in the standard Tai Chi video. total <The second exercise intensity end threshold JB, and after the mth frame in the standard Tai Chi video, the continuous F jb Frame D total If both are smaller than JB, the mth image in the standard Tai Chi video is determined to be the end frame of the standard action:

[0094] A4. In a standard Tai Chi video, extract the video from the start frame to the end frame of the standard action to obtain a unit video covering a standard action.

[0095] Embodiment 3, this embodiment provides a further technical solution based on embodiment 1, a competition intelligent timing and scoring system also includes an athlete terminal, and the processor completes the calculation of the final score ZZD and also includes the following steps:

[0096] Step 6. After completing the calculation of the final score ZZD, the processor will package the divided unit videos in sequence according to the order in which the standard movements appear in the standard Tai Chi video to form a learning data packet. At the same time, the standard values ​​BZ of all real-time movements are arranged in the order in which they appear in the real-time action video, and then a report is made, and the learning data packet and the report are sent to the athlete terminal.

[0097] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent timing and scoring system for competitions, characterized in that: It includes a first data acquisition port, a second data acquisition port, a scoring terminal, a display screen and a processor: A standard Tai Chi video used as a scoring standard is obtained through the first data acquisition port, a real-time action video of an athlete performing Tai Chi competition is obtained through the second data acquisition port, and timing data is displayed on a display screen to show the athlete; The processor divides the standard Tai Chi video into a plurality of unit videos, each unit video covers a standard action, calculates the difficulty value ND of each standard action, compares the real-time action of the real-time action video with the standard action to determine the standard value BZ of the real-time action, and calculates the total score ZF of the athlete's real-time action video according to the difficulty value ND and the standard value BZ; The scoring terminal obtains the referee's scoring results for the athlete's real-time actions, and the processor combines the referee's score and the total score ZF to calculate and output the athlete's final score ZZD.

2. The intelligent timing and scoring system for events according to claim 1, characterized in that: The processor completes the calculation of the final score ZZD by specifically including the following steps: Step 1: Divide the standard Tai Chi video into multiple unit videos, each unit video covers a standard action; Step 2: Based on the multiple unit videos obtained by division, the difficulty value ND of the standard action covered by the unit videos is calculated; Step 3: compare the real-time action video with all the unit videos, and use the standard action in the unit video to determine the standard value BZ of the real-time action corresponding to the standard action in the real-time action video; Step 4: Calculate the total score ZF of the athlete's real-time action video according to the standard values ​​BZ of all real-time actions in the obtained real-time action video and the difficulty values ​​ND of the corresponding standard actions; Step 5. Calculate the athlete's final score ZZD using the total score ZF and the referee's score.

3. The intelligent timing and scoring system for events according to claim 2, characterized in that: The step 1 specifically includes the following steps: A1. Calculate the total pixel sum S between adjacent frames of the standard Tai Chi video; A11, calculating the absolute frame difference image JZT between adjacent frames of the standard Tai Chi video; A12, converting the absolute frame difference image JZT into a grayscale frame difference image HZT; A13, binarizing the grayscale frame difference image HZT; A14, calculating the pixel sum S of the binary absolute frame difference image EZT; A2, start traversing from the beginning of the standard Tai Chi video, in the standard Tai Chi video, there is S> first motion intensity starting threshold KA between the nth frame and the n+1th frame, and before the nth frame, continuous F ka If the S between adjacent frames is less than KA, the nth frame in the standard Tai Chi video is determined to be the starting frame of the standard action; A3. After finding the start frame of the standard action, continue to traverse backwards. In the standard Tai Chi video, S between the mth frame and the m+1th frame is less than the first motion intensity end threshold JA, and after the mth frame, continuous F ja If the S between adjacent frames is less than JA, the mth frame in the standard Tai Chi video is determined to be the end frame of the standard action; A4. In a standard Tai Chi video, extract the video from the start frame to the end frame of the standard action to obtain a unit video covering a standard action.

4. The intelligent timing and scoring system for events according to claim 2, characterized in that: The step 1 specifically includes the following steps: A1. Detect the key joint positions in each frame of the standard Tai Chi video, obtain the key joint coordinate information in each frame, and calculate the total displacement D of all key joints between adjacent frames. total ; A2. Start traversing from the beginning of the standard Tai Chi video, and find the D in the nth frame in the standard Tai Chi video. total >The second motion intensity threshold KB, and before the nth frame appears in the standard Tai Chi video, the continuous F kb Frame D total If both are less than KB, the nth frame in the standard Tai Chi video is determined to be the starting frame of the standard action; A3. After finding the starting frame of the standard action, continue to traverse backwards and find D of the mth frame in the standard Tai Chi video. total <The second exercise intensity end threshold JB, and after the mth frame in the standard Tai Chi video, the continuous F jb Frame D total are all smaller than JB, then the mth image in the standard Tai Chi video is determined to be the end frame of the standard action; A4. In a standard Tai Chi video, extract the video from the start frame to the end frame of the standard action to obtain a unit video covering a standard action.

5. The intelligent timing and scoring system for events according to claim 2, characterized in that: The step 2 specifically includes the following steps: Detect the key joint positions in each frame of the standard Tai Chi video and calculate the amplitude complexity C of the standard movements in each unit video. a , speed change complexity C v And the balance stability complexity C b , use the following formula to calculate the difficulty value ND of the standard action: ND=ε*C a +ζ*C v +η*C b ; Among them, ε, ζ and η are all weight coefficients, and ε+ζ+η=100%.

6. The intelligent timing and scoring system for events according to claim 5, characterized in that: The total score ZF calculation formula is as follows: Among them, ND v Indicates the difficulty value of the standard action in the vth unit video, BZ v It represents the standard value of the real-time action corresponding to the standard action in the vth unit video, and t is the total number of unit videos split from the standard Tai Chi video.

7. The intelligent timing and scoring system for events according to claim 6, characterized in that: The final score ZZD is calculated as follows: in, and φ are weight factors, and CPF k represents the scoring result of the kth referee on the athlete’s real-time action video, and j represents the total number of referees.

8. The intelligent timing and scoring system for events according to claim 7, characterized in that: The amplitude complexity C a The calculation steps are as follows: The following formula is used to calculate the range R of the key joint points in the unit video: Among them, R i represents the activity range of the i-th key joint point in the unit video, and Respectively represent the average horizontal coordinate and average vertical coordinate of the i-th key joint point in all frames of the unit video, x i，u represents the horizontal coordinate of the i-th key joint point in the u-th frame of the unit video, y i，u Represents the vertical coordinate of the i-th key joint point in the u-th frame of the unit video; The amplitude complexity C of the standard action is calculated using the following formula a : Among them, M is the total number of key joints, w i Represents the range of motion R of the i-th key joint point i The weight of 9. The intelligent timing and scoring system for events according to claim 8, characterized in that: The speed variation complexity C v The calculation steps are as follows: The following formula is used to calculate the displacement da of the key joint points between adjacent frames in the unit video: Among them, i，u represents the displacement of the i-th key joint point between the u-th frame and the u+1-th frame in the unit video, x i，u represents the horizontal coordinate of the i-th key joint point in the u-th frame of the unit video, y i，u represents the ordinate of the i-th key joint point in the u-th frame of the unit video, x i，u+1 represents the horizontal coordinate of the i-th key joint point in the u+1-th frame of the unit video, y i，u+1 Represents the ordinate of the i-th key joint point in the u+1-th frame of the unit video; The average rate of change of standard actions in the unit video is calculated using the following formula: d : Among them, N represents the total number of frames of the unit video, and M is the total number of key joint points; The maximum value is determined by comparing the displacement d of all key joints between all adjacent frames in the unit video, which is recorded as the peak value P d , using the following formula: The speed change complexity C of the standard action in the unit video v : C v =α*r d +β*P d ; Wherein α and β are weight coefficients, and α+β=100%.

10. The intelligent timing and scoring system for events according to claim 9, characterized in that: The equilibrium stability complexity C b The calculation steps are as follows: Determine the position of the human body's center of gravity, and use the following formula to calculate the length L of the trajectory of the human body's center of gravity in the standard action in the entire unit video: Among them, x g，u and g，u They represent the horizontal and vertical coordinates of the center of gravity of the human body in the u-th frame of the unit video, respectively. g，u+1 and g，u+1 They respectively represent the horizontal and vertical coordinates of the center of gravity of the human body in the u+1th frame in the unit video, and N represents the total number of frames in the unit video; Determine the maximum offset distance D of the center of gravity of the human body relative to the center of gravity of the human body in the starting frame of the unit video in all frames of the unit video g , the following formula is used to calculate the standard action balance stability complexity C b : C b =γ*L g +δ*D g ; Wherein, γ and δ are both weight coefficients, and γ+δ=100%.

Images