CN104504400A - Detection method of driver's abnormal behavior modeled based on online behavior - Google Patents

Detection method of driver's abnormal behavior modeled based on online behavior Download PDF

Info

Publication number
CN104504400A
CN104504400A CN201510009061.9A CN201510009061A CN104504400A CN 104504400 A CN104504400 A CN 104504400A CN 201510009061 A CN201510009061 A CN 201510009061A CN 104504400 A CN104504400 A CN 104504400A
Authority
CN
China
Prior art keywords
driver
model
behavior
detection method
normal
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
Application number
CN201510009061.9A
Other languages
Chinese (zh)
Other versions
CN104504400B (en
Inventor
李远钱
周曦
周祥东
石宇
颜卓
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chongqing Institute of Green and Intelligent Technology of CAS
Original Assignee
Chongqing Institute of Green and Intelligent Technology of CAS
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Chongqing Institute of Green and Intelligent Technology of CAS filed Critical Chongqing Institute of Green and Intelligent Technology of CAS
Priority to CN201510009061.9A priority Critical patent/CN104504400B/en
Publication of CN104504400A publication Critical patent/CN104504400A/en
Application granted granted Critical
Publication of CN104504400B publication Critical patent/CN104504400B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation

Landscapes

  • Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Emergency Alarm Devices (AREA)

Abstract

The invention relates to a detection method of a driver's abnormal behavior modeled based on an online behavior and belongs to the technical field of image identification and monitoring. The method is based on a video analysis technology, the driver's abnormal behavior is detected by modeling the driver's normal behavior, and the method mainly comprises the following steps of: firstly, initializing and creating an initial model or updating the existing model; secondly, extracting driver's behavior characteristics in a driving process; thirdly, judging whether the driver's behavior is normal according to the initial model and the driver's behavior characteristics; fourthly, updating the model. The detection method is characterized in that the abnormal behavior is detected by a novelty detection way, multiple normal and abnormal behaviors can be treated by adopting a multi-mode modeling method, and false alarms are eliminated by a manual labeling method, thus the steadiness of the solution is increased and the error rate of the solution is reduced.

Description

A kind of driver's anomaly detection method based on online behavior modeling
Technical field
The present invention relates to a kind of driver's anomaly detection method, particularly a kind of driver's anomaly detection method based on online behavior modeling.
Background technology
Vehicle-mounted vision system has become the emerging application direction in image procossing, video analysis field, to the research of driver's unusual checking, belong to intelligent transportation field, it is the auxiliary gordian technique of driving of intelligence, this technology is by detecting driver's abnormal behaviour and giving a warning, the generation avoided traffic accident, has significant application value and social effect.
The research and development of the outer many countries of Current Domestic in this field are all limited to abnormal behaviour own characteristic, and such as abnormal behaviour kind is many, comprises driver and closes one's eyes for a long time, yawns, rubs one's eyes, puts first-class fatigue behaviour; Improper driving behavior classification is complicated, comprises making a phone call, eat a piece, smoke, glancing right and left; Data acquisition is more difficult, and because driver's overwhelming majority time behavior is normal, therefore abnormal behaviour occurrence number is generally less.Therefore, prior art is divided into two classes substantially: one just detects for certain specific exceptions behavior, and this method only detects for fatigue driving behavior, but the abnormal behaviour concrete manifestation of different driver may be different; Another kind is that the mode of training is single, because abnormal behaviour number of times is few, sample collection is more difficult, so be absolutely scheme be mostly all the data utilizing artificial simulation, offline mode is adopted to train, but the data of simulation are usually true not again, and the possibility that the model of therefore training in use goes wrong is comparatively large, this causes difficulty also to the use of off-line training method.
Summary of the invention
In view of this, the object of the present invention is to provide a kind of driver's anomaly detection method based on online behavior modeling.
For achieving the above object, the invention provides following technical scheme:
Based on driver's anomaly detection method of online behavior modeling, the method, based on Video Analysis Technology, by detecting the abnormal behaviour of driver to the normal behaviour modeling of driver, specifically comprises the following steps:
Step one: initialization, in initialization procedure, the behavior of driver is normal behaviour, carries out feature extraction to the behavior of driver, to model training, sets up initial model or upgrades existing model;
Step 2: feature extraction is carried out to the behavior of driver in driving procedure;
Step 3: the behavioural characteristic in conjunction with initial model and driver judges that whether the behavior of driver is normal;
Step 4: model is upgraded.
Further, the initial model in described step one is:
Wherein, P (F t) be t feature F tthe probability occurred, K is distributed quantity, η (F t, μ i,t, Σ i,t) be moment t i-th Gaussian distribution, ω i,tfor its weight, μ i,t, Σ i,tbe respectively its average and covariance matrix.
Further, in described step 2 to the method that feature extraction is carried out in the behavior of driver be, feature extracted to continuous n two field picture, n>=1, the feature F at t pixel (x, y) place t(x, y)=[I t(x, y), I t-1(x, y) ..., I t-n+1(x, y)], wherein I t(x, y) is the pixel value of moment t position (x, y), altogether n dimension.
Further, in described step 2 to the method that feature extraction is carried out in the behavior of driver can also be, image is divided into and has overlapping image block, feature (as SIFT feature) is extracted to n frame continuous in each block region, then Feature Combination is got up to form last feature.
Further, described step 3 specifically comprises the following steps:
1) F is judged by following formula twhether mate with i-th Gauss model in K gauss hybrid models, i=1 ..., K;
(F t-μ i,t) TΣ -1(F t-μ i,t)≤τ；
τ is matching threshold, is less than or equal to this threshold value characterization and belongs to this Gauss model;
If fruit meets above-mentioned formula, then jump to step 2), otherwise jump to step 3);
2) judgement and F twhether the model i of coupling is normal behaviour model, if then judge that the behavior of driver is normal behaviour;
Otherwise, judge that the behavior of driver is as abnormal behaviour, sends warning;
3) F tnot with any one Model Matching in K gauss hybrid models, judge that driving behavior is as abnormal behaviour, sends
Report to the police, and increase a Gauss model and its weight is set to one and do not meet smaller value.
Further, normal behaviour model acquisition methods is: choose K gauss hybrid models and sort from high to low according to ω/trace (Σ), and getting a front B Gauss model is normal behaviour model, and wherein B meets: t is the threshold value choosing normal behaviour model, namely normal behaviour model weight and must T be greater than.
Further, described step 4 specifically comprises the following steps:
1) average and the covariance matrix of i-th model is upgraded,
μ i,t＝(1-α)μ i,t-1+αF t，
Σ i , t 2 = ( 1 - α ) Σ i , t - 1 2 + α ( F t - μ i , t ) T ( F t - μ i , t ) ,
Wherein α is learning rate;
2) weight of i-th model is upgraded,
ω i,t＝(1-β)ω i,t-1+β(M i,t)，
Wherein β is learning rate, M i,t=1, i is by Matching Model; M j,t=0, j is other behavior model, i.e. j ≠ i.
Further, when the behavior of driver is normal, and when being detected as abnormal behaviour, manually F can be set tfor false-alarm, and the model i it and its matched is set to normal model, improves weight and makes it meet when the abnormal behavior of driver, and when being detected as normal behaviour, manually can set the behavior is abnormal behaviour, and by the Model Weight reduced and it mates.
Further, the method that model is set up comprises GMM modeling method, online boosting, nonparametric technique.
Further, anomaly detection method can also by judging that in probability model probability junior is as abnormal behaviour.
Beneficial effect of the present invention is: a kind of driver's anomaly detection method based on online behavior modeling provided by the invention, have the following advantages: (1) due to abnormal behaviour data few, it is difficult to gather, and this method does not rely on the collection of abnormal behaviour data, and the mode of novelty detection is adopted to detect abnormal behaviour; (2) adopt the method for multi-modal modeling to make model describe multiple behavior, therefore can process multiple normal behaviour and multiple abnormal behaviour, and be not limited to and only certain abnormal behaviour reported to the police; (3) method being aided with artificial mark gets rid of false-alarm, adds the stability of scheme, makes the system of employing this programme more fewer by error rate.
Accompanying drawing explanation
In order to make the object, technical solutions and advantages of the present invention clearly, below in conjunction with accompanying drawing, the present invention is described in further detail, wherein:
Fig. 1 is the process flow diagram of the method for the invention.
Embodiment
Below in conjunction with accompanying drawing, the preferred embodiments of the present invention are described in detail.
The present invention is based on the method for the detection prospect by background modeling in Video Analysis Technology, propose a kind of driver's anomaly detection method based on online behavior modeling, the method is by detecting driver's abnormal behaviour to the normal behaviour modeling of driver.
This method considers following situation: most behaviors of (1) driver are normal behaviours, only has minority behavior to be abnormal; (2) normal behaviour and abnormal behaviour can be divided in feature space.Therefore, suppose that major part input data belong to this class of normal behaviour, carry out modeling to the data of normal behaviour class, whether the abnormal behaviour of such minority just can by meeting normal behaviour model to judge.
The main thought of the method is as follows: set up initial model; Feature extraction is carried out to the behavior of driver in driving procedure; Behavioural characteristic in conjunction with initial model and driver judges that whether the behavior of driver is normal; Then model is upgraded.
It should be noted that to make the situations such as the change of model energy adaptive environment, driver's change, needing to introduce initialization procedure and on-line study renewal process.In initialization procedure, suppose that all behavior acts of driver are normally, utilize these data Modling model or upgrade existing model.
The flow process of the method for the invention is as shown in figure (1).
This method mainly comprises the following steps:
Step one: initialization, in initialization procedure, the behavior of driver is normal behaviour, carries out feature extraction to the behavior of driver, to model training, sets up initial model or upgrades existing model;
Initial model is:
P ( F t ) = Σ i = 1 K ω i , t · η ( F t , μ i , t , Σ i , t ) - - - ( 1 )
Wherein, K is distributed quantity, can be estimated, η (F by normal behaviour species number t, μ i,t, Σ i,t) be moment t i-th Gaussian distribution, ω i,tfor its weight, μ i,t, Σ i,tbe respectively its average and covariance matrix.In order to convenience of calculation, Σ i,t=Λ tΛ
Step 2: feature extraction is carried out to the behavior of driver in driving procedure;
Suppose that the basic act of driver completes in n frame, then feature is extracted to continuous n (n>=1) two field picture: the feature F at t pixel (x, y) place t(x, y)=[I t(x, y), I t+1(x, y) ..., I t+n-1(x, y)], wherein I t(x, y) is the pixel value of moment t position (x, y), altogether n dimension.
In addition, image can also be divided into and have overlapping image block, feature (as SIFT feature) be carried to n frame continuous in each block region, then Feature Combination is got up to form last feature.
Step 3: the behavioural characteristic in conjunction with initial model and driver judges that whether the behavior of driver is normal;
1) F is judged by following formula twhether mate with i-th Gauss model in K gauss hybrid models, i=1 ..., K;
(F t-μ i,t) TΣ -1(F t-μ i,t)≤τ (3)
τ is matching threshold, is less than or equal to this threshold value characterization and model i matches;
If fruit meets above-mentioned formula, then jump to step 2), otherwise jump to step 3);
2) judgement and F twhether the model i of coupling is normal behaviour model, if then judge that the behavior of driver is normal behaviour; Otherwise, judge that the behavior of driver is as abnormal behaviour, sends warning;
Normal behaviour model acquisition methods is: choose K gauss hybrid models and sort from high to low according to ω/trace (Σ), and getting a front B Gauss model is normal behaviour model, and wherein B meets:
B = arg min b ( Σ i = 1 b ω i > T ) - - - ( 3 )
T is the threshold value choosing normal behaviour model, namely normal behaviour model weight and must T be greater than.
3) F tnot with any one Model Matching in K gauss hybrid models, judge that driving behavior is as abnormal behaviour, sends warning, and increase a Gauss model and its weight is set to one and do not meet smaller value.
Step 4: model is upgraded.
1) average and the covariance matrix of i-th model is upgraded,
μ i,t＝(1-α)μ i,t-1+αF t(4)
Σ i , t 2 = ( 1 - α ) Σ i , t - 1 2 + α ( F t - μ i , t ) T ( F t - μ i , t ) - - - ( 5 )
Wherein α is learning rate;
2) weight of i-th model is upgraded,
ω i,t＝(1-β)ω i,t-1+α(M i,t) (6)
Wherein β is learning rate, M i,t=1, i is by Matching Model; M j,t=0, j is other behavior model, i.e. j ≠ i.
If the behavior of driver is normal, and when being detected as abnormal behaviour, manually can set F tfor false-alarm (warning of mistake), and the model i it and its matched is set to normal model, improves weight and makes it meet when the abnormal behavior of driver, and when being detected as normal behaviour, manually can set the behavior is abnormal behaviour, and by the Model Weight reduced and it mates.
It should be noted that to make the situations such as the change of model energy adaptive environment, driver's change, needing to introduce initialization procedure and on-line study renewal process.In initialization procedure, suppose that all behavior acts of driver are normally, utilize these data Modling model or upgrade existing model.
The present invention program proposes a kind of driver's anomaly detection method based on online behavior modeling, the mode of noveltydetection is adopted to detect abnormal behaviour, multi-frame video data are utilized to extract the feature of description behavior, and this feature is carried out to the multi-modal modeling of on-line study, make normal behaviour feature have higher fractional in this model, abnormal behavior has lower mark in this model.
Feature extracting method is not limited to the feature of illustrating and proposing above, and the feature that any one describes behavior can be applied to this framework.
The method of modeling is not limited to the GMM modeling method of illustrating and proposing above, and any one modeling method can be applied to this framework, as online boosting, and nonparametric technique (histogram) etc.
Anomaly detection method can also be the detection method of other novelty detection, and as judged in probability model, probability junior is as abnormal behaviour.
The method mark of artificial setting can be adopted for false-alarm, and the weight of the corresponding model of false-alarm is provided automatically, avoid next false-alarm.
Utilize the method for on-line study that model descriptive power is strengthened, treatable behavior increases, and false-alarm is fewer and feweri, and the change that can conform.
What finally illustrate is, above preferred embodiment is only in order to illustrate technical scheme of the present invention and unrestricted, although by above preferred embodiment to invention has been detailed description, but those skilled in the art are to be understood that, various change can be made to it in the form and details, and not depart from claims of the present invention limited range.

Claims (10)

1. based on driver's anomaly detection method of online behavior modeling, it is characterized in that: the method, based on Video Analysis Technology, by detecting the abnormal behaviour of driver to the normal behaviour modeling of driver, specifically comprises the following steps:
Step one: initialization, in initialization procedure, the behavior of driver is normal behaviour, carries out feature extraction to the behavior of driver, to model training, sets up initial model or upgrades existing model;
Step 2: feature extraction is carried out to the behavior of driver in driving procedure;
Step 3: the behavioural characteristic in conjunction with initial model and driver judges that whether the behavior of driver is normal;
Step 4: model is upgraded.
2. a kind of driver's anomaly detection method based on online behavior modeling according to claim 1, is characterized in that: the model in described step one is: P ( F t ) = Σ i = 1 K ω i , t · η ( F t , μ i , t , Σ i , t ) ,
Wherein, P (F t) be t feature F tthe probability occurred, K is distributed quantity, η (F t, μ i,t, Σ i,t) be moment t i-th Gaussian distribution, ω i,tfor its weight, μ i,t, Σ i,tbe respectively its average and covariance matrix.
3. a kind of driver's anomaly detection method based on online behavior modeling according to claim 1, it is characterized in that: in described step 2 to the method that feature extraction is carried out in the behavior of driver in driving procedure be, feature is extracted to continuous n two field picture, n>=1, the feature F at t pixel (x, y) place t(x, y)=[I t(x, y), I t-1(x, y) ..., I t-n+1(x, y)], wherein I t(x, y) is the pixel value of moment t position (x, y), altogether n dimension.
4. a kind of driver's anomaly detection method based on online behavior modeling according to claim 1, it is characterized in that: in described step 2 to the method that feature extraction is carried out in the behavior of driver in driving procedure can also be, image is divided into and has overlapping image block, feature is extracted to n two field picture continuous in each block region, then Feature Combination is got up to form last feature.
5. a kind of driver's anomaly detection method based on online behavior modeling according to claim 1, is characterized in that: described step 3 specifically comprises the following steps:
1) F is judged by following formula twhether mate with i-th Gauss model in K gauss hybrid models, i=1 ..., K;
(F t-μ i,t) TΣ -1(F t-μ i,t)≤τ；
τ is matching threshold, is less than or equal to this threshold value characterization and model i matches;
If fruit meets above-mentioned formula, then jump to step 2), otherwise jump to step 3);
2) judgement and F twhether the model i of coupling is normal behaviour model, if then judge that the behavior of driver is normal behaviour; Otherwise, judge that the behavior of driver is as abnormal behaviour, sends warning;
3) F tnot with any one Model Matching in K gauss hybrid models, judge that driving behavior is as abnormal behaviour, sends warning, and increase a Gauss model and its weight is set to one and do not meet smaller value.
6. a kind of driver's anomaly detection method based on online behavior modeling according to claim 5, it is characterized in that: normal behaviour model acquisition methods is: in a model in K gauss hybrid models, sort from high to low according to ω/trace (Σ), getting a front B Gauss model is normal behaviour model, and wherein B meets: t is the threshold value choosing normal behaviour model.
7. a kind of driver's anomaly detection method based on online behavior modeling according to claim 1, is characterized in that: described step 4 specifically comprises the following steps:
1) average and the covariance matrix of i-th model is upgraded,
μ i,t＝(1-α)μ i,t-1+αF t，
Σ i , t 2 = ( 1 - α ) Σ i , t - 1 2 + α ( F t - μ i , t ) T ( F t - μ i , t ) ,
Wherein α is learning rate;
2) weight of i-th model is upgraded,
ω i,t＝(1-β)ω i,t-1+β(M i,t)，
Wherein β is learning rate, M i,t=1, i is by Matching Model; M j,t=0, j is other behavior model, i.e. j ≠ i.
8. a kind of driver's anomaly detection method based on online behavior modeling according to claim 1, is characterized in that: when the behavior of driver is normal, and when being detected as abnormal behaviour, manually can set F tfor false-alarm, and the model i it and its matched is set to normal model, improves weight and makes it meet when the abnormal behavior of driver, and when being detected as normal behaviour, manually can set the behavior is abnormal behaviour, and by the Model Weight reduced and it mates.
9. a kind of driver's anomaly detection method based on online behavior modeling according to claim 1, is characterized in that: the method that model is set up comprises GMM modeling method, online boosting, nonparametric technique.
10. a kind of driver's anomaly detection method based on online behavior modeling according to claim 1, is characterized in that: anomaly detection method can also by judging that in probability model probability junior is as abnormal behaviour.
CN201510009061.9A 2015-01-08 2015-01-08 A kind of driver's anomaly detection method based on online behavior modeling Expired - Fee Related CN104504400B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510009061.9A CN104504400B (en) 2015-01-08 2015-01-08 A kind of driver's anomaly detection method based on online behavior modeling

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510009061.9A CN104504400B (en) 2015-01-08 2015-01-08 A kind of driver's anomaly detection method based on online behavior modeling

Publications (2)

Publication Number Publication Date
CN104504400A true CN104504400A (en) 2015-04-08
CN104504400B CN104504400B (en) 2018-07-17

Family

ID=52945796

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510009061.9A Expired - Fee Related CN104504400B (en) 2015-01-08 2015-01-08 A kind of driver's anomaly detection method based on online behavior modeling

Country Status (1)

Country Link
CN (1) CN104504400B (en)

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106022291A (en) * 2016-05-31 2016-10-12 北京理工大学 Method of detecting braking intention of driver in emergency state based on neural signal
WO2017124683A1 (en) * 2016-01-21 2017-07-27 杭州海康威视数字技术股份有限公司 Method and device for updating online self-learning event detection model
CN108009463A (en) * 2016-10-27 2018-05-08 ***通信有限公司研究院 A kind of method and device of identification
CN109241994A (en) * 2018-07-31 2019-01-18 顺丰科技有限公司 A kind of user's anomaly detection method, device, equipment and storage medium
CN109858553A (en) * 2019-01-31 2019-06-07 深圳市赛梅斯凯科技有限公司 Monitoring model update method, updating device and the storage medium of driving condition
CN109886016A (en) * 2018-12-27 2019-06-14 慧安金科（北京）科技有限公司 For detecting the method, equipment and computer readable storage medium of abnormal data
WO2019127271A1 (en) * 2017-12-28 2019-07-04 深圳市锐明技术股份有限公司 Warning method, device, storage medium and server regarding physical conflict behavior
CN110533328A (en) * 2019-08-30 2019-12-03 广州广电城市服务集团股份有限公司 A kind of project scene traffic control method, apparatus, medium and terminal device
EP3468469A4 (en) * 2016-06-13 2020-02-19 Xevo Inc. Method and system for providing behavior of vehicle operator using virtuous cycle
CN111301428A (en) * 2018-12-11 2020-06-19 上汽通用汽车有限公司 Motor vehicle driver distraction detection warning method and system and motor vehicle
US10950132B2 (en) 2016-12-22 2021-03-16 Xevo Inc. Method and system for providing artificial intelligence analytic (AIA) services using operator fingerprints and cloud data
CN108885133B (en) * 2016-04-01 2021-05-14 日本电信电话株式会社 Abnormal sound detection learning device, method thereof, and recording medium
CN113570473A (en) * 2021-06-25 2021-10-29 深圳供电局有限公司 Equipment fault monitoring method and device, computer equipment and storage medium
CN114202994A (en) * 2021-12-03 2022-03-18 广州小鹏汽车科技有限公司 Intelligent driving behavior-assisted safety education system and electronic equipment

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102799863A (en) * 2012-07-02 2012-11-28 中国计量学院 Method for detecting group crowd abnormal behaviors in video monitoring
US20140241619A1 (en) * 2013-02-25 2014-08-28 Seoul National University Industry Foundation Method and apparatus for detecting abnormal movement

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102799863A (en) * 2012-07-02 2012-11-28 中国计量学院 Method for detecting group crowd abnormal behaviors in video monitoring
US20140241619A1 (en) * 2013-02-25 2014-08-28 Seoul National University Industry Foundation Method and apparatus for detecting abnormal movement

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
杨盼: "基于驾驶行为的驾驶状态分析***", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *
高凯亮等: "一种混合高斯北京模型下的像素分类运动目标检测方法", 《南京大学学报（自然科学）》 *

Cited By (27)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2017124683A1 (en) * 2016-01-21 2017-07-27 杭州海康威视数字技术股份有限公司 Method and device for updating online self-learning event detection model
CN106991072A (en) * 2016-01-21 2017-07-28 杭州海康威视数字技术股份有限公司 Automatic measure on line event detection model update method and device
CN106991072B (en) * 2016-01-21 2022-12-06 杭州海康威视数字技术股份有限公司 Online self-learning event detection model updating method and device
EP3407200A4 (en) * 2016-01-21 2019-01-16 Hangzhou Hikvision Digital Technology Co., Ltd. Method and device for updating online self-learning event detection model
US11030886B2 (en) 2016-01-21 2021-06-08 Hangzhou Hikvision Digital Technology Co., Ltd. Method and device for updating online self-learning event detection model
CN108885133B (en) * 2016-04-01 2021-05-14 日本电信电话株式会社 Abnormal sound detection learning device, method thereof, and recording medium
CN106022291A (en) * 2016-05-31 2016-10-12 北京理工大学 Method of detecting braking intention of driver in emergency state based on neural signal
EP3468469A4 (en) * 2016-06-13 2020-02-19 Xevo Inc. Method and system for providing behavior of vehicle operator using virtuous cycle
US11068728B2 (en) 2016-06-13 2021-07-20 Xevo Inc. Method and system for providing behavior of vehicle operator using virtuous cycle
US10956758B2 (en) 2016-06-13 2021-03-23 Xevo Inc. Method and system for providing auto space management using virtuous cycle
CN108009463B (en) * 2016-10-27 2020-10-30 ***通信有限公司研究院 Identity recognition method and device
CN108009463A (en) * 2016-10-27 2018-05-08 ***通信有限公司研究院 A kind of method and device of identification
US11335200B2 (en) 2016-12-22 2022-05-17 Xevo Inc. Method and system for providing artificial intelligence analytic (AIA) services using operator fingerprints and cloud data
US10950132B2 (en) 2016-12-22 2021-03-16 Xevo Inc. Method and system for providing artificial intelligence analytic (AIA) services using operator fingerprints and cloud data
WO2019127271A1 (en) * 2017-12-28 2019-07-04 深圳市锐明技术股份有限公司 Warning method, device, storage medium and server regarding physical conflict behavior
CN109241994A (en) * 2018-07-31 2019-01-18 顺丰科技有限公司 A kind of user's anomaly detection method, device, equipment and storage medium
CN111301428A (en) * 2018-12-11 2020-06-19 上汽通用汽车有限公司 Motor vehicle driver distraction detection warning method and system and motor vehicle
CN112685735B (en) * 2018-12-27 2024-04-12 慧安金科（北京）科技有限公司 Method, apparatus and computer readable storage medium for detecting abnormal data
CN112685735A (en) * 2018-12-27 2021-04-20 慧安金科（北京）科技有限公司 Method, apparatus, and computer-readable storage medium for detecting abnormal data
CN109886016A (en) * 2018-12-27 2019-06-14 慧安金科（北京）科技有限公司 For detecting the method, equipment and computer readable storage medium of abnormal data
CN109858553B (en) * 2019-01-31 2023-12-12 锦图计算技术(深圳)有限公司 Method, device and storage medium for updating driving state monitoring model
CN109858553A (en) * 2019-01-31 2019-06-07 深圳市赛梅斯凯科技有限公司 Monitoring model update method, updating device and the storage medium of driving condition
CN110533328A (en) * 2019-08-30 2019-12-03 广州广电城市服务集团股份有限公司 A kind of project scene traffic control method, apparatus, medium and terminal device
CN113570473A (en) * 2021-06-25 2021-10-29 深圳供电局有限公司 Equipment fault monitoring method and device, computer equipment and storage medium
CN113570473B (en) * 2021-06-25 2024-02-09 深圳供电局有限公司 Equipment fault monitoring method, device, computer equipment and storage medium
CN114202994A (en) * 2021-12-03 2022-03-18 广州小鹏汽车科技有限公司 Intelligent driving behavior-assisted safety education system and electronic equipment
CN114202994B (en) * 2021-12-03 2024-04-19 广州小鹏汽车科技有限公司 Intelligent driving behavior assisting safety education system and electronic equipment

Also Published As

Publication number Publication date
CN104504400B (en) 2018-07-17

Similar Documents

Publication Publication Date Title
CN104504400A (en) Detection method of driver's abnormal behavior modeled based on online behavior
CN107402921B (en) Event time sequence data processing method, device and system for identifying user behaviors
CN103259962B (en) A kind of target tracking method and relevant apparatus
CN103258213B (en) A kind of for the dynamic vehicle model recognizing method in intelligent transportation system
CN108287989B (en) Sliding verification code man-machine identification method based on track
CN103077347B (en) A kind of hybrid intrusion detection method based on improving the fusion of kernel vector machine data
CN109446936A (en) A kind of personal identification method and device for monitoring scene
CN103096185B (en) A kind of video abstraction generating method and device
CN103150579A (en) Abnormal human behavior detecting method based on video sequence
CN110247910A (en) A kind of detection method of abnormal flow, system and associated component
CN104835179A (en) Improved ViBe background modeling algorithm based on dynamic background self-adaption
CN110288823B (en) Traffic violation misjudgment identification method based on naive Bayesian network
CN106991400A (en) A kind of fire hazard smoke detecting method and device
CN111382605A (en) Video content auditing method and device, storage medium and computer equipment
CN110738080A (en) method, device and electronic equipment for identifying modified motor vehicle
CN105469054A (en) Model construction method of normal behaviors and detection method of abnormal behaviors
CN103092925B (en) A kind of video abstraction generating method and device
CN104537354A (en) Target color recognition method and system for video surveillance
CN104077571A (en) Method for detecting abnormal behavior of throng by adopting single-class serialization model
Lin et al. FireMatch: A semi-supervised video fire detection network based on consistency and distribution alignment
CN108171187A (en) A kind of abnormal behaviour automatic identifying method and device based on the extraction of bone point
CN116204821B (en) Vibration evaluation method and system for rail transit vehicle
CN113807541B (en) Fairness repair method, system, equipment and storage medium for decision system
CN105451235A (en) Wireless sensor network intrusion detection method based on background updating
CN101814187A (en) Video tracking method based on multi-stage characteristics

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20180717

Termination date: 20210108

CF01 Termination of patent right due to non-payment of annual fee