CN108563992B - A kind of vision survey system based on recognition of face - Google Patents
A kind of vision survey system based on recognition of face Download PDFInfo
- Publication number
- CN108563992B CN108563992B CN201810203408.7A CN201810203408A CN108563992B CN 108563992 B CN108563992 B CN 108563992B CN 201810203408 A CN201810203408 A CN 201810203408A CN 108563992 B CN108563992 B CN 108563992B
- Authority
- CN
- China
- Prior art keywords
- picture
- face
- video camera
- entourage
- mouth
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000009182 swimming Effects 0.000 claims abstract description 37
- 230000003044 adaptive effect Effects 0.000 claims abstract description 13
- 210000001508 eye Anatomy 0.000 claims description 45
- 238000000034 method Methods 0.000 claims description 14
- 238000004458 analytical method Methods 0.000 claims description 13
- 230000006399 behavior Effects 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 8
- 210000000887 face Anatomy 0.000 claims description 7
- 210000003128 head Anatomy 0.000 claims description 6
- 238000001514 detection method Methods 0.000 claims description 5
- 238000005259 measurement Methods 0.000 claims description 4
- 230000000474 nursing effect Effects 0.000 claims description 4
- 230000003321 amplification Effects 0.000 claims description 3
- 238000003199 nucleic acid amplification method Methods 0.000 claims description 3
- 238000004321 preservation Methods 0.000 claims description 3
- 230000009471 action Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 5
- 230000001815 facial effect Effects 0.000 description 4
- 238000006243 chemical reaction Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000005192 partition Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 235000001674 Agaricus brunnescens Nutrition 0.000 description 1
- 241001446467 Mama Species 0.000 description 1
- 241001442654 Percnon planissimum Species 0.000 description 1
- 238000003287 bathing Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000005252 bulbus oculi Anatomy 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Human Computer Interaction (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of vision survey system based on recognition of face, including ring wall, gate is provided on ring wall, swimming pool is provided among ring wall, video camera there are three being uniformly arranged on ring wall, distance measuring sensor is provided on video camera, arc-shaped guide rail is provided between video camera and ring wall, arc-shaped guide rail is connected equipped with video camera, and it is connect by adaptive connection component with power slide, adaptive connection component includes push rod and throw-out collar, push rod one end is fixed on sliding shoe, throw-out collar inner shaft, which is connected with, is equipped with pushing sleeve, push rod one end is arranged across pushing sleeve, ball there are two being symmetrical arranged in pushing sleeve, push rod is arranged between two balls, shifting chute is provided on the corresponding push rod of ball, shifting chute inner wall is arcuate structure.The present invention can play effective booster action, and caregiver or parent's moment is reminded to pay close attention to situation.
Description
Technical field
A kind of the present invention relates to a kind of vision survey system based on recognition of face, and in particular to prospecting of accompanying and attending to of infant
System.
Background technique
Infant swimming refer to newborn's (the birth same day was new life to 28 days) or in 2 one full year of life baby in care-giving professional or
Under nurse by training baby parent, in specific interim infant's water of professional infant swimming equipment progress
Early stage health care activity, be divided into it is orderly, have position, have the passive Swim exercise of the baby of skill and two parts of independently swimming.
Hundreds of thousands of names 0-1 year old baby of most domestic key city is in swimming center, hospital, infant swimming at present
In shop, family carry out infant swimming training, infant swimming to baby's bring constitution, at heart, good influences of intelligence order youth
Papa and Mama be profuse in praise infant swimming.
With the increase of demand, many infant swimming places emerge out like the mushrooms after rain, while offering convenience,
Also there is the accident being drowned in swimming along with the close of danger, many babies or children, it is big absolutely from the analysis of accident
Part be all a lack of accompany and attend to keep an eye on and keep an eye on it is not in place and occur.
Summary of the invention
The technical problem to be solved in the present invention is to provide a kind of vision survey system based on recognition of face, can play
Effective booster action reminds caregiver or parent's moment to pay close attention to situation.
In order to solve the above-mentioned technical problems, the present invention provides a kind of vision survey system based on recognition of face, including
Ring wall is provided with gate on the ring wall, is provided with swimming pool among the ring wall, on the ring wall
Even setting is provided with distance measuring sensor there are three video camera on the video camera, one of video camera is arranged on gate opposite
Ring wall on, be provided with arc-shaped guide rail between the video camera and ring wall, the sliding shoe of the arc-shaped guide rail with take the photograph
As instrument connection, the sliding shoe is connect by adaptive connection component with power slide, and the adaptive connection component includes pushing away
Bar and throw-out collar, described push rod one end are fixed on sliding shoe, and the throw-out collar inner shaft, which is connected with, is equipped with pushing sleeve, described push rod one end
It is arranged across pushing sleeve, is symmetrically arranged with the first ball and the second ball in the pushing sleeve, first ball and the second ball can
On pushing sleeve inner wall, the push rod is arranged between the first ball and the second ball for the setting of rotation, first ball and the
Shifting chute is provided on the corresponding push rod of two balls, the shifting chute inner wall is arcuate structure.
Further, including following prospecting step:
A, video image acquisition is carried out by three video cameras, the video image of entourage is subjected to typing preservation, and
The distance between video camera and entourage data are recorded by distance measuring sensor simultaneously；
B, when infant swims, picture interception is carried out to the video image acquired in step a；
First in same timing node, while the video image of three video cameras acquisition is intercepted, obtains three
Picture, and analyze, the face information in picture is obtained by Haar eigenface recognition methods, is existed when in an only picture
When face information, next step analysis is carried out to the picture；When, there are when face information, passing through survey in two or three pictures
Video camera and the distance between entourage away from sensor detection, with determine entourage between video camera and swimming pool or
Person's swimming pool is between video camera and entourage, when the entourage in judgement picture is respectively positioned between video camera and swimming pool
When, it sounds an alarm, when determining that the swimming pool in picture is between video camera and entourage, is carried out in next step using this picture
Analysis；
C, the picture for needing to analyze in step b is analyzed, is obtained in picture by Haar eigenface recognition methods
Face location information, i.e. positioning human eye and mouth region judges to accompany and attend to by human eye in picture and mouth position relationship
Whether personnel nurse, the specific steps are as follows:
C1, positioning position of human eye, mark the position of two eyes using two rectangles, and two human eye rectangles are calculated
Central point obtain eye reference line by connecting two central points；
C2, positioning mouth position, the position of mouth are marked using rectangle, and the central point of mouth rectangle is calculated；
C3, position in the picture of screenshot human eye at a distance from mouth, by the central point of mouth rectangle and eye reference line into
Row is vertically connected, and obtains being separated by height H1；
C4, height H will be separated by1With calibrated altitude HNComparison, works as H1<HNWhen, warning is issued, H is worked as1≥HNWhen, jump to step
Rapid b carries out picture interception to the video image acquired in step a again and analyzes, when being wherein divided into first between picture interception
Between.
Further, in stepb, when face can not be navigated in picture, step b is repeated, and with the second time
Face detection is carried out to be spaced quickly interception picture, when being more than that the third time can not also navigate to face, issues warning, when the
After navigating to face in three times, step b is repeated, and to be at the first time that interval intercepts picture；
Wherein, the second time < first time < third time.
Further, the essential information that typing entourage is needed before step a passes through Haar eigenface recognition methods
Entourage is obtained with the face information under standard stance, face, human eye and mouth are marked by rectangle and record rectangular dimension,
Rectangular dimension is recorded as standard faces frame, standard human eye frame and standard mouth frame；
Spacing between the distance measuring sensor measurement entourage and video camera；Entourage is being obtained with standard stance
Under face information when, measure the spacing between entourage and video camera, be recorded as LN, in stepb, to video image into
While row picture intercepts, the spacing between entourage and video camera is measured, L is recorded as1, in step c, first to screenshot
Picture amplifies, amplification factor L1/LN, then using standard faces frame, standard human eye frame and standard mouth frame to picture
In face location positioning.
