Leg force training image-based trainer adaptive detection method and system

Abstract

This invention discloses a method and system for adaptive detection of a trainer based on leg strength training images, belonging to the field of image data processing technology. The method includes the following steps: monitoring the initial acquisition qualification of leg strength training images; monitoring the accuracy of leg feature recognition; and monitoring the trainer's reception. This invention improves the accuracy of the trainer's adaptive resistance feedback by evaluating the initial acquisition qualification of leg strength training images and determining whether to optimize the initial acquisition; quantifying the accuracy of leg feature recognition and determining whether to optimize it based on the quantification results; and finally monitoring the trainer's reception accuracy and determining whether to analyze the trainer's feedback accuracy. This solves the problem in existing technologies where low accuracy in adaptive resistance feedback is caused by interference from leg strength training image recognition.

Description

Technical Field

[0001] This invention relates to the field of image data processing technology, and in particular to a trainer adaptive detection method and system based on leg strength training images. Background Technology

[0002] Adaptive detection methods can automatically adapt to image changes under different environments and conditions, providing feedback for subsequent training effects. In filming athletes performing leg strength training in a track and field stadium, the process begins by capturing images of the trainee's leg movements using an autofocus-enabled camera and preprocessing them (e.g., grayscale conversion, binarization) to highlight the leg contours. Then, computer vision techniques (e.g., OpenCV or MediaPipe) are used to detect key points (e.g., knees, ankles) in the leg movement images to analyze the trainee's posture. Next, a data acquisition card connected to the trainer acquires the corresponding data for the leg movement images, transferring the data to the trainer. Finally, image processing algorithms (e.g., contour detection, skeletonization) are used to analyze the data. The system extracts and uses machine learning algorithms (such as support vector machines and deep learning models) to analyze the data and features corresponding to the received leg movement images, such as edge information, joint angles, and movement speed. Based on the analysis results of the data corresponding to the leg movement images, tactile feedback is finally provided through a vibration motor (the vibration motor is placed at the corresponding position of the leg support structure of the trainer, such as installing a vibration motor on the lower leg support structure of the trainer). For example, the vibration motor provides tactile feedback through different vibration modes and intensities. The vibration motor can emit continuous vibrations to prompt athletes to dynamically adjust training parameters, such as resistance and target trajectory, and provide real-time feedback.

[0003] In existing technology, during the data acquisition process corresponding to leg movement images via a data acquisition card connected to the trainer, the data acquisition card acquires leg movement images, such as the muscle strength information of the athlete's left and right legs, and obtains the real-time motor current of the left and right rotating wheels. The data acquisition card is then connected to an electromagnetic resistance control module, which includes a left rotating wheel resistance control submodule and a right rotating wheel resistance control submodule. The left rotating wheel resistance control submodule is used to adjust the motor current of the left rotating wheel to achieve resistance adjustment, and the right rotating wheel resistance control submodule is used to adjust the motor current of the right rotating wheel to achieve resistance adjustment. A twin adaptive training network is built to receive the data corresponding to the leg movement images sent by the data acquisition card. The twin adaptive training network can be called by the electromagnetic resistance control module. For example, based on the athlete's leg muscle strength information and the real-time motor current of the left and right rotating wheels, the twin adaptive training network is called to calculate the resistance balance fitness with the goal of balancing the muscle strength of the left and right legs. Based on the analysis results of the data corresponding to the leg movement images, the trainer outputs the adjustment resistance to adjust the resistance of the left and right rotating wheels of the trainer.

[0004] For example, the invention patent announcement CN118628620B discloses an interactive LED display system for smart sports, which includes: a multimodal motion posture capture unit that captures the athlete's posture data and contact force data with the ground or equipment in real time; an environmental perception unit that monitors and captures real-time environmental data such as temperature, humidity, and light intensity through environmental sensors, and works in conjunction with an adaptive motion standard database; this database stores standard posture data for various sports movements; a posture comparison and analysis unit that uses machine learning algorithms to analyze the captured data and generate accurate posture analysis results; a feedback generation unit that generates posture correction instructions based on these results; and an LED display unit that displays the athlete's current posture image, the adjusted standard posture image, and specific posture adjustment suggestions based on these instructions.

[0005] For example, the invention patent application with publication number CN117237491A discloses a method and medium for generating motion platform state parameters based on video images, which includes: acquiring a set of target state parameters of a target virtual object corresponding to the motion platform in a preset video image; recording the control data generated by the operator on the motion platform while observing the preset video image to form a set of actual real-time state parameters of the platform for the motion platform; obtaining the control quantity that the motion platform needs to adjust based on the set of target state parameters and the set of actual real-time state parameters of the platform; performing saturation processing on the control quantity; converting the saturated control quantity into a control command and sending it to the actuator on the motion platform for execution; and once the latest state parameter fed back by the actuator reaches the target state parameter value, using the set of state parameters of the motion platform at the current moment as the generated state parameters of the motion platform.

[0006] However, in the process of implementing the inventive technical solution in the embodiments of this application, it was found that the above-mentioned technology has at least the following technical problems:

[0007] In existing technologies, when filming athletes performing leg strength training in a track and field stadium, the images captured by the camera may be affected by factors such as lighting conditions. Track and field stadium lighting systems typically consist of high-mast lights installed on poles above the stands or around the perimeter of the field. The light distribution characteristics of these lights result in significant differences in light intensity across different areas of the field. These high-mast lights are usually installed at the four corners or perimeter of the field, and their light coverage forms a cone-shaped area. Areas closer to the lights, such as the edges of the field near the lights, receive more light and have higher intensity, while the center of the field or areas farther from the lights have relatively weaker light intensity. Because the light distribution may be uneven, some areas are brighter than others. This mixed lighting leads to significant brightness differences in different areas of the captured image, resulting in uneven lighting. Under uneven lighting conditions, some areas in the image may be too bright, while others may be too dark.

[0008] Furthermore, histogram equalization enhances contrast by adjusting the histogram distribution of an image to a uniform distribution. This global adjustment ignores the characteristics of different local regions in the image, which may lead to over-enhancing the contrast of some local regions while weakening the contrast of others, resulting in an imbalance in local contrast. The grayscale values ​​of contrasting pixels may be mapped to a wider grayscale range, making them more noticeable. The reduced contrast makes the image less sharp, leading to a decrease in image quality. This can obscure key points of the legs when detected by computer vision techniques (such as OpenCV or MediaPipe), thus interfering with the recognition of leg contours and textures in the image. When the trainer dynamically adjusts training parameters and provides real-time feedback based on the interfered image data received from the data acquisition card, such as the user's posture, movement range, and speed, there is a problem that the trainer's accuracy in adaptive resistance feedback is low due to interference from leg strength training image recognition. Summary of the Invention

[0009] This application provides a trainer adaptive detection method and system based on leg strength training images, which solves the problem in the prior art where the accuracy of the trainer's adaptive resistance feedback is low due to interference from leg strength training image recognition, and improves the accuracy of the trainer's adaptive resistance feedback.

