CN109460780A - Safe driving of vehicle detection method, device and the storage medium of artificial neural network - Google Patents
Safe driving of vehicle detection method, device and the storage medium of artificial neural network Download PDFInfo
- Publication number
- CN109460780A CN109460780A CN201811209118.XA CN201811209118A CN109460780A CN 109460780 A CN109460780 A CN 109460780A CN 201811209118 A CN201811209118 A CN 201811209118A CN 109460780 A CN109460780 A CN 109460780A
- Authority
- CN
- China
- Prior art keywords
- parameter
- neural network
- driving
- vehicle
- artificial neural
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Traffic Control Systems (AREA)
Abstract
The present embodiments relate to technical field of automobile electronic control, disclose a kind of safe driving of vehicle detection method, device and computer readable storage medium based on artificial neural network, in the present invention, which comprises the feature samples parameter of acquisition history driving procedure；The artificial nerve network model including prime neural network and rear class neural network is established, and utilizes the feature samples parameter training artificial nerve network model of dangerous driving；Obtain the first kind characteristic parameter and the second category feature parameter of driver in real time in the current driving procedure of vehicle；First kind characteristic parameter and the second category feature parameter are inputted into artificial nerve network model, obtain dangerous driving tagsort probability value, dangerous driving state is in judgement driver and sends prompting message.Safe driving of vehicle detection method, device and computer readable storage medium provided by the invention based on artificial neural network can be improved the reliability and accuracy rate of dangerous driving evaluation.
Description
Technical field
The present embodiments relate to technical field of automobile electronic control, in particular to a kind of vehicle based on artificial neural network
Safe driving detection method, device and computer readable storage medium.
Background technique
With the development and the improvement of people's living standards of social economy, demand of the entire society to communications and transportation increasingly increases
Add.The rapid growth of the volume of traffic deteriorates road traffic condition, leads to traffic accidents, environmental pollution, traffic congestion, economy
A series of serious social concerns such as loss are even more especially to cause society with the particularly serious traffic accident for being related to passenger stock
Extensive concern.One of the main reason for driver's dangerous driving is particularly serious traffic accident.Statistical data shows: by driving fatigue
Caused traffic accident accounts for 40% or more of 20% of total number of accident or so, Zhan Te great traffic accident.In China, 2008 because of danger
Road traffic accident totally 2568 caused by danger drives, wherein dead 1353 people, injured 3129 people, caused by direct losses about
57380000 yuan.Dangerous driving accident rate is high, and consequence is serious, threatens the traffic safety in China always.It releases in the prior art
Driver fatigue monitor system acquires driver's human face expression state by camera in real time, is same as comparing with preset data model,
Discover whether that there are fatigue drivings with this.
At least there are the following problems in the prior art for inventor's discovery: traditional dangerous driving prevention uses the same inspection
Model is surveyed, therefore for different drivers, since individual difference causes further decreasing for precision and universality, leads to danger
Reliability and the accuracy rate for driving evaluation be not high, for example traditional dangerous driving detection model blinks data by user to analyze use
Whether family occurs fatigue, but somebody also can frequently blink in energetic state, which can also be detected as danger
It drives.
Summary of the invention
Embodiment of the present invention is designed to provide a kind of safe driving of vehicle detection side based on artificial neural network
Method, device and computer readable storage medium can be improved the reliability and accuracy rate of dangerous driving evaluation.
In order to solve the above technical problems, embodiments of the present invention provide a kind of vehicle peace based on artificial neural network
It is complete to drive detection method, comprising: the feature samples parameter of acquisition history driving procedure, wherein the feature samples parameter includes
In history driving data for characterize driver status first kind feature samples parameter, for characterizing vehicle running state
Two category feature sample parameters；The artificial nerve network model including prime neural network and rear class neural network is established, and is utilized
Artificial nerve network model described in the feature samples parameter training of the dangerous driving, wherein utilize the prime neural network
Fusion Features are carried out to the first kind feature samples parameter and the second category feature sample parameter respectively, after Fusion Features
The first kind feature samples parameter and the second category feature sample parameter input training in the rear class neural network, obtain
Take the artificial nerve network model comprising default dangerous driving tagsort probability value；It is obtained in real time in the current driving procedure of vehicle
Take the first kind characteristic parameter and the second category feature parameter of the driver；By the first kind characteristic parameter and second class
The characteristic parameter input artificial nerve network model comprising default dangerous driving tagsort probability value, obtains dangerous driving
Tagsort probability value, judges whether the dangerous driving tagsort probability value is greater than the default dangerous driving tagsort
Probability value, when determining that the dangerous driving tagsort probability value is greater than the default dangerous driving tagsort probability value,
Determine that the driver is in dangerous driving state and sends prompting message.
The safe driving of vehicle detection device based on artificial neural network that the present invention also provides a kind of, comprising: at least one
A processor；And the memory being connect at least one described processor communication；Wherein, be stored with can quilt for the memory
The instruction that at least one described processor executes, described instruction is executed by least one described processor, so that described at least one
A processor is able to carry out the above-mentioned safe driving of vehicle detection method based on artificial neural network.
The present invention also provides a kind of computer readable storage mediums, are stored with computer program, the computer program
The above-mentioned safe driving of vehicle detection method based on artificial neural network is realized when being executed by processor.
In terms of existing technologies, acquisition history driving procedure is for characterizing driver status for embodiment of the present invention
First kind feature samples parameter, the second category feature sample parameter for characterizing vehicle running state, with artificial neural network
Prime neural network in model carries out Fusion Features to these two types of feature samples parameters from two information sources respectively, then will
The first kind feature samples parameter and the second category feature sample parameter after Fusion Features input the rear class nerve net
In network, to realize the training to artificial nerve network model, the artificial nerve network model after training is integrated and is examined
The first kind feature samples parameter and the second category feature sample parameter of amount reflection dangerous driving, so as to comprehensive, objective, accurate
Measurement driver driving condition, the default dangerous driving tagsort probability value enabled overcomes based on unitary class
Other feature or single piece of information source determine driver whether the limitation of dangerous driving, reduce its false detection rate and omission factor, improve
The reliability and accuracy rate of dangerous driving evaluation；By obtain the driver in real time in the current driving procedure of vehicle
A kind of characteristic parameter and the second category feature parameter, then be inputted in the artificial nerve network model, it is special to obtain dangerous driving
Class probability value is levied, and is determining that it is general that the dangerous driving tagsort probability value is greater than the default dangerous driving tagsort
When rate value, determine that the driver is in dangerous driving state and sends prompting message, so as to prevent peril,
Protection is provided for driver safety trip.
