A human standing long jump action evaluation method, device, medium and equipment
By performing coordinate transformation, filtering and smoothing, and applying a multiple linear regression model to the standing long jump video, the problem of unstable key point data in standing long jump posture estimation was solved, resulting in more accurate posture estimation and motion ability evaluation.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- JIANGXI COLLEGE OF APPLIED TECH
- Filing Date
- 2026-04-15
- Publication Date
- 2026-06-23
AI Technical Summary
In existing standing long jump attitude estimation methods, the stability of key point data is insufficient, and it is affected by jitter, occlusion and changes in lighting, resulting in inaccurate attitude estimation results.
By acquiring standing long jump videos, the raw coordinate data of key human body nodes are extracted, coordinate transformation and filtering are performed, the vertical velocity of the feet is determined using the first-order derivative, the take-off and landing times are identified, a multiple linear regression model is constructed, and the human motion trajectory during the hang-off phase is fitted by combining posture angles and body indicators.
It improves the accuracy and applicability of standing long jump posture estimation, provides reliable assessment of athletic ability for different age groups, and enhances the model's generalization ability.
Smart Images
Figure CN122265404A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of motion assessment, and in particular to a method, apparatus, medium, and equipment for assessing human standing long jump motion. Background Technology
[0002] The standing long jump is a core item in the physical fitness testing system for primary, secondary, and higher education institutions, and an important indicator for assessing students' lower limb explosive power and body coordination. To enhance students' awareness of their own health at different learning stages and effectively motivate them to actively participate in physical exercise, physical fitness testing is gradually developing towards informatization and intelligentization. Intelligent testing of the standing long jump can improve the accuracy and objectivity of the test using modern technology, providing more reliable data for evaluating individual students' athletic abilities and providing strong support for schools and education departments to conduct physical fitness monitoring, thereby promoting the digital upgrade of the physical fitness testing system.
[0003] With the rapid development of artificial intelligence technology, deploying cameras around sports venues can collect students' movement data in real time during physical fitness tests. Combined with AI-powered human posture estimation algorithms, this allows for continuous tracking of posture changes during the standing long jump, generating frame-by-frame motion images and key pixel coordinates. Based on this, further analysis of students' force application patterns at takeoff, their trajectory during hang time, and their landing posture angles can provide teachers with more targeted teaching suggestions and help students optimize their training programs, thereby improving athletic performance.
[0004] Traditional standing long jump performance evaluations in school physical tests, exams, and daily training typically involve using intelligent standing long jump testing devices, electronic mats, infrared or pressure sensors, or standard anti-slip mats with a distance measuring ruler to measure the athlete's standing long jump performance. Simultaneously, the evaluation criteria are used to determine whether the athlete stepped on or crossed the line, took a preparatory step, had an approach run, performed consecutive jumps, and whether both feet took off and landed simultaneously. However, these existing technologies still have at least the following problems: current evaluation methods mostly focus on performance measurement, violation judgment, or qualitative analysis based on a four-stage standard. Even when using human posture recognition technology, it lacks targeted processing for key point noise caused by factors such as shaking, occlusion, and changes in lighting during video acquisition, resulting in insufficient stability of key point data and consequently affecting the accuracy of subsequent standing long jump posture estimation results. Summary of the Invention
[0005] This invention provides a method, apparatus, medium, and device for evaluating human standing long jump movements, to solve the aforementioned problems existing in the prior art, namely, how to improve the accuracy of standing long jump posture estimation in the prior art. This invention provides a method for evaluating human standing long jump movements, the method comprising: Acquire the standing long jump video of the athlete to be tested, extract the original coordinate data of the key human body nodes corresponding to the standing long jump video, perform coordinate transformation on the original coordinate data, and determine the coordinate data consistent with the actual height direction; Based on coordinate data, the average vertical coordinates of the toe and heel nodes corresponding to the foot are extracted as the overall position of the foot. The vertical velocity of the foot is determined by performing numerical first-order derivative on the overall position of the foot in multiple frames of the standing long jump video. Based on the vertical velocity of the feet, the take-off time and landing time are determined, and the human motion trajectory during the hang time phase within the time interval between the take-off time and landing time is extracted; the average horizontal coordinates of the toe node and heel node of the athlete in each frame of the human motion trajectory are extracted, and the average horizontal coordinates of the toe node and heel node of each frame are arranged in chronological order to generate an average value sequence, which is used as the real motion trajectory. Posture angles, body indices, and motion indices are extracted from the actual motion trajectory. These parameters are used as independent variables, and the final posture estimation result of the standing long jump is used as the dependent variable to construct a multiple linear regression model for estimating the final posture of the standing long jump. The final posture estimation result of the standing long jump is the horizontal displacement of the human body's center of mass during the hang time phase.
