Sit-up detection method and device and electronic equipment

Abstract

The invention provides a sit-up detection method and device and electronic equipment, and relates to the field of data processing. The method comprises the following steps: acquiring motion image data, sent by a 3D depth camera, of a user during sit-up; track identification is carried out on the motion image data, and the head motion track and the shoulder lifting height of the user are determined; acquiring back pressure data sent by the pressure sensor equipment; generating a detection result based on the back pressure data, the head movement track and the shoulder lifting height; if it is determined that the detection result indicates that the sit-up action of the user is qualified, the sit-up action of the user is counted in real time, and a statistical result is generated; and marking a statistical result into the moving image data in real time, and generating an AR sit-up image corresponding to the user. By implementing the technical scheme provided by the invention, it is convenient to detect sit-ups.

Description

Technical Field

[0001] The present application relates to the technical field of data processing, and in particular to a sit-up detection method, device and electronic equipment. Background Art

[0002] In modern society, with the improvement of people's health awareness and the change of lifestyle, more and more people begin to pay attention to physical exercise and physical health. Sit-ups, as a simple and effective way of fitness, are naturally widely welcomed and sought after. Whether in the gym, park, school or home, you can see people doing sit-ups. In addition, sit-ups are also widely used in various sports competitions and physical fitness tests, becoming one of the important indicators for evaluating personal physical fitness level and training effect.

[0003] At present, the traditional sit-up detection method mainly relies on manual counting. Although this method can record the number of sit-ups performed by users to a certain extent, it has many limitations, especially in terms of data management. First, manual counting relies on the subjective judgment of the observer, which is prone to inaccurate data due to distraction, visual fatigue or counting errors. This inaccuracy not only affects the evaluation of training effects, but also may mislead users' judgment of their own physical condition. Secondly, the manual counting method is difficult to achieve fast and effective management of sit-up data for a large number of users. In collective training or physical fitness tests, facing many participants, manual counting is not only inefficient, but also prone to errors, and it is difficult to meet the accuracy requirements of modern sports training for data management. Therefore, the above method is not conducive to the detection of sit-ups.

[0004] Therefore, a detection method, device and electronic equipment for sit-ups are badly needed. Summary of the invention

[0005] The present application provides a sit-up detection method, device and electronic equipment, which are convenient for detecting the sit-up of a user.

[0006] In a first aspect of the present application, a sit-up detection method is provided, the method comprising: obtaining motion image data of a user performing sit-ups sent by a 3D depth camera; performing trajectory recognition on the motion image data to determine the head motion trajectory and shoulder lifting height of the user; obtaining back pressure data sent by a pressure sensor device; generating a detection result based on the back pressure data, the head motion trajectory and the shoulder lifting height; if it is determined that the detection result indicates that the user's sit-up action is qualified, then performing statistics on the user's sit-up action in real time and generating statistical results; annotating the statistical results to the motion image data in real time to generate an AR sit-up image corresponding to the user.

[0007] By adopting the above technical solution, the motion image data obtained by the 3D depth camera can capture the three-dimensional spatial motion details of the user when doing sit-ups, including the precise position changes of the head and shoulders. This high-precision data provides a solid foundation for subsequent trajectory recognition. The trajectory recognition technology can accurately analyze the user's head movement trajectory and shoulder lifting height, ensure strict judgment of the sit-up action standard, and reduce misjudgment caused by human factors. In addition to visual data, back pressure data provided by pressure sensor devices is also introduced. This fusion of multi-source data makes the detection method more comprehensive, which can comprehensively consider the user's action completion and force output, and improve the accuracy and reliability of detection. Once a qualified sit-up action is detected, the system can immediately perform real-time statistics and generate statistical results. This instant feedback mechanism allows users to understand their training progress and effects in a timely manner, so as to adjust their training plans or enhance their training motivation. The real-time statistics function also makes it easier for coaches or trainers to monitor and evaluate the user's training situation, providing data support for the formulation of personalized training plans. The statistical results are annotated to the motion image data in real time to generate the corresponding AR sit-up image for the user, providing users with a new interactive experience. This visual feedback method is not only intuitive and easy to understand, but also can stimulate users' interest and participation, making the training process more interesting and vivid. The application of AR technology also makes it possible to visualize training data. Users can intuitively see their training results and progress trajectory through AR images, thereby enhancing their self-confidence and sense of achievement.

[0008] Optionally, acquiring motion image data of a user doing sit-ups sent by a 3D depth camera specifically includes: receiving original image data sent by the 3D depth camera; performing image segmentation on the original image data to obtain transition image data of the user doing sit-ups; and preprocessing the transition image data to obtain the motion image data, wherein the preprocessing includes denoising, filtering and normalization.

[0009] By adopting the above technical solution, the original image data sent by the 3D depth camera is received to ensure the originality and integrity of the data. The original image data is not processed in any way, and all the detailed information is retained, providing a rich data source for subsequent analysis. The original image data is segmented to obtain the transition image data of the user when doing sit-ups. This step can remove background information irrelevant to the sit-up action and focus on the key parts of the user's body, such as the head, shoulders and back, thereby improving the pertinence and efficiency of data analysis. The transition image data is preprocessed, including denoising, filtering and normalization. Denoising can eliminate random noise and interference in the image and improve the clarity of the image; filtering reduces the tiny fluctuations and details in the image through smoothing, making the image smoother and easier to analyze; normalization converts all image data to the same scale, which is convenient for subsequent unified processing and analysis. These preprocessing steps significantly improve the quality of motion image data and lay a solid foundation for subsequent trajectory recognition and data analysis. The preprocessed motion image data is clearer and smoother, reducing the influence of noise and interference factors, thereby improving the accuracy of trajectory recognition. Accurate trajectory recognition is the key to evaluating whether the sit-up action is qualified and is essential for generating reliable detection results.

