CN114602155B - Swimming information statistical method, computer-readable storage medium and electronic device - Google Patents

Abstract

The application relates to the technical field of wearable equipment, and discloses a swimming information statistical method, a computer readable storage medium and electronic equipment, which are applied to a system comprising first electronic equipment and second electronic equipment, wherein the first electronic equipment is worn on a first part of a user, the second electronic equipment is worn on a second part of the user, and the difference between the motion tracks of the first part under different swimming postures is smaller than the difference between the motion tracks of the second part. The method comprises the following steps: the method comprises the steps that first electronic equipment obtains first motion data of a user; the first electronic equipment acquires second motion data acquired by the second electronic equipment in the swimming process of the user; the first electronic equipment integrates the first motion data and the second motion data to obtain integrated motion data; the first electronic equipment counts swimming information of the user based on the integrated motion data. According to the technical scheme, the swimming information of the user counted by the first electronic equipment is more accurate, and the user experience is favorably improved.

Description

Swimming information statistical method, computer-readable storage medium and electronic device

technical field

The present application relates to the technical field of wearable devices, in particular to a swimming information statistics method, a computer-readable storage medium and electronic devices.

Background technique

With the improvement of material and cultural level, people's awareness of sports and health has gradually improved. As a sport that can improve cardiopulmonary function, improve body coordination, and strengthen respiratory muscle strength, more and more people choose swimming. , to enhance physical fitness. At the same time, people also hope to accurately monitor and record their own swimming-related data, so as to have a more comprehensive understanding of their own sports conditions.

People can monitor swimming data through smart watches. Smart watches are cheap and easy to wear, and are widely used by the public. However, due to their limited sensors and differences in users' wearing habits, it is easy to affect the accuracy of the monitored swimming data.

Contents of the invention

In view of this, the embodiments of the present application provide a swimming information statistics method, a computer-readable storage medium, and an electronic device.

In the technical solution of the present application, the first electronic device worn on the user's first part (for example, a sports watch worn on the user's wrist), the user's first motion data collected by the first electronic device, and the user's first motion data collected from the user's second part The user's second exercise data acquired by the second electronic device (for example, a sports earphone worn on the user's ear) is integrated to perform statistics on the user's swimming information. Since the difference between the motion trajectories of the first parts is smaller than the difference between the motion trajectories of the second parts under different swimming styles, for example, for different swimming styles, even if the user's hand movements are consistent, but the user's head Actions vary greatly. Therefore, after integrating the user's exercise data collected by the first electronic device worn on the user's first part and the user's exercise data collected by the second electronic device worn on the user's second part, the user's swimming information is calculated More accurate, for example, the user's swimming style, number of strokes, number of strokes, frequency of strokes, etc. can be more accurately distinguished.

In the first aspect, the embodiment of the present application provides a swimming information statistics method, which is applied to a system including a first electronic device and a second electronic device, wherein the first electronic device is worn on the first part of the user, and the second The second electronic device is worn on the second part of the user, and the difference between the motion trajectories of the first part under different swimming styles is smaller than the difference between the motion trajectories of the second parts. The method includes:

The first electronic device acquires first motion data of the user during swimming;

The first electronic device acquires second motion data collected by the second electronic device during the user's swimming;

The first electronic device integrates the first exercise data and the second exercise data to obtain integrated exercise data;

The first electronic device calculates the user's swimming information based on the integrated exercise data.

In a possible implementation of the first aspect above, the first electronic device integrates the first exercise data and the second exercise data to obtain integrated exercise data, including:

The first electronic device frames the first motion data and the second motion data to obtain integrated motion data.

In a possible implementation of the first aspect above, the first electronic device frames the first motion data and the second motion data to obtain integrated motion data, including:

The first electronic device performs down-sampling processing on the first motion data according to the frame rate of the second motion data to obtain the down-sampled first motion data, wherein the frame rate of the down-sampled first motion data and the second motion data The frame rate of the data is the same.

In a possible implementation of the first aspect above, the first electronic device frames the first motion data and the second motion data to obtain integrated motion data, including:

The first electronic device performs interpolation processing on the second motion data according to the frame rate of the first motion data to obtain the second motion data after the interpolation processing, wherein the frame rate of the second motion data after the interpolation processing and the first motion data The frame rate is the same.

In a possible implementation of the first aspect above, the user's swimming information includes the user's swimming style, and the first electronic device calculates the user's swimming information based on the integrated exercise data, including:

The first electronic device extracts features from the integrated motion data to obtain time-domain features of the integrated motion data;

The first electronic device inputs the time-domain features of the integrated motion data into a preset stroke recognition model to obtain a stroke recognition result;

The first electronic device counts the swimming strokes of the user based on the recognition result of the swimming strokes.

In a possible implementation of the first aspect above, the preset stroke recognition model is a binary classification model.

In a possible implementation of the first aspect above, the user's swimming information also includes the user's stroke action, and the first electronic device calculates the user's swimming information based on the integrated exercise data, including:

According to the statistics of the user's swimming strokes, the first electronic device determines the first target data for counting the user's stroke movements in the integrated motion data, wherein the first target data includes the first target data collected by the first electronic device. Target sub-data and second target sub-data collected by the second electronic device;

The first electronic device determines a plurality of coarse sieving strokes according to the first target sub-data;

According to the second target sub-data, the first electronic device counts the stroke actions of the user from the plurality of roughly screened stroke actions.

In a possible implementation of the first aspect above, the first electronic device includes a gyroscope and an accelerometer, and the first target sub-data includes data collected by the gyroscope of the first electronic device or data collected by the accelerometer of the first electronic device. data.

In a possible implementation of the foregoing first aspect, the first electronic device includes an accelerometer, and the second target sub-data includes data collected by the accelerometer of the second electronic device.

In a possible implementation of the above first aspect, the user's swimming information includes the user's turning action, and the first electronic device calculates the user's swimming information based on the integrated exercise data, including:

The first electronic device determines the user's initial water stroke;

The first electronic device determines the user's initial swimming direction according to the second motion data corresponding to the first number of consecutive strokes including the initial stroke;

The first electronic device determines the swimming direction corresponding to each stroke after the first continuous number of strokes according to the second motion data corresponding to each stroke after the first number of consecutive strokes;

When the first electronic device judges that the user has turned around based on the difference between the swimming directions corresponding to every two adjacent strokes after the first consecutive number of strokes, the first electronic device determines that the user has turned around. target stroke;

The first electronic device determines the target swimming direction corresponding to the second number of consecutive strokes after the target stroke according to the second motion data corresponding to the second number of consecutive strokes after the target stroke;

When the first electronic device determines that the difference between the target swimming direction and the initial swimming direction reaches a threshold, the first electronic device confirms that the user turns around during the target stroke, and counts the user's turning movements.

In a possible implementation of the above first aspect, the above method further includes:

The first electronic device updates the initial swimming direction to a target swimming direction.

In a possible implementation of the first aspect above, the second electronic device includes a gyroscope, an accelerometer, and a magnetometer, and the second movement data includes data collected by the gyroscope of the second electronic device, data collected by the accelerometer of the second electronic device The collected data, the data collected by the magnetometer of the second electronic device.