[0010] This application provides a trainer adaptive detection method based on leg strength training images, including the following steps: During the trainer adaptive detection process, an initial acquisition qualification assessment of the leg strength training images is performed. Based on the assessment results, it is determined whether to optimize the initial acquisition of the leg strength training images. Optimizing the initial acquisition of the leg strength training images means improving the uniformity of the grayscale distribution of the leg strength training images by adjusting image grayscale settings and histogram equalization to enhance the qualification of the initial acquisition. After the initial acquisition qualification assessment is passed, the accuracy of leg feature recognition is quantified. Based on the quantification results, it is determined whether to optimize the leg feature recognition. Optimizing the leg feature recognition means enhancing the distinguishability of edge and contour information of the leg strength training images by using non-local mean denoising settings and gradient magnitude mapping to enhance the accuracy of leg feature recognition. After the accuracy quantification of leg feature recognition is passed, the trainer receiving accuracy is monitored. Based on the monitoring results, it is determined whether to analyze the trainer feedback accuracy to reflect the accuracy of the leg strength training image data fed back by the trainer.

[0011] This application provides a trainer adaptive detection system based on leg strength training images, employing a method similar to the one described above. The system includes a leg strength training image initial acquisition qualification monitoring module, a leg feature recognition accuracy monitoring module, and a trainer receiving monitoring module. The leg strength training image initial acquisition qualification monitoring module assesses the qualification of the initial leg strength training image acquisition during the trainer adaptive detection process. Based on the assessment results, it determines whether to optimize the initial leg strength training image acquisition. Optimizing the initial leg strength training image acquisition means improving the uniformity of the grayscale distribution in the leg strength training image through image grayscale settings and histogram equalization to enhance the initial leg strength training image acquisition. The system assesses the leg strength training images for qualification. The leg feature recognition accuracy monitoring module quantifies the accuracy of leg feature recognition after the initial leg strength training image acquisition qualification assessment is passed. Based on the quantification results, it determines whether to optimize leg feature recognition. Leg feature recognition optimization involves using non-local mean denoising and gradient magnitude mapping to enhance the distinguishability of edge and contour information in the leg strength training images, thereby improving the accuracy of leg feature recognition. The trainer receiving monitoring module monitors the trainer receiving accuracy after the leg feature recognition accuracy quantification is passed. Based on the trainer receiving accuracy monitoring results, it determines whether to perform trainer feedback accuracy analysis to reflect the accuracy of the leg strength training image data fed back by the trainer.

[0012] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0013] 1. By conducting an initial assessment of the leg strength training image acquisition qualification and determining whether to optimize the initial acquisition based on the assessment results, the leg strength training images that pass the initial acquisition can more clearly present the muscle, bone, and other features of the legs when the lighting may be uneven, thereby improving the accuracy of leg feature recognition. Then, after the initial acquisition qualification assessment of the leg strength training images passes, the accuracy of leg feature recognition is quantified, and the quantification results determine whether to optimize the leg feature recognition. This helps to compare the changes in leg features in images when the lighting may be uneven. Finally, after the accuracy of leg feature recognition quantification passes, the accuracy of the trainer receiving the image is monitored. The results of the trainer receiving the image accuracy monitoring determine whether to analyze the accuracy of the trainer feedback. This helps to ensure that the trainer can correctly receive the results of leg feature recognition when the lighting may be uneven, and ensures that the trainer feedback matches the actual leg strength training situation. This improves the accuracy of the trainer's adaptive resistance feedback and effectively solves the problem of low accuracy of the trainer's adaptive resistance feedback caused by interference from leg strength training image recognition in the existing technology.

[0014] 2. By quantifying the proportion of preset initial acquisition parameters for leg strength training images, the initial acquisition data of leg strength training images is obtained. The qualified complete interference value of leg strength training images obtained through complete interference assessment is used to obtain the qualified assessment value of the initial acquisition of leg strength training images. This helps to better adapt to different lighting conditions during the adaptive detection process of the trainer in the track and field, thereby reducing the impact of environmental factors such as uneven light distribution on the acquisition of leg strength training images. Compared with the existing analysis of a single variable, this is more conducive to comprehensively evaluating and optimizing the image acquisition process, improving the performance and effect of the entire adaptive detection of the trainer, and realizing the feasibility and adaptability of the qualification assessment of the initial acquisition of leg strength training images.

[0015] 3. When filming athletes performing leg strength training in the track and field, uneven lighting conditions may lead to differences in image data, which in turn affects the accuracy of the trainer's reception and feedback. By monitoring the trainer's reception-qualification verification value, and when the trainer's reception-qualification verification value is not greater than 0, the absolute value of the difference between the trainer's feedback resistance and the preset trainer resistance is used as the trainer feedback verification value. Based on the trainer feedback qualification value, judgment is made, which helps to ensure the accuracy of data reception and feedback under complex lighting conditions. This allows the trainer to provide precise resistance according to the athlete's actual performance, thereby enhancing the trainer's feedback parameters' adaptability to changes in lighting. Attached Figure Description

[0016] Figure 1 A schematic diagram of the overall logic of the adaptive detection method for trainers based on leg strength training images provided in this application embodiment;

[0017] Figure 2 A flowchart of a trainer adaptive detection method based on leg strength training images provided in an embodiment of this application;

[0018] Figure 3 A basic framework diagram for the initial acquisition and optimization of leg strength training images provided in the embodiments of this application;

[0019] Figure 4 A schematic diagram of the structure of the adaptive detection system for a trainer based on leg strength training images provided in an embodiment of this application;

[0020] Figure 5 This is a coordinate diagram of a leg strength training image provided in an embodiment of this application. Detailed Implementation

[0021] This application provides a trainer adaptive detection method and system based on leg strength training images, solving the problem of low accuracy in adaptive resistance feedback caused by interference from leg strength training image recognition in existing technologies. The method involves evaluating the initial acquisition qualification of leg strength training images and determining whether to optimize the initial acquisition based on the evaluation results. Optimization is performed when the initial acquisition qualification value is not greater than a preset qualification value. After passing the initial acquisition qualification evaluation, leg feature recognition accuracy is quantified, and optimization is performed based on the quantification results. Optimization is performed when the muscle positioning error in the image exceeds a preset muscle positioning error obtained from the database. Finally, after passing the leg feature recognition accuracy quantification, trainer reception accuracy is monitored, and the accuracy of trainer feedback is analyzed based on the monitoring results. This improves the accuracy of adaptive resistance feedback by the trainer.

[0022] The technical solution in this application embodiment is to solve the problem of low accuracy in adaptive resistance feedback by the trainer due to interference from leg strength training image recognition. The overall idea is as follows:

[0023] By conducting an initial assessment of the leg strength training image acquisition qualification and determining whether to optimize the initial leg strength training image acquisition based on the assessment results, and then quantifying the accuracy of leg feature recognition after the initial leg strength training image acquisition qualification assessment is passed, determining whether to optimize the leg feature recognition based on the quantification results, and finally monitoring the accuracy of the trainer reception after the accuracy of the leg feature recognition quantification is passed, determining whether to analyze the accuracy of the trainer feedback based on the trainer reception accuracy monitoring results, the effect of improving the accuracy of the trainer in adaptive resistance feedback is achieved.

