CN109740477A - Study in Driver Fatigue State Surveillance System and its fatigue detection method - Google Patents
Study in Driver Fatigue State Surveillance System and its fatigue detection method Download PDFInfo
- Publication number
- CN109740477A CN109740477A CN201811596591.8A CN201811596591A CN109740477A CN 109740477 A CN109740477 A CN 109740477A CN 201811596591 A CN201811596591 A CN 201811596591A CN 109740477 A CN109740477 A CN 109740477A
- Authority
- CN
- China
- Prior art keywords
- fatigue
- driver
- facial
- image
- driver fatigue
- 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
Landscapes
- Image Analysis (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention discloses a kind of Study in Driver Fatigue State Surveillance System, including image procossing is preset format by image pre-processing module；Face and face feature point detection module are using the first convolutional neural networks detection acquisition facial area and face feature point；Facial area adjustment normalization module forms facial positions frame according to human face characteristic point, facial area is cut using facial positions frame, by the facial area normalized after cutting；Facial feature extraction network module uses the confidence level for obtaining facial characteristics in the image of the second convolutional neural networks after normalization and obtaining each facial characteristics；Degree of fatigue and fatigue state judgment module obtain present image degree of fatigue according to tired judgment rule, obtain driver fatigue state using PERCLOS algorithm according to preceding image degree of fatigue.The invention also discloses a kind of Driver Fatigue Detections.The present invention can accurately obtain the facial information of driver, and carry out fusion decision accurate judgement driver fatigue state.
Description
Technical field
The present invention relates to automotive fields, examine more particularly to a kind of driver fatigue merged based on deep learning and information
Examining system.The invention further relates to a kind of fatigue detection methods merged based on deep learning and information.
Background technique
According to both domestic and external studies have shown that fatigue driving is that one of the three big reasons of major accident occur: fatigue driving is drawn
The traffic accident accounting of hair is up to 40% or more, secondly to drive when intoxicated, vehicle trouble, traffic violation etc..Driving fatigue
Refer to the driver's physiological function generated for various reasons or mental function when driving vehicle imbalance (i.e. in physiology or
It is psychological to generate fatigue), so that driver is made to weaken to the sensing capability of surrounding enviroment and decline to the maneuvering capability of vehicle,
Deviate normal driving behavior.Therefore, the state for how being actively directed to driver is monitored, and effectively provides driving fatigue
Early warning is a traffic safety problem urgently to be resolved.
Now to the real-time fatigue detecting technology of driver, the main equipment for using contact detects the physiology of driver
Signal, its advantage is that accurately, but it is not friendly enough to driver's driver behavior, or judged using information of vehicles, advantage
It is that driver will not be impacted, but accuracy is not high, lag is serious.Therefore, it is full-featured to be badly in need of one kind, it is safe and efficient
To driver fatigue detection technology.
Summary of the invention
The technical problem to be solved in the present invention is to provide one kind based on deep learning and information fusion in illumination, driver's appearance
The complex conditions such as state variation can accurately obtain the facial information of driver, judge the driver fatigue of driver's fatigue degree
Detection system.
The present invention also provides one kind based on deep learning and information fusion in complex conditions such as illumination, driver gestures variations
The facial information that can accurately obtain driver down, judges the Driver Fatigue Detection of driver's fatigue degree.
In order to solve the above technical problems, Study in Driver Fatigue State Surveillance System provided by the invention, comprising: image preprocessing mould
Block, face and face feature point detection module, facial area adjustment normalization module, facial feature extraction network module and tired
Labor degree judgment module；
Image procossing is preset format by image pre-processing module；
Face and face feature point detection module are detected using the first convolutional neural networks and obtain facial area and face
Characteristic point；
Facial area adjustment normalization module, is formed facial positions frame according to human face characteristic point, is cut out using facial positions frame
Facial area is cut, by the facial area normalized after cutting；
Facial feature extraction network module, it is special using face is obtained in the image of the second convolutional neural networks after normalization
Sign, and obtain the confidence level of each facial characteristics；
Degree of fatigue and fatigue state judgment module obtain present image degree of fatigue according to tired judgment rule, according to
Preceding image degree of fatigue obtains driver fatigue state using PERCLOS algorithm.
It is further improved the Study in Driver Fatigue State Surveillance System, the preset format is single channel gray level image, image point
Resolution is greater than 640 × 480.
It is further improved the Study in Driver Fatigue State Surveillance System, the gray scale conversion formula is Gray=0.299R+
0.587G+0.114B。
It is further improved the Study in Driver Fatigue State Surveillance System, first convolutional neural networks are multi-stage cascade convolution minds
Through network MTCNN.
It is further improved the Study in Driver Fatigue State Surveillance System, the institute for exporting multi-stage cascade convolutional neural networks MTCNN
There is face feature point to be located in facial area and form facial positions frame, then facial regional image is cut by facial positions frame.
It is further improved the Study in Driver Fatigue State Surveillance System, it is 96 × 96 that the normalization, which is by image conversion resolution,
Image.
It is further improved the Study in Driver Fatigue State Surveillance System, second convolutional neural networks use MobileNetV2
Network structure, loss function are made of Classification Loss and recurrence loss, and multi-tag Classification Loss is using Sigmoid cross entropy
Loss function, formula are as follows:
Loss is returned using L2 loss function, formula are as follows: L2 (x)=∑ (Yi-p(xi))2, the study of training process
Rate is set as 0.05, maximum number of iterations 100000.
It is further improved the Study in Driver Fatigue State Surveillance System, the facial characteristics includes: that right and left eyes open closed state, mouth
Closed state and driver head's pose.
It is further improved the Study in Driver Fatigue State Surveillance System, the fatigue diagnostic rule includes: to close one's eyes according to right and left eyes
The Pitch value of state confidence level, state of opening one's mouth confidence level and head pose obtains driver using Weighted Fusion algorithm and is worked as
Degree of fatigue confidence level under previous frame, it is fatigue that degree of fatigue confidence level, which is greater than the then present frame of first threshold, under previous frame.
It is further improved the Study in Driver Fatigue State Surveillance System, the Weighted Fusion algorithmic formula is as follows；
Confidence level of bowing isPitch is the angle value bowed.
It is further improved the Study in Driver Fatigue State Surveillance System, the first threshold is 0.5-0.9.
It is further improved the Study in Driver Fatigue State Surveillance System, the PERCLOS algorithm is driver's eye in the unit time
Ratio shared by eyeball closing time, calculation formula are as follows: PERCLOS=unit time eye closing frame number/unit time totalframes, when than
When value is greater than second threshold, driver is in a state of fatigue.
It is further improved the Study in Driver Fatigue State Surveillance System, the unit time is 30 seconds~180 seconds.
It is further improved the Study in Driver Fatigue State Surveillance System, the second threshold is 0.3-0.5.
The present invention provides a kind of Driver Fatigue Detection, comprising the following steps:
1) driver's face-image is shot；
It 2) is preset format by image procossing；
3) facial area and face feature point are obtained using the detection of the first convolutional neural networks；
4) facial positions frame is formed according to human face characteristic point, facial area is cut using facial positions frame, after cutting
Facial area normalized；
5) it uses in the image of the second convolutional neural networks after normalization and obtains facial characteristics, and obtain each facial characteristics
Confidence level；
6) present image degree of fatigue is obtained according to tired judgment rule, PERCLOS is utilized according to preceding image degree of fatigue
Algorithm obtains driver fatigue state.
It is further improved the Driver Fatigue Detection, implementation steps 2) when, preset format is single channel gray scale
Image, image resolution ratio are greater than 640 × 480.
It is further improved the Driver Fatigue Detection, the gray scale conversion formula is Gray=0.299R+
0.587G+0.114B。
It is further improved the Driver Fatigue Detection, implementation steps 3) when, first convolutional neural networks
It is multi-stage cascade convolutional neural networks MTCNN.
It is further improved the Driver Fatigue Detection, implementation steps 4) when, make multi-stage cascade convolutional Neural net
All face feature points of network MTCNN output, which are located in facial area, forms facial positions frame, then passes through facial positions frame opposite
Portion's regional image is cut.
It is further improved the Driver Fatigue Detection, implementation steps 4) when, the normalization is to turn image
Change the image that resolution ratio is 96 × 96.
It is further improved the Driver Fatigue Detection, implementation steps 5) when, the second convolutional neural networks use
MobileNetV2 network structure, loss function by Classification Loss and return loss constitute, multi-tag Classification Loss using
Sigmoid cross entropy loss function, formula are as follows:
Loss is returned using L2 loss function, formula are as follows: L2 (x)=∑ (Yi-p(xi))2, the study of training process
Rate is set as 0.05, maximum number of iterations 100000.
It is further improved the Driver Fatigue Detection, implementation steps 5) when, the facial characteristics includes: left and right
Eye opens closed state, closed state and the driver head's pose of mouth.
It is further improved the Driver Fatigue Detection, implementation steps 6) when, the fatigue diagnostic rule includes:
According to right and left eyes closed-eye state confidence level, the Pitch value of state of opening one's mouth confidence level and head pose utilizes Weighted Fusion algorithm
The degree of fatigue confidence level under driver institute present frame is obtained, degree of fatigue confidence level is greater than the then current of first threshold under previous frame
Frame is fatigue.
It is further improved the Driver Fatigue Detection, the Weighted Fusion algorithmic formula is as follows；
Confidence level of bowing isPitch is the angle value bowed.
It is further improved the Driver Fatigue Detection, the first threshold is 0.5-0.9.
It is further improved the Driver Fatigue Detection, the PERCLOS algorithm is driver in the unit time
Ratio shared by the eyes closed time, calculation formula are as follows: PERCLOS=unit time eye closing frame number/unit time totalframes, when
When ratio is greater than second threshold, driver is in a state of fatigue.
It is further improved the Driver Fatigue Detection, the unit time is 30 seconds~180 seconds.
It is further improved the Driver Fatigue Detection, the second threshold is 0.3-0.5.
The present invention can be reduced the traffic accident caused by driving fatigue, ensure that driver safety drives.The present invention is based on depths
The driver fatigue state detection system and method for degree study and information fusion are realized even if multiple in illumination, attitudes vibration etc.
The facial information of driver can be accurately obtained under the conditions of miscellaneous, and carries out fusion decision accurate judgement driver status and early warning,
Can effective preventing driver fatigue driving, reduce the traffic accident caused by driving fatigue.
The present invention at least has following technical effect that
(1) present invention uses the contactless mode of view-based access control model information, can be effective using depth convolutional neural networks
Overcome the variation bring interference such as various illumination, pose, realize and detection differentiated to the real-time and precise of driving fatigue.
(2) image is carried out into processing using single neural network, while obtains facial area and face points region, subtracted
Few redundant computation, effectively promotion calculating speed, and optimize face detection module using characteristic point, improve the precision of face detection.
(3) facial area is calculated using single neural network combination multi-tag Classification Loss and recurrence loss, together
When Pitch (the pitching of right and left eyes closed-eye state confidence value, mouth open configuration confidence value and head pose is calculated
Angle), Yaw (yaw angle) and Roll (roll angle) angle value, can be with the acquisition human face data of efficiently and accurately, calculating speed is fast.
(4) using the feature fusion of eyes, mouth and head pose, the facial information of driver is made full use of,
The more accurate effective differentiation driver's state in which of system.
Detailed description of the invention
Present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments:
Fig. 1 is the principle of the present invention schematic diagram.
Fig. 2 is that the present invention is facial and face feature point detection module is opened by neural network acquisition right and left eyes and closed, mouth
It closes and head pose value schematic illustration.
Fig. 3 is facial feature extraction network module of the present invention using MobileNetV2 modular concept schematic diagram.
Fig. 4 is that present invention fusion judges schematic illustration.
Specific embodiment
Refering to what is shown in Fig. 1, Study in Driver Fatigue State Surveillance System provided by the invention, comprising: image pre-processing module, face with
And face feature point detection module, facial area adjustment normalization module, facial feature extraction network module and degree of fatigue are sentenced
Disconnected module；
Image pre-processing module, converts the image into single channel gray level image, and image resolution ratio is greater than 640 × 480；
Gray scale conversion formula is Gray=0.299R+0.587G+0.114B；
Refering to what is shown in Fig. 2, face and face feature point detection module, using multi-stage cascade convolutional neural networks MTCNN
Detection obtains facial area and face feature point；All face feature points for exporting multi-stage cascade convolutional neural networks MTCNN
Facial positions frame is formed in facial area, then facial regional image is cut by facial positions frame；Facial characteristics includes:
Right and left eyes open closed state, closed state and the driver head's pose of mouth；
Facial area adjustment normalization module, is formed facial positions frame according to human face characteristic point, is cut out using facial positions frame
Facial area is cut, the image for being 96 × 96 by the facial area conversion resolution after cutting；
Refering to what is shown in Fig. 3, facial feature extraction network module, using the image of the second convolutional neural networks after normalization
Middle acquisition facial characteristics, and obtain the confidence level of each facial characteristics；
Second convolutional neural networks use MobileNetV2 network structure, and loss function is by Classification Loss and recurrence
Loss is constituted, and multi-tag Classification Loss is using Sigmoid cross entropy loss function, formula are as follows:
Loss is returned using L2 loss function, formula are as follows: L2 (x)=∑ (Yi-p(xi))2, the study of training process
Rate is set as 0.05, maximum number of iterations 100000；
Degree of fatigue and fatigue state judgment module obtain present image degree of fatigue according to tired judgment rule, according to
Preceding image degree of fatigue obtains driver fatigue state using PERCLOS algorithm；
The fatigue diagnostic rule includes: according to right and left eyes closed-eye state confidence level, state of opening one's mouth confidence level and head
The Pitch value of pose obtains the degree of fatigue confidence level under driver institute present frame, fatigue under previous frame using Weighted Fusion algorithm
The then present frame that degree confidence level is greater than first threshold is fatigue.
The Weighted Fusion algorithmic formula is as follows；
Confidence level of bowing isPitch is the angle value bowed；For example, bowing 17 degree；
The PERCLOS algorithm is ratio shared by driver's eyes closing time, calculation formula in the unit time are as follows:
PERCLOS=unit time eye closing frame number/unit time totalframes, when ratio is greater than second threshold, driver is in
Fatigue state issues giving fatigue pre-warning.
Wherein, the first threshold be 0.5-0.9, preferably 0.6.
The unit time is 30 seconds~180 seconds, preferably 90 seconds.
The second threshold be 0.3-0.5, preferably 0.4.
The present invention provides a kind of Driver Fatigue Detection, comprising the following steps:
1) driver's face-image is shot using camera/camcorder；
2) single channel gray level image is converted the image into, image resolution ratio is greater than 640 × 480；Gray scale conversion formula is,
Gray=0.299R+0.587G+0.114B；
3) facial area and face feature point are obtained using multi-stage cascade convolutional neural networks MTCNN detection；
4) all face feature points for exporting multi-stage cascade convolutional neural networks MTCNN are located at forming face in facial area
Portion position frame, then facial regional image is cut by facial positions frame, it is 96 by the facial area conversion resolution after cutting
× 96 image.
Facial characteristics includes: that right and left eyes open closed state, closed state and the driver head's pose of mouth；
5) it uses in the image of the second convolutional neural networks after normalization and obtains facial characteristics, and obtain each facial characteristics
Confidence level；
Second convolutional neural networks use MobileNetV2 network structure, and loss function is by Classification Loss and recurrence
Loss is constituted, and multi-tag Classification Loss is using Sigmoid cross entropy loss function, formula are as follows:
Loss is returned using L2 loss function, formula are as follows: L2 (x)=∑ (Yi-p(xi))2, the study of training process
Rate is set as 0.05, maximum number of iterations 100000.
6) present image degree of fatigue is obtained according to tired judgment rule, PERCLOS is utilized according to preceding image degree of fatigue
Algorithm obtains driver fatigue state；
The fatigue diagnostic rule includes: according to right and left eyes closed-eye state confidence level, state of opening one's mouth confidence level and head
The Pitch value of pose obtains the degree of fatigue confidence level under driver institute present frame, fatigue under previous frame using Weighted Fusion algorithm
The then present frame that degree confidence level is greater than first threshold is fatigue.
The Weighted Fusion algorithmic formula is as follows；
Confidence level of bowing isPitch is the angle value bowed；For example, bowing 17 degree；
The PERCLOS algorithm is ratio shared by driver's eyes closing time, calculation formula in the unit time are as follows:
PERCLOS=unit time eye closing frame number/unit time totalframes, when ratio is greater than second threshold, driver is in
Fatigue state issues giving fatigue pre-warning.
Wherein, the first threshold be 0.5-0.9, preferably 0.6.
The unit time is 30 seconds~180 seconds, preferably 90 seconds.
The second threshold be 0.3-0.5, preferably 0.4.
Above by specific embodiment and embodiment, invention is explained in detail, but these are not composition pair
Limitation of the invention.Without departing from the principles of the present invention, the technology driver of this field can also make it is many deformation and
It improves, these also should be regarded as protection scope of the present invention.
Claims (28)
1. a kind of Study in Driver Fatigue State Surveillance System characterized by comprising image pre-processing module, face and facial characteristics
Point detection module, facial area adjustment normalization module, facial feature extraction network module and degree of fatigue judgment module；
Image procossing is preset format by image pre-processing module；
Face and face feature point detection module are detected using the first convolutional neural networks and obtain facial area and facial characteristics
Point；
Facial area adjustment normalization module, forms facial positions frame according to human face characteristic point, cuts face using facial positions frame
Portion region, by the facial area normalized after cutting；
Facial feature extraction network module, using obtaining facial characteristics in the image of the second convolutional neural networks after normalization,
And obtain the confidence level of each facial characteristics；
Degree of fatigue and fatigue state judgment module obtain present image degree of fatigue according to tired judgment rule, according to preceding figure
As degree of fatigue obtains driver fatigue state using PERCLOS algorithm.
2. Study in Driver Fatigue State Surveillance System as described in claim 1, it is characterised in that: the preset format is single channel grayscale image
Picture, image resolution ratio are greater than 640 × 480.
3. Study in Driver Fatigue State Surveillance System as claimed in claim 2, it is characterised in that: the gray scale conversion formula is Gray=
0.299R+0.587G+0.114B。
4. Study in Driver Fatigue State Surveillance System as described in claim 1, it is characterised in that: first convolutional neural networks are multistage
Concatenated convolutional neural network MTCNN.
5. Study in Driver Fatigue State Surveillance System as claimed in claim 4, it is characterised in that: make multi-stage cascade convolutional neural networks
All face feature points of MTCNN output, which are located in facial area, forms facial positions frame, then by facial positions frame to face
Regional image is cut.
6. Study in Driver Fatigue State Surveillance System as described in claim 1, it is characterised in that: the normalization is to convert image to differentiate
The image that rate is 96 × 96.
7. Study in Driver Fatigue State Surveillance System as described in claim 1, it is characterised in that: second convolutional neural networks use
MobileNetV2 network structure, loss function by Classification Loss and return loss constitute, multi-tag Classification Loss using
Sigmoid cross entropy loss function, formula are as follows:
Loss is returned using L2 loss function, formula are as follows: L2 (x)=∑ (Yi-p(xi))2, the learning rate of training process sets
It is set to 0.05, maximum number of iterations 100000.
8. Study in Driver Fatigue State Surveillance System as described in claim 1, it is characterised in that: the facial characteristics includes: that right and left eyes are opened
Closed state, closed state and the driver head's pose of mouth.
9. Study in Driver Fatigue State Surveillance System as described in claim 1, it is characterised in that: it is described fatigue diagnostic rule include:
According to right and left eyes closed-eye state confidence level, the Pitch value of state of opening one's mouth confidence level and head pose utilizes Weighted Fusion
Algorithm obtains the degree of fatigue confidence level under driver institute present frame, and degree of fatigue confidence level is greater than first threshold then under previous frame
Present frame is fatigue.
10. Study in Driver Fatigue State Surveillance System as claimed in claim 9, it is characterised in that: the Weighted Fusion algorithmic formula is as follows；
Confidence level of bowing isPitch is the angle value bowed.
11. Study in Driver Fatigue State Surveillance System as claimed in claim 9, it is characterised in that: the first threshold is 0.5-0.9.
12. Study in Driver Fatigue State Surveillance System as described in claim 1, it is characterised in that: the PERCLOS algorithm is the unit time
Ratio shared by interior driver's eyes closing time, calculation formula are as follows: PERCLOS=unit time eye closing frame number/unit time
Totalframes, when ratio is greater than second threshold, driver is in a state of fatigue.
13. Study in Driver Fatigue State Surveillance System as claimed in claim 12, it is characterised in that: the unit time is 30 seconds~180
Second.
14. Study in Driver Fatigue State Surveillance System as claimed in claim 12, it is characterised in that: the second threshold is 0.3-0.5.
15. a kind of Driver Fatigue Detection, which comprises the following steps:
1) driver's face-image is shot；
It 2) is preset format by image procossing；
3) facial area and face feature point are obtained using the detection of the first convolutional neural networks；
4) facial positions frame is formed according to human face characteristic point, facial area is cut using facial positions frame, by the face after cutting
Region normalized；
5) it uses in the image of the second convolutional neural networks after normalization and obtains facial characteristics, and obtain setting for each facial characteristics
Reliability；
6) present image degree of fatigue is obtained according to tired judgment rule, PERCLOS algorithm is utilized according to preceding image degree of fatigue
Obtain driver fatigue state.
16. Driver Fatigue Detection as claimed in claim 15, it is characterised in that: implementation steps 2) when, preset format
For single channel gray level image, image resolution ratio is greater than 640 × 480.
17. Driver Fatigue Detection as claimed in claim 16, it is characterised in that: the gray scale conversion formula is,
Gray=0.299R+0.587G+0.114B.
18. Driver Fatigue Detection as claimed in claim 15, it is characterised in that: implementation steps 3) when, described first
Convolutional neural networks are multi-stage cascade convolutional neural networks MTCNN.
19. Driver Fatigue Detection as claimed in claim 15, it is characterised in that: implementation steps 4) when, make multistage grade
All face feature points of connection convolutional neural networks MTCNN output, which are located in facial area, forms facial positions frame, then passes through face
Portion position frame cuts facial regional image.
20. Driver Fatigue Detection as claimed in claim 15, it is characterised in that: implementation steps 4) when, the normalizing
Change is the image for being 96 × 96 by image conversion resolution.
21. Driver Fatigue Detection as claimed in claim 15, it is characterised in that: implementation steps 5) when, the second convolution
Neural network uses MobileNetV2 network structure, and loss function is made of Classification Loss and recurrence loss, multi-tag classification damage
It loses using Sigmoid cross entropy loss function, formula are as follows:
Loss is returned using L2 loss function, formula are as follows: L2 (x)=∑ (Yi-p(xi))2, the learning rate of training process sets
It is set to 0.05, maximum number of iterations 100000.
22. Driver Fatigue Detection as claimed in claim 15, it is characterised in that: implementation steps 5) when, the face
Feature includes: that right and left eyes open closed state, closed state and the driver head's pose of mouth.
23. Driver Fatigue Detection as claimed in claim 15, it is characterised in that: implementation steps 6) when, the fatigue
Diagnostic rule includes: the Pitch value utilization according to right and left eyes closed-eye state confidence level, state of opening one's mouth confidence level and head pose
Weighted Fusion algorithm obtains the degree of fatigue confidence level under driver institute present frame, and degree of fatigue confidence level is greater than first under previous frame
The then present frame of threshold value is fatigue.
24. Driver Fatigue Detection as claimed in claim 23, it is characterised in that: the Weighted Fusion algorithmic formula is such as
Under；
Confidence level of bowing isPitch is the angle value bowed.
25. Driver Fatigue Detection as claimed in claim 23, it is characterised in that: the first threshold is 0.5-0.9.
26. Driver Fatigue Detection as claimed in claim 23, it is characterised in that: the PERCLOS algorithm is unit
Ratio shared by driver's eyes closing time, calculation formula in time are as follows: PERCLOS=unit time eye closing frame number/unit
Time totalframes, when ratio is greater than second threshold, driver is in a state of fatigue.
27. Driver Fatigue Detection as claimed in claim 26, it is characterised in that: the unit time be 30 seconds~
180 seconds.
28. Driver Fatigue Detection as claimed in claim 26, it is characterised in that: the second threshold is 0.3-0.5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811596591.8A CN109740477B (en) | 2018-12-26 | 2018-12-26 | Driver fatigue detection system and fatigue detection method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811596591.8A CN109740477B (en) | 2018-12-26 | 2018-12-26 | Driver fatigue detection system and fatigue detection method thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109740477A true CN109740477A (en) | 2019-05-10 |
| CN109740477B CN109740477B (en) | 2022-12-06 |
Family
ID=66361206
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811596591.8A Active CN109740477B (en) | 2018-12-26 | 2018-12-26 | Driver fatigue detection system and fatigue detection method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109740477B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110232327A (en) * | 2019-05-21 | 2019-09-13 | 浙江师范大学 | A kind of driving fatigue detection method based on trapezoidal concatenated convolutional neural network |
| CN110532887A (en) * | 2019-07-31 | 2019-12-03 | 郑州大学 | A kind of method for detecting fatigue driving and system based on facial characteristics fusion |
| CN110678873A (en) * | 2019-07-30 | 2020-01-10 | 珠海全志科技股份有限公司 | Attention detection method based on cascade neural network, computer device and computer readable storage medium |
| CN110674701A (en) * | 2019-09-02 | 2020-01-10 | 东南大学 | Driver fatigue state rapid detection method based on deep learning |
| CN113343940A (en) * | 2021-07-19 | 2021-09-03 | 湖南科技大学 | Method for detecting fatigue state of operator on duty based on facial multi-feature point recognition |
| WO2022001091A1 (en) * | 2020-06-29 | 2022-01-06 | 北京百度网讯科技有限公司 | Dangerous driving behavior recognition method and apparatus, and electronic device and storage medium |
| CN114475620A (en) * | 2022-01-26 | 2022-05-13 | 南京科融数据系统股份有限公司 | Driver verification method and system for money box escort system |
| CN116912808A (en) * | 2023-09-14 | 2023-10-20 | 四川公路桥梁建设集团有限公司 | Bridge girder erection machine control method, electronic equipment and computer readable medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106361456A (en) * | 2015-07-23 | 2017-02-01 | 郭宏博 | Intelligent toothbrush tooth brushing way detection method and intelligent toothbrush tooth brushing way detection system |
| CN106781282A (en) * | 2016-12-29 | 2017-05-31 | 天津中科智能识别产业技术研究院有限公司 | A kind of intelligent travelling crane driver fatigue early warning system |
| US20180039751A1 (en) * | 2015-05-19 | 2018-02-08 | Sony Corporation | Information processing apparatus, information processing method, and program |
| CN108363968A (en) * | 2018-01-31 | 2018-08-03 | 上海瀚所信息技术有限公司 | A kind of tired driver driving monitoring system and method based on key point extraction |
| CN108703444A (en) * | 2018-06-15 | 2018-10-26 | 吉林大学 | The Intelligent bracelet and method of occupant and driver for identification |
| CN108875642A (en) * | 2018-06-21 | 2018-11-23 | 长安大学 | A kind of method of the driver fatigue detection of multi-index amalgamation |
-
2018
- 2018-12-26 CN CN201811596591.8A patent/CN109740477B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180039751A1 (en) * | 2015-05-19 | 2018-02-08 | Sony Corporation | Information processing apparatus, information processing method, and program |
| CN106361456A (en) * | 2015-07-23 | 2017-02-01 | 郭宏博 | Intelligent toothbrush tooth brushing way detection method and intelligent toothbrush tooth brushing way detection system |
| CN106781282A (en) * | 2016-12-29 | 2017-05-31 | 天津中科智能识别产业技术研究院有限公司 | A kind of intelligent travelling crane driver fatigue early warning system |
| CN108363968A (en) * | 2018-01-31 | 2018-08-03 | 上海瀚所信息技术有限公司 | A kind of tired driver driving monitoring system and method based on key point extraction |
| CN108703444A (en) * | 2018-06-15 | 2018-10-26 | 吉林大学 | The Intelligent bracelet and method of occupant and driver for identification |
| CN108875642A (en) * | 2018-06-21 | 2018-11-23 | 长安大学 | A kind of method of the driver fatigue detection of multi-index amalgamation |
Non-Patent Citations (1)
| Title |
|---|
| 杨欢等: "基于逆投影修正和眼睛凝视修正的列车驾驶员疲劳检测方法", 《铁道学报》 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110232327B (en) * | 2019-05-21 | 2023-04-21 | 浙江师范大学 | Driving fatigue detection method based on trapezoid cascade convolution neural network |
| CN110232327A (en) * | 2019-05-21 | 2019-09-13 | 浙江师范大学 | A kind of driving fatigue detection method based on trapezoidal concatenated convolutional neural network |
| CN110678873A (en) * | 2019-07-30 | 2020-01-10 | 珠海全志科技股份有限公司 | Attention detection method based on cascade neural network, computer device and computer readable storage medium |
| WO2021016873A1 (en) * | 2019-07-30 | 2021-02-04 | 珠海全志科技股份有限公司 | Cascaded neural network-based attention detection method, computer device, and computer-readable storage medium |
| CN110532887A (en) * | 2019-07-31 | 2019-12-03 | 郑州大学 | A kind of method for detecting fatigue driving and system based on facial characteristics fusion |
| CN110674701A (en) * | 2019-09-02 | 2020-01-10 | 东南大学 | Driver fatigue state rapid detection method based on deep learning |
| WO2022001091A1 (en) * | 2020-06-29 | 2022-01-06 | 北京百度网讯科技有限公司 | Dangerous driving behavior recognition method and apparatus, and electronic device and storage medium |
| CN113343940B (en) * | 2021-07-19 | 2022-05-10 | 湖南科技大学 | Fatigue state detection method for person on duty based on face multi-feature point identification |
| CN113343940A (en) * | 2021-07-19 | 2021-09-03 | 湖南科技大学 | Method for detecting fatigue state of operator on duty based on facial multi-feature point recognition |
| CN114475620A (en) * | 2022-01-26 | 2022-05-13 | 南京科融数据系统股份有限公司 | Driver verification method and system for money box escort system |
| CN114475620B (en) * | 2022-01-26 | 2024-03-12 | 南京科融数据系统股份有限公司 | Driver verification method and system for money box escort system |
| CN116912808A (en) * | 2023-09-14 | 2023-10-20 | 四川公路桥梁建设集团有限公司 | Bridge girder erection machine control method, electronic equipment and computer readable medium |
| CN116912808B (en) * | 2023-09-14 | 2023-12-01 | 四川公路桥梁建设集团有限公司 | Bridge girder erection machine control method, electronic equipment and computer readable medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109740477B (en) | 2022-12-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109740477A (en) | Study in Driver Fatigue State Surveillance System and its fatigue detection method | |
| CN108053615B (en) | Method for detecting fatigue driving state of driver based on micro-expression | |
| Ji et al. | Fatigue state detection based on multi-index fusion and state recognition network | |
| Tipprasert et al. | A method of driver’s eyes closure and yawning detection for drowsiness analysis by infrared camera | |
| CN103839379B (en) | Automobile and driver fatigue early warning detecting method and system for automobile | |
| CN110119676A (en) | A kind of Driver Fatigue Detection neural network based | |
| CN107491769A (en) | Method for detecting fatigue driving and system based on AdaBoost algorithms | |
| CN202257856U (en) | Driver fatigue-driving monitoring device | |
| CN112434611B (en) | Early fatigue detection method and system based on eye movement subtle features | |
| CN105719431A (en) | Fatigue driving detection system | |
| CN110728241A (en) | Driver fatigue detection method based on deep learning multi-feature fusion | |
| CN104013414A (en) | Driver fatigue detecting system based on smart mobile phone | |
| CN104637246A (en) | Driver multi-behavior early warning system and danger evaluation method | |
| CN111753674A (en) | Fatigue driving detection and identification method based on deep learning | |
| CN105117681A (en) | Multi-characteristic fatigue real-time detection method based on Android | |
| CN106919913A (en) | Method for detecting fatigue driving and device based on computer vision | |
| CN110781872A (en) | Driver fatigue grade recognition system with bimodal feature fusion | |
| CN109543577A (en) | A kind of fatigue driving detection method for early warning based on facial expression feature | |
| CN112528843A (en) | Motor vehicle driver fatigue detection method fusing facial features | |
| CN115937830A (en) | Special vehicle-oriented driver fatigue detection method | |
| Rani et al. | Development of an Automated Tool for Driver Drowsiness Detection | |
| Joseph et al. | Real time drowsiness detection using Viola jones & KLT | |
| Mašanović et al. | Driver monitoring using the in-vehicle camera | |
| Liu et al. | Design and implementation of multimodal fatigue detection system combining eye and yawn information | |
| CN116453094A (en) | Driver fatigue detection method based on multi-scale feature fusion convolutional neural network |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |