CN108647708A - Driver evaluation's method, apparatus, equipment and storage medium - Google Patents
Driver evaluation's method, apparatus, equipment and storage medium Download PDFInfo
- Publication number
- CN108647708A CN108647708A CN201810398792.0A CN201810398792A CN108647708A CN 108647708 A CN108647708 A CN 108647708A CN 201810398792 A CN201810398792 A CN 201810398792A CN 108647708 A CN108647708 A CN 108647708A
- Authority
- CN
- China
- Prior art keywords
- driver
- driving
- driving behavior
- behavior
- image
- 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
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- 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
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Traffic Control Systems (AREA)
Abstract
The embodiment of the invention discloses a kind of driver evaluation's method, apparatus, equipment and storage medium, this method to include：Obtain the driving image of driver；Driving image is analyzed to determine driver identity by human face discriminating device, wherein human face discriminating device includes a variety of human face recognition models, and the face recognition result that most face recognition algorithms are exported is as final result；Driving behavior is determined according to driver identity or driving image, and driving behavior and driver identity are associated；Driver is assessed according to driving behavior, assessment is including at least one kind in driving behavior assessment or performance evaluation.It solves the problems, such as that existing onboard system is unable to specification driver's driving behavior, reaches the technique effect for improving vehicle drive safety by the driving behavior of specification driver.
Description
Technical field
The present embodiments relate to a kind of image procossing more particularly to driver evaluation's method, apparatus, equipment and storages to be situated between
Matter.
Background technology
In order to ensure traffic safety, there are many onboard systems to pass through acquisition to vehicle and road information and prison at present
It controls to reduce traffic accident.Such as：Existing automobile data recorder can by record the speed of vehicle, driving time with
And the information of outside road environmental correclation prevents traffic accident；GPS navigator is then provided in conjunction with vehicle in the position of road in real time
The prompt of " front black spot please drive with caution ".These safeguards can reduce traffic accident to a certain extent
Generation, cannot but improve driver safe driving consciousness, so that driver is passed through what specification driving behavior avoided traffic accident
Occur.
Invention content
A kind of driver evaluation's method, apparatus of offer of the embodiment of the present invention, equipment and storage medium, solve existing vehicle-mounted
System is unable to the problem of specification driver's driving behavior, reaches and improves vehicle drive peace by the driving behavior of specification driver
Full technique effect.
In a first aspect, an embodiment of the present invention provides a kind of driver evaluation's methods, including：
Obtain the driving image of driver；
The driving image is analyzed to determine driver identity by human face discriminating device, wherein the face
Discriminating gear includes a variety of human face recognition models based on different face recognition algorithms, and most human face recognition models are exported
Face recognition result is as final result；
Driving behavior is determined according to the driver identity or the driving image, and the driving behavior is driven with described
The person's of sailing identity is associated；
Driver is assessed according to the driving behavior, the assessment is commented including at least driving behavior assessment or performance
One kind in estimating.
Further, the human face discriminating device includes odd number kind human face recognition model；
When the result of all people's face identification model output is all different, by the corresponding face of default human face recognition model
Recognition result is exported as final result.
Further, the human face discriminating device includes EignFace algorithms, FisherFace algorithms and LBP algorithms.
Further, driving behavior is determined according to the driver identity, including：
It is determined according to the duration of the driver identity and drives duration, the driving duration includes at least current drive
Duration, the same day accumulative one driven in duration, month to date driving duration and driving expired times.
Further, driving behavior is determined according to the driving image, and by the driving behavior and the driver identity
It is associated, including：
Based on preset direction Gradient Features extraction algorithm, the histograms of oriented gradients feature of driving image is obtained；
The histograms of oriented gradients feature is analyzed by the SVM classifier trained, to determine that driver whether there is
Mobile phone usage behavior, the SVM classifier include at least one kind in local SVM classifier or global SVM classifier；
When the driver is there are when mobile phone usage behavior, the mobile phone usage behavior and the driver identity are carried out
Association.
Further, driving behavior is determined according to the driving image, and by the driving behavior and the driver identity
It is associated, including：
Multiple characteristic points in the driving image are determined by vertical-horizontal integral projection method；
Seek the binary image of driving image；
According to the magnitude relationship of the grey scale pixel value of the characteristic point and default gray threshold, determine whether driver wears
Safety belt；
When the non-wear safety belt of the driver, by the driving behavior of non-wear safety belt and the driver identity into
Row association.
Further, driver is assessed in the driving behavior determined by the basis, when the assessment includes driving
When sailing behavior evaluation, further include：
When detecting violation driving behavior, corresponding voice prompt is exported, wherein the violation driving behavior includes super
When driving behavior, mobile phone usage behavior and non-wear safety belt behavior.
Second aspect, the embodiment of the present invention additionally provide a kind of driver evaluation's device, including：
Driving image acquisition module, the driving image for obtaining driver；
Driver identity determining module is analyzed the driving image for passing through human face discriminating device and is driven with determining
The person's of sailing identity, wherein the human face discriminating device includes a variety of human face recognition models based on different face recognition algorithms, and will
The face recognition result of most human face recognition model outputs is as final result；
Driving behavior determining module, for determining driving behavior according to the driver identity or the driving image, and
The driving behavior and the driver identity are associated；
Evaluation module, for being assessed driver according to the driving behavior, the assessment, which includes at least, drives row
For one kind in assessment or performance evaluation.
Further, further include model module, the model module is used for memory of driving behavior evaluation model, the assessment mould
Type is built based on target training data, and the target training data includes the driving behavior data of all drivers of target vehicle
Or the driving behavior data of all drivers of qualification are driven with target vehicle.
The third aspect, the embodiment of the present invention additionally provide a kind of driving behavior assessment equipment, and the equipment includes：
Camera, for obtaining driving image；
One or more processors；
Storage device, for storing one or more programs；
When one or more of programs are executed by one or more of processors so that one or more of processing
Device realizes driver evaluation's method as described in relation to the first aspect.
Fourth aspect, the embodiment of the present invention additionally provide a kind of storage medium including computer executable instructions, special
Sign is that the computer executable instructions by computer processor when being executed for executing driving as described in relation to the first aspect
Member's appraisal procedure.
The technical solution of driver evaluation provided in this embodiment, by human face discriminating device to the driving image of acquisition into
Row analysis is to determine driver identity, wherein human face discriminating device includes that a variety of faces based on different face recognition algorithms are known
Other model, and the face recognition result that most human face recognition models are exported is known as final result compared to single face
Other algorithm is carried out at the same time recognition of face by a variety of face recognition algorithms, and the corresponding face of most face recognition algorithms is known
The face recognition result of other model output can improve the accuracy of recognition of face as final result；According to driver identity
Or driving image determines driving behavior, and driving behavior and driver identity are associated, and can improve the standard of driver evaluation
True property；Driver to be assessed according to driving behavior, assessment includes at least one kind in driving behavior assessment or performance evaluation,
Accurately objectively driver can be assessed, can rapidly and accurately obtain the performance evaluation information of driver, moreover it is possible to is anti-
The only influence of human factor.
Description of the drawings
To describe the technical solutions in the embodiments of the present invention more clearly, make required in being described below to embodiment
Attached drawing does one and simply introduces, it should be apparent that, drawings in the following description are some embodiments of the invention, for this
For the those of ordinary skill of field, without creative efforts, others are can also be obtained according to these attached drawings
Attached drawing.
Fig. 1 is the flow chart for driver evaluation's method that the embodiment of the present invention one provides；
Fig. 2 is the flow chart of human face discriminating device creation method provided by Embodiment 2 of the present invention；
Fig. 3 is the flow chart for driver evaluation's method that the embodiment of the present invention three provides；
Fig. 4 is the flow chart for the method that driving behavior is determined according to driving image that the embodiment of the present invention four provides；
Fig. 5 is the flow chart for the method that driving behavior is determined according to driving image that the embodiment of the present invention five provides；
Fig. 6 is the safety belt image schematic diagram that the embodiment of the present invention five provides；
Fig. 7 is the image schematic diagram for the non-wear safety belt that the embodiment of the present invention five provides；
Fig. 8 is the bianry image for including safety belt that the embodiment of the present invention five provides；
Fig. 9 is the structure diagram for driver evaluation's device that the embodiment of the present invention six provides；
Figure 10 is the structural schematic diagram for the driving behavior assessment equipment that the embodiment of the present invention seven provides.
Specific implementation mode
To make the object, technical solutions and advantages of the present invention clearer, hereinafter with reference to attached in the embodiment of the present invention
Figure, technical scheme of the present invention is clearly and completely described by embodiment, it is clear that described embodiment is the present invention one
Section Example, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art are not doing
Go out the every other embodiment obtained under the premise of creative work, shall fall within the protection scope of the present invention.
Embodiment one
Fig. 1 is the flow chart for driver evaluation's method that the embodiment of the present invention one provides.The technical solution of the present embodiment is suitable
The case where for ride quality or performance of the driving behavior by driver to assess driver, it is particularly suitable for passenger traffic or goods
Transport the driver evaluation of vehicle.This method can be executed by driver evaluation's device provided in an embodiment of the present invention, the device
The mode that software and/or hardware may be used is realized, and is configured and applied in the processor.This method specifically comprises the following steps：
S102, the driving image for obtaining driver.
The driving image of driver is obtained by camera, camera may be disposed on the front windshield of vehicle, image
Head is preferably rotatable camera, and the approximation of the forward direction or face of being capable of Automatic-searching face is positive.The people of driving image shooting
Body range size should be according to image analysis demand, for example, when the purpose of image analysis includes recognition of face, driving image should include
Face.
S104, driving image is analyzed to determine driver identity by human face discriminating device, wherein human face discriminating
Device includes a variety of human face recognition models based on different face recognition algorithms, and the face that most human face recognition models are exported
Recognition result is as final result.
Before driver evaluation, it usually needs determine driver identity, then again close the assessment result of different piece
It is coupled to identified identity.The present embodiment determines driver identity by human face discriminating device, and human face discriminating device includes more
Kind human face recognition model, and the face recognition result that most face recognition algorithms are exported is as final result；Work as all people
The result of face identification model output is when being all different, using the corresponding face recognition result of default human face recognition model as most terminating
Fruit exports.
Optionally, the human face discriminating device in the present embodiment includes odd number kind face recognition algorithms, is sentenced to improve face
Three-type-person's face recognizer can be arranged in the recognition of face speed of other device, human face discriminating device, and specific face recognition algorithms can
To be configured according to actual needs, the preferred three-type-person's face recognizer of the present embodiment is respectively EignFace algorithm (features
Face algorithm), FisherFace algorithms and local binary patterns (LocalBinary Patterns, abbreviation LBP) algorithm.
S106, determine driving behavior according to driver identity or driving image, and by driving behavior and driver identity into
Row association.
Before to driver evaluation, the driving behavior of driver is first obtained, the present embodiment is according to the identity of driver
Or driving image obtains the driving behavior of driver, whether in violation of rules and regulations then analyzes acquired driving behavior.Work as driving behavior
For violation driving behavior when, this violation driving behavior of current driver's identity is added one.Certainly, by violation driving behavior
While adding one, the violation driving behavior parameter, such as time parameter are recorded, and by the whole in current violation driving behavior
Frame image or a certain frame image are stored in setting position, and are associated with the violation driving behavior, in order to violation driving behavior
Statistics and access.
S108, driver is assessed according to driving behavior, assessment includes at least driving behavior assessment or performance evaluation
In one kind.
When being assessed driver according to acquired driving behavior, the present embodiment first judges per in frame driving image
Driving behavior whether in violation of rules and regulations, and count the violation number of each single item driving behavior, pass through the violation number in set period of time
Driver is assessed, for example carries out driving behavior assessment or performance evaluation.
It is understood that since people does one using mobile phone or the action unlocking safety belt or fasten the safety belt usually needs
Several seconds are wanted, therefore can also judge whether the driving behavior of the driving image of predetermined interval is qualified.It is commented to driving behavior
When estimating, it can also will drive row and be divided into grade, determine the grade of acquired driving behavior, then statistics is less than predetermined level
Driving behavior number, driver is assessed as violation driving behavior number, then by violation driving behavior number.
The technical solution of driver evaluation provided in this embodiment, by human face discriminating device to the driving image of acquisition into
Row analysis is to determine driver identity, wherein human face discriminating device includes that a variety of faces based on different face recognition algorithms are known
Other model, and the face recognition result that most human face recognition models are exported is known as final result compared to single face
Other algorithm is carried out at the same time recognition of face by a variety of face recognition algorithms, and the corresponding face of most face recognition algorithms is known
The face recognition result of other model output can improve the accuracy of recognition of face as final result；According to driver identity
Or driving image determines driving behavior, and driving behavior and driver identity are associated, and can improve the standard of driver evaluation
True property；Driver to be assessed according to driving behavior, assessment includes at least one kind in driving behavior assessment or performance evaluation,
Accurately objectively driver can be assessed, can rapidly and accurately obtain the performance evaluation information of driver, moreover it is possible to is anti-
The only influence of human factor.
Embodiment two
Fig. 2 is the flow chart of human face discriminating device creation method provided by Embodiment 2 of the present invention.The embodiment of the present invention exists
Before the driving image of the acquisition driver of above-described embodiment, the step of increasing human face discriminating device creation method.
S1011, the driving image for being used for model training is obtained, and training sample data is determined according to driving image.
In order to improve the stability of human face recognition model and the accuracy of recognition result, the present embodiment obtains in all its bearings
Training set sample, such as the Static Background outside vehicle window, different illumination intensity (such as daytime and night), different expressions, difference are matched
It adorns (such as glasses, ear nail) etc..In addition, in order to reduce the data volume of model training, and improve the stability of model, this implementation
Example obtains the face part of driving image by automatic or manual mode, and the part such as the hair of driver, lower jaw is removed, only
Using the face part of driver's facial image as training sample data.
S1012, the corresponding human face recognition model of each face recognition algorithms is trained using training sample data.
The face recognition algorithms of the present embodiment preferably include EignFace algorithms, FisherFace algorithms and LBP algorithms, adopt
With human face recognition model of the training sample data training based on EignFace algorithms, the recognition of face based on FisherFace algorithms
Model and human face recognition model based on LBP algorithms.In above-mentioned human face recognition model training, identical training sample can be used
Data can also use different training sample data.
Same face recognition result that S1013 exports most human face recognition models or default human face recognition model
Face recognition result is exported as final result.
In order to improve human face discriminating device face recognition result accuracy, the present embodiment be its be provided with multiple faces
Identification model, and the same face recognition result that most human face recognition models are exported is as final result, when all people's face
When the result of identification model output is all different, the result of the corresponding human face recognition model output of default face recognition algorithms is made
For final result.
In the embodiment of the present invention, by the training sample data of acquisition to the recognition of face mould based on EignFace algorithms
Type, the human face recognition model based on FisherFace algorithms and the human face recognition model based on LBP algorithms are trained respectively, and three
A human face recognition model works independently, finally by the same face of the corresponding human face recognition model output of most face recognition algorithms
Recognition result or the face recognition result of the corresponding human face recognition model output of default face recognition algorithms are as last knot
Fruit substantially increases the accuracy of face recognition result.
Embodiment three
Fig. 3 is the flow chart for driver evaluation's method that the embodiment of the present invention three provides.The embodiment of the present invention is to above-mentioned
Any embodiment determines driving behavior according to driver identity, and driving behavior is associated with driver identity excellent
Change.As shown in figure 3, this method includes：
S102, the driving image for obtaining driver.
S104, driving image is analyzed to determine driver identity by human face discriminating device, wherein human face discriminating
Device includes a variety of human face recognition models, and the face recognition result that most face recognition algorithms are exported is as final result.
S106, driving behavior is determined according to driver identity, and driving behavior and driver identity are associated.
Determine that the driving duration of driver is to determine the fatigue state of driver, it is therefore desirable to by driver and driving
Duration is associated together, and the timing mode of the prior art is usually only applicable to a certain vehicle and there was only the case where driver, when this
When vehicle there are several drivers, the driving duration of every driver, the duration that the present embodiment passes through driver identity can not be determined
It determines and drives duration, optionally, duration, vehicle are driven by the way that in vehicle travel process, the duration of driver identity determines
Transport condition can be capable of the signal of collection vehicle location information or vehicle traveling information come really by GPS navigation system etc.
It is fixed.
In order to improve the safety of vehicle drive, when vehicle single running time is more than preset time, such as at 3 hours,
Prompt message, such as information of voice prompt are then sent out, prompts driver to have reached the safe driving time, should rest.
Further, in order to avoid driver practises fraud, or in order to ensure the good state of mind of driver, the present embodiment
It is also provided with the accumulative driving duration of whole day, that is, seeks all driving durations the sum of of each driver in one day, is driven when accumulative
Duration reaches preset duration, such as at 8 hours, then sends out prompt message, such as information of voice prompt, driver is prompted to have reached
The safe driving time should rest.Preferably, after previous driving duration reaches preset time, preset time ability need to be spaced
New driving is carried out, when carrying out second of driving in preset time, then prompt message is sent out, prompts driver that should carry out fully
Rest carries out new driving route again.
The performance appraisal of employee can obtain driver's every month or every through this embodiment usually as unit of the moon or year
The driving duration information in year.In order to save Installed System Memory, driving time-out record can be only preserved, convenient for the same of driver evaluation
When, and saving Installed System Memory can be dripped significantly, reduce system cost.
The violation of driver drives in order to prevent, such as fatigue driving, when driver's appearance driving time-out, and continues to drive
When, other than continuing playing alert tones, the information of vehicles and driver information are also sent to remote monitoring platform, remote monitoring
Platform can be fleet's maneuvering platform.
S108, driver is assessed according to driving behavior, assessment includes at least driving behavior assessment or performance evaluation
In one kind.
The present embodiment determines the driving duration of driver by the duration of driver identity, and then determines driver's
Single drives the accumulative driving duration of duration, whole day and driving expired times etc., realizes and accurately and rapidly determines driver's
Fatigue driving degree reduces the danger that driver is brought due to fatigue driving without human intervention.
Example IV
Fig. 4 is the flow chart for the method that driving behavior is determined according to driving image that the embodiment of the present invention four provides.This hair
Bright embodiment be driving behavior determined according to driving image to above-mentioned any embodiment, and by driving behavior and driver identity into
The associated optimization of row.As shown in figure 4, this method includes：
S1061, it is based on preset direction Gradient Features extraction algorithm, obtains the histograms of oriented gradients feature of driving image.
S1062, the SVM classifier analysis directions histogram of gradients feature by having trained, to determine whether driver deposits
In mobile phone usage behavior, SVM classifier includes at least one kind in local SVM classifier or global SVM classifier.
The histograms of oriented gradients feature of acquisition is input to the support vector machines (Support trained
VectorMachine, abbreviation SVM) grader, determine that driver uses row with the presence or absence of mobile phone by the SVM classifier trained
For.In order to improve the accuracy of SVM classifier, the present embodiment first establishes local classifiers, is then based on local optimum grader
Establish global classification device.
S1063, when driver is there are when mobile phone usage behavior, mobile phone usage behavior is associated with driver identity.
When driver is there are when mobile phone usage behavior, the corresponding mobile phone usage behavior number of the driver identity is added one.
SVM classifier needs to be trained it before use, which is to include：
1, training sample image is obtained
Driving image in this step can be identical or different as the model training data of human face discriminating device, no matter phase
Whether same, these driving images for being used for SVM classifier training preferably include positive sample and negative sample, for example, 400 positive samples
This, 35 negative samples, all sample sizes may be configured as 96 × 128, and positive sample includes the different appearances that driver holds mobile phone communication
Gesture, negative sample are the sectional drawing or grader for the similar driving image using mobile phone behavior that the driver manually chosen makes
The region sectional drawing of error detection in the training process, such as the image that shows area smaller of the mobile phone in driving image.
2, histograms of oriented gradients (Histogram of Oriented Gradient, the abbreviation of extraction training sample
HOG) feature
During using HOG operator extraction driver's mobile phone features, herein first it is 400 big it is small be 96 × 128
Sample image be divided into a large amount of segment (block), if the size of cell factory (cell) be 8 × 8, segment (block) size
8,10,12,14,16 are taken respectively for k × k, wherein k, and segment (block) is by the cell factory structure that 4 sizes are 0.5k × 0.5k
At the moving step length on horizontal and vertical direction is both configured to 2.There is 96/8-1=11 scanning window in horizontal direction, vertically
There is 128/8-1=15 scanning window on direction, thus obtains the HOG feature vectors of a large amount of segments (block), and HOG features
Total dimension of vector is 64 dimensions.
3, the local SVM classifier of HOG features training based on positive negative sample regional area.
The SVM classifier that the HOG feature vectors of each segment are corresponded to one 64 dimension, then each image corresponds to many offices
Classifier.Local classifiers of the present embodiment based on SVM are linear classifier, and the parameter coef values of linear kernel function are set as
0.01。
4, optimal partial grader is chosen from local classifiers.
The discrimination of local classifiers is tested using positive and negative sample image first, has picked out more than half probability
The grader that can correctly identify, as optimal partial grader.
5, classification results of the optimal partial grader to positive negative sample are obtained.
Positive negative sample is detected using local classifiers, if recognition result, which is driver, holds mobile phone, is denoted as 1, such as
Fruit recognition result is that driver does not hold mobile phone, is denoted as -1, obtains classification results of the optimal partial grader to positive negative sample.
6, local classifiers are combined into global classification device.
I-th class training sample can use vector xi=[xi1, xi2, xi3... xiO] T, (i=1,2 ... n) indicate, xqkIt indicates
Under the premise of q local classifiers, two classification results of k-th of grader output are used.yiIt is that the i-th class training sample corresponds to
Classification results.yi∈ { 1, -1 }, y=1 indicate that positive sample, y=-1 indicate negative sample.The classification results that final step obtains
{(xn, yn) be global SVM classifier classification results.
During one frame frame of video flowing plays, multiple dimensioned HOG detections constantly are carried out to the mobile phone of driver, and will
HOG features are input to global SVM classifier, and SVM global classification devices pass through the processing time of about 1.5s, you can identify driver
It is hand-held mobile phone.
SVM classifier in the present embodiment holds mobile phone by whether there is driver in HOG feature recognition driving images
Behavior, recognition speed is fast, accuracy rate is high.
Embodiment five
Fig. 5 is the flow chart for the method that driving behavior is determined according to driving image that the embodiment of the present invention five provides.This hair
Bright embodiment is to determine driving behavior according to driving image to above-mentioned any embodiment, and by driving behavior and driver identity
The optimization being associated.As shown in figure 5, this method includes：
S1071, multiple feature point coordinates in driving image are obtained by vertical-horizontal integral projection method.
In order to judge driver whether wear safety belt, driving image needs to include driver's upper body part, that is, drives
Sail the safe region of all or part when can show that the normal wear safety belt of driver on image.
In order to improve the accuracy rate of safety belt identification, it usually needs first image is pre-processed, such as light compensation etc.,
Then vertical-horizontal integral projection method is used to obtain multiple characteristic points in driving image, characteristic point is located at what safety belt was worn
Position, the feature point number in the present embodiment are chosen as three, the optional distributing positions of three characteristic points include waistband left-hand point,
Waistband right-hand point and shoulder belt point.It is understood that light compensation can carry out light compensation by image processing means,
The contrast of safety belt and background clothes can be improved by improving the light of driving image shooting environmental, to reach light benefit
The effect repaid.
S1072, the binary image for seeking driving image.
It converts current driving image to gray level image, is then based on default gray threshold and converts gray images into two-value
Change image, binary image shown in Figure 7.
S1073, the magnitude relationship according to the gray value and default gray threshold of multiple characteristic points, determine whether driver wears
Wear safety belt.
Grey scale pixel value of the characteristic point in gray level image is determined according to the coordinate of multiple characteristic points, and judging characteristic point
The magnitude relationship of grey scale pixel value and default gray threshold, when the grey scale pixel value of characteristic point is less than default gray threshold, then
Indicate driver's wear safety belt, otherwise, then it represents that driver does not have wear safety belt.
Wherein, the acquisition methods of default gray threshold can be：Acquisition includes the driving image of safety belt and does not include peace
The multipair image of the driving image of full band, such as hundreds of pairs, and aforementioned image preferably obtains under different scenes difference light, such as
Shown in Fig. 6, the gray level image of every width driving image is then sought, binaryzation is carried out to them based on identical gray threshold, so
The root-mean-square deviation for seeking the grey scale pixel value of every a pair of of driving image respective pixel afterwards, using as default gray threshold.
As shown in fig. 7, the gray scale difference value of the characteristic point in the figure is 24.1475, higher than default gray threshold, therefore this is attached
The non-wear safety belt of driver in figure.
As shown in figure 8, the gray scale difference value of the characteristic point in the figure is 23.122, it is less than default gray threshold, therefore this is attached
Driver in figure has worn safety belt.
S1074, when the non-wear safety belt of driver, the driving behavior of non-wear safety belt and driver identity are carried out
Association.
When assessing driver by driving behavior, driver is typically assessed according to violation driving behavior, therefore when driving
When the non-wear safety belt of member, the driving behavior of non-wear safety belt is associated with driver identity, while exporting voice and carrying
Show, prompts driver's wear safety belt.
In addition, in order to improve the contrast to safety belt and background clothes, fluorescent material can be used in the material of safety belt, or
Person is irradiated by LED lamplight carries out light compensation.
The present embodiment by the combination of vertical-horizontal integral projection method and gray scale difference value, determine driver whether safe wearing
Band substantially increases the recognition speed and accuracy rate of safety belt.
Embodiment six
Fig. 9 is the structure diagram for driver evaluation's device that the embodiment of the present invention six provides.The device is above-mentioned for executing
Driver evaluation's method that any embodiment is provided, the device can be by software or hardware realizations.The device includes：
Driving image acquisition module 11, the driving image for obtaining driver；
Driver identity determining module 12 analyzes with determination the driving image for passing through human face discriminating device
Driver identity, wherein the human face discriminating device includes odd number kind face recognition algorithms, and most face recognition algorithms are defeated
The face recognition result gone out is as final result；
Driving behavior determining module 13, for determining driving behavior according to the driver identity or the driving image,
And the driving behavior and the driver identity are associated；
Evaluation module 14, for being assessed driver according to the driving behavior, the assessment, which includes at least, to be driven
One kind in behavior evaluation or performance evaluation.
The technical solution of driving behavior apparatus for evaluating provided in an embodiment of the present invention, by human face discriminating device to acquisition
Driving image is analyzed to determine driver identity, wherein human face discriminating device includes based on different face recognition algorithms
A variety of human face recognition models, and the face recognition result that most human face recognition models are exported is as final result, compared to list
One face recognition algorithms are carried out at the same time recognition of face by a variety of face recognition algorithms, and by most face recognition algorithms pair
The face recognition result for the human face recognition model output answered can improve the accuracy of recognition of face as final result；According to
Driver identity or driving image determine driving behavior, and driving behavior and driver identity are associated, and can improve driving
The accuracy of member's assessment；Driver is assessed according to driving behavior, assessment is commented including at least driving behavior assessment or performance
One kind in estimating can accurately objectively assess driver, can rapidly and accurately obtain the performance evaluation of driver
Information, moreover it is possible to prevent the influence of human factor.
What the executable any embodiment of the present invention of driving behavior apparatus for evaluating that the embodiment of the present invention is provided was provided drives
Behavior evaluation method is sailed, has the corresponding function module of execution method and advantageous effect.
Embodiment seven
Figure 10 is the structural schematic diagram for the driving behavior assessment equipment that the embodiment of the present invention seven provides, as shown in Figure 10, should
Equipment includes camera 100, for obtaining driving image；Processor 101, memory 102, input unit 103 and output dress
Set 104；The quantity of processor 101 can be one or more in equipment, in Figure 10 by taking a processor 101 as an example；In equipment
Processor 101, memory 102, input unit 103 and output device 104 can pass through bus or other modes connection, figure
10 by bus for being connected.
Memory 102 is used as a kind of computer readable storage medium, can be used for storing software program, computer can perform journey
Sequence and module, if the corresponding program instruction/module of driving behavior appraisal procedure in the embodiment of the present invention is (for example, drive figure
As acquisition module 11, driver identity determining module 12, driver identity determining module 13 and evaluation module 14).Processor 101
By running storage software program, instruction and module in the memory 102, to execute the various function application of equipment with
And data processing, that is, realize above-mentioned driving behavior appraisal procedure.
Memory 102 can include mainly storing program area and storage data field, wherein storing program area can store operation system
Application program needed for system, at least one function；Storage data field can be stored uses created data etc. according to terminal.This
Outside, memory 102 may include high-speed random access memory, can also include nonvolatile memory, for example, at least one
Disk memory, flush memory device or other non-volatile solid state memory parts.In some instances, memory 102 can be into one
Step includes the memory remotely located relative to processor 101, these remote memories can pass through network connection to equipment.On
The example for stating network includes but not limited to internet, intranet, LAN, mobile radio communication and combinations thereof.
Input unit 103 can be used for receiving the number or character information of input, and generate with the user setting of equipment with
And the related key signals input of function control.
Output device 104 may include that display screen etc. shows equipment, for example, the display screen of user terminal.
Embodiment eight
The embodiment of the present invention eight also provides a kind of storage medium including computer executable instructions, and the computer can be held
When being executed by computer processor for executing a kind of driving behavior appraisal procedure, this method includes for row instruction：
Obtain the driving image of driver；
The driving image is analyzed to determine driver identity by human face discriminating device, wherein the face
Discriminating gear includes a variety of human face recognition models based on different face recognition algorithms, and most human face recognition models are defeated
The face recognition result gone out is as final result；
Driving behavior is determined according to the driver identity or the driving image, and the driving behavior is driven with described
The person's of sailing identity is associated；
Driver is assessed according to the driving behavior, the assessment is commented including at least driving behavior assessment or performance
One kind in estimating.
Certainly, a kind of storage medium including computer executable instructions that the embodiment of the present invention is provided, computer
The method operation that executable instruction is not limited to the described above, can also be performed the driver that any embodiment of the present invention is provided and comments
Estimate the relevant operation in method.
By the description above with respect to embodiment, it is apparent to those skilled in the art that, the present invention
It can be realized by software and required common hardware, naturally it is also possible to which by hardware realization, but the former is more in many cases
Good embodiment.Based on this understanding, technical scheme of the present invention substantially in other words contributes to the prior art
Part can be expressed in the form of software products, which can be stored in computer readable storage medium
In, such as the floppy disk of computer, read-only memory (Read-Only Memory, abbreviation ROM), random access memory (Random
Access Memory, abbreviation RAM), flash memory (FLASH), hard disk or CD etc., including some instructions are used so that a calculating
Machine equipment (can be personal computer, server or the network equipment etc.) executes the driving described in each embodiment of the present invention
Member's appraisal procedure.
It is worth noting that, in the embodiment of above-mentioned driver evaluation's device, included each unit and module are
It is divided according to function logic, but is not limited to above-mentioned division, as long as corresponding function can be realized；Separately
Outside, the specific name of each functional unit is also only to facilitate mutually distinguish, the protection domain being not intended to restrict the invention.
Note that above are only presently preferred embodiments of the present invention and institute's application technology principle.It will be appreciated by those skilled in the art that
The present invention is not limited to specific embodiments described here, can carry out for a person skilled in the art it is various it is apparent variation,
It readjusts and substitutes without departing from protection scope of the present invention.Therefore, although being carried out to the present invention by above example
It is described in further detail, but the present invention is not limited only to above example, without departing from the inventive concept, also
May include other more equivalent embodiments, and the scope of the present invention is determined by scope of the appended claims.
Claims (11)
1. a kind of driver evaluation's method, which is characterized in that including：
Obtain the driving image of driver；
The driving image is analyzed to determine driver identity by human face discriminating device, wherein the human face discriminating
Device includes a variety of human face recognition models based on different face recognition algorithms, and most human face recognition models are exported
Face recognition result is as final result；
Driving behavior is determined according to the driver identity or the driving image, and by the driving behavior and the driver
Identity is associated；
Driver is assessed according to the driving behavior, the assessment is including at least in driving behavior assessment or performance evaluation
One kind.
2. according to the method described in claim 1, it is characterized in that, the human face discriminating device includes odd number kind recognition of face mould
Type；
When the result of all people's face identification model output is all different, by the corresponding recognition of face of default human face recognition model
As a result it is used as final result to export.
3. according to the method described in claim 2, it is characterized in that, the human face discriminating device include EignFace algorithms,
FisherFace algorithms and LBP algorithms.
4. according to the method described in claim 1, it is characterized in that, determine driving behavior according to the driver identity, including：
It is determined according to the duration of the driver identity and drives duration, when the driving duration includes at least current driving
The long, same day accumulative one driven in duration, month to date driving duration and driving expired times.
5. according to the method described in claim 1, it is characterized in that, determine driving behavior according to the driving image, and by institute
Driving behavior is stated to be associated with the driver identity, including：
Based on preset direction Gradient Features extraction algorithm, the histograms of oriented gradients feature of driving image is obtained；
The histograms of oriented gradients feature is analyzed by the SVM classifier trained, to determine that driver whether there is mobile phone
Usage behavior, the SVM classifier include at least one kind in local SVM classifier or global SVM classifier；
When the driver is there are when mobile phone usage behavior, the mobile phone usage behavior is closed with the driver identity
Connection.
6. according to the method described in claim 1, it is characterized in that, determine driving behavior according to the driving image, and by institute
Driving behavior is stated to be associated with the driver identity, including：
Multiple characteristic points in the driving image are determined by vertical-horizontal integral projection method；
Seek the binary image of the driving image；
According to the magnitude relationship of the grey scale pixel value of the characteristic point and default gray threshold, determine driver whether safe wearing
Band；
When the non-wear safety belt of the driver, the driving behavior of non-wear safety belt and the driver identity are closed
Connection.
7. according to the method described in claim 6, it is characterized in that, the driving behavior is to driver determined by the basis
It is assessed, when the assessment includes driving behavior assessment, further includes：
When detecting violation driving behavior, corresponding voice prompt is exported, wherein the violation driving behavior is driven including time-out
Sail behavior, mobile phone usage behavior and non-wear safety belt behavior.
8. a kind of driver evaluation's device, which is characterized in that including：
Driving image acquisition module, the driving image for obtaining driver；
Driver identity determining module analyzes to determine driver the driving image for passing through human face discriminating device
Identity, wherein the human face discriminating device includes a variety of human face recognition models based on different face recognition algorithms, and will be most
The face recognition result of human face recognition model output is as final result；
Driving behavior determining module determines driving behavior according to the driver identity or the driving image, and is driven described
Behavior is sailed to be associated with the driver identity；
Evaluation module, for being assessed driver according to the driving behavior, the assessment is commented including at least driving behavior
Estimate or performance evaluation in one kind.
9. device according to claim 8, which is characterized in that further include model module, the model module is for storing
Driving behavior assessment models, the assessment models are built based on target training data, and the target training data includes target carriage
All drivers driving behavior data or with target vehicle drive qualification all drivers driving behavior number
According to.
10. a kind of driving behavior assessment equipment, which is characterized in that the equipment includes：
Camera, for obtaining driving image；
One or more processors；
Storage device, for storing one or more programs；
When one or more of programs are executed by one or more of processors so that one or more of processors are real
Now driver evaluation's method as described in any in claim 1-7.
11. a kind of storage medium including computer executable instructions, which is characterized in that the computer executable instructions by
For executing driver evaluation's method as described in any in claim 1-7 when computer processor executes.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810398792.0A CN108647708A (en) | 2018-04-28 | 2018-04-28 | Driver evaluation's method, apparatus, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810398792.0A CN108647708A (en) | 2018-04-28 | 2018-04-28 | Driver evaluation's method, apparatus, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108647708A true CN108647708A (en) | 2018-10-12 |
Family
ID=63748392
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810398792.0A Pending CN108647708A (en) | 2018-04-28 | 2018-04-28 | Driver evaluation's method, apparatus, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108647708A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109460780A (en) * | 2018-10-17 | 2019-03-12 | 深兰科技（上海）有限公司 | Safe driving of vehicle detection method, device and the storage medium of artificial neural network |
| CN109461344A (en) * | 2018-12-07 | 2019-03-12 | 黑匣子(杭州)车联网科技有限公司 | A kind of method of car steering behavior assessment |
| CN110276245A (en) * | 2019-05-09 | 2019-09-24 | 浙江吉利控股集团有限公司 | A kind of face identification method of driver, device, system and terminal |
| CN110648075A (en) * | 2019-09-27 | 2020-01-03 | 重庆大学 | Driving safety evaluation method and device |
| CN112329543A (en) * | 2020-10-12 | 2021-02-05 | 易显智能科技有限责任公司 | Method and related device for evaluating driving ability process |
| CN112712010A (en) * | 2020-12-28 | 2021-04-27 | 科大讯飞股份有限公司 | Method, device and equipment for evaluating safe driving of common vehicle driver |
| CN112800854A (en) * | 2021-01-04 | 2021-05-14 | 中南大学 | Automatic analysis system for locomotive crew operation based on image |
| CN112949015A (en) * | 2019-12-10 | 2021-06-11 | 奥迪股份公司 | Modeling apparatus, assistance system, vehicle, method, and storage medium |
| CN113997940A (en) * | 2021-12-20 | 2022-02-01 | 镁佳(北京)科技有限公司 | Driving behavior monitoring method and device |
| CN115520196A (en) * | 2022-09-19 | 2022-12-27 | 广州汽车集团股份有限公司 | Method and device for determining weight of driver, electronic equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102592143A (en) * | 2012-01-09 | 2012-07-18 | 清华大学 | Method for detecting phone holding violation of driver in driving |
| CN104331687A (en) * | 2014-10-30 | 2015-02-04 | 安徽国华光电技术有限公司 | Safety belt unfastening behavior detection method based on vehicular video analysis |
| CN105096528A (en) * | 2015-08-05 | 2015-11-25 | 广州云从信息科技有限公司 | Fatigue driving detection method and system |
| JP2017151694A (en) * | 2016-02-24 | 2017-08-31 | 株式会社デンソー | Safety confirmation diagnostic system and safety confirmation diagnostic method |
| CN107563280A (en) * | 2017-07-24 | 2018-01-09 | 南京道熵信息技术有限公司 | Face identification method and device based on multi-model |
-
2018
- 2018-04-28 CN CN201810398792.0A patent/CN108647708A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102592143A (en) * | 2012-01-09 | 2012-07-18 | 清华大学 | Method for detecting phone holding violation of driver in driving |
| CN104331687A (en) * | 2014-10-30 | 2015-02-04 | 安徽国华光电技术有限公司 | Safety belt unfastening behavior detection method based on vehicular video analysis |
| CN105096528A (en) * | 2015-08-05 | 2015-11-25 | 广州云从信息科技有限公司 | Fatigue driving detection method and system |
| JP2017151694A (en) * | 2016-02-24 | 2017-08-31 | 株式会社デンソー | Safety confirmation diagnostic system and safety confirmation diagnostic method |
| CN107563280A (en) * | 2017-07-24 | 2018-01-09 | 南京道熵信息技术有限公司 | Face identification method and device based on multi-model |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109460780A (en) * | 2018-10-17 | 2019-03-12 | 深兰科技（上海）有限公司 | Safe driving of vehicle detection method, device and the storage medium of artificial neural network |
| CN109461344A (en) * | 2018-12-07 | 2019-03-12 | 黑匣子(杭州)车联网科技有限公司 | A kind of method of car steering behavior assessment |
| CN110276245A (en) * | 2019-05-09 | 2019-09-24 | 浙江吉利控股集团有限公司 | A kind of face identification method of driver, device, system and terminal |
| CN110648075A (en) * | 2019-09-27 | 2020-01-03 | 重庆大学 | Driving safety evaluation method and device |
| CN112949015A (en) * | 2019-12-10 | 2021-06-11 | 奥迪股份公司 | Modeling apparatus, assistance system, vehicle, method, and storage medium |
| CN112329543A (en) * | 2020-10-12 | 2021-02-05 | 易显智能科技有限责任公司 | Method and related device for evaluating driving ability process |
| CN112712010A (en) * | 2020-12-28 | 2021-04-27 | 科大讯飞股份有限公司 | Method, device and equipment for evaluating safe driving of common vehicle driver |
| CN112800854A (en) * | 2021-01-04 | 2021-05-14 | 中南大学 | Automatic analysis system for locomotive crew operation based on image |
| CN113997940A (en) * | 2021-12-20 | 2022-02-01 | 镁佳(北京)科技有限公司 | Driving behavior monitoring method and device |
| CN115520196A (en) * | 2022-09-19 | 2022-12-27 | 广州汽车集团股份有限公司 | Method and device for determining weight of driver, electronic equipment and storage medium |
| CN115520196B (en) * | 2022-09-19 | 2023-10-13 | 广州汽车集团股份有限公司 | Weight determination method and device for driver, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108647708A (en) | Driver evaluation's method, apparatus, equipment and storage medium | |
| CN110532976B (en) | Fatigue driving detection method and system based on machine learning and multi-feature fusion | |
| CN103714660B (en) | System for achieving fatigue driving judgment on basis of image processing and fusion between heart rate characteristic and expression characteristic | |
| CN102289660B (en) | Method for detecting illegal driving behavior based on hand gesture tracking | |
| CN105373779B (en) | A kind of vehicle safety belt intelligent detecting method and intelligent checking system | |
| CN105488453B (en) | A kind of driver based on image procossing does not fasten the safety belt detection recognition method | |
| CN103679168B (en) | Detection method and detection device for character region | |
| CN109584507A (en) | Driver behavior modeling method, apparatus, system, the vehicles and storage medium | |
| CN104361332B (en) | A kind of face eye areas localization method for fatigue driving detection | |
| US20150116493A1 (en) | Method and system for estimating gaze direction of vehicle drivers | |
| CN101211411B (en) | Human body detection process and device | |
| CN105956548A (en) | Driver fatigue state detection method and device | |
| Kashevnik et al. | Seat belt fastness detection based on image analysis from vehicle in-abin camera | |
| CN105404862A (en) | Hand tracking based safe driving detection method | |
| CN102855500A (en) | Haar and HoG characteristic based preceding car detection method | |
| CN104268596A (en) | License plate recognizer and license plate detection method and system thereof | |
| CN108804987B (en) | Door opening and closing state detection method and device and people flow detection system | |
| CN107529659B (en) | Seatbelt wearing detection method, device and electronic equipment | |
| CN103700220A (en) | Fatigue driving monitoring device | |
| CN108108651B (en) | Method and system for detecting driver non-attentive driving based on video face analysis | |
| CN109800654A (en) | Vehicle-mounted camera detection processing method, apparatus and vehicle | |
| CN106295474A (en) | The fatigue detection method of deck officer, system and server | |
| CN104574819A (en) | Fatigued drive detection method based on mouth features | |
| Shaily et al. | Smart driver monitoring system | |
| Xu et al. | A machine learning approach to vehicle occupancy detection |
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: 20181012 |
|
| RJ01 | Rejection of invention patent application after publication |