Further, entourage is calculated to be separated by size between the human eye and mouth under standard stance, is separated by size
It is recorded as height of foundation HA, pass through height of foundation HAWith bow after the difference in height that generates bow height, i.e. calibrated altitude is calculated
HN；
Wherein, height of foundation HACalculation method and be separated by height H1Calculation method it is consistent；
When being separated by height H1Less than calibrated altitude HNWhen, the locating human face rectangular area in video image calculates its center,
To determine head center (Hx, Hy), wherein Hx and Hy is the position of head center in the picture, i.e. pixel coordinate value, record head
It is jumped when Hy has fluctuation to head center motion profile within a time-count cycle, and fluctuates to nurse portion center movement track
Step b is gone to, picture interception is carried out to the video image acquired in step a again and is analyzed；When head center motion profile exists
Hy is without fluctuation in one time-count cycle, or there is fluctuation but be not when nursing, to issue warning；
Wherein, Hy < β, to come back, Hy >=β, to bow, wherein β is to delimit the standard whether nursed in vertical direction, low
Head is considered as there is no nurse behavior, and new line is considered as that there are nurse behaviors.
Beneficial effects of the present invention:
One, by the design of ring wall and three arc-shaped guide rails, enable video camera while mobile, carry out angle
The transformation of degree, so as to multi-angle locating human face, to capture facial information, to accurately judge；
Two, adaptive connection component can cooperate transmitting power between arc track and linear type power slide, will be straight
The mobile power-conversion of line is the mobile power of arc, and video camera movement is smooth；
Three, by the analysis to video image, it is capable of detecting when entourage during accompanying and attending to the presence or absence of missing row
To avoid the occurrence of to lack to accompany and attend to and keeping an eye on and keep an eye on accident not in place and occurring to make warning；
Four, video image is able to record entire nurse process, and can transfer the files inquiry in later period retrospect, is accompanied and attended to confirmation
The problems such as in journey with the presence or absence of missing；
Five, picture is intercepted to video image, to calculate the distance height between eye and mouth, led in picture
Height change is crossed to determine degree of bowing, to define to obtain bowing or seeing the mobile phone in the case of nursing to the degree of bowing
Or the case where not nursed, which calculate quick, and error rate and rate of false alarm are low, can feed back at the first time and
It reminds, prevents the generation of bathing accident.
Detailed description of the invention
Fig. 1 is present system structural schematic diagram；
Fig. 2 is adaptive connection component diagram of the present invention；
Fig. 3 is schematic diagram between pushing sleeve and push rod of the present invention；
Fig. 4 is double roll structure schematic diagrames of the invention；
Fig. 5 is face's analysis schematic diagram of the present invention.
Specific embodiment
The present invention will be further explained below with reference to the attached drawings and specific examples, so that those skilled in the art can be with
It more fully understands the present invention and can be practiced, but illustrated embodiment is not as a limitation of the invention.
Referring to figs. 1 to shown in Fig. 3, an embodiment of the vision survey system of the invention based on recognition of face, including ring
Shape partition wall 1 is provided with gate 2 on ring wall, and swimming pool 3 is provided among ring wall, and ring structure can be improved image letter
The capturing efficiency of breath is uniformly arranged on ring wall there are three video camera 4, is provided with distance measuring sensor, video camera on video camera
Image is recorded for shooting, and distance measuring sensor is used to measure the distance between entourage and video camera, so as to quick
Determination entourage and the distance between swimming pool, select image so as to assist quickly deleting；
Arc-shaped guide rail 5 is provided between video camera and ring wall, the ring on gate opposite is arranged in one of video camera
On shape partition wall, so that video camera has maximum moving range, the sliding shoe of arc-shaped guide rail is connect with video camera, sliding shoe
It is connect by adaptive connection component 6 with power slide 7, adaptive connection component can be pushed to move by power slide movement
Dynamic, so that video camera be driven to move on arc-shaped guide rail, the thrust of linear movement can be converted to arc by adaptive connection component
The power of shape movement；
Adaptive connection component includes push rod 8 and throw-out collar 9, and push rod one end is fixed on sliding shoe 10, throw-out collar inner shaft
It being connected with and is equipped with pushing sleeve 11, push rod one end is arranged across pushing sleeve, it is symmetrically arranged with the first ball 12 and the second ball 13 in pushing sleeve, the
One ball and the second ball are spindle rotationally arranged on pushing sleeve inner wall, and push rod is arranged between the first ball and the second ball, the
Shifting chute 14 is provided on one ball and the corresponding push rod of the second ball, shifting chute inner wall is arcuate structure, when sliding shoe is moving
When dynamic, push rod is enabled to do fan shape moving, otherwise push rod does fan shape moving, movement is can be realized in sliding shoe, in order to drive
Push rod is mobile, and using pushing sleeve, the first ball and the second ball in pushing sleeve can adapt to the angle of rotation of push rod when pushing sleeve is mobile
Degree, and shifting chute is cooperated to be adapted to the movement of push rod, to meet the fan shape moving of push rod, axis even can between pushing sleeve and throw-out collar
To increase the rotational angle for adapting to push rod, throw-out collar is fixed, guarantees power output when pushing.
Referring to shown in Fig. 4, in one embodiment, the quantity of the first ball and the second ball is 2, so that pushing sleeve energy
Enough and push rod rotates synchronously, and pushing sleeve will not collide with push rod, not generate noise, not influence swimming environment, and when moving, energy
The problem of enough reactions at the first time, there is no mobile surpluses.
Referring to Figure 5, analysis method includes following prospecting step:
Video image acquisition is carried out, the video image of entourage is subjected to typing preservation, passes through three video cameras first
Shooting, collecting is carried out to video image, the video image of entourage is carried out by typing storage by video camera, to be analyzed
The raw frames that basis and subsequent query are checked, while the distance between video camera and entourage are recorded by distance measuring sensor
Data；
When infant's swimming, picture interception is carried out to the video image of above-mentioned acquisition, the interception of the image can choose
Single-frame images is wherein divided at the first time between picture interception certainly, this is to lack nurse but not accident occur at the first time
Interval time；
In picture interception, first in same timing node, while the video image of three video camera acquisitions is carried out
Interception, obtains three pictures, and analyze, and the face information in picture is obtained by Haar eigenface recognition methods, when only
There are when face information in one picture, next step analysis is carried out to the picture, without face characteristic, explanation does not have remaining two picture
There is the back for taking caregiver or taking caregiver, there is no need to use；When in two or three pictures
There are the distance between the video camera detected when face information by distance measuring sensor and entourages, to determine entourage
Between video camera and swimming pool or swimming pool is between video camera and entourage, when the entourage in judgement picture is equal
When between video camera and swimming pool, representing all entourages facing away from swimming pool, and the behavior can not accomplish to nurse, answer at this time
This is sounded an alarm, and when determining that the swimming pool in picture is between video camera and entourage, is carried out in next step using this picture
Analysis can reduce the complexity of judgement by deleting choosing；
Then the picture of above-mentioned screenshot is analyzed, the face in picture is obtained by Haar eigenface recognition methods
Location information, i.e. positioning human eye and mouth region, when operation, can first position position of human eye, mark two eyes using two rectangles
The position of eyeball, and the central point of two human eye rectangles is calculated, by obtaining eye reference line for two central points connections,
The eye reference line is height and position locating for eyes；Then mouth position is positioned, the position of mouth is marked using rectangle, and is counted
Calculation obtains the central point of mouth rectangle, and the central point of the mouth rectangle is mouth height and position；Then in the figure of positioning screenshot
The central point of mouth rectangle and eye reference line are vertically connected by human eye at a distance from mouth in piece, obtain being separated by height
H1, this is separated by height H1, it is the level in picture；Height H will be finally separated by1With calibrated altitude HNComparison, works as H1<HNWhen,
Warning is issued, needs to pay close attention to infant's swimming state to remind, works as H1≥HNWhen, indicate that caregiver is paying close attention to infant's trip
Swimming state, therefore warning is not issued, and picture interception and analysis are carried out to the video image of acquisition again；By in picture
Human eye judges whether entourage nurses with mouth position relationship, has calculation amount few, the sensitive feature of rapid reaction, institute
It needs hardware requirement also low, reduces preparation cost, it is easy to spread.
