Integrated sitting posture monitoring device for teenager eyesight protection

Abstract

The invention provides an integrated sitting posture monitoring device for teenager eyesight protection, and relates to the technical field of sitting posture monitoring. Comprising a first data acquisition module used for acquiring video stream data of a target object in real time; the second data acquisition module is used for acquiring sensor data of a target object; the real-time sitting posture parameter acquisition module is used for acquiring the real-time sitting posture parameter of the target object according to the sensor data of the video stream data of the target object; the model construction module is used for acquiring historical sitting posture data and body characteristic parameters of a target object and constructing a corresponding standard sitting posture model according to the historical sitting posture data and the body characteristic parameters; the comparison module is used for acquiring standard sitting posture parameters according to the standard sitting posture model and comparing the standard sitting posture parameters with the real-time sitting posture parameters to obtain a sitting posture state; the judgment module is used for judging whether the current sitting posture is an abnormal sitting posture or not according to the sitting posture state. The sitting posture monitoring system solves the problems of absolute monitoring result and misjudgment of the monitoring result in the existing sitting posture monitoring system.

Description

Technical Field

[0001] The invention relates to the technical field of sitting posture monitoring, and in particular to an integrated sitting posture monitoring device for protecting the eyesight of teenagers. Background Art

[0002] As teenagers spend more time using electronic devices, vision problems caused by poor sitting posture are becoming increasingly serious. Existing sitting posture monitoring systems (such as the sitting posture intelligent monitoring method, device, equipment and storage medium with publication number: CN 113378762 B) mainly rely on image processing technology and have the following technical defects:

[0003] By extracting the human shape parameters and face parameters of the target object in the video stream and comparing them with the standard sitting posture parameters, the sitting posture status is output. This method relies on fixed standard sitting posture parameters, resulting in monitoring results that are too absolute and fail to take into account individual differences and dynamic activities;

[0004] The sitting posture monitoring method determines whether the sitting posture is abnormal by analyzing the relative position and posture angle of the human figure and face. This method may misjudge normal body activities (such as turning the head and adjusting the sitting posture) as improper sitting posture during dynamic monitoring.

[0005] In summary, the existing sitting posture monitoring system has the problems of absolute monitoring results and misjudgment of monitoring results. Summary of the invention

[0006] In order to overcome the deficiencies of the prior art, the object of the present invention is to provide an integrated sitting posture monitoring device for protecting the eyesight of teenagers.

[0007] To achieve the above object, the present invention provides the following solutions:

[0008] An integrated sitting posture monitoring device for protecting the eyesight of teenagers, comprising:

[0009] A first data acquisition module, a second data acquisition module, a real-time sitting posture parameter acquisition module, a model building module, a comparison module and a judgment module;

[0010] The first data acquisition module is used to acquire the video stream data of the target object in real time; the second data acquisition module is used to acquire the sensor data of the target object; the real-time sitting posture parameter acquisition module is used to obtain the real-time sitting posture parameters of the target object according to the sensor data of the video stream data of the target object; the model construction module is used to acquire the historical sitting posture data and body feature parameters of the target object and construct a corresponding standard sitting posture model according to the historical sitting posture data and body feature parameters; the comparison module is used to acquire the standard sitting posture parameters according to the standard sitting posture model and compare them with the real-time sitting posture parameters to obtain the sitting posture state; the judgment module is used to judge whether the current sitting posture is an abnormal sitting posture according to the sitting posture state, and if so, set a time threshold according to the historical activity data of the target object, and judge the change of the activity state of the target object after reaching the time threshold to obtain a judgment result, if the judgment result is a change, adjust the standard sitting posture parameters, if the judgment result is no change, the sitting posture of the target object is abnormal.

[0011] Preferably, the second data acquisition module includes:

[0012] Pressure sensor submodule, temperature sensor submodule, humidity sensor submodule, accelerometer;

[0013] The pressure sensors are installed on the seat, tabletop, back cushion and footrest;

[0014] The pressure sensor submodule is used to obtain pressure data of the seat, desktop, back cushion and footrest positions, the temperature sensor submodule is used to obtain temperature data in the environment, the humidity sensor submodule is used to obtain humidity data in the environment, and the accelerometer is used to obtain dynamic sitting posture data of the target object.

[0015] Preferably, the real-time sitting posture parameter acquisition module includes:

[0016] Marking submodule, video data conversion submodule, filtering submodule, key point detection submodule, posture estimation submodule, data fusion submodule;

[0017] The marking submodule is used to timestamp the sensor data and the video stream data, the video conversion submodule is used to convert the video stream data into a grayscale image, the filtering submodule is used to filter the grayscale image to obtain preprocessed image data, the key point detection submodule is used to extract key points based on the preprocessed image data and a real-time computer vision algorithm to obtain key point position data of the target object, the posture estimation submodule is used to calculate the initial sitting posture parameters of the target image based on the key point position data, and the data fusion submodule is used to integrate the initial coordinate parameters and the sensor data to obtain the current sitting posture parameters.

[0018] Preferably, the model building module comprises:

[0019] Historical data acquisition submodule, body feature acquisition submodule, model generation submodule and feedback submodule;

[0020] The historical data acquisition submodule is used to acquire the historical sitting posture data of the current target object, the body feature acquisition submodule is used to acquire the body feature data of the target object, and the body feature data includes: height, weight, leg length, waist circumference, the model generation submodule is used to use a nonlinear regression model to construct a standard sitting posture parameter model according to the body feature data and the historical sitting posture data, and the feedback submodule is used to acquire the target object's perception evaluation in real time and adjust the parameters of the standard sitting posture parameter model.

[0021] Preferably, the judging module comprises:

[0022] Initial state judgment submodule, time threshold setting submodule, activity state judgment submodule;

[0023] The initial state judgment submodule is used to determine whether the relative position of the human figure and the face in the sitting state meets the offset requirements, and if the posture angle exceeds the offset range, the sitting posture is considered abnormal, or if the relative position of the human figure and the face exceeds the offset range, the sitting posture is considered abnormal; the time threshold setting submodule is used to use a clustering algorithm to determine the type of activity in the current time period, obtain an activity database, identify the current activity type, and set the corresponding time threshold in combination with the historical time corresponding to the current activity in the activity database; the activity judgment submodule is used to determine whether the current activity state has changed and determine whether the sitting posture is abnormal based on the change result.

[0024] Preferably, the activity status determination submodule includes:

[0025] Data acquisition unit, data preprocessing unit, model building unit;

[0026] Used to obtain sensor data and video stream data marked with a timestamp, the data preprocessing unit is used to divide the sensor data and video stream data marked with a timestamp into multiple samples according to a fixed time window to obtain window data, and convert the window data into a feature vector, the model construction unit is used to construct an RNN model, and the RNN model recognizes the feature vector to obtain the activity state of the target object.

[0027] Preferably, the RNN model is a bidirectional structure.

[0028] The present invention discloses the following technical effects:

[0029] The present invention provides an integrated sitting posture monitoring device for protecting the eyesight of teenagers, comprising:

[0030] A first data acquisition module, a second data acquisition module, a real-time sitting posture parameter acquisition module, a model building module, a comparison module and a judgment module;

[0031] The first data acquisition module is used to acquire the video stream data of the target object in real time; the second data acquisition module is used to acquire the sensor data of the target object; the real-time sitting posture parameter acquisition module is used to obtain the real-time sitting posture parameters of the target object according to the sensor data of the video stream data of the target object; the model construction module is used to acquire the historical sitting posture data and body feature parameters of the target object and construct the corresponding standard sitting posture model according to the historical sitting posture data and body feature parameters; the comparison module is used to acquire the standard sitting posture parameters according to the standard sitting posture model and compare them with the real-time sitting posture parameters to obtain the sitting posture state; the judgment module is used to judge whether the current sitting posture is an abnormal sitting posture according to the sitting posture state, and if so, set a time threshold according to the historical activity data of the target object, and judge the change of the activity state of the target object after reaching the time threshold to obtain a judgment result, if the judgment result is a change, adjust the standard sitting posture parameters, if the judgment result is no change, the sitting posture of the target object is abnormal. The present invention can establish sitting posture parameters that are more in line with their physical characteristics and habits for each user by introducing a personalized standard sitting posture model, so that all deviations will not be interpreted as abnormal too absolutely. This method will not only improve the accuracy of monitoring, but also better adapt to the needs of adolescents' physical changes during growth and development; by using the first data acquisition module and the second data acquisition module to simultaneously collect video streams and sensor data, it can comprehensively analyze the sitting habits and dynamic activities of the target object. By integrating multiple data sources, it not only provides a more comprehensive perspective, but also improves the accuracy of the system's sitting posture assessment in complex situations. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0033] Figure 1 An integrated sitting posture monitoring device for protecting teenagers' vision is provided in an embodiment of the present invention.

[0034] Description of reference numerals:

[0035] 1-first data acquisition module, 2-second data acquisition module, 3-real-time sitting posture parameter acquisition module, 4-model building module, 5-comparison module 6-judgment module. DETAILED DESCRIPTION

[0036] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0037] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0038] like Figure 1 As shown, the present invention provides an integrated sitting posture monitoring device for protecting the eyesight of teenagers, comprising:

[0039] A first data acquisition module 1, a second data acquisition module 2, a real-time sitting posture parameter acquisition module 3, a model building module 4, a comparison module 5 and a judgment module 6;

[0040] The first data acquisition module 1 is used to acquire the video stream data of the target object in real time; the second data acquisition module 2 is used to acquire the sensor data of the target object; the real-time sitting posture parameter acquisition module 3 is used to obtain the real-time sitting posture parameters of the target object according to the sensor data of the video stream data of the target object; the model construction module 4 is used to acquire the historical sitting posture data and body feature parameters of the target object and construct a corresponding standard sitting posture model according to the historical sitting posture data and body feature parameters; the comparison module 5 is used to obtain the standard sitting posture parameters according to the standard sitting posture model and compare them with the real-time sitting posture parameters to obtain the sitting posture state; the judgment module 6 is used to judge whether the current sitting posture is an abnormal sitting posture according to the sitting posture state, and if so, set a time threshold according to the historical activity data of the target object, and judge the change of the activity state of the target object after reaching the time threshold to obtain a judgment result, if the judgment result is a change, adjust the standard sitting posture parameters, if the judgment result is no change, the sitting posture of the target object is abnormal.

[0041] Specifically, during class, students sit near the monitoring device and the system monitors their sitting posture in real time. If the system detects a bad sitting posture, it will immediately give a warm reminder to encourage students to adjust their posture and protect their eyesight. In a home learning environment, parents can view their children's sitting posture and activity records through the mobile phone APP, and set appropriate reminders to help their children develop good habits.

[0042] Specifically, a high-definition webcam: Its resolution should be at least 1080p, support real-time video processing, and be able to capture clear sitting details.

[0043] Depth camera: such as Microsoft Kinect, can obtain more additional spatial data (depth information), which is helpful for sitting posture monitoring.

[0044] Sports cameras: such as GoPro, etc., are suitable for different usage scenarios, especially those that require flexible positioning.

[0045] Viewing angle coverage: The camera should be installed in a position that covers the entire sitting area. For example, install it above the table or on the wall to ensure that the sitting situation can be fully captured.

[0046] Height and tilt angle: The camera should be mounted slightly above the target subject’s head height (approximately 1.5-2 meters) and tilted slightly downward (15-30 degrees) to provide the best viewing angle for the user’s sitting posture.

[0047] Fixed bracket: Ensures the camera is fixed in position to avoid inconsistent data due to movement.

[0048] Furthermore, the second data acquisition module 2 includes:

[0049] Pressure sensor submodule, temperature sensor submodule, humidity sensor submodule, accelerometer;

[0050] The pressure sensors are installed on the seat, tabletop, back cushion and footrest;

[0051] The pressure sensor submodule is used to obtain pressure data of the seat, desktop, back cushion and footrest positions, the temperature sensor submodule is used to obtain temperature data in the environment, the humidity sensor submodule is used to obtain humidity data in the environment, and the accelerometer is used to obtain dynamic sitting posture data of the target object.

[0052] Specifically, the installation location of the pressure sensor submodule is:

[0053] Seats: Pressure sensors are embedded in the seat cushions and distributed in the main load-bearing areas of the seat.

[0054] Desktop: Pressure sensors are installed in the center and on both sides of the desktop to monitor the placement of the user's arms.

[0055] Back cushion: Place the sensor on the backrest area of ​​the back cushion to monitor the pressure distribution on the back.

[0056] Pedals: Install pressure sensors on the pedals to monitor the support of the foot.

[0057] The humidity sensor should be installed together with the temperature sensor, usually in the sensor compartment to ensure that it is not affected by direct water sources.

[0058] The accelerometer can be built into a device worn by the target subject, such as a smart bracelet, backpack, etc.; or directly integrated into a seat or desktop device.

[0059] Furthermore, the real-time sitting posture parameter acquisition module 3 includes:

[0060] Marking submodule, video data conversion submodule, filtering submodule, key point detection submodule, posture estimation submodule, data fusion submodule;

[0061] The marking submodule is used to timestamp the sensor data and the video stream data, the video conversion submodule is used to convert the video stream data into a grayscale image, the filtering submodule is used to filter the grayscale image to obtain preprocessed image data, the key point detection submodule is used to extract key points based on the preprocessed image data and a real-time computer vision algorithm to obtain key point position data of the target object, the posture estimation submodule is used to calculate the initial sitting posture parameters of the target image based on the key point position data, and the data fusion submodule is used to integrate the initial coordinate parameters and the sensor data to obtain the current sitting posture parameters.

[0062] Specifically, this module first needs to timestamp the sensor data and video stream data to ensure the time alignment between the two data sources.

[0063] In practical applications, assuming that the video stream is acquired at a frequency of 30 frames per second and the sensor data is acquired 10 times per second, the system timestamps each frame of video and the corresponding sensor data according to the minimum time unit (for example, every 100 milliseconds) to maintain data consistency.

[0064] Calculate the initial sitting posture parameters based on the key point position data, such as the inclination angle of the shoulders, the uprightness of the back, etc.

[0065] Furthermore, the model building module 4 includes:

[0066] Historical data acquisition submodule, body feature acquisition submodule, model generation submodule and feedback submodule;

[0067] The historical data acquisition submodule is used to acquire the historical sitting posture data of the current target object, the body feature acquisition submodule is used to acquire the body feature data of the target object, and the body feature data includes: height, weight, leg length, waist circumference, the model generation submodule is used to use a nonlinear regression model to construct a standard sitting posture parameter model according to the body feature data and the historical sitting posture data, and the feedback submodule is used to acquire the target object's perception evaluation in real time and adjust the parameters of the standard sitting posture parameter model.

[0068] Specifically, this module usually requires users to provide physical characteristic data, including height, weight, leg length and waist circumference, when using it for the first time. It can be obtained through input forms, questionnaires or sensors (such as smart scales).

[0069] Data storage: The collected physical characteristic data is stored in the user's profile (a database can also be used).

[0070] The historical data acquisition submodule, body feature acquisition submodule, model generation submodule and feedback submodule interact and cooperate to realize a dynamically updated standard sitting posture parameter model. This system can not only self-learn and adapt to individual differences, but also provide timely feedback on user experience, improve the sitting posture monitoring effect, and promote the protection of young people's vision and the formation of healthy sitting posture.

[0071] Furthermore, the judging module 6 includes:

[0072] Initial state judgment submodule, time threshold setting submodule, activity state judgment submodule;

[0073] The initial state judgment submodule is used to determine whether the relative position of the human figure and the face in the sitting state meets the offset requirements, and if the posture angle exceeds the offset range, the sitting posture is considered abnormal, or if the relative position of the human figure and the face exceeds the offset range, the sitting posture is considered abnormal; the time threshold setting submodule is used to use a clustering algorithm to determine the type of activity in the current time period, obtain an activity database, identify the current activity type, and set the corresponding time threshold in combination with the historical time corresponding to the current activity in the activity database; the activity judgment submodule is used to determine whether the current activity state has changed and determine whether the sitting posture is abnormal based on the change result.

[0074] Specifically, a set of offset ranges are set, and the corresponding offset values ​​(such as the distance between the head and the center point of the body, the distance between the shoulder width and the edge of the cushion, etc.) are calculated according to the standard sitting posture model; if the relative position meets the requirements but the posture angle (such as the body tilt angle) exceeds the allowable range, the sitting posture is marked as abnormal.

[0075] If the relative position of the human figure and the face exceeds the set offset range, it is also marked as an abnormality.

[0076] Specifically, historical activity data collection: extract activity data for the last month from the user's sensor data log, including the duration of activities such as sitting, standing, and walking.

[0077] Activity type tag: Classify each activity type and record it in the activity database, for example:

[0078] Sitting and studying: 30 minutes per session on average

[0079] Rest (standing or walking): 5 minutes each time on average;

[0080] Historical data analysis: Determine the appropriate time threshold based on the clustering results. For an activity like "sitting and studying", if the center point of the clustering is 30 minutes, the time threshold can be set as:

[0081] Maximum continuous sitting time threshold: Currently set to 30 minutes. If this time is exceeded, the user needs to be reminded to change posture.

[0082] Furthermore, the activity status determination submodule includes:

[0083] Data acquisition unit, data preprocessing unit, model building unit;

[0084] Used to obtain sensor data and video stream data marked with a timestamp, the data preprocessing unit is used to divide the sensor data and video stream data marked with a timestamp into multiple samples according to a fixed time window to obtain window data, and convert the window data into a feature vector, the model construction unit is used to construct an RNN model, and the RNN model recognizes the feature vector to obtain the activity state of the target object.

[0085] Specifically, set a reasonable time window, such as splitting the data every 5 seconds, to ensure that dynamic information about changes in user activities can be captured.

[0086] Data sample segmentation: The program records the data timestamp and segments the sensor data and video data at fixed intervals to form samples. Each sample contains the sensor data and video data within the last 5 seconds.

[0087] Extract key information from each sample as a feature vector, for example:

[0088] Sensor characteristics: activity duration (cumulative duration, maximum pressure, average acceleration, etc.).

[0089] Video data features: key point detection results (such as head position, shoulder position, etc.) and their relative position distance.

[0090] Specifically, the RNN model has a bidirectional structure.

[0091] Using the RNN model to process time series data can effectively capture the dynamic changes of activity states and has higher response speed and accuracy compared to traditional static analysis.

[0092] Bidirectional information flow: Bidirectional RNN considers both forward and backward information of the input sequence. The forward RNN processes past information, and the backward RNN processes future information. This bidirectional feature enables the model to utilize two kinds of contextual information, thereby providing a more comprehensive background understanding.

[0093] Better feature extraction: When dealing with dynamic processes such as changes in sitting posture, the current state often depends on the past state and the upcoming state. Bidirectional RNN can better capture this dynamic relationship, thereby improving the expressiveness of features.

[0094] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0095] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. An integrated sitting posture monitoring device for protecting the eyesight of teenagers, characterized in that: include: A first data acquisition module, a second data acquisition module, a real-time sitting posture parameter acquisition module, a model building module, a comparison module and a judgment module; The first data acquisition module is used to acquire video stream data of the target object in real time; the second data acquisition module is used to acquire sensor data of the target object; The real-time sitting posture parameter acquisition module is used to obtain the real-time sitting posture parameters of the target object according to the sensor data of the video stream data of the target object; the model construction module is used to obtain the historical sitting posture data and body feature parameters of the target object and construct a corresponding standard sitting posture model according to the historical sitting posture data and body feature parameters; The comparison module is used to obtain standard sitting posture parameters according to the standard sitting posture model and compare them with the real-time sitting posture parameters to obtain the sitting posture state; the judgment module is used to judge whether the current sitting posture is an abnormal sitting posture according to the sitting posture state. If so, a time threshold is set according to the historical activity data of the target object, and after the time threshold is reached, the activity state change judgment of the target object is performed to obtain a judgment result. If the judgment result is a change, the standard sitting posture parameters are adjusted. If the judgment result is no change, the sitting posture of the target object is abnormal.

2. The integrated sitting posture monitoring device for teenagers' vision protection according to claim 1 is characterized in that: The second data acquisition module includes: Pressure sensor submodule, temperature sensor submodule, humidity sensor submodule, accelerometer; The pressure sensors are installed on the seat, tabletop, back cushion and footrest; The pressure sensor submodule is used to obtain pressure data of the seat, desktop, back cushion and footrest positions, the temperature sensor submodule is used to obtain temperature data in the environment, the humidity sensor submodule is used to obtain humidity data in the environment, and the accelerometer is used to obtain dynamic sitting posture data of the target object.

3. The integrated sitting posture monitoring device for teenagers' vision protection according to claim 1 is characterized in that: The real-time sitting posture parameter acquisition module includes: Marking submodule, video data conversion submodule, filtering submodule, key point detection submodule, posture estimation submodule, data fusion submodule; The marking submodule is used to timestamp the sensor data and the video stream data, the video conversion submodule is used to convert the video stream data into a grayscale image, the filtering submodule is used to filter the grayscale image to obtain preprocessed image data, the key point detection submodule is used to extract key points based on the preprocessed image data and a real-time computer vision algorithm to obtain key point position data of the target object, the posture estimation submodule is used to calculate the initial sitting posture parameters of the target image based on the key point position data, and the data fusion submodule is used to integrate the initial coordinate parameters and the sensor data to obtain the current sitting posture parameters.

4. The integrated sitting posture monitoring device for protecting the eyesight of teenagers according to claim 3 is characterized in that: The model building module includes: Historical data acquisition submodule, body feature acquisition submodule, model generation submodule and feedback submodule; The historical data acquisition submodule is used to acquire the historical sitting posture data of the current target object, the body feature acquisition submodule is used to acquire the body feature data of the target object, and the body feature data includes: height, weight, leg length, waist circumference, the model generation submodule is used to use a nonlinear regression model to construct a standard sitting posture parameter model according to the body feature data and the historical sitting posture data, and the feedback submodule is used to acquire the target object's perception evaluation in real time and adjust the parameters of the standard sitting posture parameter model.

5. The integrated sitting posture monitoring device for teenagers' vision protection according to claim 3 is characterized in that: The judging module comprises: Initial state judgment submodule, time threshold setting submodule, activity state judgment submodule; The initial state judgment submodule is used to determine whether the relative position of the human figure and the face in the sitting state meets the offset requirements, and if the posture angle exceeds the offset range, the sitting posture is considered abnormal, or if the relative position of the human figure and the face exceeds the offset range, the sitting posture is considered abnormal; the time threshold setting submodule is used to use a clustering algorithm to determine the type of activity in the current time period, obtain an activity database, identify the current activity type, and set the corresponding time threshold in combination with the historical time corresponding to the current activity in the activity database; the activity judgment submodule is used to determine whether the current activity state has changed and determine whether the sitting posture is abnormal based on the change result.

6. The integrated sitting posture monitoring device for teenagers' vision protection according to claim 5, characterized in that: The activity status determination submodule includes: Data acquisition unit, data preprocessing unit, model building unit; Used to obtain sensor data and video stream data marked with a timestamp, the data preprocessing unit is used to divide the sensor data and video stream data marked with a timestamp into multiple samples according to a fixed time window to obtain window data, and convert the window data into a feature vector, the model construction unit is used to construct an RNN model, and the RNN model recognizes the feature vector to obtain the activity state of the target object.

7. The integrated sitting posture monitoring device for protecting the eyesight of teenagers according to claim 6, characterized in that: The RNN model has a bidirectional structure.

Images