Motion monitoring method and device, electronic equipment and storage medium

Abstract

The invention relates to a motion monitoring method and device, electronic equipment and a storage medium. The motion monitoring method comprises the steps that an original motion data set in the motion process is acquired; based on each original motion data set, determining a motion period corresponding to each original motion data set and a target data set corresponding to each motion period; obtaining a classification result of the action corresponding to each motion period; determining similarity information of the action corresponding to each motion period and the target action; and based on the classification result and the similarity information, determining an identification result of the target action in the motion process. According to the invention, the classification result is obtained by classifying the motion period, the similarity information is calculated according to the standard motion data set, and the target action is identified and judged by integrating the classification result and the similarity information, so that the monitoring precision of motion monitoring is ensured, the error rate of the identification result is reduced, and the user experience is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of motion monitoring, and in particular to a motion monitoring method, device, electronic equipment and storage medium. Background Art

[0002] In recent years, with the rapid development of smart devices, wearable devices of different forms, such as smart watches and smart bracelets, have been widely used in people's daily lives. Users can monitor their health or exercise status in real time by wearing these devices. In order to meet the experience and needs of exercise participants, it is usually necessary to monitor and identify specific movements during exercise to intuitively reflect the amount of exercise and exercise effect.

[0003] However, the use of related motion monitoring methods to monitor and identify specific actions has problems such as low monitoring accuracy and large errors in identification results, resulting in poor user experience. Summary of the invention

[0004] In order to overcome the problems existing in the related art, the present disclosure provides a motion monitoring method, device, electronic device and storage medium.

[0005] According to a first aspect of an embodiment of the present disclosure, a motion monitoring method is provided, which is applied to an electronic device. The motion monitoring method includes:

[0006] Acquire at least one original motion data set during the motion process, wherein the original motion data set includes original motion data representing a motion direction and / or a motion speed;

[0007] Based on each of the original motion data sets, determining at least one motion cycle corresponding to each of the original motion data sets and a target data set corresponding to each of the motion cycles;

[0008] Classifying the motion cycles corresponding to the target data sets of each of the original motion data sets to obtain classification results of actions corresponding to each of the motion cycles, wherein the classification results represent the probability that the action corresponding to the motion cycle is the target action;

[0009] Determining similarity information between the action corresponding to each of the motion cycles and the target action based on each of the target data sets and a standard motion data set corresponding to the target action;

[0010] Based on the classification result and the similarity information, a recognition result of the target action in the motion process is determined.

[0011] In some embodiments of the present disclosure, determining, based on each of the original motion data sets, at least one motion cycle corresponding to each of the original motion data sets and a target data set corresponding to each of the motion cycles includes:

[0012] The following determination process is performed for each of the original motion data sets:

[0013] Performing data preprocessing on the original motion data set to obtain a preprocessed data set;

[0014] determining at least one of the motion cycles based on the preprocessed data set;

[0015] Based on each of the motion cycles, determining a first data set corresponding to each of the motion cycles, wherein the first data set includes first target data representing feature information corresponding to the motion cycle;

[0016] Based on each of the motion cycles, a second data set corresponding to each of the motion cycles is determined, wherein the second data set includes second target data representing the action information within the corresponding motion cycle.

[0017] In some embodiments of the present disclosure, determining at least one of the motion cycles based on the pre-processed data set includes:

[0018] Based on the preprocessed data set, determining a set of data to be detected, wherein each set of data to be detected in the set of data to be detected corresponds one-to-one to a data acquisition moment of at least part of the original motion data during the motion process;

[0019] Based on the data set to be detected, determining at least one peak data;

[0020] Based on each peak data and the data acquisition time corresponding to each peak data, the motion cycle corresponding to each peak data is determined.

[0021] In some embodiments of the present disclosure, determining at least one peak data based on the to-be-detected data set includes:

[0022] The data to be detected that meets the preset detection conditions is determined as the peak data, and the preset detection conditions include:

[0023] is greater than a first threshold;

[0024] Greater than or equal to the previous data to be tested and greater than the next data to be tested;

[0025] The difference between the corresponding data acquisition time and the data acquisition time corresponding to the previous peak data is greater than the second threshold.

[0026] In some embodiments of the present disclosure, the pre-processed data set includes a plurality of three-axis motion data groups, and determining the data set to be detected based on the pre-processed data set includes:

[0027] The composite amount of each of the three-axis motion data groups in the pre-processed data set is calculated, and the composite amount is used as the data to be detected at the corresponding data acquisition time.

[0028] In some embodiments of the present disclosure, determining, based on each of the motion cycles, a first data set corresponding to each of the motion cycles includes:

[0029] The following determination process is performed for each of the motion cycles:

[0030] Determine at least one statistical characteristic value of the data to be detected within the motion cycle;

[0031] The peak data, the data acquisition time and the at least one statistical characteristic value corresponding to the motion cycle are used as the first data set corresponding to the motion cycle.

[0032] In some embodiments of the present disclosure, determining, based on each of the motion cycles, a second data set corresponding to each of the motion cycles includes:

[0033] The following determination process is performed for each of the motion cycles:

[0034] taking each of the to-be-detected data in the motion cycle as the second data set corresponding to the motion cycle; or,

[0035] The data of the preprocessed data set within the motion cycle is used as the second data set corresponding to the motion cycle.

[0036] In some embodiments of the present disclosure, the motion cycles corresponding to the target data sets of each of the original motion data sets are classified to obtain classification results of actions corresponding to each of the motion cycles, including:

[0037] Based on the first data sets corresponding to the original motion data sets and the classification model, the classification results of the actions corresponding to the motion cycles are determined, and the classification results include a first result and a second result. The first result represents that the probability that the action corresponding to the motion cycle is the target action is 1, and the second result represents that the probability that the action corresponding to the motion cycle is the target action is 0.

[0038] In some embodiments of the present disclosure, there are multiple original motion data sets, and determining the classification results of the actions corresponding to each motion cycle based on each of the first data sets corresponding to each of the original motion data sets and the classification model includes:

[0039] Merging the first data sets whose difference in the data acquisition time of the peak data between the original motion data sets is less than or equal to a third threshold to obtain a first combined data set;

[0040] Based on each of the first combined data sets and the classification model, a classification result of the action corresponding to each of the motion cycles is determined.

[0041] In some embodiments of the present disclosure, determining similarity information between the action corresponding to each of the motion cycles and the target action based on each of the target data sets and the standard motion data set corresponding to the target action includes:

[0042] Based on each of the second data sets corresponding to each of the original motion data sets and the standard motion data set, similarity information between the action corresponding to each of the motion cycles and the target action is determined.

[0043] In some embodiments of the present disclosure, determining similarity information between the action corresponding to each of the motion cycles and the target action based on each of the second data sets corresponding to each of the original motion data sets and the standard motion data set includes:

[0044] The following determination process is performed for each of the second data sets:

[0045] Determining a target matching distance between the second data set and the standard motion data set according to a dynamic time warping algorithm;

[0046] The target matching distance is used as the similarity information.

[0047] In some embodiments of the present disclosure, there are multiple original motion data sets, and determining similarity information between the action corresponding to each of the motion cycles and the target action based on each of the second data sets corresponding to each of the original motion data sets and the standard motion data set includes:

[0048] Merging the second data sets whose difference in the data acquisition time of the peak data between the original motion data sets is less than or equal to a fourth threshold to obtain a second combined data set;

[0049] Based on each of the second combined data sets and the standard motion data set, similarity information between the action corresponding to each of the motion cycles and the target action is determined.

[0050] In some embodiments of the present disclosure, there are multiple standard motion data sets, and determining the target matching distance between the second data set and the standard motion data set according to the dynamic time warping algorithm includes:

[0051] According to the dynamic time warping algorithm, respectively calculating the matching distance between the second data set and each of the standard motion data sets;

[0052] The smallest matching distance among the multiple matching distances is determined as the target matching distance.

[0053] In some embodiments of the present disclosure, the step of acquiring at least one set of original motion data during the motion process includes:

[0054] A plurality of the original motion data sets are acquired during the motion process, wherein different original motion data sets are detected by different sensors.

[0055] In some embodiments of the present disclosure, after determining the recognition result of the target action during the motion process, the motion monitoring method further includes:

[0056] Based on each of the target data sets and the target action recognition result, the force situation of the recognized target action is determined.

[0057] According to a second aspect of an embodiment of the present disclosure, a motion monitoring device is provided, which is applied to an electronic device, and the motion monitoring device includes:

[0058] An acquisition module, the acquisition module is used to acquire at least one original motion data set during the motion process, the original motion data set including original motion data representing the motion direction and / or motion speed;

[0059] A first determination module, the first determination module is used to determine, based on each of the original motion data sets, at least one motion cycle corresponding to each of the original motion data sets and a target data set corresponding to each of the motion cycles;

[0060] A second determination module, the second determination module is used to classify the motion cycles corresponding to the target data sets of each of the original motion data sets, and obtain classification results of actions corresponding to each of the motion cycles, wherein the classification results represent the probability that the action corresponding to the motion cycle is the target action;

[0061] a third determination module, the third determination module being used to determine similarity information between the action corresponding to each of the motion cycles and the target action based on each of the target data sets and a standard motion data set corresponding to the target action;

[0062] A fourth determination module is used to determine a recognition result of the target action in the motion process based on the classification result and the similarity information.

[0063] According to a third aspect of an embodiment of the present disclosure, an electronic device is provided, the electronic device comprising:

[0064] processor;

[0065] a memory for storing processor-executable instructions;

[0066] Wherein, the processor is configured to execute the motion monitoring method as described in the first aspect.

[0067] According to a fourth aspect of an embodiment of the present disclosure, a non-temporary computer-readable storage medium is provided. When instructions in the storage medium are executed by a processor of an electronic device, the electronic device can perform the motion monitoring method described in the first aspect.

[0068] The technical solution provided by the embodiments of the present disclosure may include the following beneficial effects: obtaining classification results by classifying motion cycles, calculating similarity information based on a standard motion data set, and identifying and judging target actions by combining the classification results and similarity information, thereby ensuring the monitoring accuracy of motion monitoring, reducing the error rate of recognition results, and improving user experience.

[0069] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0071] Figure 1 is a flow chart of a motion monitoring method according to an exemplary embodiment.

[0072] Figure 2 The present invention is a flowchart showing, based on each original motion data set, determining at least one motion cycle corresponding to each original motion data set and a target data set corresponding to each motion cycle according to an exemplary embodiment.

[0073] Figure 3 The present invention is a flowchart of performing data preprocessing on an original motion data set to obtain a preprocessed data set according to an exemplary embodiment.

[0074] Figure 4 The present invention is a flowchart of determining at least one motion cycle based on a pre-processed data set according to an exemplary embodiment.

[0075] Figure 5 The present invention is a flowchart of a process of determining a first data set corresponding to each motion cycle based on each motion cycle according to an exemplary embodiment.

[0076] Figure 6 It is a flowchart showing a method of determining a classification result of an action corresponding to each motion cycle based on each first data set corresponding to each original motion data set and a classification model according to an exemplary embodiment.

[0077] Figure 7 The present invention is a flowchart showing a method of determining similarity information between actions corresponding to each motion cycle and a target action based on each second data set and a standard motion data set corresponding to each original motion data set according to an exemplary embodiment.

[0078] Figure 8 It is a flowchart showing, according to another exemplary embodiment, determining similarity information between an action corresponding to each motion cycle and a target action based on each second data set corresponding to each original motion data set and a standard motion data set.

[0079] Fig. 9 It is a flowchart showing, according to an exemplary embodiment, determining a target matching distance between a second data set and a standard motion data set based on a dynamic time warping algorithm.

[0080] Fig.10 is a flow chart of a motion monitoring method according to another exemplary embodiment.

[0081] Fig.11 is a block diagram of a motion monitoring device according to an exemplary embodiment.

[0082] Fig.12 is a block diagram of an electronic device according to an exemplary embodiment.

[0083] In the figure:

[0084] 10-acquisition module; 20-first determination module; 30-second determination module; 40-third determination module; 50-fourth determination module; 101-processing component; 102-memory; 103-power component; 104-multimedia component; 105-audio component; 106-input / output interface; 107-sensor component; 108-communication component; 109-processor. DETAILED DESCRIPTION

[0085] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are merely examples of devices and methods consistent with some aspects of the present invention as detailed in the appended claims.

[0086] In recent years, with the rapid development of smart device-related hardware and human-computer interaction functions, wearable devices of different forms, such as smart bracelets and smart watches, have been widely used in people's daily lives. Users can monitor their health or exercise status in real time by wearing these wearable devices. As more and more people are involved in fitness exercises, wearable devices are equipped with exercise monitoring functions for exercise participants during exercise. In order to meet the experience and needs of exercise participants, it is usually necessary to detect and identify specific actions during exercise, such as throwing a frisbee, so as to intuitively reflect exercise results such as the number of frisbee throws.

[0087] In the related art, specific actions are identified by identifying the peak value of the original motion data, for example, the number of peak points in the collected acceleration data is used as the number of frisbee throws. However, the use of the related art motion monitoring method to monitor and identify specific actions is limited and interfered by the use scenario, and it is easy to identify actions in non-sports scenes or other actions during the motion process as specific actions such as frisbee throwing, resulting in low monitoring accuracy, large errors in recognition results, and poor user experience.

[0088] Based on this, the exemplary embodiment of the present disclosure provides a motion monitoring method, which obtains the original motion data set and determines the motion cycle and target data set corresponding to the original motion data set, and then classifies the motion cycle according to the target data set, and determines the classification result and similarity information corresponding to the motion cycle according to the target data set and the standard motion data set. The recognition result of the target action can be determined according to the classification result and similarity information. The target action is identified and judged by combining the classification result and similarity information, which ensures the monitoring accuracy of motion monitoring, reduces the error rate of the recognition result, and improves the user experience.

[0089] In an exemplary embodiment, a motion monitoring method is provided, which is applied to an electronic device, and the motion monitored may be, for example, throwing sports such as frisbee and discus, or other sports that require specific movements of the arms and wrists. The electronic device may be a wearable device, such as a smart watch, a smart bracelet, etc., or other devices with data processing capabilities connected to the wearable device, such as a mobile phone, a tablet computer, etc. connected to the wearable device. Figure 1 As shown, the motion monitoring method includes:

[0090] S100: Acquire at least one original motion data set during a motion process, where the original motion data set includes original motion data representing a motion direction and / or a motion speed.

[0091] In step S100, a raw data set in the motion process is obtained, and a raw motion data set can be formed by a certain sampling frequency or raw motion data at different times, and the sampling frequency can be, for example, 25 Hz. The raw data set can be one or more. When the raw data set is multiple, different raw data sets can include different types of raw motion data, such as data detected by different sensors. The raw motion data can characterize the direction of motion, or the raw motion data can characterize the speed of motion, or the raw motion data can characterize the direction of motion and the speed of motion. Exemplarily, the raw motion data can include, for example, speed data, acceleration data, angular velocity data, angular acceleration data, and corresponding direction data.

[0092] S200: Based on each original motion data set, determine at least one motion cycle corresponding to each original motion data set and a target data set corresponding to each motion cycle.

[0093] In step S200, for each original motion data set, the motion cycle corresponding to the original motion data set and the target data set corresponding to the motion cycle can be determined according to the original motion data set. Each original motion data set may correspond to one or more motion cycles, and the motion cycle can represent the time range of making an action, which may be a target action or a non-target action. The target data set is a data set obtained according to the original motion data set for subsequent identification, and each motion cycle corresponds to a target data set.

[0094] S300, classifying the motion cycles corresponding to the target data sets of the original motion data sets to obtain classification results of the actions corresponding to the motion cycles, wherein the classification results represent the probability that the actions corresponding to the motion cycles are the target actions.

[0095] In step S300, the motion cycle corresponding to the target data set of each original motion data set is classified, if the action corresponding to the motion cycle is more inclined to be the target action, then the classification result characterizes that the action corresponding to the motion cycle is the probability of the target action is larger, if the action corresponding to the motion cycle is inclined to be the non-target action, then the classification result characterizes that the action corresponding to the motion cycle is the probability of the target action is smaller. The target action corresponds to the action that needs to be monitored and identified, and illustratively, when monitoring and identifying the flying disc throwing action, the target action is the standard action of throwing the flying disc. For an original motion data set, the target data set corresponding to the original motion data set can be input into the classification model in turn, and the action classification result corresponding to each motion cycle is determined. The classification result, for example, can include two possible results, and when the motion cycle corresponding to an original motion data set is multiple, a classification result corresponding to it is determined for each motion cycle.

[0096] It should be noted that when there are multiple original motion data sets, the target data sets corresponding to the original motion data sets can be merged and integrated, and then the motion cycles can be classified according to the merged data sets to determine the classification results of the actions corresponding to each motion cycle. At this time, the classification results can be used as the classification results of the motion cycles corresponding to different original data sets at the same time.

[0097] S400 : Determine similarity information between the actions corresponding to each motion cycle and the target action based on each target data set and a standard motion data set corresponding to the target action.

[0098] In step S400, the standard motion data set includes the motion data corresponding to the target action, and the standard motion data set may be one or more. When the standard motion data set is multiple, different standard motion data sets may correspond to different target actions. The target action corresponds to the action that needs to be monitored and identified. For example, when monitoring and identifying the flying disc throwing action, the target action is the standard action of throwing the flying disc. The similarity information between the action corresponding to each motion cycle and the target action can be determined based on each target data set and the standard motion data set. The similarity information can be, for example, a numerical value representing the similarity between the action corresponding to the motion cycle and the target action.

