CN102013013A - Fatigue driving monitoring method - Google Patents
Fatigue driving monitoring method Download PDFInfo
- Publication number
- CN102013013A CN102013013A CN2010106025466A CN201010602546A CN102013013A CN 102013013 A CN102013013 A CN 102013013A CN 2010106025466 A CN2010106025466 A CN 2010106025466A CN 201010602546 A CN201010602546 A CN 201010602546A CN 102013013 A CN102013013 A CN 102013013A
- Authority
- CN
- China
- Prior art keywords
- image
- sorter
- driver
- monitoring method
- fatigue driving
- 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
Images
Landscapes
- Emergency Alarm Devices (AREA)
Abstract
The invention discloses a fatigue driving monitoring method, which comprises the following steps of: A, acquiring images by using a camera, and filtering the acquired images through a classifier in a monitor, wherein the image which passes through the classifier is an eye image; B, analyzing the filtered eye image and finding the maximum distance L1 between the upper and lower points of an eyeball in set time; C, monitoring the eyeball of a driver, calculating the difference of the distance L2 between the upper and lower points of the eyeball and the maximum distance L1 in the set time, and alarming if the difference exceeds a set range. The monitoring method has higher accuracy.
Description
Technical field
The present invention relates to a kind of fatigue driving monitoring method.
Background technology
In the world, driver's fatigue driving has become one of major reason that causes the traffic safety accident.Traffic hazard causes enormous economic loss and casualties to country, has increased factors leading to social instability.
A large amount of at present fatigue driving monitors that adopt mainly are contacts, monitor driver's fatigue state by the physiological signal that detects the driver, such as paste discharge electrode monitoring EEG signals or electrocardiosignal etc. at head, this monitoring mode is subjected to the interference of external environment easily, and the difference between the different drivers also can influence monitoring result, and it is not high to cause monitoring efficient.Contact with driver's health simultaneously and have influence on driver's behavior, real-time is not high, and the time-delay that causes reporting to the police, and has restricted its promotion and application.
Contactless technology for present employing, mostly be by camera collection driver's facial image information is handled, judge whether the driver tired feature occurs, and all be simple depend on the feature (as the Haar feature) that training (as the Adaboost training method) obtains and carry out Target Recognition to the technology of Flame Image Process, its monitoring efficient is not high, and the probability of flase drop is very big.
Summary of the invention
The objective of the invention is to overcome the defective of prior art, a kind of fatigue driving monitoring method is provided, the present invention overcomes the defective of prior art, and monitoring accuracy is higher.
Its technical scheme is as follows.
A kind of fatigue driving monitoring method, this method comprise the steps: that A, camera grasp image, and the image that is grasped is filtered through the sorter in the monitor, are eye image by the image behind the sorter; Eye image after B, the analysis and filter finds eyeball upper and lower ultimate range L1 in setting-up time at 2; C, driver's eyeball upper and lower distance L 2 and the difference between the ultimate range L1 in setting-up time are monitored and calculated to driver's eyeball at 2, if this difference surpasses the scope of setting, the processing of then reporting to the police.
The technical scheme of the further refinement of aforementioned techniques scheme can be.
Described sorter comprises first order sorter and second level sorter at least, in the step A, after camera grasps image, filters through first order sorter earlier, filters through second level sorter by the image behind the first order sorter again.
In the step A, camera filters through sorter after grasping image, is facial image by the image behind the sorter, grasps eye image again in facial image.
Also comprise at least three visual angle detecting devices, the detection angles scope difference of each visual angle detecting device, in the step A, the image that camera grasped is detected by each visual angle detecting device, if the image that is grasped is by corresponding visual angle detecting device, then the angular range that is detected by this visual angle detecting device is determined the angular range that driver face is departed from, if the angular range that driver face is departed from surpasses the angular range of setting, the processing of then reporting to the police.
In the step A, camera filters through sorter after grasping image, if this image is by sorter, the colour of skin information with colour of skin information in this image and setting compares again, if meet, then be judged to be facial image,, then be judged to be non-face image if do not meet.
In sum, advantage of the present invention is.
1, by analyzing in the eye image, upper and lower 2 distance of eyeball judges whether the driver is tired, as driver fatigue, its eyeball is blocked by eyelid, distance between upper and lower 2 of the eyeball can diminish, if the distance between upper and lower 2 of the eyeball is dwindled above certain limit the processing of then reporting to the police, avoid occurring traffic hazard, monitoring accuracy is higher.
2, after camera grasps head portrait, filter by multistage classifier, can comprise one or more features in each grade sorter, behind first order sorter, if image is filtered, assert that then this image is not a target image, if image is by first order sorter, then this image is again through second level sorter (even third level sorter, fourth stage sorter ...), be target image by the image behind all sorters, filter by multistage classifier, monitoring velocity is fast, the precision height.
3, in the process that grasps image, grasp face image earlier, reduce the scope again, in face image, grasp eye image, to improve the speed that grasps image.
4, further, in the process that the driver is monitored, also the angle that driver's head is departed from is monitored, with driver's head is 0 degree (benchmark angle) when the dead ahead, if driver's head departs from the benchmark angle and surpasses setting range, this moment, driver's sight line very likely departed from the dead ahead, can bring harm to driving safety, processings of reporting to the police this moment further improves the reliability and the security of monitoring.
Description of drawings
Fig. 1 is the constitutional diagram of driver's eyeball under normal condition.
Fig. 2 is the constitutional diagram of driver's eyeball under fatigue state.
Embodiment
Below in conjunction with accompanying drawing embodiments of the invention are elaborated.
A kind of fatigue driving monitoring method, this method comprise the steps: that A, camera grasp image, and the image that is grasped is filtered through the sorter in the monitor, are eye image by the image behind the sorter; Eye image after B, the analysis and filter, upper and lower 2 ultimate range L1(is as shown in Figure 1 in setting-up time to find eyeball); C, driver's eyeball is monitored and calculated to driver's eyeball, and upper and lower 2 distance L 2(is as shown in Figure 2 in setting-up time) and ultimate range L1 between difference, if this difference surpasses the scope of setting, the processing of then reporting to the police.
Wherein, described sorter comprises first order sorter, second level sorter and third level sorter, in the step A, after camera grasps image, filter through first order sorter earlier, filter through second level sorter, third level sorter again by the image behind the first order sorter, colour of skin information with colour of skin information in this image and setting compares again, if meet colour of skin information, then be judged to be facial image, and and then in facial image, grasp eye image, if do not meet colour of skin information, then be judged to be non-face image.
Also comprise six visual angle detecting devices, the detection angles scope difference of each visual angle detecting device, be respectively [0,30 °], [30 °, 60 °], [60 °, 90 °], [0 ,-30 °], [30 ° ,-60 °], [60 °,-90 °], in the step A, the image that camera grasped is detected by each visual angle detecting device, if the image that is grasped is by corresponding visual angle detecting device, then the angular range that is detected by this visual angle detecting device is determined the angular range that driver face is departed from, if the angular range that driver face is departed from surpasses the angular range of setting (in the present embodiment, setting range is 30 °), the processing of then reporting to the police.
Below in the present embodiment, when carrying out image-capture, describe based on the gentleboost training method and the sorter of MBLBP feature.
People's face and human eye are because its special structure causes there are differences between itself and the general scenery picture, and this species diversity is that face's (eye) architectural feature owing to the people determines.We can come the composition and classification device to distinguish people's face (human eye) and non-face (non-human eye) according to the feature of pixel on its structure.
LBP is a kind of uncorrelated operator of the description image texture characteristic based on gray scale, and it is by a bit coming the local grain feature of token image arbitrarily with gray-scale value magnitude relationship of point around it to image.The LBP algorithm is commonly defined as 3 * 3 window, pixel value with the central point of window is compared as eight points except the center in standard and the window, if this pixel value is greater than mean pixel, then this two-value is turned to 1, otherwise the value that will change the time becomes 0.The numerical value that draws successively is the coding of this LBP feature.
But LBP does not have the ability that large scale is described.The MBLBP feature of being expanded out by LBP has been improved the weakness of basic LBP, its basic idea is: big or small arbitrarily image region average mark is slit into 9 sub-pieces of rectangle, use the LBP operator that the average gray value of these 9 rectangular blocks is encoded then, encoding with this characterizes this image region.The MBLBP feature since relatively be magnitude relationship between the piece, therefore have the descriptive power of large-scale structure, it is fast and accuracy rate is high to be used for detected image speed.
The core concept of gentleboost algorithm is to select the very little crucial visual signature of a part from a very big feature set, thereby produces an extremely effectively sorter.It utilizes a large amount of classification capacities that general simple classification device is stacked up by certain method, constitutes the strong classifier that classification capacity is very strong, again several strong classifier series connection becoming cascade classifiers is finished picture search and detects.Subimage by the first collection sorter is given second level sorter, gives third level sorter by the subimage of the second collection sorter, by that analogy.The gentleboost training method is to insensitive for noise with respect to the characteristics of Adaboost training method maximum, can improve the detection effect significantly.
Present embodiment has following advantage.
1, by analyzing in the eye image, upper and lower 2 distance of eyeball judges whether the driver is tired, as driver fatigue, its eyeball is blocked by eyelid, distance between upper and lower 2 of the eyeball can diminish, if the distance between upper and lower 2 of the eyeball is dwindled above certain limit the processing of then reporting to the police, avoid occurring traffic hazard, monitoring accuracy is higher.
2, after camera grasps head portrait, filter by multistage classifier, can comprise one or more features in each grade sorter, behind first order sorter, if image is filtered, assert that then this image is not a target image, if image is by first order sorter, then this image is again through second level sorter (even third level sorter, fourth stage sorter ...), be target image by the image behind all sorters, filter by multistage classifier, monitoring velocity is fast, the precision height.
3, in the process that grasps image, grasp face image earlier, reduce the scope again, in face image, grasp eye image, to improve the speed that grasps image.
4, further, in the process that the driver is monitored, also the angle that driver's head is departed from is monitored, with driver's head is 0 degree (benchmark angle) when the dead ahead, if driver's head departs from the benchmark angle and surpasses setting range, this moment, driver's sight line very likely departed from the dead ahead, can bring harm to driving safety, processings of reporting to the police this moment further improves the reliability and the security of monitoring.
Be specific embodiments of the invention only below, do not limit protection scope of the present invention with this; Any replacement and the improvement done on the basis of not violating the present invention's design all belong to protection scope of the present invention.
Claims (5)
1. a fatigue driving monitoring method is characterized in that, this method comprises the steps: that A, camera grasp image, and the image that is grasped is filtered through the sorter in the monitor, is eye image by the image behind the sorter; Eye image after B, the analysis and filter finds eyeball upper and lower ultimate range L1 in setting-up time at 2; C, driver's eyeball upper and lower distance L 2 and the difference between the ultimate range L1 in setting-up time are monitored and calculated to driver's eyeball at 2, if this difference surpasses the scope of setting, the processing of then reporting to the police.
2. fatigue driving monitoring method according to claim 1, it is characterized in that, described sorter comprises first order sorter and second level sorter at least, in the step A, after camera grasps image, filter through first order sorter earlier, filter through second level sorter again by the image behind the first order sorter.
3. fatigue driving monitoring method according to claim 1 is characterized in that, in the step A, camera filters through sorter after grasping image, is facial image by the image behind the sorter, grasps eye image again in facial image.
4. state the fatigue driving monitoring method as claim 3, it is characterized in that, also comprise at least three visual angle detecting devices, the detection angles scope difference of each visual angle detecting device, in the step A, the image that camera grasped is detected by each visual angle detecting device, if the image that is grasped is by corresponding visual angle detecting device, then the angular range that is detected by this visual angle detecting device is determined the angular range that driver face is departed from, if the angular range that driver face is departed from surpasses the angular range of setting, the processing of then reporting to the police.
5. as fatigue driving monitoring method as described in the claim 3, it is characterized in that, in the step A, camera filters through sorter after grasping image, if this image passes through sorter, colour of skin information with colour of skin information in this image and setting compares again, if meet, then is judged to be facial image, if do not meet, then be judged to be non-face image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010106025466A CN102013013B (en) | 2010-12-23 | 2010-12-23 | Fatigue driving monitoring method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010106025466A CN102013013B (en) | 2010-12-23 | 2010-12-23 | Fatigue driving monitoring method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102013013A true CN102013013A (en) | 2011-04-13 |
| CN102013013B CN102013013B (en) | 2013-07-03 |
Family
ID=43843185
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2010106025466A Expired - Fee Related CN102013013B (en) | 2010-12-23 | 2010-12-23 | Fatigue driving monitoring method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102013013B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103310187A (en) * | 2012-03-13 | 2013-09-18 | 霍尼韦尔国际公司 | Face image prioritization based on face quality analysis |
| CN103679669A (en) * | 2012-09-20 | 2014-03-26 | 上海联影医疗科技有限公司 | Image fusion method based on Lab space |
| CN105069976A (en) * | 2015-07-28 | 2015-11-18 | 南京工程学院 | Integrated fatigue detection and driving record system and fatigue detection method |
| CN108021911A (en) * | 2018-01-04 | 2018-05-11 | 重庆公共运输职业学院 | A kind of driver tired driving monitoring method |
| WO2020078464A1 (en) * | 2018-10-19 | 2020-04-23 | 上海商汤智能科技有限公司 | Driving state detection method and apparatus, driver monitoring system, and vehicle |
| CN111291590A (en) * | 2018-12-06 | 2020-06-16 | 广州汽车集团股份有限公司 | Driver fatigue detection method, driver fatigue detection device, computer equipment and storage medium |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104680120B (en) * | 2013-12-02 | 2018-10-19 | 华为技术有限公司 | A kind of generation method and device of the strong classifier of Face datection |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6927694B1 (en) * | 2001-08-20 | 2005-08-09 | Research Foundation Of The University Of Central Florida | Algorithm for monitoring head/eye motion for driver alertness with one camera |
| US20070286457A1 (en) * | 2006-06-13 | 2007-12-13 | Hammoud Riad I | Dynamic eye tracking system |
| CN101375796A (en) * | 2008-09-18 | 2009-03-04 | 浙江工业大学 | Real-time detection system of fatigue driving |
| CN101872419A (en) * | 2010-06-09 | 2010-10-27 | 谭台哲 | Method for detecting fatigue of automobile driver |
-
2010
- 2010-12-23 CN CN2010106025466A patent/CN102013013B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6927694B1 (en) * | 2001-08-20 | 2005-08-09 | Research Foundation Of The University Of Central Florida | Algorithm for monitoring head/eye motion for driver alertness with one camera |
| US20070286457A1 (en) * | 2006-06-13 | 2007-12-13 | Hammoud Riad I | Dynamic eye tracking system |
| CN101375796A (en) * | 2008-09-18 | 2009-03-04 | 浙江工业大学 | Real-time detection system of fatigue driving |
| CN101872419A (en) * | 2010-06-09 | 2010-10-27 | 谭台哲 | Method for detecting fatigue of automobile driver |
Non-Patent Citations (3)
| Title |
|---|
| 《公路交通科技》 20080731 林土胜等 基于类Haar特征的驾驶者人眼疲劳状态的检测方法研究 128-131 1-5 第25卷, 第7期 * |
| 《公路交通科技》 20100531 郭克友等 基于视觉的驾驶人疲劳及注意力监测方法 104-109 3-4 第27卷, 第5期 * |
| 《计算机与数字工程》 20091231 王秀等 驾驶疲劳的眼睛状态检测算法 135-137,182 1-5 第37卷, 第8期 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103310187A (en) * | 2012-03-13 | 2013-09-18 | 霍尼韦尔国际公司 | Face image prioritization based on face quality analysis |
| CN103310187B (en) * | 2012-03-13 | 2017-10-10 | 智航知识产权运营管理有限公司 | Face-image based on facial quality analysis is prioritized |
| CN103679669A (en) * | 2012-09-20 | 2014-03-26 | 上海联影医疗科技有限公司 | Image fusion method based on Lab space |
| CN103679669B (en) * | 2012-09-20 | 2017-02-01 | 上海联影医疗科技有限公司 | Image fusion method based on Lab space |
| CN105069976A (en) * | 2015-07-28 | 2015-11-18 | 南京工程学院 | Integrated fatigue detection and driving record system and fatigue detection method |
| CN105069976B (en) * | 2015-07-28 | 2017-10-24 | 南京工程学院 | A kind of fatigue detecting and traveling record integrated system and fatigue detection method |
| CN108021911A (en) * | 2018-01-04 | 2018-05-11 | 重庆公共运输职业学院 | A kind of driver tired driving monitoring method |
| WO2020078464A1 (en) * | 2018-10-19 | 2020-04-23 | 上海商汤智能科技有限公司 | Driving state detection method and apparatus, driver monitoring system, and vehicle |
| US11308723B2 (en) | 2018-10-19 | 2022-04-19 | Shanghai Sensetime Intelligent Technology Co., Ltd. | Driving state detection method and apparatus, driver monitoring system and vehicle |
| CN111291590A (en) * | 2018-12-06 | 2020-06-16 | 广州汽车集团股份有限公司 | Driver fatigue detection method, driver fatigue detection device, computer equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102013013B (en) | 2013-07-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102013013B (en) | Fatigue driving monitoring method | |
| CN109670446B (en) | Abnormal behavior detection method based on linear dynamic system and deep network | |
| Tipprasert et al. | A method of driver’s eyes closure and yawning detection for drowsiness analysis by infrared camera | |
| CN108053427B (en) | Improved multi-target tracking method, system and device based on KCF and Kalman | |
| CN108009473B (en) | Video structuralization processing method, system and storage device based on target behavior attribute | |
| CN108053615B (en) | Method for detecting fatigue driving state of driver based on micro-expression | |
| CN103824070B (en) | A kind of rapid pedestrian detection method based on computer vision | |
| CN101477625B (en) | Upper half of human body detection method and system | |
| CN105354985B (en) | Fatigue driving monitoring apparatus and method | |
| Berri et al. | A pattern recognition system for detecting use of mobile phones while driving | |
| Qiao et al. | A smartphone-based driver fatigue detection using fusion of multiple real-time facial features | |
| CN103839065A (en) | Extraction method for dynamic crowd gathering characteristics | |
| CN107911663A (en) | A kind of elevator passenger hazardous act intelligent recognition early warning system based on Computer Vision Detection | |
| CN104966059B (en) | Driver based on intelligent monitor system drives detection method of making a phone call | |
| CN103049740A (en) | Method and device for detecting fatigue state based on video image | |
| CN101620673A (en) | Robust face detecting and tracking method | |
| CN108229245A (en) | Method for detecting fatigue driving based on facial video features | |
| CN110751051A (en) | Abnormal driving behavior detection method based on machine vision | |
| CN112381870B (en) | Binocular vision-based ship identification and navigational speed measurement system and method | |
| CN108960107A (en) | A kind of overcrowding recognition methods of small mini van and device | |
| CN112396658A (en) | Indoor personnel positioning method and positioning system based on video | |
| CN102542246A (en) | Abnormal face detection method for ATM (Automatic Teller Machine) | |
| CN104077594A (en) | Image recognition method and device | |
| CN104068868A (en) | Method and device for monitoring driver fatigue on basis of machine vision | |
| CN107563346A (en) | One kind realizes that driver fatigue sentences method for distinguishing based on eye image processing |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20130703 Termination date: 20161223 |