CN104269028A - Fatigue driving detection method and system - Google Patents
Fatigue driving detection method and system Download PDFInfo
- Publication number
- CN104269028A CN104269028A CN201410568886.XA CN201410568886A CN104269028A CN 104269028 A CN104269028 A CN 104269028A CN 201410568886 A CN201410568886 A CN 201410568886A CN 104269028 A CN104269028 A CN 104269028A
- Authority
- CN
- China
- Prior art keywords
- face
- state
- fatigue driving
- value
- driver
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
- G08B21/02—Alarms for ensuring the safety of persons
- G08B21/06—Alarms for ensuring the safety of persons indicating a condition of sleep, e.g. anti-dozing alarms
Landscapes
- Business, Economics & Management (AREA)
- Emergency Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a fatigue driving detection method and system. The fatigue driving detection method comprises the steps of S1, image acquisition and preprocessing; S2, face localization and detection; S3, face tracking; S4, eye detection and state recognition; S5, calculation of the fatigue PERCLOS value; S6, comparison between the obtained PERCLOS value with a preset threshold value and judgment of fatigue driving of drivers. By adopting the fatigue driving detection method and system, the microcosmic change of the faces of the drivers can be automatically captured, the state of the drivers can be judged through scientific probability calculation, the drivers who are in the fatigue driving state can be warned in time, and therefore driving safety can be guaranteed.
Description
Technical field
Design vehicle security technology area of the present invention, a kind of method for detecting fatigue driving of special design and system.
Background technology
Along with the development of economic society, motor vehicles sharply increase, and the traffic hazard that therefore fatigue driving causes also gets more and more.For this problem, create many fatigue-driving detection technologies, be summed up three kinds.
(1) based on the detection technique of physiological signal, according to the blood pressure of driver, the physiological signals such as brain wave judge whether fatigue, need high-precision detecting instrument, and contact driver, have a certain impact and cost is high to driver.
(2) based on the fatigue detecting technology of running information, when bearing circle does not revise direction for a long time, or left and right vehicle wheel sway frequency or amplitude excessive, speed wobble is just judged as fatigue driving, but for the driver of different driving habits, and the driving conditions on different road, be difficult to driving unified standard, degree of accuracy is not high.
(3) based on the fatigue detecting technology of physiological characteristic, when driver fatigue time, the physiological characteristics such as frequency increase of bowing and close one's eyes can be shown.By the method for machine vision, detect the above-mentioned physiological characteristic of driver, can judge that whether driver is tired.Based on the method for Machine Vision Detection physiological driver feature, have noncontact, cost is low, and degree of accuracy advantages of higher, is widely adopted in current fatigue detecting system.
In above-mentioned three kinds of methods, the third method is obviously better than first two method.But it also has shortcoming, the third method is mostly by image processing techniques at present, locating human face, then the state of scope inner analysis eyes at face, judges whether fatigue.Wherein locating human face is the hunting zone in order to reduce eyes, and improving treatment effeciency, is a step of fatigue detecting key.The method that wherein locating human face and eyes are conventional has two kinds, and a kind of is method based on morphological image, and a kind of is method based on machine learning and sorter.The former calculated amount is large, and speed is slow, large by illumination effect.The latter's computing velocity is fast, little by illumination effect, is widely adopted with Face detection technology.But in the technology of the sorter locating human face adopted at present, when driver head's action is excessive, when cannot collect positive face image, can not correct locating human face position, there is fatigue detecting algorithm Problem of Failure.
Summary of the invention
In order to solve the problems of the technologies described above, the present invention discloses a kind of method for detecting fatigue driving and system, and the present invention adopts following technical scheme to solve above-mentioned technical matters:
A kind of method for detecting fatigue driving, comprises the steps:
S1, collection image and pre-service, gather driver's image information by image capture interface and transfer to hub processor, and hub processor carries out pre-service to collection image;
S2, Face detection and detection, adopt the face classification device based on Haar feature, detects driver's human face region;
S3, face tracking, adopt the Face tracking algorithm of Kalman filter algorithm, track human faces;
S4, eye detection and state recognition, adopt eye opening sorter location driver's eyes and identify eye state, record recognition result;
S5, calculate tired PERCLOS value, after obtaining the state recognition result of eyes, by the number percent in the unit of account time shared by the eyes closed time and PERCLOS (Percentage of Eyelid Closure Over Time):
S6, the PERCLOS value obtained and the threshold value preset to be compared, judge driver's whether fatigue driving.
