Rope skipping action evaluation method, device and system, server and storage medium

Abstract

The invention provides a rope skipping action evaluation method, device and system, a server and a storage medium, which are applied to the technical field of image processing, and the method comprises the following steps: obtaining a rope skipping video of a target user; determining skeleton point information of the target user in each frame of image of the rope skipping video; target information is determined based on skeleton point information of the target user in each frame of image, and the target information comprises at least two of landing distribution, rope skipping postures and rope skipping counting results of the target user; based on at least two pieces of information of the landing distribution, the rope skipping posture and the rope skipping counting result, the rope skipping action of the target user is evaluated. According to the rope skipping action evaluation method, device and system, the server and the storage medium provided by the invention, the purpose of evaluating the rope skipping action of the user from multiple dimensions is achieved, and the comprehensiveness of the evaluation result is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a method, device, system, server and storage medium for evaluating a rope skipping action. Background Art

[0002] Rope skipping is a common physical exercise and is also one of the optional items in the physical education examinations for junior and senior high schools. As the importance of physical education in national education gradually increases, rope skipping has also received much attention.

[0003] In the prior art, after obtaining the skipping rope data of the user during the skipping rope process, the skipping rope data is compared with the template data corresponding to the pre-set standard skipping rope action template. If the user's skipping rope data matches the template data to a high degree of match, it means that the user's skipping rope action is correct. Otherwise, it means that there is an error in the user's skipping rope action.

[0004] However, the above method can only determine whether the user's rope skipping action is correct, and the evaluation result of the user's rope skipping action is relatively simple. Summary of the invention

[0005] The present invention provides a rope skipping action evaluation method, device, system, server and storage medium, which are used to solve the defect that the evaluation result of the user's rope skipping action in the prior art is relatively single, achieve the purpose of evaluating the user's rope skipping action from multiple dimensions, and improve the comprehensiveness of the evaluation result.

[0006] The present invention provides a rope skipping action evaluation method, comprising: Get the target user's skipping rope video; Determining the skeleton point information of the target user in each frame image of the rope skipping video; Determine target information based on the skeleton point information of the target user in each frame of the image, wherein the target information includes at least two of the landing distribution, rope skipping posture and rope skipping counting result of the target user; The rope skipping action of the target user is evaluated based on at least two of the landing distribution, the rope skipping posture and the rope skipping counting result.

[0007] According to a rope skipping action evaluation method provided by the present invention, the target information includes the rope skipping counting result; The determining target information based on the skeleton point information of the target user in each frame of the image includes: Based on the skeleton point information of the target user in each frame of the image, determining the skeleton point sequence of the target user corresponding to the rope skipping video; Inputting the skeleton point sequence into a rope skipping state classification model to obtain rope skipping state classification results corresponding to each of the images output by the rope skipping state classification model, wherein the rope skipping state classification results include a flying state or a non-flying state; For each rope skipping cycle, if the rope skipping state classification results corresponding to at least two consecutive frames of images in the rope skipping cycle are both in the air state, the rope skipping count is increased by one to obtain the rope skipping count result.

[0008] According to a method for evaluating a rope skipping action provided by the present invention, when the target user is rope skipping, the rope skipping state classification result further includes an interruption detection state; The evaluating the rope skipping action of the target user based on at least two of the landing distribution, the rope skipping posture and the rope skipping counting result includes: In the case where the interruption detection states corresponding to at least two consecutive frames of images are both in the interruption state, the interruption count is increased by one to obtain an interruption count result; Determining the interruption duration based on the number of images corresponding to all target interruption detection states representing the interruption state; Determine the interruption counting result and the interruption duration as rope skipping interruption information; The rope skipping action of the target user is evaluated based on at least two of the landing distribution, the rope skipping posture and the rope skipping counting result, and the rope skipping interruption information.

[0009] According to a rope skipping action evaluation method provided by the present invention, the target information includes the rope skipping posture, and the rope skipping posture includes a take-off height; The determining target information based on the skeleton point information of the target user in each frame of the image includes: For each rope skipping cycle, determining the shoulder joint pixel coordinates and the hip joint pixel coordinates of the target user in each frame image within the rope skipping cycle, and determining the upper body pixel length of the target user; Determining a mapping coefficient based on the upper body pixel length and a preset upper body pixel length of the target user in a non-rope skipping state; Determine a pixel starting height based on the pixel coordinates of the lowest point pixel representing the target user's ankle and the pixel coordinates of the highest point pixel representing the target user's ankle during the rope skipping cycle; The jump-off height is determined based on the pixel jump-off height, the mapping coefficient and the conversion coefficient, and the conversion coefficient is used to characterize the actual distance represented by the pixel.

[0010] According to a method for evaluating a rope skipping action provided by the present invention, the target information includes the rope skipping posture, and the rope skipping posture includes the degree of leg bending; The determining target information based on the skeleton point information of the target user in each frame of the image includes: For each rope skipping cycle, determining a target lower body pixel length of the target user based on the hip joint pixel coordinates and ankle coordinates of the target user in each frame image within the rope skipping cycle; Determining an initial leg bending degree corresponding to the rope skipping cycle based on the target lower body pixel length and a preset lower body pixel length of the target user in a non-rope skipping state; The average value of the initial leg bending degrees corresponding to all the rope skipping cycles is determined as the leg bending degree.

[0011] According to a method for evaluating a rope skipping action provided by the present invention, when the target user is skipping with a rope, the target information also includes the rope length; The evaluating the rope skipping action of the target user based on at least two of the landing distribution, the rope skipping posture and the rope skipping counting result includes: Based on the maximum height of the human rope detection frame in each rope skipping cycle, determine the maximum height average value; Determine the rope length based on the ratio of the human body pixel height of the target user in a non-rope skipping state to the average value of the maximum height; The rope skipping action of the target user is evaluated based on at least two of the landing distribution, the rope skipping posture and the rope skipping counting result, and the rope length.

[0012] According to a rope skipping action evaluation method provided by the present invention, the target information includes the rope skipping posture, and the rope skipping posture includes at least one of a horizontal deviation angle of a wrist joint, a change value of an arm bending angle, and a change value of a shoulder joint abduction angle, the arm bending angle is the angle between an upper arm and a forearm of the target user, and the shoulder joint abduction angle is the angle between an upper arm and a trunk of the target user.

[0013] The present invention also provides a rope skipping action evaluation device, comprising: An acquisition module is used to acquire the rope skipping video of the target user; A determination module, used to determine the skeleton point information of the target user in each frame image of the rope skipping video; The determination module is used to determine target information based on the skeleton point information of the target user in each frame of the image, wherein the target information includes at least two of the landing distribution, rope skipping posture and rope skipping counting result of the target user; An evaluation module is used to evaluate the rope skipping action of the target user based on at least two information of the landing distribution, the rope skipping posture and the rope skipping counting result.

[0014] The present invention also provides a rope skipping action evaluation system, comprising: an image acquisition device, a server and a display device; The image acquisition device is used to acquire a target user's rope skipping video; The server is used to determine the skeleton point information of the target user in each frame of the rope skipping video, determine the target information based on the skeleton point information of the target user in each frame of the image, the target information includes at least two of the landing distribution, the rope skipping posture and the rope skipping counting result of the target user, and evaluate the rope skipping action of the target user based on at least two of the landing distribution, the rope skipping posture and the rope skipping counting result; The display device is used to output the rope skipping action evaluation result.

[0015] The present invention also provides a server, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, any of the above-mentioned rope skipping action evaluation methods is implemented.

[0016] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the skipping action evaluation method as described in any one of the above is implemented.

[0017] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above-mentioned rope skipping action evaluation methods.

[0018] The rope skipping action evaluation method, device, system, server and storage medium provided by the present invention obtain a rope skipping video of a target user, determine the skeleton point information of the target user in each frame image of the rope skipping video, and then determine at least two of the target user's landing distribution, rope skipping posture and rope skipping counting results based on the skeleton point information of the target user in each frame image, so that the rope skipping action of the target user can be evaluated based on the at least two of the landing distribution, rope skipping posture and rope skipping counting results. Since the rope skipping action of the target user can be evaluated based on at least two of the target user's landing distribution, rope skipping posture and rope skipping counting results, the evaluation result is more comprehensive by evaluating the rope skipping action from multiple dimensions. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0020] Figure 1 This is one of the schematic diagrams of the rope skipping action evaluation system provided in an embodiment of the present invention.

[0021] Figure 2 A schematic flow chart of a method for evaluating a rope skipping action provided in an embodiment of the present invention.

[0022] Figure 3 A schematic diagram of rope skipping counting provided by an embodiment of the present invention.

[0023] Figure 4 A schematic diagram of skeleton point information of a target user provided by an embodiment of the present invention.

[0024] Figure 5 A schematic diagram of the structure of a rope skipping action evaluation device provided in an embodiment of the present invention.

[0025] Figure 6 The second schematic diagram of the rope skipping action evaluation system provided in an embodiment of the present invention.

[0026] Figure 7 A schematic diagram of the physical structure of a server provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] It should be noted that the serial numbers assigned to the objects described in the present invention, such as "first", "second", etc., are only used to distinguish the objects described and do not have any order or technical meaning.

[0029] At present, rope skipping is becoming more and more popular among people, and correct rope skipping skills are crucial to achieving the best exercise effect. In the prior art, a standard rope skipping action template is pre-set. When evaluating whether the user's rope skipping action is correct, the collected rope skipping data of the user is compared with the template data corresponding to the standard rope skipping action template to determine whether the user has an incorrect rope skipping action. However, the above method can only determine whether the user's rope skipping action is standard, and the dimension of evaluating the user's rope skipping action is relatively single.

[0030] In view of the above problems, an embodiment of the present invention provides a method for evaluating a rope skipping action. In the method, after a rope skipping video for a target user is collected, at least two of the target user's landing distribution, rope skipping posture, and rope skipping count results can be determined based on the target user's skeletal point information in each frame of the rope skipping video, so that the target user's rope skipping action can be evaluated based on the target user's landing distribution, rope skipping posture, and rope skipping count results. Since the target user's rope skipping action can be evaluated from multiple dimensions, the evaluation result is relatively comprehensive, so that the user's rope skipping action can be guided based on the relatively comprehensive evaluation result.

[0031] Combine the following Figures 1 to 4 The rope skipping action evaluation method provided by the embodiment of the present invention is described. The execution subject of the method can be an electronic device such as a mobile phone, a tablet computer, a smart camera, a computer, a rope skipping action evaluation system, a server or a specially designed rope skipping action evaluation device, or a rope skipping action evaluation device set in the electronic device, and the rope skipping action evaluation device can be implemented by software, hardware or a combination of the two.

[0032] Figure 1 One of the schematic diagrams of the rope skipping action evaluation system provided by an embodiment of the present invention is as follows Figure 1 As shown, the system includes an image collector, a display and a rope skipping action evaluation device. In addition, before the implementation of the scheme of the embodiment of the present invention, the corresponding rope skipping area can be depicted in the target site, for example, the rope skipping area can be marked out by using white paint or the like, and the target user will subsequently jump rope in the rope skipping area. The rope skipping video of the target user is collected by the image collector, wherein the image collector can also be understood as an image acquisition device, for example, it can be a mobile phone, a camera or a camera. After recording the rope skipping video, the image collector sends the collected rope skipping video to the rope skipping action evaluation device, and the rope skipping action evaluation device analyzes the target information of the target user in the rope skipping process based on each frame image in the rope skipping video, such as landing distribution, rope skipping posture, rope skipping counting result, interruption detection state, rope length and other information, thereby evaluating the rope skipping action of the target user based on this information and obtaining an evaluation result. The rope skipping action evaluation device sends the obtained evaluation result to the display, and the display displays the evaluation result. In addition, the rope skipping action evaluation device can also send the above-mentioned target information to the display for display, and the user can see the problems existing in the rope skipping process at a glance based on these process quantities, thereby improving the user experience.

[0033] Figure 2 A schematic diagram of a process flow of a method for evaluating a rope skipping action provided by an embodiment of the present invention is shown in FIG. Figure 2 As shown, the method includes the following: Step 201: Obtain a target user's rope skipping video.

[0034] In this step, when the target user is detected in the designated skipping area, the image collector will start capturing video of the target user. When it is detected that the target user is not in the skipping area, or when the target user finishes skipping, the video capture will be stopped. Based on the above method, the skipping video of the target user can be captured.

[0035] Step 202: Determine the skeleton point information of the target user in each frame image of the rope skipping video.

[0036] In this step, after obtaining the rope skipping video, the position and skeleton point information of the target user in each frame image of the rope skipping video can be determined by the detection model. Wherein, the detection model can be, for example, a Yolopose model. After determining the skeleton point information of the target user, the action of the target user can be drawn based on the position of these skeleton point information, thereby reflecting the rope skipping action of the target user. Wherein, the skeleton points at least include key nodes such as the head, wrist joint, hip joint, knee joint and ankle.

[0037] Among them, the Yolopose model is fine-tuned on the basis of the original Yolopose model by collecting images of rope skipping in different scenes to optimize the detection effect of human skeleton points in rope skipping scenes. Better human skeleton points indirectly affect the rope skipping state classification effect based on skeleton points, thereby affecting the final counting effect and the accuracy of process quantity diagnosis.

[0038] Step 203: Determine target information based on the skeleton point information of the target user in each frame image, where the target information includes at least two of the landing distribution, rope skipping posture and rope skipping counting result of the target user.

[0039] In this step, after determining the target user's multiple skeleton point information, the target user's landing position can be determined based on the ankle skeleton point, for example, so that the target user's landing distribution can be determined based on the landing position. Specifically, the minimum heel skeleton point position can be used as the landing point, and the landing point corresponding to each rope skipping cycle can be recorded in sequence, and drawn into an image of a fixed size to view the landing distribution. It should be understood that the more discrete the landing distribution is, the more it will affect the target user's rope skipping speed to a certain extent, and the landing distribution can reflect the stability of the target user's rope skipping technique. In this embodiment, the discrete degree of the landing point is represented by calculating the variance of all landing points. When the variance is greater than the preset variance, such as greater than 100, it means that the landing is not stable enough and is relatively scattered, indicating that the landing stability needs to be further improved, so that targeted output suggestions can be given to prompt the target user to make improvements.

[0040] The rope skipping posture can also be understood as the rope skipping posture and state of the target user, such as the horizontal deviation of the wrist joint, the degree of leg bending, the arm bending angle or the shoulder joint abduction angle, etc., and can also include any other posture that can affect the rope skipping speed. The rope skipping posture can clearly and intuitively reflect whether the target user has a wrong posture during the rope skipping process, so as to diagnose the rope skipping posture of the target user.

[0041] In addition, the skipping rope counting result can also be determined based on the target user's skeleton point information. The skipping rope counting result is an intuitive reflection of the target user's skipping rope result. Therefore, the larger the skipping rope counting result, the more standardized the target user's skipping rope action is. Conversely, it means that the target user's skipping rope action needs to be corrected.

[0042] It should be noted that at least two of the landing distribution, rope skipping posture and rope skipping counting results can be determined through the target user's skeletal point information, and the at least two pieces of information can be fused, so that rope skipping action evaluation results in multiple dimensions can be obtained based on the fusion results.

[0043] Step 204: Evaluate the rope skipping action of the target user based on at least two of the information of the landing distribution, the rope skipping posture and the rope skipping counting result.

[0044] In this step, at least two of the information of the landing distribution, the rope skipping posture, and the rope skipping count result can be displayed to the user in a visual display manner, such as displaying a heat map of the landing distribution by color coding, or displaying a simulation of the rope skipping posture by animation, and displaying the final rope skipping count result of the target user. After displaying the above information in a visual manner, the target user can intuitively see the problems he or she has in the rope skipping process, so that the user experience is better.

[0045] In addition, a comprehensive evaluation model can be constructed to comprehensively evaluate the target user's rope skipping action based on at least two pieces of information including landing distribution, rope skipping posture, and rope skipping counting results, as well as a pre-constructed comprehensive evaluation model. Furthermore, based on the obtained evaluation results, improvement suggestions for rope skipping actions can be provided to the user, such as adjusting the landing method, improving body posture, etc.

[0046] The rope skipping action evaluation method provided by the embodiment of the present invention obtains the rope skipping video of the target user, and after determining the skeleton point information of the target user in each frame image of the rope skipping video, determines at least two information of the target user's landing distribution, rope skipping posture and rope skipping counting result based on the skeleton point information of the target user in each frame image, so that the rope skipping action of the target user can be evaluated based on at least two information of the landing distribution, rope skipping posture and rope skipping counting result. Since the rope skipping action of the target user can be evaluated based on at least two information of the landing distribution, rope skipping posture and rope skipping counting result of the target user, the rope skipping action is evaluated from multiple dimensions, so that the evaluation result is more comprehensive.

[0047] Exemplarily, the above-mentioned target information includes a skipping rope counting result. When the target information is determined based on the skeleton point information of the target user in each frame image, the skeleton point sequence of the target user corresponding to the skipping rope video can be determined based on the skeleton point information of the target user in each frame image, and the skeleton point sequence is input into the skipping rope state classification model to obtain the skipping rope state classification results corresponding to each image output by the skipping rope state classification model. The skipping rope state classification results include a floating state or a non-floating state. For each skipping rope cycle, if there are at least two consecutive frames of images within the skipping rope cycle and the skipping rope state classification results corresponding to both are the floating state, the skipping rope count is increased by one to obtain the skipping rope counting result.

[0048] Specifically, after the skeleton point information of the target user in each frame image is obtained through the trained Yolopose model, based on the skeleton point information, the single-person skeleton point sequence of the target user corresponding to the rope skipping video can be determined through point matching.

[0049] Furthermore, the obtained skeleton point sequence can be input into the trained rope skipping state classification model to obtain the rope skipping state classification result corresponding to each frame image output by the rope skipping state classification model, and the rope skipping state classification result includes a floating state or a non-floating state, wherein the floating state can be understood as a state where the feet of the target user leave the ground, and the non-floating state can be understood as a state where the feet of the target user do not leave the ground. The rope skipping state classification model can be, for example, a gated recurrent unit (GRU). The GRU model is a high-efficiency model based on a recurrent neural network (RNN) and has the ability to obtain historical information. Combining historical information can obtain better classification results. The skeleton point sequence is used as the input of the trained GRU model, so that the rope skipping state classification result corresponding to each image output by the GRU model can be obtained, that is, whether the target user is in a floating state or a non-floating state in each frame image. Among them, by whether the target user is in a floating state or a non-floating state, it can also be judged whether the target user is currently in a rope skipping state.

[0050] For each rope skipping cycle, each frame image in the rope skipping cycle will obtain the rope skipping state classification result through the rope skipping state classification model. Since the rope skipping state may be repeated, the embodiment of the present invention uses deduplication logic to obtain the unique state of the current rope skipping cycle, and finally sums up all the states of all rope skipping cycles to obtain the overall rope skipping count result. Figure 3 A schematic diagram of rope skipping counting provided by an embodiment of the present invention, such as Figure 3 As shown, assuming that time T to time T+9 belongs to a rope skipping cycle, among which frames T+3 to T+7 are all in the air state (1 represents the air state, and 0 represents the non-air state). After deduplication, the actual number of rope skipping in the current rope skipping cycle can be obtained as 1. The number of all rope skipping cycles is added together to get the final rope skipping counting result.

[0051] Furthermore, in order to prevent occasional misidentification of the rope skipping state classification model and increase the accuracy of the count, the rope skipping count is increased when the rope skipping state exists for at least 2 consecutive frames, otherwise it is not counted. In addition, if an interruption information is detected in the current rope skipping cycle, but the rope skipping state occurs, the count is not performed, thereby effectively avoiding the problem of inaccurate counts caused by jumping during the interruption.

[0052] The skipping cycle may be understood as the duration from the start of a jump to the next jump of the target user.

[0053] In this embodiment, the skeleton point sequence of the target user corresponding to the rope skipping video can be determined based on the skeleton point information of the target user in each frame image, and the skeleton point sequence can be input into the rope skipping state classification model. After the rope skipping state classification result is obtained, in each rope skipping cycle, when there are at least two consecutive frames of images corresponding to the rope skipping state classification results that are both in the air state, the rope skipping count is increased by one, thereby obtaining the final rope skipping count result. Since the rope skipping count is increased only when there are at least two consecutive frames of images corresponding to the rope skipping state classification results that are both in the air state, the situation where the counting statistics are wrong due to misidentification can be avoided, thereby improving the accuracy of the rope skipping count result.

[0054] Exemplarily, based on the above embodiments, when the target user is skipping with a rope, the skipping state classification result also includes an interruption detection state. Then, when evaluating the target user's skipping action based on at least two information among the landing distribution, the skipping posture and the skipping count result, it may include: when there are at least two consecutive frames of images whose corresponding interruption detection states are both interruption states, the interruption count is increased by one to obtain an interruption count result, the interruption duration is determined based on the number of images corresponding to all target interruption detection states that are characterized as being in an interruption state, and the interruption count result and the interruption duration are determined as the skipping interruption information, and the target user's skipping action is evaluated based on the landing distribution, the skipping posture and at least two information among the skipping count result, and the skipping interruption information.

[0055] Specifically, after the skeleton point sequence is input into the rope skipping state classification model, the obtained rope skipping state classification result also includes an interruption detection state, and the interruption detection state includes whether an interruption occurs or whether an interruption does not occur.

[0056] After obtaining the interruption detection state corresponding to each frame image through the skipping rope state classification model, similar to the skipping rope counting process, since the interruption state may be repeated, the embodiment of the present invention uses deduplication logic. When there are at least two consecutive frames of images corresponding to the interruption detection state that are all interruption states within a skipping rope cycle, the interruption state will be merged to increase the interruption count by one. By counting the results in all skipping rope cycles, the interruption counting result can be obtained, that is, the total number of interruptions of the target user from the start to the end of skipping rope.

[0057] In addition, the interruption duration can also be determined based on the interruption detection state corresponding to each frame of the image by accumulating the number of all images that are in the interruption state. The determined interruption count result and the interruption duration are determined as the rope skipping interruption information, so that the rope skipping action of the target user can be evaluated based on the landing distribution, the rope skipping posture and at least two of the rope skipping count results, and the rope skipping interruption information. It should be understood that the more interruptions and the longer the interruption duration, the more the target user's rope skipping skills need to be further improved to reduce the number of interruptions and shorten the interruption duration.

[0058] In this embodiment, by determining the skipping rope interruption information including the interruption counting result and the interruption duration, and based on at least two information of the landing distribution, the skipping rope posture and the skipping rope counting result, and the skipping rope interruption information, the skipping rope action of the target user is evaluated, thereby achieving the purpose of evaluating the skipping rope action of the target user from multiple dimensions and angles, and improving the comprehensiveness of the evaluation results.

[0059] Exemplarily, based on the above embodiments, the above target information includes a rope skipping posture, and the rope skipping posture includes a jumping height; when determining the target information based on the skeleton point information of the target user in each frame image, it can be performed in the following manner: For each skipping cycle, the shoulder joint pixel coordinates and hip joint pixel coordinates of the target user in each frame image within the skipping cycle are determined, the upper body pixel length of the target user is determined, and the mapping coefficient is determined based on the upper body pixel length and the preset upper body pixel length of the target user in the non-skipping state. The pixel take-off height is determined based on the pixel coordinates of the lowest point pixel representing the target user's ankle and the pixel coordinates of the highest point pixel representing the target user's ankle within the skipping cycle, so as to determine the take-off height based on the pixel take-off height, the mapping coefficient and the conversion coefficient. The conversion coefficient is used to characterize the actual distance represented by the pixel.

[0060] Specifically, the take-off height refers to the distance between the target user's ankle point and the ground after taking off during rope skipping.

[0061] In actual application, the pixel length of the upper body of the target user in the image coordinate system during rope skipping can be determined based on the following formula (1): : (1)

[0062] in, represents the pixel coordinates of the shoulder joint, Represents the pixel coordinates of the hip joint.

[0063] Determine the pixel length of the target user's upper body during rope skipping After that, the mapping coefficient can be determined based on formula (2) : (2)

[0064] Among them, the mapping coefficient It can also be understood as a mapping coefficient that is larger when near and smaller when far, which can prevent calculation errors caused by larger when near and smaller when far. Indicates the preset upper body pixel length of the target user when not skipping rope. It can be determined by an image captured before the target user starts jumping rope, or by an image captured when the target user finishes jumping rope. The preset upper body pixel length can be obtained by subtracting the hip joint pixel coordinates from the shoulder joint pixel coordinates when the target user is not in a rope skipping state.

[0065] In addition, the difference between the ankle point at the highest moment of take-off and the ankle point at the lowest moment of landing in a rope skipping cycle can be obtained to represent the image take-off height, that is, the pixel take-off height. Specifically, the pixel take-off height can be determined by the following formula (3): : (3)

[0066] in, Indicates the pixel coordinates of the highest point pixel of the target user's ankle, Represents the pixel coordinates of the lowest point pixel of the target user's ankle.

[0067] In order to change the pixel jump height in the image coordinate system To restore to the real-world coordinate system, we also need to determine the conversion coefficient from the image to the real-world coordinate system. Among them, the conversion factor It can be determined according to the following formula (4): (4)

[0068] in, It represents the diameter of the rope skipping area in the image coordinate system, which can be obtained by the maximum vertical coordinate pixel value of the rope skipping area. Subtract the minimum vertical coordinate pixel value of the skipping area get, It represents the diameter of the skipping rope area in the real world. This value can be obtained when the skipping rope area is defined.

[0069] In addition, in order to automatically obtain the location of the rope skipping area in the image, it is necessary to collect landmark data of different scenes in advance and train an automatic site recognition model based on Convolutional Neural Networks (CNN). Through this site recognition model, the maximum vertical coordinate pixel value of the rope skipping area can be determined. and the minimum vertical coordinate pixel value of the skipping area It should be understood that since the site recognition model is usually only used once after the site deployment is completed, the determined conversion coefficient can be used directly later. Therefore, in order to obtain a more accurate detection effect, in this embodiment, the YOLOV9-E detection model with a larger number of parameters can be used for training and testing.

[0070] Furthermore, the take-off height in the actual coordinate system can be determined based on the following formula (5): : (5)

[0071] Determine the starting height corresponding to each skipping cycle After that, the starting height of multiple rope skipping cycles can be further determined. The average value of continuous rope skipping is If the average value is greater than the preset height, the skipping speed will be affected to a certain extent, and the target user needs to be reminded to lower the skipping height. For example, if the jump height is higher than 10cm within 1 second, the target user is advised to lower the jump height.

[0072] In this embodiment, the mapping coefficient is determined by the pixel length of the upper body of the target user in the skipping state and the preset upper body pixel length in the non-skipping state. The mapping coefficient can avoid the problem of inaccurate jumping height due to the change of the distance between the target user and the image collector caused by the movement of the target user during the skipping process. By correcting the jumping height, the accuracy of the jumping height can be improved.

[0073] Exemplarily, based on the above embodiment, when the target information includes rope skipping posture, and the rope skipping posture includes the degree of leg bending, the target information can be determined based on the skeleton point information of the target user in each frame image in the following manner: For each skipping cycle, the target lower body pixel length of the target user is determined based on the hip joint pixel coordinates and ankle coordinates of the target user in each frame image within the skipping cycle, and the initial leg bending degree corresponding to the skipping cycle is determined based on the target lower body pixel length and the preset lower body pixel length of the target user in the non-skipping state, and the average value of the initial leg bending degrees corresponding to all skipping cycles is determined as the leg bending degree.

[0074] Specifically, for each rope skipping cycle, the hip joint pixel coordinates and ankle coordinates of the target user in each frame image in the current rope skipping cycle can be obtained in the rope skipping stage, the lower body pixel length of the target user in each frame image can be determined, and the minimum lower body pixel length among the lower body pixel lengths corresponding to all images can be determined as the target lower body pixel length. In addition, based on the images collected by the user in the preparation stage or the end of the rope skipping stage, the preset lower body pixel length of the target user in the non-rope skipping state can be obtained, and the ratio of the target lower body pixel length to the preset lower body pixel length can be determined as the initial leg bending degree corresponding to the rope skipping cycle.

[0075] Furthermore, the average value of the initial leg bending degrees corresponding to all rope skipping cycles may be used as the final leg bending degree, and the leg bending degree may be used as an evaluation index for the rope skipping of the target user.

[0076] It should be understood that the evaluation indicator of leg bending degree is closely related to the take-off height. Generally speaking, the higher the leg bending degree, the higher the take-off height, but the higher the take-off height does not necessarily mean a greater degree of leg bending (such as upright take-off). Therefore, the two indicators are used in combination. For example, when the take-off height is high and the leg bending degree is less than the preset bending degree, such as 0.6, it is recommended to reduce the take-off height and the degree of leg bending when jumping.

[0077] Exemplarily, based on the above embodiments, when the target user is skipping with a rope, the target information also includes the length of the rope. When evaluating the skipping action of the target user based on at least two of the information of the landing distribution, the skipping posture and the skipping count result, the following method can be used: Based on the maximum height of the human rope detection frame in each skipping cycle, the maximum height average value is determined; based on the ratio of the human pixel height of the target user in the non-skipping state and the maximum height average value, the rope length is determined; based on at least two information of the landing distribution, skipping posture and skipping counting results, and the rope length, the skipping action of the target user is evaluated.

[0078] Specifically, when the skeleton point information is obtained through the Yolopose model, the union of the target user area and the skipping rope area is also output through the Yolopose model, that is, the target user and the skipping rope can be selected in the output target frame. It should be noted that in order to prevent cheating in ropeless skipping, the output detection frame contains attribute information of whether the rope is held. If the attribute information indicates holding the rope, it means that the target user and the skipping rope are included in the detection frame. If the attribute information indicates not holding the rope, it means that only the target user is included in the detection frame.

[0079] When the target user is holding a rope and skipping, the maximum height of the human rope detection frame output by the Yolopose model in each skipping cycle will be obtained, and the maximum height average of the human rope detection frame in all skipping cycles will be determined. It should be understood that when the target user is skipping, when the skipping rope reaches the position above the target user's head, the height of the human rope detection frame should be the largest. In addition, based on the image collected by the target user in the preparation stage or at the end of the skipping stage, the human body pixel height of the target user in the non-skipping state can be obtained, wherein the human body pixel height can be determined by the difference between the pixel coordinates of the target user's head bone point and the pixel coordinates of the ankle bone point.

[0080] Furthermore, the ratio between the human body pixel height and the average value of the maximum height may be determined, and the rope length may be determined by the ratio.

[0081] When evaluating the target user's rope skipping action, the evaluation can be based on at least two of the information of landing distribution, rope skipping posture and rope skipping counting results, as well as the rope length. Taking the rope length as an evaluation indicator, when the above ratio is less than a preset value, such as less than 0.7, the target user can be reminded that the rope length needs to be further adjusted, because a too long rope affects the speed and rhythm, and increases the risk of stepping on the rope during jumping, thereby causing tripping or loss of balance and increasing the risk of injury.

[0082] In this embodiment, the rope length can be determined based on the ratio of the target user's human pixel height in the non-rope skipping state and the average maximum height of the human rope detection frame, and the rope length can be used as an evaluation indicator of the rope skipping action, thereby making the evaluation result more comprehensive and accurate.

[0083] Exemplarily, based on the above embodiments, when the target information includes a rope skipping posture, the rope skipping posture may also include at least one of a horizontal deviation angle of the wrist joint, a change value of an arm bending angle, and a change value of a shoulder joint abduction angle, the arm bending angle being the angle between the upper arm and forearm of the target user, and the shoulder joint abduction angle being the angle between the upper arm and torso of the target user.

[0084] in, Figure 4 A schematic diagram of the skeleton point information of a target user provided by an embodiment of the present invention is shown in FIG. Figure 4 As shown, in addition to the jumping height and the degree of leg bending, the rope skipping posture also includes at least one of the horizontal deviation angle of the wrist joint, the change value of the arm bending angle and the change value of the shoulder joint abduction angle.

[0085] For the horizontal deviation angle of the wrist joint, the cosine value of the vector angle of the wrist bone points of both hands in each frame image is calculated to obtain the horizontal deviation angle of the wrist joint in each frame image, and the average value of the horizontal deviation angle of the wrist joint corresponding to all frame images is used as the final measurement indicator. Since the inconsistent height of the left and right hands will affect the skipping effect, for example, when the average value of the horizontal deviation angle of the wrist joint is greater than 0.5 degrees, it is necessary to output a prompt message to the target user to remind the target user to pay attention to keeping both hands on the same horizontal line as much as possible.

[0086] For the change value of the arm bending angle, for each frame of the image, the angle between the upper arm and the forearm of the target user can be determined based on the position information of the shoulder bone point, the elbow bone point and the wrist bone point in the image, and the difference between the maximum angle and the minimum angle in each skipping cycle is calculated. The difference is used as the change value of the arm bending angle in the skipping cycle, and the change values ​​of the arm bending angle in all cycles are averaged to obtain the change value of the average arm bending angle in the entire skipping stage.

[0087] For the change value of the shoulder abduction angle, for each frame of the image, the angle between the target user's upper arm and torso can be determined based on the position information of the shoulder bone point, the elbow bone point and the hip bone point in the image, and the difference between the maximum angle and the minimum angle in each skipping cycle is calculated. The difference is used as the change value of the shoulder abduction angle in the skipping cycle, and the change values ​​of the shoulder abduction angle in all cycles are averaged to obtain the average change value of the shoulder abduction angle in the entire skipping stage.

[0088] The larger the change value of the shoulder joint abduction angle and the change value of the arm bending angle are, the more likely it is that the target user is not stable enough when jumping rope with both hands, which will affect the final result of the rope jumping to a certain extent. Therefore, at least one of the wrist joint horizontal deviation angle, the change value of the arm bending angle, and the change value of the shoulder joint abduction angle can be output in a visual display manner, and corresponding suggestion information can be generated to remind the target user to make corrections.

[0089] In this embodiment, the rope skipping posture also includes at least one of the change value of the horizontal deviation angle of the wrist joint, the change value of the arm bending angle, and the change value of the shoulder joint abduction angle. The rope skipping action of the target user is evaluated through this information, making the evaluation result more comprehensive and improving the accuracy of the evaluation result.

[0090] On the basis of the above embodiments, in addition to the above evaluation indicators, the stability of the skipping rope counting results can also be counted. Assuming that the target user skips rope for a total of one minute, the skipping rope counting results can be counted cycle by cycle with a preset time length, such as 4 seconds as a cycle, and the stability of the skipping rope counting results can be counted as the cycle progresses. This evaluation indicator can measure the physical strength distribution of the target user. For example, as time goes by, the number of skipping ropes in the same time gradually decreases, indicating that the target user is physically exhausted, and suggestion information can be output to suggest that the target user better distribute physical strength or need to further strengthen physical fitness training.

[0091] The following is a description of a rope skipping action evaluation device provided in an embodiment of the present invention. The rope skipping action evaluation device described below and the rope skipping action evaluation method described above can be referenced to each other.

[0092] Figure 5 This is a schematic diagram of the structure of the rope skipping action evaluation device provided by an embodiment of the present invention, referring to Figure 5 As shown, the rope skipping action evaluation device 500 includes: An acquisition module 11 is used to acquire a target user's rope skipping video; A determination module 12, used to determine the skeleton point information of the target user in each frame image of the rope skipping video; The determination module 12 is used to determine target information based on the skeleton point information of the target user in each frame of the image, wherein the target information includes at least two of the landing distribution, rope skipping posture and rope skipping counting result of the target user; The evaluation module 13 is used to evaluate the rope skipping action of the target user based on at least two information of the landing distribution, the rope skipping posture and the rope skipping counting result.

[0093] In an exemplary embodiment, the target information includes the rope skipping counting result; The determining module 12 is specifically used for: Based on the skeleton point information of the target user in each frame of the image, determining the skeleton point sequence of the target user corresponding to the rope skipping video; Inputting the skeleton point sequence into a rope skipping state classification model to obtain rope skipping state classification results corresponding to each of the images output by the rope skipping state classification model, wherein the rope skipping state classification results include a flying state or a non-flying state; For each rope skipping cycle, if the rope skipping state classification results corresponding to at least two consecutive frames of images in the rope skipping cycle are both in the air state, the rope skipping count is increased by one to obtain the rope skipping count result.

[0094] In an exemplary embodiment, when the target user is skipping with a rope, the skipping state classification result further includes an interruption detection state; The evaluation module 13 is specifically used to: In the case where the interruption detection states corresponding to at least two consecutive frames of images are both in the interruption state, the interruption count is increased by one to obtain an interruption count result; Determining the interruption duration based on the number of images corresponding to all target interruption detection states representing the interruption state; Determine the interruption counting result and the interruption duration as rope skipping interruption information; The rope skipping action of the target user is evaluated based on at least two of the landing distribution, the rope skipping posture and the rope skipping counting result, and the rope skipping interruption information.

[0095] In an exemplary embodiment, the target information includes the rope skipping posture, and the rope skipping posture includes a jump-off height; The determining module 12 is specifically used for: For each rope skipping cycle, determining the shoulder joint pixel coordinates and the hip joint pixel coordinates of the target user in each frame image within the rope skipping cycle, and determining the upper body pixel length of the target user; Determining a mapping coefficient based on the upper body pixel length and a preset upper body pixel length of the target user in a non-rope skipping state; Determine a pixel starting height based on the pixel coordinates of the lowest point pixel representing the target user's ankle and the pixel coordinates of the highest point pixel representing the target user's ankle during the rope skipping cycle; The jump-off height is determined based on the pixel jump-off height, the mapping coefficient and the conversion coefficient, and the conversion coefficient is used to characterize the actual distance represented by the pixel.

[0096] In an exemplary embodiment, the target information includes the rope skipping posture, and the rope skipping posture includes the degree of leg bending; The determining module 12 is specifically used for: For each rope skipping cycle, determining a target lower body pixel length of the target user based on the hip joint pixel coordinates and ankle coordinates of the target user in each frame image within the rope skipping cycle; Determining an initial leg bending degree corresponding to the rope skipping cycle based on the target lower body pixel length and a preset lower body pixel length of the target user in a non-rope skipping state; The average value of the initial leg bending degrees corresponding to all the rope skipping cycles is determined as the leg bending degree.

[0097] In one example embodiment, when the target user is skipping with a rope, the target information also includes the length of the rope; The evaluation module 13 is specifically used to: Based on the maximum height of the human rope detection frame in each rope skipping cycle, determine the maximum height average value; Determine the rope length based on the ratio of the human body pixel height of the target user in a non-rope skipping state to the average value of the maximum height; The rope skipping action of the target user is evaluated based on at least two of the landing distribution, the rope skipping posture and the rope skipping counting result, and the rope length.

[0098] In an exemplary embodiment, the target information includes the rope skipping posture, and the rope skipping posture includes at least one of a horizontal deviation angle of the wrist joint, a change value of an arm bending angle, and a change value of a shoulder joint abduction angle, wherein the arm bending angle is the angle between the upper arm and forearm of the target user, and the shoulder joint abduction angle is the angle between the upper arm and torso of the target user.

[0099] The device of this embodiment can be used to execute the method of any embodiment in the embodiment of the rope skipping action evaluation method side. Its specific implementation process and technical effects are similar to those in the embodiment of the rope skipping action evaluation method side. For details, please refer to the detailed description in the embodiment of the rope skipping action evaluation method side, which will not be repeated here.

[0100] Figure 6 The second schematic diagram of the rope skipping action evaluation system provided by the embodiment of the present invention is as follows Figure 6 As shown, the rope skipping action evaluation system includes: an image acquisition device, a server and a display device; the image acquisition device is used to collect the rope skipping video of the target user; the server is used to determine the skeleton point information of the target user in each frame of the rope skipping video, and based on the skeleton point information of the target user in each frame of the image, determine the target information, the target information includes at least two of the landing distribution, rope skipping posture and rope skipping counting result of the target user, and evaluate the rope skipping action of the target user based on the landing distribution, the rope skipping posture and the rope skipping counting result; the display device is used to output the rope skipping action evaluation result. The rope skipping action evaluation system of this embodiment can be used to execute the method of any embodiment in the embodiment of the rope skipping action evaluation method side, and its specific implementation process and technical effects are similar to those in the embodiment of the rope skipping action evaluation method side. For details, please refer to the detailed introduction in the embodiment of the rope skipping action evaluation method side, which will not be repeated here.

[0101] Figure 7 A schematic diagram of the physical structure of a server provided in an embodiment of the present invention, such as Figure 7As shown, the server may include: a processor 710, a communication interface 720, a memory 730 and a communication bus 740, wherein the processor 710, the communication interface 720 and the memory 730 communicate with each other through the communication bus 740. The processor 710 may call the logic instructions in the memory 730 to execute the rope skipping action evaluation method, which includes: obtaining a rope skipping video of a target user; determining the skeleton point information of the target user in each frame of the rope skipping video; determining target information based on the skeleton point information of the target user in each frame of the image, the target information including at least two of the landing distribution, the rope skipping posture and the rope skipping counting result of the target user; and evaluating the rope skipping action of the target user based on at least two of the landing distribution, the rope skipping posture and the rope skipping counting result.

[0102] In addition, the logic instructions in the above-mentioned memory 730 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0103] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the skipping action evaluation method provided by the above-mentioned methods, which includes: obtaining a skipping video of a target user; determining the skeleton point information of the target user in each frame image of the skipping video; based on the skeleton point information of the target user in each frame of the image, determining target information, the target information including at least two of the landing distribution, skipping posture and skipping count results of the target user; based on at least two of the landing distribution, the skipping posture and the skipping count results, evaluating the skipping action of the target user.

[0104] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the skipping rope action evaluation method provided by the above-mentioned methods, the method comprising: obtaining a skipping rope video of a target user; determining the skeleton point information of the target user in each frame image of the skipping rope video; determining target information based on the skeleton point information of the target user in each frame of the image, the target information comprising at least two of the landing distribution, skipping rope posture and skipping rope counting result of the target user; evaluating the skipping rope action of the target user based on at least two of the landing distribution, the skipping rope posture and the skipping rope counting result.

[0105] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0106] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0107] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for evaluating a rope skipping action, characterized in that: include: Get the target user's skipping rope video; Determining the skeleton point information of the target user in each frame image of the rope skipping video; Determine target information based on the skeleton point information of the target user in each frame of the image, wherein the target information includes at least two of the landing distribution, rope skipping posture and rope skipping counting result of the target user; The rope skipping action of the target user is evaluated based on at least two of the landing distribution, the rope skipping posture and the rope skipping counting result.

2. The rope skipping action evaluation method according to claim 1, characterized in that: The target information includes the rope skipping counting result; The determining target information based on the skeleton point information of the target user in each frame of the image includes: Based on the skeleton point information of the target user in each frame of the image, determining the skeleton point sequence of the target user corresponding to the rope skipping video; Inputting the skeleton point sequence into a rope skipping state classification model to obtain rope skipping state classification results corresponding to each of the images output by the rope skipping state classification model, wherein the rope skipping state classification results include a flying state or a non-flying state; For each rope skipping cycle, if the rope skipping state classification results corresponding to at least two consecutive frames of images in the rope skipping cycle are both in the air state, the rope skipping count is increased by one to obtain the rope skipping count result.

3. The rope skipping action evaluation method according to claim 2, characterized in that: When the target user is skipping with a rope, the skipping state classification result further includes an interruption detection state; The evaluating the rope skipping action of the target user based on at least two of the landing distribution, the rope skipping posture and the rope skipping counting result includes: In the case where the interruption detection states corresponding to at least two consecutive frames of images are both in the interruption state, the interruption count is increased by one to obtain an interruption count result; Determining the interruption duration based on the number of images corresponding to all target interruption detection states representing the interruption state; Determine the interruption counting result and the interruption duration as rope skipping interruption information; The rope skipping action of the target user is evaluated based on at least two of the landing distribution, the rope skipping posture and the rope skipping counting result, and the rope skipping interruption information.

4. The rope skipping action evaluation method according to claim 1, characterized in that: The target information includes the rope skipping posture, and the rope skipping posture includes a jumping height; The determining target information based on the skeleton point information of the target user in each frame of the image includes: For each rope skipping cycle, determining the shoulder joint pixel coordinates and the hip joint pixel coordinates of the target user in each frame image within the rope skipping cycle, and determining the upper body pixel length of the target user; Determining a mapping coefficient based on the upper body pixel length and a preset upper body pixel length of the target user in a non-rope skipping state; Determine a pixel starting height based on the pixel coordinates of the lowest point pixel representing the target user's ankle and the pixel coordinates of the highest point pixel representing the target user's ankle during the rope skipping cycle; The jump-off height is determined based on the pixel jump-off height, the mapping coefficient and the conversion coefficient, and the conversion coefficient is used to characterize the actual distance represented by the pixel.

5. The rope skipping action evaluation method according to claim 1, characterized in that: The target information includes the rope skipping posture, and the rope skipping posture includes the degree of leg bending; The determining target information based on the skeleton point information of the target user in each frame of the image includes: For each rope skipping cycle, determining a target lower body pixel length of the target user based on the hip joint pixel coordinates and ankle coordinates of the target user in each frame image within the rope skipping cycle; Determining an initial leg bending degree corresponding to the rope skipping cycle based on the target lower body pixel length and a preset lower body pixel length of the target user in a non-rope skipping state; The average value of the initial leg bending degrees corresponding to all the rope skipping cycles is determined as the leg bending degree.

6. The rope skipping action evaluation method according to claim 1, characterized in that: When the target user is skipping with a rope, the target information also includes the length of the rope; The evaluating the rope skipping action of the target user based on at least two of the landing distribution, the rope skipping posture and the rope skipping counting result includes: Based on the maximum height of the human rope detection frame in each rope skipping cycle, determine the maximum height average value; Determine the rope length based on the ratio of the human body pixel height of the target user in a non-rope skipping state to the average value of the maximum height; The rope skipping action of the target user is evaluated based on at least two of the landing distribution, the rope skipping posture and the rope skipping counting result, and the rope length.

7. The method for evaluating the rope skipping action according to any one of claims 1 to 6, characterized in that: The target information includes the rope skipping posture, and the rope skipping posture includes at least one of a horizontal deviation angle of a wrist joint, a change value of an arm bending angle, and a change value of a shoulder joint abduction angle, wherein the arm bending angle is an angle between an upper arm and a forearm of the target user, and the shoulder joint abduction angle is an angle between an upper arm and a trunk of the target user.

8. A rope skipping action evaluation device, characterized in that: include: An acquisition module is used to acquire the rope skipping video of the target user; A determination module, used to determine the skeleton point information of the target user in each frame image of the rope skipping video; The determination module is used to determine target information based on the skeleton point information of the target user in each frame of the image, wherein the target information includes at least two of the landing distribution, rope skipping posture and rope skipping counting result of the target user; An evaluation module is used to evaluate the rope skipping action of the target user based on at least two information of the landing distribution, the rope skipping posture and the rope skipping counting result.

9. A rope skipping action evaluation system, characterized in that: include: Image acquisition devices, servers and display equipment; The image acquisition device is used to acquire a target user's rope skipping video; The server is used to determine the skeleton point information of the target user in each frame of the rope skipping video, determine the target information based on the skeleton point information of the target user in each frame of the image, the target information includes at least two of the landing distribution, the rope skipping posture and the rope skipping counting result of the target user, and evaluate the rope skipping action of the target user based on at least two of the landing distribution, the rope skipping posture and the rope skipping counting result; The display device is used to output the rope skipping action evaluation result.

10. A server comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the rope skipping action evaluation method according to any one of claims 1 to 7 is implemented.

11. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the rope skipping action evaluation method according to any one of claims 1 to 7 is implemented.

Images