Smart home controller and control method

Abstract

The present invention relates to the field of intelligent control, and in particular to an intelligent home controller and a control method. The present invention stores preset hand motion data of a monitoring subject and synchronously identifies morphological features corresponding to the preset hand motion; compares the morphological features corresponding to the preset hand motion with pre-stored morphological features corresponding to a number of normal motions, so as to determine motion matching features for the morphological features corresponding to the number of normal motions based on the comparison results; calculates motion fuzzy characterization coefficients for the preset hand motion based on the motion matching features, so as to divide the fuzzy misjudgment tendency for the preset hand motion; collects the current hand motion data of the monitoring subject in real time, so as to continuously obtain recognition results for the hand motion; in response to the recognition result being the preset hand motion, analyzes the preset hand motion based on the fuzzy misjudgment tendency for the preset hand motion to determine control instructions for preset home appliances. The present invention can improve the accuracy of the preset hand motion in controlling home appliances.

Description

Technical Field

[0001] The present invention relates to the field of intelligent control, and in particular to an intelligent home controller and a control method. Background Art

[0002] Before the concept of smart home emerged, most traditional home appliances used manual mechanical control methods, such as operating the light switch and adjusting the gear of electrical appliances through physical buttons. This method requires people to operate at the location where the control device is set, which is less convenient. Smart home uses cameras, infrared sensors and other devices to capture user gestures and converts them into control instructions through image processing and motion recognition algorithms, becoming a new interactive method in smart home control. For example, the opening and closing of smart curtains can be controlled by waving, which is more intuitive and convenient than traditional button control, providing users with more diverse control options. Especially in some scenarios where it is inconvenient to operate physical buttons with both hands, hand motion control has played a unique advantage.

[0003] Chinese patent application publication number: CN117492373A, discloses a smart home system based on gesture recognition control, including a recognition response analysis module, a depth operation status analysis module, an environment judgment module and an available warning module; the availability of gesture recognition is evaluated by collecting recognition response information and depth sensor stability information, and the gesture recognition available warning coefficient is calculated by normalizing the gesture conversion jolt evaluation value and the depth frame rate variation index. By comparing the gesture recognition available warning coefficient with the gesture recognition evaluation threshold, the availability warning module can detect potential performance degradation or failure in advance before gesture recognition problems occur, so that maintenance personnel can take timely measures to check and repair the gesture recognition system; users can enjoy high-quality gesture recognition and smart home interaction experience, which helps to improve user satisfaction.

[0004] However, there are still the following problems in the prior art:

[0005] Gesture control is often used in the field of smart home control. For example, preset hand movements are set in advance, and corresponding control instructions are generated when the sensor detects that the monitored subject has made preset hand movements. However, the prior art does not take into account that the pre-set hand movements may be relatively routine and lack specificity, which leads to the problem that the daily movements of the monitored subject are prone to mistakenly triggering control instructions. Summary of the invention

[0006] To this end, the present invention provides a smart home controller and control method to overcome the problem in the prior art that the pre-set preset hand movements may be relatively routine and lack specificity, which in turn leads to the problem that the daily movements of the monitored subject are easily triggered by mistake.

[0007] To achieve the above object, the present invention provides a smart home control method, which includes:

[0008] Storing preset hand motion data of the monitored subject, and simultaneously identifying morphological features corresponding to the preset hand motions, including trajectory features of the preset hand motions and action duration of the preset hand motions;

[0009] Comparing the preset hand action corresponding morphological features with a plurality of pre-stored normal action corresponding morphological features, so as to determine the action matching features for the plurality of normal action corresponding morphological features based on the comparison result, including the action matching degree and the action time-consuming deviation value;

[0010] Calculating a motion fuzzy representation coefficient for the preset hand motion according to the motion matching feature to classify the fuzzy misjudgment tendency for the preset hand motion;

[0011] Collect the current hand motion data of the monitored subject in real time to continuously obtain recognition results for hand motions;

[0012] In response to the recognition result being a preset hand motion, the preset hand motion is analyzed based on the fuzzy misjudgment tendency of the preset hand motion to determine a control instruction for a preset home appliance, including:

[0013] Determine the time domain segment in which the preset hand action is generated, extract the morphological features corresponding to the hand action data in the time domain segment and the corresponding adjacent time domain segment, analyze the action coherence based on each morphological feature to determine whether the preset hand action is valid, and determine the control instruction associated with the valid preset hand action;

[0014] Or, determine the control instruction associated with the preset hand movement.

[0015] Further, the process of determining the action matching features for the action features of the normal action behaviors includes:

[0016] The average of the trajectory similarities between the trajectory features corresponding to the preset hand movements and the trajectory features corresponding to each normal movement is taken as the movement matching degree;

[0017] The average of the difference between the action time required to form the preset hand action and the action time required to form each normal action is taken as the action time deviation value;

[0018] Among them, the trajectory features are generated based on the contour markers of the hand contour.

[0019] Furthermore, the process of calculating the motion fuzzy representation coefficient for the preset hand motion according to the motion matching feature includes:

[0020] The ratio of the action matching degree threshold to the action matching degree is used as the first action feature;

[0021] The ratio of the action time deviation value to the action time deviation threshold is used as the second action feature;

[0022] The first motion feature and the second motion feature are weightedly summed to determine a motion fuzzy characterization coefficient for the preset hand motion.

[0023] Furthermore, the fuzzy misjudgment tendency of the hand action is divided into:

[0024] If the motion blur representation coefficient is greater than or equal to the motion blur representation coefficient threshold, determining that the blur misjudgment tendency of the preset hand motion is high;

[0025] If the motion blur representation coefficient is less than the motion blur representation coefficient threshold, it is determined that the blur misjudgment tendency of the preset hand motion is low.

[0026] Furthermore, the process of continuously obtaining the recognition result for the hand movement includes:

[0027] Determine the trajectory characteristics corresponding to the current hand movement;

[0028] Comparing the trajectory feature with the trajectory feature corresponding to the preset hand movement to determine the similarity, so as to determine whether the hand movement is the preset hand movement;

[0029] If the similarity is greater than a preset similarity threshold, the hand movement is determined to be a preset hand movement.

[0030] Further, based on the fuzzy misjudgment tendency of the preset hand action, the preset hand action is analyzed to determine the control instruction for the preset home appliance, including:

[0031] If the fuzzy misjudgment tendency of the preset hand action is high, then determine the time domain segment in which the preset hand action is generated, extract the morphological features corresponding to the hand action data in the generated time domain segment and the corresponding adjacent time domain segment, analyze the action coherence based on each morphological feature to determine whether the preset hand action is valid, and determine the control instruction associated with the valid preset hand action;

[0032] If the fuzzy misjudgment tendency of the preset hand action is low, a control instruction associated with the preset hand action is determined.

[0033] Furthermore, the process of analyzing the action coherence based on each morphological feature includes:

[0034] Determine the range of motion corresponding to each morphological feature;

[0035] The average difference between the action amplitude corresponding to the generated time domain segment and the action amplitude corresponding to the adjacent time domain segment is calculated, and the reciprocal of the average difference is determined as the action coherence.

[0036] Furthermore, the process of determining the movement amplitude includes:

[0037] Determine the starting point of the trajectory feature, and construct a plurality of rays passing the trajectory feature with the starting point as an endpoint;

[0038] Determine the angle between each ray and the horizontal plane, and determine the maximum angle as the motion amplitude.

[0039] Further, determining whether the preset hand movement is valid includes:

[0040] If the motion continuity is less than the motion continuity threshold, it is determined that the preset hand motion is valid.

[0041] Furthermore, a controller for applying the smart home control method is provided, which includes:

[0042] An action memory, used to store preset hand action data of the monitored subject, and simultaneously identify the morphological features corresponding to the preset hand actions, including the trajectory features of the preset hand actions and the action duration of the preset hand actions;

[0043] a feature identifier connected to the action memory, for comparing the preset hand action corresponding morphological features with a plurality of pre-stored normal action corresponding morphological features, so as to determine action matching features for the plurality of normal action corresponding morphological features based on the comparison result, including action matching degree and action time-consuming deviation value;

[0044] A feature analyzer connected to the feature identifier, for calculating a motion fuzzy characterization coefficient for the preset hand motion according to the motion matching feature, so as to classify the fuzzy misjudgment tendency for the preset hand motion;

[0045] A motion monitor connected to the motion memory for collecting the current hand motion data of the monitoring subject in real time to continuously obtain recognition results for the hand motion;

[0046] A response processor is connected to the feature analyzer and the action monitor respectively, and is used to analyze the preset hand action based on the fuzzy misjudgment tendency of the preset hand action in response to the recognition result being the preset hand action to determine the control instruction for the preset home appliance, including:

[0047] Determine the time domain segment in which the preset hand action is generated, extract the morphological features corresponding to the hand action data in the time domain segment and the corresponding adjacent time domain segment, analyze the action coherence based on each morphological feature to determine whether the preset hand action is valid, and determine the control instruction associated with the valid preset hand action;

[0048] Or, determine the control instruction associated with the preset hand movement.