[0024] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0025] like Figure 1 The diagram shown is a schematic representation of the overall logic of the adaptive detection method for a trainer based on leg strength training images provided in this application embodiment. Figure 1It can be seen that by monitoring the initial acquisition qualification of leg strength training images, the integrity interference value and the initial acquisition qualification evaluation value of leg strength training images are obtained. The larger the integrity interference value of the leg strength training image, the greater the degree of interference to the integrity of the leg strength training image. When the initial acquisition qualification evaluation value of the leg strength training image is not greater than the preset initial acquisition qualification value, the initial acquisition of the leg strength training image is optimized. The optimization of the initial acquisition of the leg strength training image includes setting the image grayscale and histogram equalization clarity optimization settings. The histogram equalization clarity optimization settings also include setting the leg training image sharpening optimization and the Gaussian blur radius. Conversely, if the image muscle positioning error is not greater than the preset image muscle positioning error, leg feature recognition is optimized. Leg feature recognition optimization includes setting non-local mean denoising and obtaining gradient magnitude mapping values. Alternatively, if the image muscle positioning error is not greater than the preset image muscle positioning error, the trainer signal synchronization error is obtained through trainer reception monitoring. If the trainer signal synchronization error is greater than the preset trainer signal synchronization error, trainer reception is optimized. Otherwise, a trainer feedback verification value is obtained. If the obtained trainer feedback pass value is greater than 0, a trainer feedback accuracy pass prompt is sent. Otherwise, a level 5 alarm prompt is sent.

[0026] As a first aspect of the implementation, such as Figure 2 The diagram shows a flowchart of a trainer adaptive detection method based on leg strength training images provided in this application embodiment. The method includes the following steps:

[0027] First, the initial acquisition of leg strength training images is monitored for quality: During the adaptive detection process of the trainer, the initial acquisition of leg strength training images is evaluated for quality. Based on the evaluation results, it is determined whether to optimize the initial acquisition of leg strength training images. Optimization of the initial acquisition of leg strength training images means improving the uniformity of grayscale distribution in the leg strength training images by optimizing image grayscale settings and histogram equalization, thereby improving the quality of the initial acquisition. Monitoring the quality of the initial acquisition of leg strength training images helps to ensure the quality of leg strength training image acquisition, providing clear and accurate basic data for subsequent leg feature recognition, and avoiding recognition errors caused by image quality issues.

[0028] like Figure 3 The diagram shown is a basic framework diagram for the initial acquisition and optimization of leg strength training images provided in this application embodiment. Figure 3It can be seen that when the initial qualified assessment value of the monitored leg strength training image is greater than the preset initial qualified value, the accuracy of leg feature recognition is quantified. Otherwise, the image grayscale is set first, and then the histogram equalization and clarity optimization settings are performed. The image grayscale setting includes obtaining the qualified image grayscale value and obtaining the histogram equalization assessment value. The histogram equalization and clarity optimization settings include sharpening the leg training image and setting the Gaussian blur radius.

[0029] Next, leg feature recognition accuracy monitoring: After the initial acquisition of leg strength training images passes the qualification assessment, the accuracy of leg feature recognition is quantified. Based on the quantification results, it is determined whether to optimize leg feature recognition. Leg feature recognition optimization means using non-local mean denoising and gradient magnitude mapping to enhance the distinguishability of edge and contour information in leg strength training images, thereby enhancing the accuracy of leg feature recognition. Monitoring the accuracy of leg feature recognition helps to improve the recognition accuracy of leg features and accurately extract key features such as leg muscles and bones.

[0030] Finally, trainer reception monitoring: After the accuracy of leg feature recognition is quantified and deemed acceptable, trainer reception accuracy monitoring is performed. Based on the monitoring results, it is determined whether to perform trainer feedback accuracy analysis to reflect the accuracy of the leg strength training image data fed back by the trainer. Leg strength training image data includes leg joint position and angle, image illumination intensity, etc. Trainer reception monitoring helps to ensure the accuracy and completeness of the image data received by the trainer, ensuring that the trainer can correctly feed back training information.

[0031] In this embodiment, when photographing athletes performing leg strength training in the track and field, uneven lighting conditions can lead to differences in image data, thus affecting the accuracy of the trainer's reception and feedback. By interconnecting and interacting with the initial acquisition qualification monitoring of leg strength training images, the accuracy monitoring of leg feature recognition, and the trainer's reception monitoring, accurate leg feature data, such as image joint position coordinates and image muscle attachment point coordinates, is provided for the trainer's reception monitoring. This ensures that the trainer can receive accurate training information, which is beneficial for providing more precise feedback resistance to athletes in the track and field. In this way, the accuracy of the trainer's adaptive resistance feedback is improved.

[0032] It should be added that a database storing various settings data was established before the design of the adaptive detection method for the trainer based on leg strength training images provided in this application. The database includes, but is not limited to, preset leg strength training image resolution deviation, preset leg image light uniformity interference value, preset leg strength training image initial acquisition delay, and preset qualified leg strength training image completeness interference value, etc. The various values ​​are directly set by the technicians; for example, the number of successful recognitions of preset leg feature points is represented by the average number of successful recognitions of preset leg feature points over a historical time period.

[0033] Furthermore, the specific process for evaluating the initial acquisition qualification of leg strength training images is as follows: S1, conduct a complete interference assessment of leg strength training images: by quantifying the proportion of the total number of preset leg feature points recognized in the leg strength training image and the number of successfully recognized preset leg feature points in the leg strength training image, a complete interference value of the leg strength training image is obtained. If the complete interference value of the leg strength training image is not greater than the preset complete interference value, then S2 is executed; otherwise, a level one alarm is sent to indicate that the leg strength training image acquisition is abnormal. The initial acquisition qualification assessment of leg strength training images is used to evaluate the qualification of the initial acquisition of leg strength training images.

[0034] S2, Obtain the initial acquisition qualification value of leg strength training images: S21, quantify the proportion of the preset initial acquisition parameters of leg strength training images and the initial acquisition parameters of leg strength training images to obtain the initial acquisition data of leg strength training images. In this application, the proportion quantification means performing a ratio calculation. The initial acquisition data of leg strength training images includes resolution deviation - initial acquisition qualification index, uniform interference - initial acquisition qualification index, delay - initial acquisition qualification index and image integrity - initial acquisition qualification index.

[0035] Specifically, the expression for the discrimination deviation - initial acquisition qualification index is as follows: , a(W) represents the resolution deviation of the leg strength training image during the Wth preset initial acquisition time period, X1(W) represents the resolution deviation minus the initial acquisition pass index during the Wth preset initial acquisition time period, a(0) represents the preset leg strength training image resolution deviation, and F represents the preset constant. The preset constant is used to avoid the denominator of the initial acquisition data of leg strength training images being meaningless. Specifically, it is used to ensure the effectiveness of evaluating the deviation between the number of pixels in the preset area and the preset number of pixels, the difference between the maximum and minimum brightness of the leg strength training image, the deviation between the duration of the leg strength training image and the preset initial acquisition duration, and the degree of interference to the integrity of the leg strength training image, while providing a basis for data standardization. The standard is defined as W=1,2,...,S, where W represents the number of the preset initial acquisition time period and S represents the total number of preset initial acquisition time periods. The image processor monitors the difference between the average number of pixels in the preset area of ​​the leg strength training image during the preset initial acquisition time period and the preset number of pixels. This difference is used as the resolution deviation of the leg strength training image. Neither the resolution deviation of the leg strength training image nor the preset resolution deviation has a unit. The larger the resolution deviation of the leg strength training image, the less clear or detailed the leg strength training image is compared with the preset state. The resolution deviation of the leg strength training image has a stronger effect on the initial acquisition qualification of the leg strength training image, which in turn leads to a smaller resolution deviation - initial acquisition qualification index.

