It is a kind of to use eye health condition monitoring method and system
Technical field
The invention belongs to use eye health monitoring field, and in particular to a kind of to use eye health condition monitoring method and system.
Background technique
In recent years, the rate of myopia of student is higher and higher, this study habit with student usually, posture, with eye distance from there is pole
Important Relations.Wherein, eye posture of student etc. with eye health condition for eyes health whether have significant effect, for improve
This case finds having very much with eye abnormal conditions and timely being corrected in eye posture of the early stage to student for student in time
Necessity has many eye protectors, such as eye-protecting desk lamp on the market at present, all brag about have detection eye distance from requirement, it is most of
Apply following technology:
1. range sensor is arranged in portion in the device, when user is when using eye protector, when sensor detects
It when having object desired distance less than pre-determined distance, can be fed back to user, but this method only measures object and dress
The distance set, thus can not accurate measurements user with eye distance from.For example user is reading a book, and range sensor detects
Books are too close to report by mistake with desk lamp.
2. portion is arranged external gravity sensor and needs gravity sensor being worn on use when in use in the device
With family, it is able to detect that user leans forward or askew seat passes through vibration or nothing when gravity sensor detects abnormal state
Line technology is fed back to user, and this technology can detect whether user sits straight, and is if being worn on to detect on head
No torticollis.But it cannot be detected with eye distance from for example user reading, and it is to be unable to monitor that books are close from eyes
's.And it wears gravity sensor and charging to need all make troubles to the use of user.
3. camera is arranged in portion in the device, when user is when using desk lamp, camera will do it recognition of face, and
Then the eyes for finding user calculate at a distance from desktop, when distance is less than setting value, carry out prompting feedback.But it is this kind of
Technology can only detect eyes to the distance in predetermined position at present, and identification can not be carried out to desktop object or is carried out to the behavior of user
Judgement, so that reminding feedback mechanism very inflexible, rate of false alarm is high, is unable to reach the good effect for being monitored and being corrected with eye
Fruit.
From the foregoing, it will be observed that the prior art is when monitored with eye health condition, generally existing monitoring mechanism is inflexible, rate of false alarm
Height, the low problem of degree of intelligence.
Summary of the invention
In order to overcome the above technical defects, the present invention provides a kind of eye health condition monitoring method and system, breaches
The limitation for carrying out the high rate of false alarm monitored with eye using single sensor in the past, substantially increases the accuracy monitored with eye,
It can achieve good eye monitoring effect.
To solve the above-mentioned problems, the present invention is achieved by following technical scheme:
It is a kind of to use eye health condition monitoring method, comprising the following steps:
Obtain human body head image and with eye scene image;
The characteristic point with eye scene image is extracted, scene information is calculated;The scene information includes that desktop is flat
Face, object profile and object and desktop angle;The object is paper reading object or electronic reading device;
The characteristic point for extracting the human body head image, is calculated face information;The face information includes face device
Official's coordinate and facial orientation angle;
Using preset scene judgment rule, according to the scene information and face information, judge that user's uses eye scene;
Eye judgment rule is used using preset, according to the scene information, face information and with eye scene, calculates user's
With eye distance from or backlight reading conditions, judge user use eye health condition.
Compared with the existing technology, this method has the beneficial effect that
It, can be from the human body head image and use got by this method being monitored with eye health condition to user
Required scene information and face information are obtained in eye scene image in real time, and according to preset scene judgment rule, it is right in advance
User's judged with eye scene, further according to it is different with eye scene combine with eye judgment rule to user with eye distance from or
Backlight reading conditions are judged, by these steps, it is special that this method can get comprehensive image in eye-using environment in real time
Reference breath, and user monitor comprehensively and timely with eye health condition with eye scene for user is different, it breaks through
The limitation of the previous high rate of false alarm for monitored with eye using single sensor, substantially increase monitored with eye it is accurate
Degree can achieve good eye monitoring effect.
Further, the scene information is calculated by following steps:
Extract the desktop features point and object characteristic point in eye scene image;
Desktop plane is fitted according to the desktop features point;
It is to identify object profile according to the object characteristic point;
The angle of object and desktop is calculated according to the desktop plane and the object profile.
Through the above steps, the object on desktop plane and table can be known according to each characteristic point in realtime graphic
, and do not calculate the angle of object and desktop, convenient for it is subsequent according to these information to user carried out with eye health condition it is more smart
True judgement and monitoring.
Further, the face information is calculated by following steps:
Obtain the characteristic point in the human body head image;
The coordinate of each organ of face is determined according to the characteristic point;
Using fitting algorithm, according to the coordinate of each organ, obtain face in three dimensions towards angle.
Through the above steps, the coordinate of each organ in face can be obtained according to each characteristic point of realtime graphic
It takes and being fitted towards angle information to face, user is carried out with eye health condition according to these information convenient for subsequent
More accurately judgement and monitoring.
Further, the scene judgment rule are as follows:
According to the pupil coordinate in the facial orientation information and human face coordinate, the direction of visual lines of user is calculated;
Judge whether to recognize the object profile:
If recognizing the object profile, judge whether the object profile is on the direction of visual lines, if so, judgement
User is to read scene with eye scene;
If not recognizing the object profile, judge whether the desktop plane is on the direction of visual lines, if so,
Judge user is to write scene with eye scene.
By above-mentioned scene judgment rule, eye scene can be used in real time with accurate judgement user, facilitated subsequent according to eye
Scene is monitored the different eye posture information of user.
It is further, described to use eye judgment rule are as follows:
When it is described with eye scene be write scene when:
According to the eyes coordinates and the desktop plane, calculate eyes to desktop Writing distance;When the writing
When distance is less than preset Writing distance threshold value, judge user is with eye exception with eye health condition;
And/or the tilt angle on head is calculated using the nose of face and the coordinate of eyes, and calculate head inclination
The angle of the Z axis of angle and predetermined camera coordinate system;When the angle is less than preset slanted angle threshold value, user is judged
With eye health condition be with eye exception;
When it is described with eye scene be read scene when:
According to the eyes coordinates and the object profile calculate eyes to object reading distance, when it is described read away from
When from being less than preset reading distance threshold value, judge user is with eye exception with eye health condition;
And/or the angle according to the object and desktop calculates backlight folder in conjunction with the angle of preset light and desktop
Angle;The backlight angle is that the light is that initial edge rotates the angle generated to the object;When the backlight angle is less than in advance
If backlight angle threshold value when, judge user is that backlight eye is abnormal with eye health condition.
Use eye judgment rule by above-mentioned, according to it is different with eye scene to user with eye distance from or backlight reading conditions
Judged, user monitor comprehensively and timely with eye health condition with eye scene for user is different.
The present invention is also corresponding to disclose a kind of eye health condition monitoring system, comprising:
Image acquiring device, for obtaining human body head image and with eye scene image;
Image processing apparatus is calculated for extracting the human body head image and the characteristic point with eye scene image
Obtain scene information and face information;The scene information includes the angle of desktop plane, object profile and object and desktop;Institute
Stating object is paper reading object or electronic reading device;The face information includes human face coordinate and facial orientation angle;
Scene judgment means, for being sentenced using preset scene judgment rule according to the scene information and face information
Disconnected user's uses eye scene;
With eye judgment means, it is used to use eye judgment rule using preset, according to the scene information, face information and use
Eye scene, calculate user with eye distance from or backlight reading conditions, judge user use eye health condition.
Compared with the existing technology, this system has the beneficial effect that
It, can be from the human body head image and use got by this system being monitored with eye health condition to user
Required scene information and face information are obtained in eye scene image in real time, and according to preset scene judgment rule, it is right in advance
User's judged with eye scene, further according to it is different with eye scene combine with eye judgment rule to user with eye distance from or
Backlight reading conditions are judged, by these steps, it is special that this method can get comprehensive image in eye-using environment in real time
Reference breath, and user monitor comprehensively and timely with eye health condition with eye scene for user is different, it breaks through
The limitation of the previous high rate of false alarm for monitored with eye using single sensor, substantially increase monitored with eye it is accurate
Degree can achieve good eye monitoring effect.
Further, described image processing unit calculates the scene information by following steps:
Extract the desktop features point and object characteristic point in eye scene image;
Desktop plane is fitted according to the desktop features point;
It is to identify object profile according to the object characteristic point;
The angle of object and desktop is calculated according to the desktop plane and the object profile.
Through the above steps, image processing apparatus can be according to each characteristic point in realtime graphic on desktop plane and table
Object identified, and calculate the angle of object and desktop, convenient for it is subsequent according to these information to user's eye health
Situation carries out more accurately judgement and monitoring.
Further, described image processing unit calculates the face information by following steps:
Obtain the characteristic point in the human body head image;
The coordinate of each organ of face is determined according to the characteristic point;
Using fitting algorithm, according to the coordinate of each organ, obtain face in three dimensions towards angle.
Through the above steps, image processing apparatus can be according to each characteristic point of realtime graphic to each organ in face
Coordinate carry out obtain and being fitted towards angle information to face, eye is used to user according to these information convenient for subsequent
Health condition carries out more accurately judgement and monitoring.
Further, the scene judgment means judge the eye scene by following steps:
According to the pupil coordinate in the facial orientation information and human face coordinate, the direction of visual lines of user is calculated;
Judge whether to recognize the object profile:
If recognizing the object profile, judge whether the object profile is on the direction of visual lines, if so, judgement
User is to read scene with eye scene;
If not recognizing the object profile, judge whether the desktop plane is on the direction of visual lines, if so,
Judge user is to write scene with eye scene.
By above-mentioned scene judgment means, eye scene can be used in real time with accurate judgement user, facilitated subsequent according to eye
Scene is monitored the different eye posture information of user.
Further, described to judge the eye health condition by following steps with eye judgment means:
When it is described with eye scene be write scene when, according to the eyes coordinates and the desktop plane, calculate eyes
To the Writing distance of desktop;When the Writing distance is less than preset Writing distance threshold value, judge that user's uses eye health feelings
Condition is with eye exception;
When it is described with eye scene be read scene when:
According to the eyes coordinates and the object profile calculate eyes to object reading distance, when it is described read away from
When from being less than preset reading distance threshold value, judge user is with eye exception with eye health condition;
And/or the angle according to the object and desktop calculates backlight folder in conjunction with the angle of preset light and desktop
Angle;The backlight angle is that the light is that initial edge rotates the angle generated to the object;When the backlight angle is less than in advance
If backlight angle threshold value when, judge user is that backlight eye is abnormal with eye health condition.
Use eye judgment means by above-mentioned, according to it is different with eye scene to user with eye distance from or backlight reading conditions
Judged, user monitor comprehensively and timely with eye health condition with eye scene for user is different.
Detailed description of the invention
Fig. 1 is the step schematic diagram that eye health condition monitoring method is used described in the embodiment of the present invention 1;
Fig. 2 is an irradiation schematic diagram of a scenario of light described in the embodiment of the present invention 1 and object;
Fig. 3 is another irradiation schematic diagram of a scenario of light described in the embodiment of the present invention 1 and object;
Fig. 4 is described in the embodiment of the present invention 1 with the module diagram of eye health condition monitoring system;
Fig. 5 is described in the embodiment of the present invention 1 with the structural schematic diagram of eye health condition monitoring desk lamp.
Specific embodiment
Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings, it should be understood that preferred reality described herein
Apply example only for the purpose of illustrating and explaining the present invention and is not intended to limit the present invention.
Embodiment 1
As shown in Figure 1, the present embodiment 1 discloses a kind of eye health condition monitoring method, comprising the following steps:
S1, human body head image is obtained and with eye scene image;
Specifically, can carry out three-dimensional imaging by multiple cameras forms 3-D image;Specifically, being wrapped in 3-D image
The image of human body head and the image of desktop object are included;Specifically, before camera may be provided at such as desk lamp lighting apparatus
Side obtains the real-time imaging of user;
S2, the characteristic point for using eye scene image is extracted, scene information is calculated;Scene information includes desktop plane, object
The angle of part profile and object and desktop;Object is paper reading object or electronic reading device;Specifically, object can be books
Or the electronic display unit that can be read such as tablet computer or mobile phone;Specifically, step S2 includes:
S21, it extracts with the desktop features point and object characteristic point in eye scene image;Specifically, characteristic point can be passed through
Detection algorithm is to the characteristic point and object of the desktop in 3-D image, such as books or electronic equipment, Edge Feature Points mentioned
It takes;Specifically, extracting the desktop features point of desktop in image using Corner Detection Algorithm;Specifically, using Corner Detection Algorithm
It is combined with edge detection algorithm, extracts the candidate feature point of object in image;
S22, desktop plane is fitted according to desktop features point;Specifically, using plane fitting algorithm to desktop features point
It is fitted and is eliminated with noise, obtain desktop plane equation;
It S23, according to object characteristic point is to identify object profile;Specifically, using object identification classifier, i.e., it is small-sized
Neural network classifier identifies determine whether object is books or electronic equipment to the candidate feature point of object;
S24, the angle that object and desktop are calculated according to desktop plane and object profile.
Through the above steps, the object on desktop plane and table can be known according to each characteristic point in realtime graphic
, and do not calculate the angle of object and desktop, convenient for it is subsequent according to these information to user carried out with eye health condition it is more smart
True judgement and monitoring.
S3, the characteristic point for extracting human body head image, are calculated face information;Face information includes human face coordinate
With facial orientation angle;Specifically, step S3 includes:
Characteristic point in S31, acquisition human body head image;Specifically, using neural network algorithm in 3-D image
Multiple characteristic points of human body head and face extract;Specifically, extracting LAB the and SURF feature in image, later use
Multiple Edge Feature Points of MLP (multilayer perceptron) neural network positioning head and face;
S32, the coordinate that each organ of face is determined according to characteristic point;Specifically, determining includes eyes, eyebrow, nose, lip
The three-dimensional coordinate of the characteristic point of the organs such as portion;Specifically, directly positioning eyes if in human body head image including eye image
Coordinate, if eye image is blocked or because angle problem can not directly take, the organ that is effectively recognized by other
Characteristic point extrapolates the position coordinates of eyes in conjunction with human body standard facial model, achievees the effect that positioning.
S33, using fitting algorithm, according to the coordinate of each organ, obtain face in three dimensions towards angle;Tool
Body, using fitting algorithm filter error, face is obtained in the main direction of three-dimensional space.
Through the above steps, the coordinate of each organ in face can be obtained according to each characteristic point of realtime graphic
It takes and being fitted towards angle information to face, user is carried out with eye health condition according to these information convenient for subsequent
More accurately judgement and monitoring.
S4, the use eye scene of user is judged according to scene information and face information using preset scene judgment rule;
Specifically, scene judgment rule are as follows:
According to the pupil coordinate in facial orientation information and human face coordinate, the direction of visual lines of user is calculated;Specifically,
Coordinate modification is carried out to facial orientation angle using pupil coordinate, obtains direction of visual lines;
Judge whether to recognize object profile:
If recognizing object profile, judge whether object profile is on direction of visual lines, if so, judging that user's uses eye field
Scape is to read scene;
If not recognizing object profile, judge whether desktop plane is on direction of visual lines, if so, judging the use of user
Eye scene is to write scene.
By above-mentioned scene judgment rule, eye scene can be used in real time with accurate judgement user, facilitated subsequent according to eye
Scene is monitored the different eye posture information of user.
S5, eye judgment rule is used using preset, according to scene information, face information and with eye scene, calculates user's
With eye distance from or backlight reading conditions, judge user use eye health condition, specifically, with eye judgment rule are as follows:
When being to write scene with eye scene:
According to eyes coordinates and desktop plane, calculate eyes to desktop Writing distance;It is preset when Writing distance is less than
Writing distance threshold value when, judge user with eye health condition be with eye exception;
And/or the tilt angle on head is calculated using the nose of face and the coordinate of eyes, and calculate head inclination
The angle of the Z axis of angle and predetermined camera coordinate system;When angle is less than preset slanted angle threshold value, i.e., user is reading
When head inclination, judge user with eye health condition be with eye exception;
When being to read scene with eye scene:
According to eyes coordinates and object profile calculate eyes to object reading distance, when reading distance be less than it is preset
When reading distance threshold value, judge user is with eye exception with eye health condition;
And/or the angle according to object and desktop calculates backlight angle in conjunction with the angle of preset light and desktop;Back
Light angle is that light is that initial edge rotates the angle generated to object;When backlight angle is less than preset backlight angle threshold value, sentence
Disconnected user is that backlight eye is abnormal with eye health condition.
Specifically, the information of the angle of preset light and desktop, it can be according to light information known in environment come really
It is fixed, such as when using the camera that table lamp device is arranged in as image acquisition equipment, the relative position of camera and light source
It can be obtained by calibration, i.e. the light irradiation of table lamp device itself is Given information, then can be easy by geometry calculating
Know that light falls in the front or the back side of object, if falling in the back side, backlight angle is negative at this time, then is judged as back
Light, as shown in Figure 2;If falling in front and the included angle A of light and object less than 30 degree, also it is judged as backlight, as shown in Figure 3;
Specifically, the information of the angle of preset light and desktop, can be used sensor detection to determine, passes through setting
Sensor in fixed position is detected the irradiating angle to determine practical light to light;Specifically, if light source position not
Know or that using light source, then the astigmatism that default light source is penetrated under being vertically is not directly counted by the angle of object and desktop at this time
Backlight angle is calculated, as long as angle is greater than 45 degree, is judged as backlight.
Use eye judgment rule by above-mentioned, according to it is different with eye scene to user with eye distance from or backlight reading conditions
Judged, user monitor comprehensively and timely with eye health condition with eye scene for user is different.
S6, when determine user with eye situation is with eye exception, generate the alarm signal for triggering alarm.Tool
Body, alarm can be the audible alarm units such as buzzer, loudspeaker, it is also possible to physical alert device, such as sprayer, or
Indicative LED light carries out mute prompt by lighting for indicator light.
The present embodiment is disclosed to use eye health condition monitoring method, can be from the human body head image got and with eye field
Required scene information and face information are obtained in scape image in real time, and according to preset scene judgment rule, in advance to user
Judged with eye scene, further according to it is different with eye scene combine with eye judgment rule to user with eye distance from or backlight
Reading conditions are judged, by these steps, this method can get comprehensive characteristics of image letter in eye-using environment in real time
Breath, and for user it is different with eye scene to user with eye health condition carries out comprehensively with timely monitor, breach with
Toward the limitation for the high rate of false alarm for monitored with eye using single sensor, the accuracy monitored with eye is substantially increased, it can
Eye monitoring effect is used to reach good.
Embodiment 2
As shown in figure 4, the present embodiment, which corresponds in embodiment 1, uses eye health condition monitoring method, a kind of use is disclosed
Eye health condition monitors system, comprising:
Image acquiring device 1, for obtaining human body head image and with eye scene image;
Scene is calculated for extracting human body head image and with the characteristic point of eye scene image in image processing apparatus 2
Information and face information;Scene information includes the angle of desktop plane, object profile and object and desktop;Object is paper reading
Object or electronic reading device;Face information includes human face coordinate and facial orientation angle;
Specifically, being provided with execution code in image processing apparatus 2, scene information is calculated for executing following steps:
It extracts with the desktop features point and object characteristic point in eye scene image;
Desktop plane is fitted according to desktop features point;
It is to identify object profile according to object characteristic point;
The angle of object and desktop is calculated according to desktop plane and object profile.
Through the above steps, image processing apparatus can be according to each characteristic point in realtime graphic on desktop plane and table
Object identified, and calculate the angle of object and desktop, convenient for it is subsequent according to these information to user's eye health
Situation carries out more accurately judgement and monitoring.
Specifically, being provided with execution code in image processing apparatus 2, face information is calculated for executing following steps:
Obtain the characteristic point in human body head image;
The coordinate of each organ of face is determined according to characteristic point;
Using fitting algorithm, according to the coordinate of each organ, obtain face in three dimensions towards angle.
Scene judgment means 3, for using preset scene judgment rule, according to scene information and face information, judgement
User's uses eye scene;
Specifically, execution code is provided in scene judgment means 3, for executing following steps judgement eye scene:
According to the pupil coordinate in facial orientation information and human face coordinate, the direction of visual lines of user is calculated;
Judge whether to recognize object profile:
If recognizing object profile, judge whether object profile is on direction of visual lines, if so, judging that user's uses eye field
Scape is to read scene;
If not recognizing object profile, judge whether desktop plane is on direction of visual lines, if so, judging the use of user
Eye scene is to write scene.
According to scene information, face information and eye is used for using eye judgment rule using preset with eye judgment means 4
Scene, calculate user with eye distance from or backlight reading conditions, judge user use eye health condition.
Specifically, with execution code is provided in eye judgment means 4, for executing following steps judgement eye health feelings
Condition:
When being to write scene with eye scene:
According to eyes coordinates and desktop plane, calculate eyes to desktop Writing distance;It is preset when Writing distance is less than
Writing distance threshold value when, judge user with eye health condition be with eye exception;
And/or the tilt angle on head is calculated using the nose of face and the coordinate of eyes, and calculate head inclination
The angle of the Z axis of angle and predetermined camera coordinate system;When angle is less than preset slanted angle threshold value, the use of user is judged
Eye health condition is with eye exception;
When being to read scene with eye scene:
According to eyes coordinates and object profile calculate eyes to object reading distance, when reading distance be less than it is preset
When reading distance threshold value, judge user is with eye exception with eye health condition;
And/or the angle according to object and desktop calculates backlight angle in conjunction with the angle of preset light and desktop;Back
Light angle is that light is that initial edge rotates the angle generated to object;When backlight angle is less than preset backlight angle threshold value, sentence
Disconnected user is that backlight eye is abnormal with eye health condition.
Specifically, alarm signal can be generated when judge user with eye health condition is with eye exception with eye judgment means 4
Number.
Specifically, monitoring system with eye health condition disclosed in the present embodiment further includes a warning device, receiving
It can automatic alarm when to above-mentioned alarm signal.Specifically, warning device can be the audible alarm units such as buzzer, loudspeaker,
It is also possible to physical alert device, such as sprayer or indicative LED light, mute prompt is carried out by lighting for indicator light.
It is supervised disclosed in eye health condition monitoring system and embodiment 1 with eye health condition disclosed in the present embodiment 2
Survey method is corresponding, and specific technical detail is also similar with technical effect, and details are not described herein.
Embodiment 3
As shown in figure 5, the present embodiment corresponds to technical solution disclosed in Examples 1 and 2, a kind of eye health feelings are disclosed
Condition monitors desk lamp comprising Desk lamp feet 1, desk lamp support 2, adjustable support component 3 and illuminace component 4;Specifically, in support
At vertical direction setting, there are four cameras 5 on 2;Camera 5 for being obtained to the real-time imaging of user, carry out it is three-dimensional at
As forming 3-D image;Specifically, including the image of human body head and the image of desktop object in 3-D image;
It connect specifically, being provided with processor in desk lamp with camera, is provided with code form as implemented in processor
Image processing apparatus described in example 2, scene judgment means and with eye judgment means, can be performed method disclosed in embodiment 1,
The image that camera obtains is handled in real time and judges that user's uses eye situation;Specifically, desk lamp is additionally provided with loudspeaker, can sentence
Disconnected user with eye situation be with eye exception when, make a sound and alarm, user is reminded, with timely correction user's
Mistake is accustomed to eye.
More than, it is only presently preferred embodiments of the present invention, is not intended to limit the present invention in any form, thus it is all
Without departing from technical solution of the present invention content, any modification to the above embodiments, equivalent according to the technical essence of the invention
Variation and modification, still fall within the range of technical solution of the present invention.