[0099] It should be noted that when there are multiple original motion data sets, the target data sets corresponding to the original motion data sets can be merged and integrated, and then the similarity information can be determined based on the merged data sets. At this time, the similarity information can also be used as the similarity information of the motion cycles corresponding to different original data sets.

[0100] S500: Determine a recognition result of a target action in a motion process based on the classification result and the similarity information.

[0101] In step S500, as mentioned above, each motion cycle can determine the corresponding classification result and similarity information, and can determine whether the action corresponding to the motion cycle is the target action based on the classification result and similarity information. If the action corresponding to the motion cycle is the target action, the statistical number of the target action can be increased by 1 as the recognition result of the target action.

[0102] Exemplarily, the classification result and the similarity information are numerical values ​​R1 and R2, respectively. The larger the numerical value R1 corresponding to the classification result, the greater the probability that the action corresponding to the motion cycle is the target action. The larger the numerical value R2 corresponding to the similarity information, the higher the degree of similarity between the action corresponding to the motion cycle and the target action. A judgment index value R can be determined based on the classification result R1 and the similarity information R2, R=R1 / (R2+0.001), then it can be determined whether the action corresponding to the motion cycle is the target action based on the magnitude relationship between the judgment index value R and the fifth threshold value. When the judgment index value R is greater than or equal to the fifth threshold value, the action corresponding to the motion cycle can be determined to be the target action. When the judgment index value R is less than the fifth threshold value, it can be determined that the action corresponding to the motion cycle is not the target action.

[0103] In this embodiment, by acquiring the original motion data set and determining the motion cycle and target data set corresponding to the original motion data set, and then classifying the motion cycle according to the target data set, and determining the classification result and similarity information corresponding to the motion cycle according to the target data set and the standard motion data set, the recognition result of the target action can be determined according to the classification result and similarity information. The recognition and judgment of the target action are performed by combining the classification result and similarity information, which ensures the monitoring accuracy of motion monitoring, reduces the error rate of the recognition result, and improves the user experience.

[0104] In some embodiments, based on each original motion data set, determining at least one motion cycle corresponding to each original motion data set and a target data set corresponding to each motion cycle includes:

[0105] refer to Figure 2 As shown, the following determination process is performed for each original motion data set:

[0106] S210 , performing data preprocessing on the original motion data set to obtain a preprocessed data set.

[0107] In step S210, the original motion data set is preprocessed to obtain a preprocessed data set corresponding to the original motion data set. The amount of preprocessed data in the preprocessed data set may be smaller than the amount of original motion data in the original motion data set.

[0108] S220. Determine at least one motion cycle based on the preprocessed data set.

[0109] In step S220, one or more motion cycles corresponding to the preprocessed data set are determined based on the obtained preprocessed data set. Since the preprocessed data set corresponds to the original motion data set, the determined motion cycle also corresponds to the original motion data set.

[0110] S230. Based on each motion cycle, determine a first data set corresponding to each motion cycle, where the first data set includes first target data representing feature information within the corresponding motion cycle.

[0111] In step S230, according to the determined motion cycle, the first data set corresponding to each motion cycle is determined. For example, the first data set corresponding to the motion cycle can be determined by performing certain calculations on the preprocessed data in the preprocessed data set within the time range corresponding to the motion cycle. The first data set includes first target data representing feature information within the corresponding motion cycle. The feature information represented by the first target data can be, for example, data statistical features within the motion cycle.

[0112] S240. Based on each motion cycle, determine a second data set corresponding to each motion cycle, where the second data set includes second target data representing action information within the corresponding motion cycle.

[0113] In step S240, according to the determined motion cycle, the second data set corresponding to each motion cycle is determined. For example, the second data set corresponding to the motion cycle can be determined by extracting the preprocessed data in the preprocessed data set within the time range corresponding to the motion cycle. The second data set includes second target data representing the motion information in the corresponding motion cycle. The motion information represented by the second target data can be, for example, the acceleration, angular velocity, etc. of the motion corresponding to the motion cycle.

[0114] It should be noted that both the first data set and the second data set can be used as part of the target data set. The first data set and the second data set corresponding to the same motion cycle together constitute a target data set. When determining the classification result and similarity information based on the target data set, the classification result and similarity information can be determined based on the same or different components in the target data set. Exemplarily, the classification result can be determined based on the first data set in the target data set, and the similarity information can be determined based on the second data set in the target data set.

[0115] In this embodiment, by sequentially performing data preprocessing, determining the corresponding motion cycle, determining the first data set corresponding to each motion cycle, and determining the second data set, the determination of at least one motion cycle corresponding to each original motion data set and the target data set corresponding to each motion cycle is achieved, providing a basis for the subsequent determination of classification results and similarity information. Performing data preprocessing on the original motion data set can improve the accuracy of the determination of the motion cycle, and determining the first data set and the second data set based on the motion cycle can obtain the first target data and the second target data representing different information, which can be used for the determination of classification results and similarity information, respectively, reducing the interference of non-target actions on the monitoring process, ensuring the monitoring accuracy of motion monitoring, reducing the error rate of recognition results, and improving user experience.

[0116] In some embodiments, reference Figure 3 As shown, the original motion data set is preprocessed to obtain a preprocessed data set, including:

[0117] S211 . Perform smoothing and denoising processing on the original motion data set to obtain a third data set.

[0118] In step S211, the original motion data set is smoothed and denoised. Exemplarily, the original motion data in the original motion data set can be filtered by a finite impulse response (FIR) low-pass filter, an infinite impulse response (IIR) low-pass filter or a mean filter to remove high-frequency noise data and make the data smoother.

[0119] S212: Perform trend item elimination processing on the third data set to obtain a preprocessed data set.

[0120] In step S212, the obtained third data set is processed to eliminate trend items. Exemplarily, the data in the third data set can be subjected to trend item removal by means of a finite impulse response high-pass filter, an infinite impulse response high-pass filter and subtracting the mean filtering result to remove the influence of DC or ultra-low frequency components on subsequent identification.

[0121] In this embodiment, the original motion data set is smoothed and denoised to obtain a third data set, and the third data set is subjected to trend item elimination processing to obtain a preprocessed data set, thereby achieving data preprocessing of the original motion data set. The smoothing and denoising processing removes high-frequency noise data and makes the data smoother. The trend item elimination processing can remove the influence of direct current or ultra-low frequency components on subsequent recognition, thereby improving the accuracy of subsequent motion cycle determination, further reducing the error rate of the final recognition result, and improving the user experience.

[0122] In some embodiments, reference Figure 4 As shown, based on the preprocessed data set, determining at least one motion cycle includes:

[0123] S221 . Determine a set of data to be detected based on the preprocessed data set, wherein each set of data to be detected in the set of data to be detected corresponds one-to-one to a data acquisition moment of at least part of the original motion data during the motion process.

[0124] In step S221, the data set to be detected is determined based on the preprocessed data, so that each original motion data set corresponds to a data set to be detected. The data set to be detected is used for subsequent peak detection. Each data to be detected in the data set to be detected can correspond one-to-one with all data acquisition moments of the original motion data in the original motion data set. Since the original motion data has undergone a data preprocessing process, each data to be detected in the data set to be detected can also correspond one-to-one with part of the data acquisition moments of the original motion data in the original motion data set, so that one data acquisition moment corresponds to at most one data, which is convenient for the subsequent determination of peak data.

[0125] S222. Determine at least one peak data based on the data set to be detected.

[0126] In step S222, at least one peak data is determined in the data to be detected according to the data set to be detected. The peak data can be used to filter the data to be detected by, for example, whether it meets a preset detection condition.

[0127] S223. Determine the motion cycle corresponding to each peak data based on each peak data and the data acquisition time corresponding to each peak data.

[0128] In step S223, since the peak data is determined based on the data set to be detected, each peak data has its corresponding data acquisition time, and the motion cycle corresponding to the peak data can be determined based on each peak data and the data acquisition time corresponding to the peak data.

[0129] Exemplarily, after a peak value data is determined, the data acquisition time corresponding to the peak data can be subtracted and increased by 0.5s respectively as the start time and end time of the motion cycle corresponding to the peak data, so as to determine the motion cycle corresponding to the peak data.

[0130] In this embodiment, the data set to be detected is determined based on the preprocessed data set, and the peak data is determined based on the data set to be detected, and then the motion cycle corresponding to each peak data is determined based on the peak data and the corresponding data acquisition time, thereby realizing the determination of the motion cycle. Determining the data set to be detected based on the preprocessed data set can ensure that one data acquisition time corresponds to at most one data, which is convenient for the subsequent determination of the peak data. Determining the motion cycle based on the peak data allows each motion cycle to have an action corresponding to the motion cycle, completing the division between multiple actions in the complete motion process, facilitating the subsequent identification of the actions corresponding to each motion cycle one by one, ensuring the monitoring accuracy of motion monitoring, reducing the error rate of the identification results, and improving the user experience.