[0010] Optionally, performing trajectory recognition on the motion image data to determine the user's head motion trajectory and shoulder lifting height specifically includes: inputting the motion image data into a preset deep learning model to obtain the user's head area; using a preset feature point detection algorithm to perform feature point detection on the head area to obtain target head feature points; using a preset feature point tracking algorithm to track the target head feature points in multiple consecutive frames to obtain the user's head motion trajectory.

[0011] By adopting the above technical solution, the motion image data is input into the preset deep learning model, and the model's powerful learning and feature extraction capabilities can be used to accurately identify the user's head area. After training and optimization with a large amount of data, the deep learning model can capture subtle image features, thereby improving the accuracy of recognition. The preset feature point detection algorithm is used to detect feature points in the head area, and the key points of the head (such as the corners of the eyes, the corners of the mouth, the tip of the nose, etc.) can be accurately located. Subsequently, the feature point tracking algorithm is used to track these key points in multiple consecutive frames, which can accurately depict the user's head movement trajectory. This feature point-based method can effectively reduce recognition errors caused by factors such as image blur and occlusion. Modern deep learning algorithms and feature point detection and tracking algorithms have been continuously optimized to achieve faster processing speeds while ensuring accuracy. This means that the system can process and analyze the user's motion image data in real time and give timely feedback. The entire trajectory recognition process is highly automated without manual intervention. This not only improves processing efficiency, but also reduces errors and deviations caused by human factors.

[0012] Optionally, the performing trajectory recognition on the motion image data to determine the user's head motion trajectory and shoulder lifting height specifically includes: inputting the motion image data into the preset deep learning model to obtain the upper body area of ​​the user; using a preset contour extraction algorithm to perform contour extraction on the upper body area to obtain a shoulder contour line; marking reference points on the shoulder contour line to obtain a shoulder reference point; determining the vertical distance between the shoulder reference point and the ground based on the motion image data; and obtaining the shoulder lifting height based on the vertical distance.

[0013] By adopting the above technical solution, the user's upper body area is identified through the deep learning model, providing a precise range for subsequent contour extraction and shoulder analysis. This helps to exclude irrelevant parts of the image and improve the accuracy and efficiency of the analysis. The preset contour extraction algorithm can accurately depict the contour of the user's upper body, especially the contour line of the shoulder. This helps to analyze the shoulder movement more finely and avoid misjudgment due to image blur or interference. Marking the reference point on the shoulder contour line provides a clear benchmark for quantifying the height of the shoulder lift. This quantification method makes the evaluation results more objective and comparable. By calculating the vertical distance between the shoulder reference point and the ground, the height of the shoulder lift can be accurately obtained. This is crucial for evaluating the standardization and training effect of the sit-up movement.

[0014] Optionally, after the detection result is generated, whether the user's sit-up action is qualified is judged according to the detection result; the judging whether the user's sit-up action is qualified according to the detection result specifically includes: determining the pressure value of the user relative to the ground when performing sit-ups according to the back pressure data; comparing the pressure value with a preset pressure threshold to obtain a first comparison result; determining the curve of the head movement trajectory according to the head movement trajectory; calculating the head lifting height, the head lifting speed and the head lifting acceleration according to the curve of the head movement trajectory; comparing the head lifting height, the head lifting speed and the head lifting acceleration with a preset head lifting height threshold, a preset head lifting speed threshold and a preset head lifting acceleration threshold respectively to obtain a second comparison result; comparing the shoulder lifting height with a preset shoulder lifting height threshold to obtain a third comparison result, and the detection result includes the first comparison result, the second comparison result and the third comparison result; judging whether the user's sit-up action is qualified based on the first comparison result, the second comparison result and the third comparison result.

[0015] By adopting the above technical solution, a comprehensive and detailed evaluation of the user's sit-up action is performed by combining the back pressure data, head movement trajectory and shoulder lift height information. This multi-dimensional evaluation method can more accurately reflect the user's action completion and avoid misjudgment caused by a single indicator. By setting a preset pressure threshold, a preset head lift height / speed / acceleration threshold and a preset shoulder lift height threshold, a clear quantitative standard is provided for the evaluation. This quantitative standard makes the evaluation results more objective and comparable, and reduces the influence of human subjective factors on the evaluation results.

[0016] Optionally, the real-time statistics of the user's sit-up movements and the generation of statistical results specifically include: counting the number of qualified sit-up movements among the user's sit-up movements to obtain a first statistical result; scoring the completion quality of the user's sit-up movements according to preset standard sit-up movements to obtain a second statistical result, and the statistical results include the first statistical result and the second statistical result.

[0017] By adopting the above technical solution, by counting the number of qualified sit-ups performed by the user, i.e., the first statistical result, the user can intuitively see how many standard sit-ups he has completed. This way of quantifying the results can bring an immediate sense of achievement to the user, motivating them to continue working hard and strive to achieve greater breakthroughs in quantity. According to the preset standard sit-up action, the completion quality score of each action of the user is scored, i.e., the second statistical result. This quality assessment method not only allows users to pay attention to the number of actions, but also pays more attention to the quality of the actions. By understanding where they are deficient, users can make targeted improvements to improve their overall skill level. The statistical results include both quantitative statistics and quality scores, providing users with comprehensive feedback. This comprehensive feedback not only allows users to see their training results, but also guides them on how to train more effectively. For example, if the user's quantitative statistics are high but the quality score is low, then more attention may need to be paid to the standardization and normativeness of the action; conversely, if the quality score is high but the quantitative statistics are low, then the intensity and frequency of training may need to be increased. Based on the statistical results, the system or coach can customize a personalized training plan for the user. For example, according to the user's performance in different aspects, a training plan can be designed to improve both quantity and quality. This customized suggestion can better meet the needs of users and improve training effects.

[0018] Optionally, the method also includes: binding the AR sit-up image with the user and storing it in a preset database; obtaining a viewing request sent by an administrator wearing AR glasses; and displaying the AR sit-up image to the administrator through the AR glasses according to the viewing request.

