Sitting posture recognition and correction control method and device and intelligent seat

Abstract

The invention provides a sitting posture recognition and correction control method and device and an intelligent seat. The sitting posture recognition and correction control method is applied to the intelligent seat, a plurality of piezoelectric sensors are arranged on the intelligent seat, and the method comprises the steps that actual piezoelectric signals of the piezoelectric sensors are collected, and the existence condition of people on the intelligent seat is judged according to the actual piezoelectric signals; entering a sitting posture recognition and correction mode under the condition that a person exists on the intelligent seat; in the sitting posture recognition and correction mode, according to the actual piezoelectric signal and based on a pre-established sitting posture recognition model, determining the actual sitting posture of the user corresponding to the actual piezoelectric signal; and performing sitting posture correction reminding on the user according to the actual sitting posture of the user. According to the method, the user can be reminded in time when the poor sitting posture of the user is accurately recognized, so that the user is helped to improve living habits and increase exercises, the problem of poor posture of the user is fundamentally solved, and in addition, the steps are few, the logic is simple, and large-scale popularization and use are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of smart home, and in particular, to a sitting posture recognition and correction control method, a control device, and a smart seat. Background Art

[0002] In the related art, among the current sitting posture correctors on the market, such as correction belts, the (publicized) functions are usually stretching the cervical vertebra, opening the shoulders and spreading the back, gathering the chest, and straightening the posture, etc. Simply put, it is to use an external force to open the user's shoulders, push the user's back, and tie the user's waist tightly, so as to gradually correct the stooped posture of the user. However, the effect of the external force on the human body is very limited, and it cannot really solve the problem of bad posture.

[0003] In terms of the muscle principle, the problem of bad posture is that the muscles in the front are too tight, such as the pectoralis minor, anterior deltoid, middle deltoid, trapezius muscle, etc. being tense. When the front is overly tense, it will cause the back to be overly stretched. That is to say, the muscles in the back are stretched, which is equivalent to stretching the back muscle group at all times. There are numerous news reports about the "thoracic lordosis and kyphosis" caused by the back brace. The main reason for the user's stooped posture is due to living habits, lack of exercise, etc. If the muscles that should exert force are not allowed to exert force, they will gradually become weak, and the final result is only harmful to the user's body and not beneficial. Therefore, there is an urgent need for a smart home in the prior art to timely remind the user when it accurately recognizes that the user has a bad sitting posture, so as to help the user improve living habits, increase exercise, and thus fundamentally solve the problem of the user's bad posture. Summary of the Invention

[0004] The present invention provides a sitting posture recognition and correction control method, a control device, and a smart seat to solve the defects in the prior art and achieve the following technical effects: it can timely remind the user when it accurately recognizes that the user has a bad sitting posture, so as to help the user improve living habits, increase exercise, and thus fundamentally solve the problem of the user's bad posture. In addition, the steps are fewer, the logic is simple, and it occupies less computing power of the controller and has lower requirements for hardware, which is suitable for large-scale popular use.

[0005] The sitting posture recognition and correction control method according to the first aspect embodiment of the present invention is applied to a smart seat, and a plurality of piezoelectric sensors are provided on the smart seat. The method includes:

[0006] Collect the actual piezoelectric signals of the piezoelectric sensors, and judge the presence of a person on the smart seat according to the actual piezoelectric signals;

[0007] When there is a person on the smart seat, enter the sitting posture recognition and correction mode;

[0008] In the sitting posture recognition and correction mode, based on the actual piezoelectric signal and the pre-established sitting posture recognition model, determine the actual sitting posture of the user corresponding to the actual piezoelectric signal;

[0009] Give a sitting posture correction reminder to the user according to the actual sitting posture of the user.

[0010] According to an embodiment of the present invention, the sitting posture recognition model refers to a deep learning model in which the piezoelectric signal and the sitting posture image are in one-to-one correspondence, and the steps for establishing the sitting posture recognition model are as follows:

[0011] Let the testers assume various sitting postures on the intelligent seat;

[0012] Use the piezoelectric sensor to collect and obtain the piezoelectric signals in each different sitting posture, and use the image recognition device to collect and obtain the sitting posture images in each different sitting posture;

[0013] Based on all the collected piezoelectric signals, all the sitting posture images, and the corresponding relationships between the above two and different sitting postures, establish the sitting posture recognition model based on the deep learning method;

[0014] Among them, in the sitting posture recognition model, the piezoelectric signal, the sitting posture image, and the sitting posture are in one-to-one correspondence.

[0015] According to an embodiment of the present invention, the step of using the image recognition device to collect and obtain the sitting posture images in each different sitting posture specifically includes:

[0016] Use a camera to take sitting posture pictures in each different sitting posture, and perform image processing operations on the sitting posture pictures to generate the sitting posture images.