[0006] Optionally, determining the vertical velocity of the foot by performing a numerical first-order differential on the overall position of the foot across multiple frames in the standing long jump video specifically includes: The formula for calculating the vertical velocity of the foot is as follows: ; in, For the first i Vertical velocity of the foot in frame For the first i Average vertical coordinate of the foot in frame. For the first i Frame corresponds to time, For the first i +1 frame average of foot vertical coordinates For the first i -1 frame average of foot vertical coordinates.
[0007] Optionally, the posture angles include the angle between the thigh and the ground at takeoff, the torso tilt angle at takeoff, the lower limb joint angle at landing, the torso tilt angle at landing, the angle between the upper and lower limbs at landing, and the angle between the thigh and the ground at landing; wherein, the angle between the thigh and the ground is the angle between the line connecting the hip and ankle joints and the horizontal direction, and the angle between the thigh and the ground... The acquisition of, specifically includes: ; The trunk tilt angle is the angle between the line connecting the shoulder and hip joints and the vertical direction. The acquisition of, specifically includes: ; The lower limb joint angle is the angle formed by the hip, knee, and ankle joints of the human body. The acquisition of, specifically includes: ; The angle between the upper limb and lower limb is the angle formed by the shoulder joint, hip joint, and knee joint of the human body. The acquisition of, specifically includes: ; Where v is the vertical unit vector and g is the horizontal vector. This represents the average of the coordinates of the left and right hip joints. The average coordinates of the knee joint. s represents the average coordinates of the ankle joint, and s represents the average coordinates of the shoulder joint.
[0008] Optionally, obtaining the average horizontal coordinates of the athlete's toe and heel nodes specifically includes: The average horizontal coordinates of the athlete's toe and heel nodes are obtained using the following formula: ; in, This represents the average horizontal coordinate of the toe and heel nodes of the athlete in the i-th frame. The fitting coefficients are quadratic terms. The coefficients for the first-order term are the fitting coefficients. The fitting coefficients are constant terms. This represents the time corresponding to the i-th frame.
[0009] Optionally, the coordinate data can be smoothed using the Savitzky-Golay filtering method.
[0010] Optionally, the body metrics include height, weight, lean body mass, and muscle mass; the athletic metrics include maximum wrist speed.
[0011] Optionally, the key human body nodes specifically include: Head node, torso node, arm node, and leg node.
[0012] This invention provides a human standing long jump motion assessment device, comprising: The acquisition module is used to acquire the standing long jump video of the athlete to be tested, extract the original coordinate data of the key human body nodes corresponding to the standing long jump video, perform coordinate transformation on the original coordinate data, and determine the coordinate data that is consistent with the actual height direction. The extraction module is used to extract the average vertical coordinates of the toe node and heel node corresponding to the foot as the overall position of the foot based on the coordinate data; the vertical velocity of the foot is determined by performing the first-order numerical derivative of the overall position of the foot in multiple frames of the standing long jump video. The real motion trajectory determination module is used to determine the take-off time and landing time based on the vertical velocity of the feet, and extract the human motion trajectory during the hang time phase within the time interval between the take-off time and the landing time; extract the average horizontal coordinates of the athlete's toe node and heel node in each frame of the human motion trajectory, arrange the average horizontal coordinates of the toe node and heel node in each frame in chronological order to generate an average value sequence, and use the average value sequence as the real motion trajectory. A construction module is used to extract posture angles, body indices, and motion indices from the actual motion trajectory. The posture angles, body indices, and motion indices are used as independent variables, and the final posture estimation result of the standing long jump is used as the dependent variable to construct a multiple linear regression model for the final posture estimation of the standing long jump. The final posture estimation result of the standing long jump is the horizontal displacement of the human body's center of mass during the hang time.
[0013] The present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for evaluating human standing long jump movements.
[0014] The present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the above-described human standing long jump action evaluation method.
[0015] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention provides a method for evaluating the standing long jump action of the human body. This method uses a quadratic polynomial to fit the human body's movement trajectory, which can determine the true movement trajectory and provide accurate data for subsequent standing long jump posture estimation. Then, by extracting posture angles, body indicators, and motion indicators from the true movement trajectory, and combining the characteristic angle parameters of key human body parts with individual body indicators and motion indicators, a multiple linear regression model for the final posture estimation of the standing long jump is proposed. By introducing multiple parameters of the athlete, the final posture estimation result of the standing long jump is obtained, namely the horizontal displacement of the human body's center of mass during the hang time, which improves the applicability of the model to different age groups (including primary and secondary school students and young people) and enhances the accuracy of posture estimation. Attached Figure Description
[0016] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.
[0017] Figure 1 A flowchart of a method for evaluating human standing long jump movements provided in an embodiment of the present invention; Figure 2 A schematic diagram of key human body node division provided in an embodiment of the present invention; Figure 3 This is a comparison chart of Savitzky-Golay filtering results provided in an embodiment of the present invention; Figure 4 The diagram shows the calculation results of takeoff and landing using the first-order differential method provided in this embodiment of the invention. Figure 5 A schematic diagram of the take-off and landing times of key human body nodes provided in an embodiment of the present invention (athlete 1). Figure 6 A motion trajectory curve during the airborne phase provided in an embodiment of the present invention (athlete 1). Figure 7 A schematic diagram of the take-off and landing times of key human body nodes provided in an embodiment of the present invention (athlete 2). Figure 8 A motion trajectory curve during the airborne phase provided in an embodiment of the present invention (athlete 2); Figure 9 A schematic diagram of four characteristic angles of a key human body part provided for an embodiment of the present invention; Figure 10 A schematic diagram of a computer device for evaluating the standing long jump motion of a human body, as provided in an embodiment of the present invention. Detailed Implementation
[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.
[0019] The technical solution of the present invention and how the technical solution of the present invention solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of the present invention will now be described with reference to the accompanying drawings.
[0020] Example 1 Figure 1 This is a flowchart of a method for evaluating the standing long jump motion of a human body, as provided in an embodiment of the present invention. Figure 1 As shown in this embodiment, a method for evaluating the standing long jump motion of a human body includes: S1: Obtain the standing long jump video of the athlete to be tested, extract the original coordinate data of the key human body nodes corresponding to the standing long jump video, perform coordinate transformation on the original coordinate data, and determine the coordinate data consistent with the actual height direction.
[0021] For example, this embodiment uses a standing long jump video of a single athlete and its corresponding human body key point coordinate data as the object to illustrate how to apply the method of the present invention for analysis. First, the frame-by-frame key point coordinate data of the athlete is obtained. The human body key nodes are divided into head, torso, arms, and legs. There are a total of 33 human body key points, which are divided into four parts according to the human physiological structure: head, torso, arms, and legs. Head (corresponding key points: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10); torso (corresponding key points: 11, 12, 23, 24); arms (corresponding key points: 13, 14, 15, 16, 17, 18, 19, 20, 21, 22); legs (corresponding key points: 25, 26, 27, 28, 29, 30, 31, 32). See [link to specific key point locations] for details. Figure 2 .
[0022] After obtaining the original keypoint coordinates, the coordinate system is transformed. Next, the coordinate system transformation ensures that the Y-axis coordinates are consistent with the actual height, and data smoothing is performed to reduce noise interference. Since the Y-axis direction in the video image coordinate system is opposite to the actual height direction of the human body, at a resolution of 1280×720, the Y-axis coordinates of each frame's keypoints are flipped by subtracting the original Y-coordinate value from 720 to obtain coordinate data consistent with the actual height direction. After coordinate transformation, the height changes during the human body's take-off, flight, and landing can be more realistically reflected.