[0019] By adopting the above technical solution, the AR sit-up image is bound to the user, which means that the training process and results of each user can be recorded and presented in a unique way. This personalized experience not only enhances the user's sense of participation and belonging, but also makes the training process more interesting and vivid. The bound AR sit-up image is stored in the preset database to achieve centralized management and long-term preservation of data. This provides users and managers with convenient data traceability capabilities, allowing them to view past training records and results at any time, providing a basis for subsequent training plan formulation and effect evaluation. Managers receive viewing requests by wearing AR glasses and view the user's AR sit-up image in real time. This remote monitoring method breaks the limitations of time and space, allowing managers to pay attention to the user's training situation anytime and anywhere, identify problems in time and give guidance. At the same time, the immersive experience of AR glasses also allows managers to understand the user's training status more intuitively and improve the accuracy and effectiveness of guidance. Through a simple viewing request and response mechanism, an efficient communication channel is established between managers and users. Users can request managers' attention and guidance at any time, and managers can respond quickly and give feedback. This instant communication method promotes collaboration and cooperation between the two parties and improves training efficiency and effectiveness.

[0020] In a second aspect of the present application, a sit-up detection device is provided, the detection device comprising an acquisition module and a processing module, wherein the acquisition module is used to acquire motion image data of a user performing sit-ups sent by a 3D depth camera; the processing module is used to perform trajectory recognition on the motion image data to determine the head motion trajectory and shoulder lifting height of the user; the acquisition module is also used to acquire back pressure data sent by a pressure sensor device; the processing module is also used to generate a detection result based on the back pressure data, the head motion trajectory and the shoulder lifting height; the processing module is also used to count the user's sit-up actions in real time and generate statistical results if it is determined that the detection result indicates that the user's sit-up action is qualified; the processing module is also used to annotate the statistical results to the motion image data in real time to generate an AR sit-up image corresponding to the user.

[0021] In the third aspect of the present application, an electronic device is provided, which includes a processor, a memory, a user interface and a network interface, the memory is used to store instructions, the user interface and the network interface are both used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device performs the method described above.

[0022] In a fourth aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores instructions, and when the instructions are executed, the method described above is executed.

[0023] In summary, one or more technical solutions provided by this application have at least the following technical effects or advantages: The motion image data obtained by the 3D depth camera can capture the three-dimensional spatial motion details of the user when doing sit-ups, including the precise position changes of the head and shoulders. This high-precision data provides a solid foundation for subsequent trajectory recognition. The trajectory recognition technology can accurately analyze the user's head movement trajectory and shoulder lifting height, ensuring strict judgment of the sit-up action standard and reducing misjudgment caused by human factors. In addition to visual data, back pressure data provided by pressure sensor devices is also introduced. This fusion of multi-source data makes the detection method more comprehensive, which can comprehensively consider the user's action completion and force output, and improve the accuracy and reliability of detection. Once a qualified sit-up action is detected, the system can immediately perform real-time statistics and generate statistical results. This instant feedback mechanism allows users to understand their training progress and effects in a timely manner, so as to adjust their training plans or enhance their training motivation. The real-time statistics function also makes it easier for coaches or trainers to monitor and evaluate the user's training situation, providing data support for the formulation of personalized training plans. The statistical results are annotated to the motion image data in real time to generate the corresponding AR sit-up image for the user, providing users with a new interactive experience. This visual feedback method is not only intuitive and easy to understand, but also stimulates users' interest and participation, making the training process more interesting and vivid. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A schematic flow chart of a sit-up detection method provided in an embodiment of the present application.

[0025] Figure 2 Another flowchart of a sit-up detection method provided in an embodiment of the present application.

[0026] Figure 3 A module schematic diagram of a sit-up detection device provided in an embodiment of the present application.

[0027] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0028] Explanation of the reference numerals: 31, acquisition module; 32, processing module; 41, processor; 42, communication bus; 43, user interface; 44, network interface; 45, memory. DETAILED DESCRIPTION

[0029] In order to enable technicians in this field to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments.

[0030] In the description of the embodiments of the present application, words such as "for example" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "for example" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "for example" or "for example" is intended to present related concepts in a specific way.

[0031] In the description of the embodiments of the present application, the meaning of the term "multiple" refers to two or more. For example, multiple systems refer to two or more systems, and multiple screen terminals refer to two or more screen terminals. In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. The terms "include", "comprise", "have" and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0032] In today's society, with the deepening of public health awareness and the transformation of life patterns, physical exercise and physical health have become the focus of public attention. Sit-ups, with their simplicity and effectiveness, stand out among many fitness methods and are widely favored. From professional fitness venues to public leisure spaces and even home environments, sit-ups are everywhere and have become the first choice for many people's daily exercise. At the same time, it also occupies a place in the field of sports competition and physical fitness assessment, and is a key indicator for measuring individual physical fitness and training effectiveness.

[0033] However, the current means of detecting sit-ups mainly relies on the traditional manual counting method. Although this method has its basic role, it has exposed obvious limitations in multiple dimensions of data processing. The primary problem is that it is too subjective. The counting results are easily affected by the observer's attention fluctuations, visual fatigue and even simple mistakes, which greatly reduces the accuracy of the data. This may not only distort the true feedback of training effectiveness, but also mislead individuals about their own physical fitness level. In addition, in the face of large-scale training or testing scenarios, the manual counting method seems to be incapable of doing so, with low efficiency and rising error rates, and it is difficult to meet the urgent needs of modern sports training for efficient and accurate data management.

[0034] In order to solve the above technical problems, the present application provides a sit-up detection method, referring to Figure 1 , Figure 1 A flowchart of a sit-up detection method provided in an embodiment of the present application. The detection method is applied to a server and includes steps S110 to S160, which are as follows: S110: Acquire motion image data of a user performing sit-ups, sent by a 3D depth camera.