[0036] Specifically, the expression for uniform interference-initial acquisition qualification index is as follows: X2(W) represents the uniform interference-initial acquisition qualification index for the Wth preset initial acquisition time period, b(W) represents the uniform interference value of leg image light in the Wth preset initial acquisition time period, and b(0) represents the preset uniform interference value of leg image light. The average value of the difference between the maximum and minimum brightness values ​​of the leg strength training image in the preset initial acquisition time period is monitored by the image processor and used as the uniform interference value of leg image light. The unit of the uniform interference value of leg image light and the preset uniform interference value of leg image light are both nits. The larger the uniform interference value of leg image light, the worse the light uniformity. The stronger the effect of the uniform interference value of leg image light on the initial acquisition qualification of leg strength training image, the smaller the uniform interference-initial acquisition qualification index will be.

[0037] Specifically, the expression for the delay-initial data collection qualification index is as follows: X3(W) represents the delay-initial acquisition qualification index of the Wth preset initial acquisition time period, c(W) represents the initial acquisition delay of the leg strength training image in the Wth preset initial acquisition time period, and c(0) represents the preset initial acquisition delay of the leg strength training image. The average value of the difference between the duration of the leg strength training image in the preset initial acquisition time period and the preset initial acquisition duration is monitored by a timer and used as the initial acquisition delay of the leg strength training image. The units of the initial acquisition delay of the leg strength training image and the preset initial acquisition delay of the leg strength training image are both milliseconds. The larger the initial acquisition delay of the leg strength training image, the greater the deviation between the duration of the leg strength training image and the preset initial acquisition duration. The stronger the effect of the initial acquisition delay of the leg strength training image on the initial acquisition qualification of the leg strength training image, the smaller the delay-initial acquisition qualification index.

[0038] Specifically, the expression for the image integrity - initial acquisition qualification index is as follows: X4(W) represents the image integrity-initial acquisition qualification index for the Wth preset initial acquisition time period, d(W) represents the qualified leg strength training image integrity interference value for the Wth preset initial acquisition time period, and d(0) represents the preset qualified leg strength training image integrity interference value. The larger the qualified leg strength training image integrity interference value, the greater the possibility of incomplete leg strength training image acquisition. The stronger the effect of the qualified leg strength training image integrity interference value on the initial acquisition qualification of leg strength training images, the smaller the image integrity-initial acquisition qualification index will be.

[0039] Then, the initial acquisition data of leg strength training images are weighted and coupled with the preset control indicators for initial acquisition of leg strength training images obtained from the database to obtain the qualified evaluation value of the initial acquisition of leg strength training images.

[0040] The initial pass assessment value for leg strength training images was obtained using the following method:

[0041] ;

[0042] In the formula, X(W) represents the initial acquisition qualification assessment value of the leg strength training image in the Wth preset initial acquisition time period, X1 represents the preset resolution deviation-initial acquisition control index, X2 represents the preset uniform interference-initial acquisition control index, X3 represents the preset delay-initial acquisition control index, and X4 represents the preset image integrity-initial acquisition control index.

[0043] Specifically, the initial acquisition parameters for leg strength training images include leg strength training image resolution deviation, leg image light uniformity interference value, initial acquisition delay of leg strength training images, and complete interference value of qualified leg strength training images.

[0044] The resolution deviation of the leg strength training image reflects the degree of deviation between the number of pixels in the preset area and the preset number of pixels. It is represented by the average value of the number of pixels in the preset area of ​​the leg strength training image during the preset initial acquisition time period and the difference between the preset number of pixels. The light uniformity interference value of the leg image reflects the degree of difference between the maximum and minimum brightness values ​​of the leg strength training image. It is represented by the average value of the difference between the maximum and minimum brightness values ​​of the leg strength training image during the preset initial acquisition time period. The initial acquisition delay of the leg strength training image reflects the degree of deviation between the duration of the leg strength training image and the preset initial acquisition duration. It is represented by the average value of the difference between the duration of the leg strength training image during the preset initial acquisition time period and the preset initial acquisition duration. The complete interference value of the qualified leg strength training image is represented by the complete interference value of the leg strength training image that is not greater than the preset complete interference value.

[0045] In addition, the preset initial acquisition parameters for leg strength training images include preset resolution deviation, preset uniformity interference value for leg images, preset initial acquisition delay for leg strength training images, and preset completeness interference value for qualified leg strength training images. These preset initial acquisition parameters are represented by the average value of the initial acquisition parameters for leg strength training images over a historical time period. Preset initial acquisition control indicators for leg strength training images are used to reflect the influence of the initial acquisition data on the qualified assessment value of the initial acquisition of leg strength training images. Specifically, these include: preset resolution deviation - initial acquisition control indicator, preset uniformity interference - initial acquisition control indicator, preset delay - initial acquisition control indicator, and preset image completeness - initial acquisition control indicator. The qualified assessment value for the initial acquisition of leg strength training images reflects the combined effect of the initial acquisition parameters and the preset initial acquisition parameters on the qualification level of the initial acquisition of leg strength training images. The preset initial acquisition time period represents the preset time period for performing the initial acquisition of leg strength training images.

[0046] It should be added that the embodiments of this application provide a set of preset mapping relationships set in advance by preset personnel. The mapping relationships in the mapping set can be one-to-one or many-to-one. The initial acquisition data of leg strength training images and the preset control indicators of the initial acquisition of leg strength training images are mapped one-to-one or many-to-one. The mapping set is obtained from the database and contains the mapping set. The preset control indicators of the initial acquisition of leg strength training images are represented by the proportion of the corresponding initial acquisition data of leg strength training images. By inputting the real-time initial acquisition data of leg strength training images into the corresponding mapping set, the corresponding preset control indicators of the initial acquisition of leg strength training images can be obtained, and the value range is 0-1.

[0047] In this embodiment, the initial acquisition quality of leg strength training images is quantified by analyzing the initial acquisition data to obtain an initial acquisition quality assessment value. A larger initial acquisition data size indicates a stronger influence of the initial acquisition parameters on the initial acquisition quality. Specifically, a larger deviation between the number of pixels within a preset area and the preset number of pixels, a larger difference between the maximum and minimum brightness values ​​of the leg strength training image, a larger deviation between the duration of the leg strength training image and the preset initial acquisition duration, and a larger interference during the initial acquisition process all contribute to a smaller initial acquisition data size, which in turn leads to a smaller initial acquisition quality assessment value. Therefore, in this embodiment, the initial acquisition data and the initial acquisition quality assessment value of leg strength training images are positively correlated.