[0049] Compared with the prior art, the present invention stores the preset hand motion data of the monitored subject and synchronously identifies the morphological features corresponding to the preset hand motion; compares the morphological features corresponding to the preset hand motion with the morphological features corresponding to several pre-stored normal motions to determine the motion matching features for the morphological features corresponding to the several normal motions based on the comparison results; calculates the motion fuzzy characterization coefficient for the preset hand motion according to the motion matching features to divide the fuzzy misjudgment tendency for the preset hand motion; collects the current hand motion data of the monitored subject in real time to continuously obtain the recognition results for the hand motion; in response to the recognition result being the preset hand motion, analyzes the preset hand motion based on the fuzzy misjudgment tendency for the preset hand motion to determine the control instructions for the preset home appliance. The present invention can improve the accuracy of the control of home appliances by the preset hand motion and avoid miscontrol caused by similar motions without control awareness.

[0050] In particular, the present invention calculates the motion fuzzy representation coefficient for the preset hand motion according to the motion matching feature. In actual situations, the preset hand motion set by the monitoring subject may be a conventional simple behavior motion. Therefore, these preset hand motions will be similar or consistent with the body motions made by the monitoring target in life to a certain extent, which will cause the smart home device to mistakenly recognize the body motions in life as preset hand motions, resulting in the situation of mistakenly triggering control instructions.

[0051] Furthermore, the present invention compares the morphological features corresponding to the preset hand movements with the morphological features corresponding to several pre-stored normal movements, and evaluates the differences between the preset hand movements and the normal movements from the dimension of morphological features. It can be understood that the normal movements represent the movements that may exist in the daily life of the monitored subject. Furthermore, if the movement consistency is low and the movement time deviation value is large, it represents a situation where the preset hand movement and the normal movement are quite different. In this case, the preset hand movement has strong specificity, and it is not easy to make misjudgments when subsequently identifying control instructions for the preset hand movement. On the contrary, if the preset hand movement is slightly different from the normal movement, misjudgments are likely to occur. Therefore, quantitative analysis is performed through the motion fuzzy representation coefficient to provide data support for the subsequent classification of the fuzzy misjudgment tendency for the preset hand movement, so as to facilitate the subsequent adaptive analysis of the preset hand movement and determine the control instructions for the preset home appliances, thereby improving the accuracy of identifying the control instructions.

[0052] In particular, the present invention characterizes the difference between the preset hand movements and the normal movements by dividing the fuzzy misjudgment tendency for the preset hand movements, and further characterizes the tendency of subsequent fuzzy misjudgment for the preset hand movements. For hand movements with a high tendency of fuzzy misjudgment, since the preset hand movements take a short time and the difference between them and the normal hand movements is not obvious enough, the preset hand movements are likely to become part of a group of coherent movements of the monitored subject. In this case, the purpose of issuing the preset hand movements is not to enable the smart home device to execute the control command, that is, the above preset hand movements may be invalid. Smart home devices do not need to execute control instructions. In the case of accidental touches, if the preset hand movement is likely to become a component of a set of coherent movements occurring in the monitored subject, it is recognized that the trajectory features corresponding to the preset hand movement are highly coherent with the trajectory features corresponding to the adjacent time domain segments. Therefore, by analyzing the degree of coherence between the preset hand movement in the occurrence time domain segment and other hand movements in the adjacent time domain segments, the tendency of misjudgment is comprehensively considered to filter out invalid situations of the preset hand movement, thereby avoiding erroneous triggering of smart home device control instructions due to misjudgment, greatly improving the accuracy of control instruction execution, and accurately controlling the use of smart home devices. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 A schematic diagram of the steps of a smart home control method according to an embodiment of the invention;

[0054] Figure 2 A logical decision diagram for dividing the fuzzy misjudgment tendency of hand movements for the embodiment of the invention;

[0055] Figure 3 A logic decision diagram for determining whether a preset hand movement is valid for an embodiment of the invention;

[0056] Figure 4 A functional module diagram of a smart home controller according to an embodiment of the invention;

[0057] Figure 5 It is a schematic diagram of the corresponding motion amplitude of the morphological features of an embodiment of the invention. DETAILED DESCRIPTION

[0058] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0059] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0060] It should be noted that, in the description of the present invention, the terms such as "on" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it should not be understood as a limitation on the present invention.

