Athletic training analysis system and analysis method based on artificial intelligence
Through an artificial intelligence-based motion training analysis system, combined with image and physiological parameter detection, action feature extraction and comparison analysis are carried out, the problem of inaccurate motion analysis in the existing technology is solved, and accurate motion training analysis and personalized suggestions are provided. , improves the exercise effect.
Patent Information
- Application Number
- CN202311233522.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-09-22
- Publication Date
- 2024-12-06
AI Technical Summary
The existing sports training analysis system is inaccurate in the motion analysis results of non-professional personnel and is affected by physical physiological parameters, making it difficult to provide reliable analysis results.
Using an artificial intelligence-based motion training analysis system, the motion information is obtained through image detection and physiological parameter detection devices, and the data processor performs keyframe positioning, feature extraction and comparison analysis, and combines standard action models and physiological parameter models to provide action technology and Comprehensive analysis of physical fitness.
Accurate analysis of exercise training movements is achieved, the reliability and generalization of non-professional movement analysis results are improved, personalized exercise plans and healthy diet suggestions are provided, and exercise effects are improved.
Smart Images

Figure CN119092043A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of motion analysis, and in particular relates to a motion training analysis system and analysis method based on artificial intelligence. Background Art
[0002] At present, artificial intelligence technology is widely used in sports-related fields such as auxiliary sports training and intelligent motion analysis. The sports training analysis system currently implemented mainly conducts quantitative analysis of the executed training movements through preset rules, studies various data indicators in sports training, and thus improves the technical level of executing sports training movements. However, for the results of sports training analysis, on the one hand, it is an analysis of the standardization of sports training movements. The execution quality of the movements is analyzed through established rules, and the results of the movements performed by most non-professionals are inaccurate. On the other hand, the results of sports training analysis are also affected by the physiological parameters of the trainees during the sports training process. With the continuous development of artificial intelligence technology, the experience and professional knowledge of motion analysis are converted into motion analysis models, which can improve the reliability and generalization of the results of non-professional motion analysis in sports training. Based on the above problems existing in the field of sports training analysis, there is an urgent need for a sports training analysis system and method based on artificial intelligence. Summary of the invention
[0003] In view of the above shortcomings, the technical problem to be solved by the present invention is to provide a sports training analysis system and analysis method based on artificial intelligence, which are used to provide reliable analysis results for various sports training movements.
[0004] In order to solve the above technical problems, the technical solution adopted by the present invention is:
[0005] A sports training analysis system based on artificial intelligence includes an image detection device for acquiring image information, a physiological parameter detection device for acquiring physiological information of athletes, and a data processor. The image detection device and the physiological parameter detection device are respectively connected to the data processor. The data processor includes a data processing unit, a feature extraction unit, and an analysis unit. The data processing unit performs key frame positioning and training action segment interception on the image information, the feature extraction unit extracts motion features of the intercepted training action segment, and the analysis unit compares and analyzes the extracted motion features with standard action motion models.
[0006] As a preferred solution of the present invention, the data processor further comprises a modeling unit, which establishes a standard motion model of the motion feature according to the extracted motion feature.
[0007] As a preferred solution of the present invention, the data processor further includes a visualization unit, which is used to display the analysis results of the data processing unit and the images and physiological information acquired by the image detection device and the physiological parameter detection device.
[0008] A sports training analysis method based on artificial intelligence, based on the above-mentioned sports training analysis system based on artificial intelligence, comprises the following steps:
[0009] S1, collecting the motion information of the professional sports trainer through the image detection device and the physiological parameter detection device, and establishing a standard training action library through a data processor and storing it in a memory;
[0010] S2, obtaining the user's motion information and physiological information through the image detection device and the physiological parameter detection device, and performing key frame positioning and training action segment selection on the continuous motion training action sequence;
[0011] S3, comparing and analyzing the user's motion information and physiological information with the standard training action model library through the data processing unit.
[0012] As a preferred solution of the present invention, in step S1, the standard training action model library includes a standard training action database A and a healthy exercise physiological index library B, and the training action item name information in the standard training action database A and the healthy exercise physiological index library B are the same.