[0048] In this embodiment, the initial acquisition parameters of the leg strength training images are not isolated but interconnected, requiring correlation analysis to describe their combined effects. A larger resolution deviation in the leg strength training image means the image may be blurry or distorted. A blurry or distorted image may be considered incomplete, leading to a larger interference value for a qualified leg strength training image. A larger uniform light interference value may cause uneven brightness distribution in some areas of the image, making it difficult to clearly display details, resulting in a larger resolution deviation. A larger uniform light interference value in the leg image may affect the acquisition delay, such as automatic exposure adjustment, leading to a larger initial acquisition delay. Through comprehensive analysis of the initial acquisition parameters of the leg strength training images, the movement details during leg strength training can be captured more accurately, achieving precise evaluation of the initial acquisition qualification of the leg strength training images, thereby improving the accuracy of the trainer's adaptive resistance feedback.

[0049] Furthermore, based on the evaluation results, it is determined whether to perform initial acquisition optimization of leg strength training images. The specific process is as follows: The obtained initial acquisition pass evaluation value of leg strength training images is compared with the preset initial acquisition pass value obtained from the database. The preset initial acquisition pass value is represented by the average value of the initial acquisition pass evaluation values ​​of leg strength training images over a historical time period. In this embodiment, the range of the initial acquisition pass evaluation value of leg strength training images is 0.5-1. If the initial acquisition pass evaluation value of leg strength training images is greater than the preset initial acquisition pass value, an image initial acquisition pass prompt is sent, and the accuracy of leg feature recognition is quantified. Conversely, if the image initial acquisition fails, an image initial acquisition fail prompt is sent, and initial acquisition optimization of leg strength training images is performed. The initial acquisition optimization of leg strength training images involves setting image grayscale and histogram equalization clarity optimization settings in sequence.

[0050] The steps for setting the image grayscale are as follows: Step 1: Send a prompt to the designated personnel to input the grayscale values ​​of the unqualified leg strength training images and the initial qualified assessment values ​​of the leg strength training images into the database for grayscale value correction to obtain qualified image grayscale values. Unqualified leg strength training images refer to leg strength training images whose initial qualified assessment values ​​are not greater than the preset initial qualified value. The average grayscale value of each unqualified leg strength training image within a preset time period is monitored by an industrial camera and used as the grayscale value of the unqualified leg strength training image. The database in this application contains a mapping set used to reflect the grayscale values ​​of the unqualified leg strength training images and the initial qualified assessment values ​​of the leg strength training images. The first step involves estimating the value and mapping it to the corresponding qualified image grayscale value. The second step involves sending a prompt to a pre-defined user to set the grayscale of the leg strength training image based on the qualified image grayscale value and histogram equalization. When the prompt for setting the image grayscale is received, setting the grayscale of the leg strength training image based on the qualified image grayscale value and histogram equalization helps to make the grayscale distribution of the leg strength training image more uniform and reasonable. The qualified image grayscale value is within the preset image grayscale range. By controlling the qualified image grayscale value to be within the preset image grayscale range, it helps to avoid loss of image details and insufficient contrast. The preset image grayscale range is preset by a pre-defined user and includes the endpoints of the upper and lower limits of the range.

[0051] In addition, image grayscale settings also include: obtaining a histogram equalization evaluation value, specifically represented by the difference between the initial qualified evaluation value of the leg strength training image at the end of the image grayscale setting and the initial qualified evaluation value of the leg strength training image at the beginning of the image grayscale setting; comparing the obtained histogram equalization evaluation value with the preset equalization evaluation value obtained from the database; if the histogram equalization evaluation value is greater than the preset equalization evaluation value, a prompt indicating that the image grayscale setting is qualified is sent, and the accuracy of leg feature recognition is quantified; otherwise, histogram equalization and clarity optimization settings are performed.

[0052] It should be added that the specific process of histogram equalization and sharpness optimization settings is as follows: T1, perform leg training image sharpening optimization: send a prompt to the preset personnel to process the leg strength training image based on the unsharpening masking method to enhance the clarity and leg texture features of the leg strength training image after image grayscale setting; T2, perform Gaussian blur radius setting: send a prompt to the preset personnel to input the grayscale value of the leg strength training image at the final state of image grayscale setting and the initial qualified evaluation value of the leg strength training image into the database for mapping to obtain a qualified Gaussian blur radius adjustment value. The qualified Gaussian blur radius adjustment value is not greater than the preset maximum image Gaussian blur radius adjustment value obtained from the database. The database of this application contains a set of mapping relationships between the grayscale value of the leg strength training image at the final state of image grayscale setting and the initial qualified evaluation value of the leg strength training image, and the corresponding qualified Gaussian blur radius adjustment value. The preset maximum image Gaussian blur radius adjustment value is set in advance by the preset personnel; T3, send a prompt to the preset personnel to perform step-by-step adjustment based on the magnitude corresponding to the obtained qualified Gaussian blur radius adjustment value. Increasing the Gaussian blur radius of the leg image corresponding to the leg strength training image, by gradually increasing the Gaussian blur radius in increments corresponding to the acceptable Gaussian blur radius adjustment value, helps avoid excessive blurring or loss of detail caused by a sudden increase in the blur radius. The Gaussian blur radius of the leg image should not exceed the preset maximum Gaussian blur radius obtained from the database. If the initial acquisition pass evaluation value of the leg strength training image obtained after histogram equalization clarity optimization settings is still not greater than the preset initial acquisition pass value, controlling the Gaussian blur radius of the leg image to not exceed the preset maximum Gaussian blur radius helps prevent excessive Gaussian blur processing, ensuring that the image retains sufficient detail and information, and sending a secondary alarm to indicate an abnormality in the histogram equalization clarity optimization settings. Conversely, if the histogram equalization clarity optimization settings are received, the accuracy of leg feature recognition is quantified. When the histogram equalization clarity optimization settings are received, the leg training image is first sharpened and optimized, and then the Gaussian blur radius is set, which can effectively enhance the image clarity and leg texture features, making the details in the image more prominent and obvious.

[0053] In this embodiment, by sequentially setting the image grayscale and histogram equalization clarity optimization settings, the leg strength training image is made clearer. The histogram equalization clarity optimization setting can further enhance the details and texture features of the image, thereby improving the pass rate of the initial acquisition of the leg strength training image. This application achieves effective noise suppression and avoids excessive blurring that would lead to the loss of details in the leg strength training image by setting the image grayscale, histogram equalization clarity optimization, leg training image sharpening optimization, and Gaussian blur radius in an orderly and interrelated manner.

[0054] Furthermore, the accuracy of leg feature recognition is quantified, and the quantification result is used to determine whether to optimize leg feature recognition. The specific process is as follows: Judgment is made based on the muscle positioning error of the acquired image: If the muscle positioning error of the acquired image is not greater than the preset muscle positioning error obtained from the database, a leg feature recognition accuracy pass prompt is sent, and the accuracy of the trainer reception is monitored. The preset muscle positioning error is represented by the average value of the muscle positioning error of images over a historical time period; otherwise, a fail prompt is sent, and leg feature recognition optimization is performed. When the detected muscle positioning error is greater than the preset muscle positioning error obtained from the database, it indicates that the current leg feature recognition result is not accurate enough. Optimizing leg feature recognition helps reduce error sources and makes the recognition result closer to reality. The average distance between each successfully identified muscle location point and its corresponding preset muscle location point within the preset recognition time period is monitored by an industrial camera and used as the image muscle positioning error. The preset recognition time period represents the preset time period corresponding to the leg feature recognition accuracy quantification. The image muscle positioning error reflects the accuracy of feature recognition of leg strength training images within the preset recognition time period.

[0055] The specific process of leg feature recognition optimization is as follows: Step 1, Non-local mean denoising setting: A prompt is sent to a pre-defined personnel to process the leg strength training image based on the non-local mean denoising method; Step 2, A prompt is sent to a pre-defined personnel: The grayscale values ​​of qualified image pixels and the muscle positioning error of the image are input into the database for mapping to obtain gradient amplitude mapping values. If the obtained gradient amplitude mapping value is greater than the preset maximum gradient amplitude, a gradient amplitude mapping alarm is sent; otherwise, the corresponding gradient amplitude mapping value is marked as a qualified gradient amplitude mapping value, and a prompt is sent to the pre-defined personnel to gradually increase the gradient amplitude of the leg strength training image according to the amplitude corresponding to the qualified gradient amplitude mapping value. The preset maximum gradient amplitude is set in advance by the pre-defined personnel, and the database contains a mapping set that reflects the mapping relationship between qualified image pixel grayscale values ​​and muscle positioning errors and the corresponding qualified gradient amplitude mapping values. When the leg feature recognition optimization prompt is received, the gradient amplitude of the leg strength training image is gradually increased according to the amplitude corresponding to the qualified gradient amplitude mapping value. The gradient amplitude is related to the edge and detail features of the image. Closely related, the gradual increase in gradient amplitude can make the edges and details in leg strength training images more prominent, enhancing the visibility of leg strength training image features. The gradient amplitude is within a preset qualified gradient amplitude range, which is pre-set by preset personnel and includes the endpoints of the corresponding upper and lower limits. Controlling the gradient amplitude within the preset qualified gradient amplitude range helps maintain the accuracy of leg feature recognition. The qualified image pixel grayscale value represents the grayscale value of the leg strength training image corresponding to the initial qualified evaluation value of the leg strength training image that is not greater than the preset initial qualified value. The accuracy of feature recognition of the leg strength training image is evaluated by obtaining the image muscle localization error. If it is inaccurate, that is, the image muscle localization error is greater than the preset image muscle localization error, then optimization operations such as non-local mean denoising and gradient amplitude mapping adjustment are performed in sequence. First, non-local mean denoising is used to reduce image noise interference and make the image clearer. Then, gradient amplitude mapping adjustment is performed to enhance the edge and detail features of the image, making leg features more obvious, which is convenient for subsequent feature recognition and analysis.

[0056] Meanwhile, leg feature recognition optimization also includes feature recognition-qualification verification. The specific process of feature recognition-qualification verification is as follows: The difference between the quantitative value of leg feature recognition accuracy before optimization and the quantitative value of leg feature recognition accuracy after optimization is obtained and recorded as the feature recognition-qualification verification value. A judgment is made based on the obtained feature recognition-qualification verification value. If the obtained feature recognition-qualification verification value is not greater than 0, a level 3 alarm is sent to indicate an abnormality in leg feature recognition optimization; otherwise, a feature recognition qualification prompt is sent, and the accuracy of the trainer reception is monitored. The feature recognition-qualification verification value reflects the qualification level of leg feature recognition optimization within a preset recognition time period. By controlling the feature recognition-qualification verification value to be greater than 0, it helps to promptly detect and avoid invalid or erroneous optimization operations.

[0057] In this embodiment, the accuracy of leg feature recognition is improved by quantifying the accuracy of leg feature recognition, optimizing leg feature recognition, and verifying the qualification of feature recognition. After optimizing the leg feature recognition, the accuracy of feature recognition is verified again to ensure that the features of the leg strength training image can be accurately recognized.

[0058] Furthermore, after the accuracy of leg feature recognition is quantified as qualified, the accuracy of the trainer reception is monitored. The specific process is as follows: the time when the trainer receives the training signal is monitored by a timer, and the difference between the time and the preset trainer reception time is quantified. In this application, the difference quantification means performing difference calculation, which is used as the trainer signal synchronization error. In this embodiment, the range of the trainer signal synchronization error is 10-20 milliseconds. The trainer signal synchronization error is used to reflect the qualification of the trainer receiving leg strength training image data during the preset trainer reception time period. The trainer signal synchronization error is compared with the preset trainer signal synchronization error obtained from the database. If the trainer signal synchronization error is greater than the preset trainer signal synchronization error, a trainer reception accuracy failure prompt is sent, and the trainer reception is optimized. Otherwise, a trainer reception accuracy qualification prompt is sent, and the trainer feedback accuracy analysis is performed.

[0059] Specifically, trainer reception optimization refers to improving the accuracy of the trainer's reception of leg strength training image data by setting the data bitrate. The data bitrate setting involves sending a prompt to a preset user to gradually increase the trainer's data bitrate by the increment corresponding to the trainer's data bitrate mapping value. When the trainer reception optimization prompt is detected, the trainer's data bitrate is increased step-by-step by the increment corresponding to the trainer's data bitrate mapping value. This allows for the transmission of more leg strength training image data per unit time, better assisting in the monitoring and analysis of leg strength training effects. The trainer's data bitrate does not exceed a preset maximum trainer data bitrate, which is pre-set by a preset user. The trainer's data bitrate mapping value is obtained by inputting the number of transmission requests, the trainer's data transmission bandwidth, and the trainer's signal synchronization error into a database. The database then performs mapping processing to obtain the trainer's data bitrate mapping value. The database contains a mapping set reflecting the mapping relationship between the number of transmission requests, the trainer's data transmission bandwidth, and the trainer's signal synchronization error, and the corresponding data bitrate mapping value.

[0060] It should be added that trainer reception optimization also includes trainer reception accuracy monitoring qualification verification. The specific process is as follows: Obtain the difference between the trainer signal synchronization error before and after trainer reception optimization, and record it as the trainer reception qualification verification value; Based on the obtained trainer reception qualification verification value, a judgment is made: If the trainer reception qualification verification value is greater than 0, a level 4 alarm is sent to indicate that the trainer reception accuracy monitoring is abnormal; otherwise, a trainer reception qualification prompt is sent, and trainer feedback accuracy analysis is performed. By controlling the trainer reception qualification verification value to not be greater than 0, it helps to ensure that the signal synchronization error after trainer reception optimization is not greater than before optimization, avoiding situations where optimization is ineffective or leads to a decrease in reception accuracy, thereby ensuring the qualification of trainer reception and ensuring that the trainer can stably and reliably receive leg strength training image data.

[0061] In this embodiment, by monitoring and optimizing the accuracy of trainer reception, after confirming that the leg features in the image can be accurately identified, the system further checks whether the image data received by the trainer is synchronized and accurate. If it is inaccurate, that is, when the trainer signal synchronization error is greater than the preset trainer signal synchronization error, optimization is performed by adjusting the data bit rate, etc., so that the trainer can stably and accurately receive leg strength training image data; thereby improving the accuracy of the trainer's adaptive resistance feedback.