[0061] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installation" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0062] See also Figures 1 to 4 As shown, Figure 1 Schematic diagram of the steps of the smart home control method according to an embodiment of the present invention. Figure 2 A logical decision diagram for dividing the fuzzy misjudgment tendency of hand movements according to an embodiment of the present invention, Figure 3 The following is a logic decision diagram for determining whether a preset hand movement is valid according to an embodiment of the present invention. Figure 4 : is a functional module diagram of a smart home controller according to an embodiment of the present invention. The smart home control method according to the embodiment of the present invention includes:

[0063] Step S1, storing preset hand motion data of the monitored subject, and synchronously identifying morphological features corresponding to the preset hand motion, including trajectory features of the preset hand motion and action duration of the preset hand motion;

[0064] Step S2, comparing the preset hand action corresponding morphological features with a plurality of pre-stored normal action corresponding morphological features, so as to determine the action matching features for the plurality of normal action corresponding morphological features based on the comparison result, including the action matching degree and the action time deviation value;

[0065] Step S3, calculating a motion fuzzy representation coefficient for the preset hand motion according to the motion matching feature, so as to classify the fuzzy misjudgment tendency for the preset hand motion;

[0066] Step S4, collecting the current hand motion data of the monitoring subject in real time to continuously obtain the recognition results for the hand motion;

[0067] Step S5, in response to the recognition result being a preset hand motion, analyzing the preset hand motion based on the fuzzy misjudgment tendency of the preset hand motion to determine a control instruction for a preset home appliance, including:

[0068] Determine the time domain segment in which the preset hand action is generated, extract the morphological features corresponding to the hand action data in the time domain segment and the corresponding adjacent time domain segment, analyze the action coherence based on each morphological feature to determine whether the preset hand action is valid, and determine the control instruction associated with the valid preset hand action;

[0069] Or, determine the control instruction associated with the preset hand movement.

[0070] It can be understood that usually a single preset hand movement corresponds to one control instruction. When the preset hand movement is recognized, the associated control instruction can be determined, and then an electrical signal is generated to control the corresponding smart home device. This will not be repeated.

[0071] Specifically, the preset hand motion data may be image data, and then the trajectory features are determined by identifying the hand contour in the image data.

[0072] In some possible implementations, the hand contour in each frame image can be identified through an image segmentation algorithm, and then after selecting a marker point in the hand contour, the moving trajectory of the marker point is used as the trajectory feature. The selection of the marker point is not limited. For example, the center of the hand contour or any fingertip in the hand contour can be selected as the marker point. Technical personnel in this field can make their own selection, which will not be elaborated here.

[0073] Specifically, the purpose of storing the morphological features corresponding to normal actions is to characterize the morphological features corresponding to several actions performed by the monitored subject in daily life. The morphological features corresponding to normal actions can be obtained in advance. For example, the hand movement data of the monitored subject when performing various actions can be obtained in advance in a laboratory environment, and the morphological features corresponding to each hand movement data can be obtained synchronously, and then the hand movement data and morphological feature data can be stored in a database.

[0074] Specifically, there is no specific limitation on the method of real-time acquisition of the current hand motion data of the monitoring subject, and the data can be acquired through a pre-installed visual acquisition component, for example,

[0075] In some possible implementations, the visual acquisition component may be installed in an independent master control module, and the master control module integrates a logic component for processing the hand motion data collected by the visual acquisition component to complete the analysis of the hand motion;

[0076] In some possible implementations, the visual acquisition component can be installed on a household device, and a logic component for processing the hand motion data collected by the visual acquisition component can be integrated on the household device to complete the analysis of the hand motion, which will not be elaborated here.