[0023] For example, in actual video capture, the human keypoint recognition algorithm is affected by factors such as shaking, occlusion, and changes in lighting, causing noise fluctuations in the original keypoint coordinates. To reduce noise interference and preserve the keypoint change trend as much as possible, this embodiment uses the Savitzky-Golay filtering method to smooth the coordinate sequence of all 33 keypoints. Preferably, a second-order polynomial fitting is used, the sliding window length is set to 7 frames, and the processed result is retained to two significant decimal places. Taking the foot keypoint motion trajectory as an example, the comparison results before and after filtering are as follows: Figure 3 As shown. By Figure 3 It is evident that the key point trajectory is smoother after filtering, which can provide a stable data foundation for subsequent identification of take-off and landing moments.
[0024] S2: Based on coordinate data, extract the average vertical coordinates of the toe node and heel node corresponding to the foot as the overall position of the foot; determine the vertical velocity of the foot by performing numerical first-order differential on the overall position of the foot in multiple frames of the standing long jump video.
[0025] For example, in identifying the take-off and landing moments, this embodiment does not use manual visual inspection, but instead combines the dynamic characteristics of the standing long jump and uses the abrupt changes in the vertical velocity of the foot for automatic identification. Because the athlete's foot rapidly leaves the ground at the moment of take-off, its vertical coordinate rises rapidly, resulting in a negative abrupt change in vertical velocity; conversely, at the moment of landing, the foot rapidly contacts the ground, causing its vertical coordinate to drop rapidly, resulting in a positive abrupt change in vertical velocity. Therefore, the average vertical coordinates of four key points on the foot (29, 30, 31, and 32) are selected as the overall foot position, and their first-order derivative is performed to obtain the vertical velocity of the foot in each frame. The formula for calculating the vertical velocity of the foot is:
[0026] ; in, Let be the vertical velocity of the foot in the i-th frame. The average vertical coordinate of the foot in the i-th frame. The time corresponding to the i-th frame is calculated using a one-sided difference method for the first and last frames to ensure that the speed sequence is consistent with the frame number.
[0027] S3: Based on the vertical velocity of the feet, determine the take-off time and landing time, and extract the human motion trajectory during the hang time phase within the time interval between the take-off time and landing time; extract the average horizontal coordinates of the athlete's toe node and heel node in each frame of the human motion trajectory, arrange the average horizontal coordinates of the toe node and heel node in each frame in chronological order to generate an average value sequence, and use the average value sequence as the real motion trajectory.
[0028] For example, by combining video frame rate with keypoint fluctuations, the threshold for sudden speed changes during takeoff and landing can be set to ±150px / s. When the vertical foot velocity first falls below -150px / s, it is determined as the takeoff moment; after identifying the takeoff moment, within the subsequent 1-second timeframe, when the vertical foot velocity exceeds 150px / s, it is determined as the landing moment. This method ensures the temporal consistency and reliability of jump action recognition. The takeoff and landing recognition results are as follows: Figure 4 As shown, the corresponding human key node poses are as follows: Figure 5 As shown in the figure, during takeoff, the athlete's feet leave the ground, the body leans forward, and the arms swing; upon landing, the athlete's legs extend forward and the center of gravity shifts downward, conforming to the basic movement characteristics of the standing long jump.
[0029] After identifying the takeoff and landing times, the hang time data of the athlete is extracted based on the time interval between them. Considering that the overall trajectory of the human body in the air during the standing long jump is approximately parabolic, but the actual trajectory may deviate slightly due to posture adjustments near landing, this embodiment uses a quadratic polynomial to fit the hang time trajectory. The acquisition of the average horizontal coordinates of the athlete's toe and heel nodes specifically includes:
[0030] The average horizontal coordinates of the athlete's toe and heel nodes are obtained using the following formula: ; in, This represents the average horizontal coordinate of the toe and heel nodes of the athlete in the i-th frame. The fitting coefficients are quadratic terms. The coefficients for the first-order term are the fitting coefficients. The fitting coefficients are constant terms. This represents the time corresponding to the i-th frame. The fitting result is as follows: Figure 6 As shown. By Figure 6 It can be seen that the trajectory during the hang time phase generally follows a smooth parabolic trend, which can well describe the actual movement path of the athlete during the take-off process.