[0035] Specifically, a 3D depth camera is a device that can obtain three-dimensional spatial information of objects in a scene. It measures the distance from each point on the surface of an object to the camera through specific technologies (such as structured light, binocular stereo vision, TOF, etc.), thereby constructing a three-dimensional model or point cloud data of the object. In the sit-up scene, the 3D depth camera can capture the user's body posture and motion trajectory in real time, including the ups and downs of the body, the position changes of the head, etc. These data are of great significance for analyzing the user's motion posture, evaluating the motion effect, and providing personalized motion guidance. The server plays the role of data reception, processing, and storage here. It receives motion image data from the 3D depth camera, analyzes and processes these data through specific algorithms to extract useful information (such as the user's motion trajectory, speed, acceleration, etc.). The processed data can be used for a variety of purposes, such as motion analysis, health monitoring, and motion guidance. After receiving the data sent by the 3D depth camera, the server first decodes and verifies the data to ensure the integrity and accuracy of the data. Then, the image processing technology and machine learning algorithm are used to perform in-depth analysis of the data to extract key motion features. Finally, the processing results are presented to the user in a visual form or stored in a database for subsequent use. Motion image data refers to a series of images or video frames generated by a 3D depth camera when capturing the user's movements. These data contain the three-dimensional spatial information of the user's body, such as the distance from each pixel to the camera, the shape and outline of the object, etc. In the sit-up scene, motion image data can be used to analyze whether the user's movement posture is correct, whether the movement range is sufficient, and whether the movement speed is appropriate. By analyzing these data, users can be provided with personalized exercise suggestions and guidance to help them achieve better exercise results.

[0036] In a possible implementation, obtaining motion image data of a user doing sit-ups sent by a 3D depth camera specifically includes: receiving original image data sent by the 3D depth camera; performing image segmentation on the original image data to obtain transition image data of the user doing sit-ups; and preprocessing the transition image data to obtain motion image data, wherein the preprocessing includes denoising, filtering, and normalization.

[0037] Specifically, when a 3D depth camera captures a user doing sit-ups, it continuously generates a series of images containing depth information (i.e., raw image data). These data are sent to a server or data processing unit via a communication protocol (such as USB, Wi-Fi, Bluetooth, etc.). Since the raw image data may contain information about the entire scene (not just the user), image segmentation is required to extract the part related to the user's sit-up movement. This usually involves using image processing techniques (such as background subtraction, edge detection, image segmentation, etc.) to identify and separate the user area. The segmented image data (transition image data) is more focused on the user's movement, reducing unnecessary computation and storage space requirements. Preprocessing is a key step in data preparation, which aims to improve data quality and make it more suitable for subsequent analysis and processing. For sit-up motion image data, preprocessing includes the following aspects: Denoising: Remove noise in the image, such as random pixel errors, sensor noise, etc., to improve the clarity and accuracy of the image. Filtering: Apply filters to smooth the image, reduce high-frequency noise and details, while retaining important edge and contour information. This helps to better analyze the user's motion trajectory and posture. Normalization: Converting image data to a uniform format or range to facilitate comparison across time or across users. For example, depth information can be normalized to a range of 0 to 1, or images can be resized to a fixed resolution.

[0038] For example, suppose a student is doing sit-ups in a training field, and the teacher uses a 3D depth camera to track the student's sit-up movements and provide real-time feedback. Here is a specific example of the process: When the student starts doing sit-ups, the 3D depth camera starts capturing depth images of the scene. These images contain three-dimensional information of the student, the floor, the wall, and other objects that may be present. These raw image data are sent to a smartphone or tablet connected to the sit-up device via Wi-Fi. After receiving the raw image data, the server first uses a background subtraction algorithm to identify and remove fixed backgrounds (such as the floor and walls) in the image. Then, through edge detection and image segmentation techniques, the application separates the student area from the image and generates transition image data that only contains the student and its motion trajectory. Next, the server preprocesses the transition image data. It first applies a median filter to remove noise in the image, such as sensor noise or random pixel errors. Then, a Gaussian filter is used to smooth the image, reducing unnecessary details while maintaining the clarity of the student's motion trajectory. Finally, in order to unify the data format, the server normalizes the depth information to a range of 0 to 1 and resizes the image to a fixed resolution suitable for analysis. After the above processing, the server generates high-quality motion image data. This data can now be used for further analysis, such as calculating the student's movement speed, posture accuracy, muscle strength, etc., to provide personalized feedback and suggestions to students.

[0039] S120: Perform trajectory recognition on the motion image data to determine the user's head movement trajectory and shoulder lifting height.

[0040] Specifically, trajectory recognition is an important technology in the field of computer vision and image processing, which aims to extract the motion path of an object from an image or video sequence. In this scenario, the server uses advanced algorithms and models to conduct in-depth analysis of the pre-processed motion image data to identify the motion trajectory of the user's head. The head motion trajectory refers to the movement path of the user's head from the starting position to the highest point and then back to the starting position when the user is doing sit-ups. The server can draw this trajectory by analyzing the position changes of the head in the image. This helps to evaluate the user's movement speed, rhythm and consistency. The shoulder lift height is a key indicator in the sit-up exercise, which reflects the degree of contraction of the user's abdominal muscles and the effect of the exercise. The server can determine the height of the shoulder lift by comparing the distance between the user's shoulder and the ground at different time points. This information is crucial for evaluating the user's exercise quality and training effect.

[0041] In one possible implementation, trajectory recognition is performed on motion image data to determine the user's head motion trajectory and shoulder lift height, specifically including: inputting the motion image data into a preset deep learning model to obtain the user's head area; using a preset feature point detection algorithm to perform feature point detection on the head area to obtain target head feature points; using a preset feature point tracking algorithm to track the target head feature points in multiple consecutive frames to obtain the user's head motion trajectory.

[0042] Specifically, this step involves feeding the captured motion image (usually a series of consecutive video frames) as input into a pre-trained deep learning model. This model may be based on a convolutional neural network (CNN) or other neural network structures suitable for image processing. The role of the model is to identify the user's head area from complex image data. By analyzing the color, texture, shape and other features in the image, the deep learning model can distinguish different parts in the image and separate the head area from the background and other body parts. This is usually achieved by providing a large number of image samples with the head area labeled during the model training phase. After obtaining the head area, the next step is to use a feature point detection algorithm to identify the key points on the head (also called feature points). These feature points are usually points on the head that are easy to identify and relatively stable in different postures, such as the corners of the eyes, the corners of the mouth, the tip of the nose, the chin, etc. The feature point detection algorithm can be implemented based on a variety of technologies such as template matching, shape context, and deep learning methods. After being processed by the feature point detection algorithm, a set of feature points representing the key positions of the head will be obtained. These feature points will serve as the basis for the subsequent tracking algorithm to describe the posture and movement of the head. Next, a feature point tracking algorithm is used to track the position changes of these feature points in consecutive video frames. The tracking algorithm can handle small differences between images and head movements, ensuring that feature points can be accurately tracked even in fast movements or large posture changes. By tracking the target head feature points in multiple consecutive frames, the user's head movement trajectory can be drawn. This trajectory describes the position and posture of the head at different time points, providing an important basis for subsequent motion analysis, evaluation or feedback.

[0043] In one possible implementation, trajectory recognition is performed on motion image data to determine the user's head motion trajectory and shoulder lift height, which specifically includes: inputting the motion image data into a preset deep learning model to obtain the user's upper body area; using a preset contour extraction algorithm to extract the contour of the upper body area to obtain a shoulder contour line; marking reference points on the shoulder contour line to obtain a shoulder reference point; determining the vertical distance between the shoulder reference point and the ground based on the motion image data; and obtaining the shoulder lift height based on the vertical distance.

[0044] Specifically, similar to the previous step, this step involves feeding the captured motion images (video frames) as input into a pre-trained deep learning model. However, this time, the model aims to identify the user's upper body area, not just the head. This usually requires the model to have higher complexity and more extensive training data so that it can accurately distinguish the upper body from the background and other body parts. The deep learning model can identify and segment the user's upper body area by analyzing features in the image, such as color, texture, shape, and human body proportions. This process may involve techniques such as image segmentation, object detection, or human pose estimation. After obtaining the upper body area, a contour extraction algorithm is used to further extract the contour of the upper body. The contour extraction algorithm can be implemented based on edge detection, morphological operations, or deep learning methods. These algorithms can identify the boundaries of the upper body area and generate one or more contours, including the contour of the shoulder. In order to quantify the height of the shoulder lift, one or more reference points need to be marked on the shoulder contour. These reference points are usually representative or easily identifiable points on the shoulder contour, such as the acromion (the highest point of the shoulder) or the junction of the shoulder and the neck. Reference point annotation can be achieved by manual selection, automatic detection, or a combination of both. Next, it is necessary to determine the vertical distance between the shoulder reference point and the ground in each frame of the image. This usually involves techniques such as perspective transformation in the image, depth estimation, or reference points in known scenes (such as markers on the ground). By calculating the pixel distance between the reference point and the ground, and combining the camera's intrinsic and extrinsic parameters (such as focal length, position, and orientation), it can be converted into a vertical distance in the real world. Finally, by comparing the vertical distance between the shoulder reference point and the ground in different frames, the height of the shoulder lift can be calculated. This height can be an increment relative to the starting position (such as the shoulder position when lying on your back) or an absolute height relative to a fixed reference point (such as the ground).

[0045] S130: Acquire back pressure data sent by the pressure sensor device.

[0046] Specifically, a pressure sensor is a device that can detect and convert pressure signals into measurable electrical signals. In this scenario, the pressure sensor is designed to measure and record pressure changes on the user's back. This device may contain multiple pressure sensing points, which can accurately capture the pressure distribution and changes in different areas of the back. Back pressure data refers to information about the pressure on the user's back collected by the pressure sensor device. This data may include the size, distribution, rate of change of pressure, etc., which can reflect the force on the user's back in a specific action or state.

[0047] S140: Generate a detection result based on the back pressure data, the head movement trajectory, and the shoulder lifting height.

[0048] Specifically, during sit-ups, changes in pressure on the back (especially the waist and lower back) can reflect the user's movement completion and force. For example, when the user stands up from a supine position, the back pressure increases; and when the user returns to a supine position, the pressure decreases. These data help determine whether the user has completely stood up and the strength of the stand. The head movement trajectory provides intuitive information about the user's upper body movement. In sit-ups, the movement of the head should be coordinated with the overall movement of the body, that is, from the head close to the ground when lying on the back to the head raised when standing up. By analyzing the head movement trajectory, the user's movement fluidity and posture control can be evaluated. The shoulder lift height is one of the important indicators for measuring the completion of sit-ups. A complete sit-up usually requires the user to lift the shoulders completely off the ground. By monitoring the shoulder lift height, the user's exercise effect can be quantified and it can be determined whether the expected training goals have been achieved.

[0049] For example, by combining these data, the server generates a detailed test result, including the completion degree of each sit-up (based on the height of the shoulders), the correctness of the posture (based on the coordination of the head movement trajectory and the body movement), and the strength (based on the back pressure data). For example, the test result may show: "The completion degree of this sit-up is 90% (the shoulders are basically lifted off the ground), the posture control is good (the head and body move synchronously), but the strength is slightly insufficient (the back pressure peak is low), please try to increase the strength when standing up." Such feedback helps users adjust their training methods and improve training results.

[0050] S150: If it is determined that the detection result indicates that the user's sit-up action is qualified, the user's sit-up action is counted in real time and a statistical result is generated.

[0051] Specifically, first, the server will comprehensively evaluate whether the user's sit-up action is qualified based on the previously mentioned back pressure data, head movement trajectory, and shoulder lifting height. Qualified criteria include shoulders completely lifted off the ground, head and body movement coordination, and back pressure changes in line with expectations. Only when these conditions are met will the server consider the user's sit-up action to be qualified. Once it is determined that the user's action is qualified, the server will start to count the number of sit-ups of the user in real time. This usually means that the server will continue to monitor the user's actions and increase the count each time the user completes a qualified sit-up. During the statistical process or after the statistics are completed, the server will generate statistical results based on the collected data. These results include the total number of sit-ups completed by the user, the average completion time, the fastest / slowest completion time, etc. The statistical results may also be presented in charts, reports, or other visual forms so that users can understand their training more intuitively. In an embodiment of the present application, the statistical results are stored in the form of AR in a motion image video.

[0052] In one possible implementation, the user's sit-up movements are counted in real time and statistical results are generated, specifically including: counting the number of qualified sit-up movements among the user's sit-up movements to obtain a first statistical result; scoring the completion quality of the user's sit-up movements according to preset standard sit-up movements to obtain a second statistical result, and the statistical results include the first statistical result and the second statistical result.

[0053] Specifically, during this process, the server will continuously monitor the user's sit-up movements and determine whether each movement is qualified based on the previously set qualification standards (such as the shoulders are completely lifted off the ground, the head and body movements are coordinated, etc.). For each qualified sit-up movement, the server will count and update the count in real time during the training process so that the user can understand their training progress at any time. When the training is over, the server will provide the total number of qualified sit-ups completed by the user as the first statistical result. In addition to the quantity statistics, the server will also perform a quality assessment on each user's movement based on a preset standard sit-up movement model. This standard model may include the amplitude, speed, fluency, posture accuracy and other aspects of the movement, which is used to measure whether the user's movement is close to the ideal state. For each movement, the server will give a score based on its matching degree with the standard model. These scores can be specific numerical scores or grade ratings (such as excellent, good, average, poor, etc.). After the training is over, the server will combine the scores of all movements and give an overall completion quality score as the second statistical result.

[0054] S160 , annotate the statistical results to the motion image data in real time, and generate an AR sit-up image corresponding to the user.

[0055] Specifically, during the user's sit-up training, the server will count and calculate the user's training data in real time, such as the number of qualified movements, the quality score of movement completion, etc. These statistical results are not only numerical feedback, but also need to be presented to the user in an intuitive way so that they can quickly understand and apply this information. Therefore, the server will annotate the statistical results to the user's current motion image data in real time. This means that the statistical results will be added to the real-time video stream or static image of the user's sit-up training and displayed as part of the image. Once the statistical results are annotated to the motion image data, the server can use augmented reality technology to present these data in a visual way. The AR sit-up image is a composite image that combines real-world images and virtual information (such as statistical results). In this image, the user can see the real picture of himself doing sit-ups, and can also see the statistical result annotations generated by the server on the screen, such as the number of qualified movements, the quality score of movement completion, etc. These annotations can appear in the form of text, numbers, icons or other visual elements, and can be customized and adjusted according to the needs and preferences of the user.

[0056] In a possible implementation, after generating the detection results, judging whether the user's sit-up action is qualified according to the detection results; judging whether the user's sit-up action is qualified according to the detection results specifically includes: determining the pressure value of the user relative to the ground when performing sit-ups according to the back pressure data; comparing the pressure value with a preset pressure threshold to obtain a first comparison result; determining a curve of the user's head motion trajectory according to the head motion trajectory; calculating the head lifting height, head lifting speed and head lifting acceleration according to the curve of the user's head motion trajectory; comparing the head lifting height, head lifting speed and head lifting acceleration with a preset head lifting height threshold, a preset head lifting speed threshold and a preset head lifting acceleration threshold, respectively, to obtain a second comparison result; comparing the shoulder lifting height with a preset shoulder lifting height threshold to obtain a third comparison result, and the detection results include a first comparison result, a second comparison result and a third comparison result; judging whether the user's sit-up action is qualified based on the first comparison result, the second comparison result and the third comparison result.

[0057] Specifically, first, the server analyzes the back pressure data to determine the pressure value of the user relative to the ground when doing sit-ups. This pressure value reflects the force applied to the user's back during the action. Next, the server compares this pressure value with a preset pressure threshold to obtain a first comparison result. The preset pressure threshold is set based on the back pressure range required for a qualified sit-up action. The server generates a curve representing the head movement path by analyzing the data of the head movement trajectory. Then, based on this curve, the server calculates the height, speed, and acceleration of the head lift. These parameters reflect the amplitude, speed, and acceleration changes of the user's head movement. Next, the server compares the calculated head lift height, speed, and acceleration with the preset head lift height threshold, speed threshold, and acceleration threshold, respectively, to obtain a second comparison result. These preset thresholds are also set based on the standards for head movement in a qualified sit-up action. The server also directly compares the user's shoulder lift height with the preset shoulder lift height threshold to obtain a third comparison result. This comparison result is used to evaluate whether the user has fully lifted his shoulders to meet the requirements of a qualified sit-up. Finally, the server will comprehensively judge whether the user's sit-up action is qualified based on the above three comparison results (the first comparison result, the second comparison result and the third comparison result). If all the comparison results meet the preset qualified conditions (that is, the pressure value, the head lifting height, the speed, the acceleration and the shoulder lifting height all reach the preset threshold), the user's sit-up action is judged to be qualified; otherwise, it is judged to be unqualified.

[0058] In one possible implementation, refer to Figure 2 , Figure 2 Another flow chart of a sit-up detection method provided in an embodiment of the present application. The method further includes steps S210 to S230, which are as follows: S210, binding the AR sit-up image with the user and storing it in a preset database; S220, obtaining a viewing request sent by an administrator, who is wearing AR glasses; S230, displaying the AR sit-up image to the administrator through the AR glasses according to the viewing request.

[0059] Specifically, after a user completes a sit-up training and generates a corresponding AR sit-up image, the image is bound to the user. This means that the image is tagged or associated with the user's personal profile or training record so that it can be tracked and identified as to which user the image belongs to. This binding helps ensure the accuracy and traceability of the data, allowing managers or users themselves to easily find and view the training images of a specific user. The bound AR sit-up image is stored in a preset database. This database may be a server or cloud storage solution specifically designed to store training data, user information, and related images. The images stored in the database can be accessed through a specific query or retrieval mechanism so that they can be quickly found and displayed to users or managers when needed. When a manager wants to view a user's AR sit-up image, they send a viewing request to the server in some way (such as using a specific application, sending an email or message, etc.). This request contains enough information for the server to identify which manager wants to view which user's image. Once the server receives the manager's viewing request and verifies the legitimacy of the request, it retrieves the corresponding AR sit-up image from the database. The server then displays the images to managers through AR glasses they wear. AR glasses use augmented reality technology to superimpose virtual images or information onto real-world views, allowing managers to view and evaluate users' training without leaving their current environment.