In a possible implementation of the first aspect above, the first electronic device is a sports watch.

In a possible implementation of the foregoing first aspect, the second electronic device is a sports earphone.

In the second aspect, the implementation of the present application provides a computer-readable storage medium, and instructions are stored on the computer-readable storage medium, and when the instructions are executed on the electronic equipment, the electronic equipment executes any of the above-mentioned first aspect and the first aspect. A statistical method for swimming information in a possible implementation.

In a third aspect, the implementation of this application provides an electronic device, including:

memory for storing instructions to be executed by one or more processors of the electronic device, and

A processor, when the instructions are executed by one or more processors, the processor is configured to execute the above first aspect and the swimming information statistics method in any possible implementation of the first aspect.

Description of drawings

Fig. 1 shows an application scenario in which a user wears a watch 100 and earphones 200 for swimming according to some embodiments of the present application;

FIG. 2A shows a block diagram of the hardware structure of the watch 100 shown in FIG. 1 according to some embodiments of the present application;

FIG. 2B shows a hardware structural block diagram of the earphone 200 shown in FIG. 1 according to some embodiments of the present application;

Fig. 3 shows a flow chart of swimming stroke recognition of the watch 100 according to some embodiments of the present application;

Fig. 4A shows another flow chart of swimming stroke recognition of the watch 100 according to some embodiments of the present application;

Fig. 4B shows a schematic diagram of the distribution of air pressure counter values according to some embodiments of the present application;

Fig. 4C shows a schematic structural diagram of a swimming stroke recognition model according to some embodiments of the present application;

Fig. 4D shows a schematic diagram of several standard swimming strokes according to some embodiments of the present application;

Fig. 5A shows a flow chart of a watch 100 performing water stroke recognition according to some embodiments of the present application;

Fig. 5B shows a schematic flowchart of a method for realizing the coarse screening action in Fig. 5A according to some embodiments of the present application;

FIG. 5C shows a time-domain distribution diagram of a signal at the watch 100 terminal and a signal at the earphone 200 according to some embodiments of the present application;

FIG. 6A shows a schematic diagram of a turn-around recognition process of a watch 100 according to some embodiments of the present application;

Fig. 6B shows a schematic diagram of the time-domain distribution of the yaw angle value of an earphone 200 according to some embodiments of the present application;

Fig. 7 shows a schematic user interface of swimming information statistics of the watch 100 according to some embodiments of the present application;

Fig. 8 shows a software logic block diagram of the watch 100 shown in Fig. 1 according to some embodiments of the present application.

Detailed ways

Illustrative embodiments of the present application include, but are not limited to, a swimming information statistics method, a computer-readable storage medium, and an electronic device. Various aspects of the illustrative embodiments are described below using terms commonly employed by those skilled in the art.

Fig. 1 shows an exemplary application scenario of the embodiment of the present application. In FIG. 1 , a user wears a watch 100 on his arm during swimming. The watch 100 can monitor the user's swimming data, for example, monitor the user's swimming style, number of strokes, etc., so that the user can check it later.

The watch 100 usually measures the motion data of the user's arm through built-in sensors (eg, acceleration sensor, gyroscope), and determines swimming data such as swimming strokes and stroke times based on the measured motion data. However, due to the limitation of the measurement capability of the watch 100, the accuracy of the measured swimming data is limited.

For example, the swimming style of the user may be breaststroke, butterfly, freestyle, backstroke, etc. when swimming. When the user is swimming freestyle or backstroke, there is little difference in the movement of the user's arms. At this time, it may be difficult to accurately distinguish between freestyle and backstroke based on the motion data measured by the watch 100 , resulting in low accuracy of the swimming data.

To this end, the embodiment of the present application provides a swimming data monitoring system based on the combination of multiple devices, the system includes a first wearable device that is easy to put on and take off, such as a sports watch (hereinafter referred to as a watch), and a second wearable device, For example, sports earphones (hereinafter referred to as earphones). The user needs to wear these two wearable devices during swimming. For example, in the application scenario shown in FIG. 1 , the watch 100 is worn on the user's wrist, and the earphone 200 is worn on the user's ear. Both wearable devices are equipped with accelerometer (acceleration transducer, ACC) meter, magnetometer (magnetometer, MAG), gyroscope (gyroscope, GYRO) and other sensors, and these two wearable devices can realize real-time data communication through Bluetooth . During the swimming process of the user, the two wearable devices obtain sensor data in real time, and one of the wearable devices transmits the real-time collected sensor data to the other wearable device with stronger processing capability in real time. The wearable device jointly analyzes the data collected by its own sensor and the sensor data received from another transmittable device to determine swimming information such as the user's swimming posture, number of strokes, number of strokes, stroke frequency, etc.

Among them, the swimming posture can include breaststroke, butterfly stroke, freestyle, backstroke four standard strokes and medley, etc.; the number of strokes can be the number of movements of swinging arms during swimming; the number of swimming times can be the number of times of turning back in the swimming lane during swimming ; The stroke frequency may be the frequency of the arm swing during swimming. It should be understood that the watch and earphones involved in the embodiments of the present application may have a waterproof function.

In this embodiment of the present application, the user's motion data collected by a wearable device worn on the user's wrist and the user's motion data collected by a wearable device worn on the user's head (for example, worn on the user's ear) are jointly analyzed. For different swimming styles, even if the user's hand movements are consistent, the user's head movements are quite different. Therefore, using the user's motion data collected by the wearable device worn on the user's wrist and the user's motion data collected by the wearable device worn on the user's head to conduct a joint analysis of swimming information, the user's swimming information obtained is more accurate. For example, the user's swimming style, number of strokes, number of strokes, frequency of strokes, etc. of the user can be more accurately distinguished.

In addition, in some embodiments, the swimming information monitoring equipment is professional equipment, which is expensive and difficult to wear. Compared with this embodiment, the above-mentioned first/second wearable devices in this application may be wearable devices that are light, easy to wear, and low in cost. The technical solution of the present application can not only provide the user with a light and easy-to-wear experience, but also provide accurate swimming information during the swimming process of the user, which helps to improve the user experience.

It should be understood that the wearable devices to which the technical solution of the present application is applicable include but are not limited to sports watches and sports earphones, wherein the first wearable device may also be a smart bracelet, a smart ring, and other devices. The second wearable device may also be a head-mounted device, such as an augmented reality device.

It should be noted that the application scenario shown in FIG. 1 is only to illustrate an exemplary application scenario of the technical solution of the present application, which only includes a watch 100 worn on the user's wrist and a pair of earphones worn on the user's ear 200. In some embodiments of the present application, the user can also wear more than two wearable devices. For example, in addition to the watch 100 worn on the user's wrist and a pair of earphones 200 worn on the user's ears, it can also include Wrist and other parts of the foot ring, etc.

In order to facilitate the understanding of the technical solution of the present application, the technical solution of the present application will be described in detail below by taking the application scenario shown in FIG. 1 where the user wears the watch 100 and the earphone 200 for swimming as an example.

First, the hardware structures of the watch 100 and the earphone 200 shown in FIG. 1 are introduced respectively with reference to FIG. 2A and FIG. 2B . Figure 2A is a hardware structure diagram of a watch 100 shown in the embodiment of the present application. The hardware structure of the watch 100 includes: accelerometer 101, gyroscope 102, magnetometer 103, barometer 104, processor 105, memory 106, communication Module 107, power supply module 108.

Although not shown, the watch 100 may also include a display screen, an ambient temperature sensor, an antenna, a wireless-fidelity (Wi-Fi) module, a near field communication technology (near filed communication, NFC) module, a global positioning System (global positioning system, GPS) module, speaker, etc.

The accelerometer 101 can detect the acceleration of the watch 100 in various directions (generally three axes). During the user's swimming process, in different swimming styles, with the swing of the user's arm, the signals of the data of the three coordinate axes of the accelerometer 101 behave differently. Therefore, the signals of the three coordinate axes of the accelerometer 101 can be analyzed. Distinguish the user's swimming style, stroke actions corresponding to different swimming styles, etc.

The gyroscope 102 can detect angular velocity or angular acceleration of the watch 100 in various directions (for example, three-axis directions). During the user's swimming process, in different swimming styles, with the swing of the user's arm, the signals of the data of the three coordinate axes of the gyroscope 102 behave differently. Therefore, it is also necessary to analyze the signals of the three coordinate axes of the gyroscope 102 It can distinguish the user's swimming style, stroke actions corresponding to different swimming styles, etc.

The magnetometer 103 can detect the direction of the earth's magnetic field. The movement direction of the watch 100 can be determined based on the direction of the earth's magnetic field. During the swimming process of the user, since the user will turn around when the user completes each swimming and starts the next swimming, so that the movement direction of the watch 100 worn on the user's wrist will change greatly, therefore, the magnetometer 103 is used to The collected movement direction data of the watch 100 is beneficial to identify the user's turning action.

The barometer 104 can be used to detect the pressure in the environment where the watch 100 is located. Since the pressure values collected by the barometer 104 are different when the watch 100 is in water and in air, analyzing the data collected by the barometer 104 is convenient for the user to detect water entry and exit.

The processor 105 is used for system scheduling, and can be used to control the aforementioned sensors such as the accelerometer 101 , the gyroscope 102 , the magnetometer 103 , and the barometer 104 , and to perform operations on the data collected by these sensors. For example, in some embodiments of the present application, the processor 105 is used to perform framing and feature extraction based on the data collected by the accelerometer 101, the gyroscope 102, and the sensor data obtained from the earphone 200, and then perform swimming strokes based on the extracted features. identify. As another example, in some embodiments of the present application, the processor 105 is used to identify the stroke action according to the data collected by the accelerometer 101 and/or the gyroscope 102 of the watch 100 and the accelerometer data collected by the earphone 200 . As another example, in some embodiments of the present application, the processor 105 is configured to perform turn recognition according to accelerometer data, gyroscope data, and magnetometer data acquired from the earphone 200 .

The memory 106 is used to store software programs and data, and the processor 105 executes various functional applications and data processing of the watch 100 by running the software programs and data stored in the memory. In some embodiments of the present application, the memory 2 may store data collected by the accelerometer 101 , gyroscope 102 , magnetometer 103 , barometer 104 and other sensors, as well as some sensor data obtained from the earphone 200 .

The communication module 107, the watch 100 can establish a connection with other electronic devices through the communication module 107. For example, in the embodiment of the present application, the watch 100 can establish a Bluetooth connection with the user's earphone 200, and receive some data collected by the earphone 200 from the earphone 200. sensor data.

The power supply module 108 can be used to supply power to various components of the watch 100 mentioned above.

It can be understood that the structure shown in the embodiment of the present invention does not constitute a specific limitation on the watch 100 . In other embodiments of the present application, the watch 100 may include more or fewer components than shown in the illustration, or combine some components, or separate some components, or arrange different components. The illustrated components can be realized in hardware, software or a combination of software and hardware.

FIG. 2B is a hardware structure diagram of an earphone 200 provided by an embodiment of the present application. The hardware structure of the earphone 200 includes: an accelerometer 201, a gyroscope 202, a magnetometer 203, a memory 204, a processor 205, a communication module 206, Power module 207, audio module 210. Wherein, the audio module 210 includes a speaker 211 and a microphone 212 .

The accelerometer 201 can detect the acceleration of the earphone 200 in various directions (generally three axes).

The gyroscope 202 can be used to determine the motion posture of the earphone 200 .

The magnetometer 203 can be used to detect the movement direction of the hand headset 200 .

The memory 204 is used to store software programs and data, and the processor 205 implements various functional applications and data processing of the earphone 200 by running the software programs and data stored in the memory. In some embodiments of the present application, the memory 204 may store data collected by the accelerometer 201 , the gyroscope 202 , the magnetometer 203 , and other sensors.

The processor 205 is used for system scheduling, and can be used to control the accelerometer 201, gyroscope 202, magnetometer 203 and other sensors to collect data.

Communication module 206, the headset 200 can establish a connection with other electronic devices through the communication module 206, for example, in the embodiment of the application, the headset 200 can establish a Bluetooth connection with the user's watch 100, and send some sensor data collected by the headset 200 Give the watch 100.

The power module 207 can be used to supply power to various components of the earphone 200 .

The speaker 211 is also called a "horn", and is used to convert audio electrical signals into sound signals. The headset 200 can listen to music through the speaker 211, or listen to hands-free calls. When receiving a phone call or voice information through the earphone 200, the speaker 211 can be placed close to the human ear to receive the voice.

The microphone 212 is also called "microphone" or "microphone", and is used to convert sound signals into electrical signals. When making a call or sending a voice message, the user can put his mouth close to the call microphone 212 to make a sound, and input the sound signal to the call microphone 212 . The headset 200 can be provided with at least one call microphone 212 . In other embodiments, the earphone 212 may be provided with two call microphones, which may also implement a noise reduction function in addition to collecting sound signals.

It can be understood that, the structure illustrated in the embodiment of the present invention does not constitute a specific limitation on the earphone 200 . In other embodiments of the present application, the earphone 200 may include more or fewer components than shown in the figure, or combine some components, or separate some components, or arrange different components. The illustrated components can be realized in hardware, software or a combination of software and hardware.

The functions such as swimming stroke recognition, stroke recognition, and turn recognition that can be realized by the technical solution of the present application will be further introduced in detail below in conjunction with the hardware structure block diagrams of the above-mentioned watch 100 and earphone 200 .

First, an exemplary implementation of stroke recognition is introduced. FIG. 3 is a flowchart of swimming stroke recognition performed by a watch 100 provided by some embodiments of the present application. The execution subject of each step in the flowchart shown in FIG. 3 may be the watch 100 . With reference to Fig. 3, a kind of stroke recognition method provided by the present application includes the following steps:

S301: Determine to enter the water.

For example, according to the air pressure data collected from the barometer 104 of the watch 100, it is determined whether the user has entered the water. When the air pressure value satisfies a certain threshold condition, it can be determined that the user enters the water. The details will be described below.

S302: Collect motion data of the user.

For example, the watch 100 uses the accelerometer 101 to collect acceleration data of the watch 100 in various directions (generally three axes). The gyroscope 102 is used to collect angular velocity or angular acceleration of the watch 100 in various directions (for example, three-axis directions). The direction of the earth's magnetic field can be collected by using the magnetometer 103 .

S303: Obtain the motion data of the user collected by the earphone 200 .

For example, watch 100 sends a data acquisition request to earphone 200 , and earphone 200 responds to the request by sending the user's motion data collected by earphone 200 through accelerometer 201 , gyroscope 202 , and magnetometer 203 to watch 100 .

S304: Integrate the motion data collected by the sensor of the watch 100 and the motion data obtained from the earphone 200 to obtain integrated motion data.

For example, the watch 100 obtains accelerometer data (hereinafter referred to as ACC data) and gyroscope data (hereinafter referred to as GYRO data) collected by the earphone 100, and the ACC data and GYRO data collected by the watch 100 for data grouping to obtain the combined data , how to group packages will be described in detail below.

S305: Determine the user's swimming style based on the integrated exercise data.

For example, the watch 100 performs time-domain feature extraction on the data obtained by combining the above data, and obtains time-domain features such as window mean value, variance, mean absolute deviation, and peak-to-valley features of the gyroscope and accelerometer data. Then these extracted time-domain features are input into the pre-trained stroke recognition model to determine the user's stroke. The specific implementation details will be described in detail below.

FIG. 3 introduces a method flow of a swimming stroke recognition method according to an embodiment of the present application. The following will use the specific example shown in FIG. 4A to exemplarily introduce the detailed flow of the wristwatch 100 performing stroke recognition. With reference to the flow chart of a wrist watch 100 swimming stroke recognition shown in FIG. 4A, the swimming stroke recognition method of the present application includes the following steps:

S401: Water in and out detection.

For example, water entry and exit detection is performed using air pressure data collected from the barometer 104 of the wristwatch 100 . When the watch 100 enters the water, the air pressure value collected by the barometer 104 will suddenly increase, and when the watch 100 comes out of the water, the air pressure value collected by the barometer 104 will suddenly decrease.

S402: Judging whether the watch 100 has entered the water, if it is judged that the watch 100 has entered the water, proceed to S406, and perform time-domain feature extraction on the ACC data and GYRO data collected by the watch 100 and the ACC data and GYRO data collected by the earphone 200 . If it is judged that the watch 100 has not entered the water, it means that the user has not started swimming yet, then enter S403 and wait for entering the water.

Exemplarily, as shown in FIG. 4B , the air pressure count value (that is, air pressure data) collected by the barometer 104 of the watch 100 suddenly increases and is greater than the threshold P1, and it is preliminarily judged that the watch 100 has entered the water, that is, it is preliminarily judged that the user has started swimming. The air pressure data collected over a period of time can then be further processed to verify whether the watch 100 is submerged in water again. For example, continuous collection of air pressure data of T1 duration, a total of 50 frames of air pressure data in one second, with 4 seconds as a window, a total of 200 frames of air pressure data in one window, and then averaging the 200 frames of air pressure data in each window value, get the average air pressure value in each window, and then calculate the difference between the average air pressure values in adjacent windows, if the average air pressure values in each window are greater than the threshold P1, and the difference between the average air pressure values in adjacent windows is greater than When the threshold is P2, it is determined that the watch 100 has entered the water, that is, the user is reconfirmed to start swimming; if the average air pressure values in each window are greater than the threshold value Threshold1, and the difference between the average air pressure values of adjacent windows is greater than the threshold value Threshold2. , it is determined that the watch 100 has not entered the water, that is, it is finally determined that the user has not started swimming.

In some embodiments, the barometer data returns to 1.5 times the air pressure before entering the water for 30 seconds, and lasts for T2, for example, 30 seconds, then it is determined that the watch 100 is out of the water.

S403: Wait for entering the water.

That is to say, if it is judged that the watch 100 has not entered the water, it waits for entering the water. During this process, the barometer 104 of the watch 100 can continue to collect air pressure data.

S404: Data framing.

That is to say, the swimming stroke recognition module 111 of the watch 100 determines that the watch 100 has entered the water through the data collected by the barometer 104, which means that the user starts swimming. The received data (respectively denoted as ACC data and GYRO data), and the data collected by the accelerometer 201 and the gyroscope 202 of the earphone 200 returned by the earphone 200 in real time, are framed.

For example, the watch 100 collects 100 frames of ACC data and GYRO data in 1 second, and the earphone 200 collects 50 frames of ACC data and GYRO data in 1 second. The ACC data is averaged every two frames to obtain a new 50-frame ACC data, and the 100-frame GYRO data collected by itself is averaged every two frames to obtain a new 50-frame GYRO data. Then align the new 50 frames of ACC data and the new 50 frames of GYRO data with the 50 frames of ACC data and GYRO data returned by the headset 200 based on time to obtain 50 frames of combined data. As another example, since the data sampling rate of ACC data and GYRO data collected by watch 100 is higher than that of earphone 100, the ACC data and GYRO data of watch 100 can be down-sampled. The watch 100 collects 50 frames of data every 1 second; or interpolates the ACC data and GYRO data of the earphone 200, so that the ACC data and GYRO data of the two devices, the watch 100 and the earphone 200, can be aligned in time and maintain the number of data consistent for subsequent processing.

S405: Sliding window processing is performed on the numbers after framing.

Exemplarily, a set time length (for example, 4 seconds) is used as a window, and a set step size (for example, 0.5 seconds) is used to perform sliding window processing on the data obtained through S404.

It should be understood that in a specific application, the above-mentioned window is a data window with a fixed duration customized according to needs. For example, for the combined data of 50 frames obtained above, data with a time span of 5 seconds, 6 seconds, etc. can also be used. As a window, the step size can also be customized according to needs, for example, the step size can also be 1 second. Because in the foregoing embodiment, the data duration obtained after framing is 1 second, and there are 50 frames of ACC data and GYRO data, a window is made in 4 seconds, and a window has 200 frames of ACC data and GYRO data. In this case, In the follow-up, a feature extraction is performed on the data in a window (that is, the data within a period of time), and the accuracy rate is higher than that of one frame of data.

S406: Time-domain feature extraction.

For example, time-domain features such as window mean, variance, mean absolute deviation, and peak-to-valley features of the gyroscope and accelerometer data are extracted from the combined data obtained above.

S407: Scoring the decision tree.

That is to say, after extracting some time-domain features of ACC data and GYRO data through S406, the swimming stroke recognition module 111 can use these time-domain features and adopt a scoring strategy of multiple decision trees (such as binary classification models) to obtain the user's The result of scoring strokes. In other words, a preset stroke recognition model is used to perform stroke recognition on the above-mentioned time-domain features. The preset stroke recognition model can be the decision tree model shown in FIG. 4C, because the standard stroke usually has the There are four types of breaststroke, backstroke, butterfly, and freestyle, so the decision tree model needs to include at least "breaststroke model", "butterfly model", "freestyle model", and "backstroke model", corresponding to other strokes other than these four standard strokes Swimming style can be scored using the "unknown swimming style model". That is to say, since there are 5 types of results to be recognized in total, 5 types of stroke recognition models need to be trained in total.

Each of these five models is used to identify one of breaststroke, backstroke, butterfly, freestyle, and unknown strokes. Each model will have multiple trees, and each tree will output the probability of a certain result. And the probability of not a certain result (for example, the probability of breaststroke is 0.7, the probability of not breaststroke is 0.3), and the probability of recognizing a certain stroke in each tree is added to get the total number of strokes recognized as this stroke. probability. Then select the stroke with the highest total probability among the five models as the recognition result.

S408: Obtain a recognition result.

For example, in some embodiments, some time-domain features corresponding to the ACC data and GYRO data within 4s can be input into the above decision tree model, and the stroke with the highest total probability is taken as the recognition result. For example, the total probabilities of breaststroke, backstroke, butterfly, freestyle, and unknown strokes obtained by the above models are respectively 50%, 60%, 65%, 85%, and 20%, and then the freestyle with the highest total probability can be determined as The user's stroke.

It can be understood that the above execution order of steps S401 to S408 is just an example, and in other embodiments, other execution orders may also be adopted, and some steps may also be split or combined, which is not limited here. For example, S404 and S405 can be executed before S401, and can also be executed simultaneously with S401. S401 and S402 can also be combined into one step.

An exemplary implementation of water stroke recognition will be introduced below. FIG. 5A is a flow chart of water stroke recognition provided by some embodiments of the present application. The execution subject of each step in the flow chart shown in FIG. 5A may be the watch 100 . Referring to Figure 5A, the process includes the following:

S501: Determine the swimming style.

For example, the user's swimming style is determined according to the method provided in steps S401-S408, so as to accurately determine the user's swimming style. In other embodiments, the swimming style may also be determined by other methods (for example, a method of determining the swimming style based on the sensor data of the watch 100 ).

S502: Perform axis selection according to the stroke.

Since the GYRO (gyroscope) data includes three coordinate axes of X, Y, and Z, the signals of different coordinate axes of the GYRO data corresponding to different swimming styles are different, and the user's hand movements are quite different when swimming with different swimming styles. , and the difference in head movements is small, so in order to improve the recognition accuracy, the stroke recognition module 112 can select the GYRO data collected by the watch 100 (including the three coordinate axes of X, Y and Z) according to the determined stroke The data of one of the coordinate axes is analyzed.

In some embodiments, if it is determined that the user's swimming style is breaststroke, the Y-axis signal of the GYRO data collected by the watch 100 is selected for analysis; Analyze the Z-axis signal of the data; if it is determined that the user's swimming style is butterfly, then choose to analyze the Z-axis signal of the GYRO data collected by the watch 100; if it is determined that the user's swimming style is backstroke or an unknown swimming style, then select Analyze the modulus values of the signals of the three coordinate axes of the ACC data (data collected by acceleration) collected by the watch 100, for example, after performing square operations on the data of the three coordinate axes of the ACC data, and then the three coordinate axes The results of the square operation are summed, and then the square root operation is performed on the summation result, and the modulus value obtained by the square root operation is used as a data source for subsequent waveform analysis of the watch 100 data analysis.

S503: Analyze the peak of watch data.

That is to analyze the GYRO (gyroscope) data or the coordinate axis data of the ACC data selected in S502 according to the determined stroke. The peaks and valleys of the filtered signal.

S504: Coarse screening action.

That is to say, after the peaks and valleys are determined for the above-mentioned filtered signal, the stroke action can be preliminarily screened according to the determined peaks and valleys. For example, the flow chart of a rough screening action shown in Figure 6B includes the following:

S5041: Determine the extreme point.

For example, the extreme points A1 and A2 of the signal at the watch 100 terminal shown in FIG. 5C are determined.

S5042: Determine whether the extreme point is greater than a threshold. If it is judged that the extreme point is greater than the threshold, then enter S5043, and determine the extreme point greater than the threshold as the peak point of the preliminary screening; if it is judged that the extreme point is less than or equal to the threshold, then enter S4044, and discard the corresponding extreme point.

S5043: Determine the peak point of the preliminary screening.

In some embodiments, a peak threshold may be preset, and when the searched extreme point is greater than the threshold, the corresponding extreme point is used as an action point to be selected, that is, determined as a peak point for preliminary screening. For example, in the embodiment shown in FIG. 5C , if the extreme points A1 and A2 are found, and the extreme points A1 and A2 are greater than the peak threshold Threshold3, then the extreme points A1 and A2 can be used as the peak points for preliminary screening.

S5044: Discard corresponding extreme points.

That is, if the searched extreme point is smaller than the aforementioned peak value threshold, the extreme point is discarded, that is to say, the extreme point is not used in the subsequent stroke recognition.

S5045: Determine whether the peak distance is greater than the time threshold. If it is judged that the peak distance is greater than the time threshold, go to S5046 to determine the primary screening action point; if it is judged that the peak distance is less than or equal to the time threshold, go to S4067 and discard the corresponding peak point.

S5046: Determine the primary screening action point.

That is, according to the time interval between wave peaks, the initially determined stroke action points are further screened out. Exemplarily, the time interval t between two adjacent peaks is judged. When t is greater than the time threshold T0, this peak is regarded as an effective peak. When the time interval t between two adjacent peaks is less than T0, the adjacent peak is selected The larger value is taken as the effective peak. For example, as shown in Figure 5C, the time interval t1 between two adjacent extreme points (that is, two adjacent peaks) A1 and A2 is greater than the time threshold T0, then the extreme points A1 and A2 can be used as the primary screening action point.

S5047: Discard the corresponding peak points.

That is to say, among the previously determined preliminary screening peak points, the primary screening peak points whose distance between peaks does not meet the time threshold are selected as invalid peak points.

According to the solution of this embodiment, it is possible to avoid the interference of invalid strokes on statistical data, which is beneficial to accurately count the number of strokes.

S505: Analyze the peak of the earphone 200 data.

The data source of the headphone 200 data peak analysis is the modulus value of the ACC data collected by the headphone 200. The method of finding the peak is similar to that in S504, except that the peak threshold and time threshold involved are different, and will not be described here.

S506: Action confirmation.

Although the wave peak analysis of the watch 100 data can obtain the initially screened stroke action point, but when the user's hand shakes in the swimming pool and the user is not swimming, the watch 100 may mistakenly judge that the user's hand shake is the user's swimming, thereby erroneously Therefore, it is necessary to combine the ACC data or GYRO data of the watch 100 and the ACC data of the earphone 200 again to further confirm the rough screening action determined in S504.

Exemplarily, it can be judged whether there is a peak point of the ACC data of the earphone 200 between the two rough screening action points obtained from the ACC data or GYRO data of the watch 100, and if so, the two rough screening action points Confirmed as a complete stroke. For example, in the embodiment shown in FIG. 5C, between the peak points A1 and A2 of the data at the watch 100 terminal, there is a peak point B1 corresponding to the signal at the earphone 200 terminal, and it can be confirmed that there is a complete peak point B1 between the peak points A1 and A2. paddling action.

From the above introduction about water stroke recognition, we can know that after analyzing the ACC data or GYRO data of the watch 100 to obtain the initial screening peak point, combined with the ACC data of the earphone 200, further confirm the stroke motion, and the screened stroke Water movements are more accurate.

It can be understood that the above execution order of steps S501 to S506 is just an example, and in other embodiments, other execution orders may also be adopted, and some steps may also be split or combined, which is not limited here.

An exemplary implementation of turn recognition will be introduced below. FIG. 6A is a flow chart of turning around recognition provided by some embodiments of the present application. The execution subject of each step in FIG. 6A may be the watch 100 . With reference to Figure 6A, the turn-around recognition process includes the following:

S601: Identify a stroke action.

For example, after the water stroke recognition module 112 of the watch 100 recognizes the user's water stroke, it sends the recognition result data to the turn recognition module 113 of the watch 100, and the turn recognition module 113 recognizes the user's stroke according to the received result data. stroke action. In other embodiments, other methods may also be used to identify stroke actions.

S602: Determine whether the number of actions is greater than the set number N1. In the following, N1=3 is taken as an example for introduction. But the present application is not limited thereto. In other embodiments, N can be 2, 5 and other numbers.

If it is judged that there are more than 3 strokes, go to S603 and calculate the mean value of the yaw angle of the earphone 200 corresponding to the first 3 strokes. If it is determined that the number of strokes is less than or equal to 3, return to S601.

S603: Record the average yaw angle Angle0 of the earphone 200 in the first three actions. Therefore, the starting orientation of the user's swimming can be determined according to the average yaw angle Angle0 of the earphone 200 corresponding to the first three strokes (for details, refer to FIG. 6B ).

For example, when the stroke movement recognition module 112 recognizes the stroke movement, it can also calculate the start time of each stroke movement during the swimming process of the user, so that when 3 stroke movements are recognized, take the 3 times The average value of the yaw angle Angle0 of the earphone 100 within the continuous stroke time period is used as the starting orientation of swimming.

Among them, the calculation method of the mean value of the yaw angle of the earphone 100 can be as follows: ACC data (including three axes), GYRO data (including three axes), and MAG data (including three axes) collected by the earphone 200 are calculated in nine axes. The attitude angle of the earphone 200 during swimming is obtained through fusion calculation, and the attitude angle is used as the yaw angle of the earphone 200 . The yaw angle of the user's head can also be calculated according to the attitude angle.

S604: Obtain the mean value Angle1 of the yaw angle of the earphone 200 during each subsequent stroke.

For example, the nine-axis fusion calculation is performed on the ACC data, GYRO data, and MAG data corresponding to each stroke after the first three strokes of the user's swimming, and the results of each stroke after the first three strokes are obtained. Corresponding to the average value of the yaw angle Angle1 of the earphone 200 .

S605: Determine whether the difference between the mean values of yaw angles of two adjacent actions is greater than a threshold value delta0. If it is determined that the difference between the average yaw angles of the two adjacent actions is greater than the threshold value delta0, then go to S606 and preliminarily determine that the user has turned around between the two adjacent actions. Otherwise, return to S604.

S606: It is pre-determined to turn around. That is to say, after judging that there are more than 3 strokes, you can compare the mean value of the yaw angle of each adjacent two strokes after the 3 strokes. If the difference of the mean value of the flight angle is greater than the difference threshold delta0, it can be preliminarily inferred that the swimmer has turned around.

S607: Judging whether the number of actions after the predicted turn is greater than the set number N2. In the following, N2=3 is taken as an example for introduction. But the present application is not limited thereto. In other embodiments, N can be 2, 4 and other numbers.

That is, it is judged whether there are more than 3 strokes after the turning motion of the aforementioned preliminary judgment. If it is determined that the number of strokes after the predicted turn is greater than 3, go to S608 to obtain the average value of the yaw angles of the three strokes after the predicted turn. Otherwise, enter S609 and wait for the number of actions to reach 3 after the predicted turn.

S608: Obtain the mean value of the yaw angle Angle2 within 3 consecutive strokes after the turn.

For example, the nine-axis fusion calculation is performed on the ACC data, GYRO data, and MAG data corresponding to the three consecutive strokes after the aforementioned predicted turn to obtain the yaw angle of the earphone 200 corresponding to the three strokes, and then calculate the The average value of the yaw angles corresponding to these three strokes is calculated to obtain the average value of the yaw angle Angle2 (for details, refer to FIG. 6B ).

S609: Wait for the number of actions to reach 3 after the predicted turn.

S610: Determine |Angle2-Angle0|>delta1. That is to determine whether the absolute value of the difference between the mean value of yaw angle Angle2 during the 3 consecutive strokes after turning around and the mean value of yaw angle Angle0 of the user's initial 3 consecutive strokes is greater than the difference threshold delta1. If the difference is greater than the difference threshold delta1, go to S611 to determine that a turning action has occurred, otherwise go to S612 to determine that the user has turned around falsely, that is, no turning has occurred.

S611: Make sure to turn around.

In some embodiments, when it is determined that the swimmer turns around, the average yaw angle Angle2 of three consecutive strokes after the turn can be used as the new swimming orientation, that is, the yaw angle Angle0 that originally represented the initial orientation of the earphone 100 The value is updated to the value of Angle2, so that the direction represented by the updated Angle0 is used as the reference direction for subsequent judgment of turning.

S612: Determine a false turn.

From the above introduction about the recognition of turning movements, it can be seen that when the user turns around during swimming, the signals of ACC data, GYRO data, and MAG data collected by the earphone 200 have a large difference. The ACC data, GYRO data, and MAG data are used to calculate the yaw angle, and a more accurate yaw angle can be obtained, so that the turn recognition result can be more accurate. In some embodiments, after the user's turning action is recognized, the total number of turning times can also be calculated, and then the user's swimming times can be calculated according to the total number of turning times. For example, if turning around is recognized n times, the total number of swimming times is counted as n+1 times.

In some embodiments, the number of strokes per stroke can also be calculated. For example, the number of strokes for the Nth (assuming N>1) stroke can be calculated from the time between the end of the N-1th turn and the start of the Nth turn. The sum of stroke times is obtained.

In some embodiments, it is also possible to make statistics on the strokes used in each swim, for example, to determine how many strokes are included in a swim, and the duration ratio of each stroke, and the stroke with the largest duration as the main stroke for the trip.

In some embodiments, the user's stroke frequency can also be calculated, for example, the sum of the duration of each stroke action is obtained as totalTime, and the total stroke times are determined as strokeCnt, then the average single stroke Water duration meanStrokeTime = totalTime/strokeCnt. Then convert the average single stroke duration into stroke frequency per minute: strokeFreq = (60 * 1000) / meanStrokeTime.

In some embodiments, after the watch 100 counts the number of strokes of the user, the number of strokes per stroke, the stroke, and the frequency of strokes, it can also be shown to the user through a display program. For details, refer to the interface shown in FIG. 7 schematic diagram.

It can be understood that the above execution order of steps S601 to S612 is just an example, and in other embodiments, other execution orders may also be adopted, and some steps may also be split or combined, which is not limited here.

The software logic block diagram of the watch 100 described above will be introduced below with reference to FIG. 8 . As shown in FIG. 8 , the watch 100 includes a swimming stroke recognition module 111 , a stroke motion recognition module 112 , a turn recognition module 113 , and a result statistics module 114 .

Among them, the swimming style recognition module 111 may be used to identify the user's swimming style, for example, to recognize that the user's certain swimming style is one of breaststroke, butterfly, backstroke, and freestyle as shown in FIG. 4D . When the user swims with these standard strokes, the accelerometer signal and gyroscope signal collected by the earphone 200 and the watch 100 will have obvious characteristics, and the stroke recognition model can be trained based on these characteristics, so as to use the trained model for stroke recognition. In some embodiments, the stroke recognition model may be a decision tree model.

The swimming stroke recognition module 111 may include a data group packet submodule M101, a feature extraction submodule M102, and a decision tree swimming stroke recognition submodule M103.

Exemplarily, the data packet sub-module M101 can be used for data packet (or data frame), that is, the wristwatch 100 determines that the user starts swimming through the data collected by the barometer 104, and the swimming stroke recognition module 111 in the wristwatch 100 The data collected by the accelerometer 101 and the gyroscope 102 (respectively denoted as ACC data and GYRO data), and the data collected by the accelerometer 201 and the gyroscope 202 of the earphone 200 returned in real time by the earphone 200 are packaged ( Also called framing). For example, the watch 100 collects 100 frames of ACC data and GYRO data in 1 second, and the earphone 200 collects 50 frames of ACC data and GYRO data in 1 second. The ACC data is averaged every two frames to obtain a new 50-frame ACC data, and the 100-frame GYRO data collected by itself is averaged every two frames to obtain a new 50-frame GYRO data. Then align the new 50 frames of ACC data and the new 50 frames of GYRO data with the 50 frames of ACC data and GYRO data returned by the headset 200 based on time to obtain 50 frames of combined data.

The feature extraction sub-module M102 can perform feature extraction on the combined data obtained above. For example, the window mean value of gyroscope and accelerometer data (that is, the mean value of ACC data and the mean value of GYRO data in a window), variance, mean absolute deviation, peak-to-valley characteristics, etc. are extracted from the combined data obtained above. feature. Among them, the window is a data window with a fixed duration that can be customized according to needs. For example, for the combined data of 50 frames obtained above, the data with a time span of 4 seconds is used as a window, and the step size can also be customized according to needs, for example The step size can be 0.5 seconds. Because in the foregoing embodiment, the data duration obtained after framing is 1 second, and there are 50 frames of ACC data and GYRO data, a window is made in 4 seconds, and a window has 200 frames of ACC data and GYRO data. In this case, Performing a feature extraction on the data in a window (that is, the data within a period of time) has a higher accuracy than one frame of data. In some embodiments, the feature data related to the gyroscope and accelerometer extracted above are extracted separately, for example, the ACC data and GYRO data in each window are respectively subjected to feature extraction to obtain the gyroscope and accelerometer in each window. feature data.

The decision tree stroke recognition sub-module M103 can perform decision tree stroke recognition based on the features extracted above. That is to say, the decision tree stroke recognition sub-module M103 performs stroke recognition based on the aforementioned extracted features, for example, using the aforementioned extracted feature data as input, using a pre-trained decision tree model to perform inference on these extracted feature data, The result of inference is the result of stroke recognition.

The stroke recognition module 112 may include a peak analysis submodule M104 and a stroke confirmation submodule M105. Exemplarily, the peak analysis sub-module M104 can be used to identify the user's water stroke. For example, after the swimming stroke recognition module 111 recognizes the swimming stroke of the user, the recognition result is sent to the peak analysis sub-module M104. Since the GYRO (gyroscope) data includes three coordinate axes of X, Y, and Z, the signals of different coordinate axes of the GYRO data corresponding to different swimming styles are different, and the user's hand movements are quite different when swimming with different swimming styles. , and the difference in head movements is small, so in order to improve the recognition accuracy, the peak analysis sub-module M104 can compare the GYRO data (including the three coordinate axes of X, Y, and Z) collected by the watch 100 according to different swimming strokes. The peak analysis of the data is carried out to initially identify the stroke action, and then the stroke action confirmation sub-module M105 performs the stroke action confirmation on the peak analysis of the ACC data collected by the earphone 200 . It should be understood that each time the user strokes the water, there will be a stroke action, so the recognition of the stroke action makes it easy to count how many times the user strokes the water in one swim.

The turn recognition module 113 may include a nine-axis fusion pose calculation submodule M106 and a turn recognition submodule M107. The turn-turn recognition module 113 can be used to recognize the user's turn-turn behavior. It should be understood that, usually, the user will turn around once after completing one swim and performing the next swim. Therefore, recognizing the user's turning behavior can facilitate counting how many times the user swam. When the user is swimming, the watch 100 keeps moving in space along with the user's arm, and since the earphone 200 is worn on the user's ear, and the movement of the user's head relative to the arm is relatively small during swimming, And when the user turns around, the posture of the earphone 200 changes greatly. Therefore, the turn recognition module 113 can use the ACC data, GYRO data, and MAG data (data collected by the magnetometer 203 of the earphone 200) of the earphone 200 in the aforementioned combined data for analysis. data, GYRO data, and MAG data through nine-axis fusion algorithm for nine-axis fusion attitude calculation, that is, the ACC data (three coordinate axes), GYRO data (three coordinate axes) and MAG data (three coordinate axes) of the headset 200 ) for fusion to calculate the attitude angle of the earphone 200 corresponding to each stroke, and then the turn recognition sub-module M107 performs turn recognition according to the calculated attitude angles of each stroke. It should be understood that if the attitude angle of the earphone 200 corresponding to two adjacent strokes changes greatly, it indicates that the user has turned around.

The result statistics module 114 can be used to count the swimming strokes of the user according to the swimming strokes identified by the aforementioned swimming stroke recognition module 111, the stroke strokes recognized by the stroke recognition module 112, and the turning behavior recognized by the turn recognition module 113. , stroke frequency, number of strokes, number of strokes (that is, stroke count). In order to facilitate the watch 100 to display the statistical results to the user for reference.

In some cases, the disclosed embodiments may be implemented in hardware, firmware, software, or any combination thereof. The disclosed embodiments can also be implemented as instructions carried by or stored on one or more transitory or non-transitory machine-readable (e.g., computer-readable) storage media, which can be executed by one or more processors read and execute. For example, instructions may be distributed over a network or via other computer-readable storage media. Thus, a machine-computer-readable storage medium may include any mechanism for storing or transmitting information in a form readable by a machine (eg, a computer), including but not limited to, floppy disks, compact disks, compact disks, read-only memories (CD-ROMs), , magneto-optical disk, read-only memory (Read Only Memory, ROM), random access memory (Random Access Memory, RAM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, EPROM), electrically erasable programmable Read-only memory (Electrically Erasable Programmable Read-Only Memory, EEPROM), magnetic or optical card, flash memory, or used to transmit information by means of electricity, light, sound or other forms of propagation signals using the Internet (for example, carrier waves, infrared signals, digital signals etc.) tangible machine-readable storage. Thus, a machine-computer-readable storage medium includes any type of machine-readable storage medium suitable for storing or transmitting electronic instructions or information in a form readable by a machine (eg, a computer).

In the drawings, some structural or methodological features may be shown in a particular arrangement and/or order. However, it should be understood that such specific arrangements and/or orderings may not be required. Rather, in some embodiments, these features may be arranged in a different manner and/or order than shown in the illustrative figures. Additionally, the inclusion of structural or methodological features in a particular figure does not imply that such features are required in all embodiments, and in some embodiments these features may not be included or may be combined with other features.

It should be noted that each unit/module mentioned in each device embodiment of this application is a logical unit/module. Physically, a logical unit/module can be a physical unit/module, or a physical unit/module. A part of the module can also be realized with a combination of multiple physical units/modules, the physical implementation of these logical units/modules is not the most important, the combination of functions realized by these logical units/modules is the solution The key to the technical issues raised. In addition, in order to highlight the innovative part of this application, the above-mentioned device embodiments of this application do not introduce units/modules that are not closely related to solving the technical problems proposed by this application, which does not mean that the above-mentioned device embodiments do not exist other units/modules.

It should be noted that in the examples and descriptions of this patent, relative terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply There is no such actual relationship or order between these entities or operations. Furthermore, the term "comprises", "comprises" or any other variation thereof is intended to cover a non-exclusive inclusion such that a process, method, article, or apparatus comprising a set of elements includes not only those elements, but also includes elements not expressly listed. other elements of or also include elements inherent in such a process, method, article, or device. Without further limitations, an element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, method, article or apparatus comprising said element.

Although this application has been shown and described with reference to certain preferred embodiments thereof, those skilled in the art will understand that various changes in form and details may be made therein without departing from this disclosure. The spirit and scope of the application.

Claims (13)

1. A swimming information statistics method, applied to a system including a first electronic device and a second electronic device, characterized in that the first electronic device is worn on the first part of the user, and the second electronic device is worn For the second part of the user, the difference between the motion trajectories of the first part under different swimming styles is smaller than the difference between the motion trajectories of the second parts, and the method includes: The first electronic device acquires first motion data of the user during swimming; The first electronic device acquires second motion data collected by the second electronic device during swimming of the user; The first electronic device performs data framing of the first motion data and the second motion data to obtain framed motion data, wherein the data framing method includes: Perform downsampling processing on the first motion data according to the frame rate of the second motion data, or perform interpolation processing on the second motion data according to the frame rate of the first motion data, so that the The frame rate of the first motion data is the same as that of the second motion data; The first electronic device performs sliding window processing on the framed motion data and extracts time domain features; The first electronic device determines a swimming stroke recognition result based on the extracted time-domain features and a preset swimming stroke recognition model; The first electronic device counts the user's swimming style based on the swimming style recognition result, wherein the user's swimming information includes the user's swimming style. 2. The method according to claim 1, wherein the first electronic device determines the swimming stroke recognition result based on the extracted time-domain features and the preset swimming stroke recognition model, including: The first electronic device inputs the time-domain features of the framed motion data into a preset stroke recognition model to obtain a stroke recognition result. 3. The method according to claim 2, wherein the preset stroke recognition model is composed of at least one binary classification decision tree, and, The first electronic device determines a swimming stroke recognition result based on the extracted time-domain features and a preset swimming stroke recognition model, including: Using the at least one binary classification decision tree and a preset scoring strategy to score the extracted time-domain features respectively to obtain a scoring result for the user's swimming stroke; The stroke corresponding to the maximum score in the stroke scoring results is determined as the stroke recognition result. 4. The method according to claim 3, wherein the user's swimming information also includes the user's stroke action, and the first electronic device calculates the user's swimming action based on the framed motion data. Swimming information, including: The first electronic device determines first target data for counting the user's stroke movements in the framed motion data according to the counted swimming strokes of the user, wherein the first The target data includes first target sub-data collected by the first electronic device and second target sub-data collected by the second electronic device; The first electronic device determines a plurality of coarse sieving strokes according to the first target sub-data; According to the second target sub-data, the first electronic device counts the stroke actions of the user from the plurality of roughly screened stroke actions. 5. The method according to claim 4, wherein the first electronic device includes a gyroscope and an accelerometer, and the first target sub-data includes data collected by the gyroscope of the first electronic device or the The data collected by the accelerometer of the first electronic device. 6. The method according to claim 4, wherein the first electronic device includes an accelerometer, and the second target sub-data includes data collected by the accelerometer of the second electronic device. 7. The method according to any one of claims 4 to 6, wherein the user's swimming information includes the user's turning action, and the first electronic device calculates based on the framed motion data The user's swimming information, including: The first electronic device determines the user's initial water stroke; The first electronic device determines the user's initial swimming direction according to the second motion data corresponding to the first number of consecutive strokes including the initial stroke; The first electronic device determines each stroke after the first number of strokes in succession according to the second motion data corresponding to each stroke after the first number of strokes in succession Corresponding swimming direction; When the first electronic device judges that the user turns around according to the difference between the swimming directions corresponding to every two adjacent strokes after the first number of consecutive strokes, the first electronic device The device determines the corresponding target stroke action when the user turns around; The first electronic device determines, according to the second motion data corresponding to the second number of consecutive strokes after the target stroke, the stroke corresponding to the second number of consecutive strokes after the target stroke. target swimming direction; When the first electronic device determines that the difference between the target swimming direction and the initial swimming direction reaches a threshold, the first electronic device confirms that the user has turned around during the target stroke, and counts all Describe the user's turning action. 8. The method according to claim 7, further comprising: The first electronic device updates the initial swimming direction to the target swimming direction. 9. The method according to claim 8, wherein the second electronic device includes a gyroscope, an accelerometer, and a magnetometer, and the second motion data includes data collected by the gyroscope of the second electronic device . Data collected by the accelerometer of the second electronic device, and data collected by the magnetometer of the second electronic device. 10. The method according to claim 1, wherein the first electronic device is a sports watch. 11. The method according to claim 1, wherein the second electronic device is a sports earphone. 12. A computer-readable storage medium, wherein instructions are stored on the computer-readable storage medium, and when the instructions are executed on the electronic equipment, the electronic equipment executes the method described in any one of claims 1-11. Statistical method of swimming information. 13. An electronic device, characterized in that it comprises: memory for storing instructions to be executed by one or more processors of the electronic device, and, A processor, when the instructions are executed by one or more processors, the processor is used to execute the swimming information statistics method according to any one of claims 1-11.

Images