Swimming action analysis method and device, wearable device and storage medium

Abstract

The present disclosure provides a swimming action analysis method and device, a wearable device and a storage medium. The method comprises: determining that a swimming state is entered; obtaining an action sensing signal of a swimming action; obtaining initial trajectory information according to the action sensing signal; and performing centroid processing on the initial trajectory information to obtain corrected trajectory information. The initial trajectory information of the swimming action is processed by the centroid processing to obtain the corrected trajectory information, which can reduce the error of the swimming action analysis and effectively improve the accuracy and analysis effect of the swimming action analysis.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of data processing, and particularly relates to a swimming action analysis method and device, a wearable device and a storage medium. BACKGROUND

[0002] In recent years, the popularity of wearable devices continues to rise, and the technology is becoming more and more mature. People's demand for fitness is also becoming stronger. Swimming is currently the third largest sport, and swimming analysis as a basic function of wearable devices is becoming more and more popular in the market. The public acceptance and use frequency are also becoming higher and higher.

[0003] In the related art, a geomagnetic sensor is usually used to analyze swimming actions.

[0004] In this way, the analysis result of the swimming action is inaccurate, and the analysis effect is poor. SUMMARY

[0005] The present disclosure aims to at least partially solve one of the technical problems in the related art.

[0006] To this end, the purpose of the present disclosure is to propose a swimming action analysis method and device, a wearable device, a storage medium and a computer program product. Since the initial trajectory information of the swimming action is processed by the centroid to obtain the corrected trajectory information, the swimming action analysis error can be reduced, and the accuracy and analysis effect of the swimming action analysis can be effectively improved.

[0007] The swimming action analysis method proposed by the first aspect of the present disclosure comprises: determining the entry into the swimming state; obtaining the action sensing signal of the swimming action; obtaining the initial trajectory information according to the action sensing signal; and processing the initial trajectory information by the centroid to obtain the corrected trajectory information.

[0008] The swimming action analysis method proposed by the first aspect of the present disclosure determines the entry into the swimming state, obtains the action sensing signal of the swimming action, obtains the initial trajectory information according to the action sensing signal, and processes the initial trajectory information by the centroid to obtain the corrected trajectory information. Since the initial trajectory information of the swimming action is processed by the centroid to obtain the corrected trajectory information, the swimming action analysis error can be reduced, and the accuracy and analysis effect of the swimming action analysis can be effectively improved.

[0009] The swimming action analysis device provided in the second aspect of the present disclosure comprises: a judgment module configured to judge whether a swimming state is entered; a first acquisition module configured to acquire an action sensing signal of a swimming action; a second acquisition module configured to obtain initial trajectory information according to the action sensing signal; and a processing module configured to perform centroid processing on the initial trajectory information to obtain corrected trajectory information.

[0010] The swimming action analysis device provided in the second aspect of the present disclosure judges whether a swimming state is entered, acquires an action sensing signal of a swimming action, obtains initial trajectory information according to the action sensing signal, and performs centroid processing on the initial trajectory information to obtain corrected trajectory information. Since the corrected trajectory information is obtained by performing centroid processing on the initial trajectory information of the swimming action, the swimming action analysis error can be reduced, and the accuracy and analysis effect of the swimming action analysis can be effectively improved.

[0011] According to the third aspect of the present disclosure, a wearable device is provided, which comprises at least one processor and a memory connected with the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the swimming action analysis method of the first aspect of the present disclosure.

[0012] According to the fourth aspect of the present disclosure, a non-transitory computer readable storage medium storing computer instructions is provided, and the computer instructions are used to enable the computer to execute the swimming action analysis method of the first aspect of the present disclosure.

[0013] According to the fifth aspect of the present disclosure, a computer program product is provided, which comprises a computer program, and the computer program, when executed by a processor, implements the swimming action analysis method of the first aspect of the present disclosure.

[0014] According to the above-mentioned solutions of the present disclosure, by judging whether a swimming state is entered, acquiring an action sensing signal of a swimming action, obtaining initial trajectory information according to the action sensing signal, and performing centroid processing on the initial trajectory information to obtain corrected trajectory information, since the corrected trajectory information is obtained by performing centroid processing on the initial trajectory information of the swimming action, the swimming action analysis error can be reduced, and the accuracy and analysis effect of the swimming action analysis can be effectively improved.

[0015] Additional aspects and advantages of the present disclosure will be in part apparent and in part pointed out hereinafter. BRIEF DESCRIPTION OF DRAWINGS

[0016] The above mentioned and / or additional aspects and advantages of the present disclosure will become apparent and more readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings, in which:

[0017] Figure 1 is a flowchart of a swimming action analysis method according to an embodiment of the present disclosure;

[0018] Figure 2 is a flowchart of a swimming action analysis method according to another embodiment of the present disclosure;

[0019] Figure 3 is a schematic diagram of a swimming action analysis model according to another embodiment of the present disclosure;

[0020] Figure 4 is a flowchart of a swimming action analysis method according to another embodiment of the present disclosure;

[0021] Figure 5 is a flowchart of a swimming action analysis method according to another embodiment of the present disclosure;

[0022] Figure 6 is a flowchart of a swimming action analysis method according to another embodiment of the present disclosure;

[0023] Figure 7 is a flowchart of a swimming action analysis method according to another embodiment of the present disclosure;

[0024] Figure 8 is a flowchart of a swimming action analysis method according to another embodiment of the present disclosure;

[0025] Figure 9 is a structural schematic diagram of a swimming action analysis device according to an embodiment of the present disclosure;

[0026] Figure 10 is a structural schematic diagram of a swimming action analysis device according to another embodiment of the present disclosure;

[0027] Figure 11 shows a block diagram of an exemplary wearable device suitable for use in implementing embodiments of the present disclosure. DETAILED DESCRIPTION

[0028] Embodiments of the present disclosure are described in detail below with reference to the attached drawing figures, wherein the same or like reference numerals are used throughout the drawing figures to refer to the same or like elements or elements having the same or similar functionality. The embodiments described below are exemplary and are intended to be illustrative of the present disclosure, and are not to be construed as limiting thereof. Conversely, embodiments of the present disclosure encompass all changes, modifications and alterations of the embodiments falling within the spirit and scope of the appended claims.

[0029] Figure 1is a flowchart of a swimming action analysis method according to an embodiment of the present disclosure.

[0030] It should be noted that the execution subject of the swimming action analysis method in this embodiment is a swimming action analysis device, which can be implemented in the form of software and / or hardware, and can be configured in a wearable device such as a smart watch, a smart bracelet, etc., without limitation.

[0031] The wearable device is a portable device that can be directly worn on the body or integrated into the user's clothes or accessories. The wearable device is not only a hardware device, but also realizes corresponding functions through software support and data interaction and cloud interaction.

[0032] It should be noted that the signals and data related to the swimming action in the embodiments of the present disclosure are obtained after authorization by the relevant user, and the obtaining process complies with relevant laws and regulations and does not violate public order and good customs.

[0033] As shown in Figure 1 The swimming action analysis method comprises the following steps.

[0034] S101: Determine whether to enter a swimming state.

[0035] The state information of the measured person in swimming can be referred to as a swimming state, which can be used to determine that the measured person is performing a swimming action in the water, without limitation.

[0036] In the embodiments of the present disclosure, a sensor or the like can be used to detect whether the measured person is in a swimming state, or a swimming state switching button can be provided in the corresponding system, and whether the measured person enters the swimming state can be determined according to the operation of the measured person on the switching button, or any other possible implementation manner can be used to determine whether the measured person enters the swimming state, without limitation.

[0037] For example, when the measured person enters the swimming pool and performs a swimming action, the sensor on the measured person records the action of the measured person to determine whether the measured person enters the swimming state, or the measured person can operate the state switching button in the swimming action analysis device to switch the state to the swimming state, without limitation.

[0038] S102: Obtain an action sensing signal of the swimming action.

[0039] The signal for action analysis of the swimming action can be referred to as an action sensing signal. The action sensing signal can be a signal collected by a wearable device with a sensor in real time through the sensor for sensing the swimming action. Alternatively, the action sensing signal can be an action sensing signal of the swimming action stored in a big data platform. No limitation is made in this regard.

[0040] In the embodiments of the present disclosure, the sensor can be one or more of a six-axis sensor, a visual sensor, a fluid sensor, a geomagnetic sensor, etc. No limitation is made in this regard.

[0041] In the embodiments of the present disclosure, the sensor signal collected by each sensor for sensing the swimming action can be directly acquired and used as the action sensing signal. Alternatively, the acquired sensor signal can be preprocessed, for example, by data denoising, median filtering, etc., and the processed signal can be used as the action sensing signal. Alternatively, any other possible implementation manner can be used to acquire the action sensing signal of the swimming action. No limitation is made in this regard.

