A method for short-range scene recognition and switching based on wearable devices

Abstract

The present invention discloses a method for close-range scene recognition switching based on wearable devices, which relates to the field of perceptual interaction technology, including: obtaining operation data of historical user wearable devices; screening the operation data of historical user wearable devices to obtain historical operation time characteristic parameter data; analyzing the change trend of the wearable device scene mode in the historical operation time characteristic data to determine the interval vector range of the wearable device scene mode; obtaining the operation data of real-time user wearable devices; performing correlation mapping on the operation data of real-time user wearable devices according to the wearable device scene mode to obtain the operation vector matrix of the user wearable device; performing matching verification based on the operation vector matrix of the user wearable device and the interval vector range of the wearable device scene mode, and screening out the scene mode with the largest matching degree as the switching scene. The advantages of the present invention are: improving the accuracy and real-time performance of scene recognition, and realizing personalized services for wearable devices.

Description

Technical Field

[0001] The present invention relates to the field of perception interaction technology, and in particular to a method for close-range scene recognition switching based on a wearable device. Background Art

[0002] The close-range scene recognition switching of wearable devices refers to the use of built-in sensors in the device to monitor the environmental data around the user in real time, combine preset scene modes with algorithm analysis, automatically identify the close-range scene the user is currently in, and automatically adjust the device's function settings or trigger corresponding services accordingly to improve user experience and convenience.

[0003] The close-range scene recognition and switching of existing wearable devices all use a single sensor to judge the actual change value of environmental data, resulting in limited recognition accuracy. In addition, due to the low sensor accuracy of wearable devices, misjudgment may occur when recognizing certain similar or complex scenes, resulting in poor accuracy and real-time performance of close-range scene recognition and switching. Summary of the invention

[0004] In order to solve the above technical problems, a method for close-range scene recognition and switching based on wearable devices is provided. This technical solution solves the problem that the close-range scene recognition and switching of the above-mentioned existing wearable devices all use a single sensor to judge the actual change value of environmental data, resulting in limited recognition accuracy. In addition, due to the low sensor accuracy of wearable devices, misjudgment may occur when recognizing certain similar or complex scenes, resulting in poor accuracy and real-time performance of close-range scene recognition and switching.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0006] The method for close-range scene recognition switching based on a wearable device includes:

[0007] Based on the environmental sensors of wearable devices, obtain the historical operation data of users' wearable devices;

[0008] Filter the historical operation data of the user's wearable device according to the wearable device scene mode to obtain the historical operation sequence characteristic parameter data of the wearable device scene mode;

[0009] Based on the historical running time series feature data of the wearable device scene mode, analyzing the change trend of the wearable device scene mode in the historical running time series feature data, and determining the interval vector range of the wearable device scene mode;

[0010] Obtain real-time operating data of user wearable devices;

[0011] According to the wearable device scene mode, the operation data of the real-time user wearable device is correlated and mapped to obtain the operation vector matrix of the user wearable device;

[0012] The matching degree is verified based on the operation vector matrix of the user's wearable device and the interval vector range of the wearable device scene mode, and the scene mode with the greatest matching degree is selected as the switching scene.

[0013] Preferably, filtering the historical operation data of the user's wearable device according to the wearable device scene mode to obtain the historical operation sequence characteristic parameter data of the wearable device scene mode specifically includes:

[0014] Based on the historical operation data of the user's wearable device, the correlation coefficient between each operation parameter in the operation data and the wearable device scene mode is calculated according to the scene mode that affects the wearable device;

[0015] Clustering the correlation coefficient between each operating parameter and the wearable device scene mode to obtain a characteristic array of associated operating parameters of the wearable device scene mode;

[0016] Based on the operation timestamp in the operation data of the historical user's wearable device, the operation timestamp is used as the observation window, and the associated operation parameter feature array of the wearable device scene mode is used as the observation object to establish the historical operation sequence feature parameter data of the wearable device scene mode.

[0017] Preferably, the calculation of the correlation coefficient between each operating parameter in the operating data and the scene mode of the wearable device based on the historical operating data of the wearable device of the user according to the influencing scene mode of the wearable device specifically includes:

[0018] Based on the wearable device scene mode, mark the wearable device scene mode to pay attention to the operating parameters;

[0019] According to the concerned operating parameters of each scene mode, a weighted value is assigned to the mean of the concerned operating parameters of each scene mode;

[0020] Perform standard processing on the operation data of historical user wearable devices and construct a scatter plot of the operation of historical user wearable devices;

[0021] Based on the historical user wearable device operation scatter diagram, the wearable device scene mode focus operation parameters and the weighted assignment of the focus operation parameters of each scene mode, calculate the correlation coefficient between each scatter point in the historical user wearable device operation scatter diagram and the focus operation parameters of each scene mode per unit time;

[0022] Wherein, the correlation coefficient is specifically: ;

[0023] In the formula, S jk is the correlation coefficient between the kth operating parameters concerned by the jth scene mode, X ikis the kth operating parameter corresponding to the i-th scatter point in the operating scatter plot, W jk is the mean weight of the k operating parameters concerned by the jth scene mode of the wearable device, and n is the total number of operating parameters.

[0024] Preferably, clustering the correlation coefficient between each operating parameter and the wearable device scene mode to obtain the associated operating parameter feature array of the wearable device scene mode specifically includes:

[0025] Based on the correlation coefficient between each operating parameter and the scene mode of the wearable device, a correlation coefficient matrix between the operating parameter and the scene mode of the wearable device is established;

[0026] ;

[0027] Among them, A is the correlation coefficient matrix between the operating parameters and the scene mode of the wearable device, S jk is the correlation coefficient between the kth operating parameters concerned by the jth scene mode, m is the total number of scene modes, and n is the total number of operating parameters;

[0028] Based on the wearable device scene mode, a corresponding wearable device scene mode category set is established;

[0029] According to the wearable device scene mode category set, select the operation parameters corresponding to the number of the wearable device scene mode category set in the historical user wearable device operation data as the centroid;

[0030] According to the historical operation data of the user's wearable device, the operation parameter corresponding to the number of scene mode category sets of the wearable device is selected as the centroid;

[0031] Based on the correlation coefficient between each element in the correlation coefficient matrix between the operating parameters and the scene modes of the wearable device and the centroid of the wearable device scene mode category set, iterative clustering is performed through the mean clustering algorithm to obtain the associated operating parameter feature array of the wearable device scene mode;

[0032] The iterative clustering using the mean clustering algorithm specifically includes:

[0033] ;

[0034] In the formula, is the associated operating parameter feature array of the jth scene mode after the vth iteration of the wearable device, is the centroid of the j-th scene mode category set after the v-1th iteration.

[0035] Preferably, based on the historical running time sequence feature data of the wearable device scene mode, analyzing the change trend of the wearable device scene mode in the historical running time sequence feature data, determining the interval vector range of the wearable device scene mode specifically includes:

[0036] Based on the autoregressive moving average ARIMA, a wearable device scene mode trend conversion analysis model is constructed;

[0037] Substitute the historical running sequence feature data based on the wearable device scene mode into the wearable device scene mode trend conversion analysis model to predict the wearable device scene mode in the future unit time as the output, analyze the confidence interval of the error function between the predicted value and the actual value, and determine the interval vector range of the wearable device scene mode;

[0038] The wearable device scene mode trend conversion analysis model is specifically:

[0039] ;

[0040] Where P is the interval vector range of the wearable device scene mode, is the t+i'th hysteresis learning order of the associated operating parameter feature array of the jth scene mode after the vth iteration of the wearable device, SE is the standard deviation of the prediction error, and p is the total number of hysteresis learning orders.

[0041] Preferably, performing matching verification based on the operation vector matrix of the user's wearable device and the interval vector range of the scene mode of the wearable device, and selecting the scene mode with the greatest matching degree as the switching scene specifically includes:

[0042] Based on the running vector matrix of the user's wearable device, calculate the recursive score value of each scene mode per unit time;

[0043] Verify the matching degree of the recursive score value of each scene mode under unit time with the interval vector range of the scene mode of the wearable device to obtain the matching degree of the scene mode of the wearable device;

[0044] Based on the matching degree of the scene modes of the wearable device in a unit time, the scene mode with the greatest matching degree is selected as the switching scene.

[0045] The scene mode with the greatest matching degree is selected as the switching scene as follows:

[0046] ;

[0047] In the formula, Score j (t) is the recursive score value of the jth scene mode at the tth unit time, p - is the lower limit of the interval vector range of the wearable device scene mode, p+ is the upper limit of the interval vector range of the wearable device scene mode, r tk is the kth operating parameter of the operating vector matrix of the user's wearable device at the tth unit time, ∈ is the error term, D is the maximum matching degree of the wearable device scene mode, min is the minimum function, and max is the maximum function.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] The present invention proposes a close-range scene recognition switching solution based on wearable devices. The environmental sensor of the wearable device collects and analyzes the historical operation sequence feature data in the historical operation data, determines the interval vector range of different scene modes, and maps the real-time operation data into an operation vector matrix, and verifies the matching degree with the interval vector range of the scene mode, so as to automatically select and switch to the scene mode that matches the user's current activity the most. The accuracy and real-time performance of scene recognition are improved, and personalized services of wearable devices are realized. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 A flow chart of a method for close-range scene recognition switching based on a wearable device;

[0051] Figure 2 A flow chart of a method for obtaining historical operating sequence characteristic parameter data of a wearable device scene mode;

[0052] Figure 3 A flow chart of the method for analyzing the correlation coefficient between operating parameters and scene modes of wearable devices;

[0053] Figure 4 A flow chart of a method for obtaining a characteristic array of associated operating parameters of a wearable device scene mode;

[0054] Figure 5 A flow chart of a method for determining the range of interval vectors of a wearable device scene mode;

[0055] Figure 6 The flowchart of the scene switching method is to select the scene mode with the greatest matching degree. DETAILED DESCRIPTION

[0056] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0057] Reference Figure 1 As shown, the method for close-range scene recognition switching based on a wearable device includes:

[0058] Based on the environmental sensors of wearable devices, obtain the historical operation data of users' wearable devices;

[0059] Filter the historical operation data of the user's wearable device according to the wearable device scene mode to obtain the historical operation sequence characteristic parameter data of the wearable device scene mode;

[0060] Based on the historical running time series feature data of the wearable device scene mode, analyzing the change trend of the wearable device scene mode in the historical running time series feature data, and determining the interval vector range of the wearable device scene mode;

[0061] Obtain real-time operating data of user wearable devices;

[0062] According to the wearable device scene mode, the operation data of the real-time user wearable device is correlated and mapped to obtain the operation vector matrix of the user wearable device;

[0063] The matching degree is verified based on the operation vector matrix of the user's wearable device and the interval vector range of the wearable device scene mode, and the scene mode with the greatest matching degree is selected as the switching scene.

[0064] This solution collects and analyzes historical operation sequence feature data in historical operation data through the environmental sensors of wearable devices, determines the interval vector ranges of different scene modes, and maps the real-time operation data into an operation vector matrix, verifies the matching degree with the interval vector range of the scene mode, and automatically selects and switches to the scene mode that best matches the user's current activity. This improves the accuracy and real-time performance of scene recognition and realizes personalized services for wearable devices.

[0065] Reference Figure 2 As shown, the operation data of historical user wearable devices are filtered according to the wearable device scene mode, and the historical operation sequence characteristic parameter data of the wearable device scene mode is obtained, which specifically includes:

[0066] Based on the historical operation data of the user's wearable device, the correlation coefficient between each operation parameter in the operation data and the wearable device scene mode is calculated according to the scene mode that affects the wearable device;

[0067] Clustering the correlation coefficient between each operating parameter and the wearable device scene mode to obtain a characteristic array of associated operating parameters of the wearable device scene mode;

[0068] Based on the operation timestamp in the operation data of the historical user's wearable device, the operation timestamp is used as the observation window, and the associated operation parameter feature array of the wearable device scene mode is used as the observation object to establish the historical operation sequence feature parameter data of the wearable device scene mode.

[0069] It is understandable that although users have different living habits and living environments, the operating parameters that the scene modes of wearable devices focus on are relatively small. For example, when the target user is in motion, the operating parameters that the wearable device's preset sports scenes focus on include but are not limited to: heart rate parameters, device posture parameters, and duration parameters. Therefore, by calculating the correlation coefficient between the wearable device scene mode and each operating parameter in the historical user's wearable device operating data, clustering and screening are performed according to the mean of the correlation coefficient to improve the accuracy of scene recognition.

[0070] Reference Figure 3 As shown, the operation data based on the historical user wearable device and the correlation coefficient between each operation parameter in the operation data and the wearable device scene mode are calculated according to the scene mode affecting the wearable device, which specifically includes:

[0071] Based on the wearable device scene mode, mark the wearable device scene mode to pay attention to the operating parameters;

[0072] According to the concerned operating parameters of each scene mode, a weighted value is assigned to the mean of the concerned operating parameters of each scene mode;

[0073] Perform standard processing on the operation data of historical user wearable devices and construct a scatter plot of the operation of historical user wearable devices;

[0074] Based on the historical user wearable device operation scatter diagram, the wearable device scene mode focus operation parameters and the weighted assignment of the focus operation parameters of each scene mode, calculate the correlation coefficient between each scatter point in the historical user wearable device operation scatter diagram and the focus operation parameters of each scene mode per unit time;

[0075] Wherein, the correlation coefficient is specifically: ;

[0076] In the formula, S jk is the correlation coefficient between the kth operating parameters concerned by the jth scene mode, X ik is the kth operating parameter corresponding to the i-th scatter point in the operating scatter plot, W jk is the mean weight of the k operating parameters concerned by the jth scene mode of the wearable device, and n is the total number of operating parameters.

[0077] It can be understood that by analyzing historical user wearable device data, marking and weighting the mean of the focus operating parameters of each scene mode, constructing a scatter plot after standardizing the operating data, and then calculating the correlation coefficient between each scatter point and the scene mode focus parameter, the operating status of the wearable device in different scenarios is quantitatively evaluated to determine the degree of match between the preset mode and the running status of the wearable device in different scenarios. This method can accurately identify the key operating characteristics of the wearable device in different usage scenarios, and improve the accuracy and efficiency of scene recognition through quantitative analysis.

[0078] Reference Figure 4 As shown, clustering the correlation coefficient between each operating parameter and the wearable device scene mode to obtain the associated operating parameter feature array of the wearable device scene mode specifically includes:

[0079] Based on the correlation coefficient between each operating parameter and the scene mode of the wearable device, a correlation coefficient matrix between the operating parameter and the scene mode of the wearable device is established;

[0080] ;

[0081] Among them, A is the correlation coefficient matrix between the operating parameters and the scene mode of the wearable device, S jk is the correlation coefficient between the kth operating parameters concerned by the jth scene mode, m is the total number of scene modes, and n is the total number of operating parameters;

[0082] Based on the wearable device scene mode, a corresponding wearable device scene mode category set is established;

[0083] According to the wearable device scene mode category set, select the operation parameters corresponding to the number of the wearable device scene mode category set in the historical user wearable device operation data as the centroid;

[0084] According to the historical operation data of the user's wearable device, the operation parameter corresponding to the number of scene mode category sets of the wearable device is selected as the centroid;

[0085] Based on the correlation coefficient between each element in the correlation coefficient matrix between the operating parameters and the scene modes of the wearable device and the centroid of the wearable device scene mode category set, iterative clustering is performed through the mean clustering algorithm to obtain the associated operating parameter feature array of the wearable device scene mode;

[0086] The iterative clustering using the mean clustering algorithm specifically includes:

[0087] ;

[0088] In the formula, is the associated operating parameter feature array of the jth scene mode after the vth iteration of the wearable device, is the centroid of the j-th scene mode category set after the v-1th iteration.

[0089] This solution constructs a correlation coefficient matrix by analyzing the correlation between the operating parameters of wearable devices and scene modes; defines a set of scene mode categories; selects representative operating parameters as the initial centroid for each type of scene from historical data; uses the mean clustering algorithm combined with the correlation coefficient to iteratively cluster to extract characteristic parameter combinations under different scenarios to form a correlated operating parameter feature array. This solution can effectively mine and identify the key operating parameters of wearable devices in different usage scenarios, and provide data support for the intelligent management and personalized recommendation of wearable devices.

[0090] Reference Figure 5 As shown, based on the historical running time characteristic data of the wearable device scene mode, analyzing the change trend of the wearable device scene mode in the historical running time characteristic data, determining the interval vector range of the wearable device scene mode specifically includes:

[0091] Based on the autoregressive moving average ARIMA, a wearable device scene mode trend conversion analysis model is constructed;

[0092] Substitute the historical running sequence feature data based on the wearable device scene mode into the wearable device scene mode trend conversion analysis model to predict the wearable device scene mode in the future unit time as the output, analyze the confidence interval of the error function between the predicted value and the actual value, and determine the interval vector range of the wearable device scene mode;

[0093] The wearable device scene mode trend conversion analysis model is specifically:

[0094] ;

[0095] Where P is the interval vector range of the wearable device scene mode, is the t+i'th hysteresis learning order of the associated operating parameter feature array of the jth scene mode after the vth iteration of the wearable device, SE is the standard deviation of the prediction error, and p is the total number of hysteresis learning orders.

[0096] Reference Figure 6 As shown, the matching degree is verified according to the operation vector matrix of the user's wearable device and the interval vector range of the wearable device scene mode, and the scene mode with the largest matching degree is selected as the switching scene, which specifically includes:

[0097] Based on the running vector matrix of the user's wearable device, calculate the recursive score value of each scene mode per unit time;

[0098] Verify the matching degree of the recursive score value of each scene mode under unit time with the interval vector range of the scene mode of the wearable device to obtain the matching degree of the scene mode of the wearable device;

[0099] Based on the matching degree of the scene modes of the wearable device in a unit time, the scene mode with the greatest matching degree is selected as the switching scene;

[0100] The scene mode with the greatest matching degree is selected as the switching scene as follows:

[0101] ;

[0102] In the formula, Score j (t) is the recursive score value of the jth scene mode at the tth unit time, p - is the lower limit of the interval vector range of the wearable device scene mode, p + is the upper limit of the interval vector range of the wearable device scene mode, r tk is the kth operating parameter of the operating vector matrix of the user's wearable device at the tth unit time, ∈ is the error term, D is the maximum matching degree of the wearable device scene mode, min is the minimum function, and max is the maximum function.

[0103] This solution analyzes the real-time operation data of the wearable device, calculates its matching degree with the preset scene mode interval, integrates all matching degrees within a unit time, selects the scene mode with the highest overall matching degree, accurately identifies the user's scene, and automatically switches to the corresponding mode.

[0104] The above shows and describes the basic principle, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited by the above embodiments, and the above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention, and these changes and improvements all fall within the scope of the present invention to be protected.

Claims

1. A method for close-range scene recognition switching based on a wearable device, characterized in that: include: Based on the environmental sensors of wearable devices, obtain the historical operation data of users' wearable devices; Filter the historical operation data of the user's wearable device according to the wearable device scene mode to obtain the historical operation sequence characteristic parameter data of the wearable device scene mode; Based on the historical running time series feature data of the wearable device scene mode, analyzing the change trend of the wearable device scene mode in the historical running time series feature data, and determining the interval vector range of the wearable device scene mode; Obtain real-time operating data of user wearable devices; According to the wearable device scene mode, the operation data of the real-time user wearable device is correlated and mapped to obtain the operation vector matrix of the user wearable device; The matching degree is verified based on the operation vector matrix of the user's wearable device and the interval vector range of the wearable device scene mode, and the scene mode with the greatest matching degree is selected as the switching scene; Among them, it is characterized in that the operation data of historical user wearable devices are screened according to the wearable device scene mode, and the historical operation sequence characteristic parameter data of the wearable device scene mode is obtained, which specifically includes: Based on the historical operation data of the user's wearable device, the correlation coefficient between each operation parameter in the operation data and the wearable device scene mode is calculated according to the scene mode that affects the wearable device; Clustering the correlation coefficient between each operating parameter and the wearable device scene mode to obtain a characteristic array of associated operating parameters of the wearable device scene mode; Based on the operation timestamp in the operation data of the historical user's wearable device, the operation timestamp is used as the observation window, and the associated operation parameter feature array of the wearable device scene mode is used as the observation object to establish the historical operation sequence feature parameter data of the wearable device scene mode; The operation data of the wearable device of the historical user is used to calculate the correlation coefficient between each operation parameter in the operation data and the scene mode of the wearable device according to the scene mode of the wearable device, which specifically includes: Based on the wearable device scene mode, mark the wearable device scene mode to pay attention to the operating parameters; According to the concerned operating parameters of each scene mode, a weighted value is assigned to the mean of the concerned operating parameters of each scene mode; Perform standard processing on the operation data of historical user wearable devices and construct a scatter plot of the operation of historical user wearable devices; Based on the historical user wearable device operation scatter diagram, the wearable device scene mode focus operation parameters and the weighted assignment of the focus operation parameters of each scene mode, calculate the correlation coefficient between each scatter point in the historical user wearable device operation scatter diagram and the focus operation parameters of each scene mode per unit time; Wherein, the correlation coefficient is specifically: ; In the formula, S jk is the correlation coefficient between the kth operating parameters concerned by the jth scene mode, X ik is the kth operating parameter corresponding to the i-th scatter point in the operating scatter plot, W jk is the mean weight of the k operating parameters concerned by the jth scene mode of the wearable device, and n is the total number of operating parameters; Among them, based on the historical running time characteristic data of the wearable device scene mode, analyzing the change trend of the wearable device scene mode in the historical running time characteristic data, determining the interval vector range of the wearable device scene mode specifically includes: Based on the autoregressive moving average ARIMA, a wearable device scene mode trend conversion analysis model is constructed; Substitute the historical running sequence feature data based on the wearable device scene mode into the wearable device scene mode trend conversion analysis model to predict the wearable device scene mode in the future unit time as the output, analyze the confidence interval of the error function between the predicted value and the actual value, and determine the interval vector range of the wearable device scene mode; The wearable device scene mode trend conversion analysis model is specifically: ； Where P is the interval vector range of the wearable device scene mode, is the t+i'th hysteresis learning order of the associated operating parameter feature array of the jth scene mode after the vth iteration of the wearable device, SE is the standard deviation of the prediction error, and p is the total number of hysteresis learning orders.

2. The method for close-range scene recognition switching based on a wearable device according to claim 1, characterized in that: The step of clustering the correlation coefficient between each operating parameter and the scene mode of the wearable device to obtain the associated operating parameter feature array of the scene mode of the wearable device specifically includes: Based on the correlation coefficient between each operating parameter and the scene mode of the wearable device, a correlation coefficient matrix between the operating parameter and the scene mode of the wearable device is established; ； Among them, A is the correlation coefficient matrix between the operating parameters and the scene mode of the wearable device, S jk is the correlation coefficient between the kth operating parameters concerned by the jth scene mode, m is the total number of scene modes, and n is the total number of operating parameters; Based on the wearable device scene mode, a corresponding wearable device scene mode category set is established; According to the wearable device scene mode category set, select the operation parameters corresponding to the number of the wearable device scene mode category set in the historical user wearable device operation data as the centroid; According to the historical operation data of the user's wearable device, the operation parameter corresponding to the number of scene mode category sets of the wearable device is selected as the centroid; Based on the correlation coefficient between each element in the correlation coefficient matrix between operating parameters and wearable device scene modes and the centroid of the wearable device scene mode category set, iterative clustering is performed through the mean clustering algorithm to obtain the associated operating parameter feature array of the wearable device scene mode.

3. The method for close-range scene recognition switching based on a wearable device according to claim 2, characterized in that: The iterative clustering of the mean value clustering algorithm specifically includes: ； In the formula, is the associated operating parameter feature array of the jth scene mode after the vth iteration of the wearable device, is the centroid of the j-th scene mode category set after the v-1th iteration.

4. The method for close-range scene recognition switching based on a wearable device according to claim 3, characterized in that: According to the operation vector matrix of the user's wearable device and the interval vector range of the wearable device scene mode, the matching degree is verified, and the scene mode with the largest matching degree is selected as the switching scene, which specifically includes: Based on the running vector matrix of the user's wearable device, calculate the recursive score value of each scene mode per unit time; Verify the matching degree of the recursive score value of each scene mode under unit time with the interval vector range of the scene mode of the wearable device to obtain the matching degree of the scene mode of the wearable device; Based on the matching degree of the scene modes of the wearable device in a unit time, the scene mode with the greatest matching degree is selected as the switching scene.

5. The method for close-range scene recognition switching based on a wearable device according to claim 4, characterized in that: The scene mode with the greatest matching degree is selected as the switching scene: ； In the formula, Score j (t) is the recursive score value of the jth scene mode at the tth unit time, p - is the lower limit of the interval vector range of the wearable device scene mode, p + is the upper limit of the interval vector range of the wearable device scene mode, r tk is the kth operating parameter of the operating vector matrix of the user's wearable device at the tth unit time, ∈ is the error term, D is the maximum matching degree of the wearable device scene mode, min is the minimum function, and max is the maximum function.

Images