[0013] As a preferred solution of the present invention, the method for establishing the standard training action database A in step S1 includes the following steps:
[0014] S1.1A forms a sequence according to the coordinates of the two-dimensional key points of the human body in each image frame of the extracted continuous image frames;
[0015] S1.2A uses dimensionality reduction method to extract motion features of standard actions;
[0016] S1.3A The motion characteristics of the standard action are modeled by a data modeling method to obtain a standard action model.
[0017] As a preferred solution of the present invention, the method for establishing a healthy exercise physiological indicator library B in step S1 includes the following steps: according to the stored physiological parameter information during the execution of training movements, the stored healthy exercise physiological parameters are modeled by a data modeling method to obtain a healthy exercise physiological parameter model B.
[0018] As a preferred solution of the present invention, step S3 includes the following specific steps:
[0019] S3.1 inputting the standard training action information stored in the standard training action database A in the above-mentioned standard training action library into the modeling unit, and establishing a standard action model for the training action items corresponding to the above-mentioned standard training action sub-library;
[0020] S3.2 Input the effective training action image frame segments obtained by the data processing unit into the feature extraction unit to obtain the human body training action information corresponding to the effective training action image frame segments; at the same time, obtain the body physiological parameter segments collected by the body physiological parameter detection device within the time period of the effective training action image frame segments through the data processing unit;
[0021] S3.3 Input the human body training action information obtained by the feature extraction unit and the standard action model established for the corresponding training action project into the analysis unit to obtain the action analysis result of the effective training action image frame segment; input the body physiological parameter information in the image frame segment obtained by the data processing unit and the healthy exercise physiological model of the corresponding training action project into the analysis unit to obtain the physiological analysis result of the body physiological parameter segment;
[0022] S3.4 Combine the above-mentioned action analysis results and physiological analysis results to obtain comprehensive analysis results of action techniques and physical fitness in sports training.
[0023] As a preferred solution of the present invention, the analysis unit in step S3.3 analyzes the sports training in combination with the action techniques and physical fitness in the sports training, specifically including the following steps:
[0024] S3.3.1 inputting the human body training movement information and the standard movement model established for the corresponding training movement items into the analysis unit to obtain the movement analysis results of the above body physiological parameter segments;
[0025] S3.3.2 inputting the body physiological parameter information in the image frame segment and the healthy sports physiological model of the corresponding training action item into the analysis unit to obtain the physiological analysis result in the above-mentioned effective action image frame segment;
[0026] S3.3.3 Combine the results of movement analysis and physiological analysis to obtain comprehensive analysis results of movement techniques and physical fitness in sports training.
[0027] The beneficial effects of the present invention are: (1) the sports training analysis system provided by the present invention has a simple structure, can realize the collection and modeling of standard action information and the analysis and feedback of input actions, and has strong practicality and high accuracy.
[0028] (2) The sports training analysis method of the present invention combines the sports information and physiological parameter information of the sports trainees during the execution of training movements to perform sports training analysis, which meets the requirements of the sports trainees for analysis of both technical level and physical fitness, and is convenient for providing a set of reasonable diet recommendations and personalized exercise plans. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] Figure 1 It is a schematic diagram of the sports training analysis system.
[0030] Figure 2 It is a schematic diagram of analysis of the sports training analysis unit of the present invention. DETAILED DESCRIPTION
[0031] The present invention will be further described below in conjunction with the accompanying drawings.
[0032] A sports training analysis system based on artificial intelligence includes an image detection device for acquiring image information, a physiological parameter detection device for acquiring physiological information of athletes, and a data processor. The image detection device and the physiological parameter detection device are respectively connected to the data processor. The image information and the physiological information of the user's body are sent to the data processor through the image detection device and the physiological parameter detection device. The image information and the physiological information are analyzed and processed by the data processor to obtain reliable sports training analysis results.
[0033] The data processor is equipped with a corresponding memory, which is used to store a standard training action library, a monitoring sports physiological index library and various data records in the analysis process.
[0034] As a specific implementation, the image detection device is a camera for acquiring the user's motion images, the physiological parameter detection device includes one or more of a heart rate collector, a blood oxygen detector and a blood pressure detector, so as to acquire various physiological information of the user, and the data processor is a calculator.
[0035] The data processor includes a data processing unit, a feature extraction unit, a modeling unit, an analysis unit, and a visualization unit, wherein the data processing unit is used to locate key frames and extract training action segments for input continuous image frames; the feature extraction unit is used to extract motion features of the selected continuous frame segments of training actions; the modeling unit is used to model the extracted standard action motion features to form a standard action model; the analysis unit is used to compare and analyze the specific motion features of the executed training actions with the standard action motion model based on the human skeleton sequence of the sports trainee during the exercise process; the visualization unit is used to display the analysis results of the data processing unit, the images obtained by the image detection device and the physiological parameter detection device, the physiological information change process, and the motion scene, two-dimensional human posture, and motion feature analysis during the exercise process.
[0036] The data processing process of the data processing unit is as follows:
[0037] (1) locating the image frame at the beginning of the training action as the start frame, and locating the image frame at the end of the training action as the end frame, to obtain a valid training action image frame segment between the start frame and the end frame;
[0038] (2) Obtaining body physiological parameter segments collected by a body physiological parameter detection device during the time period between the start and end of the training action.
[0039] The extraction process of the feature extraction unit is as follows:
[0040] (1) Obtain the topological structure of the human skeleton through the posture estimation algorithm, which mainly includes the two-dimensional coordinates of the key points of the human body in the image frame;
[0041] (2) extracting posture information according to the two-dimensional coordinates, wherein the posture information includes at least two kinds of static posture information, namely, key posture angle and key space angle;
[0042] (3) Extracting static posture information from each frame in the image frame segment to form action information in the effective motion image frame segment.
[0043] The modeling unit models the action information of the standard training actions stored in the standard training action database to form a standard action model. The specific steps are as follows:
[0044] (1) Reduce the dimension of the stored action information through principal component analysis;
[0045] (2) The reduced-dimensional motion information is modeled through GMM to obtain a feature description of the standard training motion. The feature description at least includes the movement process of each organ of the human body during the training motion and the name of the training motion.
[0046] The gap process of the analysis unit is as follows,
[0047] (1) the analysis unit compares and analyzes the action information extracted within the time period of the effective motion image frame segment with the standard training action model formed by the modeling unit to obtain the action technology analysis result;
[0048] (2) comparing and analyzing the body physiological parameter information collected during the time period of the effective motion image frame segment with the health training action physiological parameter indicators stored in the system to obtain the physical fitness analysis results;
[0049] (3) Combine the above-mentioned movement technique analysis results and physical fitness analysis results to obtain sports training analysis results; and based on the analysis results, based on the corresponding healthy diet theory and sports fitness theory, provide a corresponding set of scientific and effective healthy diet plans, personalized exercise plans and work and rest strategies to achieve the purpose of improving exercise effects.
[0050] A sports training analysis method based on artificial intelligence comprises the following steps:
[0051] Step S1, collect sports training action image frames and sports training body physiological parameters, establish a standard training action library, use the image detection device, the physiological parameter detection device and the data processor to establish the standard sports training library, and store it in the memory; wherein the standard training action library includes a standard training action database A and a healthy sports physiological index library B, the training action item name information in the standard training action database A and the healthy sports physiological index library B is the same, and the training action item names include pull-ups, squats, sit-ups, etc.
[0052] Standard training action image frame segments of professional sports trainers and human body training action information obtained after feature extraction are collected by an image acquisition device and a data processor, and stored in a memory, so as to establish a standard training action database A; the standard training action image frame segments stored in the standard training action sub-libraries (wherein the sub-libraries include pull-up sub-libraries, squat sub-libraries, sit-up sub-libraries, etc.) contained in the standard training action database A, and the motion information in the standard training action image frame segments is extracted by a feature extraction unit, and the motion information is a skeleton sequence of human body motion.
[0053] The method for establishing the standard training action database A comprises the following steps:
[0054] S1.1A forms a sequence according to the coordinates of the two-dimensional key points of the human body in each image frame of the extracted continuous image frames;
[0055] S1.2A uses dimensionality reduction method to extract motion features of standard actions;
[0056] S1.3A The motion characteristics of the standard action are modeled by a data modeling method to obtain a standard action model.
[0057] The physiological parameter detection device and the data processor are used to collect the physiological parameter records obtained by professional sports trainers during the execution of standard training movements and store them in the memory. The stored healthy sports physiological parameters are modeled by the data modeling method to obtain the healthy sports physiological parameter model B.
[0058] Step S2, acquisition of motion information and physiological information; temporal positioning and segmentation of the action of the continuous motion training action sequence: the image detection device transmits the continuous motion image frames and the physiological parameter detection device transmits the continuous physiological parameter information to the data analysis module respectively, and the key image frames in the continuous motion image frames are positioned by the data processing unit to obtain effective training action image frame segments and execute the physiological parameter information in the effective training action image frame segments.
[0059] Specifically, the image frames are transmitted to the data processor through the camera, and the effective training action image frame segments are obtained through the data processing unit; the heart rate parameters during the training action are detected by the heart rate collector, the blood oxygen parameters during the training action are detected by the blood oximeter, and the blood pressure parameters during the training action are detected by the blood pressure detector.
[0060] Step S3, comparing and analyzing the user's motion information and physiological information with the standard training action model library through the data processing unit.
[0061] Step S3 includes the following specific steps:
[0062] S3.1 inputting the standard training action information stored in the standard training action database A in the above-mentioned standard training action library into the modeling unit, and establishing a standard action model for the training action items corresponding to the above-mentioned standard training action sub-library;
[0063] S3.2 Input the effective training action image frame segments obtained by the data processing unit into the feature extraction unit to obtain the human body training action information corresponding to the effective training action image frame segments; at the same time, obtain the body physiological parameter segments collected by the body physiological parameter detection device within the time period of the effective training action image frame segments through the data processing unit;
[0064] S3.3 Input the human body training action information obtained by the feature extraction unit and the standard action model established for the corresponding training action project into the analysis unit to obtain the action analysis result of the effective training action image frame segment; input the body physiological parameter information in the image frame segment obtained by the data processing unit and the healthy exercise physiological model of the corresponding training action project into the analysis unit to obtain the physiological analysis result of the body physiological parameter segment;
[0065] S3.4 Combine the above-mentioned action analysis results and physiological analysis results to obtain comprehensive analysis results of action techniques and physical fitness in sports training.
[0066] In step S3.3, the analysis unit analyzes the sports training in combination with the action techniques and physical fitness in the sports training, which specifically includes the following steps:
[0067] S3.3.1 inputting the human body training movement information and the standard movement model established for the corresponding training movement items into the analysis unit to obtain the movement analysis results of the above body physiological parameter segments;
[0068] S3.3.2 inputting the body physiological parameter information in the image frame segment and the healthy sports physiological model of the corresponding training action item into the analysis unit to obtain the physiological analysis result in the above-mentioned effective action image frame segment;
[0069] S3.3.3 Combine the results of movement analysis and physiological analysis to obtain comprehensive analysis results of movement techniques and physical fitness in sports training.
[0070] Step S4, feedback and suggestions on exercise training; using the data processor to provide corresponding diet suggestions, personalized exercise training and work and rest strategy suggestions for the analysis results output by the above-mentioned analysis unit according to the analysis results and the test person's own physiological parameters such as height and weight, and the normal required protein, fat and carbohydrates, so as to improve the test person's exercise effect.
[0071] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention; therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.
[0072] Although this article uses more terms corresponding to the figure marks in the figures, it does not exclude the possibility of using other terms; these terms are used only to more conveniently describe and explain the essence of the present invention; interpreting them as any additional limitations is contrary to the spirit of the present invention.
Claims
1. A sports training analysis system based on artificial intelligence, characterized in that: The invention comprises an image detection device for acquiring image information, a physiological parameter detection device for acquiring physiological information of athletes, and a data processor. The image detection device and the physiological parameter detection device are respectively connected to the data processor. The data processor comprises a data processing unit, a feature extraction unit, and an analysis unit. The data processing unit performs key frame positioning and training action segment interception on the image information. The feature extraction unit extracts motion features of the intercepted training action segment. The analysis unit compares and analyzes the extracted motion features with the standard action motion model.
2. The sports training analysis system based on artificial intelligence according to claim 1, characterized in that: The data processor also includes a modeling unit, which establishes a standard motion model of the motion feature according to the extracted motion feature.
3. The sports training analysis system based on artificial intelligence according to claim 1, characterized in that: The data processor also includes a visualization unit, which is used to display the analysis results of the data processing unit and the images and physiological information acquired by the image detection device and the physiological parameter detection device.
4. A sports training analysis method based on artificial intelligence, characterized in that: A sports training analysis system based on artificial intelligence according to any one of claims 1 to 3 comprises the following steps: S1, collecting the motion information of the professional sports trainer through the image detection device and the physiological parameter detection device, and establishing a standard training action library through a data processor and storing it in a memory; S2, obtaining the user's motion information and physiological information through the image detection device and the physiological parameter detection device, and performing key frame positioning and training action segment selection on the continuous motion training action sequence; S3, comparing and analyzing the user's motion information and physiological information with a standard training action model library through a data processing unit; S4, feedback and suggestions on sports training.
5. The sports training analysis method based on artificial intelligence according to claim 4 is characterized in that: In step S1, the standard training action model library includes a standard training action database A and a healthy exercise physiological index library B, and the training action item name information in the standard training action database A and the healthy exercise physiological index library B are the same.
6. The sports training analysis method based on artificial intelligence according to claim 5, characterized in that: The method for establishing the standard training action database A in step S1 includes the following steps: S1.1A forms a sequence based on the coordinates of the two-dimensional key points of the human body in each image frame of the extracted continuous image frames; S1.2A uses a dimensionality reduction method to extract the motion features of the standard action; S1.3A The motion characteristics of the standard action are modeled by a data modeling method to obtain a standard action model.
7. The sports training analysis method based on artificial intelligence according to claim 5, characterized in that: The method for establishing the healthy exercise physiological index library B in step S1 includes the following steps: according to the stored physiological parameter information during the execution of training movements, the stored healthy exercise physiological parameters are modeled by a data modeling method to obtain a healthy exercise physiological parameter model B.
8. The sports training analysis method based on artificial intelligence according to claim 5, characterized in that: Step S3 includes the following specific steps: S3.1 inputting the standard training action information stored in the standard training action database A in the above-mentioned standard training action library into the modeling unit, and establishing a standard action model for the training action items corresponding to the above-mentioned standard training action sub-library; S3.2 Input the effective training action image frame segments obtained by the data processing unit into the feature extraction unit to obtain the human body training action information corresponding to the effective training action image frame segments; at the same time, obtain the body physiological parameter segments collected by the body physiological parameter detection device within the time period of the effective training action image frame segments through the data processing unit; S3.3 Input the human body training action information obtained by the feature extraction unit and the standard action model established for the corresponding training action project into the analysis unit to obtain the action analysis result of the effective training action image frame segment; input the body physiological parameter information in the image frame segment obtained by the data processing unit and the healthy exercise physiological model of the corresponding training action project into the analysis unit to obtain the physiological analysis result of the body physiological parameter segment; S3.4 Combine the above-mentioned action analysis results and physiological analysis results to obtain comprehensive analysis results of action techniques and physical fitness in sports training.
9. The sports training analysis method based on artificial intelligence according to claim 8, characterized in that: In step S3.3, the analysis unit analyzes the sports training in combination with the action techniques and physical fitness in the sports training, which specifically includes the following steps: S3.3.1 inputting the human body training movement information and the standard movement model established for the corresponding training movement items into the analysis unit to obtain the movement analysis results of the above body physiological parameter segments; S3.3.2 inputting the body physiological parameter information in the image frame segment and the healthy sports physiological model of the corresponding training action item into the analysis unit to obtain the physiological analysis result in the above-mentioned effective action image frame segment; S3.3.3 Combine the results of movement analysis and physiological analysis to obtain comprehensive analysis results of movement techniques and physical fitness in sports training.