[0062] It should be added that the specific process of trainer feedback accuracy analysis is as follows: Obtain the trainer feedback verification value, represented by the absolute value of the difference between the trainer resistance fed back by the trainer and the preset trainer resistance; quantify the degree of difference between the trainer feedback verification value and the preset trainer resistance feedback deviation obtained from the database to obtain the trainer feedback pass value, where the preset trainer resistance is represented by the average value of the trainer resistance fed back over historical time periods; based on the trainer feedback pass value, a judgment is made: if the obtained trainer feedback pass value is greater than 0, a trainer feedback accuracy pass prompt is sent; otherwise, a level 5 alarm prompt is sent to indicate an anomaly in the trainer feedback accuracy analysis; the trainer feedback pass value is used to reflect the accuracy of the trainer's adaptive feedback resistance to the preset personnel after receiving leg strength training image data.

[0063] As a second implementation method, such as Figure 4 The diagram shown is a structural schematic of the adaptive detection system for trainers based on leg strength training images provided in this application embodiment. The adaptive detection system for trainers based on leg strength training images provided in this application embodiment applies a method for adaptive detection of trainers based on leg strength training images, including a leg strength training image initial acquisition qualification monitoring module, a leg feature recognition accuracy monitoring module, and a trainer receiving monitoring module.

[0064] like Figure 5 The image shown is a coordinate diagram of a leg strength training image provided in an embodiment of this application. Figure 5 As can be seen, a Cartesian coordinate system is set up with the center point of the collected leg strength training images as the origin to quantify the spatial position of joint movement. The nodes marked in the figure are the corresponding position points of key parts such as the hip, knee, and ankle joints, which facilitates the analysis of the coordinated force exertion of each joint and muscle.

[0065] The initial acquisition qualification monitoring module for leg strength training images is used to evaluate the qualification of the initial acquisition of leg strength training images during the adaptive detection process of the trainer. Based on the evaluation results, it determines whether to optimize the initial acquisition of leg strength training images. Optimizing the initial acquisition of leg strength training images means improving the uniformity of grayscale distribution of leg strength training images by optimizing image grayscale settings and histogram equalization to improve the qualification of the initial acquisition of leg strength training images.

[0066] The leg feature recognition accuracy monitoring module is used to quantify the leg feature recognition accuracy after the initial acquisition of leg strength training images has passed the qualification assessment. Based on the quantification results, it determines whether to optimize the leg feature recognition. Leg feature recognition optimization means using non-local mean denoising and gradient magnitude mapping to enhance the distinguishability of edge and contour information in the leg strength training images, thereby improving the accuracy of leg feature recognition.

[0067] The trainer receiving monitoring module is used to monitor the accuracy of the trainer receiving after the accuracy of leg feature recognition has been quantified and qualified. Based on the monitoring results, it determines whether to perform trainer feedback accuracy analysis to reflect the accuracy of the leg strength training image data fed back by the trainer.

[0068] In this embodiment, the acquisition and monitoring equipment used in this application includes an image processor, a data acquisition card, an industrial camera, a timer, etc. It monitors the qualified value of the trainer's feedback and sends a five-level alarm when the qualified value of the trainer's feedback is not greater than 0, so as to prompt the trainer's feedback accuracy analysis to be abnormal. This ensures that the feedback function of the trainer is always under effective monitoring, reduces failures caused by abnormal feedback, ensures that the training can be carried out normally and stably, and will not be frequently interrupted due to feedback problems. Once an abnormality occurs, it can respond quickly, thereby realizing the improvement of the accuracy of the trainer's adaptive resistance feedback.

[0069] In summary, this embodiment of the application assesses the initial acquisition qualification of leg strength training images and determines whether to optimize the initial acquisition based on the assessment results. Successfully acquired leg strength training images can more clearly present the muscle and bone features of the legs even when the lighting distribution is uneven, thereby improving the accuracy of leg feature recognition. After the initial acquisition qualification assessment is passed, the accuracy of leg feature recognition is quantified, and the quantification results determine whether to optimize the leg feature recognition. This helps to compare changes in leg features in images when the lighting distribution is uneven. Finally, after the accuracy quantification of leg feature recognition is passed, the accuracy of the trainer's reception is monitored. The monitoring results determine whether to analyze the accuracy of the trainer's feedback. This helps ensure that the trainer can correctly receive the leg feature recognition results even when the lighting distribution is uneven, ensuring that the trainer's feedback matches the actual leg strength training situation. This improves the accuracy of the trainer's adaptive resistance feedback and effectively solves the problem in the prior art where the accuracy of the trainer's adaptive resistance feedback is low due to interference from leg strength training image recognition.

[0070] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0071] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0072] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0073] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0074] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0075] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for trainer adaptive detection based on leg strength training images, characterized in that, Comprising the following steps: During the trainer adaptive detection process, leg strength training image initial collection conformity evaluation is performed, and whether to perform leg strength training image initial collection optimization is determined according to the evaluation result, wherein the leg strength training image initial collection optimization means that the leg strength training image gray scale distribution uniformity is improved through image gray scale setting and histogram equalization clear optimization; After the leg strength training image initial collection conformity evaluation is qualified, leg feature recognition accuracy quantification is performed, and whether to perform leg feature recognition optimization is determined according to the quantification result, wherein the leg feature recognition optimization means that the leg strength training image edge and contour information differentiation is enhanced through non-local mean denoising setting and gradient amplitude mapping; After the leg feature recognition accuracy quantification is qualified, trainer receiving accuracy monitoring is performed, and whether to perform trainer feedback accuracy analysis is determined according to the trainer receiving accuracy monitoring result, so as to reflect the accuracy degree of the trainer feedback leg strength training image data.

2. The method of claim 1, wherein the method further comprises: The specific process of the leg strength training image initial collection conformity evaluation is as follows: S1, the total recognition number and the successful recognition number of the preset leg feature points in the leg strength training image are quantified by the proportion, and the leg strength training image complete interference value is obtained, if the leg strength training image complete interference value is not greater than the preset recognition complete interference value, S2 is executed, otherwise a first level alarm prompt is sent to prompt that the leg strength training image collection is abnormal; S2, the preset leg strength training image initial collection parameter and the leg strength training image initial collection parameter are quantified by the proportion, and then the preset leg strength training image initial collection control index obtained from the database is coupled with the weight, to obtain the leg strength training image initial collection conformity evaluation value; The leg strength training image initial collection parameter includes leg strength training image resolution deviation, leg image light uniformity interference value, leg strength training image initial collection delay and qualified leg strength training image complete interference value; The leg strength training image resolution deviation is used to reflect the deviation degree of the pixel number in the preset area and the preset pixel number; The leg image light uniformity interference value is used to reflect the difference degree of the maximum value of the brightness of the leg strength training image and the minimum value of the brightness; The leg strength training image initial collection delay is used to reflect the deviation degree of the time length of the leg strength training image and the preset initial collection time length; The qualified leg strength training image complete interference value is represented by the leg strength training image complete interference value which is not greater than the preset recognition complete interference value; The preset leg strength training image initial collection control index is used to reflect the influence degree of the leg strength training image initial collection data on the leg strength training image initial collection conformity evaluation value; The leg strength training image initial collection conformity evaluation value is used to reflect the action of the leg strength training image initial collection parameter and the preset leg strength training image initial collection parameter on the initial collection conformity degree of the leg strength training image.

3. The method of claim 2, wherein the method further comprises: The specific process of determining whether to perform leg strength training image initial collection optimization according to the evaluation result is as follows: If the initial image collection quality evaluation value of the leg strength training image is greater than the preset initial collection quality value, an image initial collection quality prompt is sent, and leg feature recognition accuracy quantification is performed, otherwise, an image initial collection unqualified prompt is sent and leg strength training image initial collection optimization is performed; The leg strength training image initial collection optimization indicates that image grayscale setting and histogram equalization clear optimization setting are sequentially performed; The image grayscale setting steps are as follows: An unqualified leg strength training image is sent to a preset personnel to input the grayscale value and the leg strength training image initial collection quality evaluation value into a database for grayscale value correction to obtain a qualified image grayscale value, wherein the unqualified leg strength training image represents a leg strength training image corresponding to a leg strength training image initial collection quality evaluation value that is not greater than the preset initial collection quality value; An unqualified prompt is sent and leg feature recognition optimization is performed; The image muscle positioning error is represented by the average value of the distances between each muscle position point successfully identified in the preset identification time period and the corresponding preset muscle position point; 4. The method of claim 3, wherein the method further comprises: The image muscle positioning error is represented by the average value of the distances between each muscle position point successfully identified in the preset identification time period and the corresponding preset muscle position point; ​ ​ ​ 5. The method of claim 1, wherein the method further comprises: determining a leg force training image based on the leg force training image; and determining a leg force training image based on the leg force training image. ​ ​ ​ ​ The image muscle positioning error is used to reflect the accuracy of feature recognition of the leg strength training image in a preset recognition time period.

6. The method of claim 5, wherein the method further comprises: The specific process of the leg feature recognition optimization is as follows: Step one, send a prompt to the preset personnel to process the leg strength training image based on the non-local mean denoising method; Step two, input the qualified image pixel gray value and the image muscle positioning error into the database for mapping to obtain the gradient amplitude mapping value, if the obtained gradient amplitude mapping value is greater than the preset maximum gradient amplitude value, send a gradient amplitude mapping alarm prompt, otherwise, mark the corresponding gradient amplitude mapping value as a qualified gradient amplitude mapping value, and send a prompt to the preset personnel to gradually increase the gradient amplitude of the leg strength training image with the amplitude corresponding to the qualified gradient amplitude mapping value; The qualified image pixel gray value represents the gray value of the leg strength training image corresponding to the leg strength training image initial collection qualified evaluation value not greater than the preset initial collection qualified value.

7. The method of claim 6, wherein the method further comprises: determining a leg force training image based on the leg force training image; and determining a leg force training image based on the leg force training image. The leg feature recognition optimization also includes feature recognition-eligibility verification. The specific process of the feature recognition-eligibility verification is as follows: Obtain the difference between the leg feature recognition accuracy quantization value before leg feature recognition optimization and the leg feature recognition accuracy quantization value after leg feature recognition optimization, denoted as the feature recognition-eligibility verification value; If the obtained feature recognition-eligibility verification value is not greater than 0, send a three-level alarm prompt to prompt the leg feature recognition optimization abnormality, otherwise, send a feature recognition eligibility prompt and perform training device receiving accuracy monitoring; The feature recognition-eligibility verification value is used to reflect the eligibility degree of leg feature recognition optimization in a preset recognition time period.

8. The method of claim 1, wherein the method further comprises: determining a leg force training image based on the leg force training image; and determining a leg force training image based on the leg force training image. After the leg feature recognition accuracy quantization is qualified, the training device receiving accuracy monitoring is performed, and the specific process is as follows: Quantize the difference between the training device receiving training signal time and the preset training device receiving time as the training device signal synchronization error; The training device signal synchronization error is used to reflect the eligibility degree of the training device receiving leg strength training image data in a preset training device receiving time period; If the training device signal synchronization error is greater than the preset training device signal synchronization error, send a training device receiving accuracy unqualified prompt and perform training device receiving optimization, otherwise, send a training device receiving accuracy qualified prompt and perform training device feedback accuracy analysis; The training device receiving optimization means that the training device receiving leg strength training image data accuracy is improved by data code rate setting, and the data code rate setting means sending a prompt to the preset personnel to gradually increase the training device data code rate with the amplitude corresponding to the training device data code rate mapping value; The training device data code rate mapping value is obtained by inputting the transmission request number, the training device data transmission bandwidth and the training device signal synchronization error into the database, and mapping processing in the database to obtain the training device data code rate mapping value; The training device receiving optimization also includes training device receiving accuracy monitoring eligibility verification, and the specific process is as follows: The difference between the training device receiving pre-optimization training device signal synchronization error and the training device receiving post-optimization training device signal synchronization error is obtained, and is recorded as a training device receiving-eligibility verification value; If the training device receiving-eligibility verification value is greater than 0, a fourth-level alarm prompt is sent to prompt that the training device receiving accuracy monitoring is abnormal, otherwise, a training device receiving eligibility prompt is sent, and training device feedback accuracy analysis is performed.

9. The method of claim 1, wherein the method further comprises: determining a leg force training image based on the leg force training image; and determining a leg force training image based on the leg force training image. The specific process of the training device feedback accuracy analysis is as follows: A training device feedback verification value is obtained, which is represented by the absolute value of the difference between the training device feedback training device resistance and the preset training device resistance; The training device feedback verification value and the preset training device resistance feedback deviation obtained from the database are subjected to difference degree quantification to obtain a training device feedback eligibility value; If the obtained training device feedback eligibility value is greater than 0, a training device feedback accuracy eligibility prompt is sent, otherwise, a fifth-level alarm prompt is sent to prompt that the training device feedback accuracy analysis is abnormal; The training device feedback eligibility value is used to reflect the feedback accuracy of the adaptive feedback resistance to the preset person after the training device receives the leg strength training image data.

10. A trainer adaptive detection system based on leg strength training images, applying the trainer adaptive detection method based on leg strength training images according to any one of claims 1-9, characterized in that, It comprises a leg strength training image initial collection eligibility monitoring module, a leg feature recognition accuracy monitoring module, and a training device receiving monitoring module: The leg strength training image initial collection eligibility monitoring module is used to perform leg strength training image initial collection eligibility evaluation during the training device adaptive detection process, and to determine whether to perform leg strength training image initial collection optimization according to the evaluation result, wherein the leg strength training image initial collection optimization represents that the leg strength training image gray scale distribution uniformity is improved through image gray scale setting and histogram equalization clarity optimization; The leg feature recognition accuracy monitoring module is used to perform leg feature recognition accuracy quantification after the leg strength training image initial collection eligibility evaluation is qualified, and to determine whether to perform leg feature recognition optimization according to the quantification result, wherein the leg feature recognition optimization represents that the leg strength training image edge and contour information differentiation is enhanced through non-local mean denoising setting and gradient amplitude mapping; The training device receiving monitoring module is used to perform training device receiving accuracy monitoring after the leg feature recognition accuracy quantification is qualified, and to determine whether to perform training device feedback accuracy analysis according to the training device receiving accuracy monitoring result, so as to reflect the accuracy degree of the training device feedback leg strength training image data.

Images