In addition, the first kind characteristic parameter for obtaining the driver and vehicle in real time in the current driving procedure of vehicle
And the second category feature parameter, it specifically includes: acquiring the direct picture of the driver in real time；To the direct picture of acquisition into
Row Face datection is to obtain the facial image of the driver；Human eye in the facial image is positioned, described in identification
The state of human eye is to obtain the first kind characteristic parameter according to the state of the human eye；It is connect in real time with the vehicle communication,
To obtain the second category feature parameter.
In addition, the human eye in the facial image positions, identifies the state of the human eye, specifically includes:
Facial image input human face characteristic point is returned into convolutional neural networks model, regression forecasting is in a down-sampled low resolution
The boundary position of human eye in rate image.
In addition, the state of the human eye specifically includes: the opening and closing size of the human eye and the opening and closing frequency of the human eye.
In addition, the first kind feature samples parameter is facial behavioural characteristic sample parameter, the second category feature sample
Parameter is vehicle driving sample parameter；It is described to establish the artificial neural network mould including prime neural network and rear class neural network
Type specifically includes: constructing the first artificial neural network for merging the facial behavioural characteristic sample parameter, constructs for melting
The second artificial neural network for closing the vehicle driving sample parameter, manually by first artificial neural network and described second
For neural network as the prime neural network, establishing includes first artificial neural network, second artificial neural network
The artificial nerve network model of network and the rear class neural network.
In addition, the first kind characteristic parameter specifically includes: human eye number of winks in first time interval, second
Eye slackness in time interval and the gazing direction of human eyes in third time interval；The second category feature parameter is specific
It include: the velocity variations number of the driving vehicle in the 4th time interval, the steering wheel turn in the 5th time interval
Dynamic angle, the acceleration change number in the 6th time interval.By setting the different time for different characteristic parameters
Interval, so that each different characteristic parameter data more can reflect the driving shape of driver in matching time interval
State further improves the reliability and accuracy rate of dangerous driving evaluation.
In addition, special to the first kind feature samples parameter and second class respectively using the prime neural network
Sign sample parameter carries out before Fusion Features, further includes: to the human eye number of winks, the eye slackness, the driving
The velocity variations number of vehicle, the gazing direction of human eyes, the steering wheel angle variable, the acceleration change number when
Data synchronization processing is carried out on domain；Respectively to the human eye number of winks, the eye slackness, institute after data synchronization processing
State gazing direction of human eyes, the velocity variations number for driving vehicle, the steering wheel angle variable, the acceleration change time
Several or characterize data is normalized；Respectively to the human eye number of winks after normalized, eye relaxation
Degree, the gazing direction of human eyes, velocity variations number, the steering wheel angle variable, the acceleration for driving vehicle
Change frequency carries out data normalization processing；It is described that the first kind feature samples are joined respectively using the prime neural network
The several and described second category feature sample parameter carries out Fusion Features, specifically includes: based on the first artificial neural network logarithm
Fusion Features are carried out according to the human eye number of winks after standardization, the eye slackness, the gazing direction of human eyes,
Velocity variations number, the side based on second artificial neural network to the driving vehicle after data standardization
Fusion Features are carried out to disk corner variable, the acceleration change number.By right respectively using the prime neural network
The first kind feature samples parameter and the second category feature sample parameter carry out before Fusion Features to these parameters
Data processing, the driving condition for the embodiment driver for enabling these feature samples parameters after data processing more accurate,
Further improve the reliability and accuracy rate of dangerous driving evaluation.
In addition, the acquisition dangerous driving feature samples parameter, specifically includes: obtaining vehicle current driver's, history
Facial cybernetics control number and vehicle operation characteristic parameter in driving data are using as the dangerous driving feature samples parameter
First kind feature samples parameter and second feature sample parameter, or obtain history driver, in history driving data
Facial cybernetics control number and vehicle operation characteristic parameter are using as the dangerous driving feature samples parameter.By obtaining vehicle
Current driver's, facial cybernetics control number in history driving data and vehicle operation characteristic parameter are using as the danger
The first kind feature samples parameter and second feature sample parameter of driving characteristics sample parameter, i.e., the history periodically driven user
Data input artificial nerve network model be trained, the self-learning function of implementation model and optimize and revise model parameter so that
Safe driving detection more targetedly, is more able to satisfy personalized safe driving detection demand.
Detailed description of the invention
Fig. 1 is the safe driving of vehicle detection side based on artificial neural network that first embodiment provides according to the present invention
The flow chart of method；
Fig. 2 is the safe driving of vehicle detection side based on artificial neural network that second embodiment provides according to the present invention
The flow chart of method；
Fig. 3 is the safe driving of vehicle detection dress based on artificial neural network that third embodiment provides according to the present invention
The structural schematic diagram set.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with attached drawing to the present invention
Each embodiment be explained in detail.However, it will be understood by those skilled in the art that in each embodiment party of the present invention
In formula, many technical details are proposed in order to make reader more fully understand the present invention.But even if without these technical details
And various changes and modifications based on the following respective embodiments, claimed technical solution of the invention also may be implemented.
The safe driving of vehicle detection method based on artificial neural network that the first embodiment of the present invention is related to a kind of, such as
Shown in Fig. 1, comprising the following steps:
S101: the feature samples parameter of acquisition history driving procedure.
About step S101, specifically, the feature samples parameter includes driving in history driving data for characterizing
The first kind feature samples parameter of member's state, the second category feature sample parameter for characterizing vehicle running state, as shown in table 1
Reference standard, given a mark according to driving and driver status to it, establish dangerous driving sample data, the first kind is special
Sign sample parameter specifically includes: the human eye number of winks in first time interval, the relaxation of the eye in the second time interval
Degree and the gazing direction of human eyes in third time interval；The second category feature sample parameter specifically includes: in the 4th time
The velocity variations number of the driving vehicle in interval, the steering wheel rotational angle in the 5th time interval, the 6th when
Between interval in acceleration change number.By setting different time intervals for different characteristic parameters so that it is each not
Same characteristic parameter data more can reflect the driving condition of driver in matching time interval, further improve
The reliability and accuracy rate of dangerous driving evaluation.
Table 1
It is noted that in the present embodiment, the acquisition dangerous driving feature samples parameter is specifically included: being obtained
A current driver's of picking up the car, facial cybernetics control number in history driving data and vehicle operation characteristic parameter are using as institute
The first kind feature samples parameter and second feature sample parameter of dangerous driving feature samples parameter are stated, or obtains history driver
Member, facial cybernetics control number in history driving data and vehicle operation characteristic parameter are using as the dangerous driving feature
Sample parameter.By obtaining vehicle current driver's, facial cybernetics control number and vehicle driving in history driving data
Characteristic parameter using the first kind feature samples parameter and second feature sample parameter as the dangerous driving feature samples parameter,
I.e. periodically by user drive historical data input artificial nerve network model be trained, the self-learning function of implementation model and
Model parameter is optimized and revised, so that safe driving detection more targetedly, is more able to satisfy personalized safe driving detection and needs
It asks.
S102: the artificial nerve network model including prime neural network and rear class neural network is established, and utilizes danger
Artificial nerve network model described in the feature samples parameter training of driving obtains general comprising default dangerous driving characteristic parameter classification
The artificial nerve network model of rate value.
It is specifically, described to establish the artificial mind including prime neural network and rear class neural network about step S102
It through network model, specifically includes: constructing the first artificial neural network for merging the facial behavioural characteristic sample parameter, structure
The second artificial neural network for merging the vehicle driving sample parameter is built, by first artificial neural network and described
For second artificial neural network as the prime neural network, establishing includes first artificial neural network, second people
The artificial nerve network model of artificial neural networks and the rear class neural network.
It should be noted that in the present embodiment, rear class neural network is BP neural network, BP neural network is a kind of
By the Multi-layered Feedforward Networks of error back propagation (abbreviation error-duration model) training, algorithm is known as BP algorithm, its basic thought
Gradient descent method, using gradient search technology, to make network real output value and desired output error mean square it is poor
For minimum.
Basic BP algorithm includes two processes of backpropagation of the propagated forward and error of signal.When calculating error output
It is carried out by from the direction for being input to output, and adjusts weight and threshold value and then carried out from the direction for being output to input.When forward-propagating,
Input signal acts on output node by hidden layer, by nonlinear transformation, generates output signal, if reality output and expectation
Output is not consistent, then is transferred to the back-propagation process of error.Error-duration model be by output error by hidden layer to input layer by
Layer anti-pass, and give error distribution to all units of each layer, from the error signal that each layer obtains as adjustment each unit weight
Foundation.By adjusting the linking intensity and hidden node of input node and hidden node and the linking intensity and threshold of output node
Value declines error along gradient direction, by repetition learning training, determines network parameter (weight corresponding with minimal error
And threshold value), training stops stopping.Trained neural network can voluntarily be handled the input information of similar sample at this time
The smallest information by non-linear conversion of output error.
S103: the first kind characteristic parameter and the second category feature ginseng of driver are obtained in real time in the current driving procedure of vehicle
Number.
It is specifically, described to obtain the of the driver in real time in the current driving procedure of vehicle about step S103
A kind of characteristic parameter and the second category feature parameter, specifically include: acquiring the direct picture of the driver in real time；To the institute of acquisition
It states direct picture and carries out Face datection to obtain the facial image of the driver；Human eye in the facial image is determined
Position identifies the state of the human eye to obtain the first kind characteristic parameter according to the state of the human eye；In real time with the vehicle
Communication connection, to obtain the second category feature parameter.It is understood that a kind of vehicle-mounted by designing in present embodiment
Robot is connect with vehicle network data interface communication, and on-vehicle machines people can obtain vehicle running state by the data-interface
(other driving states information such as speed, acceleration, real-time direction information, steering wheel special project angle information, i.e. the second class are special for information
Levy parameter)；Camera is installed on on-vehicle machines people, the direct picture of the driver can be acquired in real time by camera.
Preferably, in the present embodiment, the human eye in the facial image positions, and identifies the human eye
State, specifically include: by the facial image input human face characteristic point return convolutional neural networks model, regression forecasting is one
The boundary position of human eye in a down-sampled low-resolution image.For example, in down-sampled low-resolution image, human eye and/
Or mouth can be the image block that resolution ratio is 72*72.In one embodiment, facial image is input to convolutional Neural net
Network, so that obtaining includes two pairs of data for characterizing the two-dimensional coordinate of two eye positions.In another embodiment, by face figure
As being input to convolutional neural networks, so that obtaining includes three pairs of data for characterizing the two-dimensional coordinate of two eyes and mouth position.
Conventional images analysis method can also be used, for example, Pupil Segmentation, SHAPE DETECTION etc. orient the boundary position of human eye, in turn
Determine the state of human eye.The state for determining human eye includes the area of pupil of human, human eye opening and closing size and its variation.For example, returning
The one human eye area area changed, the relative motion of pupil region.It is understood that the state of the human eye specifically includes: institute
State the opening and closing size of human eye and the opening and closing frequency of the human eye.
S104: first kind characteristic parameter and the input of the second category feature parameter are included into default dangerous driving tagsort probability
The artificial nerve network model of value obtains dangerous driving tagsort probability value, judges that dangerous driving tagsort probability value is
It is no to be greater than default dangerous driving tagsort probability value, dangerous driving is preset determining that dangerous driving tagsort probability value is greater than
When tagsort probability value, determine that driver is in dangerous driving state and sends prompting message.
In order to make it easy to understand, below to the safe driving of vehicle detection method in present embodiment based on artificial neural network
Be specifically described: 1, vehicle: equipped with unified vehicle network data interface, robot can obtain vehicle by the data-interface
Running condition information (other driving states information such as speed, acceleration, real-time direction information, steering wheel special project angle information) 2,
User: user's expression (morbid state, tired, be fascinated by, the expression informations such as absent-minded) is obtained by robot camera；Obtain user's limbs
Status information；The detection of above-mentioned status information can be easily realized based on existing deep learning neural network recognization detection algorithm
Identification and higher level semantic analysis 3, server end: each new engine people's equipment is loaded with universal detection model, should
Server is used to handle the true driving history data of individual subscriber, and training generates personalized safe driving detection model and a
People drives driving archives 4, robot: there is certain calculation power to realize that the data processing-of local side is substantially carried out at video image
Reason and other driving states data；Voice interactive function-plays music by car networking system, opens air-conditioning, interior light, rain
Brush etc.；Vision system: driver's expression, limbs status information are obtained, while people-car interaction can be carried out by gesture identification；Together
When built-in highly sensitive acceleration sensor vehicle perceived with this accelerate and the data such as lateral drift.
Non-security driving condition: 1, eye muscle relaxation, frequent blinking, eyelid open amplitude and become smaller and close one's eyes for a long time；
2, it is difficult to keep head normal attitude, is always looked smug or conceited；3, it is still drank after a night state；4, speed wobble；5, steering wheel angle becomes
Change frequent；6, direction of traffic control is bad, and there are lateral drifts etc. for driving locus.
Data processing model: versatility neural network carries out three classification-dangerous driving-poor risk-safety.
(1) model training: acquisition training dataset: it includes but is not limited to blink, mood table that this model training, which inputs parameter,
Feelings, whens watching the traversing state of state, car speed, vehicle, wheel steering angle, eyes slackness, rate of acceleration change attentively etc.
Ordinal number evidence；Collecting training data normal driving marking data, poor risk driving driving marking data, be in extreme danger data of giving a mark
Three classes；The sampling of each characteristic parameter (refers to a certain characteristic parameter in a certain period of time to peace by the sampling of optimal time window principle
The influence conspicuousness highest of full driving condition: if watching area deviating road visual range in a people 2s, causes danger
Probability can be multiplied, therefore the data are acquired as unit of two seconds；If steering wheel adjusts steering frequency and angle in 20 seconds
It is relatively low or higher, then illustrate that the driver attention does not concentrate, control wheel grip reduces, and dangerous probability occurs and increases), therefore
Time interval, that is, time window of different characteristic parameter samplings is different, by the parameter optimization scheme, improves system dangerous detection
Conspicuousness；Data are synchronous: all kinds of characteristic parameters of acquisition being synchronized in the time domain and (are referred to when driver's blink is frequent
Between in section the expression of corresponding driver, vehicle driving states significant change can also occur, have between each characteristic parameter very high
Correlation, it is therefore desirable to carry out data synchronization processing in the time domain)；Data slicer and normalized: by the danger after synchronizing
Dangerous state drives each characteristic parameter and carries out slicing treatment according to optimal time window, simultaneously because the value range of different characteristic parameter
It is inconsistent, it is therefore desirable to which that input data is normalized；Data slicer point (T, T+ △ Tn), △ Tn are n-th of feature
The optimal time window size of parameter；Data normalization: since above-mentioned data time window size is different, the spy of every class is caused
It is different to levy parameter dimensions, it is therefore desirable to data in addition to being standardized after normalized, present embodiment by with
Unit segment of the time window size of the longest characteristic parameter of time window as input data；With the shortest data of time window
The step-length that is updated as each input data of time window size；Characteristic parameter tentatively merges: obtaining the spy of continuous data slice
Levy the value of vector (t, Xn)-a certain moment a certain characteristic parameter；First high relevance data is tentatively melted into neural network is crossed
Conjunction-Sn (frequency of wink and human face expression marking are highly relevant)；Then the feature vector of the output of the neural network Sn is defeated
Enter subsequent neural network to be trained, until model is restrained.
(2) test: vehicle-mounted camera obtains driver's face image, obtains face characteristic point data and background server number
According to verifying matching is carried out, the corresponding safe driving detection model of the driver is loaded if successful match；If non-registered personnel
Driving then loads general detection model, and is automatically performed new user's registration, which is stored in server, is established personal
Drive archives；On-vehicle machines people obtains vehicle running state parameter, driver status parameter and is input in detection model, if defeated
Result is that dangerous fatigue driving then reminds driver by modes such as acousto-optic-electrics out, and silent status is in if safety；Upload is driven
The person's of sailing current road segment running data carries out individuation data training study；Above-mentioned time series data synchronizes place by timestamp
Reason.
In terms of existing technologies, acquisition history driving procedure is for characterizing driver status for embodiment of the present invention
First kind feature samples parameter, the second category feature sample parameter for characterizing vehicle running state, with artificial neural network
Prime neural network in model carries out Fusion Features to these two types of feature samples parameters from two information sources respectively, then will
The first kind feature samples parameter and the second category feature sample parameter after Fusion Features input the rear class nerve net
In network, to realize the training to artificial nerve network model, the artificial nerve network model after training is integrated and is examined
The first kind feature samples parameter and the second category feature sample parameter of amount reflection dangerous driving, so as to comprehensive, objective, accurate
Measurement driver driving condition, the default dangerous driving tagsort probability value enabled overcomes based on unitary class
Other feature or single piece of information source determine driver whether the limitation of dangerous driving, reduce its false detection rate and omission factor, improve
The reliability and accuracy rate of dangerous driving evaluation；By obtain the driver in real time in the current driving procedure of vehicle
A kind of characteristic parameter and the second category feature parameter, then be inputted in the artificial nerve network model, it is special to obtain dangerous driving
Class probability value is levied, and is determining that it is general that the dangerous driving tagsort probability value is greater than the default dangerous driving tagsort
When rate value, determine that the driver is in dangerous driving state and sends prompting message, so as to prevent peril,
Protection is provided for driver safety trip.
Second embodiment of the present invention is related to a kind of safe driving of vehicle detection method based on artificial neural network, this
Embodiment is that further improvement has been done on the basis of first embodiment, is specifically theed improvement is that, in this embodiment party
In formula: being joined respectively to the first kind feature samples parameter and the second category feature sample using the prime neural network
Number carries out before Fusion Features, further includes: to the human eye number of winks, the eye slackness, the speed for driving vehicle
Degree change frequency, the gazing direction of human eyes, the steering wheel angle variable, the acceleration change number carry out in the time domain
Data synchronization processing；The human eye number of winks after data synchronization processing, the eye slackness, the human eye are infused respectively
Apparent direction, velocity variations number, the steering wheel angle variable, the acceleration change number for driving vehicle are returned
One change processing；Respectively to the human eye number of winks after normalized, the eye slackness, the human eye side of watching attentively
Data mark is carried out to, the velocity variations number of vehicle, the steering wheel angle variable, the acceleration change number of driving
Quasi-ization processing.By special to the first kind feature samples parameter and second class respectively using the prime neural network
It levies before sample parameter carries out Fusion Features and data processing is carried out to these parameters, so that these feature samples after data processing
Parameter can be more accurate embodiment driver driving condition, further improve the reliability of dangerous driving evaluation and accurate
Rate.It is understood that above-mentioned default dangerous driving tagsort probability value can be (early, middle and late, high according to vehicle running environment
Speed etc.) adjustment in real time is 90% as preset dangerous driving class probability daytime, preset value is adjusted to 70% after morning.
The detailed process of present embodiment is as shown in Figure 2, comprising:
S201: the feature samples parameter of acquisition history driving procedure.
S202: to human eye number of winks, eye slackness, velocity variations number, the gazing direction of human eyes, side for driving vehicle
Data synchronization processing is carried out in the time domain to disk corner variable, acceleration change number.
About step S202, specifically, i.e., all kinds of characteristic parameters of acquisition is synchronized in the time domain, such as driven
Significant change, each feature ginseng can also occur for the expression of corresponding driver, vehicle driving states in member's blink frequent period
There is very high correlation, it is therefore desirable to carry out data synchronization processing in the time domain between number.
S203: respectively to the human eye number of winks after data synchronization processing, eye slackness, gazing direction of human eyes, driving
Velocity variations number, steering wheel angle variable, the acceleration change number of vehicle are normalized.
About step S203, specifically, the precarious position after synchronizing drives each characteristic parameter according to optimal time window
Slicing treatment is carried out, simultaneously because the value range of different characteristic parameter is inconsistent, it is therefore desirable to which normalizing is carried out to input data
Change processing.
S204: respectively to the human eye number of winks after normalized, eye slackness, gazing direction of human eyes, driving vehicle
Velocity variations number, steering wheel angle variable, acceleration change number carry out data normalization processing.
About step S204, specifically, since above-mentioned data time window size is different, the feature of every class is caused
Parameter dimensions are different, it is therefore desirable to data in addition to being standardized after normalized.
S205: based on the first artificial neural network to after data standardization human eye number of winks, eye slackness,
Gazing direction of human eyes carries out Fusion Features, based on the second artificial neural network to the speed of the driving vehicle after data standardization
It spends change frequency, steering wheel angle variable, acceleration change number and carries out Fusion Features, by the first category feature after Fusion Features
In sample parameter and the second category feature sample parameter input rear class neural network, obtain general comprising default dangerous driving tagsort
The artificial nerve network model of rate value.
S206: the first kind characteristic parameter and the second category feature ginseng of driver are obtained in real time in the current driving procedure of vehicle
Number.
S207: inputting artificial nerve network model for first kind characteristic parameter and the second category feature parameter, obtains danger and drives
Tagsort probability value is sailed, judges whether dangerous driving tagsort probability value is greater than default dangerous driving tagsort probability
Value determines at driver when determining that dangerous driving tagsort probability value is greater than default dangerous driving tagsort probability value
In dangerous driving state and send prompting message.
Step S201, step S206 to step S207 in present embodiment and the step S101 in first embodiment,
Step S103 is similar to step S104, and in order to avoid repeating, details are not described herein again.
In terms of existing technologies, acquisition history driving procedure is for characterizing driver status for embodiment of the present invention
First kind feature samples parameter, the second category feature sample parameter for characterizing vehicle running state, with artificial neural network
Prime neural network in model carries out Fusion Features to these two types of feature samples parameters from two information sources respectively, then will
The first kind feature samples parameter and the second category feature sample parameter after Fusion Features input the rear class nerve net
In network, to realize the training to artificial nerve network model, the artificial nerve network model after training is integrated and is examined
The first kind feature samples parameter and the second category feature sample parameter of amount reflection dangerous driving, so as to comprehensive, objective, accurate
Measurement driver driving condition, the default dangerous driving tagsort probability value enabled overcomes based on unitary class
Other feature or single piece of information source determine driver whether the limitation of dangerous driving, reduce its false detection rate and omission factor, improve
The reliability and accuracy rate of dangerous driving evaluation；By obtain the driver in real time in the current driving procedure of vehicle
A kind of characteristic parameter and the second category feature parameter, then be inputted in the artificial nerve network model, it is special to obtain dangerous driving
Class probability value is levied, and is determining that it is general that the dangerous driving tagsort probability value is greater than the default dangerous driving tagsort
When rate value, determine that the driver is in dangerous driving state and sends prompting message, so as to prevent peril,
Protection is provided for driver safety trip.
Third embodiment of the invention is related to a kind of safe driving of vehicle detection device based on artificial neural network, such as schemes
It is shown, comprising:
At least one processor 301；And
With the memory 302 of at least one processor 301 communication connection；Wherein,
Memory 302 is stored with the instruction that can be executed by least one processor 301, instructs by least one processor
301 execute, so that at least one processor 301 is able to carry out the above-mentioned safe driving of vehicle detection side based on artificial neural network
Method.
Wherein, memory 302 is connected with processor 301 using bus mode, and bus may include any number of interconnection
Bus and bridge, bus is by one or more processors 301 together with the various circuit connections of memory 302.Bus may be used also
With by such as peripheral equipment, voltage-stablizer, together with various other circuit connections of management circuit or the like, these are all
It is known in the art, therefore, it will not be further described herein.Bus interface provides between bus and transceiver
Interface.Transceiver can be an element, be also possible to multiple element, such as multiple receivers and transmitter, provide for
The unit communicated on transmission medium with various other devices.The data handled through processor 301 pass through antenna on the radio medium
It is transmitted, further, antenna also receives data and transfers data to processor 301.
Processor 301 is responsible for management bus and common processing, can also provide various functions, including timing, periphery connects
Mouthful, voltage adjusting, power management and other control functions.And memory 302 can be used for storage processor 301 and execute
Used data when operation.
Four embodiment of the invention is related to a kind of computer readable storage medium, is stored with computer program.Computer
Above method embodiment is realized when program is executed by processor.
That is, it will be understood by those skilled in the art that implement the method for the above embodiments be can be with
Relevant hardware is instructed to complete by program, which is stored in a storage medium, including some instructions are to make
It obtains an equipment (can be single-chip microcontroller, chip etc.) or processor (processor) executes side described in each embodiment of the application
The all or part of the steps of method.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only
Memory), random access memory (RAM, Random Access Memory), magnetic or disk etc. are various can store journey
The medium of sequence code.
It will be understood by those skilled in the art that the respective embodiments described above are to realize specific embodiments of the present invention,
And in practical applications, can to it, various changes can be made in the form and details, without departing from the spirit and scope of the present invention.
Claims (10)
1. a kind of safe driving of vehicle detection method based on artificial neural network characterized by comprising
Acquire the feature samples parameter of history driving procedure, wherein the feature samples parameter includes using in history driving data
Join in the first kind feature samples parameter of characterization driver status, the second category feature sample for characterizing vehicle running state
Number；
The artificial nerve network model including prime neural network and rear class neural network is established, and utilizes the dangerous driving
Artificial nerve network model described in feature samples parameter training, wherein using the prime neural network respectively to described first
Category feature sample parameter and the second category feature sample parameter carry out Fusion Features, and the first kind after Fusion Features is special
It levies sample parameter and the second category feature sample parameter inputs training in the rear class neural network, obtain comprising default danger
The artificial nerve network model of driving characteristics class probability value；
Obtain the first kind characteristic parameter and the second category feature of the driver and vehicle in real time in the current driving procedure of vehicle
Parameter；
The first kind characteristic parameter and the second category feature parameter input is described comprising presetting dangerous driving tagsort
The artificial nerve network model of probability value obtains dangerous driving tagsort probability value, judges the dangerous driving tagsort
Whether probability value is greater than the default dangerous driving tagsort probability value, is determining dangerous driving tagsort probability value
When greater than the default dangerous driving tagsort probability value, determine that the driver is in dangerous driving state and sends prompting
Information.
2. the safe driving of vehicle detection method according to claim 1 based on artificial neural network, which is characterized in that institute
State the first kind characteristic parameter for obtaining the driver and vehicle in real time in the current driving procedure of vehicle and the second category feature ginseng
Number, specifically includes:
The direct picture of the driver is acquired in real time；
Face datection is carried out to obtain the facial image of the driver to the direct picture of acquisition；
Human eye in the facial image is positioned, identifies the state of the human eye to obtain according to the state of the human eye
The first kind characteristic parameter；
It is connect in real time with the vehicle communication, to obtain the second category feature parameter.
3. the safe driving of vehicle detection method according to claim 2 based on artificial neural network, which is characterized in that institute
It states and the human eye in the facial image is positioned, identify the state of the human eye, specifically include:
Facial image input human face characteristic point is returned into convolutional neural networks model, regression forecasting is down-sampled low at one
The boundary position of human eye in image in different resolution.
4. the safe driving of vehicle detection method according to claim 2 based on artificial neural network, which is characterized in that institute
The state for stating human eye specifically includes: the opening and closing size of the human eye and the opening and closing frequency of the human eye.
5. the safe driving of vehicle detection method according to claim 1 based on artificial neural network, which is characterized in that institute
First kind feature samples parameter is stated as facial behavioural characteristic sample parameter, the second category feature sample parameter is vehicle driving sample
This parameter；It is described to establish the artificial nerve network model including prime neural network and rear class neural network, it specifically includes:
The first artificial neural network for merging the facial behavioural characteristic sample parameter is constructed, is constructed for merging the vehicle
Traveling sample parameter the second artificial neural network, by first artificial neural network and second artificial neural network
As the prime neural network, establishing includes first artificial neural network, second artificial neural network and described
The artificial nerve network model of rear class neural network.
6. the safe driving of vehicle detection method according to claim 5 based on artificial neural network, which is characterized in that institute
It states first kind feature samples parameter to specifically include: human eye number of winks in first time interval, in the second time interval
Eye slackness and the gazing direction of human eyes in third time interval；The second category feature sample parameter specifically includes:
Velocity variations number, the steering wheel angle of rotation in the 5th time interval of the driving vehicle in the 4th time interval
Degree, the acceleration change number in the 6th time interval.
7. the safe driving of vehicle detection method according to claim 6 based on artificial neural network, which is characterized in that
The first kind feature samples parameter and the second category feature sample parameter are carried out respectively using the prime neural network
Before Fusion Features, further includes:
The human eye number of winks, the eye slackness, velocity variations number, the human eye for driving vehicle are watched attentively
Direction, the steering wheel angle variable, the acceleration change number carry out data synchronization processing in the time domain；
Respectively to the human eye number of winks after data synchronization processing, the eye slackness, the gazing direction of human eyes, institute
The velocity variations number for driving vehicle, the steering wheel angle variable, the acceleration change number or characterize data is stated to carry out
Normalized；
Respectively to the human eye number of winks, the eye slackness, the gazing direction of human eyes, described after normalized
The velocity variations number, the steering wheel angle variable, the acceleration change number for driving vehicle carry out at data normalization
Reason；
It is described to utilize the prime neural network respectively to the first kind feature samples parameter and the second category feature sample
Parameter carries out Fusion Features, specifically includes:
Based on first artificial neural network to the human eye number of winks after data standardization, eye relaxation
Degree, the gazing direction of human eyes carry out Fusion Features, based on second artificial neural network to data standardization after
The velocity variations number for driving vehicle, the steering wheel angle variable, the acceleration change number carry out Fusion Features.
8. the safe driving of vehicle detection method according to claim 1 based on artificial neural network, which is characterized in that institute
Acquisition dangerous driving feature samples parameter is stated, is specifically included:
Obtain vehicle current driver's history driving data in facial cybernetics control number and vehicle operation characteristic parameter with
As the first kind feature samples parameter and second feature sample parameter of the dangerous driving feature samples parameter, or
Facial cybernetics control number and vehicle operation characteristic parameter in the history driving data of acquisition history driver is to make
For the dangerous driving feature samples parameter.
9. a kind of safe driving of vehicle detection device based on artificial neural network characterized by comprising
At least one processor；And
The memory being connect at least one described processor communication；Wherein,
The memory is stored with the instruction that can be executed by least one described processor, and described instruction is by described at least one
It manages device to execute, so that at least one described processor is able to carry out as described in any item of the claim 1 to 8 based on artificial mind
Safe driving of vehicle detection method through network.
10. a kind of computer readable storage medium, is stored with computer program, which is characterized in that the computer program is located
Reason device realizes the safe driving of vehicle detection side described in any item of the claim 1 to 8 based on artificial neural network when executing
Method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811209118.XA CN109460780A (en) | 2018-10-17 | 2018-10-17 | Safe driving of vehicle detection method, device and the storage medium of artificial neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811209118.XA CN109460780A (en) | 2018-10-17 | 2018-10-17 | Safe driving of vehicle detection method, device and the storage medium of artificial neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109460780A true CN109460780A (en) | 2019-03-12 |
Family
ID=65607805
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811209118.XA Pending CN109460780A (en) | 2018-10-17 | 2018-10-17 | Safe driving of vehicle detection method, device and the storage medium of artificial neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109460780A (en) |
Cited By (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109934171A (en) * | 2019-03-14 | 2019-06-25 | 合肥工业大学 | Driver's passiveness driving condition online awareness method based on layered network model |
| CN110143202A (en) * | 2019-04-09 | 2019-08-20 | 南京交通职业技术学院 | A kind of dangerous driving identification and method for early warning and system |
| CN110188655A (en) * | 2019-05-27 | 2019-08-30 | 上海蔚来汽车有限公司 | Driving condition evaluation method, system and computer storage medium |
| CN110427871A (en) * | 2019-07-31 | 2019-11-08 | 长安大学 | A kind of method for detecting fatigue driving based on computer vision |
| CN110443219A (en) * | 2019-08-13 | 2019-11-12 | 树根互联技术有限公司 | Driving behavior method for detecting abnormality, device and industrial equipment |
| CN110516658A (en) * | 2019-09-06 | 2019-11-29 | 山东理工大学 | A kind of recognizer design of driver's mood based on face-image and vehicle operating information |
| CN110667593A (en) * | 2019-09-06 | 2020-01-10 | 中国平安财产保险股份有限公司 | Driving reminding method, device and equipment based on deep learning and storage medium |
| CN110765980A (en) * | 2019-11-05 | 2020-02-07 | 中国人民解放军国防科技大学 | Abnormal driving detection method and device |
| CN111387940A (en) * | 2020-03-12 | 2020-07-10 | 泰康保险集团股份有限公司 | Fatigue detection method and device and electronic equipment |
| CN111523766A (en) * | 2020-03-27 | 2020-08-11 | 中国平安财产保险股份有限公司 | Driving risk assessment method and device, electronic equipment and readable storage medium |
| WO2020186883A1 (en) * | 2019-03-18 | 2020-09-24 | 北京市商汤科技开发有限公司 | Methods, devices and apparatuses for gaze area detection and neural network training |
| CN112114671A (en) * | 2020-09-22 | 2020-12-22 | 上海汽车集团股份有限公司 | Human-vehicle interaction method and device based on human eye sight and storage medium |
| CN112232525A (en) * | 2020-12-15 | 2021-01-15 | 鹏城实验室 | Driving mode characteristic construction and screening method and device and storage medium |
| CN112336349A (en) * | 2020-10-12 | 2021-02-09 | 易显智能科技有限责任公司 | Method and related device for recognizing psychological state of driver |
| CN112396235A (en) * | 2020-11-23 | 2021-02-23 | 浙江天行健智能科技有限公司 | Traffic accident occurrence time prediction modeling method based on eyeball motion tracking |
| CN112489425A (en) * | 2020-11-25 | 2021-03-12 | 平安科技（深圳）有限公司 | Vehicle anti-collision early warning method and device, vehicle-mounted terminal equipment and storage medium |
| CN112597825A (en) * | 2020-12-07 | 2021-04-02 | 深延科技(北京)有限公司 | Driving scene segmentation method and device, electronic equipment and storage medium |
| CN112630496A (en) * | 2020-12-01 | 2021-04-09 | 江苏博沃汽车电子系统有限公司 | Method and device for improving accuracy of current sensor |
| CN112630509A (en) * | 2020-12-01 | 2021-04-09 | 江苏博沃汽车电子系统有限公司 | Method and device for improving sensitivity of voltage sensor |
| CN112798300A (en) * | 2021-02-07 | 2021-05-14 | 柳州龙燊汽车部件有限公司 | Anti-collision detection method and system for automobile compartment |
| CN113569674A (en) * | 2021-07-16 | 2021-10-29 | 深圳昌恩智能股份有限公司 | Driving behavior identification method based on intelligent vehicle-mounted terminal |
| CN113602287A (en) * | 2021-08-13 | 2021-11-05 | 吉林大学 | Man-machine driving system for female drivers with low driving ages |
| CN114103988A (en) * | 2020-08-31 | 2022-03-01 | 奥迪股份公司 | Safety monitoring device, vehicle comprising same, and corresponding method, equipment and medium |
| CN115512511A (en) * | 2021-06-07 | 2022-12-23 | 中移物联网有限公司 | Early warning method, early warning device, mobile terminal and readable storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104207791A (en) * | 2014-08-26 | 2014-12-17 | 江南大学 | Fatigue driving detection method |
| CN105139070A (en) * | 2015-08-27 | 2015-12-09 | 南京信息工程大学 | Fatigue driving evaluation method based on artificial nerve network and evidence theory |
| CN107341468A (en) * | 2017-06-30 | 2017-11-10 | 北京七鑫易维信息技术有限公司 | Driver status recognition methods, device, storage medium and processor |
| CN108021875A (en) * | 2017-11-27 | 2018-05-11 | 上海灵至科技有限公司 | A kind of vehicle driver's personalization fatigue monitoring and method for early warning |
| CN108647708A (en) * | 2018-04-28 | 2018-10-12 | 清华-伯克利深圳学院筹备办公室 | Driver evaluation's method, apparatus, equipment and storage medium |
-
2018
- 2018-10-17 CN CN201811209118.XA patent/CN109460780A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104207791A (en) * | 2014-08-26 | 2014-12-17 | 江南大学 | Fatigue driving detection method |
| CN105139070A (en) * | 2015-08-27 | 2015-12-09 | 南京信息工程大学 | Fatigue driving evaluation method based on artificial nerve network and evidence theory |
| CN107341468A (en) * | 2017-06-30 | 2017-11-10 | 北京七鑫易维信息技术有限公司 | Driver status recognition methods, device, storage medium and processor |
| CN108021875A (en) * | 2017-11-27 | 2018-05-11 | 上海灵至科技有限公司 | A kind of vehicle driver's personalization fatigue monitoring and method for early warning |
| CN108647708A (en) * | 2018-04-28 | 2018-10-12 | 清华-伯克利深圳学院筹备办公室 | Driver evaluation's method, apparatus, equipment and storage medium |
Cited By (34)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109934171B (en) * | 2019-03-14 | 2020-03-17 | 合肥工业大学 | Online perception method for passive driving state of driver based on hierarchical network model |
| CN109934171A (en) * | 2019-03-14 | 2019-06-25 | 合肥工业大学 | Driver's passiveness driving condition online awareness method based on layered network model |
| WO2020186883A1 (en) * | 2019-03-18 | 2020-09-24 | 北京市商汤科技开发有限公司 | Methods, devices and apparatuses for gaze area detection and neural network training |
| CN110143202A (en) * | 2019-04-09 | 2019-08-20 | 南京交通职业技术学院 | A kind of dangerous driving identification and method for early warning and system |
| CN110188655A (en) * | 2019-05-27 | 2019-08-30 | 上海蔚来汽车有限公司 | Driving condition evaluation method, system and computer storage medium |
| CN110427871A (en) * | 2019-07-31 | 2019-11-08 | 长安大学 | A kind of method for detecting fatigue driving based on computer vision |
| CN110427871B (en) * | 2019-07-31 | 2022-10-14 | 长安大学 | Fatigue driving detection method based on computer vision |
| CN110443219A (en) * | 2019-08-13 | 2019-11-12 | 树根互联技术有限公司 | Driving behavior method for detecting abnormality, device and industrial equipment |
| CN110443219B (en) * | 2019-08-13 | 2022-02-11 | 树根互联股份有限公司 | Driving behavior abnormity detection method and device and industrial equipment |
| CN110667593A (en) * | 2019-09-06 | 2020-01-10 | 中国平安财产保险股份有限公司 | Driving reminding method, device and equipment based on deep learning and storage medium |
| CN110516658A (en) * | 2019-09-06 | 2019-11-29 | 山东理工大学 | A kind of recognizer design of driver's mood based on face-image and vehicle operating information |
| CN110765980A (en) * | 2019-11-05 | 2020-02-07 | 中国人民解放军国防科技大学 | Abnormal driving detection method and device |
| CN111387940A (en) * | 2020-03-12 | 2020-07-10 | 泰康保险集团股份有限公司 | Fatigue detection method and device and electronic equipment |
| CN111523766A (en) * | 2020-03-27 | 2020-08-11 | 中国平安财产保险股份有限公司 | Driving risk assessment method and device, electronic equipment and readable storage medium |
| CN111523766B (en) * | 2020-03-27 | 2020-11-13 | 中国平安财产保险股份有限公司 | Driving risk assessment method and device, electronic equipment and readable storage medium |
| CN114103988B (en) * | 2020-08-31 | 2024-04-19 | 奥迪股份公司 | Safety monitoring device, vehicle comprising same, and corresponding method, device and medium |
| CN114103988A (en) * | 2020-08-31 | 2022-03-01 | 奥迪股份公司 | Safety monitoring device, vehicle comprising same, and corresponding method, equipment and medium |
| CN112114671A (en) * | 2020-09-22 | 2020-12-22 | 上海汽车集团股份有限公司 | Human-vehicle interaction method and device based on human eye sight and storage medium |
| CN112336349A (en) * | 2020-10-12 | 2021-02-09 | 易显智能科技有限责任公司 | Method and related device for recognizing psychological state of driver |
| CN112396235A (en) * | 2020-11-23 | 2021-02-23 | 浙江天行健智能科技有限公司 | Traffic accident occurrence time prediction modeling method based on eyeball motion tracking |
| WO2022110737A1 (en) * | 2020-11-25 | 2022-06-02 | 平安科技（深圳）有限公司 | Vehicle anticollision early-warning method and apparatus, vehicle-mounted terminal device, and storage medium |
| CN112489425A (en) * | 2020-11-25 | 2021-03-12 | 平安科技（深圳）有限公司 | Vehicle anti-collision early warning method and device, vehicle-mounted terminal equipment and storage medium |
| CN112630509B (en) * | 2020-12-01 | 2023-03-10 | 江苏博沃汽车电子系统有限公司 | Method and device for improving sensitivity of voltage sensor |
| CN112630496B (en) * | 2020-12-01 | 2023-03-10 | 江苏博沃汽车电子系统有限公司 | Method and device for improving accuracy of current sensor |
| CN112630496A (en) * | 2020-12-01 | 2021-04-09 | 江苏博沃汽车电子系统有限公司 | Method and device for improving accuracy of current sensor |
| CN112630509A (en) * | 2020-12-01 | 2021-04-09 | 江苏博沃汽车电子系统有限公司 | Method and device for improving sensitivity of voltage sensor |
| CN112597825A (en) * | 2020-12-07 | 2021-04-02 | 深延科技(北京)有限公司 | Driving scene segmentation method and device, electronic equipment and storage medium |
| CN112232525B (en) * | 2020-12-15 | 2021-07-13 | 鹏城实验室 | Driving mode characteristic construction and screening method and device and storage medium |
| CN112232525A (en) * | 2020-12-15 | 2021-01-15 | 鹏城实验室 | Driving mode characteristic construction and screening method and device and storage medium |
| CN112798300A (en) * | 2021-02-07 | 2021-05-14 | 柳州龙燊汽车部件有限公司 | Anti-collision detection method and system for automobile compartment |
| CN115512511A (en) * | 2021-06-07 | 2022-12-23 | 中移物联网有限公司 | Early warning method, early warning device, mobile terminal and readable storage medium |
| CN113569674A (en) * | 2021-07-16 | 2021-10-29 | 深圳昌恩智能股份有限公司 | Driving behavior identification method based on intelligent vehicle-mounted terminal |
| CN113602287A (en) * | 2021-08-13 | 2021-11-05 | 吉林大学 | Man-machine driving system for female drivers with low driving ages |
| CN113602287B (en) * | 2021-08-13 | 2024-01-26 | 吉林大学 | Man-machine co-driving system for drivers with low driving ages |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109460780A (en) | Safe driving of vehicle detection method, device and the storage medium of artificial neural network | |
| Zhang et al. | A new real-time eye tracking based on nonlinear unscented Kalman filter for monitoring driver fatigue | |
| KR102297162B1 (en) | Vehicle control methods and systems, in-vehicle intelligent systems, electronic devices, media | |
| JP7273031B2 (en) | Information processing device, mobile device, information processing system and method, and program | |
| EP3712028A1 (en) | Information processing device and information processing method | |
| Ji et al. | Fatigue state detection based on multi-index fusion and state recognition network | |
| CN103839379B (en) | Automobile and driver fatigue early warning detecting method and system for automobile | |
| CN107657236A (en) | Vehicle security drive method for early warning and vehicle-mounted early warning system | |
| US11783600B2 (en) | Adaptive monitoring of a vehicle using a camera | |
| CN202257856U (en) | Driver fatigue-driving monitoring device | |
| CN106571015A (en) | Driving behavior data collection method based on Internet | |
| CN105354988A (en) | Driver fatigue driving detection system based on machine vision and detection method | |
| EP3465532A1 (en) | Control device, system and method for determining the perceptual load of a visual and dynamic driving scene | |
| Tang et al. | Real-time image-based driver fatigue detection and monitoring system for monitoring driver vigilance | |
| CN113743471B (en) | Driving evaluation method and system | |
| CN109664894A (en) | Fatigue driving safety pre-warning system based on multi-source heterogeneous data perception | |
| Zhou et al. | Driving behavior prediction considering cognitive prior and driving context | |
| CN109886148A (en) | A kind of driver's active warning system and device based on recognition of face | |
| CN110213720A (en) | Unexpected prevention method in mobile phone use process based on user behavior analysis | |
| Hofbauer et al. | Measuring driver situation awareness using region-of-interest prediction and eye tracking | |
| Shaily et al. | Smart driver monitoring system | |
| US11685399B2 (en) | Adjusting driving pattern of autonomous vehicle | |
| CN110803170A (en) | Driving assistance system with intelligent user interface | |
| Wathiq et al. | Driver safety approach using efficient image processing algorithms for driver distraction detection and alerting | |
| KR20110065304A (en) | System and method of drowsy driving recognition based on the personalized template of a driver |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190312 |
|
| RJ01 | Rejection of invention patent application after publication |