CN115393383A

CN115393383A - Myopia protection intelligent glasses based on scene recognition judges with eye health degree

Info

Publication number: CN115393383A
Application number: CN202110561171.1A
Authority: CN
Inventors: 杨旭东; 高俊超; 李航
Original assignee: Shanghai Chuanxin Information Technology Co ltd
Current assignee: Shanghai Chuanxin Information Technology Co ltd
Priority date: 2021-05-22
Filing date: 2021-05-22
Publication date: 2022-11-25

Abstract

The invention discloses a method for judging eye health degree based on scene recognition, which is applied to intelligent myopia prevention glasses; the glasses frame structurally comprises a glasses frame, a sensor component, a microprocessor, a memory, a protection component, a battery and a corresponding circuit; the sensor component comprises a wearing state sensor, an acceleration sensor, a displacement sensor, a distance sensor and an ambient light sensor; the built-in intelligent algorithm software is solidified in the memory; the method is realized by dividing the eye using scene according to different influences on the eye using health; the method comprises the steps of identifying the current eye using scene of a user through an algorithm by collecting and processing sensor data built in the glasses, so that the identifiable scene covers all eye using time of the user wearing the glasses; calculating the eye health degree of the user according to the scene recognition result through an intelligent algorithm of software; meanwhile, the calculation result of the eye health degree is intelligently matched through software and is used as the input of the protection component, a corresponding protection mechanism is triggered in a targeted manner, a user is required to take corrective measures, the user is supervised to use eyes reasonably, and the effect of protecting the eye health is achieved.

Description

Myopia protection intelligent glasses based on scene recognition judges with eye health degree

Technical Field

The invention relates to application of artificial intelligence of sensor pattern recognition to wearable products, in particular to application of a method for recognizing eye scene and judging eye health degree to myopia protection intelligent glasses.

Background

According to the data of Wei Jian of China, the myopia rate of teenagers in our country is as high as 53.6% in 2018, and the myopia rate of college students reaches 90% which is surprising. Such high rates of myopia not only affect the quality of an individual's future life, but potentially even become a limiting factor in national development. The reason behind this is not always due to the neglect of eye health, and more importantly, the lack of suitable means to protect eye health. For the definition of eye health, most people stay in the influence of distance. The real reason for this is that for children, the comprehensive assessment of key factors such as the development level of the eyeball, the fatigue degree of ciliary muscle, the illumination intensity and quality, and the amount of exercise should be included.

According to the investigation of the eyesight protection equipment in the current market, the mainstream products can not be applied to various scenes, some products only depend on the control of the eye using distance during writing operation, and the influence of other links and conditions on the eye using health is ignored, for example, electronic equipment, particularly a mobile phone is used; because the screen is smaller, the font and the image are also smaller, and better visual experience can drive the user to be closer to the screen so as to see the details clearly; the situation can lead the ciliary muscle to be in a tense state, and the damage to the vision is more than the influence of writing operation with the same duration on the vision; similarly, the cumulative effects of different eye scenes on myopia will vary, such as the need to overlook the eyes over time to relax the eye, even if the eye distance is kept at the correct distance; therefore, if the eye use scene cannot be fully considered and the evaluation of the eye health degree is lacked, the eyesight health cannot be effectively protected.

A myopia prevention glasses identifies the eye use scene of a user, comprehensively considers eye use health factors, dynamically evaluates the eye use health degree of the user in real time, can be differentially applied to a protection component, and reminds the user until the user is forcibly required to correct the eye use mode; is suitable for eye health protection in all environmental conditions and is especially suitable for preventing myopia of children with self behavior control capacity needing to be improved.

Disclosure of Invention

Aiming at the limitations of the existing myopia prevention method and equipment, the invention provides a method for realizing scene recognition based on sensor data and evaluating eye health degree through an algorithm, which is applied to intelligent glasses for preventing myopia.

The hardware at least comprises a spectacle frame, a wearing state sensor, an acceleration and displacement sensor, a distance sensor, an ambient light sensor, a microprocessor, a memory, a protection component, a battery and a corresponding management circuit; all of the hardware is built into the eye frame.

The wearing state sensor positioned at the nose support is used for judging whether the glasses are worn on the nose bridge or not; the acceleration and displacement sensor is used for collecting head and eye posture data, and acceleration and displacement vectors and corresponding scalar values in three axial lines and angles can be obtained through an algorithm; the distance sensor is arranged in the front of the glasses and faces the front and is used for collecting the distance from the front half to the object; an ambient light sensor disposed in front of the glasses and facing forward for collecting incident light intensity and color data; the software stored and operated on the microprocessor comprises an intelligent algorithm module, a data input module, a control output module and a data storage and transmission module.

Scene division and recognition based on different definitions of myopia influence degrees are realized through an intelligent algorithm, for example, the following classification method is adopted: the ciliary muscle is in a highly tense state, and the behavior is kept for a short time, which can easily cause myopia, and is defined as a category one; the ciliary muscle is in a moderately tense state, and the behavior is maintained for a medium and short time, which can cause or deepen myopia, and is defined as a category two; the ciliary muscle is in a low-grade tense state, and the behavior can be kept for a long time to cause myopia, which is defined as a third category; ciliary muscle is in a relaxed state, and maintaining this behavior for any length of time does not lead to myopia, which is defined as category four; this action can increase ocular blood circulation, relax ciliary muscles, and relieve asthenopia, and is defined as category five.

Preferably, based on the result of the eye use scene recognition and the eye use health degree of the user, the intelligent algorithm can correspondingly give a control instruction to control the behavior of the protection component, remind or force the user to make a corresponding corrective measure, and avoid excessive eye use and unhealthy eye use.

Preferably, when the user wears the glasses to do sports, the glasses can identify the intensity of the current sports of the user besides the sports scene; if walking or running, the function of recording steps can be realized.

Preferably, in order to accurately identify whether the glasses are worn on the nose bridge, the capacitive sensor or the pressure sensing sensor is designed to be placed at the nose support, so that the data collected by the equipment reflect the real eye health degree, and possible evasive behaviors of a user can be recorded.

Preferably, the distance sensor facing the front of the wearer is an optical distance sensor based on FoT, and is capable of detecting a variable-frequency and high-precision object distance in a one-time multipoint matrix detection manner.

Drawings

Fig. 1 is a schematic diagram of a system architecture design of novel intelligent protection control glasses.

Fig. 2 details a logical implementation of scene recognition and eye health recognition.

Detailed Description

Embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

The wearing state sensor is arranged in the nose pads of the glasses, and when the nose pads of the glasses are in contact with the nose bridge, the nose pads of the glasses are judged to be in place, namely the glasses are in a wearing state; all eye scene judgment needs to enter a subsequent judgment state based on the fact that the glasses are in a wearing state, otherwise, the glasses enter a standby mode, and when the glasses are worn or removed, the system records corresponding time in a memory.

The acceleration and displacement sensors are arranged in the glasses, and have the characteristics of small volume, high sensitivity, low power consumption, fast response, high reliability, flexible data output mode and the like, are used for acquiring the accelerations corresponding to the linear velocity and the angular velocity of the whole glasses in three axial directions respectively, and can acquire the characteristics of the speed, the acceleration, the displacement and the angle corresponding to the current head through a built-in algorithm; when the robot is in a motion state, the intensity of motion, the corresponding time and the corresponding duration can be further judged according to the discrete level of each direction of acceleration; in walking or running exercise mode, the built-in algorithm can identify the steps and record the number of steps in the corresponding time period.

The distance sensor is arranged in the spectacle frame, the detection direction faces to the visual direction of a wearer, the distance data of an object in a set solid angle range of the visual direction are obtained through a window on the frame, the range can be divided into a plurality of small area matrixes, the sensor can respectively obtain the distance data of each corresponding area, and can also obtain the maximum value, the minimum value or the average value of the distances of a plurality of areas according to the setting; when the acceleration and the displacement obtained by the acceleration and displacement sensor accord with the static condition criterion, the high-precision matrix type optical distance sensor scans the front object; fitting the distance matrix data at different rotation angles or pitch angles to obtain a spatial three-dimensional model taking a user as a center; by utilizing the change of the relative position and angle of the glasses and the three-dimensional model, the distance from the eyes to the object to be watched, the inclination angle of the two eyes, the inclination angle of the sight line and the object and the shape of the object to be watched can be calculated; by the same method, the three-dimensional distribution of the light intensity based on the space angle can be reconstructed according to the reading of the light intensity sensor, and the fact that the object in front is the screen of the electronic product or the desk and the book can be identified by using the geometric size characteristics of the objects such as the book, the desk, the electronic product and the desk lamp and the change rule of the brightness and the illumination intensity, particularly the change of the gradient.

Defining scenes for distinguishing the influence degree of the scenes on myopia, such as reading and writing under the condition of body rest, using electronic equipment under the condition of body rest, watching television under the condition of body rest, moving and taking vehicles; based on the difference in the degree of myopia impact, the following classification methods can be defined: the ciliary muscle is in a highly tense state, and the behavior is kept for a short time, which can easily cause myopia, and is defined as a category one; the ciliary muscle is in a medium tension state, and the behavior is continuously maintained for a medium and short time, which can cause or deepen myopia and is defined as a category two; the ciliary muscle is in a low-level tension state, and the behavior can be kept for a long time to cause myopia, which is defined as a class three; ciliary muscle is in a relaxed state, and maintaining this behavior for any length of time does not result in myopia, which is defined as category four; through this behavior, it is possible to increase ocular blood circulation or relax ciliary muscles and relieve asthenopia, and it is defined as category five.

The above definition of the category can be judged by the data of each built-in sensor; the logical relationship is shown in fig. 2: firstly, acquiring the in-place state of a wearing state sensor, if the wearing state sensor is not in place, indicating that the glasses are not placed on the bridge of the nose, recording an event and time thereof, and turning off other sensors; if the acceleration and displacement sensor is in place, starting the acceleration and displacement sensor and reading data, and judging whether the acceleration and displacement sensor is in a motion state by a statistical analysis method; if the state is in a motion state and the half-height width of statistical distribution is large, classifying the current state into a scene category five, and recording the number of motion steps; if the half-height width of the statistical distribution is smaller, classifying the current state into a scene class four; if the object is not in a motion state, starting a distance sensor and a light sensor, reading data, integrating the data of an acceleration sensor and the data of a displacement sensor, establishing a three-dimensional model for the object in the space around the front hemisphere, and finally calculating the current real eye distance for visual observation; when the eye using distance is larger than a preset threshold value, classifying the current state into a scene class III; when the eye distance is smaller than a preset threshold value, reading data of a light intensity sensor, and calculating the light intensity at a specific position and the change condition of RGB components of colors so as to judge whether the screen is a screen of an electronic product; if not, classifying the current state into a scene type two; if so, the current state is classified as scene type one.

In the algorithm, quantized near vision influence degree values are preset for different scenes, and the eye health degree of a user can be calculated by respectively utilizing methods of differentiation, integration and difference values according to the retention time of the corresponding scene; the real-time evaluation result of the eye health degree of the user is combined with the quantitative evaluation of the influence of the real-time scene on the eyesight, and the result is used for generating a control instruction and outputting the control instruction to the protection component; the protection component can activate corresponding reminding functions according to the instruction, wherein the reminding functions comprise light reminding, intermittent vision obstruction, complete vision obstruction and the like, and remind and force a user to take corresponding correction measures so as to avoid overuse of eyes and damage of vision health.

Particularly, in a motion scene, a built-in algorithm can identify steps based on data of an acceleration sensor and a displacement sensor, and further quantize the contribution of motion to the health degree according to the frequency and the number of the steps, so as to change the evaluation result of the health degree; the start and end times of the movement, and the number of steps in the corresponding time period, are also recorded in memory.

Hereinbefore, specific embodiments of the present invention are described with reference to the drawings. However, those skilled in the art will appreciate that various modifications and substitutions can be made to the specific embodiments of the present invention without departing from the spirit and scope of the invention. Such modifications and substitutions are intended to be within the scope of the present invention as defined by the appended claims.

Claims

1. A method for judging eye health based on scene recognition is applied to myopia protection intelligent glasses; the intelligent glasses for preventing myopia at least comprise an acceleration sensor, a displacement sensor, a distance sensor, an ambient light sensor, a microprocessor, a memory, a protection component, a wearing state sensor, a glasses frame and a battery; the method is characterized in that: whether the equipment is worn or not is obtained through the wearing state sensor, the head and eye postures of a user are obtained through the acceleration sensor and the displacement sensor, the ambient light intensity is sensed through the ambient light sensor, and the distance from an object to the glasses in the visual field range is detected through the distance sensor; the microprocessor acquires the sensor data in real time and identifies the current eye scene of the user through an algorithm; the definition of the scene is divided according to the influence degree of the behavior on the myopia, and the recognition is realized through an algorithm; the scene recognition result and the corresponding occurrence time and duration are recorded in the memory at the same time, the eye health degree of the user is evaluated through an algorithm, and a corresponding control instruction is output and applied to the protection component to remind or require the user to correct the eye using mode, so that the aim of effectively protecting the eye health of the user is fulfilled.

2. The method according to claim 1, characterized in that all the above-mentioned functional components are integrated in the eyeglasses frame, wherein the acceleration and displacement sensors are constituted by a triaxial fiber optic gyroscope and a triaxial accelerometer; the wearing state sensor is positioned at the nose pad, is a capacitance or pressure sensor and acquires the wearing state by judging whether the wearing state sensor is close to or in contact with the skin; the distance sensor is based on FoT matrix type optical distance sensor, which is located in front of the glasses, the detection direction is consistent with the sight line direction of the user, and the distance reading is obtained through a window on the frame of the eyes.

3. The method of claim 1, wherein the behaviors are classified according to the degree of myopia development or progression by the scene and are respectively given a weight coefficient; the more easily asthenopia is generated to cause or deepen myopia, the weight coefficient is at one extreme; the easier it is to eliminate visual fatigue, the weighting factor is at the other extreme.

4. The method according to claim 3, characterized in that in the scene of motion, the time and intensity of the motion can be judged; such as running, the number of steps may be recorded.

5. The method as claimed in claims 1 and 3, wherein the current eye health value is calculated by an algorithm for a predetermined period of time using weighting coefficients respectively defined for the various behavior classes.