Preferably, in above-mentioned a kind of method for detecting fatigue driving, described step S2 specifically comprises steps A, adopts based on Kalman filter algorithm parameter tracking face:
A1, set up system state equation;
X
(k+1)＝A
(k)X
(k)+W
(k)
A2, set up systematic observation equation;
Z
(k)＝H
(k)X
(k)+V
(k)
A3, prediction face location, according to the predictive equation group of Kalman filter, wave filter uses the estimation of state on last stage, makes the estimation to current state;
Predictive equation group:
(1) predicted state
X
(k|k-1)＝A
(k)X
(k|k-1)
(2) predicted estimate covariance matrix
P
(k|k-1)＝A
(k)P
(k-1|k-1)A
T (k)+Q
(k)
A4, detect actual face location, behind the position predicting current face place, call Haar classifier and detect face in estimation range, obtain current state information; When can't detect face in estimation range, just detect face at full figure; In time also can't detect face at full figure, illustrate that driver head's action is excessive, cannot face be detected;
A5, use current actual face location to upgrade Kalman filter by renewal equation group, allow the predicted value of Kalman filter and actual value more and more close, reach the effect of tracking;
Renewal equation group:
(3) optimum kalman gain
K
(k)＝P
(k|k-1)H
T k[H
(k)P
(k|k-1)H
T (k)+R
(k)]
-1
(4) state estimation upgraded
X
(k|k)＝X
(k|k-1)+K
(k)[Z
(k+1)-H
(k)X]
(k|k-1)
(5) covariance upgraded is estimated
P
(k|k)＝[E-K
(k)H
(k)]P
(k|k-1)
A6, mark human face region in the input image.
Wherein X
kfor in the state estimation of face in the k moment, A
(k)act on X
(k)on state-transition matrix, H
(k)the observing matrix of system, Z
(k)the observational variable of system, W
(k)be process noise, and suppose it to meet average be null covariance matrix to be Q
(k)gaussian distribution, i.e. W
k~ N (0, Q
(k)); V
(k)be observation noise, also suppose it to meet average be null covariance matrix to be R
(k)gaussian distribution, i.e. V
k~ N (0, R
(k)); P
kfor error correlation matrix, the levels of precision of tolerance estimated value;
Preferably, in above-mentioned a kind of method for detecting fatigue driving, the eye opening sorter in described step S3 is trained through following steps:
The sample source of B1, collection face, intercepts the human eye area of face sample source and non-human eye area, is positive sample and negative sample;
B2, use machine learning algorithm training eye opening sorter.
Preferably, in above-mentioned a kind of method for detecting fatigue driving, in step S5, PERCLOS value adopts following formulae discovery:
PERCLOS value=(frame number/fixing frame number of eyes closed) * 100%
Preferably, in above-mentioned a kind of method for detecting fatigue driving, described step B2 comprises:
B21, generate positive and negative pattern representation file;
B22, generate the vector description file of positive sample;
B23, the executable file using OpenCV to carry are trained.
Compared with prior art, the present invention has following technique effect:
Adopt design of the present invention, can face's micro-variations of automatic capturing driver, differentiate driver status by the probability calculation of science, and warning is in the driver of fatigue driving state in time, ensures traffic safety.Simultaneously the present invention uses Kalman filter and related algorithm through constantly prediction, upgrades, will make predicted value and actual value more and more close, within the scope of tolerance, predicted value can be considered as actual value.When driver head's action is excessive, when cannot pass through sorter locating human face position, Kalman filter is used to follow the tracks of, real face location is replaced by the face location of prediction, continue fatigue driving to detect, avoiding cannot locating human face, the problem that fatigue detecting algorithm lost efficacy.
Accompanying drawing explanation
Fig. 1 is principle of the invention schematic diagram;
Fig. 2 is face tracking principle schematic in the present invention;
Fig. 3 is Kalman filter recursive process in the present invention;
Fig. 4 is the training method flow process of sorter of opening eyes in the present invention.