In above-mentioned analysis, it is also necessary to which locating human face indicates no caregiver when that can not navigate to face in the picture
In the presence of or caregiver be, when there is such situation, to need extremely to pay close attention to, therefore rapidly to acquisition back to swimming pool
Video image carries out picture interception, and is intercepted with this picture as start time, and set the third time as deadline, in third
Picture is quickly intercepted in time using the second time as interval and carries out Face detection, when can not also navigate to people more than the third time
Face illustrates that caregiver is not concerned with infant's swimming state for a long time, needs immediately to issue warning；When default in the third time
When position is to face, then explanation only it is temporary turn one's head or back to etc., there are nurse movements, therefore fast as interval using the second time
The operation that speed interception picture carries out Face detection terminates, and continues to position human eye and mouth region with the above-mentioned image for navigating to face
And analyze, picture interception then is carried out to the video image of acquisition again, is needed at this time to be at the first time that interval intercepts picture；
Wherein, it the second time < first time < third time, saves and has no effect in time after the third time is overboard
Safety time section is at the first time any time period in safety time, and the second time was smaller than at the first time, reaches and quickly divides
The effect distinguished avoids interval time too long and misses caregiver and turn one's head the image checked or irregularly occur checking, reduces
Rate of false alarm improves and uses sensitivity.
Of course, it is possible to which the essential information of typing entourage passes through Haar eigenface recognition methods before admission swimming
Entourage is obtained with the face information under standard stance, face, human eye and mouth are marked by rectangle and record rectangular dimension,
Rectangular dimension is recorded as standard faces frame, standard human eye frame and standard mouth frame；It further include distance measuring sensor, distance measuring sensor
The spacing between entourage and video camera is measured, when obtaining entourage with face information under standard stance, measurement is accompanied
Spacing between shield personnel and video camera, the distance can be demarcated directly, and entourage is allowed to stand on uniform location, this
Distance is recorded as LN, while carrying out picture interception to video image, measure the spacing between entourage and video camera, note
Record is L1, then the picture of screenshot is amplified, amplification factor L1/LN, then use standard faces frame, standard human eye frame
And standard mouth frame positions the face location in picture, and under the method positioning of which, measured obtained data energy
It is enough by facial image taken under spacing different under any time period also to primary standard facial image size, therefore have more
Accurately to compare basis, comparing result is reliable, reduces rate of false alarm.
In another embodiment, first calculate entourage to be separated by size between the human eye and mouth under standard stance,
It is separated by size and is recorded as height of foundation HA, pass through height of foundation HAWith bow after the difference in height that generates height of bowing is calculated, i.e.,
Calibrated altitude HN；Height of foundation HACalculation method and be separated by height H1Calculation method it is consistent, when being separated by height H1Less than standard
Height HNWhen, the locating human face rectangular area in video image calculates its center, to determine head center (Hx, Hy), wherein Hx
It is the position of head center in the picture, i.e. pixel coordinate value with Hy, records head center movement track, when head center moves
There is fluctuation in track Hy within a time-count cycle, and fluctuating is to carry out picture when nursing to the video image of acquisition again and cut
It takes and analyzes；When Hy or has fluctuation but when for nurse without fluctuation to head center motion profile within a time-count cycle,
Issue warning；
Wherein, Hy < β, to come back, Hy >=β, to bow, β is that the standard whether nursed delimited in vertical direction, view of bowing
For there is no nurse behavior, new line is considered as there are nurse behavior, under standard stance, can set and bow in front of entourage
Nurse calibration object checks that when bowing nurse calibration object be that there are the minimum angles of bowing of nurse behavior, in standard faces information
After acquisition, entourage, which bows, to be checked to bow and demarcates object, is bowed the image of process by video camera acquisition, and using above-mentioned
β standard is calculated in method, according to everyone face information, can record everyone different β standard respectively, improve analysis
Precision.
In one embodiment, between the distance measuring sensor of setting can measure between the head of entourage and video camera
Away from practical spacing being recorded as, can first typing or by between distance measuring sensor measurement swimming pool and video camera before use
Spacing, be recorded as normal pitch, through practical spacing compared with normal pitch, when the two is close, can assert caregiver
Member is located on pool edge, then by the shooting of video camera, captures facial image, and when that can capture face, representative is nursed
Personnel pay close attention at station infant's swimming state on pool edge, therefore can not have to positioning human eye and mouth region, can directly make
Judge out, reduces the complexity of calculating.
Embodiment described above is only to absolutely prove preferred embodiment that is of the invention and being lifted, protection model of the invention
It encloses without being limited thereto.Those skilled in the art's made equivalent substitute or transformation on the basis of the present invention, in the present invention
Protection scope within.Protection scope of the present invention is subject to claims.
Claims (5)
1. a kind of vision survey system based on recognition of face, which is characterized in that including ring wall, set on the ring wall
It is equipped with gate, swimming pool is provided among the ring wall, is uniformly arranged on the ring wall there are three video camera, it is described to take the photograph
As being provided with distance measuring sensor on instrument, one of video camera is arranged on the ring wall on gate opposite, the video camera with
Arc-shaped guide rail is provided between ring wall, the sliding shoe of the arc-shaped guide rail is connect with video camera, and the sliding shoe passes through certainly
It adapts to connection component to connect with power slide, the adaptive connection component includes push rod and throw-out collar, and described push rod one end is fixed
It is arranged on sliding shoe, the throw-out collar inner shaft, which is connected with, is equipped with pushing sleeve, and described push rod one end is arranged across pushing sleeve, right in the pushing sleeve
Title is provided with the first ball and the second ball, and first ball and the second ball are spindle rotationally arranged on pushing sleeve inner wall, institute
It states push rod to be arranged between the first ball and the second ball, is provided with shifting on first ball and the corresponding push rod of the second ball
Dynamic slot, the shifting chute inner wall are arcuate structure；
Adaptive connection component can be pushed mobile by power slide movement, so that video camera be driven to move up in arc-shaped guide rail
Dynamic, the thrust of linear movement can be converted to the power of arcuate movement by adaptive connection component.
2. the vision survey system based on recognition of face as described in claim 1, which is characterized in that walked including following prospecting
It is rapid:
A, video image acquisition is carried out by three video cameras, the video image of entourage is subjected to typing preservation, and simultaneously
The distance between video camera and entourage data are recorded by distance measuring sensor；
B, when infant swims, picture interception is carried out to the video image acquired in step a；First in same timing node
In, while the video image of three video cameras acquisition is intercepted, three pictures are obtained, and analyze, passes through Haar feature people
Face recognition method obtains the face information in picture, when, there are when face information, carrying out down to the picture in an only picture
The analysis of one step；When the video camera there are when face information, detected by distance measuring sensor in two or three pictures with accompany
The distance between shield personnel, to determine entourage between video camera and swimming pool or swimming pool is located at video camera and accompanies and attends to people
It between member, when determining that the entourage in picture is respectively positioned between video camera and swimming pool, sounds an alarm, when in judgement picture
When swimming pool is between video camera and entourage, next step analysis is carried out using this picture；
C, the picture for needing to analyze in step b is analyzed, the people in picture is obtained by Haar eigenface recognition methods
Face location information, i.e. positioning human eye and mouth region judge entourage with mouth position relationship by the human eye in picture
Whether nursed, the specific steps are as follows:
C1, positioning position of human eye, mark the position of two eyes using two rectangles, and be calculated in two human eye rectangles
Heart point obtains eye reference line by connecting two central points；
C2, positioning mouth position, the position of mouth are marked using rectangle, and the central point of mouth rectangle is calculated；C3, positioning
The central point of mouth rectangle and eye reference line are vertically connected by human eye at a distance from mouth in the picture of screenshot, are obtained
It is separated by height H1；
C4, height H will be separated by1With calibrated altitude HNComparison, works as H1< HNWhen, warning is issued, H is worked as1≥HNWhen, step b is jumped to,
Again picture interception is carried out to the video image acquired in step a and analyzed, is wherein divided at the first time between picture interception.
3. the vision survey system based on recognition of face as claimed in claim 2, which is characterized in that in stepb, when can not
When navigating to face in picture, step b is repeated, and quickly intercepts picture using the second time as interval and carries out Face detection, when
Face can not be also navigated to more than the third time, issues warning, after navigating to face within the third time, repeats step b, and
To be at the first time that interval intercepts picture；Wherein, the second third time < time, < first time.
4. the vision survey system based on recognition of face as claimed in claim 2, which is characterized in that need to record before step a
The essential information for entering entourage is obtained entourage by Haar eigenface recognition methods and is believed with the face under standard stance
Breath, marks face, human eye and mouth by rectangle and records rectangular dimension, rectangular dimension is recorded as standard faces frame, standard people
Eye frame and standard mouth frame；Spacing between the distance measuring sensor measurement entourage and video camera；It accompanies and attends to people in acquisition
When member is with face information under standard stance, the spacing between entourage and video camera is measured, L is recorded asN, in stepb,
While carrying out picture interception to video image, the spacing between entourage and video camera is measured, L is recorded as1, in step c
In, first the picture of screenshot is amplified, amplification factor L1/LN, then use standard faces frame, standard human eye frame and mark
Quasi- mouth frame positions the face location in picture.
5. the vision survey system based on recognition of face as claimed in claim 4, which is characterized in that calculate entourage to mark
It is separated by size between human eye and mouth under quasi- stance, is separated by size and is recorded as height of foundation HA, pass through height of foundation HAWith it is low
Height of bowing, i.e. calibrated altitude H is calculated in the difference in height generated after headN；
Wherein, height of foundation HACalculation method and be separated by height H1Calculation method it is consistent；
When being separated by height H1Less than calibrated altitude HNWhen, the locating human face rectangular area in video image calculates its center, with true
Determine head center (Hx, Hy), wherein Hx and Hy is the position of head center in the picture, i.e. pixel coordinate value, is recorded in head
Heart motion profile is jumped to when Hy has fluctuation to head center motion profile within a time-count cycle, and fluctuates to nurse
Step b carries out picture interception to the video image acquired in step a again and analyzes；When head center motion profile is at one
Hy is without fluctuation in time-count cycle, or there is fluctuation but be not when nursing, to issue warning；
Wherein, Hy < β, to come back, Hy >=β, to bow, wherein β is that the standard whether nursed delimited in vertical direction, view of bowing
For there is no nurse behavior, new line is considered as that there are nurse behaviors.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810203408.7A CN108563992B (en) | 2018-03-13 | 2018-03-13 | A kind of vision survey system based on recognition of face |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810203408.7A CN108563992B (en) | 2018-03-13 | 2018-03-13 | A kind of vision survey system based on recognition of face |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108563992A CN108563992A (en) | 2018-09-21 |
| CN108563992B true CN108563992B (en) | 2019-11-15 |
Family
ID=63531639
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810203408.7A Active CN108563992B (en) | 2018-03-13 | 2018-03-13 | A kind of vision survey system based on recognition of face |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108563992B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10964187B2 (en) * | 2019-01-29 | 2021-03-30 | Pool Knight, Llc | Smart surveillance system for swimming pools |
| CN110852150B (en) * | 2019-09-25 | 2022-12-20 | 珠海格力电器股份有限公司 | Face verification method, system, equipment and computer readable storage medium |
| CN111414789A (en) * | 2019-10-20 | 2020-07-14 | 上海辰努智能科技有限公司 | Video gridding-based human body turn-back behavior judgment method |
| CN111428668B (en) * | 2020-03-31 | 2024-03-08 | 中科人工智能创新技术研究院(青岛)有限公司 | Multi-view variable-pitch type face data acquisition system and method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102436715A (en) * | 2011-11-25 | 2012-05-02 | 大连海创高科信息技术有限公司 | Fatigue driving detection method |
| CN103578226A (en) * | 2012-07-31 | 2014-02-12 | 青岛联合创新技术服务平台有限公司 | Video capturing alarm device used for swimming pool and warning method thereof |
| CN204447228U (en) * | 2015-01-28 | 2015-07-08 | 王冬生 | Swimming teaching is with instructing surveillance equipment |
| CN105485198A (en) * | 2015-12-29 | 2016-04-13 | 武汉正通传动技术有限公司 | Push-pull force overload protective device |
| CN107595307A (en) * | 2017-10-23 | 2018-01-19 | 湖南科技大学 | Fatigue driving detection device and detection method based on machine vision eye recognition |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101877051A (en) * | 2009-10-30 | 2010-11-03 | 江苏大学 | Driver attention state monitoring method and device |
| CN102693606B (en) * | 2011-03-24 | 2016-07-06 | 中国人民解放军国防科学技术大学 | A kind of swimming pool anti-drowning method for early warning based on shooting network and system |
| CN104344017B (en) * | 2014-10-28 | 2016-05-25 | 华中科技大学 | A kind of flat closed type reversal valve |
-
2018
- 2018-03-13 CN CN201810203408.7A patent/CN108563992B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102436715A (en) * | 2011-11-25 | 2012-05-02 | 大连海创高科信息技术有限公司 | Fatigue driving detection method |
| CN103578226A (en) * | 2012-07-31 | 2014-02-12 | 青岛联合创新技术服务平台有限公司 | Video capturing alarm device used for swimming pool and warning method thereof |
| CN204447228U (en) * | 2015-01-28 | 2015-07-08 | 王冬生 | Swimming teaching is with instructing surveillance equipment |
| CN105485198A (en) * | 2015-12-29 | 2016-04-13 | 武汉正通传动技术有限公司 | Push-pull force overload protective device |
| CN107595307A (en) * | 2017-10-23 | 2018-01-19 | 湖南科技大学 | Fatigue driving detection device and detection method based on machine vision eye recognition |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108563992A (en) | 2018-09-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108563992B (en) | A kind of vision survey system based on recognition of face | |
| US20210049353A1 (en) | Ai-based physical function assessment system | |
| Stone et al. | Fall detection in homes of older adults using the Microsoft Kinect | |
| CN102103663B (en) | Ward visit service robot system and target searching method thereof | |
| JPWO2019239813A1 (en) | Information processing method, information processing program and information processing system | |
| CN109215806A (en) | A kind of public place health monitoring systems and method based on recognition of face | |
| RU2618172C2 (en) | Device and method for device movement and orientation tracking | |
| Lohmann et al. | Skeleton timed up and go | |
| US20210212595A1 (en) | Apparatus for sensing | |
| CN108549838A (en) | A kind of back-up surveillance method of view-based access control model system | |
| Negishi et al. | Stable contactless sensing of vital signs using RGB-thermal image fusion system with facial tracking for infection screening | |
| CN108596022A (en) | A kind of mobile image identifying system | |
| Duncan et al. | Camera-based short physical performance battery and timed up and go assessment for older adults with cancer | |
| Pueo et al. | Video-based system for automatic measurement of barbell velocity in back squat | |
| Formosa et al. | Quantifying stroke coordination during the breathing action in front-crawl swimming using an instantaneous net drag force profile | |
| CN113069088B (en) | Artificial intelligence interaction device | |
| US20210282646A1 (en) | Medical data measurement system | |
| Khan et al. | Joint use of a low thermal resolution thermal camera and an rgb camera for respiration measurement | |
| Inoue et al. | Bed exit action detection based on patient posture with long short-term memory | |
| Hill et al. | Touchless respiratory monitor preliminary data and results | |
| WO2022019001A1 (en) | Evaluation device, evaluation method, and program | |
| Liao et al. | A vision-based walking posture analysis system without markers | |
| de Albuquerque et al. | Kinect sensor used as a support tool in clinical analysis | |
| US20210307616A1 (en) | Fever Detector | |
| Vitanza et al. | Skeleton Timed Up and Go on MARIO Robot |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20190929 Address after: 201700 Shanghai Qingpu District heavy Guzhen Zhao heavy road 2778 Lane 26, 201 room. Applicant after: Shanghai Shen Xue Supply Chain Management Co., Ltd. Address before: 215000 room 502-1, lotus building, 393 Chunshen Middle Road, Xiangcheng District, Suzhou, Jiangsu. Applicant before: Suzhou Ao Ke De Rui Intelligent Technology Co., Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |