Intelligent sensing monitoring method suitable for onboard sports equipment

By installing IMU sensors and motion capture suits on the onboard motion equipment, combined with deep learning algorithms, efficient and accurate monitoring of motion postures is achieved, and the problems of expensive equipment and insufficient accuracy in the existing technology are solved, and are suitable for a variety of sports Scene.

CN119015685APending Publication Date: 2024-11-26MATWORTH (HANGZHOU) DIGITAL TECH CO LTD
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202410091235.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-01-23
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

In prior art, in sports or extreme sports, wearable sensor solution equipment is expensive and has a great impact on the motion process. Machine vision solution equipment and algorithm accuracy are insufficient, and it cannot be applied to motion posture monitoring in a wide range of scenarios.

Method used

Two IMU sensor units are installed on the onboard motion equipment, and feature data is collected and motion posture monitoring is performed through deep learning algorithm pre-trained models. Combined with the motion capture suit and camera to collect joint data, achieving non-influence and simple motion posture perception.

Benefits of technology

It realizes efficient and accurate monitoring of sports postures, is suitable for a variety of sports systems and scenarios, simplifies equipment installation and data acquisition, and is suitable for large-scale data acquisition and method iteration.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119015685A_ABST
    Figure CN119015685A_ABST
Patent Text Reader

Abstract

The invention discloses an intelligent sensing monitoring method suitable for onboard sports equipment, and belongs to the technical field of sensing monitoring, and the intelligent sensing monitoring method suitable for onboard sports equipment comprises the steps of sampling, data processing, data application and artificial intelligence model pre-training, and monitoring and sensing result integration and application. The spatial-temporal data of the motion process and the motion posture of the motion load structure body are dynamically monitored and recorded, the motion posture and state of a sporter (or a load) are deduced by adopting a deep learning algorithm based on the spatial-temporal data, and the method can be effectively expanded, is simple, efficient and easy to implement, is suitable for various motion systems and application scenes, and has wide application prospects. The method has the advantages of no influence on sporters, simplicity and convenience in equipment installation and data acquisition, convenience in large-scale data acquisition and continuous iteration of the method.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of perception monitoring technology, and in particular to an intelligent perception monitoring method suitable for board-mounted sports equipment. Background Art

[0002] According to research, in most current sports or extreme sports, in order to obtain the athlete's (human body) movement posture, the mainstream approach is to set up close-fitting sensors on the athlete's body parts or entire body (such as wearing a full-body motion capture suit, wearing local sensors, etc.), and collect movement posture data and restore the posture through sensor data; or to shoot a video or take photos of the athlete's movement process, and extract and restore the athlete's movement posture through a vision-based machine learning algorithm.

[0003] The above-mentioned technical solutions, such as the wearable sensor solution, are expensive, have complex initial configuration, and have a significant impact on the athlete's movement process. Although this solution can obtain relatively accurate data, it cannot be used in a wider range of scenarios due to the above limitations. The machine vision solution also has the problem of high equipment and shooting costs (drone shooting or follow-up shooting of high-level athletes, etc.). At the same time, due to the inherent limitations of the machine vision algorithm itself, the extraction of motion posture features also has the problem of insufficient accuracy.

[0004] Therefore, it is necessary to provide an intelligent sensing monitoring method suitable for board-mounted sports equipment to solve the above problems. Summary of the invention

[0005] In view of the deficiencies in the prior art, an embodiment of the present invention aims to provide an intelligent sensing monitoring method applicable to board-mounted sports equipment to solve the problems in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] An intelligent sensing monitoring method applicable to board-mounted sports equipment comprises the following steps:

[0008] Step S1, sampling, wherein the sampling includes installing two IMU sensor units on the board-mounted sports equipment, and continuously collecting feature data at different positions of the board-mounted sports equipment when the IMU sensor is turned on, wherein the static collection of the IMU sensor unit is used to establish a database, and the dynamic collection of the IMU sensor unit is used to update the database;

[0009] Step S2, data processing, wherein the data processing includes cleaning the original spatiotemporal data generated during the movement of the board-mounted sports equipment, eliminating invalid data, correcting deviations, arranging and aligning the time axis, and outputting a data format that meets the processing requirements of the pre-trained artificial intelligence model;

[0010] Step S3, data application and artificial intelligence model pre-training, the data application and artificial intelligence model pre-training include using the motion data as input data for the artificial intelligence model pre-training of the deep learning framework, and performing artificial intelligence model pre-training for human body posture deduction of board-mounted sports through the deep learning algorithm architecture, so as to form monitoring and perception of various types of motion data generated during the movement of board-mounted sports equipment;

[0011] Step S4, integration and application of monitoring and perception results, the integration and application of monitoring and perception results include combining the motion data, the corresponding athlete motion label data and the artificial intelligence model pre-trained by the deep learning algorithm, obtaining the monitoring data of the board-mounted sports equipment, and deducing and simulating the corresponding athlete motion posture, that is, completing the perception of the motion posture of the board-mounted sports athlete through the monitoring of the sports equipment.

[0012] As a further solution of the present invention, the collected data is transmitted through the information transceiver module and the wireless network module, and needs to be stored before data processing. The data storage module is used to store the information data of the athlete.

[0013] As a further solution of the present invention, the distance between the two IMU sensor units is 8 cm, and the two IMU sensor units are parallel to the length direction of the board-mounted sports equipment.

[0014] As a further solution of the present invention, the alignment timeline includes time alignment of two IMU sensor units, and in the training data collection phase, time alignment of IMU sensor unit data with athlete motion capture suit data and athlete gliding posture video data.

[0015] As a further solution of the present invention, the motion data includes the motion trajectory, average speed, maximum speed, and directional linear acceleration and angular acceleration of the board-mounted sports equipment.

[0016] As a further solution of the present invention, the sampling also includes a motion capture suit worn by the athlete and a camera arranged on a board-mounted sports equipment, and / or a fixed-distance aerial camera of a drone, a manually followed gliding camera, etc. The motion capture suit collects the athlete's joint data and motion morphology data by cooperating with the camera.

[0017] As a further solution of the present invention, the data processing module also includes an information transceiver module and a wireless network module. The information transceiver module is used to receive data collected by the micro sensor and send it to the cloud server through the wireless network module. The wireless network module is remotely connected to the cloud server.

[0018] In summary, compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0019] The present invention can perform static collection of motion load structures and establish a database through sampling, dynamically collect and update the database, collect joint data and motion morphology data through a motion capture suit and a camera, input the collected data into the AI ​​pre-training model through the AI ​​pre-training model for convolution calculation, quantize weight data of different types of data, and update the AI ​​pre-training model through continuous data collection. The data collection process can realize dynamic monitoring and recording of the spatiotemporal data of the motion process and motion posture of the motion load structure. The method is simple, efficient, and easy to implement, and is suitable for a variety of motion systems and application scenarios. It has the advantages of no impact on athletes, simple and convenient equipment installation and data collection, and convenient large-scale data collection and continuous iteration of methods.

[0020] In order to more clearly illustrate the structural features and effects of the present invention, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 The data collection flow chart of the embodiment of the invention.

[0022] Figure 2 Schematic diagram of the acquisition process of an embodiment of the invention.

[0023] Figure numerals: 1-ski board, 2-IMU sensor unit, 3-power supply and data processing module, 4-foot strap, 5-motion capture suit, 6-camera. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0025] The specific implementation of the present invention is described in detail below in conjunction with specific embodiments.

[0026] In one embodiment of the present invention, the intelligent sensing monitoring method applicable to board-mounted sports equipment is characterized by comprising the following steps:

[0027] Step S1, sampling;

[0028] Among them, for different board-mounted sports equipment (such as snowboard 1), two IMU sensor units 2 are used and installed on the equipment, the unit spacing is fixed at 8cm, the connection line of the two IMU sensor units 2 is parallel to the length direction of the equipment, and sampling starts. After the two IMU sensor units 2 are powered on, they self-check and calibrate and synchronize the time axis (GPS time or local time service); the IMU sensor unit 2 used in the present invention adopts 2G acceleration, 250 range, 100khz frequency, and 30hz sampling frequency; continuous sampling is performed in the power-on state; the present invention is installed on the board-mounted sports equipment and can be regarded as a part of the sports equipment, and performs rigid body motion and a certain degree of deformation together with the sports equipment; by collecting feature data at different positions of the sports equipment, the rigid body motion and deformation of the board-mounted sports equipment can be accurately reflected, wherein the static collection of the IMU sensor unit 2 is used to establish a database, and the dynamic collection of the IMU sensor unit 2 is used to update the database, and the sampling also includes a motion capture suit 5 worn on the athlete and a camera 6 set on the board-mounted sports equipment, and the motion capture suit 5 collects the athlete's joint data and motion morphology data in cooperation with the camera 6.

[0029] Step S2, data processing;

[0030] Among them, for the original spatiotemporal data generated during the movement of the motion load structure, data cleaning is performed, invalid data is eliminated, deviations are corrected, and the time axis is sorted and aligned (including the time alignment of the two IMU sensors, and the time alignment of the IMU data and the athlete's motion capture suit 5 data during the training data collection phase), and then the data file in the standard RAW / Text format is output, and the cleaned original spatiotemporal data is classified, sorted, and converted (such as converted to Euler angle data through differential calculation) and used as input data for deep learning and artificial intelligence model training, and saved as a backup;

[0031] Step S3: data application and artificial intelligence model pre-training;

[0032] Among them, the spatiotemporal data of onboard sports equipment that are collected and cleaned are further processed by software and conventional algorithms to generate multi-dimensional motion data feature representations, including but not limited to the trajectory, average speed, maximum speed, and change of direction acceleration of onboard sports; the further processed data are used as input data for pre-training of artificial intelligence models of deep learning frameworks (corresponding labels of synchronous motion data collected by the athlete's motion capture suit 5), and pre-training of artificial intelligence models for human posture deduction of onboard sports is carried out through deep learning algorithm architectures such as RNN, CNN, LSTM, and Transformer.

[0033] Step S4: Integration and application of monitoring and perception results;

[0034] Among them, through the above process, on the one hand, the monitoring and perception of various types of sports data generated by board-mounted sports equipment during exercise are formed; on the other hand, through this data and the corresponding athlete motion label data, the artificial intelligence model obtained by pre-training with the deep learning algorithm can deduce and simulate the corresponding athlete motion posture from the monitoring data of the board-mounted sports equipment, that is, through the monitoring of the sports equipment, the perception of the motion posture of the board-mounted sports athletes is completed.

[0035] The present invention can perform static collection of motion load structures and establish a database through sampling, dynamically collect and update the database, collect joint data and motion morphology data through the motion capture suit 5 and the camera 6, and realize dynamic monitoring and recording of the spatiotemporal data of the motion process and motion posture of the motion load structure. It is simple, efficient and easy to operate, suitable for a variety of motion systems and application scenarios, has no impact on athletes, and is easy and convenient to install equipment and collect data, and is convenient for large-scale data collection and continuous iteration of methods.

[0036] In one embodiment of the present invention, see Figure 1 The intelligent sensing monitoring method suitable for board-mounted sports equipment takes snowboarding as an example. At this time, the sports load structure is a snowboard 1, and the athlete's feet are fixed on the snowboard 1 through a fastener 4. The movement process mainly slides from a higher position to a lower position under the action of gravity, and the athlete's feet and whole body limbs exert force to perform technical movements, thereby controlling the snowboard 1 and generating various sliding postures.

[0037] Taking this as an example, for this sport (snowboarding), the motion posture of the athlete (skier) is deduced with the motion load structure (ski 1). The specific plan is: equipment installation, data collection and export, and the hardware module for monitoring the spatiotemporal data of the motion of the motion load structure is mainly composed of two IMU sensor units 2. The connection line of the two sensor units 2 is parallel to the length direction of the ski 1 and is attached to the surface between the two "fixtures" of the ski 1. The inside of the module is used to record spatial position and time information according to the GPS function module, and a power supply and data processing module 3 is connected between the two sensor units 2.

[0038] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent sensing monitoring method applicable to board-mounted sports equipment, characterized in that: The following steps are involved: Step S1, sampling, wherein the sampling includes using a data processing module and two IMU sensor units, the two IMU sensor units are installed on the board-mounted sports equipment, and the feature data of different positions of the board-mounted sports equipment are continuously collected when the IMU sensor is turned on, wherein the static collection of the IMU sensor unit is used to establish a database, and the dynamic collection of the IMU sensor unit is used to update the database; Step S2, data processing, wherein the data processing includes cleaning the original spatiotemporal data generated during the movement of the board-mounted sports equipment, eliminating invalid data, correcting deviations, arranging and aligning the time axis, and outputting a data format that meets the processing requirements of the pre-trained artificial intelligence model; Step S3, data application and artificial intelligence model pre-training, the data application and artificial intelligence model pre-training include using the motion data as input data for the artificial intelligence model pre-training of the deep learning framework, and performing artificial intelligence model pre-training for human body posture deduction of board-mounted sports through the deep learning algorithm architecture, so as to form monitoring and perception of various types of motion data generated during the movement of board-mounted sports equipment; Step S4, integration and application of monitoring and perception results, the integration and application of monitoring and perception results include combining the motion data, the corresponding athlete motion label data and the artificial intelligence model pre-trained by the deep learning algorithm, obtaining the monitoring data of the board-mounted sports equipment, and deducing and simulating the corresponding athlete motion posture, that is, completing the perception of the motion posture of the board-mounted sports athlete through the monitoring of the sports equipment.

2. The intelligent sensing monitoring method for board-mounted sports equipment according to claim 1, characterized in that: The distance between the two IMU sensor units is 8 cm, and the two IMU sensor units are parallel to the length direction of the board-mounted sports equipment.

3. The intelligent sensing monitoring method for board-mounted sports equipment according to claim 1, characterized in that: The collection also includes motion capture suit data collection and video capture collection of the athlete's gliding posture. The motion capture suit data collection establishes a database through static collection.

4. The intelligent sensing monitoring method for board-mounted sports equipment according to claim 1, characterized in that: The alignment timeline includes the time alignment of two IMU sensor units, and in the training data collection phase, the time alignment of the IMU sensor unit data with the athlete's motion capture suit data and the athlete's gliding posture video data.

5. The intelligent sensing monitoring method for board-mounted sports equipment according to claim 1, characterized in that: The motion data includes the motion trajectory, average speed, maximum speed, and direction-changing linear acceleration and angular acceleration of the board-mounted sports equipment.

6. The intelligent sensing monitoring method for board-mounted sports equipment according to claim 1, characterized in that: The sampling also includes a motion capture suit worn by the athlete and a camera installed on a board-mounted sports equipment, and / or a fixed-distance aerial camera of a drone, and a manually followed gliding camera. The motion capture suit collects the athlete's joint data and motion morphology data by cooperating with the camera.

7. The intelligent sensing monitoring method for board-mounted sports equipment according to claim 1, characterized in that: The data processing module also includes an information transceiver module and a wireless network module. The information transceiver module is used to receive data collected by the micro sensor and send it to the cloud server through the wireless network module. The wireless network module is remotely connected to the cloud server.

Citation Information

Patent Citations

  • Snow motion information monitoring method and related device

    CN115346273A

  • High-speed motion capture and recognition method and system for motion scene

    CN116030533A

  • AI-based swimming stroke foul action judgment method and system

    CN117095472A

  • Methods and system for training and improving machine learning models

    US20210319337A1