[0017] According to an embodiment of the present invention, in the step of using a camera to take sitting posture pictures in each different sitting posture and performing image processing operations on the sitting posture pictures to generate the sitting posture images, the image processing operations specifically include:

[0018] Extract the basic human figure in the sitting posture picture corresponding to the sitting posture;

[0019] Extract the skeletal key points of the basic human figure;

[0020] Generate a sitting posture image corresponding to the sitting posture according to the skeletal key points.

[0021] According to an embodiment of the present invention, the step of generating a sitting posture image corresponding to the sitting posture according to the skeletal key points specifically includes:

[0022] Based on all the bone key points and the human body structure framework, obtain the forward contour and lateral contour of the spinal trunk of the person in this sitting posture.

[0023] Based on the forward contour and lateral contour of the spinal trunk, generate a sitting posture image corresponding to this sitting posture, and the sitting posture image is a three-dimensional sitting posture image.

[0024] According to an embodiment of the present invention, in the sitting posture recognition and correction mode, the step of determining the actual sitting posture of the user corresponding to the actual piezoelectric signal based on the actual piezoelectric signal and a pre-established sitting posture recognition model specifically includes:

[0025] In the sitting posture recognition and correction mode, search for a sitting posture image corresponding to the actual piezoelectric signal in the sitting posture recognition model, and use this sitting posture image as the actual sitting posture of the user at present.

[0026] According to an embodiment of the present invention, after the step of searching for a sitting posture image corresponding to the actual piezoelectric signal in the sitting posture recognition model and using this sitting posture image as the actual sitting posture of the user at present in the sitting posture recognition and correction mode, it specifically further includes:

[0027] Use an image recognition device to collect the actual sitting posture picture of the user;

[0028] According to the actual sitting posture picture, correct the sitting posture image corresponding to the actual piezoelectric signal, and use the corrected sitting posture image as the actual sitting posture of the user.

[0029] According to an embodiment of the present invention, the step of giving a sitting posture correction reminder to the user according to the actual sitting posture of the user specifically includes:

[0030] According to the actual sitting posture of the user, determine the spinal trunk information of the user at present;

[0031] If it is determined that the spinal trunk information meets the bad sitting posture condition, then give different sitting posture correction reminders to the user according to the difference of the spinal trunk information.

[0032] According to an embodiment of the present invention, the spinal trunk information includes the degree of spinal trunk curvature, the degree of spinal trunk torsion, and the relative height of the spinal trunk;

[0033] And the bad sitting posture condition includes: the degree of spinal trunk curvature is greater than the set standard curvature, the degree of spinal trunk torsion is greater than the set standard torsion, and / or the relative height of the spinal trunk is not within the set standard height range.

[0034] A sitting posture recognition and correction control device according to an embodiment of the second aspect of the present invention is applied to an intelligent seat. A plurality of piezoelectric sensors are provided on the intelligent seat. The control device includes:

[0035] An acquisition module, configured to collect the actual piezoelectric signals of the piezoelectric sensors and determine the presence of a person on the intelligent seat according to the actual piezoelectric signals;

[0036] A first control module, configured to enter a sitting posture recognition and correction mode when there is a person on the intelligent seat;

[0037] A second control module, configured to determine the actual sitting posture of the user corresponding to the actual piezoelectric signal according to the actual piezoelectric signal and based on a pre-established sitting posture recognition model in the sitting posture recognition and correction mode;

[0038] A reminder module, configured to give a sitting posture correction reminder to the user according to the actual sitting posture of the user.

[0039] An intelligent seat according to an embodiment of the third aspect of the present invention includes:

[0040] A base and a backrest, and a plurality of piezoelectric sensors are provided on the base;

[0041] A memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the sitting posture recognition and correction control method according to the embodiment of the first aspect of the present invention.

[0042] The present invention provides a sitting posture recognition and correction control method, which is mainly applied to smart homes such as intelligent seats. By collecting the actual piezoelectric signals of different parts of the intelligent seat through piezoelectric sensors, when it is determined that there is a person sitting on the intelligent seat, it enters the sitting posture recognition and correction mode, and in this mode, based on a pre-established sitting posture recognition model and the previously obtained actual piezoelectric signals, the current actual sitting posture of the user (i.e., the actual sitting posture of the user) is obtained, so as to give a corresponding sitting posture correction reminder to the user according to the actual sitting posture of the user. That is, this method can timely remind the user when it accurately recognizes that the user has a bad sitting posture, thereby helping the user improve living habits, increase exercise, and fundamentally solve the problem of the user's poor body posture.

[0043] In addition, it should be noted that this method does not need to continuously collect and analyze the current posture pictures of the user, but by the method of pre-establishing a sitting posture recognition model, an association is established between the piezoelectric signal and the user's sitting posture, so that in actual applications, the piezoelectric signal can be directly collected and the current actual sitting posture of the user can be directly and simply obtained according to the piezoelectric signal. In this way, this method has fewer steps, simple logic, and less occupation of the computing power of the controller, and has low requirements for hardware, and is suitable for large-scale popularization and use. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0045] Figure 1 is a schematic flowchart of the sitting posture recognition and correction control method provided by the present invention;

[0046] Figure 2 is a schematic structural diagram of the sitting posture recognition and correction control device provided by the present invention;

[0047] Figure 3 is a schematic structural diagram of the intelligent seat provided by the present invention;

[0048] Figure 4 is a schematic structural diagram of the electronic device provided by the present invention.

[0049] Reference numerals:

[0050] 1. Intelligent seat; 2. Piezoelectric sensor. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0051] To make the objectives, technical solutions and advantages of the present invention clearer, the following will clearly and completely describe the technical solutions in the present invention in conjunction with the drawings in the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments in the present invention belong to the scope of protection of the present invention.

[0052] In the description of this specification, the descriptions referring to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. mean that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the embodiments of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without conflict, those skilled in the art can combine the different embodiments or examples described in this specification and the features of the different embodiments or examples.

[0053] The following describes the sitting posture recognition and correction control method, sitting posture recognition and correction control device, and intelligent seat proposed by the present invention with reference to the accompanying drawings. Among them, before elaborating on the embodiments of the present invention in detail, the entire application scenario will be described first. The sitting posture recognition and correction control method, sitting posture recognition and correction control device, electronic device, and computer-readable storage medium of the embodiments of the present invention can be applied not only to the local intelligent seat but also to the cloud platform in the Internet field, or the cloud platform in other types of Internet fields, or can also be applied to third-party devices. Among them, the third-party devices may include various different types such as mobile phones, tablet computers, notebooks, in-vehicle computers, and other intelligent terminals.

[0054] The following takes the sitting posture recognition and correction control method applicable to an intelligent seat as an example for illustration. It should be understood that the control method of the embodiments of the present invention can also be applicable to the cloud platform and third-party devices.

[0055] Before introducing the sitting posture recognition and correction control method of the first aspect embodiment of the present invention, it should also be pointed out that the structural basis for implementing this sitting posture recognition and correction control method generally lies in the intelligent seat 1, and a number of piezoelectric sensors 2 need to be provided on the intelligent seat 1 to achieve the detection and transmission of piezoelectric signals.

[0056] As Figure 1 shown, according to the sitting posture recognition and correction control method of the embodiments of the present invention, it includes:

[0057] Step S1, collect the actual piezoelectric signals of the piezoelectric sensors 2, and judge the presence of a person on the intelligent seat 1 according to the actual piezoelectric signals;

[0058] Step S2, when there is a person on the intelligent seat 1, enter the sitting posture recognition and correction mode;

[0059] Step S3, in the sitting posture recognition and correction mode, according to the actual piezoelectric signals and based on the pre-established sitting posture recognition model, determine the actual sitting posture of the user corresponding to the actual piezoelectric signals;

[0060] Step S4, give a sitting posture correction reminder to the user according to the actual sitting posture of the user.

[0061] Taking the example of applying the sitting posture recognition and correction control method according to an embodiment of the present invention to an intelligent seat 1 having a number of piezoelectric sensors 2, its specific working process is as follows: The piezoelectric sensors 2 are arranged at various different positions of the intelligent seat 1. The piezoelectric sensors 2 detect the actual piezoelectric signals at different positions on the intelligent seat 1 and transmit the actual piezoelectric signals to the controller of the intelligent seat 1. The controller receives the actual piezoelectric signals and further analyzes and judges the actual piezoelectric signals, thereby judging the presence of a person on the current intelligent seat 1, that is, judging whether there is a person sitting on the current intelligent seat 1.

[0062] Further, when it is judged that there is a person sitting on the current intelligent seat 1, the controller immediately controls the intelligent seat 1 to enter the sitting posture recognition and correction mode. In the sitting posture recognition and correction mode, the controller will continuously obtain and monitor the actual sitting posture of the user and give corresponding reminders according to whether the actual sitting posture of the user is standard. When the actual sitting posture of the user is not standard or bad, the controller will further give different degrees and different forms of sitting posture correction reminders according to the degree or level of the bad actual sitting posture of the user.

[0063] For example, when the actual sitting posture of the user is standard, the intelligent seat 1 does not give a sitting posture correction reminder to the user; while when the actual sitting posture of the user is not standard or bad, the intelligent seat 1 will give a sitting posture correction reminder to the user based on its own reminder module, and the controller will also classify the actual sitting posture of the user into different types of bad sitting postures and different degrees of bad sitting postures, and give different reminders according to different types of bad sitting postures and different degrees of bad sitting postures. Specifically, when the actual sitting posture of the user is that the trunk spine has too high a bending degree, the reminder module will specifically send a reminder signal to remind the user to straighten the trunk; when the actual sitting posture of the user is that the relative height of the trunk spine is too high, the reminder module will specifically send a reminder signal to remind the user to sit lower appropriately.

[0064] Among them, the method for the controller to obtain the actual sitting posture of the user is as follows: Obtain the current actual piezoelectric signal of the user, and based on a pre-established sitting posture recognition model, find the posture corresponding to the actual piezoelectric signal in the above-mentioned sitting posture recognition model as the actual sitting posture of the user.

[0065] It should be explained that the above-mentioned sitting posture recognition model refers to a model established based on deep learning algorithms and various statistical data. The sitting posture recognition model includes an independent variable parameter of piezoelectric signals and a dependent variable parameter of user postures. When the specific piezoelectric signal is known, through the one-to-one correspondence relationship between the piezoelectric signals and postures in the sitting posture recognition model, the user posture corresponding to the actual piezoelectric signal can be obtained as the actual user posture.

[0066] In the related art, the functions (publicized) of current sitting posture correctors on the market, such as correction belts, are usually to stretch the cervical vertebrae, open the shoulders and expand the back, gather the chest, and straighten the posture, etc. Simply put, it is to use external force to open the user's shoulders, push the user's back and tie the user's waist tightly, so as to gradually correct the stooped posture of the user. However, the effect of external force on the human body is very limited, and it cannot really solve the problem of bad posture.

[0067] In terms of the muscle principle, the problem of bad posture is that the muscles in the front are too tight, such as the pectoralis minor, anterior deltoid, middle deltoid, trapezius muscle, etc. being tense. When the front is overly tense, it will cause the back to be overly stretched. That is to say, the muscles in the back are stretched, which is equivalent to stretching the back muscle group at all times. There are numerous news reports about "thoracic lordosis and thoracic kyphosis" caused by back braces. The main reason for a person to be stooped is due to living habits, lack of exercise, etc. If the muscles that should exert force are not allowed to exert force, they will gradually become weak, and the final result will only do harm to the user's body and bring no benefits. Therefore, there is an urgent need for a smart home device in the existing technology that can timely remind the user when it accurately identifies that the user has a bad sitting posture, so as to help the user improve living habits, increase exercise, and fundamentally solve the problem of the user's bad posture.

[0068] In summary, to solve the technical defects in the above-mentioned related art, the present invention provides a sitting posture recognition and correction control method, which is mainly applied to smart home devices such as smart seats 1. The method collects the actual piezoelectric signals of different parts of the smart seat 1 through piezoelectric sensors 2. When it is determined that someone is sitting on the smart seat 1, it enters the sitting posture recognition and correction mode, and in this mode, based on the pre-established sitting posture recognition model and the previously obtained actual piezoelectric signals, it obtains the user's current actual sitting posture (i.e., the user's actual sitting posture), so as to give corresponding sitting posture correction reminders to the user according to the user's actual sitting posture. That is to say, this method can timely remind the user when it accurately identifies that the user has a bad sitting posture, so as to help the user improve living habits, increase exercise, and fundamentally solve the problem of the user's bad posture.

[0069] In addition, it should be pointed out that this method does not need to continuously collect and analyze the user's current posture pictures. Instead, by pre-establishing a sitting posture recognition model, it establishes an association between piezoelectric signals and the user's sitting posture. Thus, in actual applications, it can directly collect piezoelectric signals and simply obtain the user's current actual sitting posture according to the piezoelectric signals. In this way, this method has fewer steps, simple logic, occupies less computing power of the controller, and has lower requirements for hardware, making it suitable for large-scale popularization and use.

[0070] According to some embodiments of the present invention, the sitting posture recognition model refers to a pre-established deep learning model in which piezoelectric signals and sitting posture images correspond one by one. The steps for establishing the sitting posture recognition model are as follows:

[0071] The tester assumes various sitting postures on the intelligent seat 1;

[0072] The piezoelectric sensor 2 is used to collect and obtain the piezoelectric signals in each different sitting posture, and the image recognition device is used to collect and obtain the sitting posture images in each different sitting posture;

[0073] According to all the collected piezoelectric signals, all the sitting posture images, and the corresponding relationships between the above two and different sitting postures, and based on the deep learning method, a sitting posture recognition model is established.

[0074] Among them, in the sitting posture recognition model, the piezoelectric signals, the sitting posture images, and the sitting postures correspond one by one.

[0075] As can be seen from the above, the establishment of the sitting posture recognition model requires simulating the sitting postures of multiple groups of users on the intelligent seat 1 in advance, collecting and obtaining the piezoelectric signals and sitting posture images in the state of each sitting posture when constructing each group of sitting postures, and then training based on the data in which the above multiple groups of piezoelectric signals and sitting posture images correspond one by one. Finally, based on the deep learning method, the sitting posture recognition model is established to achieve the one-to-one correspondence of the piezoelectric signals, the sitting posture images, and the sitting postures in the model.

[0076] Furthermore, the step of using the image recognition device to collect and obtain the sitting posture images in each different sitting posture specifically includes:

[0077] Using a camera to take pictures of the sitting postures in each different sitting posture, and performing image processing operations on the sitting posture pictures to generate sitting posture images.

[0078] It should be noted that the above sitting posture images and sitting posture pictures are two completely different concepts. Among them, the sitting posture images are not the sitting posture pictures directly taken by the camera, but the calibrated images of the refined user postures that need to be processed based on the sitting posture pictures.

[0079] For the convenience of understanding, a specific embodiment of a method for generating sitting posture images will be given below.

[0080] In a specific embodiment of the present invention, in the step of using a camera to take pictures of the sitting postures in each different sitting posture and performing image processing operations on the sitting posture pictures to generate sitting posture images, the image processing operations specifically include:

[0081] Extracting the basic portrait in the sitting posture picture corresponding to the sitting posture;

[0082] Extracting the skeletal key points of the basic portrait;

[0083] Generate a sitting posture image corresponding to the sitting posture based on the skeletal key points.

[0084] In this embodiment, after the sitting posture picture corresponding to the sitting posture is captured by the camera, the sitting posture picture needs to be processed to finally obtain the sitting posture image. Among them, in the image processing operation, first, the outline of the basic portrait in the sitting posture picture needs to be outlined completely and the basic portrait needs to be extracted. After the basic portrait is extracted, further capture the skeletal key points in the basic portrait. Specifically, the skeletal key points are all the key points that can completely represent the user's posture obtained after image analysis and recognition. The skeletal key points include but are not limited to: left ear bone, right ear bone, left eye socket, right eye socket, cervical vertebra, spine, lumbar vertebra, left shoulder, left elbow, left wrist, right shoulder, right elbow, right wrist, left hip, middle hip, left knee, right knee, left ankle and right ankle, etc. The present invention does not make specific limitations on the type and quantity of the skeletal key points, and the extraction of the skeletal key points needs to be determined according to the actual situation.

[0085] After the extraction of the skeletal key points is completed, the controller further generates a sitting posture image corresponding to the sitting posture according to the skeletal key points. At this time, the sitting posture image deletes the useless image data and only retains the image data that most directly and accurately reflects the user's sitting posture. In this way, the recognition of the user's sitting posture is more accurate and concise, and the concise sitting posture image can also better reflect whether the user's sitting posture is bad, as well as the degree and level of badness, which is convenient for subsequent reminder and correction operations.

[0086] Furthermore, the steps of generating a sitting posture image corresponding to the sitting posture according to the skeletal key points specifically include:

[0087] According to all the skeletal key points and based on the human body structure framework, obtain the forward contour and lateral contour of the spinal trunk of the person in this sitting posture;

[0088] Based on the forward contour and lateral contour of the spinal trunk, generate a sitting posture image corresponding to this sitting posture, and the sitting posture image is a three-dimensional sitting posture image.

[0089] In this way, through the above-mentioned forward contour and lateral contour of the spinal trunk, a sitting posture image in the form of a three-dimensional image can be drawn. Compared with the sitting posture picture in the plane structure, the sitting posture image in the three-dimensional form can more accurately reflect whether the user's sitting posture is bad, and can also reflect various details of the user's sitting posture, which is convenient for the reminder module to give more targeted sitting posture correction reminders according to the different sitting posture details.

[0090] It should also be noted that there may be more than one sitting posture picture corresponding to a sitting posture, but there is definitely only one sitting posture image corresponding to a sitting posture. Specifically, cameras can be placed at multiple positions of the intelligent seat 1, and multiple groups of sitting posture pictures can be taken using the cameras at different positions. Each group of sitting posture pictures can extract a group of skeletal key points. Finally, based on the multiple groups of skeletal key points obtained from the sitting posture pictures taken from multiple positions, the sitting posture image corresponding to this sitting posture can be accurately obtained.

[0091] That is to say, one posture corresponds to several sitting posture pictures, but one posture only corresponds to one sitting posture image, and one posture also only corresponds to one piezoelectric signal.

[0092] The above embodiments are the introductions to the methods for obtaining sitting posture images in the present invention. However, the above embodiments do not limit the scope of the present invention. Sitting posture images can also be obtained based on other technologies, and the present invention does not make special limitations here.

[0093] According to some embodiments of the present invention, in the sitting posture recognition and correction mode, the steps of determining the actual sitting posture of the user corresponding to the actual piezoelectric signal according to the actual piezoelectric signal and based on the pre-established sitting posture recognition model specifically include:

[0094] In the sitting posture recognition and correction mode, find the sitting posture image corresponding to the actual piezoelectric signal in the sitting posture recognition model, and use this sitting posture image as the actual sitting posture of the user at present.

[0095] Further, after the step of finding the sitting posture image corresponding to the actual piezoelectric signal in the sitting posture recognition model and using this sitting posture image as the actual sitting posture of the user at present in the sitting posture recognition and correction mode, it specifically further includes:

[0096] Use the image recognition device to collect the actual sitting posture pictures of the user;

[0097] According to the actual sitting posture pictures, correct the sitting posture image corresponding to the actual piezoelectric signal, and use the corrected sitting posture image as the actual sitting posture of the user.

[0098] In this way, by collecting the actual sitting posture pictures of the user again and correcting the sitting posture image according to the actual sitting posture pictures, the accurate recognition of the actual sitting posture of the user can be further realized, and the accuracy of the program result can be further ensured.

[0099] According to some embodiments of the present invention, the steps of giving sitting posture correction reminders to the user according to the actual sitting posture of the user specifically include:

[0100] According to the actual sitting posture of the user, determine the current spinal trunk information of the user;

[0101] If it is determined that the spinal trunk information meets the bad sitting posture conditions, different sitting posture correction reminders will be given to the user according to the different spinal trunk information.

[0102] It should be noted that since the actual sitting posture of the user is determined based on the sitting posture images in the sitting posture recognition model, and the sitting posture images contain information such as skeletal key points, the forward contour of the spinal trunk, and the lateral contour of the spinal trunk, the corresponding spinal trunk information can be easily obtained according to the actual sitting posture of the user.

[0103] Furthermore, the spinal trunk information includes the degree of spinal trunk curvature, the degree of spinal trunk torsion, and the relative height of the spinal trunk.

[0104] And the bad sitting posture conditions include: the degree of spinal trunk curvature is greater than the set standard curvature, the degree of spinal trunk torsion is greater than the set standard torsion, and / or the relative height of the spinal trunk is not within the set standard height range.

[0105] Specifically, when it is determined that at least one of the following three bad sitting posture conditions is met, it is proved that the spinal trunk information meets the bad sitting posture conditions. Among them, the three bad sitting posture conditions are: the degree of spinal trunk curvature is greater than the set standard curvature, the degree of spinal trunk torsion is greater than the set standard torsion, and the relative height of the spinal trunk is not within the set standard height range.

[0106] After determining that the spinal trunk information meets the bad sitting posture conditions, the controller will further give targeted sitting posture correction reminders to the user according to the different degrees of the spinal trunk curvature, the spinal trunk torsion, and / or the relative height of the spinal trunk.

[0107] For example, when the actual sitting posture of the user is that the curvature of the trunk spine is too high, the reminder module will specifically send a reminder signal to remind the user to straighten the trunk; when the actual sitting posture of the user is that the relative height of the trunk spine is too high, the reminder module will specifically send a reminder signal to remind the user to sit lower appropriately; when the actual sitting posture of the user is that the degree of spinal trunk torsion is too large, the reminder module will specifically send a reminder signal to remind the user to twist appropriately to straighten the body.

[0108] Even further, on the premise of meeting the bad sitting posture conditions, based on the different ranges of the intervals where the spinal trunk curvature, the spinal trunk torsion, and the relative height of the spinal trunk are located, the controller will determine different bad user sitting postures as different degrees of badness, and the reminder module will also send reminder signals of different degrees according to different degrees of badness.

[0109] Taking the curvature of the torso spine as an example, if the curvature of the torso spine is greater than the first set curvature and less than or equal to the second set curvature, then the degree of badness of the user's actual sitting posture is relatively low at this time. Therefore, the reminder module issues a relatively mild sitting posture correction reminder (such as a yellow light staying on); if the curvature of the torso spine is greater than the second set curvature and less than or equal to the third set curvature, then the degree of badness of the user's actual sitting posture is relatively high at this time. Therefore, the reminder module issues a relatively high-level sitting posture correction reminder (such as a red light staying on); if the curvature of the torso spine is greater than the third set curvature, then the degree of badness of the user's actual sitting posture is extremely high at this time. Therefore, the reminder module issues an extremely high-level sitting posture correction reminder (such as a red light flashing and a harsh alarm sound).

[0110] Among them, the above-mentioned first set curvature, second set curvature, and third set curvature are all greater than the above-mentioned set standard curvature.

[0111] The sitting posture recognition and correction control device provided by the present invention will be described below. The sitting posture recognition and correction control device described below can be mutually referred to the sitting posture recognition and correction control method described above.

[0112] As Figure 2 shown, the sitting posture recognition and correction control device according to the second aspect embodiment of the present invention is applied to the intelligent seat 1. A plurality of piezoelectric sensors 2 are provided on the intelligent seat 1. The control device includes:

[0113] An acquisition module 110, configured to collect the actual piezoelectric signals of the piezoelectric sensors 2, and judge the presence of a person on the intelligent seat 1 according to the actual piezoelectric signals;

[0114] A first control module 120, configured to enter the sitting posture recognition and correction mode when there is a person on the intelligent seat 1;

[0115] A second control module 130, configured to determine the user's actual sitting posture corresponding to the actual piezoelectric signal according to the actual piezoelectric signal and based on a pre-established sitting posture recognition model in the sitting posture recognition and correction mode;

[0116] A reminder module 140, configured to give a sitting posture correction reminder to the user according to the user's actual sitting posture.

[0117] As Figure 3 shown, the intelligent seat 1 according to the third aspect embodiment of the present invention includes a base and a backrest. A plurality of piezoelectric sensors 2 are provided on the base, and further includes a control device. The control device is configured to execute the sitting posture recognition and correction control method described in the first aspect embodiment of the present invention.

[0118] The intelligent seat 1 and the sitting posture recognition and correction control device according to the embodiments of the present invention collect the actual piezoelectric signals of different parts of the intelligent seat 1 through the piezoelectric sensor 2. When it is determined that there is a person sitting on the intelligent seat 1, it enters the sitting posture recognition and correction mode. In this mode, based on the pre-established sitting posture recognition model and the previously obtained actual piezoelectric signals, the current actual sitting posture of the user (i.e., the actual sitting posture of the user) is obtained. Thus, corresponding sitting posture correction reminders are given to the user according to the actual sitting posture of the user. That is to say, this method can timely remind the user when it accurately recognizes that the user has a bad sitting posture, thereby helping the user improve living habits, increase exercise, and fundamentally solve the problem of the user's poor body posture.

[0119] In addition, it should be pointed out that the present invention does not need to continuously collect and analyze the current posture pictures of the user. Instead, by the method of pre-establishing a sitting posture recognition model, the piezoelectric signal is associated with the user's sitting posture. Thus, in practical applications, the piezoelectric signal can be directly collected and the current actual sitting posture of the user can be simply obtained directly according to the piezoelectric signal. In this way, this method has fewer steps, simple logic, occupies less computing power of the controller, and has lower requirements for hardware, and is suitable for large-scale popular use.

[0120] Figure 4 An example of the schematic physical structure of an electronic device is shown as Figure 4 shown. The electronic device may include: a processor 810, a communication interface 820, a memory 830, and a communication bus 840. Among them, the processor 810, the communication interface 820, and the memory 830 complete mutual communication through the communication bus 840. The processor 810 can call the logical instructions in the memory 830 to execute the sitting posture recognition and correction control method, and this method includes: collecting the actual piezoelectric signals of the piezoelectric sensor 2, and judging the presence of personnel on the intelligent seat 1 according to the actual piezoelectric signals; when there is a person on the intelligent seat 1, entering the sitting posture recognition and correction mode; in the sitting posture recognition and correction mode, according to the actual piezoelectric signals and based on the pre-established sitting posture recognition model, determining the actual sitting posture of the user corresponding to the actual piezoelectric signals; and giving sitting posture correction reminders to the user according to the actual sitting posture of the user.

[0121] In addition, when the logical instructions in the above-mentioned memory 830 can be implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes.

[0122] On the other hand, the present invention also provides a computer program product. The computer program product includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the sitting posture recognition and correction control method provided by the above-mentioned various methods. The method includes: collecting the actual piezoelectric signal of the piezoelectric sensor 2 and judging the presence of a person on the intelligent seat 1 according to the actual piezoelectric signal; when there is a person on the intelligent seat 1, entering the sitting posture recognition and correction mode; in the sitting posture recognition and correction mode, according to the actual piezoelectric signal and based on a pre-established sitting posture recognition model, determining the actual sitting posture of the user corresponding to the actual piezoelectric signal; and giving a sitting posture correction reminder to the user according to the actual sitting posture of the user.

[0123] On another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it realizes the sitting posture recognition and correction control method provided by the above-mentioned various methods. The method includes: collecting the actual piezoelectric signal of the piezoelectric sensor 2 and judging the presence of a person on the intelligent seat 1 according to the actual piezoelectric signal; when there is a person on the intelligent seat 1, entering the sitting posture recognition and correction mode; in the sitting posture recognition and correction mode, according to the actual piezoelectric signal and based on a pre-established sitting posture recognition model, determining the actual sitting posture of the user corresponding to the actual piezoelectric signal; and giving a sitting posture correction reminder to the user according to the actual sitting posture of the user.

[0124] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0125] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of each embodiment or some parts of the embodiments.

[0126] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features. And these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A sitting posture recognition and correction control method is applied to an intelligent seat, and a plurality of piezoelectric sensors are provided on the intelligent seat. Characterized in that, The method includes: Collect the actual piezoelectric signals of the piezoelectric sensors, and judge the presence of personnel on the intelligent seat according to the actual piezoelectric signals; When there is a person on the intelligent seat, enter the sitting posture recognition and correction mode; In the sitting posture recognition and correction mode, according to the actual piezoelectric signals and based on a pre-established sitting posture recognition model, determine the actual sitting posture of the user corresponding to the actual piezoelectric signals; Give a sitting posture correction reminder to the user according to the actual sitting posture of the user.

2. The sitting posture recognition and correction control method according to claim 1, Characterized in that, The sitting posture recognition model refers to a deep learning model in which the piezoelectric signals and sitting posture images are in one-to-one correspondence established in advance. The steps for establishing the sitting posture recognition model are as follows: The testers assume various sitting postures on the intelligent seat; Use the piezoelectric sensors to collect and obtain the piezoelectric signals in each different sitting posture, and use an image recognition device to collect and obtain the sitting posture images in each different sitting posture; According to all the collected piezoelectric signals, all the sitting posture images, and the corresponding relationships between the above two and different sitting postures, and based on the deep learning method, establish the sitting posture recognition model; Among them, in the sitting posture recognition model, the piezoelectric signals, the sitting posture images, and the sitting postures are in one-to-one correspondence.

3. The sitting posture recognition and correction control method according to claim 2, Characterized in that, The step of using an image recognition device to collect and obtain the sitting posture images in each different sitting posture specifically includes: Use a camera to take sitting posture pictures in each different sitting posture, and perform image processing operations on the sitting posture pictures to generate the sitting posture images.

4. The sitting posture recognition and correction control method according to claim 3, Characterized in that, In the step of using a camera to take sitting posture pictures in each different sitting posture and performing image processing operations on the sitting posture pictures to generate the sitting posture images, the image processing operations specifically include: Extract the basic portrait in the sitting posture picture corresponding to the sitting posture; Extract the bone key points of the basic portrait; Generate a sitting posture image corresponding to the sitting posture according to the bone key points.

5. The sitting posture recognition and correction control method according to claim 4, Characterized in that, The step of generating a sitting posture image corresponding to the sitting posture according to the bone key points specifically includes: According to all the bone key points and based on the human body structure framework, obtain the forward contour and lateral contour of the spinal trunk of the person in this sitting posture; Based on the forward contour and lateral contour of the spinal trunk, generate a sitting posture image corresponding to this sitting posture, and the sitting posture image is a three-dimensional sitting posture image.

6. The sitting posture recognition and correction control method according to any one of claims 2 to 5, Characterized in that, In the sitting posture recognition and correction mode, according to the actual piezoelectric signal and based on a pre-established sitting posture recognition model, the step of determining the actual sitting posture of the user corresponding to the actual piezoelectric signal specifically includes: In the sitting posture recognition and correction mode, search for a sitting posture image corresponding to the actual piezoelectric signal in the sitting posture recognition model, and use this sitting posture image as the actual sitting posture of the user at present.

7. The sitting posture recognition and correction control method according to item 6, characterized in that, After the step of searching for a sitting posture image corresponding to the actual piezoelectric signal in the sitting posture recognition model and using this sitting posture image as the actual sitting posture of the user at present in the sitting posture recognition and correction mode, it specifically further includes: Use an image recognition device to collect an actual sitting posture picture of the user; According to the actual sitting posture picture, correct the sitting posture image corresponding to the actual piezoelectric signal, and use the corrected sitting posture image as the actual sitting posture of the user.

8. The sitting posture recognition and correction control method according to any one of claims 1 to 5, characterized in that, The step of giving a sitting posture correction reminder to the user according to the actual sitting posture of the user specifically includes: According to the actual sitting posture of the user, determine the current spinal trunk information of the user; If it is determined that the spinal trunk information meets the bad sitting posture condition, then according to the difference of the spinal trunk information, give different sitting posture correction reminders to the user.

9. The sitting posture recognition and correction control method according to claim 8, characterized in that, The spinal trunk information includes the size of the spinal trunk curvature, the spinal trunk torsion degree, and the relative height of the spinal trunk; And the bad sitting posture condition includes: the spinal trunk curvature is greater than the set standard curvature, the spinal trunk torsion degree is greater than the set standard torsion degree, and / or the relative height of the spinal trunk is not within the set standard height range.

10. A sitting posture recognition and correction control device is applied to an intelligent seat, and a plurality of piezoelectric sensors are provided on the intelligent seat, characterized in that, The control device includes: An acquisition module, configured to collect the actual piezoelectric signals of the piezoelectric sensors and judge the presence of a person on the intelligent seat according to the actual piezoelectric signals; A first control module, configured to enter the sitting posture recognition and correction mode when there is a person on the intelligent seat; A second control module, configured to, in the sitting posture recognition and correction mode, according to the actual piezoelectric signal and based on a pre-established sitting posture recognition model, determine the actual sitting posture of the user corresponding to the actual piezoelectric signal; A reminder module, configured to give a sitting posture correction reminder to the user according to the actual sitting posture of the user.

11. An intelligent seat, characterized in that, includes: A base and a backrest, and a plurality of piezoelectric sensors are provided on the base; A memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the program, it implements the sitting posture recognition and correction control method according to any one of claims 1 to 9.

Images