[0131] In some embodiments, based on the set of data to be detected, at least one peak data is determined, including: determining the data to be detected that meets preset detection conditions as peak data, the preset detection conditions including: greater than a first threshold; greater than or equal to the previous data to be detected and greater than the next data to be detected; the difference between the corresponding data acquisition time and the data acquisition time corresponding to the previous peak data is greater than a second threshold.

[0132] Among them, the data to be detected that meets all the preset detection conditions in the data set to be detected is determined as the peak data, and the first threshold is used as the basic low threshold of the peak data. The data to be detected must first meet the preset detection condition greater than the first threshold. Secondly, the data to be detected should not be less than the previous data to be detected and greater than the next data to be detected, so as to ensure that the data to be detected is the maximum value among the adjacent data to be detected, and when continuous equal maximum values ​​appear, it is ensured that the peak data is the last data to be detected among the continuous equal maximum values, that is, the selection of peak data is realized when a flat-top peak occurs. Finally, the difference between the data acquisition time corresponding to the data to be detected and the data acquisition time corresponding to the previous peak data should be greater than the second threshold, so as to avoid simultaneously determining the data to be detected with too small a difference in the two data acquisition times as peak data, resulting in miscalculation of the motion cycle.

[0133] In this embodiment, the peak data is determined by setting preset detection conditions and determining the data to be detected that meets all the preset detection conditions as peak data. The setting of the preset detection conditions ensures that the peak data is greater than the basic low threshold value when it is used as peak data, and ensures that the peak data is the maximum value among the adjacent data to be detected. At the same time, it ensures that the selection of peak data can be realized when a flat peak occurs, and it can also avoid determining the data to be detected with too small a difference in the time of data acquisition as peak data at the same time, which makes the determination of peak data more reasonable, ensures the accuracy of the subsequent determination of the motion cycle based on the peak data, and thus ensures the monitoring accuracy of motion monitoring, reduces the error rate of the recognition result, and improves the user experience.

[0134] In some embodiments, the preprocessed data set includes multiple three-axis motion data groups. Based on the preprocessed data set, the data set to be detected is determined, including: calculating the composite amount of each three-axis motion data group in the preprocessed data set, and using the composite amount as the data to be detected at the corresponding data acquisition time.

[0135] The preprocessed data set includes a plurality of three-axis motion data groups, each of which corresponds to a different data acquisition time, and each of which includes motion data in three different axes, so that the preprocessed data included in the preprocessed data set can more comprehensively and truly reflect the user's motion state, and the motion data in three different axes can be, for example, acceleration data or angular velocity data in three mutually perpendicular directions. When determining the data set to be detected based on the preprocessed data set, the composite amount of each three-axis motion data group in the preprocessed data set is calculated, and the composite amount is used as the data to be detected at the corresponding data acquisition time.

[0136] Exemplarily, the pre-processed data set includes a plurality of three-axis motion data sets (x1, y1, z1), (x2, y2, z2), ..., (x n ,y n , z n ), n is a positive integer greater than 2, and each three-axis motion data group corresponds to the data acquisition time 1, 2, ..., n, then the synthesis amount of each three-axis motion data group can be calculated separately And the synthetic quantities amp1, amp2, ..., amp n As the data to be detected at the corresponding data acquisition time.

[0137] In this embodiment, by calculating the composite amount of each three-axis motion data group in the preprocessing data set, the composite amount can be used as the data to be detected at the corresponding data acquisition time, thereby realizing the determination of the data set to be detected. The calculation of the composite amount enables multiple motion data in the three-axis motion data set to be synthesized into one data, ensuring that each data to be detected in the data set to be detected corresponds to a different data acquisition time, and the composite amount can also characterize the motion state embodied by the three-axis motion data set. The subsequent determination of the peak data based on the composite amount can also ensure the rationality of the peak data and the corresponding motion cycle.

[0138] In some embodiments, based on each motion cycle, determining a first data set corresponding to each motion cycle includes:

[0139] refer to Figure 5 As shown, the following determination process is performed for each motion cycle:

[0140] S231. Determine at least one statistical characteristic value of the data to be detected within a motion cycle.

[0141] In step S231, the statistical characteristic value of the data to be detected in the motion cycle of the data set to be detected is determined. The statistical characteristic value may be one or more, and the statistical characteristic value can characterize the characteristics of the connection between the data to be detected or the overall distribution of each data to be detected. The data to be detected can be calculated according to a specific calculation method corresponding to the statistical characteristic to achieve the determination of the statistical characteristic value.

[0142] S232: Take the peak data, data acquisition time and at least one statistical characteristic value corresponding to the motion cycle as the first data set corresponding to the motion cycle.

[0143] In step S232, the peak data corresponding to the motion cycle, the data acquisition time corresponding to the peak data, and the statistical characteristic value determined according to the motion cycle are used as the first data set of the motion cycle. When there are multiple statistical characteristic values, the multiple statistical characteristic values ​​are all used as components of the first data set.

[0144] In this embodiment, by determining at least one statistical characteristic value of the data to be detected in each motion cycle, the peak data, data acquisition time and at least one statistical characteristic value corresponding to the motion cycle can be used as the first data set corresponding to the motion cycle, thereby realizing the determination of the first data set and providing a basis for the determination of the target data set. The first data set includes peak data, acquisition time and statistical characteristic values, which can more comprehensively and specifically characterize the characteristics of the corresponding action cycle, facilitate the determination of classification results and similarity information, reduce the interference of non-target actions on the monitoring process, ensure the monitoring accuracy of motion monitoring, reduce the error rate of recognition results, and improve user experience.

[0145] In some embodiments, the at least one statistical characteristic value includes any one or a combination of any multiple of mean, variance, kurtosis and skewness.

[0146] The statistical characteristic value can be any one of the mean, variance, kurtosis and skewness, or any combination of multiple ones, where the mean is used to characterize the average value of each data to be detected within the motion cycle, the variance is used to characterize the degree of dispersion of each data to be detected within the motion cycle, the kurtosis is used to characterize the peak state of the probability distribution of each data to be detected within the motion cycle, and the skewness is used to characterize the direction and degree of distribution skewness of each data to be detected within the motion cycle.

[0147] Exemplarily, the mean, variance, kurtosis and skewness may all be statistical characteristic values ​​as components of the first data set, and the statistical characteristic values ​​may also include characteristic values ​​of other statistical features such as root mean square, spectral centroid, etc.

[0148] In this embodiment, by taking any one or any combination of mean, variance, kurtosis and skewness as statistical characteristic values, the statistical characteristic values ​​can characterize certain numerical or distribution characteristics of the data to be detected within the motion cycle. When any combination of multiple values ​​is used as statistical characteristic values, the characteristics of the corresponding motion cycle can be more comprehensively and specifically characterized, which facilitates the determination of classification results and similarity information, reduces the interference of non-target actions on the monitoring process, ensures the monitoring accuracy of motion monitoring, reduces the error rate of recognition results, and improves user experience.

[0149] In some embodiments, based on each motion cycle, a second data set corresponding to each motion cycle is determined, including: performing the following determination process for each motion cycle: taking each data to be detected within the motion cycle as the second data set corresponding to the motion cycle, or, taking the data of the preprocessed data set within the motion cycle as the second data set corresponding to the motion cycle.

[0150] For each motion cycle, the data to be detected in the motion cycle can be used as the second data set corresponding to the motion cycle, and the data of the preprocessed data set in the motion cycle can also be used as the second data set corresponding to the motion cycle.

[0151] Exemplarily, when the preprocessing data set includes multiple three-axis motion data groups, the three-axis motion data groups within the motion cycle can be directly used as the second data set corresponding to the motion cycle. At this time, each second target data of the second data set is a three-axis motion data group corresponding to at least part of the data acquisition time, and each data acquisition time corresponds to motion data in three different axes. It is also possible to use each to-be-detected data within the motion cycle as the second data set corresponding to the motion cycle. Each to-be-detected data within the motion cycle can be a composite quantity determined based on each three-axis motion data group. At this time, each second target data of the second data set is a composite quantity corresponding to at least part of the data acquisition time.

[0152] In this embodiment, each data to be detected within the motion cycle or the data of the pre-processed data set within the motion cycle is used as the second data set corresponding to the motion cycle, thereby realizing the determination of the second data set corresponding to each motion cycle, and the components or synthetic quantities in each direction that characterize the motion state can be used as the second target data of the second data set, which facilitates the subsequent determination of classification results and similarity information based on the target data set, reduces the interference of non-target actions on the monitoring process, ensures the monitoring accuracy of motion monitoring, reduces the error rate of recognition results, and improves user experience.

[0153] In some embodiments, the motion cycles corresponding to the target data sets of each original motion data set are classified to obtain classification results of the actions corresponding to each motion cycle, including: based on the first data sets corresponding to the original motion data sets and the classification model, the classification results of the actions corresponding to each motion cycle are determined, the classification results including a first result and a second result, the first result representing that the probability that the action corresponding to the motion cycle is the target action is 1, and the second result representing that the probability that the action corresponding to the motion cycle is the target action is 0.

[0154] Among them, the classification model can be a decision tree model, and the decision tree model can present decision rules based on real data with gold label to solve the classification problem. Each internal node of the decision tree model corresponds to a decision judgment, and each leaf node corresponds to a judgment result. The classification result of the action corresponding to each motion cycle can be determined according to the first data set in the target data set corresponding to each original motion data set and the decision tree model. The classification result includes a first result and a second result. The first result represents that the probability of the action corresponding to the motion cycle being the target action is 1, and the second result represents that the probability of the action corresponding to the motion cycle being the target action is 0.

[0155] Exemplarily, the model parameter depth of the decision tree model, i.e., the number of decision judgments, may be, for example, 4. The decision tree model may be obtained, for example, by supervised training, and the minimum number of samples corresponding to each leaf node may be, for example, 10. After the first data set is input into the decision tree model, the first result that characterizes the probability that the action corresponding to the motion cycle is the target action may be, for example, a value of 1, and the second result that characterizes the probability that the action corresponding to the motion cycle is the target action may be, for example, a value of 0.

[0156] In this embodiment, the decision tree model is used as a classification model, and the classification results of the actions corresponding to each motion cycle can be determined according to each first data set corresponding to each original motion data set, thereby achieving the determination of the classification results. By determining the classification results through the decision tree model and the first data set, multiple decisions and judgments can be made according to the characteristic information of the motion cycle, ensuring the accuracy of the classification of the motion cycles corresponding to the target action and the non-target action, facilitating the subsequent determination of the recognition results of the target action according to the classification results and similarity information, ensuring the monitoring accuracy of motion monitoring, reducing the error rate of the recognition results, and improving the user experience.

[0157] In some embodiments, there are multiple sets of original motion data, refer to Figure 6 As shown, based on each first data set corresponding to each original motion data set and the classification model, the classification result of the action corresponding to each motion cycle is determined, including:

[0158] S310 , merging first data sets whose differences in the data acquisition times of the peak data between the original motion data sets are less than or equal to a third threshold to obtain a first combined data set.

[0159] In step S310, when there are multiple original motion data sets, each original motion data set corresponds to at least one motion cycle, and each motion cycle corresponds to a first data set, so that each original motion data set corresponds to one or more different first data sets. At this time, the first data sets whose difference in the data acquisition time of the peak data between the original motion data sets is less than or equal to the third threshold are merged to obtain a first combined data set. As mentioned above, each first data set can include peak data, and the acquisition time of the peak data here is the data acquisition time of the peak data corresponding to each first data set.

[0160] S320: Determine the classification results of the actions corresponding to each motion cycle based on each first combined data set and the classification model.

[0161] In step S320, the classification results of the actions corresponding to each motion cycle are determined according to each first combined data set and a classification model such as a decision tree model. The manner of determining the classification results according to the decision tree model and the first combined data set is similar to the manner of determining the classification results according to the decision tree model and the first data set as described above, and will not be described in detail here. The obtained classification results can simultaneously characterize the probability that the actions corresponding to the motion cycles of different original motion data sets corresponding to the first combined data set are the target actions.

[0162] It should be noted that when there are multiple original motion data sets, the decision tree model used to determine the classification result based on the first combined data set and the decision tree model may be different from the decision tree model used to determine the classification result based on the first data set and the decision tree model when there is one original motion data set.

[0163] Exemplarily, there are two original motion data sets, which are original motion data sets including three-axis acceleration data and three-axis angular velocity data. The first data sets corresponding to the original motion data set of three-axis acceleration and the original motion data set of three-axis angular velocity are determined respectively. The differences in the acquisition time of the peak data of each first data set corresponding to the original motion data set of three-axis acceleration and each first data set corresponding to the original motion data set of three-axis angular velocity are calculated in sequence, and a first data set corresponding to the three-axis acceleration and a first data set corresponding to the three-axis angular velocity whose difference is less than a third threshold are merged to obtain a first combined data set, thereby determining all the first combined data sets. Finally, the first combined data set is input into the decision tree model to obtain the classification results of the actions corresponding to each motion cycle.

[0164] In this implementation, when there are multiple original motion data sets, some of the first data sets between the original motion data sets are merged to obtain a first combined data set, and the classification result is determined based on the first combined data set and the classification model, thereby achieving the determination of the classification result in the case of multiple original motion data sets, comprehensively considering the impact of different original motion data sets on the classification result, further improving the accuracy of the determination of the classification result, thereby ensuring the monitoring accuracy of motion monitoring, reducing the error rate of the recognition result, and improving the user experience. In addition, the first data sets whose difference in the data acquisition time of the peak data is less than or equal to the third threshold are merged, avoiding the merging of the first data sets whose data acquisition time difference is too large, and ensuring that the motion cycle corresponding to the merged first data set is the motion cycle corresponding to the same action.

[0165] In some embodiments, based on each target data set and the standard motion data set corresponding to the target action, the similarity information between the action corresponding to each motion cycle and the target action is determined, including: based on each second data set corresponding to each original motion data set and the standard motion data set, the similarity information between the action corresponding to each motion cycle and the target action is determined.

[0166] Among them, each original motion data set corresponds to at least one motion cycle, and each motion cycle corresponds to a second data set. The similarity information between the actions in each motion cycle and the target action can be determined based on the second data set in the target data set corresponding to each motion cycle and the standard motion data set corresponding to the target action.

[0167] In this embodiment, based on the second data set and the standard motion data set, the similarity information between the action corresponding to each motion cycle and the target action can be determined, thereby achieving the determination of the similarity information. By determining the similarity information based on the second target data in the second data set used to characterize the action information corresponding to the motion cycle and the standard motion data set, the similarity between the action corresponding to the motion cycle and the target action can be judged based on the action information of the motion cycle, thereby ensuring the accuracy of the classification of the motion cycles corresponding to the target action and the non-target action, facilitating the subsequent determination of the recognition result of the target action based on the classification result and the similarity information, thereby ensuring the monitoring accuracy of motion monitoring, reducing the error rate of the recognition result, and improving the user experience.

[0168] In some embodiments, based on each second data set corresponding to each original motion data set and the standard motion data set, determining similarity information between the motion corresponding to each motion cycle and the target motion includes:

[0169] refer to Figure 7 As shown, the following determination process is performed for each second data set:

[0170] S410: Determine a target matching distance between the second data set and the standard motion data set according to a dynamic time warping algorithm.

[0171] In step S410, a target matching distance between each second data set and the standard motion data set is determined according to a dynamic time warping (DTW) algorithm. The time warping algorithm can evaluate the matching distance between two data sets, and can dynamically stretch or compress the timing trend to obtain a matching distance that can evaluate the similarity between the two data sets.

[0172] S420: Use the target matching distance as similarity information.

[0173] In step S240, the determined target matching distance is used as the similarity information of the motion cycle corresponding to the second data set, and the target matching distance is used to characterize the similarity between the action corresponding to the motion cycle and the target action.

[0174] In this embodiment, the target matching distance between each second data set and the standard motion data set is determined according to the dynamic time warping algorithm, and the target matching distance can be used as the similarity information between the motion cycle corresponding to the second data set and the target motion, thereby realizing the determination of similarity. The dynamic time warping algorithm can dynamically stretch or compress the timing trend, and is suitable for the dynamic adjustment of data sets with different data quantities. The target matching distance is determined according to the dynamic time warping algorithm, which expands the adaptation range between the second data set and the standard motion data set, and ensures that the second data sets with different data quantities can determine the corresponding similarity according to the same standard motion data set.

[0175] In some embodiments, there are multiple sets of original motion data, refer to Figure 8 As shown, based on each second data set corresponding to each original motion data set and the standard motion data set, determining the similarity information between the action corresponding to each motion cycle and the target action includes:

[0176] S430: Merge the second data sets whose differences in the data acquisition times of the peak data between the original motion data sets are less than or equal to the fourth threshold to obtain a second combined data set.

[0177] In step S430, when there are multiple original motion data sets, each original motion data set corresponds to at least one motion cycle, and each motion cycle corresponds to a second data set, so that each original motion data set corresponds to one or more different second data sets. At this time, the second data sets whose difference in the data acquisition time of the peak data between the original motion data sets is less than or equal to the fourth threshold are merged to obtain a second combined data set, where the acquisition time of the peak data is the data acquisition time of the peak data corresponding to each second data set.

[0178] S440: Determine similarity information between the action corresponding to each motion cycle and the target action based on each second combined data set and the standard motion data set.

[0179] In step S440, the method of determining similarity information based on the standard motion data set and the second combined data set is similar to the method of determining similarity information based on the standard motion data set and the second data set as described above, and is not described in detail here. The obtained similarity information can simultaneously characterize the similarity between the motion cycles of different original data sets corresponding to the first combined data set and the target motion.

[0180] It should be noted that when there are multiple original motion data sets, and the standard motion data set used to determine the similarity information based on the second combined data set and the standard motion data set is the same as the original motion data set, the standard motion data set used to determine the similarity information based on the second data set and the standard motion data set may be different.

[0181] Exemplarily, there are two original motion data sets, which are original motion data sets including three-axis acceleration data and three-axis angular velocity data. The second data sets corresponding to the original motion data set of three-axis acceleration and the original motion data set of three-axis angular velocity are determined respectively. The differences in the acquisition time of the peak data of each second data set corresponding to the original motion data set of three-axis acceleration and each second data set corresponding to the original motion data set of three-axis angular velocity are calculated in sequence, and a second data set corresponding to the three-axis acceleration and a second data set corresponding to the three-axis angular velocity whose difference is less than a fourth threshold are merged to obtain a second combined data set, thereby determining all the second combined data sets. Finally, based on the second combined data set and the standard motion data set, the similarity information between the action corresponding to each motion cycle and the target action is obtained.

[0182] In this embodiment, when there are multiple original motion data sets, some of the second data sets between the original motion data sets are merged to obtain a second combined data set, and similarity information is determined based on the second combined data set and the standard motion data set, thereby achieving the determination of similarity information in the case of multiple original motion data sets, comprehensively considering the results of similarity information for different original motion data sets, further improving the accuracy of the determination of similarity information, and thus ensuring the monitoring accuracy of motion monitoring, reducing the error rate of recognition results, and improving user experience. In addition, the second data sets whose difference in the data acquisition time of the peak data is less than or equal to the fourth threshold are merged, avoiding the merging of second data sets whose data acquisition time difference is too large, and ensuring that the motion cycle corresponding to the merged second data set is the motion cycle corresponding to the same action.

[0183] In some embodiments, the standard motion data set is multiple, reference Fig. 9 As shown, according to the dynamic time warping algorithm, determining the target matching distance between the second data set and the standard motion data set includes:

[0184] S411. Calculate the matching distance between the second data set and each standard motion data set respectively according to a dynamic time warping algorithm.

[0185] In step S411, there are multiple standard motion data sets, and the multiple standard motion data sets can correspond to different target actions. At this time, according to the dynamic time warping algorithm, the matching distance between the second data set and each standard motion data set is calculated respectively to obtain multiple matching distances corresponding to the multiple standard motion data sets.

[0186] S412: Determine the smallest matching distance among the multiple matching distances as the target matching distance.

[0187] In step S412, the smallest matching distance among the obtained multiple matching distances is determined as the target matching distance, so that the target matching distance can represent the similarity between the action of the action cycle corresponding to the second data set and the closest target action.

[0188] Exemplarily, there are three standard motion data sets, which correspond to target actions performed by amateur level, enthusiast level and professional level, respectively. The matching distances D1, D2 and D3 between the second data set and the three standard motion data sets are calculated respectively, and the minimum value D among the three is taken. min is the target matching distance, which is used as the final similarity information.

[0189] In this embodiment, when there are multiple standard motion data sets, the matching distances between the second data set and each standard motion data set are calculated respectively according to the dynamic time warping algorithm, and the smallest matching distance among the multiple matching distances is determined as the target matching distance, so that the target matching distance can characterize the degree of similarity between the action of the action cycle corresponding to the second data set and the closest target action, thereby avoiding the target action corresponding to a single standard motion data set being too different from the target action of other levels of sports participants and being unable to accurately identify the target action, thereby ensuring the monitoring accuracy of motion monitoring, reducing the error rate of the recognition results, and improving the user experience.

[0190] In some embodiments, acquiring at least one original motion data set during the motion process includes: acquiring multiple original motion data sets during the motion process, where different original motion data sets are detected by different sensors.

[0191] In this embodiment, at least one original motion data set acquired during the motion process can be detected by different sensors to obtain the original motion data sets corresponding to each sensor, so that the original motion data sets included in the multiple original motion data sets can represent the motion direction and motion speed through different data, thereby improving the accuracy of subsequent target motion recognition results.

[0192] In some embodiments, the sensor includes an accelerometer and a gyroscope.

[0193] The acceleration sensor and the gyroscope are respectively used as sensors for obtaining two different sets of raw motion data. The acceleration sensor can obtain acceleration data in multiple axes as raw motion data, and the gyroscope can obtain angular velocity data in different axes as raw motion data.

[0194] In this embodiment, the acceleration sensor and the gyroscope are used as different sensors for obtaining multiple sets of original motion data, which can respectively obtain acceleration data and angular velocity data in multiple axes, so that the original motion data in different sets of original motion data can characterize the motion speed and motion direction in different ways, thereby improving the accuracy of subsequent target motion recognition results.

[0195] In some embodiments, after determining the recognition result of the target action during the exercise process, the exercise monitoring method further includes: determining the force of the recognized target action based on each target data set and the recognition result of the target action.

[0196] Among them, after determining the recognition result of the target action corresponding to each motion cycle, the force of the identified target action can also be determined based on the target data set corresponding to each motion cycle and the recognition result of the target action. Exemplarily, when the recognition result of the target action corresponding to a motion cycle characterizes the action corresponding to the motion cycle as the target action, the force of the target action can be determined based on the peak data included in the first data set in the target data set corresponding to the motion cycle. The force of the target action can be characterized by the force level, for example. The higher the force level, the more sufficient the force of the target action.

[0197] For example, there may be two sets of original motion data, which are detected by an acceleration sensor and a gyroscope respectively. The original motion data sets detected by the acceleration sensor are all three-axis acceleration data, and the original motion data sets detected by the gyroscope are all three-axis angular velocity data. The acceleration peak data of the first data set corresponding to the three-axis acceleration data and the angular velocity peak data of the second data set corresponding to the three-axis angular velocity data can be determined respectively. When the acceleration peak data is greater than the sixth threshold and less than the seventh threshold, and the angular velocity peak data is greater than the eighth threshold and less than the ninth threshold, the force level characterizing the force of the target action is classified as low. When the acceleration peak data is greater than the seventh threshold and the angular velocity peak data is greater than the eighth threshold and less than the ninth threshold, or when the acceleration peak data is greater than the sixth threshold and less than the seventh threshold and the angular velocity peak data is greater than the ninth threshold, the force level characterizing the force of the target action is classified as intermediate. When the acceleration peak data is greater than the seventh threshold and the angular velocity peak data is greater than the ninth threshold, the force level characterizing the force of the target action is classified as high.

[0198] In this embodiment, after determining the recognition result of the target action during the exercise process, the force of the identified target action can be determined based on the target data set and the target action recognition result corresponding to each exercise cycle, which can realize the division of the force level of the target action. The force of the target action can be counted after the exercise, which expands the function of exercise monitoring and improves the user experience.

[0199] In an exemplary embodiment, reference Fig.10 As shown, a motion monitoring method is provided, which is applied to an electronic device, and the motion monitoring method includes:

[0200] S1, obtaining two sets of original motion data during the motion process, the two sets of original motion data are detected by an acceleration sensor and a gyroscope respectively;

[0201] S2. Based on the two original motion data sets, obtaining at least one motion cycle corresponding to each original motion data set, and a first data set and a second data set corresponding to each motion cycle;

[0202] S3, merging the first data sets whose difference in the data acquisition time of the peak data between the two original motion data sets is less than or equal to the third threshold to obtain a first combined data set;

[0203] S4, determining the classification results of the actions corresponding to each motion cycle based on each first combined data set and the decision tree model;

[0204] S5, merging the second data sets whose difference in the data acquisition time of the peak data between the two original motion data sets is less than or equal to the fourth threshold to obtain a second combined data set;

[0205] S6. Determine similarity information between the action corresponding to each motion cycle and the target action based on each second combined data set and the standard motion data set;

[0206] S7. Determine the recognition result of the target action during the motion process based on the classification result and the similarity information.

[0207] In this embodiment, the motion cycle corresponding to the original motion data set and the first data set and the second data set corresponding to each motion cycle are determined by acquiring the original motion data set and merging them, and then the classification result of the action corresponding to the motion cycle and the similarity information between the action corresponding to the motion cycle and the target action are determined by the classification of the motion cycle and the standard motion data set respectively according to the first data set and the second data set, and the recognition result of the target action can be determined according to the classification result and the similarity information. The two original motion data sets can characterize the motion direction and motion speed through different data, which improves the accuracy of the subsequent target action recognition results. The classification results are calculated by the classification model, the similarity information is calculated according to the standard motion data set, and the target action is identified and judged by combining the classification results and the similarity information, which ensures the monitoring accuracy of motion monitoring, reduces the error rate of the recognition results, and improves the user experience.

[0208] In an exemplary embodiment, reference Fig.11 As shown, a motion monitoring device is provided, which is applied to electronic equipment. The motion monitoring device includes an acquisition module 10, a first determination module 20, a second determination module 30, a third determination module 40 and a fourth determination module 50. The acquisition module 10 is used to acquire at least one original motion data set in the motion process, and the original motion data set includes original motion data representing the motion direction and / or motion speed. The first determination module 20 is used to determine at least one motion cycle corresponding to each original motion data set and a target data set corresponding to each motion cycle based on each original motion data set. The second determination module 30 is used to classify the motion cycle corresponding to the target data set of each original motion data set, and obtain the classification result of the action corresponding to each motion cycle, and the classification result represents the probability that the action corresponding to the motion cycle is the target action. The third determination module 40 is used to determine the similarity information between the action corresponding to each motion cycle and the target action based on the standard motion data set corresponding to each target data set and the target action. The fourth determination module 50 is used to determine the recognition result of the target action in the motion process based on the classification result and the similarity information.

[0209] In this embodiment, the original motion data set is acquired by the acquisition module 10 and the motion cycle and target data set corresponding to the original motion data set are determined by the first determination module 20, and then the classification result of the action corresponding to the motion cycle and the similarity information between the action corresponding to the motion cycle and the target action are determined by the second determination module 30 and the third determination module 40 according to the target data set and the standard motion data set, and the recognition result of the target action can be determined by the fourth determination module 50 according to the classification result and the similarity information. The classification result is obtained by classifying the motion cycle, the similarity information is calculated according to the standard motion data set, and the target action is identified and judged based on the comprehensive classification result and the similarity information, which ensures the monitoring accuracy of motion monitoring, reduces the error rate of the recognition result, and improves the user experience.

[0210] In one embodiment, the acquisition module 10 is further used to: acquire multiple original motion data sets during the motion process, where different original motion data sets are detected by different sensors.

[0211] In one embodiment, the first determination module 20 is also used to: execute the following determination process for each original motion data set: perform data preprocessing on the original motion data set to obtain a preprocessed data set; determine at least one motion cycle based on the preprocessed data set; based on each motion cycle, determine a first data set corresponding to each motion cycle, the first data set includes first target data representing characteristic information within the corresponding motion cycle; based on each motion cycle, determine a second data set corresponding to each motion cycle, the second data set includes second target data representing action information within the corresponding motion cycle.

[0212] In one embodiment, the first determination module 20 is further used to: perform smoothing and denoising processing on the original motion data set to obtain a third data set; and perform trend item elimination processing on the third data set to obtain a preprocessed data set.

[0213] In one embodiment, the first determination module 20 is also used to: determine a data set to be detected based on a preprocessed data set, wherein each data to be detected in the data set to be detected corresponds one-to-one to a data acquisition moment of at least part of the original motion data during the motion process; determine at least one peak data based on the data set to be detected; and determine a motion cycle corresponding to each peak data based on each peak data and the data acquisition moment corresponding to each peak data.

[0214] In one embodiment, the first determination module 20 is also used to: determine the data to be detected that meets the preset detection conditions as peak data, and the preset detection conditions include: greater than a first threshold; greater than or equal to the previous data to be detected and greater than the next data to be detected; the difference between the corresponding data acquisition time and the data acquisition time corresponding to the previous peak data is greater than a second threshold.

[0215] In one embodiment, the preprocessing data set includes multiple three-axis motion data groups, and the first determination module 20 is further used to: calculate the composite amount of each three-axis motion data group in the preprocessing data set, and use the composite amount as the data to be detected at the corresponding data acquisition time.

[0216] In one embodiment, the first determination module 20 is also used to: perform the following determination process for each motion cycle: determine at least one statistical characteristic value of the data to be detected within the motion cycle; and use the peak data, data acquisition time and at least one statistical characteristic value corresponding to the motion cycle as the first data set corresponding to the motion cycle.

[0217] In one embodiment, the first determination module 20 is also used to: perform the following determination process for each motion cycle: use each data to be detected within the motion cycle as the second data set corresponding to the motion cycle; or, use the data of the preprocessed data set within the motion cycle as the second data set corresponding to the motion cycle.

[0218] In one embodiment, the second determination module 30 is also used to: determine the classification results of the actions corresponding to each motion cycle based on the first data sets corresponding to each original motion data set and the classification model, the classification results including a first result and a second result, the first result representing that the probability that the action corresponding to the motion cycle is the target action is 1, and the second result representing that the probability that the action corresponding to the motion cycle is the target action is 0.

[0219] In one embodiment, there are multiple original motion data sets, and the second determination module 30 is also used to: merge the first data sets whose differences in the data acquisition times of the peak data between the original motion data sets are less than or equal to the third threshold to obtain a first combined data set; and determine the classification results of the actions corresponding to each motion cycle based on the first combined data sets and the classification model.

[0220] In one embodiment, the third determination module 40 is further used to determine similarity information between the action corresponding to each motion cycle and the target action based on each second data set corresponding to each original motion data set and the standard motion data set.

[0221] In one embodiment, the third determination module 40 is also used to: perform the following determination process for each second data set: determine the target matching distance between the second data set and the standard motion data set according to the dynamic time warping algorithm; and use the target matching distance as similarity information.

[0222] In one embodiment, there are multiple original motion data sets, and the third determination module 40 is also used to: merge the second data sets whose differences in the data acquisition times of the peak data between the original motion data sets are less than or equal to the fourth threshold to obtain a second combined data set; based on the second combined data sets and the standard motion data set, determine the similarity information between the action corresponding to each motion cycle and the target action.

[0223] In one embodiment, there are multiple standard motion data sets, and the third determination module 40 is also used to: calculate the matching distances between the second data set and each standard motion data set according to the dynamic time warping algorithm; and determine the smallest matching distance among the multiple matching distances as the target matching distance.

[0224] In some embodiments, the motion monitoring device further includes a fifth determination module, which is used to determine the force of the identified target action based on the recognition results of each target data set and the target action.

[0225] In an exemplary embodiment, an electronic device is provided. The electronic device may be a wearable device, such as a smart watch, a smart bracelet, etc., or may be other devices with data processing capabilities connected to the wearable device, such as a mobile phone, a tablet computer, etc. connected to the wearable device.

[0226] refer to Fig.12 As shown, the electronic device may include one or more of the following components: a processing component 101 , a memory 102 , a power component 103 , a multimedia component 104 , an audio component 105 , an input / output (I / O) interface 106 , a sensor component 107 , and a communication component 108 .

[0227] The processing component 101 generally controls the overall operation of the electronic device, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing component 101 may include one or more processors 109 to execute instructions to complete all or part of the steps of the above-mentioned method. In addition, the processing component 101 may include one or more modules to facilitate the interaction between the processing component 101 and other components. For example, the processing component 101 may include a multimedia module to facilitate the interaction between the multimedia component 104 and the processing component 101.

[0228] The memory 102 is configured to store various types of data to support operations on the electronic device. Examples of such data include instructions for any application or method operating on the electronic device, contact data, phone book data, messages, pictures, videos, etc. The memory 102 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0229] The power component 103 provides power to various components of the electronic device. The power component 103 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the electronic device.

[0230] The multimedia component 104 includes a screen that provides an output interface between the electronic device and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundaries of the touch or slide action, but also detect the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 104 includes a front camera and / or a rear camera. When the electronic device is in an operating mode, such as a shooting mode or a video mode, the front camera and / or the rear camera may receive external multimedia data. Each front camera and rear camera may be a fixed optical lens system or have a focal length and optical zoom capability.

[0231] The audio component 105 is configured to output and / or input audio signals. For example, the audio component 105 includes a microphone (MIC), and when the electronic device is in an operation mode, such as a call mode, a recording mode, and a speech recognition mode, the microphone is configured to receive an external audio signal. The received audio signal can be further stored in the memory 102 or sent via the communication component 108. In some embodiments, the audio component 105 also includes a speaker for outputting audio signals.

[0232] I / O interface 106 provides an interface between processing component 101 and peripheral interface modules, which may be keyboards, click wheels, buttons, etc. These buttons may include but are not limited to: a home button, a volume button, a start button, and a lock button.

[0233] The sensor assembly 107 includes one or more sensors for providing various aspects of status assessment for the electronic device. For example, the sensor assembly 107 can detect the on / off state of the electronic device, the relative positioning of components, such as the display and keypad of the electronic device, and the sensor assembly 107 can also detect the position change of the electronic device or a component of the electronic device, the presence or absence of user contact with the electronic device, the orientation or acceleration / deceleration of the electronic device and the temperature change of the electronic device. The sensor assembly 107 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. The sensor assembly 107 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 107 may also include an accelerometer, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0234] The communication component 108 is configured to facilitate wired or wireless communication between the electronic device and other devices. The device can access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof. In an exemplary embodiment, the communication component 108 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 108 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.

[0235] In an exemplary embodiment, the electronic device may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to execute the above-mentioned motion monitoring method applied to the electronic device.

[0236] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 102 including instructions, and the above instructions can be executed by a processor 109 of an electronic device to complete the above-mentioned motion monitoring method applied to the electronic device. For example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a tape, a floppy disk, and an optical data storage device. When the instructions in the storage medium are executed by the processor 109 of the electronic device, the electronic device is enabled to perform the motion monitoring method shown in the above-mentioned embodiment.

[0237] Those skilled in the art will readily appreciate other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art that are not disclosed in this disclosure. The specification and examples are to be considered as exemplary only, and the true scope and spirit of the present invention are indicated by the following claims.

[0238] It should be understood that the present invention is not limited to the exact construction that has been described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.

Claims

1. A motion monitoring method, applied to an electronic device, characterized in that: The motion monitoring method comprises: Acquire at least one original motion data set during the motion process, wherein the original motion data set includes original motion data representing a motion direction and / or a motion speed; Based on each of the original motion data sets, determining at least one motion cycle corresponding to each of the original motion data sets and a target data set corresponding to each of the motion cycles; Classifying the motion cycles corresponding to the target data sets of each of the original motion data sets to obtain classification results of actions corresponding to each of the motion cycles, wherein the classification results represent the probability that the action corresponding to the motion cycle is the target action; Determining similarity information between the action corresponding to each of the motion cycles and the target action based on each of the target data sets and a standard motion data set corresponding to the target action; Based on the classification result and the similarity information, a recognition result of the target action in the motion process is determined.

2. The motion monitoring method according to claim 1, characterized in that: The determining, based on each of the original motion data sets, at least one motion cycle corresponding to each of the original motion data sets and a target data set corresponding to each of the motion cycles comprises: The following determination process is performed for each of the original motion data sets: Performing data preprocessing on the original motion data set to obtain a preprocessed data set; determining at least one of the motion cycles based on the preprocessed data set; Based on each of the motion cycles, determining a first data set corresponding to each of the motion cycles, wherein the first data set includes first target data representing feature information corresponding to the motion cycle; Based on each of the motion cycles, a second data set corresponding to each of the motion cycles is determined, wherein the second data set includes second target data representing the action information within the corresponding motion cycle.

3. The motion monitoring method according to claim 2, characterized in that: The determining, based on the pre-processed data set, at least one of the motion cycles comprises: Based on the preprocessed data set, determining a set of data to be detected, wherein each set of data to be detected in the set of data to be detected corresponds one-to-one to a data acquisition moment of at least part of the original motion data during the motion process; Based on the data set to be detected, determining at least one peak data; Based on each peak data and the data acquisition time corresponding to each peak data, the motion cycle corresponding to each peak data is determined.

4. The motion monitoring method according to claim 3, characterized in that: The step of determining at least one peak value data based on the set of data to be detected comprises: The data to be detected that meets the preset detection conditions is determined as the peak data, and the preset detection conditions include: is greater than a first threshold; Greater than or equal to the previous data to be tested and greater than the next data to be tested; The difference between the corresponding data acquisition time and the data acquisition time corresponding to the previous peak data is greater than the second threshold.

5. The motion monitoring method according to claim 3, characterized in that: The pre-processed data set includes a plurality of three-axis motion data groups, and determining the data set to be detected based on the pre-processed data set includes: The composite amount of each of the three-axis motion data groups in the pre-processed data set is calculated, and the composite amount is used as the data to be detected at the corresponding data acquisition time.

6. The motion monitoring method according to claim 3, characterized in that: The determining, based on each of the motion cycles, a first data set corresponding to each of the motion cycles comprises: The following determination process is performed for each of the motion cycles: Determine at least one statistical characteristic value of the data to be detected within the motion cycle; The peak data, the data acquisition time and the at least one statistical characteristic value corresponding to the motion cycle are used as the first data set corresponding to the motion cycle.

7. The motion monitoring method according to claim 3, characterized in that: The determining, based on each of the motion cycles, a second data set corresponding to each of the motion cycles comprises: The following determination process is performed for each of the motion cycles: taking each of the to-be-detected data in the motion cycle as the second data set corresponding to the motion cycle; or, The data of the preprocessed data set within the motion cycle is used as the second data set corresponding to the motion cycle.

8. The motion monitoring method according to any one of claims 2 to 7, characterized in that: The classifying the motion cycles corresponding to the target data sets of each of the original motion data sets to obtain classification results of the actions corresponding to each of the motion cycles includes: Based on the first data sets corresponding to the original motion data sets and the classification model, the classification results of the actions corresponding to the motion cycles are determined, and the classification results include a first result and a second result. The first result represents that the probability that the action corresponding to the motion cycle is the target action is 1, and the second result represents that the probability that the action corresponding to the motion cycle is the target action is 0.

9. The motion monitoring method according to claim 8, characterized in that: There are a plurality of original motion data sets, and determining the classification result of the action corresponding to each motion cycle based on each of the first data sets corresponding to each of the original motion data sets and the classification model includes: Merging the first data sets whose difference in the data acquisition time of the peak data between the original motion data sets is less than or equal to a third threshold to obtain a first combined data set; Based on each of the first combined data sets and the classification model, a classification result of the action corresponding to each of the motion cycles is determined.

10. The motion monitoring method according to any one of claims 2 to 7, characterized in that: The determining, based on each of the target data sets and the standard motion data set corresponding to the target motion, similarity information between the motion corresponding to each of the motion cycles and the target motion comprises: Based on each of the second data sets corresponding to each of the original motion data sets and the standard motion data set, similarity information between the action corresponding to each of the motion cycles and the target action is determined.

11. The motion monitoring method according to claim 10, characterized in that: The determining, based on each of the second data sets corresponding to each of the original motion data sets and the standard motion data set, similarity information between the motion corresponding to each of the motion cycles and the target motion comprises: The following determination process is performed for each of the second data sets: Determining a target matching distance between the second data set and the standard motion data set according to a dynamic time warping algorithm; The target matching distance is used as the similarity information.

12. The motion monitoring method according to claim 11, characterized in that: There are a plurality of original motion data sets, and determining similarity information between an action corresponding to each of the motion cycles and the target action based on each of the second data sets corresponding to each of the original motion data sets and the standard motion data set, comprises: Merging the second data sets whose difference in the data acquisition time of the peak data between the original motion data sets is less than or equal to a fourth threshold to obtain a second combined data set; Based on each of the second combined data sets and the standard motion data set, similarity information between the action corresponding to each of the motion cycles and the target action is determined.

13. The motion monitoring method according to claim 11, characterized in that: There are multiple standard motion data sets, and determining the target matching distance between the second data set and the standard motion data set according to the dynamic time warping algorithm includes: According to the dynamic time warping algorithm, respectively calculating the matching distance between the second data set and each of the standard motion data sets; The smallest matching distance among the multiple matching distances is determined as the target matching distance.

14. The motion monitoring method according to any one of claims 1 to 7, characterized in that: The obtaining of at least one set of original motion data during the motion process includes: A plurality of the original motion data sets are acquired during the motion process, wherein different original motion data sets are detected by different sensors.

15. The motion monitoring method according to any one of claims 1 to 7, characterized in that: After determining the recognition result of the target action in the motion process, the motion monitoring method further includes: Based on each of the target data sets and the target action recognition result, the force situation of the recognized target action is determined.

16. A motion monitoring device, applied to an electronic device, characterized in that: The motion monitoring device comprises: An acquisition module, the acquisition module is used to acquire at least one original motion data set during the motion process, the original motion data set including original motion data representing the motion direction and / or motion speed; A first determination module, the first determination module is used to determine, based on each of the original motion data sets, at least one motion cycle corresponding to each of the original motion data sets and a target data set corresponding to each of the motion cycles; A second determination module, the second determination module is used to classify the motion cycles corresponding to the target data sets of each of the original motion data sets, and obtain classification results of the actions corresponding to each of the motion cycles, wherein the classification results represent the probability that the action corresponding to the motion cycle is the target action; a third determination module, the third determination module being used to determine similarity information between the action corresponding to each of the motion cycles and the target action based on each of the target data sets and a standard motion data set corresponding to the target action; A fourth determination module is used to determine a recognition result of the target action in the motion process based on the classification result and the similarity information.

17. An electronic device, characterized in that: The electronic device comprises: processor; a memory for storing processor-executable instructions; Wherein, the processor is configured to execute the motion monitoring method as described in any one of claims 1-15.

18. A non-transitory computer-readable storage medium, characterized in that: When the instructions in the storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the motion monitoring method as described in any one of claims 1-15.

Images