[0077] Specifically, the process of determining the motion matching features for several normal motion behavior motion features includes:

[0078] The average of the trajectory similarities between the trajectory features corresponding to the preset hand movements and the trajectory features corresponding to each normal movement is taken as the movement matching degree;

[0079] The average of the difference between the action time required to form the preset hand action and the action time required to form each normal action is taken as the action time deviation value;

[0080] Among them, the trajectory features are generated based on the contour markers of the hand contour.

[0081] Specifically, there is no limitation on the method for determining the trajectory similarity. For example, the Euclidean distance between corresponding points of two trajectories can be calculated to obtain a distance sequence, and the average value of the distance sequence is used as the trajectory similarity. Of course, other methods can also be used, which will not be described in detail.

[0082] In implementation, the image processing algorithm or model for determining trajectory similarity may be imported into the logic component to implement the corresponding function, which will not be described in detail.

[0083] Specifically, the process of calculating the motion blur representation coefficient for the preset hand motion according to the motion matching feature includes:

[0084] The ratio of the action matching degree threshold to the action matching degree is used as the first action feature;

[0085] The ratio of the action time deviation value to the action time deviation threshold is used as the second action feature;

[0086] The first motion feature and the second motion feature are weightedly summed to determine a motion fuzzy characterization coefficient for the preset hand motion.

[0087] In this implementation, when performing weighted summation, the weight of the first action feature is set to 0.55, and the weight of the second action feature is set to 0.45;

[0088] In implementation, a number of stored preset hand motion data may be obtained in advance as samples, and the motion matching features of the morphological features corresponding to each preset hand and the morphological features corresponding to a number of normal motions may be determined to obtain the mean value of the motion matching degree and the mean value of the motion time deviation value;

[0089] The action matching threshold is determined based on the action matching mean. Since the purpose of setting the action matching threshold is to characterize the situation where the morphological features corresponding to the preset hand action are significantly different from the morphological features corresponding to several normal actions, the action matching threshold is set between 0.55 and 0.75 times the action matching mean in implementation.

[0090] The action time deviation threshold is determined based on the mean of the action time deviation values. Since the purpose of setting the action time deviation threshold is also to characterize the situation where the morphological features corresponding to the preset hand movements are significantly different from the morphological features corresponding to several normal movements, the action time deviation threshold is set to between 1.25 times and 1.5 times the mean of the action time deviation values ​​in implementation.

[0091] The present invention calculates the motion fuzzy representation coefficient for the preset hand motion according to the motion matching feature. In actual situations, the preset hand motion set by the monitoring subject may be a conventional simple behavior motion. Therefore, these preset hand motions will be similar or consistent with the body motions made by the monitoring target in life to a certain extent, which will cause the smart home device to mistakenly recognize the body motions in life as preset hand motions, resulting in the situation of mistakenly triggering control instructions.

[0092] Furthermore, the present invention compares the morphological features corresponding to the preset hand movements with the morphological features corresponding to several pre-stored normal movements, and evaluates the differences between the preset hand movements and the normal movements from the dimension of morphological features. It can be understood that the normal movements represent the movements that may exist in the daily life of the monitored subject. Furthermore, if the movement consistency is low and the movement time deviation value is large, it represents a situation where the preset hand movement and the normal movement are quite different. In this case, the preset hand movement has strong specificity, and it is not easy to make misjudgments when subsequently identifying control instructions for the preset hand movement. On the contrary, if the preset hand movement is slightly different from the normal movement, misjudgments are likely to occur. Therefore, quantitative analysis is performed through the motion fuzzy representation coefficient to provide data support for the subsequent classification of the fuzzy misjudgment tendency for the preset hand movement, so as to facilitate the subsequent adaptive analysis of the preset hand movement and determine the control instructions for the preset home appliances, thereby improving the accuracy of identifying the control instructions.

[0093] Specifically, the fuzzy misjudgment tendency for the hand action is divided into:

[0094] If the motion blur representation coefficient is greater than or equal to the motion blur representation coefficient threshold, determining that the blur misjudgment tendency of the preset hand motion is high;

[0095] If the motion blur representation coefficient is less than the motion blur representation coefficient threshold, it is determined that the blur misjudgment tendency of the preset hand motion is low.

[0096] The motion blur characterization coefficient threshold is selected in the interval [1.15,1.2].