[0031] S4: Extract the posture angle, body indicators, and motion indicators from the actual motion trajectory, use the posture angle, body indicators, and motion indicators as independent variables, and use the final posture estimation result of the standing long jump as the dependent variable to construct a multiple linear regression model for the final posture estimation of the standing long jump; wherein, the final posture estimation result of the standing long jump is the horizontal displacement of the human body's center of mass during the hang time.
[0032] For example, after obtaining the motion path during the hang-off phase, characteristic angles are calculated for key body parts to quantitatively describe the athlete's force application patterns and posture changes during the hang-off phase. This embodiment selects four types of angle features, including lower limb joint angles, trunk tilt angles, the angle between the upper and lower limbs, and the angle between the thigh and the ground.
[0033] The lower limb joint angle is the angle formed by the hip joint, knee joint, and ankle joint, reflecting the degree of flexion and extension of the lower limb. The hip joint coordinates are the average of the left and right hip coordinates, the knee joint coordinates are the average of the left and right knee coordinates, and the ankle joint coordinates are the average of the left and right ankle coordinates. The calculation formula is as follows:
[0034] ; In the formula, This represents the average coordinates of the hip joint, which is the average of the coordinates of the left and right hip joints. The average coordinates of the knee joint. This represents the average coordinates of the ankle joint. This angle, expressed as hip-knee-ankle, reflects the degree of lower limb flexion and extension. During takeoff, a decrease in angle indicates squatting and cushioning, while a rapid increase in angle at the moment of takeoff indicates the leg drive and power generation process. A diagram illustrating the lower limb joint angles is shown below. Figure 7 As shown in the diagram. The angle decreases before takeoff to indicate a squatting and cushioning motion, and rapidly increases at the moment of takeoff to indicate the downward extension and power generation process of the lower limbs. The angle is illustrated as follows: Figure 9 As shown in angle 1.
[0035] The trunk tilt angle is the angle between the line connecting the shoulder and hip joints and the vertical direction, reflecting the degree of forward trunk tilt. The shoulder joint coordinates are taken as the average of the left and right shoulder key points, and the hip joint coordinates are taken as the average of the left and right hip key points. The calculation formula is as follows:
[0036] ; In the formula, The average coordinates of the shoulder joint. The vertical direction is represented by this angle, which is the tilt angle of the shoulder-hip joint relative to the vertical direction. If the athlete is standing upright with their shoulder and hip on the same vertical line, then this angle is... An excessively large angle can affect takeoff efficiency and flight posture. A diagram illustrating the torso tilt angle is shown below. Figure 7As shown in the diagram. When the athlete is upright, this angle is close to 0°; an excessively large angle may affect takeoff efficiency and flight posture. The angle is illustrated in the diagram below. Figure 9 As shown in angle 2.
[0037] The angle between the upper and lower limbs is the angle formed by the shoulder joint, hip joint, and knee joint. It is used to characterize the relative positional relationship between the trunk and the thigh. The formula for its calculation is: ; In the formula, , , These are the average coordinates of the shoulder, hip, and knee joints, respectively. This angle represents the angle between the torso and the thigh, reflecting the athlete's body posture during takeoff and whether there is a "bending" motion in the air. A diagram illustrating the angle between the upper and lower limbs is shown below. Figure 7 As shown in the diagram. This angle reflects whether the athlete forms a noticeable "folding" abdominal contraction during takeoff and flight. The angle is illustrated in the diagram below. Figure 9 As shown in angle 3.
[0038] The angle between the thigh and the ground is the angle between the line connecting the hip and ankle joints and the horizontal direction. It is used to describe the range of leg swing and the degree of forward extension. The calculation formula is as follows: ; In the formula, , These are the average coordinates of the hip and ankle joints, respectively. This angle, represented by the angle between the thigh and the ground, reflects their relative position and can be used to determine the leg swing amplitude during takeoff and landing. A diagram illustrating the thigh-to-ground angle is shown below. Figure 7 As shown. This angle can be used to assess an athlete's leg control during takeoff and landing, as illustrated in the diagram. Figure 9 As shown in angle 4.
[0039] Based on the aforementioned angle calculation method, key features related to the final posture estimation are further extracted, including: the angle between the thigh and the ground at takeoff, the trunk tilt angle at takeoff, the lower limb joint angle at landing, the trunk tilt angle at landing, the angle between the upper and lower limbs at landing, and the angle between the thigh and the ground at landing, totaling six posture angle values. Simultaneously, five body index values are extracted: the athlete's height, weight, lean body mass, muscle mass, and maximum wrist speed, constituting the 11 input features of the final posture estimation model.
[0040] Before modeling, the 11 input features are Z-score standardized to eliminate the influence between different units. The standardization formula is:
[0041] ; in, These are the original eigenvalues. These are the standardized eigenvalues. The characteristic mean, The characteristic standard deviation is used. Then, the final posture estimation result of the standing long jump is used as the dependent variable. Establish a multiple linear regression model:
[0042] ; in, For the intercept term, For each independent variable, the regression coefficients are... This represents the error term. The least squares method is then used to solve for each regression coefficient, and the calculation formula is as follows:
[0043] ; Substituting the obtained coefficients into the regression model yields the model expression for estimating the final posture of the standing long jump. This model comprehensively reflects the influence of the athlete's posture angles and body indicators on the final performance, thus achieving an objective quantitative estimation of the standing long jump posture.
[0044] Example 2 For example, this embodiment selects long jump data from two athletes for verification. The first-order differential method is used to accurately identify the takeoff and landing times, and the results are consistent with the video data, proving the effectiveness of the method. The trajectory during the hang-time phase conforms to the expected parabolic trajectory, and the fitting results are as follows: Figure 6 As shown, after performing the coordinate transformation and Savitzky-Golay filtering on the keypoint data of athletes 1 and 2 respectively, the average vertical coordinates of four key points on the feet were extracted, and the vertical velocity of the feet was calculated using the first-order derivative. The take-off and landing times were identified based on a velocity mutation threshold of ±150px / s. The results show that athlete 1's take-off time was 4.133s, corresponding to frame 124, and the landing time was 4.567s, corresponding to frame 137, with a hang time of 0.434s; athlete 2's take-off time was 5.467s, corresponding to frame 164, and the landing time was 5.900s, corresponding to frame 177, with a hang time of 0.433s. The above identification results are consistent with their motion video performance, indicating that the first-order derivative recognition method used in this invention has good reliability and consistency.
[0045] Furthermore, by fitting the trajectory data of the two athletes during their hang time, their height change curves during the airborne phase can be obtained. Combining the analysis results in this invention, the trajectory of athlete 1 during the hang time phase conforms to a parabolic trend, with the feet rising rapidly after takeoff, reaching the highest point in the middle, and then gradually descending to the ground; athlete 2 exhibits a similar pattern. Figure 6 and Figure 8 The fitting curves of the two athletes during the hang time phase are also shown. The distribution of hang time points matches the fitting curves as a whole, indicating that the present invention can effectively characterize the motion path during the standing long jump by using polynomial fitting.
[0046] Example 3 This embodiment constructs a regression model based on long jump data from multiple athletes, combining key angles and body indicators to accurately estimate the final long jump posture. Through multiple regression, combining body indicators and posture angles further improves the model's accuracy and stability. Standing long jump test data from multiple athletes are selected as training samples, and six posture angle values and five body indicator values are extracted, forming an 11-dimensional input feature vector. This 11-dimensional feature vector is then input into a multiple linear regression model for training, obtaining the regression coefficients of each feature on the final posture estimation result. According to the correlation analysis results of this invention, in terms of body indicators, lean body mass, muscle mass, weight, and height can effectively characterize an athlete's physical characteristics and power potential; in terms of posture angles, the key angles at takeoff and landing, as well as the maximum wrist speed, are highly correlated with long jump performance. Therefore, using them together as model input variables can improve the accuracy and generalization ability of posture estimation.
[0047] Furthermore, after training, the 11 feature values of the athlete are input into the established multiple linear regression model to output the final posture estimation result of the athlete's standing long jump. This result not only reflects the athlete's current movement quality but can also be used to identify weaknesses in the movement, such as insufficient forward lean at takeoff, inadequate abdominal contraction during the flight phase, or uncoordinated leg control before landing, thus providing a basis for subsequent teaching guidance and training optimization. Compared to traditional methods that rely solely on a single posture parameter or a single body indicator, this invention can comprehensively utilize movement process information and body indicator information to achieve a more comprehensive, stable, and objective standing long jump posture estimation.
[0048] In summary, this invention constructs a standing long jump action evaluation method based on human posture and body indicators through steps such as human body key point segmentation, coordinate transformation, filtering and smoothing, first-order differential identification, hang-time trajectory fitting, feature angle extraction, and multiple linear regression modeling. This method can accurately identify key stages in the standing long jump action, quantitatively analyze the athlete's posture angles, and combine body indicators to achieve final posture estimation, demonstrating strong accuracy, objectivity, and application value.
[0049] The above are one or more embodiments of the human standing long jump action assessment method provided in this specification. Based on the same idea, this specification also provides a corresponding human standing long jump action assessment device, including: The acquisition module is used to acquire the standing long jump video of the athlete to be tested, extract the original coordinate data of the key human body nodes corresponding to the standing long jump video, perform coordinate transformation on the original coordinate data, and determine the coordinate data that is consistent with the actual height direction. The extraction module is used to extract the average vertical coordinates of the toe node and heel node corresponding to the foot as the overall position of the foot based on the coordinate data; the vertical velocity of the foot is determined by performing the first-order numerical derivative of the overall position of the foot in multiple frames of the standing long jump video. The real motion trajectory determination module is used to determine the take-off time and landing time based on the vertical velocity of the feet, and extract the human motion trajectory during the hang time phase within the time interval between the take-off time and the landing time; extract the average horizontal coordinates of the athlete's toe node and heel node in each frame of the human motion trajectory, arrange the average horizontal coordinates of the toe node and heel node in each frame in chronological order to generate an average value sequence, and use the average value sequence as the real motion trajectory. A construction module is used to extract posture angles, body indices, and motion indices from the actual motion trajectory. The posture angles, body indices, and motion indices are used as independent variables, and the final posture estimation result of the standing long jump is used as the dependent variable to construct a multiple linear regression model for the final posture estimation of the standing long jump. The final posture estimation result of the standing long jump is the horizontal displacement of the human body's center of mass during the hang time.
[0050] Specific limitations regarding the human standing long jump performance evaluation device can be found in the limitations of the human standing long jump performance evaluation method described above, and will not be repeated here. Each module in the aforementioned human standing long jump performance evaluation device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.
[0051] The present invention also provides a computer-readable storage medium storing a computer program that can be used to execute the above-described human standing long jump motion assessment method.
[0052] The present invention also provides Figure 10 The schematic diagram of the computer device shown in Figure 10 illustrates that, at the hardware level, the computer device includes a processor, an internal bus, a network interface, memory, and non-volatile memory, and may also include other hardware required for business operations. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to implement the human standing long jump action evaluation method provided in the above embodiment.
[0053] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.
[0054] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above 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 invention.
Claims
1. A method for evaluating human standing long jump movements, characterized in that, include: Acquire the standing long jump video of the athlete to be tested, extract the original coordinate data of the key human body nodes corresponding to the standing long jump video, perform coordinate transformation on the original coordinate data, and determine the coordinate data consistent with the actual height direction; Based on coordinate data, the average vertical coordinates of the toe and heel nodes corresponding to the foot are extracted as the overall position of the foot. The vertical velocity of the foot is determined by performing numerical first-order derivative on the overall position of the foot in multiple frames of the standing long jump video. Based on the vertical velocity of the feet, the take-off time and landing time are determined, and the human motion trajectory during the hang time phase within the time interval between the take-off time and landing time is extracted; the average horizontal coordinates of the toe node and heel node of the athlete in each frame of the human motion trajectory are extracted, and the average horizontal coordinates of the toe node and heel node of each frame are arranged in chronological order to generate an average value sequence, which is used as the real motion trajectory. Posture angles, body indices, and motion indices are extracted from the actual motion trajectory. These parameters are used as independent variables, and the final posture estimation result of the standing long jump is used as the dependent variable to construct a multiple linear regression model for estimating the final posture of the standing long jump. The final posture estimation result of the standing long jump is the horizontal displacement of the human body's center of mass during the hang time phase.
2. The method for evaluating human standing long jump as described in claim 1, characterized in that, The method of determining the vertical velocity of the foot by performing a numerical first-order differential on the overall position of the foot in multiple frames of a standing long jump video specifically includes: The formula for calculating the vertical velocity of the foot is as follows: ； in, For the first i Vertical velocity of the foot in frame For the first i Average vertical coordinate of the foot in frame. For the first i Frame corresponds to time, For the first i +1 frame average of foot vertical coordinates For the first i -1 frame average of foot vertical coordinates.
3. The method for evaluating human standing long jump as described in claim 1, characterized in that, The posture angles include the angle between the thigh and the ground at takeoff, the torso tilt angle at takeoff, the lower limb joint angle at landing, the torso tilt angle at landing, the angle between the upper and lower limbs at landing, and the angle between the thigh and the ground at landing; wherein, the angle between the thigh and the ground is the angle between the line connecting the hip and ankle joints and the horizontal direction, and the angle between the thigh and the ground... The acquisition of, specifically includes: ； The trunk tilt angle is the angle between the line connecting the shoulder and hip joints and the vertical direction. The acquisition of, specifically includes: ； The lower limb joint angle is the angle formed by the hip, knee, and ankle joints of the human body. The acquisition of, specifically includes: ； The angle between the upper limb and lower limb is the angle formed by the shoulder joint, hip joint, and knee joint of the human body. The acquisition of, specifically includes: ； Where v is the vertical unit vector and g is the horizontal vector. This represents the average of the coordinates of the left and right hip joints. The average coordinates of the knee joint. s represents the average coordinates of the ankle joint, and s represents the average coordinates of the shoulder joint.
4. The method for evaluating human standing long jump as described in claim 1, characterized in that, The acquisition of the average horizontal coordinates of the athlete's toe and heel nodes specifically includes: The average horizontal coordinates of the athlete's toe and heel nodes are obtained using the following formula: ； in, This represents the average horizontal coordinate of the toe and heel nodes of the athlete in the i-th frame. The fitting coefficients are quadratic terms. The coefficients for the first-order term are the fitting coefficients. The fitting coefficients are constant terms. This represents the time corresponding to the i-th frame.
5. The method for evaluating human standing long jump as described in claim 1, characterized in that, The coordinate data is smoothed using the Savitzky-Golay filtering method.
6. The method for evaluating human standing long jump as described in claim 1, characterized in that, The body metrics include height, weight, lean body mass, and muscle mass; the athletic metrics include maximum wrist speed.
7. The method for evaluating human standing long jump as described in claim 1, characterized in that, The key nodes of the human body specifically include: Head node, torso node, arm node, and leg node.
8. A human standing long jump motion assessment device, characterized in that, include: The acquisition module is used to acquire the standing long jump video of the athlete to be tested, extract the original coordinate data of the key human body nodes corresponding to the standing long jump video, perform coordinate transformation on the original coordinate data, and determine the coordinate data that is consistent with the actual height direction. The extraction module is used to extract the average vertical coordinates of the toe node and heel node corresponding to the foot as the overall position of the foot based on the coordinate data; the vertical velocity of the foot is determined by performing the first-order numerical derivative of the overall position of the foot in multiple frames of the standing long jump video. The real motion trajectory determination module is used to determine the take-off time and landing time based on the vertical velocity of the feet, and extract the human motion trajectory during the hang time phase within the time interval between the take-off time and the landing time; extract the average horizontal coordinates of the athlete's toe node and heel node in each frame of the human motion trajectory, arrange the average horizontal coordinates of the toe node and heel node in each frame in chronological order to generate an average value sequence, and use the average value sequence as the real motion trajectory. A construction module is used to extract posture angles, body indices, and motion indices from the actual motion trajectory. The posture angles, body indices, and motion indices are used as independent variables, and the final posture estimation result of the standing long jump is used as the dependent variable to construct a multiple linear regression model for the final posture estimation of the standing long jump. The final posture estimation result of the standing long jump is the horizontal displacement of the human body's center of mass during the hang time.
9. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the human standing long jump motion evaluation method according to any one of claims 1-7.
10. A computer device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the human standing long jump motion assessment method according to any one of claims 1-7.