Embodiment
Below in conjunction with accompanying drawing, the present invention is described in further detail, but the present invention is not only confined to following examples, as shown in Figure 1, Figure 2, Figure 3 shows:
A kind of method for detecting fatigue driving of the present invention, comprises the steps:
S1, collection image and pre-service, gather driver's image information by image capture interface and transfer to hub processor, and hub processor carries out the pre-service such as gray processing, histogram equalization to collection image;
S2, Face detection and detection, adopt the face classification device based on Haar feature, detects driver's human face region;
S3, face tracking, adopt the Face tracking algorithm of Kalman filter algorithm, track human faces, specifically comprises the steps:
A1, set up system state equation;
X
(k+1)＝A
(k)X
(k)+W
(k)
A2, set up systematic observation equation;
Z
(k)＝H
(k)X
(k)+V
(k)
A3, prediction face location, according to the predictive equation group of Kalman filter, wave filter uses the estimation of state on last stage, makes the estimation to current state;
Predictive equation group:
(1) predicted state
X
(k|k-1)＝A
(k)X
(k|k-1)
(2) predicted estimate covariance matrix
P
(k|k-1)＝A
(k)P
(k-1|k-1)A
T (k)+Q
(k)
A4, detect actual face location, behind the position predicting current face place, call Haar classifier and detect face in estimation range, obtain current state information; When can't detect face in estimation range, just detect face at full figure; In time also can't detect face at full figure, illustrate that driver head's action is excessive, cannot face be detected;
A5, use current actual face location to upgrade Kalman filter by renewal equation group, allow the predicted value of Kalman filter and actual value more and more close, reach the effect of tracking;
Renewal equation group:
(3) optimum kalman gain
K
(k)＝P
(k|k-1)H
T k[H
(k)P
(k|k-1)H
T (k)+R
(k)]
-1
(4) state estimation upgraded
X
(k|k)＝X
(k|k-1)+K
(k)[Z
(k+1)-H
(k)X]
(k|k-1)
(5) covariance upgraded is estimated
P
(k|k)＝[E-K
(k)H
(k)]P
(k|k-1)
A6, mark human face region in the input image.
Wherein X
kfor in the state estimation of face in the k moment, A
(k)act on X
(k)on state-transition matrix, H
(k)the observing matrix of system, Z
(k)the observational variable of system, W
(k)be process noise, and suppose it to meet average be null covariance matrix to be Q
(k)gaussian distribution, i.e. W
k~ N (0, Q
(k)); V
(k)be observation noise, also suppose it to meet average be null covariance matrix to be R
(k)gaussian distribution, i.e. V
k~ N (0, R
(k)); P
kfor error correlation matrix, the levels of precision of tolerance estimated value;
The problems such as direct sorter carries out Face detection, there is calculated amount large, for a long time consuming time, and error detection is high.In order to improve speed and the accuracy of Face detection, the present invention adopts the track algorithm based on the face of Kalman filter.
But Kalman filter algorithm is applicable to linear system, system nearly all in reality is all nonlinear, only has and real system is approximately linear system, could use Kalman filter algorithm.Consider that in video between every two frames, the time interval is short, the speed of target and position basically identical between two frames, the motion of face can be approximately linear uniform motion like this, meet Kalman filter and system linearity is supposed.
Utilize Kalman filter algorithm tracking target, first need to set up system model, then by algorithm filter tracking target.
1.1 set up Linear system model:
The state of Kalman filter is represented by following Two Variables: X
(k)for the estimation of the state in k moment; P
kfor error correlation matrix, be used for measuring the levels of precision of estimated value;
The general mathematics model of system is: X
(k+1)=A
(k)x
(k)+ W
(k)(system state equation), Z
(k)=H
(k)x
(k)+ V
(k)(systematic observation equation).X
(k)the system state variables sequence in K moment, X
(k)={ x, y, w, h, v
x, v
y, v
w, v
h)
t, the status switch that the present invention chooses is coordinate (x, y) and the speed (v of face area geometric center
x, v
y), wide and high (w, the h) of face's boundary rectangle, and wide and high pace of change (v
w, v
h).A
(k)act on X
(k)on state-transition matrix; H
(k)it is the observing matrix of system; Z
(k)it is the observational variable of system; W
(k)be process noise, and to suppose that it meets average be zero, covariance matrix is Q
(k)gaussian distribution, i.e. W
k~ N (0, Q
(k)).V
(k)be observation noise, also supposing that it meets average is zero, and covariance matrix is R
(k)gaussian distribution, i.e. V
k~ N (0, R
(k)).
Set up system model, need to set up system state equation, and observation equation.