[0097] Specifically, the process of continuously obtaining the recognition results for the hand movements includes:

[0098] Determine the trajectory characteristics corresponding to the current hand movement;

[0099] Comparing the trajectory feature with the trajectory feature corresponding to the preset hand movement to determine the similarity, so as to determine whether the hand movement is the preset hand movement;

[0100] If the similarity is greater than a preset similarity threshold, the hand movement is determined to be a preset hand movement.

[0101] During implementation, the similarity threshold is pre-set, wherein a technician in this field can repeat the same hand movement several times, determine the trajectory features accordingly, calculate the mean similarity between the trajectory features, and set the mean similarity as the similarity threshold.

[0102] Specifically, the preset hand motion is analyzed based on the fuzzy misjudgment tendency of the preset hand motion to determine the control instruction for the preset home appliance, including:

[0103] If the fuzzy misjudgment tendency of the preset hand action is high, then determine the time domain segment in which the preset hand action is generated, extract the morphological features corresponding to the hand action data in the generated time domain segment and the corresponding adjacent time domain segment, analyze the action coherence based on each morphological feature to determine whether the preset hand action is valid, and determine the control instruction associated with the valid preset hand action;

[0104] If the fuzzy misjudgment tendency of the preset hand action is low, a control instruction associated with the preset hand action is determined.

[0105] Specifically, the process of analyzing the coherence of movements based on the characteristics of each morphology includes:

[0106] Determine the range of motion corresponding to each morphological feature;

[0107] The average difference between the action amplitude corresponding to the generated time domain segment and the action amplitude corresponding to the adjacent time domain segment is calculated, and the reciprocal of the average difference is determined as the action coherence.

[0108] In this embodiment, the time domain segment in which the preset hand movement is recognized is determined as the generation time domain segment. Since the preset hand movement is relatively simple and the movement time is short, generally speaking, 1s-2s are enough to complete the implementation of the preset hand movement. For the duration of the adjacent time domain segments required for collection, 1s is preferred in this embodiment. By analyzing the movement amplitudes of other hand movements of the monitored subject in the generation time domain segment and its adjacent time domain segments before and after, the continuity of the entire continuous hand movement is comprehensively determined.

[0109] Specifically, the process of determining the amplitude of the movement includes:

[0110] Determine the starting point of the trajectory feature, and construct a plurality of rays passing the trajectory feature with the starting point as an endpoint;

[0111] Determine the angle between each ray and the horizontal plane, and determine the maximum angle as the motion amplitude.

[0112] See also Figure 5 As shown, it is a schematic diagram of the motion amplitude corresponding to the morphological features in this embodiment, wherein the corresponding motion amplitude in the time domain segment t2 is α2, the corresponding motion amplitude in the previous adjacent time domain segment t1 is α1, and the corresponding motion amplitude in the subsequent adjacent time domain segment t3 is α3.

[0113] If the trajectories corresponding to the generated time domain segment and the adjacent time domain segment are relatively coherent, the action amplitudes are relatively close. On the contrary, if the trajectories corresponding to the generated time domain segment and the adjacent time domain segment are less coherent, the action amplitudes are more different and the corresponding action coherence is lower.

[0114] Specifically, there is no limitation on the method of analyzing the trajectory characteristics, and any existing curve analysis tool, such as MATLAB, can be used to analyze the trajectory characteristics, which will not be described in detail.

[0115] Specifically, determining whether the preset hand motion is valid includes:

[0116] If the motion continuity is less than the motion continuity threshold, it is determined that the preset hand motion is valid.

[0117] In this embodiment, the motion continuity threshold is preset, and the corresponding trajectory features of a number of normal actions are obtained, the motion continuity of each trajectory feature is determined, the motion continuity mean is solved, and the motion continuity threshold is set to 1.03 times to 1.07 times the motion continuity mean.

[0118] The motion coherence threshold is selected in the interval [0.1, 0.2].

[0119] For hand movements with a high tendency of fuzzy misjudgment, since the preset hand movement takes a short time and the difference from the normal hand movement is not obvious, the preset hand movement is likely to become a part of a set of coherent movements of the monitored subject. In this case, the purpose of issuing the preset hand movement is not to make the smart home device execute the control instruction, that is, the above preset hand movement may be invalid, and the smart home device does not need to execute the control instruction. For example, when controlling the smart light switch by waving, if the waving movement is simply recognized, the waving movement when changing clothes may trigger the switch by mistake;

[0120] It is understandable that, in actual situations, compared to the case of accidental touch, when the monitoring subject intends to make a preset hand movement, the coherence between the trajectory features corresponding to the preset hand movement and the corresponding trajectory features in the adjacent time domain segment is relatively poor;

[0121] As for the case of false touch, if the preset hand movement is likely to become a component of a group of coherent movements occurring in the monitored subject, and it is recognized that the trajectory features corresponding to the preset hand movement are highly coherent with the trajectory features corresponding to the adjacent time domain segments, then by analyzing the degree of coherence between the preset hand movement in the occurrence time domain segment and other hand movements in the adjacent time domain segments, the tendency of misjudgment is comprehensively considered to filter out invalid situations of the preset hand movement, thereby avoiding erroneous triggering of smart home device control commands due to misjudgment, greatly improving the accuracy of the execution of control commands, and accurately controlling the use of smart home devices.

[0122] Specifically, a controller for applying the smart home control method is also provided, which includes:

[0123] An action memory, used to store preset hand action data of the monitored subject, and simultaneously identify the morphological features corresponding to the preset hand actions, including the trajectory features of the preset hand actions and the action duration of the preset hand actions;

[0124] a feature identifier connected to the action memory, for comparing the preset hand action corresponding morphological features with a plurality of pre-stored normal action corresponding morphological features, so as to determine action matching features for the plurality of normal action corresponding morphological features based on the comparison result, including action matching degree and action time-consuming deviation value;

[0125] A feature analyzer connected to the feature identifier, for calculating a motion fuzzy characterization coefficient for the preset hand motion according to the motion matching feature, so as to classify the fuzzy misjudgment tendency for the preset hand motion;

[0126] A motion monitor connected to the motion memory for collecting the current hand motion data of the monitoring subject in real time to continuously obtain recognition results for the hand motion;

[0127] A response processor is connected to the feature analyzer and the action monitor respectively, and is used to analyze the preset hand action based on the fuzzy misjudgment tendency of the preset hand action in response to the recognition result being the preset hand action to determine the control instruction for the preset home appliance, including:

[0128] Determine the time domain segment in which the preset hand action is generated, extract the morphological features corresponding to the hand action data in the time domain segment and the corresponding adjacent time domain segment, analyze the action coherence based on each morphological feature to determine whether the preset hand action is valid, and determine the control instruction associated with the valid preset hand action;

[0129] Or, determine the control instruction associated with the preset hand movement.

[0130] Specifically, there is no limitation on the specific structures of the action memory, feature identifier, feature parser and response processor, which themselves or each unit therein can be composed of logic components or a combination of logic components, and the logic components include a field programmable processor, a computer or a microprocessor in a computer.

[0131] Specifically, the motion monitor can be a visual acquisition component, such as a photographic device, etc. It only needs to be able to collect preset hand data, which will not be elaborated here.

[0132] Specifically, after determining the control instruction, the corresponding smart home device can be controlled based on the control instruction, which will not be described in detail.

[0133] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. A smart home control method, characterized in that: include: Storing preset hand motion data of the monitored subject, and simultaneously identifying morphological features corresponding to the preset hand motions, including trajectory features of the preset hand motions and action duration of the preset hand motions; Comparing the preset hand action corresponding morphological features with a plurality of pre-stored normal action corresponding morphological features, so as to determine the action matching features for the plurality of normal action corresponding morphological features based on the comparison result, including the action matching degree and the action time-consuming deviation value; Calculating a motion fuzzy representation coefficient for the preset hand motion according to the motion matching feature to classify the fuzzy misjudgment tendency for the preset hand motion; Collect the current hand motion data of the monitored subject in real time to continuously obtain recognition results for hand motions; In response to the recognition result being a preset hand motion, the preset hand motion is analyzed based on the fuzzy misjudgment tendency of the preset hand motion to determine a control instruction for a preset home appliance, including: Determine the time domain segment in which the preset hand action is generated, extract the morphological features corresponding to the hand action data in the time domain segment and the corresponding adjacent time domain segment, analyze the action coherence based on each morphological feature to determine whether the preset hand action is valid, and determine the control instruction associated with the valid preset hand action; or, determining a control instruction associated with the preset hand movement; The process of determining the motion matching features for several normal motion behavior motion features includes: The average of the trajectory similarities between the trajectory features corresponding to the preset hand movements and the trajectory features corresponding to each normal movement is taken as the movement matching degree; The average of the difference between the action time required to form the preset hand action and the action time required to form each normal action is taken as the action time deviation value; Among them, the trajectory features are generated based on the contour markers of the hand contour; The process of calculating the motion fuzzy representation coefficient for the preset hand motion according to the motion matching feature includes: The ratio of the action matching degree threshold to the action matching degree is used as the first action feature; The ratio of the action time deviation value to the action time deviation threshold is used as the second action feature; Determine a motion fuzzy representation coefficient for the preset hand motion by weighted summing the first motion feature and the second motion feature; Classify the tendency of fuzzy misjudgment for the hand movement, including: If the motion blur representation coefficient is greater than or equal to the motion blur representation coefficient threshold, determining that the blur misjudgment tendency of the preset hand motion is high; If the motion blur representation coefficient is less than the motion blur representation coefficient threshold, it is determined that the blur misjudgment tendency of the preset hand motion is low.

2. The smart home control method according to claim 1, characterized in that: The process of continuously obtaining the recognition result for the hand movement includes: Determine the trajectory characteristics corresponding to the current hand movement; Comparing the trajectory feature with the trajectory feature corresponding to the preset hand movement to determine the similarity, so as to determine whether the hand movement is the preset hand movement; If the similarity is greater than a preset similarity threshold, the hand movement is determined to be a preset hand movement.

3. The smart home control method according to claim 1, characterized in that: Analyze the preset hand movements based on the fuzzy misjudgment tendency of the preset hand movements to determine control instructions for preset home appliances, include, If the fuzzy misjudgment tendency of the preset hand action is high, then determine the time domain segment in which the preset hand action is generated, extract the morphological features corresponding to the hand action data in the generated time domain segment and the corresponding adjacent time domain segment, analyze the action coherence based on each morphological feature to determine whether the preset hand action is valid, and determine the control instruction associated with the valid preset hand action; If the fuzzy misjudgment tendency of the preset hand action is low, a control instruction associated with the preset hand action is determined.

4. The smart home control method according to claim 1, characterized in that: The process of analyzing the coherence of movements based on the characteristics of each morphology includes: Determine the range of motion corresponding to each morphological feature; The average difference between the action amplitude corresponding to the generated time domain segment and the action amplitude corresponding to the adjacent time domain segment is calculated, and the reciprocal of the average difference is determined as the action coherence.

5. The smart home control method according to claim 4, characterized in that: The process of determining the amplitude of the motion includes, Determine the starting point of the trajectory feature, and construct a plurality of rays passing the trajectory feature with the starting point as an endpoint; Determine the angle between each ray and the horizontal plane, and determine the maximum angle as the motion amplitude.

6. The smart home control method according to claim 1, characterized in that: Determining whether the preset hand motion is valid includes: If the motion continuity is less than the motion continuity threshold, it is determined that the preset hand motion is valid.

7. A controller using the smart home control method according to any one of claims 1 to 6, characterized in that: include: An action memory, used to store preset hand action data of the monitored subject, and simultaneously identify the morphological features corresponding to the preset hand actions, including the trajectory features of the preset hand actions and the action duration of the preset hand actions; a feature identifier connected to the action memory, for comparing the preset hand action corresponding morphological features with a plurality of pre-stored normal action corresponding morphological features, so as to determine action matching features for the plurality of normal action corresponding morphological features based on the comparison result, including action matching degree and action time-consuming deviation value; A feature analyzer connected to the feature identifier, for calculating a motion fuzzy characterization coefficient for the preset hand motion according to the motion matching feature, so as to classify the fuzzy misjudgment tendency for the preset hand motion; A motion monitor connected to the motion memory for collecting the current hand motion data of the monitoring subject in real time to continuously obtain recognition results for the hand motion; A response processor is connected to the feature analyzer and the action monitor respectively, and is used to analyze the preset hand action based on the fuzzy misjudgment tendency of the preset hand action in response to the recognition result being the preset hand action to determine the control instruction for the preset home appliance, including: Determine the time domain segment in which the preset hand action is generated, extract the morphological features corresponding to the hand action data in the time domain segment and the corresponding adjacent time domain segment, analyze the action coherence based on each morphological feature to determine whether the preset hand action is valid, and determine the control instruction associated with the valid preset hand action; Or, determine the control instruction associated with the preset hand movement.

Images