[0042] In some embodiments, the action sensing signal can include an acceleration signal of the swimming action, and / or an angular motion signal of the swimming action, and / or a derivative sensing signal obtained by fusing the acceleration signal and the angular motion signal. Since the acceleration signal, and / or the angular motion signal, and / or the derivative sensing signal obtained by fusing the acceleration signal and the angular motion signal is used as the action sensing signal, various types of action sensing signals can be collected, which effectively improves the comprehensiveness and dimensionality of the action sensing signal collection. When the swimming action is analyzed based on the action sensing signal with multiple dimensions, the comprehensiveness of the analysis consideration can be effectively improved, the analysis effect is more accurate, and the reliability and accuracy of the action sensing signal for characterizing the swimming action feature are effectively improved.

[0043] The signal related to the acceleration of the subject during the swimming process can be referred to as an acceleration signal, and the signal related to the angular velocity of the subject during the swimming process can be referred to as an angular velocity signal. Some fusion algorithm is used to fuse the acceleration and the angular velocity, and the signal obtained by fusion can be referred to as a derivative sensing signal. The acceleration signal can be used to determine the characteristics related to the acceleration of the subject during the swimming process, the angular velocity signal can be used to determine the characteristics related to the angular velocity of the subject during the swimming process, and the derivative sensing signal can be used to determine the characteristics related to the fusion of the acceleration and the angular velocity of the subject during the swimming process. No limitation is made in this regard.

[0044] In the embodiments of the present disclosure, the acceleration sensor can be used to detect the acceleration signal of the swimming action, the angular velocity sensor can be used to detect the angular velocity signal of the swimming action, and then the acceleration signal and the angular velocity are fused to obtain the derived sensing signal. At least one of the aforementioned acceleration signal, angular velocity signal and derived sensing signal can be used as the action sensing signal, or the positioning system can be used to locate the subject to obtain the positioning position, and the action tracking device can be used to track the swimming action of the subject with reference to the positioning position to obtain the corresponding acceleration signal, angular velocity signal and derived sensing signal of the subject in the swimming process, and at least one of the aforementioned acceleration signal, angular velocity signal and derived sensing signal can be used as the action sensing signal, and this is not limited.

[0045] In other embodiments, the action sensing signal can further include a frequency sensing signal describing the stroke frequency, and / or a trajectory sensing signal describing the motion trajectory, and / or a posture sensing signal describing the motion posture, and the like, and this is not limited.

[0046] S103: Obtain initial trajectory information according to the action sensing signal.

[0047] The motion trajectory of the subject in the swimming process initially determined based on the action sensing signal can be referred to as initial trajectory information, and the initial trajectory information can be the trajectory information of the arm sliding, the trajectory information of the swimming route, and the like. The trajectory information, such as the position of each trajectory point in the trajectory, the corresponding acquisition time point of the trajectory point, and the like, is not limited.

[0048] In the embodiments of the present disclosure, the corresponding trajectory information can be analyzed as the initial trajectory information by performing signal feature analysis processing on the action sensing signal, or the corresponding trajectory information can be simulated as the initial trajectory information by using a trajectory simulation model according to the acceleration, angular velocity and other information recorded in the action sensing signal, or other arbitrary possible implementation manners can be used to determine the initial trajectory information, and this is not limited.

[0049] For example, a swimming trajectory analysis model can be built, and the acceleration signal, angular velocity signal and derived sensing signal in the action sensing signal are input into the swimming trajectory analysis model to obtain the corresponding initial trajectory information.

[0050] S104: Perform centroid processing on the initial trajectory information to obtain the corrected trajectory information.

[0051] The center of mass is referred to as the center of mass, and refers to a hypothetical point on a mass system where the mass is concentrated. The center of mass processing of the initial trajectory information refers to the corresponding correction processing of a hypothetical point where the mass is concentrated in the initial trajectory information. The center of mass may be, for example, a hypothetical point where the mass of the initial trajectory information related trajectory data segment is concentrated, and is not limited in this regard.

[0052] In the embodiments of the present disclosure, the processing of the initial trajectory information may be the corresponding correction processing of the center of mass related to each trajectory data segment. For example, the initial trajectory information may be corrected and adjusted using the center of mass processing method to obtain corrected trajectory information, and the corrected trajectory information is used to analyze the swimming motion of the subject.

[0053] In some embodiments, the center of mass processing of the initial trajectory information may be to determine each center of mass point in the swimming trajectory, and then to adjust and correct the position, coordinates, etc. of each center of mass point according to the motion sensing signal to obtain corrected trajectory information, and the present disclosure is not limited in this regard.

[0054] In other embodiments, the initial trajectory information may also be input into a center of mass processing model to process the initial trajectory information using the execution processing model to obtain corrected trajectory information, and the present disclosure is not limited in this regard.

[0055] In other embodiments, the initial trajectory information may also be processed using any other possible method to obtain corrected trajectory information, and the present disclosure is not limited in this regard.

[0056] In the embodiments, the swimming state is determined, the motion sensing signal of the swimming motion is obtained, the initial trajectory information is obtained according to the motion sensing signal, and the initial trajectory information is processed to obtain corrected trajectory information. Since the corrected trajectory information is obtained by processing the initial trajectory information of the swimming motion, the swimming motion analysis error is reduced, and the accuracy and analysis effect of the swimming motion analysis are effectively improved.

[0057] Figure 2 FIG. 1 is a flowchart of a swimming motion analysis method according to another embodiment of the present disclosure.

[0058] As shown in FIG. 1, the swimming motion analysis method includes the following steps. Figure 2

[0059] S201: Determine whether to enter a swimming state.

[0060] ​Optionally, in the embodiments of the present disclosure, the swimming state of the swimming action can be determined according to the action sensing signal, or the swimming state is determined according to the software instruction, and if the swimming state meets the state condition, the swimming time range of the swimming action is determined. Since the corresponding swimming state is determined according to the action sensing signal or the software instruction, and the state condition is set to determine the swimming time range, it can be determined whether the swimming state is in the swimming state, and the accuracy and reliability of the swimming state determination are effectively improved.

[0061] The state condition can be used to determine whether the swimming state is in the swimming state, that is, when the swimming state is in the swimming state, it can be determined that the swimming state meets the state condition, and when the swimming state is not in the swimming state, it can be determined that the swimming state does not meet the state condition.

[0062] In the embodiments of the present disclosure, the action sensing signal can be analyzed and processed to determine the swimming state of the swimming action. For example, the feature comparison method, the engineering method, the model matching method, etc. can be used to determine the swimming state of the swimming action in combination with the action sensing signal, and no limitation is made thereto.

[0063] In the embodiments of the present disclosure, the threshold value judgment, the decision tree, the random forest, the multi-layer perception (MLP), etc. machine learning method can be used to determine the swimming state of the swimming action, or the deep learning method can be used to determine the swimming state of the swimming action, or other any possible implementation manner can be used to determine the swimming state of the swimming action, and no limitation is made thereto.

[0064] That is, in the embodiments of the present disclosure, it can be determined whether the swimming state meets the state condition, and when it is determined that the swimming state meets the state condition, the swimming time range of the swimming action can be triggered.

[0065] The swimming duration from determining the swimming start to determining the swimming end can be referred to as the swimming time range. In the swimming process, the time range between the swimming start time and the swimming end time of the corresponding full course, single course or single lap of the swimming can be taken as the swimming time range, and no limitation is made thereto.

[0066] In the embodiments of the present disclosure, the state condition can also be set as whether the swimming action is made. When it is detected that the subject makes the swimming action, it is determined that the swimming state meets the state condition, and when it is detected that the subject does not make the swimming action, it is determined that the swimming state does not meet the state condition. The swimming action can be, for example, a water entry action, an arm sliding action, a leg swinging action, etc. Or, a state confirmation interaction button can be set in the related device, and when it is detected that the subject triggers the button, it is confirmed that the swimming state meets the state condition, and no limitation is made thereto.

[0067] Optionally, in some embodiments, if the swimming state satisfies the state condition, it is determined whether the action scene of the swimming action is a water scene, and according to the action sensing signal, it is determined the action continuity of the swimming action, if the action scene is a water scene and / or the action continuity satisfies the continuity condition, the swimming time range of the swimming action is determined, the swimming time range is triggered by combining the action scene and the action continuity, which can effectively guarantee the objectivity, rationality and accuracy of the determination of the swimming time range.

[0068] The continuity of the swimming action of the subject in the movement process can be referred to as action continuity, and the action continuity of the subject can be determined in combination with the swimming posture. The judgment standard of the action continuity can be whether the subject produces continuous action in the movement process, such as periodic sliding arms, periodic kicking legs, etc. The action continuity is determined according to whether the movement of the subject in the movement process is interrupted or stopped, which is not limited.

[0069] In the embodiments of the present disclosure, the sensor related to liquid detection can be used to determine whether the action scene of the subject is a water scene, or the subject can set a state (such as a warm-up mode, an underwater swimming mode, etc.) to determine whether the action scene of the subject is a water scene, or any other possible implementation manner to determine whether the action scene is a water scene, which is not limited.

[0070] In the embodiments of the present disclosure, when it is detected that the action scene is not a water scene or the action continuity does not satisfy the continuity condition, the detection can be continued until the action scene is a water scene and the action continuity satisfies the continuity condition, triggering the recording of the swimming time range of the swimming action, which is not limited.

[0071] S202: Obtain an action sensing signal of the swimming action.

[0072] Thus, the action signal of the swimming action in the swimming time range can be obtained as the action sensing signal. The description of the action sensing signal can be referred to the above embodiments, which will not be repeated here.

[0073] S203: Obtain initial trajectory information according to the action sensing signal.

[0074] The description of S203 can be specifically referred to the above embodiments, which will not be repeated here.

[0075] S204: Perform trajectory division on the initial trajectory information to obtain a plurality of trajectory data segments.

[0076] The data segment obtained by performing trajectory division on the initial trajectory information can be referred to as a trajectory data segment, and the trajectory data segment can be used to represent data information of a corresponding motion trajectory.

[0077] In the embodiments of the present disclosure, the initial trajectory information in the swimming time range can be divided to determine a plurality of trajectory data segments, or the initial trajectory information in the swimming time range can be recorded in a preset period, and the initial trajectory information obtained in each preset period can be analyzed to obtain a plurality of corresponding trajectory data segments. This is not limited.

[0078] Optionally, in some embodiments, when the initial trajectory information is divided to obtain a plurality of trajectory data segments, the same number of data points can be cut from the initial trajectory information each time, and a trajectory data segment can be formed according to the same number of data points cut each time; or the initial trajectory information can be divided based on a peak point, a trough point, and / or a zero point to obtain a plurality of trajectory data segments. Since the same number of data points is cut each time, and a trajectory data segment is formed according to the same number of data points cut each time, and / or the initial trajectory information is divided based on a peak point, a trough point, and / or a zero point, the trajectory data segment can be obtained more accurately, the division of the trajectory data segment can more objectively describe the corresponding initial trajectory information, and the flexibility of the division of the trajectory data segment can be effectively improved when a plurality of trajectory data segments are obtained by using a plurality of division methods.

[0079] In some embodiments, a peak value detection method can be used to set a signal peak value, and a corresponding trajectory data segment can be determined according to the peak value. The peak value can be a peak point or a trough point, which can be adjusted according to a scene and an action. Alternatively, a zero-crossing point detection method can be used to set zero point information, and a plurality of corresponding trajectory data segments can be obtained according to a signal passing through a zero point. This is not limited.

[0080] In other embodiments, a plurality of methods such as a peak value detection method and a zero-crossing point detection method can be combined to determine a plurality of corresponding trajectory data segments. This is not limited.

[0081] S205: Centroid recognition is performed on the plurality of trajectory data segments respectively to obtain a plurality of centroid information respectively corresponding to the plurality of trajectory data segments.

[0082] The multiple trajectory data segments in the action sensing signal are subjected to centroid identification to obtain multiple centroid points (the centroid point is the centroid of the trajectory data segment), and the information corresponding to the multiple centroid points can be referred to as centroid information. The centroid information can be used to determine the motion trajectory and trajectory characteristics of the subject within the swimming time range, and the like, and no limitation is made in this regard.

[0083] In the embodiments of the present disclosure, one trajectory data segment can correspond to one centroid point, and the centroid point can be used to determine the trajectory characteristics corresponding to the motion trajectory in the trajectory data segment, such as the average speed, acceleration, time, and the like information corresponding to the centroid point, and no limitation is made in this regard.

[0084] Optionally, in the embodiments of the present disclosure, when the multiple trajectory data segments are subjected to centroid identification to obtain multiple centroid information corresponding to the multiple trajectory data segments respectively, the multiple centroid point positions corresponding to the multiple centroid points can be determined, the centroid point distance between adjacent centroid point positions can be determined, the adjacent centroid point positions refer to the first centroid point position and the second centroid point position with a time domain distance of N points, N is greater than or equal to 1, the position change information between the multiple centroid point positions can be determined, and the multiple centroid point positions, and / or the centroid point distance, and / or the position change information are used as the multiple centroid information. Since the multiple centroid point positions, and / or the centroid point distance, and / or the position change information are used as the multiple centroid information, the multiple centroid points can be accurately and multi-dimensionally represented, the accuracy of the centroid information determination is effectively guaranteed, and thus the accuracy of the overall trajectory information correction is effectively improved, and the accuracy of the swimming action analysis is effectively improved.

[0085] The position of the centroid point in the corresponding trajectory data segment can be referred to as the centroid point position, the distance between the adjacent centroid point positions can be referred to as the centroid point distance, and the position point change, the direction change of the position vector, and the like information generated by the centroid point position can be referred to as the position change information.

[0086] In the embodiments of the present disclosure, any one or a combination of multiple centroid point positions, centroid point distances, and position change information can be used as the multiple centroid information, and no limitation is made in this regard.

[0087] In some embodiments, one of the plurality of centroid points can be preset as a first centroid point, the position corresponding to the first centroid point can be referred to as a first centroid point position, a centroid point corresponding to the positions of the N points in the time domain can be referred to as a second centroid point, and the position corresponding to the second centroid point can be referred to as a second centroid point position. Thus, the centroid point distance between the first centroid point position and the second centroid point position is determined, the position change information between the plurality of centroid point positions is determined, and the centroid point position of each centroid point is determined. One or more combinations of the determined centroid point distance, position change information, and centroid point position are used as centroid information, and no limitation is imposed thereon.

[0088] In other embodiments, the swimming track can also be positioned using a positioning system, the initial track information is determined according to the positioning, the initial track information is divided to obtain a plurality of track data segments, the plurality of divided track data segments are respectively mapped to a plurality of positioning data information in the positioning system, the centroid of each track data segment is identified based on the corresponding positioning data information, the plurality of centroid point positions and / or centroid point distances and / or position change information are determined, and the centroid information is used, and no limitation is imposed thereon.

[0089] In other embodiments, any other possible implementation can be used to identify the centroid of each track data segment to obtain the plurality of centroid information corresponding to the plurality of track data segments, and no limitation is imposed thereon.

[0090] S206: The plurality of centroid information is processed to obtain the corrected track information.

[0091] In the embodiments of the present disclosure, the plurality of centroid information can be analyzed and processed, and the initial track information of the measured person within the swimming time range is corrected according to the analysis result to obtain the corrected track information.

[0092] In some embodiments, the centroid information can be processed by using a data processing method (for example, setting a big data model, etc.), or a data analysis method can also be used, for example, a swim efficiency (Swim+Golf, SWOLF) analysis model can be used to extract the corresponding centroid information and analyze and process the centroid information to obtain the corrected track information, and no limitation is imposed thereon.

[0093] Optionally, in some embodiments, according to the plurality of center of mass information, the action stroke information, and / or the swimming turn information, and / or the swimming posture information of the swimming action in the swimming time range is determined; the action stroke information, and / or the swimming turn information, and / or the swimming posture information is taken as the corrected trajectory information. Since the plurality of center of mass information is analyzed and obtained, the action stroke information, and / or the swimming turn information, and / or the swimming posture information in the swimming time range is determined according to the center of mass information, more accurate and comprehensive information in the swimming time range can be obtained, the comprehensiveness and integrity of the swimming action analysis can be effectively improved, and the application scenarios of the swimming action analysis method can be effectively expanded.

[0094] The information used to describe the action stroke of the swimming action can be referred to as action stroke information.

[0095] For example, the action stroke information can be, for example, information indicating the number of arm strokes of the measured person in the swimming process, or the number of leg swings, and the like, and is not limited in this regard.

[0096] The information used to describe the posture of the swimming action can be referred to as swimming posture information.

[0097] For example, the swimming posture information can be the swimming posture and body shape of the measured person in the swimming process, and is not limited in this regard, for example, breaststroke, backstroke, freestyle, butterfly, and the like.

[0098] The information used to describe the turn action related to the swimming action can be referred to as swimming turn information.

[0099] For example, the swimming turn information can be, for example, information indicating the number of turns of the measured person in the swimming process, and / or the turn position, and / or the turn direction, and the like, and is not limited in this regard.

[0100] In the embodiments of the present disclosure, a deep learning method can be used to build a corresponding swimming action analysis model, as shown in FIG. 1. Figure 3 Figure 3 FIG. 2 is a schematic diagram of a swimming action analysis model according to another embodiment of the present disclosure, wherein the swimming detection module is used to obtain the action sensing signal of the swimming action, process the action sensing signal, and obtain the corresponding action stroke information, swimming posture information, and swimming turn information; the stroke detection module is used to determine the action stroke information; the swimming posture recognition module is used to determine the swimming posture information; the turn judgment module is used to determine the swimming turn information; and the action stroke information, the swimming posture information, and the swimming turn information are taken as the corrected trajectory information and input to the post-processing module for analysis and processing to obtain the swimming action analysis result.

[0101] ​Optionally, according to the stroke information, the swimming turn information, and / or the swimming posture information, the swimming action can be analyzed. The stroke information, the swimming posture information, and the swimming turn information can be analyzed to obtain the current swimming lap, the swimming length, and / or the swimming distance. The stroke information, the swimming posture information, and the swimming turn information can be analyzed to obtain the stroke per lap, the stroke in a set time range, the total stroke, the stroke frequency, and / or the swimming posture statistical information. The swimming posture statistical information includes the main swimming posture information and / or the mixed swimming posture information. Since multiple information is analyzed and obtained, more accurate and comprehensive information in the swimming process can be obtained, the comprehensiveness and integrity of the swimming action analysis can be effectively improved, and the application scenarios of the swimming action analysis method can be effectively expanded.

[0102] In the embodiments of the present disclosure, the stroke information, the swimming posture information, and the swimming turn information can be combined to analyze multiple information of the measured person within the swimming time range. The analysis method can be a pre-set data analysis model, or artificial intelligence, big data analysis, etc. This is not limited.

[0103] In the embodiment, the initial trajectory information of the swimming action is processed by the center of mass to obtain the corrected trajectory information for analyzing the swimming action, which can reduce the error of the swimming action analysis and effectively improve the accuracy and analysis effect of the swimming action analysis. The swimming time range is determined by combining the action scene and the action continuity, which can effectively guarantee the objectivity, rationality and accuracy of the determination of the swimming time range. The same number of data points are intercepted, and / or the segmentation based on the peak point, and / or the segmentation based on the trough point, and / or the segmentation based on the zero point, etc. are used to obtain the plurality of trajectory data segments, which can more accurately obtain the trajectory data segment, so that the division of the trajectory data segment more objectively describes the corresponding initial trajectory information, and effectively improves the flexibility of the division of the trajectory data segment. The plurality of center of mass points, and / or the center of mass point distance, and / or the position change information are used as the plurality of center of mass information, which can accurately and multidimensionally represent the plurality of center of mass points, effectively guarantee the accuracy of the determination of the center of mass information, effectively improve the accuracy of the correction of the overall trajectory information, and effectively improve the accuracy of the swimming action analysis. The action division information, and / or the swimming turn information, and / or the swimming posture information in the swimming time range are determined according to the center of mass information, which can obtain more accurate and comprehensive information in the swimming time range, effectively improve the comprehensiveness and integrity of the swimming action analysis, and effectively expand the application scene of the swimming action analysis method. The plurality of information is analyzed and obtained, which can obtain more accurate and comprehensive information in the swimming process, effectively improve the comprehensiveness and integrity of the swimming action analysis, and effectively expand the application scene of the swimming action analysis method. The initial trajectory information is divided into a plurality of trajectory data segments, the center of mass of the plurality of trajectory data segments is recognized to obtain the plurality of center of mass information corresponding to the plurality of trajectory data segments respectively, and the center of mass of the plurality of center of mass information is processed to obtain the corrected trajectory information, which can process the initial trajectory information by the center of mass according to the plurality of trajectory data segments, correct the initial trajectory information, and then obtain more objective trajectory information, effectively improve the accuracy and objectivity of the trajectory information.

[0104] Figure 4 is a flowchart of a swimming action analysis method according to another embodiment of the present disclosure.

[0105] As shown in Figure 4 , the swimming action analysis method comprises:

[0106] S401: determining whether to enter a swimming state.

[0107] S402: obtaining an action sensing signal of a swimming action.

[0108] S403: obtaining initial trajectory information according to the action sensing signal.

[0109] S404: performing trajectory division on the initial trajectory information to obtain a plurality of trajectory data segments.

[0110] S405: respectively performing centroid identification on the plurality of trajectory data segments to obtain a plurality of centroid information corresponding to the plurality of trajectory data segments respectively.

[0111] The description of S401-S405 can be specifically referred to the above embodiments, which will not be repeated here.

[0112] S406: determining a target centroid position from the plurality of centroid positions in the swimming time range.

[0113] Among the plurality of centroid positions in the swimming time range, one or more centroid positions selected according to a pre-set screening condition can be referred to as a target centroid position, wherein the screening condition can be adaptively configured according to actual scene requirements, which is not limited.

[0114] For example, in the process of determining the swimming stroke information, the screening condition can be set as the position of the arm swing to a certain angle, and the plurality of target centroid positions can be counted, which is not limited.

[0115] In the embodiments of the present disclosure, the sliding window algorithm can be used to determine the centroid position, the target centroid position can be calculated and counted according to the peak value detection, zero-crossing detection, numerical differentiation method, or a condition judgment model can be built, the screening condition can be set, and the plurality of centroid positions meeting the screening condition can be determined as the target centroid position, or a deep learning method can be used to train a deep learning model, the plurality of centroid positions can be screened by the trained deep learning model, and the target centroid position can be obtained, which is not limited.

[0116] Optionally, in some embodiments, the peak value detection method, and / or the zero-crossing detection method, and / or the numerical differentiation method can be used to determine the target centroid position from the plurality of centroid positions in the swimming time range. Since the peak value detection method, and / or the zero-crossing detection method, and / or the numerical differentiation method are used to determine the target centroid position, the target centroid position can be more accurately obtained, the error of the target centroid position determination can be reduced, and the accuracy of the target centroid position determination can be improved.

[0117] In some embodiments, the peak value detection method can be used to set the signal peak value, and the corresponding target centroid position can be determined according to the peak value. The peak value can be a peak point or a valley point, which can be adjusted according to the scene and action, or the zero-crossing detection method can be used to set the zero point information, and the plurality of target centroid positions can be obtained according to the situation of the signal passing through the zero point, which is not limited.

[0118] In some other embodiments, multiple methods such as the peak detection method and the zero-crossing detection method can be combined to determine multiple target centroid positions, and no limitation is imposed thereon.

[0119] S407: The number of target centroid positions is taken as the stroke information of the swimming action within the swimming time range.

[0120] In the embodiments of the present disclosure, the number of target centroid positions can represent the corresponding strokes of the subject during the swimming process.

[0121] For example, when the target centroid position is determined by using the zero-crossing detection method, the number of target centroid positions can represent the number of times the subject slides to the zero point. By counting the number of target centroid positions and the time information, the stroke frequency of the subject during the swimming process can be obtained.

[0122] In the embodiments of the present disclosure, the number of target centroid positions determined within the swimming time range can be counted and taken as the stroke information of the swimming action, or the time information, distance information, and other information corresponding to the target centroid position can be collectively taken as the stroke information, and no limitation is imposed thereon.

[0123] In the embodiments, the initial trajectory information of the swimming action is processed by centroid to obtain the corrected trajectory information, so as to analyze the swimming action, which can reduce the analysis error of the swimming action and effectively improve the accuracy and analysis effect of the swimming action. The initial trajectory information is divided into multiple trajectory data segments, the centroid of each trajectory data segment is identified to obtain multiple centroid information corresponding to the multiple trajectory data segments, which can process the initial trajectory information by centroid according to the multiple trajectory data segments to obtain more objective and accurate trajectory information. The target centroid position is determined by using the peak detection method, the zero-crossing detection method, and / or the numerical difference method, which can more accurately obtain the target centroid position, reduce the error of the target centroid position, and further improve the accuracy of the target centroid position. The target centroid position is determined from the multiple centroid positions within the swimming time range, and the number of target centroid positions is taken as the stroke information of the swimming action within the swimming time range. The target signal centroid position is determined from the multiple signal centroid positions, and the number of target signal centroid positions is taken as the stroke information of the swimming action within the swimming time range, which can effectively improve the convenience of determining the stroke information and effectively guarantee the accuracy of determining the stroke information, realize the balance between the determination efficiency and the accuracy, and guarantee the applicability of the swimming action analysis method.

[0124] Figure 5 is a flowchart of a swimming action analysis method according to another embodiment of the present disclosure.

[0125] As Figure 5 shown, the swimming action analysis method comprises:

[0126] S501: determining that a swimming state is entered.

[0127] S502: obtaining an action sensing signal of a swimming action.

[0128] S503: obtaining initial trajectory information according to the action sensing signal.

[0129] S504: performing trajectory division on the initial trajectory information to obtain a plurality of trajectory data segments.

[0130] S505: respectively performing centroid recognition on the plurality of trajectory data segments to obtain a plurality of centroid information respectively corresponding to the plurality of trajectory data segments.

[0131] The description of S501-S505 can be specifically referred to the above embodiments, which will not be repeated here.

[0132] S506: fitting to obtain a posture feature of the swimming action within a swimming time range according to the plurality of centroid information.

[0133] The feature used to represent the action and posture of the measured person within the swimming time range can be referred to as a posture feature of the swimming action, and the posture feature can be, for example, an overall posture feature, an arm swing feature, a leg swing feature, etc., without limitation.

[0134] In the embodiments of the present disclosure, an information analysis model can be used to analyze and fit the overall posture feature, the arm swing feature, the leg swing feature, etc. of the measured person within the swimming time range from the plurality of centroid information, or the plurality of centroid information can be processed to extract information capable of representing the posture feature of the swimming action to analyze the posture feature, or any other possible way can be used to fit to obtain the posture feature of the swimming action within the swimming time range according to the plurality of centroid information, such as artificial intelligence, mathematical algorithm, etc., without limitation.

[0135] S507: determining a posture feature matched with the reference posture feature, wherein the reference posture feature is a feature of a reference swimming posture.

[0136] The pre-set feature of the swimming posture used as a reference can be referred to as a reference posture feature, and the reference posture feature can be used to represent the feature of a standard swimming posture, or it can be a pre-set swimming posture feature corresponding to the body shape of the measured person, without limitation.

[0137] In the embodiments of the present disclosure, the reference swimming posture can be a plurality of standard swimming postures, and thus the reference posture feature can be, for example, a posture feature corresponding to a standard butterfly swimming posture, a posture feature corresponding to a standard breaststroke, and the like, without limitation.

[0138] In the embodiments of the present disclosure, the image information of the subject in the swimming process can also be recognized according to an image recognition manner, and the corresponding posture feature is analyzed and obtained, and is matched and recognized with the plurality of reference posture features set in advance, and the corresponding reference posture feature is matched and obtained, or the reference posture feature corresponding to the swimming posture information can also be set as reference data according to data processing, and a judgment standard is set, and when the data corresponding to the posture feature of the subject in the swimming process is detected to satisfy the judgment standard, the reference posture feature matched with the posture feature is further judged, and the like, without limitation.

[0139] For example, the data information (angle information, amplitude information, and the like) corresponding to the arm swing, leg swing, and the like of a plurality of standard swimming postures is recorded in advance, and is recorded in the swimming posture judgment model as data information corresponding to the reference posture feature of the swimming action. In the swimming process of the subject, the action sensing signal is obtained, and the arm swing amplitude, angle, and the like of the subject are analyzed from the action sensing signal and are input into the swimming posture judgment model, and are compared with the plurality of reference posture features, and the reference posture feature with the highest similarity is selected as the matched reference posture feature.

[0140] Optionally, the initial trajectory information is frame-processed according to the plurality of mass center information to obtain a plurality of trajectory data frames, a frame shift between adjacent trajectory data frames is a target frame shift, and the posture feature of the swimming action is analyzed according to the action division information and the plurality of trajectory data frames. Since the initial trajectory information is frame-processed to obtain the plurality of trajectory data frames, the posture feature of the swimming action is analyzed according to the action division information and the plurality of trajectory data frames, which can effectively improve the accuracy of the posture feature recognition and analysis of the swimming action. At the same time, the frame shift between adjacent trajectory data frames is used as the target frame shift, which can effectively and accurately analyze and process the initial trajectory information frame by frame, effectively improve the efficiency of the frame-by-frame analysis and processing, and effectively improve the referenceability of the obtained posture feature of the swimming action.

[0141] In the embodiments of the present disclosure, according to the action segment information and the plurality of trajectory data frames, the posture feature of the swimming action can be analyzed. The posture feature of the swimming action can be analyzed and determined based on a pre-trained posture determination model by inputting the action segment information and the plurality of trajectory data frames into the posture determination model. Alternatively, the posture feature can be determined by using a decision tree method to perform decision processing on the action segment information and the plurality of trajectory data frames. Alternatively, the posture feature of the swimming action can be analyzed by using a multilayer perceptron, a random forest, or the like. No limitation is imposed.

[0142] For example, a decision tree model is set, and a plurality of decision schemes for recognizing swimming postures are pre-set. The effective data information carried in the action segment information and the plurality of trajectory data frames is analyzed. The corresponding analysis result is obtained based on the effective data information. The posture feature of the swimming action is determined based on the analysis result.

[0143] In the embodiments of the present disclosure, the data signal collected by the sensor can be segmented by data framing to obtain a plurality of trajectory data frames. Alternatively, the derived data collected by the sensor can be segmented by data framing to obtain a plurality of trajectory data frames. Alternatively, the data signal collected by the sensor can be preprocessed, and the processed data signal can be segmented to obtain a plurality of trajectory data frames. No limitation is imposed.

[0144] The method of data segmentation can be fixed-length segmentation or non-fixed-length segmentation. No limitation is imposed.

[0145] For example, it is assumed that the frame length of the trajectory data frame can be L, and the target frame shift can be K. The target frame shift K can be determined according to the real-time and accuracy requirements of the swimming posture determination. No limitation is imposed.

[0146] Optionally, in some embodiments, the action segment information and the plurality of trajectory data frames are input into a pre-trained posture determination model to obtain the posture feature output by the posture determination model. The posture feature of the swimming action is determined by building the pre-trained posture determination model, which can effectively improve the efficiency and accuracy of the posture determination and effectively improve the determination effect of the posture feature of the swimming action.

[0147] In the embodiments of the present disclosure, the deep learning model can be built according to the action stroke information and the trajectory data frame corresponding to different swimming postures, and the deep learning model can be pre-trained to obtain a posture determination model for outputting posture features. The posture determination model can be built by collecting trajectory data frames of multiple known posture features as samples in advance, inputting the samples into the deep learning model, and iteratively training the deep learning model. When the deep learning model converges, the trained deep learning model is used as the posture determination model. Thus, the posture features of the swimming action are determined according to the pre-trained posture determination model, and no limitation is made to this.

[0148] S508: The reference swimming posture is used as the swimming posture information.

[0149] In the embodiments of the present disclosure, since the reference posture feature is the feature of the reference swimming posture, the reference posture feature can be obtained by pre-analyzing the reference swimming posture. After the reference posture feature matching the posture feature is determined, the reference swimming posture to which the reference posture feature belongs can be directly used as the swimming posture information, or the reference swimming posture can be subjected to corresponding guidance and correction processing, and the swimming posture after the guidance and correction is used as the swimming posture information, and no limitation is made to this.

[0150] For example, when it is detected that the posture feature of the testee in the swimming process matches the reference posture feature corresponding to the butterfly stroke, the reference swimming posture corresponding to the butterfly stroke can be used as the swimming posture information, that is, it is indicated that the testee is currently performing the butterfly stroke.

[0151] In the embodiment, the initial trajectory information of the swimming action is processed by the center of mass to obtain corrected trajectory information for analyzing the swimming action, which can reduce the error of the swimming action analysis and effectively improve the accuracy and analysis effect of the swimming action analysis. The initial trajectory information is divided into multiple trajectory data segments, the multiple trajectory data segments are respectively identified by the center of mass to obtain multiple center of mass information corresponding to the multiple trajectory data segments, the initial trajectory information can be processed by the center of mass according to the multiple trajectory data segments to obtain more objective and accurate trajectory information. The posture feature of the swimming action in the swimming time range is fitted according to the multiple center of mass information, the reference posture feature of the posture feature matching is determined, and the reference swimming posture is used as the swimming posture information, which can accurately and quickly identify the corresponding swimming posture of the swimming action, and further improve the efficiency and accuracy of the swimming posture determination. The initial trajectory information is processed by frame to obtain multiple trajectory data frames, the posture feature of the swimming action is analyzed according to the action division information and the multiple trajectory data frames, which can effectively improve the accuracy of the posture feature identification and analysis of the swimming action. At the same time, the frame shift between adjacent trajectory data frames is used as the target frame shift, which can effectively and accurately analyze the initial trajectory information frame by frame, effectively improve the efficiency of the frame-by-frame analysis processing, and effectively improve the referenceability of the obtained posture feature of the swimming action. The posture determination model is pre-trained to determine the posture feature of the swimming action, which can effectively improve the efficiency and accuracy of the posture determination, and effectively improve the determination effect of the posture feature of the swimming action.

[0152] Figure 6 is a flowchart of a swimming action analysis method according to another embodiment of the present disclosure.

[0153] As shown in Figure 6 , the swimming action analysis method comprises:

[0154] S601: determining whether to enter a swimming state.

[0155] S602: obtaining an action sensing signal of a swimming action.

[0156] S603: obtaining initial trajectory information according to the action sensing signal.

[0157] S604: dividing the initial trajectory information into multiple trajectory data segments.

[0158] S605: respectively identifying the multiple trajectory data segments by the center of mass to obtain multiple center of mass information corresponding to the multiple trajectory data segments.

[0159] The description of S601-S605 can be specifically referred to the above-mentioned embodiments, which will not be repeated here.

[0160] S606: Determine the swimming turn information of the swimming action in the swimming time range according to the position change information and the centroid point distance.

[0161] In the embodiments of the present disclosure, the centroid point distance and the position change information are taken as the centroid information, and the swimming turn information of the swimming action in the swimming time range is determined according to the position change information and the centroid point distance.

[0162] In the embodiments of the present disclosure, the turn position, the number of laps, the distance accumulation value and other swimming turn information can be determined according to the position change information and the centroid point distance by methods such as decision tree method, random forest method, or the turn position, the number of laps, the distance accumulation value and other swimming turn information can also be determined by methods such as building a convolutional neural network (CNN) model, a long short-term memory (LSTM) model, by analyzing and processing the position change information and the centroid point distance, and no limitation is made to this.

[0163] For example, when the position change information and the centroid point distance are the same, it can be indicated that the measured person is performing a straight-line motion and no turn action occurs, and when the position change information is significantly greater than the centroid point distance, it can be indicated that a turn action occurs in the range, so as to determine the turn position, the number of laps, the distance accumulation value and other swimming turn information corresponding to the route.

[0164] In the embodiments, the initial trajectory information of the swimming action is processed by centroid to obtain the corrected trajectory information, so as to analyze the swimming action, which can reduce the swimming action analysis error and effectively improve the accuracy and analysis effect of the swimming action analysis. The initial trajectory information is divided into multiple trajectory data segments, the centroid of each trajectory data segment is recognized to obtain multiple centroid information corresponding to the multiple trajectory data segments, which can process the initial trajectory information by centroid according to the multiple trajectory data segments to obtain more objective and accurate trajectory information. The swimming turn information of the swimming action in the swimming time range is determined by analyzing the position change information and the centroid point distance, which can make the judgment of the swimming turn more accurate, improve the accuracy of the data of the swimming turn position and the swimming distance, and effectively improve the accuracy of the swimming action analysis.

[0165] In summary, in some embodiments, as shown in Figure 7 Figure 7 ​is a swimming action analysis flowchart proposed by another embodiment of the present disclosure. First, swimming signals are collected using multiple sensors, swimming data is obtained, and the swimming data is preprocessed to obtain action sensing signals. Then, stroke detection and turning judgment are performed, and it is judged whether the current state of the measured person is a swimming state. If it is a swimming state, the data obtained by stroke detection is subjected to stroke confirmation and swimming posture recognition, action stroke information and swimming posture information are generated, and turning confirmation is performed to generate swimming turning information. The action stroke information, swimming posture information, and swimming turning information obtained are post-processed to generate corresponding analysis results.

[0166] In some other embodiments, as shown in Figure 8 , Figure 8 is a swimming action analysis flowchart proposed by another embodiment of the present disclosure. First, swimming signals are collected using multiple sensors, swimming data is obtained, and the swimming data is preprocessed to obtain action sensing signals. Then, data framing of the action sensing signals and stroke detection are performed. Stroke detection generates action stroke information. The action sensing signals are subjected to data framing processing to obtain multiple trajectory data frames. The trajectory data frames can be used for swimming posture recognition to generate swimming posture information. The trajectory data frames can also be used for calculation of center of mass points and center of mass point distances, and based on the center of mass point positions and center of mass point distances, turning detection is realized to generate swimming turning information. The action stroke information, swimming posture information, and swimming turning information obtained are post-processed to generate corresponding analysis results.

[0167] Figure 9 is a structural schematic diagram of a swimming action analysis device according to an embodiment of the present disclosure.

[0168] As shown in Figure 9 , the swimming action analysis device 90 comprises:

[0169] The judgment module 901 is configured to judge whether to enter a swimming state.

[0170] The first acquisition module 902 is configured to acquire action sensing signals of swimming actions.

[0171] The second acquisition module 903 is configured to obtain initial trajectory information according to the action sensing signals.

[0172] The processing module 904 is configured to perform center of mass processing on the initial trajectory information to obtain corrected trajectory information.

[0173] In some embodiments of the present disclosure, as shown in Figure 10 , Figure 10 is a structural schematic diagram of a swimming action analysis device according to another embodiment of the present disclosure. The processing module 904 comprises:

[0174] The dividing sub-module 9041 is configured to divide the initial trajectory information to obtain a plurality of trajectory data segments.

[0175] The identifying sub-module 9042 is configured to identify a plurality of centroids respectively corresponding to the plurality of trajectory data segments.

[0176] The processing sub-module 9043 is configured to process the plurality of centroid information to obtain the corrected trajectory information.

[0177] In some embodiments of the present disclosure, as shown in Figure 10 the identifying sub-module 9042 is specifically configured to:

[0178] determine a plurality of centroid points respectively corresponding to the plurality of trajectory data segments;

[0179] determine a plurality of centroid point positions respectively corresponding to the plurality of centroid points;

[0180] determine a centroid point distance between adjacent centroid point positions, wherein the adjacent centroid point positions indicate that a first centroid point position and a second centroid point position with a time domain distance of N points are adjacent, and N is greater than or equal to 1;

[0181] determine position change information between the plurality of centroid point positions;

[0182] use the plurality of centroid point positions, and / or the centroid point distance, and / or the position change information as the plurality of centroid information.

[0183] In some embodiments of the present disclosure, as shown in Figure 10 the dividing sub-module 9041 is specifically configured to:

[0184] cut off the same number of data points each time, and form a trajectory data segment according to the same number of data points each time; or

[0185] divide the initial trajectory information based on a peak point, and / or a trough point, and / or a zero point to obtain the plurality of trajectory data segments.

[0186] In some embodiments of the present disclosure, as shown in Figure 10 the processing sub-module 9043 is specifically configured to:

[0187] determine, according to the plurality of centroid information, action stroke information, and / or swimming turn information, and / or swimming posture information of the swimming action within a swimming time range;

[0188] use the action stroke information, and / or the swimming turn information, and / or the swimming posture information as the corrected trajectory information.

[0189] In some embodiments of the present disclosure, as shown in Figure 10 The determining module 901 is specifically configured to:

[0190] determine a swimming state of the swimming motion according to the action sensing signal; or

[0191] determine the swimming state according to the software instruction.

[0192] If the swimming state satisfies a state condition, determine a swimming time range of the swimming motion.

[0193] The first obtaining module 902 is specifically configured to:

[0194] obtain the action sensing signal of the swimming motion in the swimming time range.

[0195] In some embodiments of the present disclosure, as shown in Figure 10 The determining module 901 is specifically configured to:

[0196] When the swimming state satisfies the state condition, determine whether an action scene of the swimming motion is an underwater scene.

[0197] determine action continuity of the swimming motion according to the action sensing signal.

[0198] When the action scene is the underwater scene, and / or the action continuity satisfies a continuity condition, determine a swimming time range of the swimming motion.

[0199] In some embodiments of the present disclosure, as shown in Figure 10 The center of mass information is a center of mass point position.

[0200] The processing submodule 9043 is specifically configured to:

[0201] determine a target center of mass position from the multiple center of mass positions in the swimming time range.

[0202] take the number of the target center of mass positions as action stroke information of the swimming motion in the swimming time range.

[0203] In some embodiments of the present disclosure, as shown in Figure 10 The processing submodule 9043 is specifically configured to:

[0204] determine the target center of mass position from the multiple center of mass positions in the swimming time range by using a peak detection method, and / or a zero-crossing point detection method, and / or a numerical differentiation method.

[0205] In some embodiments of the present disclosure, as shown in Figure 10 The processing submodule 9043 is specifically configured to:

[0206] According to the plurality of center of mass information, posture features of the swimming action in the swimming time range are fitted;

[0207] The reference posture feature matched with the posture feature is determined, wherein the reference posture feature is a feature of a reference swimming posture;

[0208] The reference swimming posture is taken as the swimming posture information.

[0209] In some embodiments of the present disclosure, as shown in Figure 10 The processing submodule 9043 is specifically configured to:

[0210] According to the plurality of center of mass information, the initial trajectory information is frame-processed to obtain a plurality of trajectory data frames, wherein the frame shift between adjacent trajectory data frames is the target frame shift;

[0211] According to the action stroke information and the plurality of trajectory data frames, posture features of the swimming action are analyzed.

[0212] In some embodiments of the present disclosure, as shown in Figure 10 The processing submodule 9043 is specifically configured to:

[0213] The action stroke information and the plurality of trajectory data frames are input into a pre-trained posture judgment model to obtain posture features output by the posture judgment model.

[0214] In some embodiments of the present disclosure, as shown in Figure 10 The center of mass information is center of mass point distance and position change information;

[0215] The processing submodule 9043 is specifically configured to:

[0216] According to the position change information and the center of mass point distance, swimming turn information of the swimming action in the swimming time range is determined.

[0217] In some embodiments of the present disclosure, as shown in Figure 1 to Figure 8 The first analysis module 905 is configured to analyze the swimming action according to the action stroke information, and / or the swimming turn information, and / or the swimming posture information;

[0218] The first analysis module 905 is configured to analyze the swimming action according to the action stroke information, and / or the swimming turn information, and / or the swimming posture information;

[0219] The second analysis module 906 is configured to analyze the action stroke information, the swimming posture information, and the swimming turn information to obtain a current swimming lap number, and / or a swimming stroke number, and / or a swimming distance.

[0220] The third analysis module 907 is configured to analyze the stroke information, the swimming posture information, and the swimming turn information to obtain each stroke, and / or stroke in a set time range, and / or total stroke, and / or stroke frequency, and / or swimming posture statistical information, wherein the swimming posture statistical information comprises main swimming posture information and / or mixed swimming posture information.

[0221] The swimming action analysis device provided in the embodiments of the present disclosure corresponds to the swimming action analysis method provided in the embodiments of the present disclosure. Figure 1 to Figure 8 The swimming action analysis device provided in the embodiments of the present disclosure corresponds to the swimming action analysis method provided in the embodiments of the present disclosure. Figure 11 The swimming action analysis device provided in the embodiments of the present disclosure corresponds to the swimming action analysis method provided in the embodiments of the present disclosure.

[0222] In the embodiment, the action sensing signal of the swimming action is obtained by judging the entering of the swimming state, the initial trajectory information is obtained according to the action sensing signal, and the corrected trajectory information is obtained by the centroid processing on the initial trajectory information. Since the corrected trajectory information is obtained by the centroid processing on the initial trajectory information of the swimming action, the swimming action analysis error can be reduced, and the accuracy and analysis effect of the swimming action analysis can be effectively improved.

[0223] In order to implement the above-mentioned embodiments, the present disclosure further provides a non-transitory computer readable storage medium having a computer program stored thereon, and the program is executed by a processor to implement the swimming action analysis method provided in the foregoing embodiments of the present disclosure.

[0224] In order to implement the above-mentioned embodiments, the present disclosure further provides a wearable device, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor executes the program to implement the swimming action analysis method provided in the foregoing embodiments of the present disclosure.

[0225] In order to implement the above-mentioned embodiments, the present disclosure further provides a computer program product, and when the instructions in the computer program product are executed by a processor, the swimming action analysis method provided in the foregoing embodiments of the present disclosure is executed.

[0226] Figure 11 A block diagram of an exemplary wearable device suitable for implementing embodiments of the present disclosure is shown. Figure 11 The displayed wearable device 12 is only an example, and should not bring any limitation to the functions and use range of the embodiments of the present disclosure.

[0227] As Figure 11As shown, the wearable device 12 is in the form of a general- purpose computing device. The components of the wearable device 12 can include, but are not limited to, one or more processors or processing units 16, a system memory 28, and a bus 18 that couples various system components including the system memory 28 to the processing unit 16. The bus 18 represents one or more of any of several bus structures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration bus, a processor or local bus using any of a variety of bus architectures, etc. Such architectures include, for example, Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MCA) bus, Enhanced ISA bus, Video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnect (PCI) bus.

[0228] The wearable device 12 typically includes a variety of computer system readable media. Such media can be any available media that is accessible by the wearable device 12 and includes both volatile and non- volatile media, removable and non-removable media.

[0229] The memory 28 can include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache memory 32. The wearable device 12 can further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, a storage system 34 can be provided for reading from and writing to non-removable, non-volatile magnetic media (e.g., a "hard drive"). Figure 11 Not shown is a generally-rectangularly-shaped "hard drive".

[0230] Although ​A disk drive, a floppy disk drive, a CD-ROM drive, a DVD-ROM drive, etc., can be provided for reading from or writing to a removable nonvolatile magnetic media (e.g., a "floppy disk"), and to a removable nonvolatile optical disk (e.g., a CD-ROM, a DVD-ROM, or other optical media). In such cases, each will be connected to the bus 18 by one or more data media interfaces. The memory 28 can include at least one program product having a set (e.g., at least one) of program modules that are configured to carry out the functions of embodiments of the disclosure.

[0231] The program / utility 40, having a set (at least one) of program modules 42, can be stored in memory 28 by way of example, and not limitation, as well as an operating system, one or more application programs, other program modules, and program data, each or some combination thereof, can include implementation of a networking environment. The program modules 42 generally carry out the functions and / or methodologies of embodiments of the disclosure as described herein.

[0232] The wearable device 12 can also communicate with one or more external devices 14 such as a keyboard or a pointing device, a display 24, etc. and also with one or more devices that enable a user to interact with the wearable device 12 and / or any devices (e.g., a networking card, a modem, etc.) that enable the wearable device 12 to communicate with one or more other computing devices. Such communication can occur via the input / output (I / O) interface 22. Still yet, the wearable device 12 can communicate with one or more networks such as a local area network (LAN), a wide area network (WAN), and / or the Internet through a network adapter 20. As presented, the network adapter 20 communicates with the other

[0233] The processing unit 16 performs various function applications and data processing by running programs stored in the system memory 28, such as implementing the swimming movement analysis method mentioned in the foregoing embodiments.

[0234] Those skilled in the art will readily understand that the disclosure is well adapted to carry out the objects and obtain the ends and advantages mentioned, as well as those inherent therein, and that the embodiments described herein are illustrative of the disclosure and are not intended to limit the scope of the disclosure. Accordingly, the disclosure is not to be restricted except in the spirit of the appended claims.

[0235] It should be understood that the present disclosure is not limited to the precise structures herein described and illustrated in the drawings, and that various modifications and changes can be made therein without departing from the scope of this present disclosure. The scope of the present disclosure is limited only by the claims that follow.

[0236] It should be noted that in the description of the present disclosure, the terms "first", "second", etc. are used only for descriptive purposes and are not to be construed as indicating or implying relative importance. In addition, in the description of the present disclosure, the meaning of "a plurality of" is two or more, unless otherwise specified.

[0237] Any process or method descriptions or any other descriptions herein can be understood as representing embodiments of the present disclosure encompassing any tangible part of a process, method, or other procedures, implemented as code, for example, and / or implemented via any combination of hardware and / or software, whether preexisting or being developed in the future. The embodiments of the present disclosure contemplate any and all such future techniques and technologies.

[0238] It should be understood that various parts of the present disclosure can be implemented in hardware, software, firmware, or a combination thereof. In the above-described embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any of the following technologies known in the art or their combinations can be used: discrete logic circuit with logic gates for implementing logical functions on data signals, application specific integrated circuit with appropriate combination logic gates, programmable gate array (PGA), field programmable gate array (FPGA), etc.

[0239] Those skilled in the art can understand that all or part of the steps carried out by the above-mentioned embodiments can be completed by programs instructing relevant hardware, and the programs can be stored in a computer readable storage medium, which includes one or a combination of steps of the method embodiments when executed.

[0240] In addition, each functional unit in each embodiment of the present disclosure can be integrated in one processing module, or each unit can be physically present separately, or two or more units can be integrated in one module. The integrated module can be realized in the form of hardware or in the form of a software functional module. When the integrated module is realized in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer readable storage medium.

[0241] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.

[0242] In the description of the present specification, the description referring to the terms "one embodiment", "some embodiments", "an example", "a specific example" or "some examples" etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present disclosure. In the present specification, the illustrative description of the above terms does not necessarily mean the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0243] Although the embodiments of the present disclosure have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present disclosure, and those of ordinary skill in the art can make changes, modifications, replacements and variations to the above embodiments within the scope of the present disclosure.

Claims

1. A method for analyzing swimming motions, characterized in that, include: Determine if you have entered a swimming state; Acquire motion sensing signals of swimming movements; Based on the motion sensing signals, initial trajectory information is obtained; The initial trajectory information is divided into multiple trajectory data segments. Determine multiple centroids corresponding to the multiple trajectory data segments respectively; Determine the positions of multiple centroids corresponding to the multiple centroids; Determine the centroid distance between adjacent centroid positions, wherein the adjacent centroid positions indicate that the first centroid position and the second centroid position, which are N points away from it in the time domain, are adjacent, and N is greater than or equal to 1; Determine the positional change information between the multiple centroids; The positions of the multiple centroids, and / or the distances between the centroids, and / or the position change information are used as multiple centroid information corresponding to the multiple trajectory data segments respectively; The multiple centroid information is processed to obtain the corrected trajectory information.

2. The method as described in claim 1, characterized in that, The initial trajectory information is divided into multiple trajectory data segments, including: Each time, the same number of data points are extracted from the initial trajectory information, and the trajectory data segment is formed based on the same number of data points extracted each time; or The initial trajectory information is divided into multiple trajectory data segments based on segmentation based on peak points, and / or segmentation based on trough points, and / or segmentation based on zero points.

3. The method as described in claim 1, characterized in that, The centroid processing of the plurality of centroid information to obtain the corrected trajectory information includes: Based on the multiple centroid information, determine the stroke count information, and / or swimming turn information, and / or swimming posture information within the swimming time range; The stroke sequence information, and / or the swimming turn information, and / or the swimming posture information are used as the corrected trajectory information.

4. The method as described in claim 3, characterized in that, The determination of entering a swimming state includes: Based on the motion sensing signal, the swimming state of entering the swimming motion is determined; or The software determines whether to enter swimming mode based on instructions. If the swimming state meets the state conditions, then the swimming time range of the swimming action is determined; The acquisition of motion sensing signals for swimming movements includes: Acquire motion sensing signals of the swimming motion within the swimming time range.

5. The method as described in claim 4, characterized in that, If the swimming state meets the state conditions, then determining the swimming time range of the swimming action includes: If the swimming state meets the state condition, then determine whether the swimming action is an underwater scene; The continuity of the swimming motion is determined based on the motion sensing signal; If the action scenario is an underwater scenario, and / or the continuity of the action satisfies the continuity condition, then the swimming time range of the swimming action is determined.

6. The method as described in claim 3, characterized in that, The centroid information is the position of the centroid point; The step of determining the stroke sequence information of swimming movements within the swimming time range based on the multiple centroid information includes: The target centroid position is determined from the plurality of centroid positions within the swimming time range; The number of target centroid positions is used as the stroke count information for the swimming action within the swimming time range.

7. The method as described in claim 6, characterized in that, Determining the target centroid position from the plurality of centroid positions within the swimming time range includes: The target centroid position is determined from the plurality of centroid positions within the swimming time range using a peak detection method, and / or a zero-crossing detection method, and / or a numerical difference method.

8. The method as described in claim 6, characterized in that, in, The step of determining the swimming posture information within the swimming time range based on the multiple centroid information includes: Based on the multiple centroid information, the posture characteristics of the swimming motion within the swimming time range are obtained by fitting. Determine a reference posture feature for matching the posture feature, wherein the reference posture feature is a feature of a reference swimming posture; The reference swimming posture is used as the swimming posture information.

9. The method as described in claim 8, characterized in that, The step of fitting the posture features of the swimming motion within the swimming time range based on the multiple centroid information includes: The initial trajectory information is segmented into frames based on the multiple centroid information to obtain multiple trajectory data frames, wherein the frame shift between adjacent trajectory data frames is the target frame shift. The posture characteristics of the swimming motion are analyzed based on the stroke count information and the multiple trajectory data frames.

10. The method as described in claim 9, characterized in that, The step of analyzing the posture characteristics of the swimming motion based on the motion stroke information and the multiple trajectory data frames includes: The action stroke information and the multiple trajectory data frames are input into a pre-trained pose determination model to obtain the pose features output by the pose determination model.

11. The method as described in claim 3, characterized in that, The centroid information includes the distance to the centroid point and the position change information. The step of determining the swimming turn information within the swimming time range based on the multiple centroid information includes: Based on the position change information and the distance to the center of mass, the swimming turn information within the swimming time range is determined.

12. The method as described in claim 3, characterized in that, Also includes: The swimming motion is analyzed based on the stroke count information, and / or the swimming turn information, and / or the swimming posture information; By analyzing the stroke count information, swimming posture information, and swimming turn information, the current swimming distance, and / or swimming laps, and / or swimming distance can be obtained. Analyze the stroke count information, swimming posture information, and swimming turn information to obtain the stroke count per stroke, and / or stroke count within a set time range, and / or total stroke count, and / or stroke frequency, and / or swimming posture statistics, wherein the swimming posture statistics include: primary swimming posture information, and / or mixed swimming posture information.

13. A swimming motion analysis device, characterized in that, include: The judgment module is used to determine whether the user has entered the swimming state. The first acquisition module is used to acquire motion sensing signals of swimming movements; The second acquisition module is used to obtain initial trajectory information based on the motion sensing signal; The processing module is used to perform centroid processing on the initial trajectory information to obtain corrected trajectory information; The processing module includes: A segmentation submodule is used to segment the initial trajectory information to obtain multiple trajectory data segments; The identification submodule is used to identify the centroids of the multiple trajectory data segments respectively, so as to obtain multiple centroid information corresponding to the multiple trajectory data segments respectively; The processing submodule is used to perform centroid processing on the multiple centroid information to obtain the corrected trajectory information; The identification submodule is specifically used for: Determine multiple centroids corresponding to the multiple trajectory data segments respectively; Determine the positions of multiple centroids corresponding to the multiple centroids; Determine the centroid distance between adjacent centroid positions, wherein the adjacent centroid positions indicate that the first centroid position and the second centroid position, which are N points away from it in the time domain, are adjacent, and N is greater than or equal to 1; Determine the positional change information between the multiple centroids; The positions of the plurality of centroids, and / or the distances between the centroids, and / or the position change information are used as the plurality of centroid information.

14. The apparatus as claimed in claim 13, characterized in that, The processing submodule is specifically used for: Based on the multiple centroid information, determine the stroke count information, and / or swimming turn information, and / or swimming posture information within the swimming time range; The stroke sequence information, and / or the swimming turn information, and / or the swimming posture information are used as the corrected trajectory information.

15. The apparatus as claimed in claim 14, characterized in that, The judgment module is specifically used for: Based on the motion sensing signal, the swimming state of entering the swimming motion is determined; or The software determines whether to enter swimming mode based on instructions. If the swimming state meets the state conditions, then the swimming time range of the swimming action is determined; The first acquisition module is specifically used for: Acquire motion sensing signals of the swimming motion within the swimming time range.

16. The apparatus as claimed in claim 14, characterized in that, The centroid information is the position of the centroid point; Specifically, the processing submodule is used for: The target centroid position is determined from the plurality of centroid positions within the swimming time range; The number of target centroid positions is used as the stroke count information for the swimming action within the swimming time range.

17. The apparatus as claimed in claim 16, characterized in that, in, The processing submodule is specifically used for: Based on the multiple centroid information, the posture characteristics of the swimming motion within the swimming time range are obtained by fitting. Determine a reference posture feature for matching the posture feature, wherein the reference posture feature is a feature of a reference swimming posture; The reference swimming posture is used as the swimming posture information; The processing submodule is specifically used for: The initial trajectory information is segmented into frames based on the multiple centroid information to obtain multiple trajectory data frames, wherein the frame shift between adjacent trajectory data frames is the target frame shift. The posture characteristics of the swimming motion are analyzed based on the stroke count information and the multiple trajectory data frames.

18. The apparatus as claimed in claim 14, characterized in that, The centroid information includes the distance to the centroid point and the position change information. Specifically, the processing submodule is used for: Based on the position change information and the distance to the center of mass, the swimming turn information within the swimming time range is determined.

19. The apparatus as claimed in claim 14, characterized in that, Also includes: The first analysis module is used to analyze the swimming motion based on the stroke count information, and / or the swimming turn information, and / or the swimming posture information; The second analysis module is used to analyze the stroke information, swimming posture information, and swimming turn information to obtain the current swimming distance, and / or swimming laps, and / or swimming distance. The third analysis module is used to analyze the stroke count information, the swimming posture information, and the swimming turn information to obtain the stroke count per stroke, and / or the stroke count within a set time range, and / or the total number of strokes, and / or the stroke frequency, and / or swimming posture statistics, wherein the swimming posture statistics include: primary swimming posture information, and / or mixed swimming posture information.

20. A wearable device, comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the swimming motion analysis method according to any one of claims 1-12.

21. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to execute the swimming motion analysis method according to any one of claims 1-12.

Images