[0060] The present application also provides a sit-up detection device, referring to Figure 3 , Figure 3 A module schematic diagram of a sit-up detection device provided in an embodiment of the present application. The detection device is a server, and the server includes an acquisition module 31 and a processing module 32, wherein the acquisition module 31 acquires the motion image data of the user performing sit-ups sent by the 3D depth camera; the processing module 32 performs trajectory recognition on the motion image data to determine the user's head motion trajectory and shoulder lifting height; the acquisition module 31 acquires the back pressure data sent by the pressure sensor device; the processing module 32 generates a detection result based on the back pressure data, the head motion trajectory and the shoulder lifting height; if the processing module 32 determines that the detection result indicates that the user's sit-up action is qualified, the user's sit-up action is counted in real time and a statistical result is generated; the processing module 32 annotates the statistical result to the motion image data in real time to generate an AR sit-up image corresponding to the user.

[0061] In a possible implementation, the acquisition module 31 acquires motion image data of a user performing sit-ups sent by a 3D depth camera, specifically including: the acquisition module 31 receives original image data sent by the 3D depth camera; the processing module 32 performs image segmentation on the original image data to obtain transition image data of the user performing sit-ups; the processing module 32 preprocesses the transition image data to obtain motion image data, and the preprocessing includes denoising, filtering, and normalization.

[0062] In one possible implementation, the processing module 32 performs trajectory recognition on the motion image data to determine the user's head motion trajectory and shoulder lifting height, specifically including: the processing module 32 inputs the motion image data into a preset deep learning model to obtain the user's head area; the processing module 32 uses a preset feature point detection algorithm to perform feature point detection on the head area to obtain target head feature points; the processing module 32 uses a preset feature point tracking algorithm to track the target head feature points in multiple consecutive frames to obtain the user's head motion trajectory.

[0063] In one possible implementation, trajectory recognition is performed on motion image data to determine the user's head motion trajectory and shoulder lifting height, which specifically includes: the processing module 32 inputs the motion image data into a preset deep learning model to obtain the user's upper body area; the processing module 32 uses a preset contour extraction algorithm to extract the contour of the upper body area to obtain a shoulder contour line; the processing module 32 marks reference points on the shoulder contour line to obtain a shoulder reference point; the processing module 32 determines the vertical distance between the shoulder reference point and the ground based on the motion image data; the processing module 32 obtains the shoulder lifting height based on the vertical distance.

[0064] In a possible implementation, after generating the detection result, the processing module 32 determines whether the user's sit-up action is qualified according to the detection result; judging whether the user's sit-up action is qualified according to the detection result specifically includes: the processing module 32 determines the pressure value of the user relative to the ground when performing the sit-up exercise according to the back pressure data; the processing module 32 compares the pressure value with the preset pressure threshold to obtain a first comparison result; the processing module 32 determines the curve of the user's head motion trajectory according to the head motion trajectory; the processing module 32 calculates the head lifting height, the head lifting speed and the head lifting acceleration according to the curve of the user's head motion trajectory; the processing module 32 compares the head lifting height, the head lifting speed and the head lifting acceleration with the preset head lifting height threshold, the preset head lifting speed threshold and the preset head lifting acceleration threshold respectively to obtain a second comparison result; the processing module 32 compares the shoulder lifting height with the preset shoulder lifting height threshold to obtain a third comparison result, and the detection result includes the first comparison result, the second comparison result and the third comparison result; the processing module 32 judges whether the user's sit-up action is qualified based on the first comparison result, the second comparison result and the third comparison result.

[0065] In a possible implementation, the processing module 32 counts the user's sit-up movements in real time and generates statistical results, specifically including: the processing module 32 counts the number of qualified sit-up movements among the user's sit-up movements to obtain a first statistical result; the processing module 32 scores the completion quality of the user's sit-up movements according to preset standard sit-up movements to obtain a second statistical result, and the statistical results include the first statistical result and the second statistical result.

[0066] In a possible implementation, the processing module 32 binds the AR sit-up image with the user and stores it in a preset database; the acquisition module 31 obtains a viewing request sent by the manager, who wears AR glasses; the processing module 32 displays the AR sit-up image to the manager through the AR glasses based on the viewing request.

[0067] It should be noted that: when the device provided in the above embodiment realizes its function, only the division of the above functional modules is used as an example. In actual application, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0068] The present application also provides an electronic device, referring to Figure 4 , Figure 4The electronic device may include: at least one processor 41 , at least one network interface 44 , a user interface 43 , a memory 45 , and at least one communication bus 42 .

[0069] The communication bus 42 is used to realize the connection and communication between these components.

[0070] The user interface 43 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 43 may also include a standard wired interface and a wireless interface.

[0071] The network interface 44 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0072] Among them, the processor 41 may include one or more processing cores. The processor 41 uses various interfaces and lines to connect various parts in the entire server, and executes various functions of the server and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory 45, and calling data stored in the memory 45. Optionally, the processor 41 can be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), and programmable logic array (Programmable Logic Array, PLA). The processor 41 can integrate one or a combination of a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU) and a modem. Among them, the CPU mainly processes the operating system, user interface and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 41, and it can be implemented separately through a chip.

[0073] Among them, the memory 45 may include a random access memory (RAM) or a read-only memory (Read-Only Memory). Optionally, the memory 45 includes a non-transitory computer-readable storage medium. The memory 45 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 45 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned method embodiments, etc.; the data storage area may store data involved in the above-mentioned method embodiments, etc. The memory 45 may also be optionally at least one storage device located away from the aforementioned processor 41. As Figure 4 As shown, the memory 45 as a computer storage medium may include an operating system, a network communication module, a user interface module and an application program for a sit-up detection method.