(1) system state equation is set up:
Because between two frames every in video, the time interval is short, the position (x of target in next frame
(k+1), y
(k+1)) and speed (v
x (k+1), v
y (k+1)), wide and high (w
(k+1), h
k+1)), and wide and high pace of change (v
w (k+1), v
h (k+1)) do not have too large change, then can be similar to and obtain following equations:
Again because make the state variable in K moment be above:
X
(k)＝{x
(k)?y
(k)?w
(k)?h
(k)?v
x(k)?v
y(k)?v
w(k)?v
h(k)}
T
The state variable of subsequent time is X
(k+1), then system of equations (1) can be converted into:
Then system of equations (2) can be converted to:
X
(k+1)＝A
(k)·X
(k)
Due to the linear movement that this is just approximate, so be not completely credible, there is certain error, this error is a random value, is also called process noise.The stochastic error of Kalman filter supposing the system meets the Gaussian distribution that average is 0.If process noise W
(k)meeting average is zero, and covariance matrix is Q
(k)gaussian distribution, i.e. W
k~ N (0, Q
(k)), get by test of many times the unit matrix that empirical value Q (k)=0.01 × E, E is 8*8.System state equation after then considering noise becomes:
X
(k+1)＝A
(k)X
(k)+W
(k)
(2) systematic observation equation is set up:
So in multisystem state variable, the value of the wide height (w, h) in position (x, y) of our final only pass target centroid, i.e. observation vector Z
(k)={ x y w h}
t.So obtain system of equations:
Formula (3) can be expressed as: Z
(k)=H
(k)x
(k)
The same with system state equation, also there is stochastic error in the observational variable of system, is called observation noise, and Kalman filter hypothesis observation noise also obeys the Gaussian distribution that average is 0.If observation noise V
(k)meeting average is zero, and covariance matrix is Q
(k)gaussian distribution, i.e. V
k~ N (0, R
(k)), get empirical value R by test of many times
(k)=0.1 × E, E are the unit matrix of 8*8.System state equation after then considering noise becomes:
Z
(k)＝H
(k)X
(k)+V
(k)
The tracing process of 1.2 Kalman filter:
The operation of Kalman filter comprises prediction and upgrades two stages.At forecast period, wave filter uses the estimation of laststate, makes the estimation to current state.In more new stage, wave filter utilizes the predicted value obtained at forecast period the observed reading optimization of current state, to obtain a more accurate new estimation value.
Predictive equation group: (a) predicted state
X
(k|k-1)＝A
(k)X
(k|k-1)
(b) predicted estimate covariance matrix
P
(k|k-1)＝A
(k)P
(k-1|k-1)A
T (k)+Q
(k)
Renewal equation group: (c) optimum kalman gain
K
(k)＝P
(k|k-1)H
T k[H
(k)P
(k|k-1)H
T (k)+R
(k)]
-1
D state estimation that () upgrades
X
(k|k)＝X
(k|k-1)+K
(k)[Z
(k+1)-H
(k)X]
(k|k-1)
E covariance that () upgrades is estimated
P
(k|k)＝[E-K
(k)H
(k)]P
(k|k-1)
Through type (a), can predict and obtain the state X of target at subsequent time
(1|0).Instantly, after being carved into for the moment, near predicted position, local is carried out to image and detects, detect the virtual condition X of target
(1), through type (b) to formula (e) completes the renewal to Kalman filter, ready for predicting the dbjective state of subsequent time again.So constantly recursion, constantly can predict the state of subsequent time moving target, as shown in Figure 3.
Can be found out by above-mentioned steps, Kalman filter, through constantly prediction, upgrades, will make predicted value and actual value more and more close, within the scope of tolerance, predicted value can be considered as actual value.When driver head's action is excessive, when cannot pass through sorter locating human face position, Kalman filter can be used to follow the tracks of, replace real face location by the face location of prediction, continue fatigue driving and detect.Doing so avoids cannot locating human face, the problem that fatigue detecting algorithm lost efficacy
S4, eye detection and state recognition, adopt eye opening sorter location driver's eyes and identify eye state, record recognition result, and the eye opening sorter of use is trained through following method, as shown in Figure 4:
B1, collecting sample as much as possible, comprise positive negative sample, and wherein positive sample is the human eye opening state, and negative sample is human eye and other non-eye opening regions of closure state;
B21, generate positive and negative pattern representation file;
B22, generate the vector description file of positive sample;
B23, the executable file using OpenCV to carry are trained.
The training of sorter can be carried out according to three steps of Fig. 4.
The description document of the generation sample of the first step, mainly refers to the path that describes positive sample and target image-eye opening image number in the picture and position.Negative sample only needs the path describing its sample.Therefore the form of positive sample is as follows,
The position > in the picture of the number >< target image of the file path >< target image of < sample.
