Upward body training method and training device

Abstract

The application provides a pull-up training method and a training device, and relates to a data processing technology, wherein the training method comprises the following steps: triggering identity collection of a collection unit, and performing primary identity verification based on obtained first collection information; in response to determining that the verification is passed, determining current strength level data based on the first collection information, and selecting a suitable training course from a course library based on the strength level data; in response to receiving a trigger signal corresponding to any training course, determining a training stage in an unplayed state as a target stage, triggering the collection unit to perform identity collection based on the target stage, and performing secondary identity verification based on obtained second collection information; in response to determining that the verification is passed based on the secondary identity verification, playing a display video corresponding to the target stage based on the display unit, and controlling the collection unit to perform posture collection, so as to determine a standard training score based on obtained posture information. The application at least improves training efficiency.

Description

Technical Field

[0001] This invention relates to data processing technology, and more particularly to a pull-up training method and training equipment. Background Technology

[0002] Pull-ups, as an exercise that combines strength training and physical fitness assessment, are widely used in fitness and shaping, military and police physical fitness training, and student physical fitness testing. Their training effect is directly related to the improvement of upper limb and core strength, as well as the achievement of training goals. Whether fitness enthusiasts are pursuing muscle growth or military and police personnel are improving their physical fitness pass rate, scientific training methods are key. However, pull-ups require high upper limb strength, and there are significant differences in the basic strength of different trainees. Moreover, the correctness of posture during training directly affects the training effect and exercise safety. How to develop a plan for individual differences and ensure the scientific nature of training has become the core demand for improving the training effect of pull-ups.

[0003] Currently, pull-up training methods are mainly unguided self-training models. Trainees rely on general fitness tutorials or experience to practice, lacking accurate assessment of their own strength level, which often leads to a mismatch between training intensity and their own ability.

[0004] Therefore, there is an urgent need to provide a pull-up training method and equipment that can improve training efficiency. Summary of the Invention

[0005] In view of the above problems, the present invention is proposed to provide a pull-up training method and training device that overcomes or at least partially solves the above problems.

[0006] According to one aspect of the present invention, a pull-up training method adapted to the pull-up training device is provided, comprising the following steps: The user terminal selects any training mode based on the display unit, triggering the collection unit to collect the user terminal's identity and perform primary identity verification based on the collected information. Once the response confirms that the verification has passed, the current strength level data of the corresponding user terminal is determined based on the information collected once, and a training course that matches the strength level data is selected from the retrieved course library based on the strength level data. Upon receiving a trigger signal corresponding to any training course, the training stage included in the training course and located first in the training order that is in an unplayed state is identified as the target stage. The trigger collection unit performs identity collection based on the target stage and performs secondary identity verification based on the obtained secondary collection information. The response determines that the verification is successful based on the secondary identity verification, plays the display video of the corresponding target stage based on the display unit, and controls the acquisition unit to perform posture acquisition, so as to determine the standard training score based on the obtained posture information.

[0007] Optionally, in the method according to the present invention, in response to the user terminal selecting any training mode based on the display unit, the acquisition unit is triggered to collect the user terminal's identity, and primary-level identity verification is performed based on the obtained primary collection information, including: When the pressure value output by the pressure sensor pre-set on the training table is greater than the preset value, the display state of the corresponding training display interface on the display unit is controlled to change from the locked state to the interactive state. When any training mode on the training display interface is continuously interacted with for a preset duration, it is determined that the user has selected the training mode. The trigger unit collects the user's identity information and identifies the obtained facial recognition information as a single collection. Retrieve historical registration information and compare the information collected at one time with the facial registration information included in the historical registration information; The response determines that if the collected information is the same as any face registration information based on the comparison, the verification is successful; otherwise, the verification is unsuccessful.

[0008] Optionally, in the method according to the present invention, triggering the acquisition unit to acquire the user's identity and determining the obtained facial recognition information as a single acquisition, includes: The acquisition unit is triggered to acquire images, and in response to the presence of a limb region indicating any human limb in the acquired image, a face frame is generated with the image center point of the corresponding acquired image as the center. If there is a pixel difference greater than a preset pixel threshold between any image pixel that overlaps with the face frame and the reference pixel of the corresponding human skin, the voice unit pre-set on the training table will be controlled to retrieve and play the posture adjustment audio. If the pixel difference between each image pixel that overlaps with the face frame and the reference pixel of the corresponding human skin is less than or equal to a preset pixel threshold, the image portion of the face frame corresponding to the captured image is determined as one acquisition information.

[0009] Optionally, in the method according to the present invention, triggering the acquisition unit to perform image acquisition, and responding to the presence of a limb region indicating any human limb in the acquired image, generating a face frame centered on the image center point of the corresponding acquired image, includes: The acquisition unit is triggered to acquire images, and an acquisition task with a continuously preset acquisition duration is established based on the acquisition unit. The response determines that the acquired image does not contain any limb area indicating any human body limb based on the acquisition task. If the locked state of the corresponding training display interface is changed from the display state to the interactive state, the response generates a face frame centered on the center point of the corresponding acquired image.

[0010] Optionally, in the method according to the invention, the response determines that the verification has passed, determines the current strength level data of the corresponding user terminal based on the information collected once, and selects a training course suitable for the strength level data from the retrieved course library based on the strength level data, including: The response determines that the primary identity verification has passed, retrieves the historical database, and iterates through each record in the historical database based on the information collected once. If any recorded data is matched with a previously collected data, the current strength level data of the corresponding user terminal is determined based on the recorded data; otherwise, the current strength level data is determined based on the retrieved strength test process. Training courses that match the strength level data are selected from a retrieved course library.

[0011] Optionally, in the method according to the invention, determining the current force level data based on the retrieved force testing procedure includes: Based on the basic information sub-process corresponding to the first test priority included in the strength test process, the physiological baseline information of the user is determined, and based on the age sub-information and physique sub-information included in the physiological baseline information, the age influence coefficient and physique influence coefficient are determined. The strength test process determines the user's strength test value based on the test information sub-process corresponding to the second test priority, and updates the strength test value based on the age influence coefficient and the physique influence coefficient to obtain the updated strength test value.

[0012] Optionally, in the method according to the invention, selecting training courses adapted to the strength level data from a retrieved course library based on the strength level data includes: Retrieve the course library, which includes each training course and the strength level range corresponding to each training course; If the strength level data falls within a strength level range, the training course corresponding to that strength level range will be determined as being adapted to the strength level data. In response to the strength level data being located in multiple strength level intervals, the appropriate age value corresponding to each strength level interval is determined, and the user age value of the corresponding user terminal is determined based on the age sub-information. Training courses that correspond to age-appropriate values ​​that are less than the user's age and have the smallest age difference are identified as those that are adapted to the strength level data.

[0013] Optionally, in the method according to the present invention, in response to determining that the verification is passed based on secondary identity verification, the corresponding target stage's display video is retrieved and played, and the acquisition unit is controlled to perform posture acquisition, so as to determine a standard training score based on the obtained posture information, including: The response is based on secondary identity verification to confirm that the verification is successful. The corresponding target stage display video is retrieved and played, and the acquisition unit is controlled to perform posture acquisition to obtain the posture information of the user terminal when playing the display video to different training segments, and to obtain the preset limb movements corresponding to each training segment. Based on posture information, the degree of motion variation of the corresponding preset limb movements is determined, and the average of all motion variation degrees in the corresponding display video is calculated to determine the standard training score.

[0014] Optionally, in the method according to the present invention, the degree of motion variation of the corresponding preset limb movement is determined based on posture information, and the average of all the degree of motion variation in the corresponding display video is calculated to determine a standard training score, including: Each action variation degree is compared with the retrieved preset variation threshold, and if any action variation degree is greater than the preset variation threshold, the action variation degree is determined to have a deviation attribute. Obtain the deviation ratio of all actions with deviation attributes, and if the deviation ratio is greater than a preset ratio threshold, determine the first score retrieved as the standard training score. If the response deviation ratio is less than or equal to the preset ratio threshold, the average of all motion changes in the corresponding video will be calculated, and the resulting second score will be determined as the standard training score.

[0015] Optionally, in the method according to the invention, the method further includes: Image recognition is performed on the posture image obtained based on posture information, and the posture image is determined based on the recognition result to include the belt body area indicating the assistive belt, and the belt body area has a contact relationship with the leg area indicating the user's leg limbs, and the user is judged to be in an assisted exercise mode. The response-assisted exercise mode is a two-leg assisted mode, and based on the posture image, it is determined that different leg limbs correspond to different assist belts with different force exertion situations. The control telescopic unit adjusts the contraction of the assist belt and obtains the extension length of the corresponding assist belt. The response-assisted training method is a single-leg assistance method, and the assistance belt is tilted based on the belt area. The extension length of the assistance belt is determined based on the tilt angle of the corresponding tilt state. Based on posture information, all posture images corresponding to each training cycle are combined into training videos, and the longest extension length obtained for each training video is determined as the effect indication length. The lengths of each effect indicator are sorted according to the time sequence of each corresponding training cycle, and the display unit is controlled to display the resulting effect sequence.

[0016] Optionally, in the method according to the present invention, the response-assisted exercise mode is a two-leg assisted mode, and based on the posture image, it is determined that different leg limbs correspond to different assistive belts with different force exertion states. The telescopic unit is controlled to adjust the contraction of the assistive belts, and the extension length of the corresponding assistive belts is obtained, including: The responsive training method is a two-legged assisted method, and an image coordinate system is established based on the posture image, wherein the Y-axis of the image coordinate system extends along the direction of gravity; The two belt areas that have contact with the two leg areas indicating different leg limbs and indicate different assistive belts are defined as the first contact belt and the second contact belt, respectively. The first belt coordinate group and the second belt coordinate group that make up the first contact belt and the second contact belt are determined based on the image coordinate system. The image coordinate point located in the first volume coordinate group and corresponding to the minimum vertical coordinate value is determined as the first minimum coordinate point, and the image coordinate point located in the second volume coordinate group and corresponding to the minimum vertical coordinate value is determined as the second minimum point; In response to the longitudinal coordinate difference between the first minimum coordinate point and the second minimum coordinate point, which have corresponding longitudinal coordinate values, it is determined that different leg limbs correspond to different assistive belts with different force exertion situations, and the telescopic unit is controlled to adjust the length of the assistive belt corresponding to the minimum coordinate point with the smaller longitudinal coordinate value. The belt coordinate system is updated accordingly based on the adjusted belt, and the extension length of the corresponding belt is obtained based on the longitudinal coordinate value corresponding to the minimum coordinate point in the updated belt coordinate system.

[0017] Optionally, in the method according to the present invention, the responsive exercise mode is a single-leg assisted mode, and the assisted belt is determined to be tilted based on the belt area, and the extension length of the corresponding assisted belt is determined based on the tilt angle of the corresponding tilt state, including: The response-assisted exercise method is a single-leg assisted method, and an image coordinate system is established based on the posture image, wherein the Y-axis of the image coordinate system extends along the direction of gravity; Based on the image coordinate system, the volume coordinate set that makes up the volume region is determined. The image coordinate point that is located in the volume coordinate set and corresponds to the minimum vertical coordinate value is determined as the minimum coordinate point, and the image coordinate point that is located in the volume coordinate set and corresponds to the maximum vertical coordinate value is determined as the maximum coordinate point. The system responds to the difference in lateral coordinates between the minimum and maximum coordinate points, which corresponds to the lateral coordinate values. The system tilts the belt and obtains the difference in longitudinal coordinates between the minimum and maximum coordinate points. The tilt angle of the corresponding booster belt is determined based on the difference in longitudinal and lateral coordinates, and the extension length of the corresponding booster belt is determined based on the tilt angle.

[0018] According to another aspect of the present invention, a training device adapted to the aforementioned pull-up training method is provided, comprising: Casing; and The display unit, processing unit, and acquisition unit are located in the housing.

[0019] According to another aspect of the present invention, a pull-up training system is provided, comprising: The master-level verification module is configured to respond to the user terminal selecting any training mode based on the display unit, trigger the collection unit to collect the user terminal's identity, and perform master-level identity verification based on the collected information. The course adaptation module is configured to respond to the verification confirmation, determine the current strength level data of the corresponding user based on the information collected once, and select training courses that are compatible with the strength level data from the retrieved course library based on the strength level data. The secondary verification module is configured to respond to the trigger signal received for any training course, identify the training stage in the training course that is in the first position according to the training order and is in an unplayed state as the target stage, trigger the collection unit to collect identity based on the target stage, and perform secondary identity verification based on the obtained secondary collection information. The training scoring module is configured to respond to secondary identity verification to confirm that the verification has passed, play the display video of the corresponding target stage based on the display unit, and control the acquisition unit to perform posture acquisition, so as to determine the standard training score based on the obtained posture information.

[0020] According to the present invention, the present invention can improve the efficiency of pull-up training and meet the scientific training needs of different groups, specifically based on the following: First, the two-level identity verification mechanism ensures the relevance and continuity of training. The primary verification locks the user's identity by collecting information and accurately retrieves historical data to determine the current strength level, avoiding the bias of blindly assessing one's own ability in unguided training. The secondary verification ensures that each stage of training is executed by the target user, preventing the training plan from becoming chaotic and laying a solid foundation for personalized training. Secondly, the strength level-adapted courses achieve personalized solutions for each individual. This invention can select suitable content from the course library based on the user's strength data. Beginners can obtain low-intensity introductory programs, while advanced learners can be matched with high-intensity breakthrough courses. This solves the problem of low efficiency caused by general tutorials and makes the training intensity and ability accurately matched. Finally, posture acquisition and scoring feedback optimize training movements in real time. This invention can provide standard demonstrations based on display videos, and the acquisition unit simultaneously captures the user's posture and generates scores, allowing trainees to intuitively understand movement defects and make timely corrections to avoid ineffective training and sports injuries. Compared with no-instruction mode, this invention can improve the scientific nature and efficiency of training, and provide a reliable guarantee for improving physical fitness and achieving assessment standards. Attached Figure Description

[0021] Figure 1 A front view schematic diagram of a pull-up training device according to an embodiment of the present invention is shown; Figure 2 A side view of the pull-up training device in this embodiment is shown. Figure 3 This is a top view schematic diagram of the pull-up training device in this embodiment; Figure 4 A flowchart of a pull-up training method according to another embodiment of the present invention is shown; Figure 5 A system block diagram of a pull-up training system according to yet another embodiment of the present invention is shown. Detailed Implementation

[0022] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0023] To address the problems existing in the prior art, the inventors proposed the solution of this invention. One embodiment of this invention provides a pull-up training device, wherein... Figure 1 A front view of the pull-up training device of this embodiment is shown, wherein, as Figure 1 As shown, this device mainly includes a display unit, a processing unit, and a data acquisition unit. These components work together to form a complete training support system.

[0024] The acquisition unit is the core component of the device, enabling precise perception. It comprehensively collects user identity information and training posture information. In practical applications, the acquisition unit can integrate components such as a high-definition wide-angle camera, an ID card reader, and a campus card reader. The high-definition wide-angle camera, fixed at an appropriate position in front of the training structure, can clearly capture the user's facial information for identity verification and capture body movements in real time during training, providing data support for subsequent posture analysis. The ID card reader and campus card reader serve as auxiliary identity collection methods, improving the convenience of identity verification. Furthermore, the comprehensive acquisition capability of the acquisition unit ensures a strong binding between user identity and training data, preventing proxy training from the source and guaranteeing the authenticity and exclusivity of the training data, allowing for accurate recording of the user's training results. The training structure can be understood as the equipment structure needed for pull-up training, such as a horizontal bar. As the core control module of the device, the processing unit plays a crucial role in data processing, logical operations, and instruction scheduling. The processing unit can be housed within a casing and can utilize an industrial-grade microcomputer or a high-performance embedded motherboard. It integrates core components such as a CPU and GPU. The GPU accelerates data processing efficiency, ensuring real-time analysis of the identity and posture information transmitted by the acquisition unit. The processing unit compares the identity information acquired by the acquisition unit with historical registration information to complete identity verification. Simultaneously, based on the acquired posture information and a preset action standard model, it analyzes the data to determine whether the actions are compliant. In short, the processing unit's efficient computing power enables the device to quickly respond to various data demands during training, ensuring a smooth training process and preventing data processing delays from affecting the user's training rhythm. As the primary medium for human-computer interaction, the display unit typically employs a large touchscreen of 32 inches or more and integrates speakers, providing users with a clear and intuitive interface and voice prompts. During training, the display unit shows key information such as training courses, follow-up videos, and training data. The follow-up videos guide users to standardize their movements and help them establish correct training patterns; the real-time displayed training data allows users to understand their training progress and the accuracy of their movements, enhancing the focus of their training. The user-friendly design of the display unit reduces the difficulty of operation, allowing users to concentrate on the training itself and improving the immersion in the training experience.

[0025] For example, in Figure 1 In the example, the pull-up training device may also include a voice unit, a hard disk recorder, and a switch. The voice unit is used to output voice content, the hard disk recorder is used to store the corresponding video content, and the switch is used to form a local area network.

[0026] Furthermore, it can be stated that, Figure 2A side view of the pull-up training device in this embodiment is shown. Figure 3 A top view of the pull-up training device in this embodiment is shown, as follows: Figure 2 as well as Figure 3 As shown, the pull-up training device may also include two embedded handles, a fixed power strip, and power and network cable ports. The embedded handles facilitate gripping and moving the pull-up training device, the fixed power strip provides different types of power supply ports, and the power and network cable ports are used to connect the power supply port and the network cable port to achieve power supply and network connection for the pull-up training device.

[0027] Figure 4 A flowchart of a pull-up training method adapted to the pull-up training device described above, according to an embodiment of the present invention, is shown. The method is suitable for execution in a computing device, wherein the computing device can be understood as a terminal with data processing capabilities, such as a mobile phone or a computer.

[0028] like Figure 4 As shown, the pull-up training method proposed in this embodiment includes steps S1 to S4, which can be specifically described as follows: S1. The user terminal selects any training mode based on the display unit, triggering the collection unit to collect the user terminal's identity and perform primary identity verification based on the collected information. S2. The response confirms that the verification has passed. Based on the information collected once, the current strength level data of the corresponding user terminal is determined, and a training course that matches the strength level data is selected from the retrieved course library. S3. Upon receiving the trigger signal corresponding to any training course, the training stage included in the training course and located first in the training order that is in an unplayed state is determined as the target stage. The trigger collection unit performs identity collection based on the target stage and performs secondary identity verification based on the obtained secondary collection information. S4. The response determines that the verification is successful based on the secondary identity verification. The display unit plays the display video of the corresponding target stage and controls the acquisition unit to perform posture acquisition, so as to determine the standard training score based on the obtained posture information.

[0029] For example, in this embodiment, multi-stage identity verification, accurate course matching, and real-time posture scoring can ensure the authenticity of training data, the suitability of courses, and the standardization of movements, thereby significantly improving the training experience. Specifically, firstly, in response to the user selecting any training mode based on the display unit, this embodiment can trigger the acquisition unit to collect the user's identity based on the server or the processing unit of the corresponding pull-up training device. Based on the acquired information, a primary-level identity verification is performed. In practical applications, the display unit can be a large touchscreen of 32 inches or more, allowing users to clearly view and select training modes, such as short-term or long-term training modes. The acquisition unit can be a high-definition wide-angle camera, which can be fixed at an appropriate position in front of the training structure to complete identity collection. Through face recognition technology throughout the process, the user's identity is strongly bound to the training data. Primary-level identity verification ensures from the beginning of training that the current training user is the actual person, effectively preventing proxy training and ensuring that all subsequent training data is associated with the user's real identity. This guarantees the exclusivity and authenticity of the training data, avoids data distortion caused by proxy training, and provides a genuine user base for subsequent accurate course recommendations. Next, once the primary identity verification is successful, the server can determine the current strength level data of the corresponding user based on the collected information. Then, based on the strength level data, it selects a training course from the retrieved course library that matches the strength level data. It can be noted that traditional training plans cannot be dynamically adjusted according to the user's strength level. However, this step accurately obtains the current strength level data by associating the user's historical training data or triggering a strength test through a single collection of information. After that, it selects a suitable course from the course library that stores courses of different difficulty levels. For example, users with zero experience are matched with introductory courses that include basic force exertion explanations, while advanced users are matched with high-intensity inter-set training courses. This can avoid the problem of users feeling frustrated due to courses that are too difficult or ineffective training due to courses that are too easy. This allows each user to improve their ability at an appropriate training intensity, significantly improving training efficiency and users' willingness to stick to training, and optimizing the training experience. Then, in response to the trigger signal received for any training course, the server can determine the training stage that is in the first position in the training order and is not yet broadcast as the target stage. The trigger collection unit collects identity information based on the target stage and performs secondary identity verification based on the obtained secondary collection information. That is, secondary identity verification is required before entering a specific training stage. It can be noted that secondary identity verification is an important part of the anti-proxy training mechanism. This verification is performed by facial recognition when entering a course and at the beginning of each project in formal training. It can prevent the situation where someone else replaces the user after selecting a course. For example, if the user leaves temporarily after selecting a course, others cannot enter the training through secondary identity verification. This ensures that the entire training process is completed by the user himself, avoids training data breakage and chaos caused by changing people in the middle of the training stage, ensures the continuity of the training process and the integrity of the data, and makes the training data more valuable for reference. Finally, upon successful verification of the secondary identity, the server can play the demonstration video corresponding to the target stage on the display unit and control the acquisition unit to capture posture information. A standard training score is then determined based on the acquired posture information. Specifically, this can be achieved through a dual-video window system, such as a follow-up video and a real-time capture video, to integrate the training process. The demonstration video in this step corresponds to the follow-up video, which may include a standard pull-up demonstration and audio explanation. Users watch the demonstration video on the display unit and follow along. Simultaneously, the acquisition unit captures the user's training posture (posture information) in real-time using a high-definition wide-angle camera and analyzes this information using a computer vision algorithm based on skeleton recognition. For example, it determines whether the user's chin is above the bar or whether the body is excessively swaying, thereby determining the standard training score. By playing the demonstration video in real-time, users can intuitively learn the standard movements, avoiding sports injuries caused by improper form. The standard training score allows users to know the accuracy of their movements in real time, adjust movement details promptly, improve the scientific nature and targeting of training, and ensure that users clearly understand the direction for improvement in each training session, gradually improving movement quality and training effectiveness, further optimizing the training experience.

[0030] Furthermore, in this embodiment, the aforementioned "responding to the user terminal selecting any training mode based on the display unit, triggering the collection unit to collect the user terminal's identity, and performing primary-level identity verification based on the obtained collection information" may further include the following steps: When the pressure value output by the pressure sensor pre-set on the training table is greater than the preset value, the display state of the corresponding training display interface on the display unit is controlled to change from the locked state to the interactive state. When any training mode on the training display interface is continuously interacted with for a preset duration, it is determined that the user has selected the training mode. The trigger unit collects the user's identity information and identifies the obtained facial recognition information as a single collection. Retrieve historical registration information and compare the information collected at one time with the facial registration information included in the historical registration information; The response determines that if the collected information is the same as any face registration information based on the comparison, the verification is successful; otherwise, the verification is unsuccessful.

[0031] For example, in this embodiment, primary-level identity verification can be performed based on the following method steps: Firstly, to address the issue of accidental touches on the training display interface, this embodiment can pre-install a pressure sensor on the training surface. The state transition of the training display interface is controlled by the change in the pressure sensor value. Specifically, the training display interface on the display unit will only switch from a locked state to an interactive state when the pressure value output by the pressure sensor is greater than a preset value. If the pressure value does not reach the preset value, the training display interface will remain locked. This can effectively prevent the interface from being accidentally touched and entering the interactive state when the device is not in use, preventing invalid operation records or incorrect mode selection. This allows users to directly enter the effective interactive process without having to clear the interface of accidental operation traces before preparing for training, saving training preparation time and improving operational convenience. Meanwhile, once the training display interface is in an interactive state, the user is only confirmed to have selected the training mode when any training mode on the training display interface is continuously interacted with for a preset duration. This prevents the user from accidentally selecting a non-target training mode due to finger touch or operational error, ensuring that the training mode selected by the user is consistent with the actual needs, reducing the tedious operation of changing modes later, and further improving the accuracy of operation and the smoothness of the training process. Secondly, in the identity collection stage, this embodiment can explicitly trigger the collection unit to collect the user's identity based on the server, and determine the obtained facial recognition information as the collection information in one go. In combination with the actual application scenario, the collection unit can use a high-definition wide-angle camera. The camera is fixed at an appropriate position in front of the training structure, which can clearly capture the user's facial features, ensuring that the obtained facial recognition information has high clarity and completeness. Using facial recognition information as the collection information in one go is more convenient than traditional card recognition methods. Users do not need to carry physical cards. They only need to complete the facial collection in front of the camera, reducing the user's carrying burden and operation steps. At the same time, the high-definition facial recognition information also lays a data foundation for subsequent accurate comparison, avoiding verification failure due to blurry collected information, reducing the number of times users collect information repeatedly, and improving the efficiency of identity collection. Finally, to ensure the accuracy of identity verification, this embodiment can retrieve historical registration information after the server obtains the collected information once. Then, it compares the collected information with each face registration information contained in the historical registration information one by one. If the comparison determines that the collected information is the same as any face registration information, the primary identity verification is determined to be successful; if the collected information does not match any face registration information, the verification is determined to be unsuccessful. This can minimize omissions or misjudgments and ensure that only the registered user can pass the primary identity verification. It effectively prevents unregistered users from occupying device resources or others from training on their behalf, and strongly binds the training data with the user's identity, ensuring the authenticity and exclusivity of the training data. At the same time, the clear comparison result determination method allows users to quickly know the verification status. If the verification fails, they can register or re-collect information in a timely manner, avoiding user waiting due to the lack of transparency in the verification process, and further improving the user experience of the identity verification process.

[0032] Furthermore, in this embodiment, the aforementioned "triggering the collection unit to collect the user's identity and determining the obtained facial recognition information as a collection instance" may further include the following steps: The acquisition unit is triggered to acquire images, and in response to the presence of a limb region indicating any human limb in the acquired image, a face frame is generated with the image center point of the corresponding acquired image as the center. If there is a pixel difference greater than a preset pixel threshold between any image pixel that overlaps with the face frame and the reference pixel of the corresponding human skin, the voice unit pre-set on the training table will be controlled to retrieve and play the posture adjustment audio. If the pixel difference between each image pixel that overlaps with the face frame and the reference pixel of the corresponding human skin is less than or equal to a preset pixel threshold, the image portion of the face frame corresponding to the captured image is determined as one acquisition information.

[0033] For example, in this embodiment, the acquisition of information in a single collection can be achieved based on the following method steps: First, in this embodiment, the acquisition unit is triggered by the server to acquire images. When the acquired image contains a limb region indicating any human limb, a face frame is generated centered on the image center point of the corresponding acquired image. In practical applications, the acquisition unit can use a high-definition wide-angle camera, which is fixed at an appropriate position in front of the training structure to fully capture images within the training area, ensuring timely recognition of human limbs. This design of first confirming the limb region and then regenerating the face frame effectively avoids invalid image acquisition by the acquisition unit when no one is present, reducing unnecessary consumption of device computing resources. It also makes the generation of the face frame more targeted. That is, the face frame generated centered on the image center point can cover the human face area with the highest probability, avoiding missed capture of face information due to frame offset, making subsequent face information extraction more efficient, reducing the need for users to readjust their positions due to improper frame placement, and saving acquisition time. Secondly, when any image pixel that overlaps with the face frame has a pixel difference greater than a preset pixel threshold with the reference pixel of the corresponding human skin, the server will control the voice unit pre-set on the training table to retrieve and play posture adjustment audio. The voice unit on the training table can integrate speaker function to clearly play posture adjustment prompts, such as: Please adjust the facial angle to ensure uniform lighting, please move closer to the camera to ensure facial clarity, etc. This allows users to know the problems with their current posture or position in a timely manner and make quick adjustments according to the prompts, avoiding repeated attempts by users due to not knowing the reason for the failure of the acquisition, reducing the complexity of the acquisition operation, improving the user experience, and at the same time, it can also pre-filter out images with substandard pixel quality, laying the foundation for obtaining high-quality acquisition information in the future. Finally, when the server has a pixel difference of less than or equal to a preset pixel threshold between each image pixel that overlaps with the face frame and the corresponding reference pixel of human skin, the image portion of the captured image corresponding to the face frame is determined as a single acquisition. This ensures that the final single acquisition information has high clarity and authenticity. The pixel difference is controlled within the preset threshold, which means that the skin color reproduction and detail recognition of the face image can meet the requirements of subsequent identity verification. This significantly reduces the probability of verification failure due to image blurring or distortion, making the primary identity verification process smoother, reducing the number of times users need to re-capture due to verification failure, shortening the preparation time before training, allowing users to enter the training phase faster, and further improving the overall training experience.

[0034] Furthermore, in this embodiment, the aforementioned "triggering the acquisition unit to perform image acquisition, and responding to the existence of a limb region indicating any human limb in the acquired image, generating a face frame centered on the image center point of the corresponding acquired image" may further include the following steps: The acquisition unit is triggered to acquire images, and an acquisition task with a continuously preset acquisition duration is established based on the acquisition unit. The response determines that the acquired image does not contain any limb area indicating any human body limb based on the acquisition task. If the locked state of the corresponding training display interface is changed from the display state to the interactive state, the response generates a face frame centered on the center point of the corresponding acquired image.

[0035] For example, in this embodiment, the generation of the face frame can be specifically implemented based on the following method steps: Firstly, in practical applications, the acquisition unit can use a high-definition wide-angle camera. This camera is fixed at an appropriate position in front of the training structure and can stably capture images within the training area within a preset acquisition time. Setting a preset acquisition time can avoid the continuous occupation of device computing resources due to the acquisition task running indefinitely, making the acquisition process more timely. For example, setting the preset acquisition time to 3 seconds can ensure that the camera captures clear images without making the user wait due to excessive acquisition time, reducing the user's time spent in the acquisition process, while also reducing unnecessary power consumption of the device and improving the device's operating efficiency. Next, after acquiring images based on the established acquisition task, the server will determine whether there is a limb region in the acquired image that indicates any human limb. If the acquisition task determines that the acquired image does not contain the limb region, it means that no user is currently in the training area preparing for training. At this time, the lock state of the corresponding training display interface is changed from the display state to the interactive state, which avoids the need for the user to manually unlock the interface after approaching the device. This allows subsequent users to directly select the training mode on the interactive interface after entering the training area, reducing operation steps and optimizing the user operation process. If the captured image confirms the presence of the limb region, it indicates the user is ready in the training area. At this point, a facial frame is generated centered on the image center point of the captured image. This ensures the frame covers the user's face to the greatest extent possible, preventing incomplete facial information capture due to frame misalignment. For example, when the user stands in the middle of the training area, the image center point and the user's face are essentially aligned, allowing the generated frame to accurately select the face. This reduces the need for the user to readjust their position due to improper frame placement, making subsequent identity capture steps smoother and further improving the user's pre-training preparation efficiency.

[0036] Furthermore, in this embodiment, the aforementioned "response confirms verification passed, determines the current strength level data of the corresponding user based on the collected information, and selects a training course that matches the strength level data from the retrieved course library based on the strength level data" may also include the following steps: The response determines that the primary identity verification has passed, retrieves the historical database, and iterates through each record in the historical database based on the information collected once. If any recorded data is matched with a previously collected data, the current strength level data of the corresponding user terminal is determined based on the recorded data; otherwise, the current strength level data is determined based on the retrieved strength test process. Training courses that match the strength level data are selected from a retrieved course library.

[0037] For example, in this embodiment, the selection of training courses can be implemented based on the following method steps: First, upon successful verification of the primary identity, the server can access the historical database and iterate through each record in the database based on the collected information. This historical database stores strength level data recorded during the user's past training sessions. The collected information serves as a core identifier for the user, accurately linking them to the individual. By iterating through each record, the system can comprehensively check for the existence of the user's historical training records, preventing omissions of historical strength level data due to incomplete data retrieval. This iterative retrieval method allows users with historical records to determine their current strength level directly using past training data without repeating strength tests, significantly saving preparation time and allowing users to quickly enter the core training phase, reducing unnecessary waiting time and improving the smoothness of the training process. Next, after traversing the recorded data, the server can determine the current strength level data based on the matching results. If any recorded data matches a previously collected set of information, it indicates that the recorded data belongs to the current user. Based on this recorded data, the server determines the corresponding user's current strength level data, ensuring that the strength level data is consistent with the user's past training ability and avoiding course recommendation bias due to data mismatch. If all recorded data does not match a previously collected set of information, it indicates that the user is using the device for the first time or has no historical training records. In this case, the server determines the current strength level data based on the retrieved strength test process. The strength test process assesses the user's basic strength through standardized test movements, ensuring that even new users can obtain strength level data that matches their actual abilities. This avoids blindly recommending courses due to a lack of data, preventing beginners from developing resistance due to overly difficult courses or experienced users from wasting training time due to overly easy courses. Test movements include, for example, the maximum number of pull-ups attempted within a specified time, and the quality of completing specific movements. Finally, the server can select training courses that match the determined strength level data from the retrieved course library. This course library stores pull-up training courses of different difficulty levels and training goals, such as basic strength activation courses for beginners and strength enhancement inter-set training courses for advanced users. In this embodiment, by accurately matching strength level data with course difficulty, each user can obtain training content that suits their ability. For example, users with weak foundations can start with simple movements and low-intensity sets to gradually build training confidence and movement foundation; users with stronger abilities can break through training bottlenecks through high-intensity, high-difficulty courses, improving the targeting and effectiveness of training. This allows users to experience appropriate challenges and progress in each training session, enhancing their sense of accomplishment and further optimizing the training experience.

[0038] Furthermore, in this embodiment, the aforementioned "determining the current strength level data based on the retrieved strength test process" may further include the following steps: Based on the basic information sub-process corresponding to the first test priority included in the strength test process, the physiological baseline information of the user is determined, and based on the age sub-information and physique sub-information included in the physiological baseline information, the age influence coefficient and physique influence coefficient are determined. The strength test process determines the user's strength test value based on the test information sub-process corresponding to the second test priority, and updates the strength test value based on the age influence coefficient and the physique influence coefficient to obtain the updated strength test value.

[0039] For example, in this embodiment, the acquisition of strength test values ​​can be achieved based on the following method steps: First, the server can determine the user's physiological baseline information based on the basic information sub-process corresponding to the first test priority in the strength test process. The setting of the first test priority ensures that key individual difference information affecting strength performance is collected before the actual strength test, avoiding subsequent data deviations due to the omission of core physiological factors. Furthermore, the physiological baseline information specifically includes age and body structure information. Age reflects the user's physiological functional status, such as muscle strength and recovery speed; for example, users aged 25-35 typically have high muscle synthesis efficiency and good strength reserves, while users over 50 need to consider the impact of muscle loss on strength. Body structure information reflects the user's body shape characteristics, such as height, weight, and limb proportions; for example, a user with a large body weight but insufficient muscle mass... Users with sufficient strength will exert significantly more force when performing pull-ups than users of moderate weight and adequate muscle mass. After obtaining these two types of information, the system further determines the age and physique influence coefficients based on a preset scientific algorithm. For example, for the age sub-information, a higher age influence coefficient can be set for users aged 20-30, that is, quantifying the positive support of their better physiological functions for strength, while a lower age influence coefficient can be set for users over 55, that is, quantifying the objective impact of changes in physiological functions on strength. For the physique sub-information, a coefficient can be set according to the height-to-weight ratio to avoid misjudgment of strength performance due to differences in physique. This quantifies the influence of age and physique on strength, providing a scientific basis for subsequent correction of strength test values, and making the calculation of strength level data more personalized. Finally, the server can then determine the user's strength test value based on the test information sub-process corresponding to the second test priority in the strength test process. It should be noted that the design of the second test priority ensures that standardized strength testing is conducted only after the individual difference influence coefficient is clearly defined, making the acquisition of raw test values ​​more targeted and valuable for reference. Specifically, the test information sub-process can obtain the user's raw strength test value through preset standardized movements. This value directly reflects the user's current strength performance but has not yet incorporated the influence of individual differences. At this point, the raw strength test value can be updated using the age influence coefficient and physique influence coefficient obtained earlier to obtain the updated strength test value. For example, if a user's raw strength test value is 6 repetitions... For pull-ups, since the target user is 38 years old, their physiological function is slightly lower than that of younger users. Therefore, the age influence coefficient can be 0.9. Similarly, since their weight is slightly higher than the standard value for their height, the exertion burden is slightly greater. Therefore, the physique influence coefficient can be 0.85. This allows the updated strength test values ​​to more accurately reflect their true strength level among peers of the same age and physique. This ensures that the strength level data is no longer a single raw test result, but a comprehensive assessment that incorporates individual differences. This ensures that the data truly matches the user's actual ability and avoids training discomfort caused by biased course recommendations. At the same time, it can also match personalized needs, allowing each user to start training based on their own true ability, improving the safety and relevance of training.

[0040] Furthermore, in this embodiment, the aforementioned "selecting training courses that match the strength level data from the retrieved course library based on the strength level data" may further include the following steps: Retrieve the course library, which includes each training course and the strength level range corresponding to each training course; If the strength level data falls within a strength level range, the training course corresponding to that strength level range will be determined as being adapted to the strength level data. In response to the strength level data being located in multiple strength level intervals, the appropriate age value corresponding to each strength level interval is determined, and the user age value of the corresponding user terminal is determined based on the age sub-information. Training courses that correspond to age-appropriate values ​​that are less than the user's age and have the smallest age difference are identified as those that are adapted to the strength level data.

[0041] For example, in this embodiment, selecting training courses based on strength level data can be achieved through the following specific method steps: First, the server can access the course library, which contains various training courses and the corresponding strength level range for each course. This avoids the confusion caused by the lack of clear strength range division in traditional course libraries. Specifically, the strength level ranges in the course library can be set according to the common ability levels of pull-up training. For example, 0-2 reps / set corresponds to beginner courses, 3-5 reps / set corresponds to basic improvement courses, and 6 reps or more / set corresponds to advanced strengthening courses. This provides clear course positioning for users with different strength levels and reduces the time cost of blindly choosing from many courses. Next, the server can design differentiated filtering logic based on different scenarios where the strength level data falls within a certain range. That is, when the strength level data falls within a specific range, the server directly identifies the training course corresponding to that range as suitable for the data, allowing users to quickly find a course precisely matching their current strength level. For example, if the strength level data is 4 reps / set, which falls within the 3-5 reps / set range, the server can directly select the corresponding basic improvement course without additional complex judgments, saving time on course selection and allowing users to enter the training phase more quickly, thus improving the smoothness of the training process. When the strength level data falls within multiple ranges... When dealing with multiple strength level ranges, such as when strength level data falls at the borderline between two ranges, or when testing errors cover multiple ranges, an optimization logic can be introduced that links age-appropriate values ​​with user age values. That is, first, the age-appropriate value corresponding to each strength level range is determined. This age-appropriate value is a range of age ranges suitable for courses based on the physiological functions and recovery abilities of users of different ages. For example, if the age-appropriate value for a certain strength level range is 18-25 years old, it means that the intensity and pace of courses in this range are more suitable for users of that age. Then, based on the age sub-information in the previously obtained physiological baseline information, the user age value for the corresponding user is determined. Finally, the server can determine the training courses that correspond to the age-appropriate values ​​of users younger than their age and have the smallest age difference as suitable for their strength level data. For example, if a user is 28 years old and their strength level data covers two different ranges: 3-5 reps / set for users aged 18-25 and 6+ reps / set for users aged 23-30, the age-appropriate value for users aged 23-30 is smaller than that for users aged 28, and the corresponding course is selected. This fully considers the impact of age on training suitability and avoids the problem of excessive or insufficient intensity caused by filtering solely based on strength level. In other words, middle-aged users will not be matched with high-intensity courses suitable for young people, and young users will not be matched with overly conservative low-intensity courses. This ensures that the courses, in addition to strength-appropriateness, further conform to the user's physiological characteristics, improving the safety and experience of training.

[0042] Furthermore, in this embodiment, the aforementioned "response based on secondary identity verification confirms successful verification, retrieves and plays the display video of the corresponding target stage, and controls the acquisition unit to perform posture acquisition, so as to determine the standard training score based on the obtained posture information" may also include the following steps: The response is based on secondary identity verification to confirm that the verification is successful. The corresponding target stage display video is retrieved and played, and the acquisition unit is controlled to perform posture acquisition to obtain the posture information of the user terminal when playing the display video to different training segments, and to obtain the preset limb movements corresponding to each training segment. Based on posture information, the degree of motion variation of the corresponding preset limb movements is determined, and the average of all motion variation degrees in the corresponding display video is calculated to determine the standard training score.

[0043] For example, in this embodiment, the determination of the standard training score can be specifically based on the following method steps: First, upon successful verification based on secondary identity verification, the server can retrieve and play the demonstration video for the corresponding target stage and control the acquisition unit to capture posture information. This allows the user to obtain posture information for different training segments corresponding to the demonstration video and acquire the preset limb movements for each training segment. Specifically, in practical applications, the demonstration video can include a standard pull-up demonstration and audio explanation, such as shoulder activation movements during warm-up and the complete pull-up exertion movement during formal training. When users watch the demonstration video through the display unit, they can intuitively learn the standard movement details of each training segment, avoiding improper force exertion due to fuzzy memory of the movements. This provides users with a standard reference at all times during training, reducing the formation of incorrect movements. Simultaneously, the acquisition unit can be fixed at an appropriate position in front of the training structure, capturing the user's posture in real time during different training segments of the demonstration video and generating corresponding posture information. For example, in the training segment corresponding to the chin-over-the-bar exercise, the camera will capture the user's neck and shoulder joint positions, generating posture information. This allows for precise matching of the movement requirements of each training segment, avoiding the omission of key movement details caused by traditional general data collection throughout the entire process, making subsequent evaluation more targeted. In addition, the server can further obtain the preset limb movements corresponding to each training segment. It can be noted that the preset limb movements are based on standard movement parameters set by sports science, such as the elbow flexion angle during pull-ups and the vertical angle between the torso and the ground, providing a clear benchmark for subsequent posture comparisons and ensuring that the evaluation is systematic. Next, the server can determine the degree of motion variation for the corresponding preset limb movements based on the acquired posture information, and calculate the average of all motion variation degrees in the corresponding displayed video to determine the standard training score. It can be explained that motion variation degree refers to the degree of difference between the user's actual posture information and the preset limb movements in the corresponding training segment. For example, if the preset limb movement requires the elbow to bend to 90 degrees during a pull-up, and the user's actual posture only bends the elbow to 120 degrees, a corresponding motion variation degree value will be generated. The greater the difference, the higher the motion variation degree value. Here, by calculating the motion variation degree of each training segment, users can clearly understand the deviations in different stages of the movement, such as whether the shoulder stretching in the warm-up stage is in place, and whether the pull-up force exertion in the formal training is standard. By averaging all motion variation degrees... Mean calculation can comprehensively evaluate the overall standard of a user's movements throughout the entire target stage, avoiding the bias caused by scoring only a single segment. For example, in a user's three training segments at a target stage, the first two segments have low movement variability, meaning the corresponding movements are standard, while the third segment has high movement variability, meaning the movements have significant deviations. Based on this, it can be seen that mean calculation can balance the performance of each segment and provide a more comprehensive standard training score. This reflects both the detailed deviations of individual movements and the overall quality of the training movements, allowing users to clearly understand their own training strengths and weaknesses. Subsequently, they can focus on improving the segments with high movement variability, improving the targeting and effectiveness of training. At the same time, it makes the evaluation of training effects more objective and scientific, enhances the user's sense of recognition of training results, and further optimizes the training experience.

[0044] Furthermore, in this embodiment, the aforementioned "determining the degree of motion variation of the corresponding preset limb movement based on posture information, and calculating the average of all motion variations in the corresponding displayed video to determine the standard training score" may also include the following steps: Each action variation degree is compared with the retrieved preset variation threshold, and if any action variation degree is greater than the preset variation threshold, the action variation degree is determined to have a deviation attribute. Obtain the deviation ratio of all actions with deviation attributes, and if the deviation ratio is greater than a preset ratio threshold, determine the first score retrieved as the standard training score. If the response deviation ratio is less than or equal to the preset ratio threshold, the average of all motion changes in the corresponding video will be calculated, and the resulting second score will be determined as the standard training score.

[0045] For example, in this embodiment, determining the standard training score based on the degree of action variation can be specifically implemented based on the following method steps: First, the server compares the degree of variation of each movement with a preset variation threshold. If any movement's degree of variation exceeds the preset threshold, it is identified as having a deviation attribute. This preset variation threshold is an acceptable range of difference based on sports science and extensive standard movement data. For example, by analyzing joint angle changes in a standard pull-up using skeletal recognition technology, a reasonable fluctuation range for key indicators such as elbow flexion angle and trunk swing amplitude is determined. The value corresponding to this range is the preset variation threshold. When the degree of variation of a user's movement in a training segment exceeds this threshold, it indicates a significant deviation from the standard movement and a lack of standardization. This precise deviation attribute determination can systematically filter out non-standard movements in the user's training, avoiding the problem of traditional mean calculations masking individual serious deviations. This allows users to clearly identify which training segments require focused correction, improving the targeted nature of their improvements. Next, the server can obtain the deviation ratio of all actions with deviation attributes. The deviation ratio refers to the ratio of the number of actions with deviation attributes to the total number of actions with deviation in all training segments corresponding to the displayed video. For example, if the displayed video contains 10 training segments, and 3 of the segments have actions with deviation attributes, then the deviation ratio is 30%. This means that by calculating the deviation ratio, the overall proportion of non-standard actions in a user's training can be intuitively reflected, avoiding the bias caused by focusing only on a single deviation action or the overall average. For example, a user may have one action with an extremely high degree of variation, but the other 9 are within the standard, with a deviation ratio of only 10%, indicating that the overall action quality is good. On the other hand, another user may have 5 actions with variations slightly above the threshold, and although the average is close to the former, the deviation ratio reaches 50%, indicating that the overall action standardization is poor. This allows for a more comprehensive scoring basis and provides key references for subsequent scenario-based scoring. Finally, the server determines the standard training score based on the comparison between the deviation ratio and the preset ratio threshold. If the deviation ratio is greater than the preset ratio threshold, it indicates that the proportion of non-standard movements in the user's training is too high, and the overall movement quality does not meet the requirements. In this case, the first score retrieved is determined as the standard training score. Here, the setting of the first score focuses more on the warning function. For example, a low value can directly remind the user that there are currently a large number of non-standard movements, and the basic movement standardization needs to be improved first to avoid the user ignoring the core defects due to the mean masking problem. If the deviation ratio is less than or equal to the preset ratio threshold, it indicates that the proportion of non-standard movements in the user's training is controllable, and the overall movement quality is good. Once the basic quality of the movements is met, the average of all movement variations in the corresponding video is calculated, and the resulting second score is used as the standard training score. This second score comprehensively reflects the overall standard of the user's movements, taking into account the impact of individual deviations without excessively lowering the score due to minor deviations. It aligns with the user's actual training performance, ensuring that the standard training score neither overlooks numerous deviations nor exaggerates the impact of minor deviations. This results in a more objective and accurate score. A high deviation rate allows for focused correction of basic movements, while a low deviation rate allows for optimization of details to address specific issues, enhancing the training's relevance and effectiveness, and further optimizing the training experience.

[0046] Furthermore, considering that many pull-up training scenarios require the assistance of lifting straps—for example, beginners with weak strength need lifting straps to lower the training threshold, heavier trainees need lifting straps to balance the exertion load, or users with unilateral strength deficiencies need targeted reinforcement—this embodiment can further include the following steps to improve the training system, enabling assisted training to achieve precise adaptation, dynamic adjustment, and visible results, ensuring the scientific and targeted nature of training for users with different needs: Image recognition is performed on the posture image obtained based on posture information, and the posture image is determined based on the recognition result to include the belt body area indicating the assistive belt, and the belt body area has a contact relationship with the leg area indicating the user's leg limbs, and the user is judged to be in an assisted exercise mode. The response-assisted exercise mode is a two-leg assisted mode, and based on the posture image, it is determined that different leg limbs correspond to different assist belts with different force exertion situations. The control telescopic unit adjusts the contraction of the assist belt and obtains the extension length of the corresponding assist belt. The response-assisted training method is a single-leg assistance method, and the assistance belt is tilted based on the belt area. The extension length of the assistance belt is determined based on the tilt angle of the corresponding tilt state. Based on posture information, all posture images corresponding to each training cycle are combined into training videos, and the longest extension length obtained for each training video is determined as the effect indication length. The lengths of each effect indicator are sorted according to the time sequence of each corresponding training cycle, and the display unit is controlled to display the resulting effect sequence.

[0047] For example, in this embodiment, based on the application scenario of using assistive straps, this embodiment can specifically perform dynamic adjustment and effect visualization based on the following method steps: First, after the user enters the training phase through secondary identity verification, the server controls the acquisition unit to continuously capture the user's posture information and generate posture images. At the same time, it performs real-time image recognition on the posture images. It can be noted that this recognition process mainly focuses on two core judgment conditions: one is whether there is a belt body area indicating the assistive strap in the posture image, and the other is whether the belt body area is in contact with the leg area indicating the user's leg limbs. When the server confirms that both conditions are met based on the recognition results, it means that the user has been using the assistive strap pre-installed in the training structure for assisted exercise. At this time, the server can directly determine that the user is currently in assisted exercise mode, and thus use this as a trigger to start subsequent targeted processes, avoiding invalid calculations when the user is not using the assistive strap, and ensuring the operating efficiency of the equipment. Secondly, in one scenario, if the server determines that the user is using a two-legged assist method, it will initiate the assist balance adjustment and data acquisition process. That is, the server can analyze the force feedback of the two assist belts through posture images. If it identifies an imbalance in the force exertion of the assist belts corresponding to different leg limbs, for example, if the stretching amplitude of one assist belt is significantly greater than that of the other side, it will control the telescopic unit to contract and adjust the assist belt on the weaker side to ensure that the assist strength on both sides tends to be balanced, and avoid the user's body tilting or uneven muscle development due to force imbalance. Furthermore, after the adjustment is completed, the extension length of the two assist belts can be obtained in real time and recorded as the core state data of this training phase. In another scenario, if the server determines that the user is using a single-leg assist method, it will execute the assist adaptation adjustment and data collection process. That is, the server can identify the shape characteristics of the belt area through posture images. If it is determined that the assist belt is tilted, it will calculate the extension length of the assist belt that is precisely matched with the current force state based on the preset mapping relationship between tilt angle and assist force. Next, after each training cycle ends, the server will combine the posture images into a complete training video based on all the posture information collected during that cycle, in chronological order. This allows users to review the training process and view the details of the movements. At the same time, the longest extension length is selected from all the extension length data of the assisting bands corresponding to the training video and determined as the effect indicator length for that training cycle. It can be said that this length directly reflects the user's maximum exertion and training intensity during that cycle, and is a core indicator for quantifying training effects. In addition, a training cycle can be understood as a complete set of pull-up training or a complete pull-up training session. Finally, the server sorts the effect indication lengths of all training cycles according to their chronological order, forming a clear effect sequence. It then controls the display unit to present this sequence in a visual format, allowing users to intuitively view the changes in their training intensity across different cycles. For example, if the effect indication length shows a decreasing trend, it indicates that the user's upper body strength is gradually improving and their reliance on the support straps is decreasing. This makes the training results immediately apparent, providing data support for subsequent adjustments to the training plan. It also intuitively reflects the significant improvement in the user's upper body strength, enhancing their sense of accomplishment and motivation to persevere.

[0048] It should be noted that, from the functional perspective of assistive straps, their purpose is to provide auxiliary support for users with insufficient strength, lowering the training threshold for pull-ups. Better training results mean stronger upper limb and core strength, and a lower dependence on the assistive straps. This indicates that the correct form can be completed without significant stretching of the straps, thus the extension length will naturally be shorter. Conversely, beginners or those with weaker strength require significant stretching of the assistive straps to provide sufficient support, resulting in a longer extension length. In this embodiment, the longest extension length is used as the effect indicator length. As the user's training effect improves, the longest extension length in each training cycle will show a decreasing trend, intuitively reflecting the user's reduced need for the assistive straps and significant training effectiveness.

[0049] Furthermore, in this embodiment, the aforementioned "responding to the exercise mode as a two-legged assisted mode, and determining that different leg limbs correspond to different assistive belts with different force exertion states based on posture images, controlling the telescopic unit to contract and adjust the assistive belts, and obtaining the extension length of the corresponding assistive belts" may also include the following steps: The responsive training method is a two-legged assisted method, and an image coordinate system is established based on the posture image, wherein the Y-axis of the image coordinate system extends along the direction of gravity; The two belt areas that have contact with the two leg areas indicating different leg limbs and indicate different assistive belts are defined as the first contact belt and the second contact belt, respectively. The first belt coordinate group and the second belt coordinate group that make up the first contact belt and the second contact belt are determined based on the image coordinate system. The image coordinate point located in the first volume coordinate group and corresponding to the minimum vertical coordinate value is determined as the first minimum coordinate point, and the image coordinate point located in the second volume coordinate group and corresponding to the minimum vertical coordinate value is determined as the second minimum point; In response to the longitudinal coordinate difference between the first minimum coordinate point and the second minimum coordinate point, which have corresponding longitudinal coordinate values, it is determined that different leg limbs correspond to different assistive belts with different force exertion situations, and the telescopic unit is controlled to adjust the length of the assistive belt corresponding to the minimum coordinate point with the smaller longitudinal coordinate value. The belt coordinate system is updated accordingly based on the adjusted belt, and the extension length of the corresponding belt is obtained based on the longitudinal coordinate value corresponding to the minimum coordinate point in the updated belt coordinate system.

[0050] For example, in this embodiment, for the dual-leg assistance method, the extension length can be obtained based on the following method steps: First, once the server determines that the user is in the bi-leg assisted mode, it can establish a unified image coordinate system based on the current posture image acquired by the acquisition unit. The Y-axis of this image coordinate system extends along the direction of gravity, and the magnitude of its vertical coordinate value is directly related to the stretching degree of the assist belt. That is, the smaller the vertical coordinate value, the longer the assist belt is stretched, and the weaker the leg force is. This provides a benchmark for the subsequent quantitative judgment of the force exertion, ensuring the uniformity and accuracy of the coordinate analysis. Next, the server can accurately identify the belt body region and leg region in the posture image. That is, the two assistive belt body regions that form contact with the leg regions of the two limbs are respectively identified as the first contact belt body and the second contact belt body. Furthermore, based on the established image coordinate system, the coordinate information of all image pixels contained in each of the two contact belt bodies is extracted to form the first belt body coordinate group and the second belt body coordinate group respectively. Here, the spatial position characteristics of the belt body are completely recorded through the coordinate group, providing data support for the subsequent selection of minimal coordinate points. Then, the server can select the image coordinate point with the minimum vertical coordinate value from the first belt coordinate group and define it as the first minimum coordinate point. At the same time, it can select the image coordinate point with the minimum vertical coordinate value from the second belt coordinate group and define it as the second minimum coordinate point. It can be explained that since the Y-axis of the image coordinate system is along the direction of gravity, the minimum coordinate point corresponds to the lowest point of the stretching of the assist belt. Its vertical coordinate value directly reflects the maximum stretching degree of the assist belt on that side, thereby quantifying the force intensity of the corresponding leg. Furthermore, the server can compare the vertical coordinate values ​​of the two minimum coordinate points. If there is a vertical coordinate difference between the two that is greater than a preset threshold, it is directly determined that there is an imbalance in the force exertion of different leg limbs corresponding to different assist belts. That is, on the side with a smaller vertical coordinate value, the assist belt is stretched longer and the leg force exertion is relatively weaker. Subsequently, the server sends control commands to the telescopic unit to adjust the length of the assistive band corresponding to the minimum coordinate point with a smaller longitudinal coordinate value. It can be noted that the length of the adjustment is determined based on the magnitude of the longitudinal coordinate difference. Under normal circumstances, it should be the same as the actual length corresponding to the longitudinal coordinate difference to ensure that the stretching degree of the assistive bands on both sides tends to be consistent after adjustment, thereby making the assistive force on both sides balanced. This avoids problems such as the user's body tilting, muscle development imbalance, or reduced training effect due to insufficient force on one side, and ensures the standardization and effectiveness of double-leg assisted training. Finally, after the assist belt is contracted and adjusted, the server updates the first or second belt coordinate set synchronously based on the actual state of the adjusted belt body to ensure that the coordinate data is consistent with the actual position of the assist belt. Furthermore, it extracts the longitudinal coordinate value corresponding to the minimum coordinate point from the updated belt coordinate set, and accurately calculates and obtains the extension length of the corresponding assist belt based on the longitudinal coordinate value. This extension length serves as the core state data representing the relationship between the assist belt state and the user's force exertion in bipedal assist training.

[0051] For example, a heavier user uses bi-leg assisted training. After the server establishes an image coordinate system along the gravity direction along the Y-axis based on the posture image, it identifies the first contact band that contacts the left leg and the second contact band that contacts the right leg. It extracts the corresponding coordinate sets of the first and second bands and further filters out the first minimum coordinate point with a vertical coordinate value of 30cm and the second minimum coordinate point with a vertical coordinate value of 22cm. It can be seen that the difference between the two vertical coordinates is 8cm, indicating that the left leg exerts less force than the right leg. At this time, the server can control the telescopic unit to retract the left assist band by 8cm. After adjustment, the coordinate set of the first band is updated, and the vertical coordinate value of its minimum coordinate point becomes 22cm, reducing the difference with the vertical coordinate value of the second minimum coordinate point to 0cm. Then, based on the updated coordinate values, the extension length of both assist bands is obtained to ensure that the calculation of the subsequent effect indication length accurately reflects the training status. Here, the telescopic unit can specifically be a motor device that is pre-installed on the training structure to control the extension and retraction of the assist band. Furthermore, in this embodiment, the aforementioned "response-assisted training method is a single-leg assisted method, and the assisted belt is determined to be tilted based on the belt area, and the extension length of the corresponding assisted belt is determined based on the tilt angle of the corresponding tilt state" may also include the following steps: The response-assisted exercise method is a single-leg assisted method, and an image coordinate system is established based on the posture image, wherein the Y-axis of the image coordinate system extends along the direction of gravity; Based on the image coordinate system, the volume coordinate set that makes up the volume region is determined. The image coordinate point that is located in the volume coordinate set and corresponds to the minimum vertical coordinate value is determined as the minimum coordinate point, and the image coordinate point that is located in the volume coordinate set and corresponds to the maximum vertical coordinate value is determined as the maximum coordinate point. The system responds to the difference in lateral coordinates between the minimum and maximum coordinate points, which corresponds to the lateral coordinate values. The system tilts the belt and obtains the difference in longitudinal coordinates between the minimum and maximum coordinate points. The tilt angle of the corresponding booster belt is determined based on the difference in longitudinal and lateral coordinates, and the extension length of the corresponding booster belt is determined based on the tilt angle.

[0052] For example, in this embodiment, when the server determines that the user is in single-leg assist mode, the server can establish a unified image coordinate system based on the current posture image acquired in real time by the acquisition unit. The Y-axis of the image coordinate system extends along the direction of gravity, and the change of the longitudinal coordinate value is directly related to the vertical stretch dimension of the assist belt. This provides a unified benchmark for subsequent belt coordinate analysis and angle calculation, ensuring the consistency of data calculation under different training scenarios and avoiding extension length errors caused by benchmark differences. Next, the server accurately identifies and extracts the coordinates of the belt body region in the posture image. That is, based on the established image coordinate system, it extracts the coordinate information of all image pixels in the belt body region to form a complete belt body coordinate set, comprehensively recording the spatial position characteristics of the assistive belt. Furthermore, it filters out key coordinate points from the belt body coordinate set. Specifically, the image coordinate point corresponding to the minimum vertical coordinate value is defined as the minimum coordinate point, and the image coordinate point corresponding to the maximum vertical coordinate value is defined as the maximum coordinate point. It can be noted that the minimum coordinate point corresponds to the end of the assistive belt that contacts the user's leg, while the maximum coordinate point should correspond to the fixed end of the assistive belt located in the training structure. The coordinates of these two endpoints can completely represent the overall stretching shape of the assistive belt, laying a data foundation for subsequent tilt state judgment and angle calculation. Then, the server determines whether the booster belt is tilted by analyzing the coordinate difference between the two polar coordinate points. That is, the horizontal coordinate values ​​of the minimum coordinate point and the maximum coordinate point can be compared. If there is a horizontal coordinate difference of more than a preset threshold, it means that the two ends of the booster belt are offset in the horizontal direction. The server directly determines that the booster belt is tilted. At this time, the server can simultaneously obtain the vertical coordinate difference between the minimum coordinate point and the maximum coordinate point. Since this difference reflects the stretching span of the booster belt in the vertical direction, it together with the horizontal coordinate difference constitutes the basis for calculating the tilt angle, ensuring that the angle calculation can fully reflect the spatial tilt characteristics of the booster belt. Subsequently, the server calculates the tilt angle of the assistive belt based on the difference between the longitudinal and lateral coordinates using trigonometric functions. This tilt angle quantifies the degree to which the assistive belt deviates from the vertical state, directly relating to the user's body posture when exerting force with one leg and the force pattern of the assistive belt. It is a key parameter characterizing the training state of single-leg assisted movement, avoiding the problem of not being able to quantify the posture of the assistive belt in traditional training. Specifically, the trigonometric function operation can be the tangent function, and the corresponding calculation method can be: tilt angle = arctan(lateral coordinate difference / longitudinal coordinate difference). Finally, based on the determined tilt angle and the difference in vertical coordinates between the maximum and minimum coordinate points, the server again uses trigonometric functions to determine the extension length of the assistive band. It can be seen that since the tilt angle is the angle between the assistive band and the vertical direction, the extension length can be derived according to the cosine theorem of trigonometric functions. This is key data representing the core state of the user during single-leg assisted training, directly reflecting the correlation between the stretching state of the assistive band and the user's force exertion during training. It provides accurate and reliable basic data support for the selection of the effect indication length for each training cycle.

[0053] For example, a user with a weakness in one side uses single-leg assisted training. During training, a slight lateral shift in the body causes the assist belt to tilt. After the server establishes an image coordinate system along the gravity direction along the Y-axis based on the posture image, it extracts the belt body coordinate set for the belt body area and filters out the minimum and maximum coordinate points. The minimum coordinate point corresponds to a horizontal coordinate value of 20cm and a vertical coordinate value of 15cm, while the maximum coordinate point corresponds to a horizontal coordinate value of 28cm and a vertical coordinate value of 85cm. Based on the calculation, the horizontal coordinate difference is 8cm, indicating that the assist belt is tilted. At the same time, the server obtains the vertical coordinate difference of 70cm and further calculates the tilt angle to be approximately 6.5° and the extension length to be approximately 70.45cm through trigonometric functions. This value is recorded as the current extension length of the assist belt, providing accurate data for determining the subsequent effect indication length.

[0054] In summary, this embodiment can improve pull-up training efficiency and meet the scientific training needs of different groups. Specifically, it can be based on the following: First, the two-level identity verification mechanism ensures the relevance and continuity of training. The primary verification locks the user's identity by collecting information and accurately retrieves historical data to determine the current strength level, avoiding the bias of blindly assessing one's own ability in unguided training. The secondary verification ensures that each stage of training is executed by the target user, preventing the training plan from becoming chaotic and laying a solid foundation for personalized training. Secondly, the strength level-adapted courses achieve personalized solutions for each individual. This embodiment can select suitable content from the course library based on the user's strength data. Beginners can obtain low-intensity introductory programs, while advanced learners can be matched with high-intensity breakthrough courses. This solves the problem of low efficiency caused by general tutorials and makes the training intensity and ability accurately matched. Finally, posture acquisition and scoring feedback optimize training movements in real time. This embodiment can provide standard demonstrations based on display videos. The acquisition unit simultaneously captures the user's posture and generates scores, allowing trainees to intuitively understand movement defects and make timely corrections to avoid ineffective training and sports injuries. Compared with the unguided mode, this embodiment can improve the scientific nature and efficiency of training, and provide a reliable guarantee for improving physical fitness and achieving assessment standards.

[0055] Figure 5 A system block diagram of a pull-up training system according to another embodiment of the present invention is shown, such as Figure 5 As shown, it includes: The master-level verification module is configured to respond to the user terminal selecting any training mode based on the display unit, trigger the collection unit to collect the user terminal's identity, and perform master-level identity verification based on the collected information. The course adaptation module is configured to respond to the verification confirmation, determine the current strength level data of the corresponding user based on the information collected once, and select training courses that are compatible with the strength level data from the retrieved course library based on the strength level data. The secondary verification module is configured to respond to the trigger signal received for any training course, identify the training stage in the training course that is in the first position according to the training order and is in an unplayed state as the target stage, trigger the collection unit to collect identity based on the target stage, and perform secondary identity verification based on the obtained secondary collection information. The training scoring module is configured to respond to secondary identity verification to confirm that the verification has passed, play the display video of the corresponding target stage based on the display unit, and control the acquisition unit to perform posture acquisition, so as to determine the standard training score based on the obtained posture information.

[0056] In the specification provided herein, the algorithms and displays are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used with the examples of this invention. The required structure for constructing such systems is apparent from the above description. Furthermore, this invention is not directed to any particular programming language. It should be understood that the contents of the invention described herein can be implemented using various programming languages, and the above description of specific languages ​​is for the purpose of disclosing preferred embodiments of the invention.

[0057] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0058] Similarly, it should be understood that, in order to streamline this disclosure and aid in understanding one or more of the various aspects of the invention, in the above description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof.

[0059] Those skilled in the art will understand that modules, units, or components of the devices disclosed in the examples herein can be arranged in the devices described in this embodiment, or alternatively, can be located in one or more devices different from the devices in this example. The modules in the foregoing examples can be combined into a single module or, in addition, can be divided into multiple sub-modules.

[0060] Those skilled in the art will understand that the modules in the device of the embodiment can be adaptively changed and placed in one or more devices different from that embodiment. Modules, units, or components in the embodiment can be combined into a single module, unit, or component, and further, they can be divided into multiple sub-modules, sub-units, or sub-components.

[0061] Furthermore, those skilled in the art will understand that although some embodiments described herein include certain features included in other embodiments but not others, combinations of features from different embodiments are meant to be within the scope of the invention and form different embodiments.

[0062] Furthermore, some of the embodiments described herein are methods or combinations of method elements that can be implemented by a processor of a computer system or by other means of performing the functions. Therefore, a processor having the necessary instructions for implementing the methods or method elements forms means for implementing the methods or method elements. Furthermore, the elements described herein in the apparatus embodiments are examples of means for implementing the functions performed by elements for the purposes of carrying out the invention.

[0063] As used herein, unless otherwise specified, the use of ordinal numbers such as “first,” “second,” “third,” etc., to describe ordinary objects merely indicates different instances of similar objects and is not intended to imply that the objects being described must have a given order in time, space, ordering, or any other manner.

[0064] Although the invention has been described with respect to a limited number of embodiments, those skilled in the art will understand from the foregoing description that other embodiments are conceivable within the scope of the invention described herein. Furthermore, it should be noted that the language used in this specification has been chosen primarily for readability and edibility purposes, and not for the purpose of explaining or limiting the subject matter of the invention.

Claims

1. A pull-up training method, characterized in that, Includes the following steps: The user terminal selects any training mode based on the display unit, triggering the collection unit to collect the user terminal's identity and perform primary identity verification based on the collected information. Once the response confirms that the verification has passed, the current strength level data of the corresponding user terminal is determined based on the information collected once, and a training course that matches the strength level data is selected from the retrieved course library based on the strength level data. Upon receiving a trigger signal corresponding to any training course, the training stage included in the training course and located first in the training order that is in an unplayed state is identified as the target stage. The trigger collection unit performs identity collection based on the target stage and performs secondary identity verification based on the obtained secondary collection information. The response determines that the verification is successful based on the secondary identity verification, plays the display video of the corresponding target stage based on the display unit, and controls the acquisition unit to perform posture acquisition, so as to determine the standard training score based on the obtained posture information.

2. The method according to claim 1, characterized in that, The user terminal selects any training mode based on the display unit, triggering the acquisition unit to collect the user's identity information. Based on the collected information, a primary-level identity verification is performed, including: When the pressure value output by the pressure sensor pre-set on the training table is greater than the preset value, the display state of the corresponding training display interface on the display unit is controlled to change from the locked state to the interactive state. When any training mode on the training display interface is continuously interacted with for a preset duration, it is determined that the user has selected the training mode. The trigger unit collects the user's identity information and identifies the obtained facial recognition information as a single collection. Retrieve historical registration information and compare the information collected at one time with the facial registration information included in the historical registration information; The response determines that if the collected information is the same as any face registration information based on the comparison, the verification is successful; otherwise, the verification is unsuccessful.

3. The method according to claim 2, characterized in that, The trigger unit collects the user's identity information and identifies the obtained facial recognition information as a single collection, including: The acquisition unit is triggered to acquire images, and an acquisition task with a continuously preset acquisition duration is established based on the acquisition unit. The response determines that the acquired image does not contain any limb area indicating any human limb based on the acquisition task, and controls the lock state of the corresponding training display interface to change from display state to interactive state; otherwise, it generates a face frame with the image center point of the corresponding acquired image as the center. If there is a pixel difference greater than a preset pixel threshold between any image pixel that overlaps with the face frame and the reference pixel of the corresponding human skin, the voice unit pre-set on the training table will be controlled to retrieve and play the posture adjustment audio. If the pixel difference between each image pixel that overlaps with the face frame and the reference pixel of the corresponding human skin is less than or equal to a preset pixel threshold, the image portion of the face frame corresponding to the captured image is determined as one acquisition information.

4. The method according to claim 1, characterized in that, Upon successful verification, the system determines the current strength level of the corresponding user based on the collected information and selects training courses from the retrieved course library that match the strength level data. These courses include: The response determines that the primary identity verification has passed, retrieves the historical database, and iterates through each record in the historical database based on the information collected once. If any recorded data is matched with a previously collected data, the current strength level data of the corresponding user terminal is determined based on the recorded data; otherwise, the current strength level data is determined based on the retrieved strength test process. Training courses that match the strength level data are selected from a retrieved course library.

5. The method according to claim 4, characterized in that, The current strength level data is determined based on the retrieved strength test process, including: Based on the basic information sub-process corresponding to the first test priority included in the strength test process, the physiological baseline information of the user is determined, and based on the age sub-information and physique sub-information included in the physiological baseline information, the age influence coefficient and physique influence coefficient are determined. The strength test process determines the user's strength test value based on the test information sub-process corresponding to the second test priority, and updates the strength test value based on the age influence coefficient and the physique influence coefficient to obtain the updated strength test value.

6. The method according to claim 5, characterized in that, Training courses that match the strength level data are selected from a retrieved course library, including: Retrieve the course library, which includes each training course and the strength level range corresponding to each training course; If the strength level data falls within a strength level range, the training course corresponding to that strength level range will be determined as being adapted to the strength level data. In response to the strength level data being located in multiple strength level intervals, the appropriate age value corresponding to each strength level interval is determined, and the user age value of the corresponding user terminal is determined based on the age sub-information. Training courses that correspond to age-appropriate values ​​that are less than the user's age and have the smallest age difference are identified as those that are adapted to the strength level data.

7. The method according to claim 1, characterized in that, The response, based on secondary identity verification confirming successful verification, retrieves and plays the demonstration video for the corresponding target stage, and controls the acquisition unit to perform posture acquisition. A standard training score is then determined based on the obtained posture information, including: The response is based on secondary identity verification to confirm that the verification is successful. The corresponding target stage display video is retrieved and played, and the acquisition unit is controlled to perform posture acquisition to obtain the posture information of the user terminal when playing the display video to different training segments, and to obtain the preset limb movements corresponding to each training segment. Based on posture information, the degree of motion variation of the corresponding preset limb movements is determined, and the average of all motion variation degrees in the corresponding display video is calculated to determine the standard training score.

8. The method according to claim 7, characterized in that, Based on posture information, the degree of motion variation for corresponding preset limb movements is determined, and the average of all motion variation degrees in the corresponding demonstration video is calculated to determine a standard training score, including: Each action variation degree is compared with the retrieved preset variation threshold, and if any action variation degree is greater than the preset variation threshold, the action variation degree is determined to have a deviation attribute. Obtain the deviation ratio of all actions with deviation attributes, and if the deviation ratio is greater than a preset ratio threshold, determine the first score retrieved as the standard training score. If the response deviation ratio is less than or equal to the preset ratio threshold, the average of all motion changes in the corresponding video will be calculated, and the resulting second score will be determined as the standard training score.

9. The method according to claim 1, characterized in that, The method further includes: Image recognition is performed on the posture image obtained based on posture information, and the posture image is determined based on the recognition result to include the belt body area indicating the assistive belt, and the belt body area has a contact relationship with the leg area indicating the user's leg limbs, and the user is judged to be in an assisted exercise mode. The response-assisted exercise mode is a two-leg assisted mode, and based on the posture image, it is determined that different leg limbs correspond to different assist belts with different force exertion situations. The control telescopic unit adjusts the contraction of the assist belt and obtains the extension length of the corresponding assist belt. The response-assisted training method is a single-leg assistance method, and the assistance belt is tilted based on the belt area. The extension length of the assistance belt is determined based on the tilt angle of the corresponding tilt state. Based on posture information, all posture images corresponding to each training cycle are combined into training videos, and the longest extension length obtained for each training video is determined as the effect indication length. The lengths of each effect indicator are sorted according to the time sequence of each corresponding training cycle, and the display unit is controlled to display the resulting effect sequence.

10. The method according to claim 9, characterized in that, The assisted training method is a bi-leg assisted method, and based on posture images, it determines that different leg limbs correspond to different assistive straps with different force exertions. The control telescopic unit adjusts the contraction of the assistive straps and obtains the extension length of the corresponding assistive straps, including: The responsive training method is a two-legged assisted method, and an image coordinate system is established based on the posture image, wherein the Y-axis of the image coordinate system extends along the direction of gravity; The two belt areas that have contact with the two leg areas indicating different leg limbs and indicate different assistive belts are defined as the first contact belt and the second contact belt, respectively. The first belt coordinate group and the second belt coordinate group that make up the first contact belt and the second contact belt are determined based on the image coordinate system. The image coordinate point located in the first volume coordinate group and corresponding to the minimum vertical coordinate value is determined as the first minimum coordinate point, and the image coordinate point located in the second volume coordinate group and corresponding to the minimum vertical coordinate value is determined as the second minimum point; In response to the longitudinal coordinate difference between the first minimum coordinate point and the second minimum coordinate point, which have corresponding longitudinal coordinate values, it is determined that different leg limbs correspond to different assistive belts with different force exertion situations, and the telescopic unit is controlled to adjust the length of the assistive belt corresponding to the minimum coordinate point with the smaller longitudinal coordinate value. The belt coordinate system is updated accordingly based on the adjusted belt, and the extension length of the corresponding belt is obtained based on the longitudinal coordinate value corresponding to the minimum coordinate point in the updated belt coordinate system.

11. The method according to claim 9, characterized in that, The assisted training method is a single-leg assisted method, and the assisted belt is tilted based on the belt area. The extension length of the assisted belt is determined based on the tilt angle of the corresponding tilt state, including: The response-assisted exercise method is a single-leg assisted method, and an image coordinate system is established based on the posture image, wherein the Y-axis of the image coordinate system extends along the direction of gravity; Based on the image coordinate system, the volume coordinate set that makes up the volume region is determined. The image coordinate point that is located in the volume coordinate set and corresponds to the minimum vertical coordinate value is determined as the minimum coordinate point, and the image coordinate point that is located in the volume coordinate set and corresponds to the maximum vertical coordinate value is determined as the maximum coordinate point. The system responds to the difference in lateral coordinates between the minimum and maximum coordinate points, which corresponds to the lateral coordinate values. The system tilts the belt and obtains the difference in longitudinal coordinates between the minimum and maximum coordinate points. The tilt angle of the corresponding booster belt is determined based on the difference in longitudinal and lateral coordinates, and the extension length of the corresponding booster belt is determined based on the tilt angle.

12. A training device adapted to the pull-up training method according to any one of claims 1 to 11, characterized in that, include: case; as well as The display unit, processing unit, and acquisition unit are located in the housing.

Images