[0074] exist Figure 4 In the electronic device shown, the user interface 43 is mainly used to provide an input interface for the user and obtain data input by the user; and the processor 41 can be used to call an application program for a sit-up detection method stored in the memory 45. When executed by one or more processors, the electronic device executes one or more methods in the above-mentioned embodiments.

[0075] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the order of the actions described, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required for the present application.

[0076] The present application also provides a computer-readable storage medium, which stores instructions. When executed by one or more processors, the electronic device executes one or more of the methods described in the above embodiments.

[0077] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0078] In the several embodiments provided in this application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0079] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0080] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0081] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a memory and includes several instructions for a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned memory includes: various media that can store program codes, such as USB flash drives, mobile hard drives, magnetic disks or optical disks.

[0082] The above is only an exemplary embodiment of the present disclosure, and the scope of the present disclosure cannot be limited thereto. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the disclosure of the specification and the truth of practice, it will be easy for those skilled in the art to think of other embodiments of the present disclosure. This application is intended to cover any modification, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the technical field that are not recorded in the present disclosure. The description and examples are regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. A sit-up detection method, characterized in that: The method comprises: Obtain motion image data of a user doing sit-ups sent by a 3D depth camera; Performing trajectory recognition on the motion image data to determine the head motion trajectory and shoulder lifting height of the user; Get back pressure data sent by the pressure sensor device; Generate a detection result based on the back pressure data, the head movement trajectory and the shoulder lifting height; If it is determined that the detection result indicates that the user's sit-up action is qualified, counting the user's sit-up action in real time and generating statistical results; The statistical results are annotated to the motion image data in real time to generate an AR sit-up image corresponding to the user.

2. The sit-up detection method according to claim 1, characterized in that: The step of obtaining the motion image data of the user performing sit-ups sent by the 3D depth camera specifically includes: Receiving raw image data sent by the 3D depth camera; Performing image segmentation on the original image data to obtain transition image data of the user performing sit-ups; The transition image data is preprocessed to obtain the motion image data, wherein the preprocessing includes denoising, filtering and normalization processing.

3. The sit-up detection method according to claim 1, characterized in that: The performing trajectory recognition on the motion image data to determine the head motion trajectory and shoulder lifting height of the user specifically includes: Inputting the motion image data into a preset deep learning model to obtain the head area of ​​the user; Using a preset feature point detection algorithm to perform feature point detection on the head area to obtain target head feature points; A preset feature point tracking algorithm is used to track the target head feature points in multiple consecutive frames to obtain the user's head movement trajectory.

4. The method for detecting sit-ups according to claim 3, characterized in that: The performing trajectory recognition on the motion image data to determine the head motion trajectory and shoulder lifting height of the user specifically includes: Inputting the motion image data into the preset deep learning model to obtain the upper body area of ​​the user; Using a preset contour extraction algorithm to extract the contour of the upper body area to obtain a shoulder contour line; Marking reference points on the shoulder contour line to obtain shoulder reference points; determining a vertical distance between the shoulder reference point and the ground according to the motion image data; The shoulder lifting height is obtained according to the vertical distance.

5. The method for detecting sit-ups according to claim 4, characterized in that: After the detection result is generated, judging whether the sit-up action of the user is qualified according to the detection result; judging whether the sit-up action of the user is qualified according to the detection result specifically includes: determining, according to the back pressure data, a pressure value of the user relative to the ground when performing sit-ups; Comparing the pressure value with a preset pressure threshold to obtain a first comparison result; Determining a curve of the head motion trajectory according to the head motion trajectory; According to the curve of the head movement trajectory, the head lifting height, the head lifting speed and the head lifting acceleration are calculated; Compare the head lifting height, the head lifting speed, and the head lifting acceleration with a preset head lifting height threshold, a preset head lifting speed threshold, and a preset head lifting acceleration threshold, respectively, to obtain a second comparison result; Comparing the shoulder lifting height with a preset shoulder lifting height threshold to obtain a third comparison result, wherein the detection result includes the first comparison result, the second comparison result and the third comparison result; Based on the first comparison result, the second comparison result and the third comparison result, it is determined whether the sit-up action of the user is qualified.

6. The method for detecting sit-ups according to claim 1, characterized in that: The real-time statistics of the sit-up actions of the user and the generation of statistical results specifically include: Counting the number of qualified sit-up movements among the sit-up movements of the user to obtain a first statistical result; According to a preset standard sit-up action, the completion quality of the user's sit-up action is scored to obtain a second statistical result, which includes the first statistical result and the second statistical result.

7. The sit-up detection method according to claim 1, characterized in that: The method further comprises: Binding the AR sit-up image to the user and storing it in a preset database; Obtaining a viewing request sent by an administrator, wherein the administrator wears AR glasses; According to the viewing request, the AR sit-up image is displayed to the manager through the AR glasses.

8. A sit-up detection device, characterized in that: The detection device comprises an acquisition module (31) and a processing module (32), wherein: The acquisition module (31) is used to acquire motion image data of a user performing sit-ups sent by a 3D depth camera; The processing module (32) is used to perform trajectory recognition on the motion image data to determine the head movement trajectory and shoulder lifting height of the user; The acquisition module (31) is also used to acquire back pressure data sent by the pressure sensor device; The processing module (32) is further used to generate a detection result based on the back pressure data, the head movement trajectory and the shoulder lifting height; The processing module (32) is further configured to count the sit-up movements of the user in real time and generate statistical results if it is determined that the detection result indicates that the sit-up movement of the user is qualified; The processing module (32) is further used to annotate the statistical results into the motion image data in real time, and generate an AR sit-up image corresponding to the user.

9. An electronic device, characterized in that: The electronic device comprises a processor (41), a memory (45), a user interface (43) and a network interface (44), wherein the memory (45) is used to store instructions, the user interface (43) and the network interface (44) are both used to communicate with other devices, and the processor (41) is used to execute the instructions stored in the memory (45) so that the electronic device executes the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed, the method according to any one of claims 1 to 7 is performed.

Images