All the other positive sample formats by that analogy.
Negative sample only needs to describe sample path in the picture, and its form is as follows:
Position > in the image of < sample
The descriptor format of all the other negative samples by that analogy
The vector description file of the positive sample of generation of second step, calls relevant file.
The executable file that the use OpenCV of the 3rd step carries is trained.
The present invention is according to the method training classifier of AdaBoost, and regulation trains the verification and measurement ratio of every one deck strong classifier to be 0.998, and false drop rate is 0.5.Start most to be use the sample collected, comprise positive negative sample totally 25000 sample sets, train, in the process of training, every layer of strong classifier must meet the verification and measurement ratio of 0.998 and the false drop rate of 0.5, thus makes to train and obtain the requirement that sorter effect can meet system.
Use AdaBoost algorithm training in advance well eye opening sorter carries out target detection, and compare other detection method, the speed of process is fast, and the impact of receiving illumination is little, and accuracy rate is high.And directly train eye opening sorter, while the eyes of location, just can judge that eyes are in the state of opening, and in time not detecting that eyes are all, are all considered as closed-eye state, like this location of eyes and state-detection are combined, improve detection efficiency.
S5, calculate tired PERCLOS value, after obtaining the state recognition result of eyes, by the number percent in the unit of account time shared by the eyes closed time and PERCLOS (Percentage of Eyelid Closure Over Time);
S6, the PERCLOS value obtained and the threshold value preset to be compared, judge driver's whether fatigue driving.
In step S4, PERCLOS value and the number percent in the unit interval shared by the eyes closed time, the fatigue state of driver is generally acknowledged most effective method at present to use PERCLOS value to differentiate, its computing formula is as follows:
PERCLOS value=(eyes closed time/set time) * 100%
For the ease of programming, the molecule denominator on the above formula left side, simultaneously divided by each averaging time detected, obtains following formula:
PERCLOS value=(frame number/fixing frame number of eyes closed) * 100%
In step S5, the threshold value preset rule of thumb is preset according to a large amount of experimental datas and environment usually, the PERCLOS value obtained and the threshold value preset are compared, when PERCLOS value is more than or equal to predetermined threshold value, system judge driver as fatigue driving and flog system give the alarm, driver is reminded to have a rest, to ensure traffic safety.
Adopt design of the present invention, can face's micro-variations of automatic capturing driver, differentiate driver status by the probability calculation of science, and warning is in the driver of fatigue driving state in time, ensures traffic safety.Simultaneously the present invention uses Kalman filter and related algorithm through constantly prediction, upgrades, will make predicted value and actual value more and more close, within the scope of tolerance, predicted value can be considered as actual value.When driver head's action is excessive, when cannot pass through sorter locating human face position, Kalman filter is used to follow the tracks of, real face location is replaced by the face location of prediction, continue fatigue driving to detect, avoiding cannot locating human face, the problem that fatigue detecting algorithm lost efficacy.
The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, all any amendments done within the spirit and principles in the present invention, equivalent replacement and improvement etc., all should be included within protection scope of the present invention.
Claims (5)
1. a method for detecting fatigue driving, is characterized in that, comprises the steps:
S1, collection image and pre-service, gather driver's image information by image capture interface and transfer to hub processor, and hub processor carries out pre-service to collection image;
S2, Face detection and detection, adopt the face classification device based on Haar feature, detects driver's human face region;
S3, face tracking, adopt the Face tracking algorithm of Kalman filter algorithm, track human faces;
S4, eye detection and state recognition, adopt eye opening sorter location driver's eyes and identify eye state, record recognition result;
S5, calculate tired PERCLOS value, after obtaining the state recognition result of eyes, by the number percent in the unit of account time shared by the eyes closed time and PERCLOS (Percentage of Eyelid Closure Over Time);
S6, the PERCLOS value obtained and the threshold value preset to be compared, judge driver's whether fatigue driving.
2. a kind of method for detecting fatigue driving according to claim 1, is characterized in that, described step S2 specifically comprises steps A, adopts based on Kalman filter algorithm parameter tracking face:
A1, set up system state equation;
X
(k+1)＝A
(k)X
(k)+W
(k)
A2, set up systematic observation equation;
Z
(k)＝H
(k)X
(k)+V
(k)
A3, prediction face location, according to the predictive equation group of Kalman filter, wave filter uses the estimation of state on last stage, makes the estimation to current state;
Predictive equation group:
(1) predicted state
X
(k|k-1)＝A
(k)X
(k|k-1)
(2) predicted estimate covariance matrix
P
(k|k-1)＝A
(k)P
(k-1|k-1)A
T (k)+Q
(k)
A4, detect actual face location, behind the position predicting current face place, call Haar classifier and detect face in estimation range, obtain current state information; When can't detect face in estimation range, just detect face at full figure; In time also can't detect face at full figure, illustrate that driver head's action is excessive, cannot face be detected;
A5, use current actual face location to upgrade Kalman filter by renewal equation group, allow the predicted value of Kalman filter and actual value more and more close, reach the effect of tracking;
Renewal equation group:
(3) optimum kalman gain
K
(k)＝P
(k|k-1)H
T k[H
(k)P
(k|k-1)H
T (k)+R
(k)]
-1
(4) state estimation upgraded
X
(k|k)＝X
(k|k-1)+K
(k)[Z
(k+1)-H
(k)X]
(k|k-1)
(5) covariance upgraded is estimated
P
(k|k)＝[E-K
(k)H
(k)]P
(k|k-1)
A6, mark human face region in the input image.
Wherein X
kfor in the state estimation of face in the k moment, A
(k)act on X
(k)on state-transition matrix, H
(k)the observing matrix of system, Z
(k)the observational variable of system, W
(k)be process noise, and suppose it to meet average be null covariance matrix to be Q
(k)gaussian distribution, i.e. W
k~ N (0, Q
(k)); V
(k)be observation noise, also suppose it to meet average be null covariance matrix to be R
(k)gaussian distribution, i.e. V
k~ N (0, R
(k)); P
kfor error correlation matrix, the levels of precision of tolerance estimated value;
State-transition matrix
Systematic observation matrix
3. a kind of method for detecting fatigue driving according to claim 1, is characterized in that: the eye opening sorter in described step S3 is trained through following steps:
The sample source of B1, collection face, intercepts the human eye area of face sample source and non-human eye area, is positive sample and negative sample;
B2, use machine learning algorithm training eye opening sorter.
4. a kind of method for detecting fatigue driving according to claim 1, is characterized in that: in step S5, and PERCLOS value adopts following formulae discovery:
PERCLOS value=(frame number/fixing frame number of eyes closed) * 100%.
5. a kind of method for detecting fatigue driving according to claim 3, is characterized in that, described step B2 comprises:
B21, generate positive and negative pattern representation file;
B22, generate the vector description file of positive sample;
B23, the executable file using OpenCV to carry are trained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410568886.XA CN104269028B (en) | 2014-10-23 | 2014-10-23 | Fatigue driving detection method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410568886.XA CN104269028B (en) | 2014-10-23 | 2014-10-23 | Fatigue driving detection method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104269028A true CN104269028A (en) | 2015-01-07 |
| CN104269028B CN104269028B (en) | 2017-02-01 |
Family
ID=52160545
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410568886.XA Expired - Fee Related CN104269028B (en) | 2014-10-23 | 2014-10-23 | Fatigue driving detection method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104269028B (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104573657A (en) * | 2015-01-09 | 2015-04-29 | 安徽清新互联信息科技有限公司 | Blind driving detection method based on head lowing characteristics |
| CN105160913A (en) * | 2015-08-17 | 2015-12-16 | 上海斐讯数据通信技术有限公司 | Method and apparatus for standardizing driving behaviors of drivers |
| CN106203394A (en) * | 2016-07-26 | 2016-12-07 | 浙江捷尚视觉科技股份有限公司 | Fatigue driving safety monitoring method based on human eye state detection |
| CN106295474A (en) * | 2015-05-28 | 2017-01-04 | 交通运输部水运科学研究院 | The fatigue detection method of deck officer, system and server |
| CN106447828A (en) * | 2016-10-10 | 2017-02-22 | 深圳市驰勇科技有限公司 | Automobile data recorder with iris recognition function |
| CN108229359A (en) * | 2017-12-26 | 2018-06-29 | 大唐软件技术股份有限公司 | A kind of face image processing process and device |
| CN108446600A (en) * | 2018-02-27 | 2018-08-24 | 上海汽车集团股份有限公司 | A kind of vehicle driver's fatigue monitoring early warning system and method |
| CN110852247A (en) * | 2019-11-07 | 2020-02-28 | 北京云迹科技有限公司 | Abnormality detection method, abnormality detection device, electronic apparatus, and computer-readable storage medium |
| CN111035096A (en) * | 2020-01-09 | 2020-04-21 | 郑州铁路职业技术学院 | Engineering constructor fatigue detection system based on safety helmet |
| CN111243233A (en) * | 2018-11-28 | 2020-06-05 | 现代自动车株式会社 | Apparatus, system and method for managing fatigue driving |
| CN112668393A (en) * | 2020-11-30 | 2021-04-16 | 海纳致远数字科技(上海)有限公司 | Fatigue degree detection device and method based on face recognition and key point detection |
| CN112950900A (en) * | 2021-01-27 | 2021-06-11 | 吉林云帆智能工程有限公司 | Driver behavior monitoring and detecting method |
| CN113536900A (en) * | 2021-05-31 | 2021-10-22 | 浙江大华技术股份有限公司 | Method and device for evaluating quality of face image and computer readable storage medium |
| DE102021111465A1 (en) | 2021-05-04 | 2022-11-10 | Bayerische Motoren Werke Aktiengesellschaft | DETERMINING A PERSON'S FATIGUE BASED ON AN ANALYSIS OF AN EYE BLINK |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH0318351A (en) * | 1989-06-16 | 1991-01-25 | Omron Corp | Driving fatigue calculating device |
| CN1680779A (en) * | 2005-02-04 | 2005-10-12 | 江苏大学 | Fatigue monitoring method and device for driver |
| CN101090482A (en) * | 2006-06-13 | 2007-12-19 | 唐琎 | Driver fatigue monitoring system and method based on image process and information mixing technology |
| CN101540090A (en) * | 2009-04-14 | 2009-09-23 | 华南理工大学 | Driver fatigue monitoring device based on multivariate information fusion and monitoring method thereof |
| CN102622600A (en) * | 2012-02-02 | 2012-08-01 | 西南交通大学 | High-speed train driver alertness detecting method based on face image and eye movement analysis |
| CN104013414A (en) * | 2014-04-30 | 2014-09-03 | 南京车锐信息科技有限公司 | Driver fatigue detecting system based on smart mobile phone |
-
2014
- 2014-10-23 CN CN201410568886.XA patent/CN104269028B/en not_active Expired - Fee Related
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH0318351A (en) * | 1989-06-16 | 1991-01-25 | Omron Corp | Driving fatigue calculating device |
| CN1680779A (en) * | 2005-02-04 | 2005-10-12 | 江苏大学 | Fatigue monitoring method and device for driver |
| CN101090482A (en) * | 2006-06-13 | 2007-12-19 | 唐琎 | Driver fatigue monitoring system and method based on image process and information mixing technology |
| CN101540090A (en) * | 2009-04-14 | 2009-09-23 | 华南理工大学 | Driver fatigue monitoring device based on multivariate information fusion and monitoring method thereof |
| CN102622600A (en) * | 2012-02-02 | 2012-08-01 | 西南交通大学 | High-speed train driver alertness detecting method based on face image and eye movement analysis |
| CN104013414A (en) * | 2014-04-30 | 2014-09-03 | 南京车锐信息科技有限公司 | Driver fatigue detecting system based on smart mobile phone |
Non-Patent Citations (1)
| Title |
|---|
| 李晶等: "基于卡尔曼滤波的人脸跟踪算法", 《西安邮电学院学报》, vol. 15, no. 3, 31 March 2010 (2010-03-31), pages 101 - 104 * |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104573657B (en) * | 2015-01-09 | 2018-04-27 | 安徽清新互联信息科技有限公司 | It is a kind of that detection method is driven based on the blind of feature of bowing |
| CN104573657A (en) * | 2015-01-09 | 2015-04-29 | 安徽清新互联信息科技有限公司 | Blind driving detection method based on head lowing characteristics |
| CN106295474A (en) * | 2015-05-28 | 2017-01-04 | 交通运输部水运科学研究院 | The fatigue detection method of deck officer, system and server |
| CN105160913A (en) * | 2015-08-17 | 2015-12-16 | 上海斐讯数据通信技术有限公司 | Method and apparatus for standardizing driving behaviors of drivers |
| CN105160913B (en) * | 2015-08-17 | 2018-02-06 | 上海斐讯数据通信技术有限公司 | A kind of method and device of specification driver driving behavior |
| CN106203394B (en) * | 2016-07-26 | 2019-04-26 | 浙江捷尚视觉科技股份有限公司 | Fatigue driving safety monitoring method based on human eye state detection |
| CN106203394A (en) * | 2016-07-26 | 2016-12-07 | 浙江捷尚视觉科技股份有限公司 | Fatigue driving safety monitoring method based on human eye state detection |
| CN106447828A (en) * | 2016-10-10 | 2017-02-22 | 深圳市驰勇科技有限公司 | Automobile data recorder with iris recognition function |
| CN108229359A (en) * | 2017-12-26 | 2018-06-29 | 大唐软件技术股份有限公司 | A kind of face image processing process and device |
| CN108446600A (en) * | 2018-02-27 | 2018-08-24 | 上海汽车集团股份有限公司 | A kind of vehicle driver's fatigue monitoring early warning system and method |
| CN111243233A (en) * | 2018-11-28 | 2020-06-05 | 现代自动车株式会社 | Apparatus, system and method for managing fatigue driving |
| CN111243233B (en) * | 2018-11-28 | 2022-11-29 | 现代自动车株式会社 | Apparatus, system and method for managing fatigue driving |
| CN110852247A (en) * | 2019-11-07 | 2020-02-28 | 北京云迹科技有限公司 | Abnormality detection method, abnormality detection device, electronic apparatus, and computer-readable storage medium |
| CN111035096A (en) * | 2020-01-09 | 2020-04-21 | 郑州铁路职业技术学院 | Engineering constructor fatigue detection system based on safety helmet |
| CN112668393A (en) * | 2020-11-30 | 2021-04-16 | 海纳致远数字科技(上海)有限公司 | Fatigue degree detection device and method based on face recognition and key point detection |
| CN112950900A (en) * | 2021-01-27 | 2021-06-11 | 吉林云帆智能工程有限公司 | Driver behavior monitoring and detecting method |
| DE102021111465A1 (en) | 2021-05-04 | 2022-11-10 | Bayerische Motoren Werke Aktiengesellschaft | DETERMINING A PERSON'S FATIGUE BASED ON AN ANALYSIS OF AN EYE BLINK |
| WO2022233527A1 (en) | 2021-05-04 | 2022-11-10 | Bayerische Motoren Werke Aktiengesellschaft | Determining the drowsiness of a person on the basis of a blinking analysis |
| CN113536900A (en) * | 2021-05-31 | 2021-10-22 | 浙江大华技术股份有限公司 | Method and device for evaluating quality of face image and computer readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104269028B (en) | 2017-02-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104269028A (en) | Fatigue driving detection method and system | |
| Ramzan et al. | A survey on state-of-the-art drowsiness detection techniques | |
| Craye et al. | Driver distraction detection and recognition using RGB-D sensor | |
| CN102289660B (en) | Method for detecting illegal driving behavior based on hand gesture tracking | |
| Mbouna et al. | Visual analysis of eye state and head pose for driver alertness monitoring | |
| CN105286802B (en) | Driver Fatigue Detection based on video information | |
| CN102054163B (en) | Method for testing driver fatigue based on monocular vision | |
| CN104021370B (en) | The driver status monitoring method and system of a kind of view-based access control model information fusion | |
| CN104794855A (en) | Driver's attention comprehensive assessment system | |
| CN105096528A (en) | Fatigue driving detection method and system | |
| Tadesse et al. | Driver drowsiness detection through HMM based dynamic modeling | |
| CN105303830A (en) | Driving behavior analysis system and analysis method | |
| CN104794856A (en) | Driver's attention comprehensive evaluation system | |
| CN104129389A (en) | Method for effectively judging and recognizing vehicle travelling conditions and device thereof | |
| CN104616438A (en) | Yawning action detection method for detecting fatigue driving | |
| Celona et al. | A multi-task CNN framework for driver face monitoring | |
| Qiao et al. | A smartphone-based driver fatigue detection using fusion of multiple real-time facial features | |
| Sun et al. | An integrated solution for lane level irregular driving detection on highways | |
| Azim et al. | Automatic fatigue detection of drivers through pupil detection and yawning analysis | |
| Pech et al. | Head tracking based glance area estimation for driver behaviour modelling during lane change execution | |
| Senaratne et al. | Comparing two video-based techniques for driver fatigue detection: classification versus optical flow approach | |
| Bergasa et al. | Visual monitoring of driver inattention | |
| CN116279561A (en) | Early warning method, device and equipment for fatigue driving of vehicle and readable storage medium | |
| Oh et al. | Intervehicle safety warning information system for unsafe driving events: Methodology and prototypical implementation | |
| Joseph et al. | Real time drowsiness detection using Viola jones & KLT |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170201 Termination date: 20171023 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |