CN107844783A - A kind of commerial vehicle abnormal driving behavioral value method and system - Google Patents
A kind of commerial vehicle abnormal driving behavioral value method and system Download PDFInfo
- Publication number
- CN107844783A CN107844783A CN201711277950.9A CN201711277950A CN107844783A CN 107844783 A CN107844783 A CN 107844783A CN 201711277950 A CN201711277950 A CN 201711277950A CN 107844783 A CN107844783 A CN 107844783A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- abnormal driving
- driver
- training
- 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
- 230000002159 abnormal effect Effects 0.000 title claims abstract description 53
- 230000003542 behavioural effect Effects 0.000 title claims abstract description 41
- 238000000034 method Methods 0.000 title claims abstract description 24
- 230000006399 behavior Effects 0.000 claims abstract description 19
- 230000011218 segmentation Effects 0.000 claims abstract description 4
- 238000012549 training Methods 0.000 claims description 40
- 230000000007 visual effect Effects 0.000 claims description 32
- 238000012544 monitoring process Methods 0.000 claims description 17
- 238000004891 communication Methods 0.000 claims description 14
- 238000001514 detection method Methods 0.000 claims description 14
- 238000013179 statistical model Methods 0.000 claims description 13
- 230000006870 function Effects 0.000 claims description 12
- 239000011159 matrix material Substances 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 8
- 230000005540 biological transmission Effects 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 7
- 238000007476 Maximum Likelihood Methods 0.000 claims description 3
- 238000005286 illumination Methods 0.000 claims description 3
- 238000005192 partition Methods 0.000 claims description 3
- 241000208340 Araliaceae Species 0.000 claims description 2
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims description 2
- 235000003140 Panax quinquefolius Nutrition 0.000 claims description 2
- 235000008434 ginseng Nutrition 0.000 claims description 2
- 230000009466 transformation Effects 0.000 claims 1
- 230000000391 smoking effect Effects 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 4
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 206010039203 Road traffic accident Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000012806 monitoring device Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000033764 rhythmic process Effects 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/59—Context or environment of the image inside of a vehicle, e.g. relating to seat occupancy, driver state or inner lighting conditions
- G06V20/597—Recognising the driver's state or behaviour, e.g. attention or drowsiness
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
-
- 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
- G06V40/161—Detection; Localisation; Normalisation
- G06V40/166—Detection; Localisation; Normalisation using acquisition arrangements
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Human Computer Interaction (AREA)
- Traffic Control Systems (AREA)
Abstract
The present invention relates to a kind of commerial vehicle abnormal driving behavioral value method and system, driver's video is gathered by the camera in front of commerial vehicle, the position of driver is partitioned into from the video image collected using car-mounted terminal, carry out skin color segmentation, using the face of adboost positioning drivers, judge whether to make a phone call with reference to face and driver behavior, smoking, speak, the abnormal driving behavior such as fatigue.Abnormal driving behavior is passed to background monitoring center by network, and voice reminder is carried out in vehicle-mounted end, the notice of driver is lifted, abnormal driving behavior is recorded, ensure the safety of vehicle drive.
Description
Technical field
The invention belongs to technical field of image processing, and in particular to a kind of commerial vehicle abnormal driving behavioral value method and
System.
Background technology
According to Ministry of Public Security's as shown by data, commerial vehicle accounts for the 70% of sum in the especially big traffic accident in the whole nation, accident rate with
The death rate remains high, and wherein operating passenger car driver tired driving behavior is the major reason for causing operation accident occurred frequently, is given
Driver and passenger's person and property safety cause serious threat.Driver fatigue state monitoring technology is broadly divided into three major types,
Detection method based on physical signs, the detection method based on the analysis of driver's behavioral trait and the detection side based on machine vision
Method.
Detection method based on physical signs uses contact type measurement mode, typically passes through the physiological signal of test drives people
To speculate the fatigue state of driver.In such method, eeg analysis are to apply the most, method of best performance at present,
It is to estimate the fatigue state of driver by analyzing the absolute or relative change of each rhythm and pace of moving things composition in EEG spectrum.In addition,
Research is found, in the case of long-time nighttime driving or fatigue driving, heart rate meeting degradation, electrocardiogram, which can be also used as, differentiates ginseng
Count to infer to driver fatigue state.But contact physiological parameter assessment method usually requires measured and wears phase
It the device (such as electrode patch) answered, can cause greatly to disturb to driving behavior, be not suitable for the application under actual environment.
It is special by analyzing the steering wheel of driver, pedal operation based on the fatigue detection method of driver's behavioral trait
Property or vehicle driving trace feature speculate the fatigue state of driver, wherein amendment operating characteristic quilt of the driver to steering wheel
Think stronger correlation be present with fatigue state.The steering wheel monitoring of Digital Installations companies of U.S. exploitation
Device S.A.M detects steering wheel angle by being placed in the magnetic strips below steering wheel, if no pair of driver in a period of time
Steering wheel carries out any amendment operation, then system judges that driver enters fatigue state and triggers alarm.Although it is based on driver
The fatigue detection method of operating characteristic can reach certain accuracy of identification, and measurement process will not bring interference to driver,
But the operation of driver with fatigue state except having outside the Pass, also by road environment, travel speed, personal habits, operative skill etc.
Influence, its accuracy still has much room for improvement with robustness.
Human body one or the shape in several regions are detected by monitoring device based on the fatigue detection method of machine vision
State, pass through the driving behavior of these condition adjudgements driver.Eye feature, the mouth motion feature of driver all can be directly used for
Detection fatigue, wherein the information related to eye state is most widely used at present.
The content of the invention
In order to solve the above-mentioned problems in the prior art, the invention provides a kind of commerial vehicle abnormal driving behavior
Detection method and system.The technical problem to be solved in the present invention is achieved through the following technical solutions：
A kind of commerial vehicle abnormal driving behavioral value method, comprises the following steps：
Step 1: 500 all ages and classes of collection, different sexes and driver's image of different illumination, manual segmentation go out institute
The area of skin color of driver's image is stated, and the broca scale picture cut out is gone into YC from rgb color spacebCrColor space, obtain
To the training data of area of skin color；
Step 2: the C of area of skin color pixel is calculated using formula (1)bCrValue is in Gaussian density function Pi(c | skin) in
Probable value：
Wherein, c is pixel in YCbCrColor vector [C corresponding to color spaceb,Cr]T, mean μ and covariance matrix Σ
Calculated by step 3；
Step 3: mean μ is estimated with covariance matrix Σ by EM algorithms, it is divided into the E- steps and formula (3) of formula (2)
M- is walked, wherein：E- walks the expectation for calculating log-likelihood function, and M- walks the parameter for selecting expectation maximum, then by described in
Parameter substitutes into E- steps, calculates and it is expected, so repeatedly, untill converging to the optimal solution in maximum likelihood meaning；
E- is walked：
M- is walked：
Wherein, ωi,jRepresent sample ciBelong to the posterior probability of jth class；The proportionality coefficient of jth class is represented,Represent
The average of jth class,Represent the covariance matrix of jth class；
Step 4: the mixed Gaussian statistical model of structure area of skin color, the probability density function of mixed Gaussian statistical model
As shown in formula (4)：
Wherein：M be Gaussian component number, πiIt is the proportion coefficient of Gaussian component, meets：πi>=0, i=1,2 ...
m；Pi(c | skin) is the Gaussian density function of ith pixel point in step 2；
Step 5: using step 1 to step 4, the mixed Gaussian statistical model built by step 4 carries out colour of skin area
The extraction in domain；
Wherein, the process for carrying out skin cluster is：Driver's image is inputted in mixed Gaussian statistical model, first will
Color space conversion is to YCbCrIn color space, the C of each pixel is obtainedbCrValue, formula (4) is recycled to calculate each pixel
CbCrIt is worth the probable value in mixed Gauss model, the pixel that probable value is more than to training data statistical threshold is labeled as skin
Color dot, and then the region that all colour of skin points are formed retains area of skin color, wipes non-colour of skin area as the area of skin color of image
Domain；
Step 6: the extraction image that step 5 is obtained, human face region is identified in the face monitor trained；
Step 7: the human face region that removal step six obtains in the area of skin color obtained from step 5, and by other areas
Domain is defined as human hand region, according to the position of human hand region in the picture, defines left-hand area and right-hand area, then according to left hand
The position relationship in region, right-hand area and human face region, judge driver whether abnormal driving.
Face monitor is trained as a preferred embodiment of the present invention, in the step 6 using collecting
What 500 driver's photos were trained, specific training process is：
Initial thick instruction is carried out in whole 180 degree scope to [- 90 °, the 90 °] visual angle rotated horizontally outside face plane first
Practice；Initial thick training is carried out in whole 60 degree of scopes to [- 30 °, the 30 °] visual angle to be turned clockwise in face plane；To face
[- 20 °, the 20 °] visual angle from top to bottom rotated outside plane carries out initial thick training in whole 40 degree of scopes；
Secondly [- 90 °, the 90 °] visual angle rotated horizontally outside face plane is divided into [- 90 °, -30 °], [- 30 °, 30 °],
[30 °, 90 °] three subintervals are finely divided training；[- 30 °, the 30 °] visual angle to be turned clockwise in face plane is divided into
[- 30 °, -10 °], [- 10 °, 10 °], [10 °, 30 °] three subintervals are finely divided training；It will be rotated outside face plane from upper
Arrive down [- 20 °, 20 °] visual angle and be divided into [- 20 °, 0 °], [0 °, 20 °] two subintervals are finely divided training；
[- 90 °, the 90 °] visual angle rotated horizontally outside face plane is divided into [- 90 °, -60 °] again, [- 60 °, -
30 °], [- 30 °, 0 °], [0 °, 30 °], [30 °, 60 °], [60 °, 90 °] six subintervals are further finely divided training；By people
[- 30 °, the 30 °] visual angle to be turned clockwise in face plane is divided into [- 30 °, -20 °], [- 20 °, -10 °], [- 10 °, 0 °],
[0 °, 10 °], [10 °, 20 °], [20 °, 30 °] six subintervals are further finely divided training；To rotate outside face plane from
Top to bottm [- 20 °, 20 °] visual angle is divided into [- 20 °, -10 °], [- 10 °, 0 °], [0 °, 10 °], [10 °, 20 °] four subintervals
Further it is finely divided training；
Finally the detector that every class visual angle is trained on different sections is layered according to above-mentioned partition order and integrated, it is initially thick
The detector of training segments the detector of training under upper, the top-down various visual angles face inspection for forming a layering cascade
Survey device.
It is abnormal using above-mentioned commerial vehicle the invention also discloses a kind of commerial vehicle abnormal driving behavioral value system
Driving behavior detection method, including：
Camera：Immediately ahead of commerial vehicle, for gathering the realtime graphic of driver；
Car-mounted terminal：It is connected with camera output end, is handled for receiving realtime graphic, and to realtime graphic, is examined
Measure after driver has abnormal driving behavior, the behavioural information of abnormal driving is sent to Surveillance center；
Monitoring unit：For receiving the behavioural information of car-mounted terminal transmission, and the behavioural information of abnormal driving is remembered
Record；
The output end of the camera and the input of car-mounted terminal connect, and the output end of the car-mounted terminal and monitoring are single
The input connection of member.
As a preferred embodiment of the present invention, prompting module is provided with the car-mounted terminal, the prompting module is used
Driven in reminding driver safety.
As a preferred embodiment of the present invention, the car-mounted terminal is provided with wireless communication module and microprocessor,
The output end of the input connection camera of the microprocessor, the output end of the microprocessor and wireless communication module it is defeated
Enter end connection, the wireless communication module carries out radio communication with radio network gateway；
The monitoring unit includes being used to receive radio network gateway, monitor that car-mounted terminal sends abnormal driving behavioural information
And memory module, the input of the monitor and the output end of radio network gateway connect, the output end of the monitor and storage
The input connection of module；
The monitor is used for the behavioural information for handling abnormal driving, to realize the behavioural information of abnormal driving and the driving
The personal information of member matches；
The memory module is used for the abnormal driving behavioural information for recording the driver.
As a preferred embodiment of the present invention, the transmission means of the wireless communication module is 2G networks, 3G network,
One or more in 4G networks.
As a preferred embodiment of the present invention, the model ARM7TDMI-S of the microprocessor.
Compared with prior art, beneficial effects of the present invention：
The present invention only need to install video monitoring equipment in front of driver's cabin, compared to other driving behavior detection modes, to driving
The driving procedure for the person of sailing does not have any influence.Method detection of the image that video monitoring equipment collects by the present invention, this
Invention can determine whether out driver with the presence or absence of make a phone call, smoking, the abnormal driving behavior such as speak.
Brief description of the drawings
Fig. 1 is the schematic diagram of commerial vehicle abnormal driving behavioral value system.
Fig. 2 is the block diagram of commerial vehicle abnormal driving behavioral value system.
Fig. 3 is the schematic diagram that mixed Gaussian statistical model carries out area of skin color extraction result.
Fig. 4 is abnormal driving differentiation figure.
In figure：1st, commerial vehicle；11st, camera；12nd, car-mounted terminal；13rd, prompting module；2nd, monitoring unit.
Embodiment
Further detailed description is done to the present invention with reference to specific embodiment, but embodiments of the present invention are not limited to
This.
In order to more easily detect whether the driving behavior of driver is abnormal, present embodiments provides a kind of commerial vehicle
Abnormal driving behavioral value method, comprises the following steps：
Step 1: 500 all ages and classes of collection, different sexes and driver's image of different illumination, manual segmentation go out to drive
The area of skin color of the person's of sailing image, and the broca scale picture cut out is gone into YC from rgb color spacebCrColor space, obtain skin
The training data in color region；
Step 2: the C of area of skin color pixel is calculated using formula (1)bCrValue is in Gaussian density function Pi(c | skin) in
Probable value：
Wherein, c is pixel in YCbCrColor vector [C corresponding to color spaceb,Cr]T, mean μ and covariance matrix Σ
Calculated by step 3；
Step 3: mean μ is estimated with covariance matrix Σ by EM algorithms, it is divided into the E- steps and formula (3) of formula (2)
M- is walked, wherein：E- walks the expectation for calculating log-likelihood function, and M- is walked for selecting it is expected maximum parameter, then by parameter
E- steps are substituted into, calculates and it is expected, so repeatedly, untill converging to the optimal solution in maximum likelihood meaning；
E- is walked：
M- is walked：
Wherein, ωi,jRepresent sample ciBelong to the posterior probability of jth class；The proportionality coefficient of jth class is represented,Represent
The average of jth class,Represent the covariance matrix of jth class；
Step 4: the mixed Gaussian statistical model of structure area of skin color, the probability density function of mixed Gaussian statistical model
As shown in formula (4)：
Wherein：M be Gaussian component number, πiIt is the proportion coefficient of Gaussian component, meets：πi>=0, i=1,2 ...
m；(all proportion coefficients and for 1), Pi(c | skin) it is close for the Gauss of ith pixel point in step 2
Spend function；It is shown experimentally that, when Gaussian component number is 8, can be preferably fitted the skin distribution of human body；
Step 5: using step 1 to step 4, the mixed Gaussian statistical model built by step 4 carries out colour of skin area
The extraction in domain；
Wherein, the process for carrying out skin cluster is：Driver's image is inputted in mixed Gaussian statistical model, first will
Color space conversion is to YCbCrIn color space, the C of each pixel is obtainedbCrValue, formula (4) is recycled to calculate each pixel
CbCrIt is worth the probable value in mixed Gauss model, the pixel that probable value is more than to training data statistical threshold is labeled as skin
Color dot, and then the region that all colour of skin points are formed retains area of skin color, wipes non-colour of skin area as the area of skin color of image
Domain (the image schematic diagram of extraction is as shown in Figure 3)；
Step 6: the extraction image that step 5 is obtained, human face region is identified in the face monitor trained；
Wherein, face monitor is trained in step 6 to be trained using the 500 driver's photos collected, is had
The training process of body is：
Initial thick instruction is carried out in whole 180 degree scope to [- 90 °, the 90 °] visual angle rotated horizontally outside face plane first
Practice；Initial thick training is carried out in whole 60 degree of scopes to [- 30 °, the 30 °] visual angle to be turned clockwise in face plane；To face
[- 20 °, the 20 °] visual angle from top to bottom rotated outside plane carries out initial thick training in whole 40 degree of scopes；
Secondly [- 90 °, the 90 °] visual angle rotated horizontally outside face plane is divided into [- 90 °, -30 °], [- 30 °, 30 °],
[30 °, 90 °] three subintervals are finely divided training；[- 30 °, the 30 °] visual angle to be turned clockwise in face plane is divided into
[- 30 °, -10 °], [- 10 °, 10 °], [10 °, 30 °] three subintervals are finely divided training；It will be rotated outside face plane from upper
Arrive down [- 20 °, 20 °] visual angle and be divided into [- 20 °, 0 °], [0 °, 20 °] two subintervals are finely divided training；
[- 90 °, the 90 °] visual angle rotated horizontally outside face plane is divided into [- 90 °, -60 °] again, [- 60 °, -
30 °], [- 30 °, 0 °], [0 °, 30 °], [30 °, 60 °], [60 °, 90 °] six subintervals are further finely divided training；By people
[- 30 °, the 30 °] visual angle to be turned clockwise in face plane is divided into [- 30 °, -20 °], [- 20 °, -10 °], [- 10 °, 0 °],
[0 °, 10 °], [10 °, 20 °], [20 °, 30 °] six subintervals are further finely divided training；To rotate outside face plane from
Top to bottm [- 20 °, 20 °] visual angle is divided into [- 20 °, -10 °], [- 10 °, 0 °], [0 °, 10 °], [10 °, 20 °] four subintervals
Further it is finely divided training；
Finally the detector that every class visual angle is trained on different sections is layered according to above-mentioned partition order and integrated, it is initially thick
The detector of training segments the detector of training under upper, the top-down various visual angles face inspection for forming a layering cascade
Survey device；
Step 7: the human face region that removal step six obtains in the area of skin color obtained from step 5, and by other areas
Domain is defined as human hand region, according to the position of human hand region in the picture, defines left-hand area and right-hand area, then according to left hand
The position relationship in region, right-hand area and human face region, judge driver whether abnormal driving (abnormal driving behavior differentiate result
Figure is as shown in figure 4, wherein, the image of Fig. 4 right-hand columns the top, is the image of driver's normal driving；Fig. 4 right-hand column bottoms
Image, image when abnormal driving behavior be present for driver).
The present embodiment also discloses a kind of commerial vehicle abnormal driving behavioral value device, as depicted in figs. 1 and 2, including：
Camera：Immediately ahead of commerial vehicle, for gathering the realtime graphic of driver；
Car-mounted terminal：It is connected with camera output end, is handled for receiving realtime graphic, and to realtime graphic, is examined
Measure after driver has abnormal driving behavior, the behavioural information of abnormal driving is sent to Surveillance center；
Monitoring unit：For receiving the behavioural information of car-mounted terminal transmission, and the behavioural information of abnormal driving is remembered
Record；
The output end of camera and the input of car-mounted terminal connect, the output end of car-mounted terminal and the input of monitoring unit
End connection.
Wherein, car-mounted terminal is provided with wireless communication module and microprocessor, the input connection camera of microprocessor
Output end, the input of the output end of microprocessor and wireless communication module connected, and wireless communication module is put into wireless network
Row radio communication；The transmission means of wireless communication module is the one or more in 2G networks, 3G network, 4G networks；It is it is preferred that micro-
The model ARM7TDMI-S- single-chip microcomputers of processor.
Wherein, monitoring unit includes being used for radio network gateway, the monitoring for receiving car-mounted terminal transmission abnormal driving behavioural information
The output end of device and memory module, the input of monitor and radio network gateway connects, the output end of monitor and memory module
Input connects；Monitor is used for the behavioural information for handling abnormal driving, to realize the behavioural information of abnormal driving and the driving
The personal information of member matches；Memory module is used for the abnormal driving behavioural information for recording the driver.
Wherein, prompting module is provided with car-mounted terminal, the prompting module is used to remind driver safety to drive, the prompting
The input of module and the output end of microprocessor connect.
The present embodiment need to only install video monitoring equipment in front of driver's cabin, right compared to other driving behavior detection modes
The driving procedure of driver does not have any influence.The image that video monitoring equipment collects is examined by the method for the present embodiment
Survey, the present embodiment can determine whether out driver with the presence or absence of make a phone call, smoking, the abnormal driving behavior such as speak.
Above content is to combine specific preferred embodiment further description made for the present invention, it is impossible to is assert
The specific implementation of the present invention is confined to these explanations.For general technical staff of the technical field of the invention,
On the premise of not departing from present inventive concept, some simple deduction or replace can also be made, should all be considered as belonging to the present invention's
Protection domain.
Claims (7)
1. a kind of commerial vehicle abnormal driving behavioral value method, comprises the following steps：
Step 1: 500 all ages and classes of collection, different sexes and driver's image of different illumination, manual segmentation go out described drive
The area of skin color of the person's of sailing image, and the broca scale picture cut out is gone into YC from rgb color spacebCrColor space, obtain skin
The training data in color region；
Step 2: the C of area of skin color pixel is calculated using formula (1)bCrValue is in Gaussian density function PiIt is general in (c | skin)
Rate value：
<mrow>
<msub>
<mi>P</mi>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>c</mi>
<mo>|</mo>
<mi>s</mi>
<mi>k</mi>
<mi>i</mi>
<mi>n</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mrow>
<mn>2</mn>
<mi>&pi;</mi>
<mo>|</mo>
<mi>&Sigma;</mi>
<msup>
<mo>|</mo>
<mrow>
<mn>1</mn>
<mo>/</mo>
<mn>2</mn>
</mrow>
</msup>
</mrow>
</mfrac>
<msup>
<mi>e</mi>
<mrow>
<mo>-</mo>
<mfrac>
<mn>1</mn>
<mn>2</mn>
</mfrac>
<msup>
<mrow>
<mo>(</mo>
<mi>c</mi>
<mo>-</mo>
<mi>&mu;</mi>
<mo>)</mo>
</mrow>
<mi>T</mi>
</msup>
<msup>
<mi>&Sigma;</mi>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>c</mi>
<mo>-</mo>
<mi>&mu;</mi>
<mo>)</mo>
</mrow>
</mrow>
</msup>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, c is pixel in YCbCrColor vector [C corresponding to color spaceb,Cr]T, mean μ passes through with covariance matrix Σ
Step 3 calculates；
Step 3: mean μ is estimated with covariance matrix Σ by EM algorithms, it is divided into the E- steps of formula (2) and the M- of formula (3)
Step, wherein：E- walks the expectation for calculating log-likelihood function, and M- is walked for selecting it is expected maximum parameter, then by the ginseng
Number substitutes into E- steps, calculates and it is expected, so repeatedly, untill converging to the optimal solution in maximum likelihood meaning；
E- is walked：
M- is walked：
<mrow>
<msub>
<mover>
<mo>&Sigma;</mo>
<mo>^</mo>
</mover>
<mi>j</mi>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</munderover>
<msub>
<mi>&omega;</mi>
<mrow>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
</mrow>
</msub>
<mo>&CenterDot;</mo>
<mrow>
<mo>(</mo>
<msub>
<mi>c</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mover>
<mi>&mu;</mi>
<mo>^</mo>
</mover>
<mi>j</mi>
</msub>
<mo>)</mo>
</mrow>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>c</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mover>
<mi>&mu;</mi>
<mo>^</mo>
</mover>
<mi>j</mi>
</msub>
<mo>)</mo>
</mrow>
<mi>T</mi>
</msup>
</mrow>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>N</mi>
</munderover>
<msub>
<mi>&omega;</mi>
<mrow>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
</mrow>
</msub>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>3</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, ωi,jRepresent sample ciBelong to the posterior probability of jth class；The proportionality coefficient of jth class is represented,Represent jth class
Average,Represent the covariance matrix of jth class；
Step 4: the mixed Gaussian statistical model of structure area of skin color, the probability density function such as public affairs of mixed Gaussian statistical model
Shown in formula (4)：
<mrow>
<mi>P</mi>
<mrow>
<mo>(</mo>
<mi>c</mi>
<mo>|</mo>
<mi>s</mi>
<mi>k</mi>
<mi>i</mi>
<mi>n</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>m</mi>
</munderover>
<msub>
<mi>&pi;</mi>
<mi>i</mi>
</msub>
<mo>&CenterDot;</mo>
<msub>
<mi>P</mi>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>c</mi>
<mo>|</mo>
<mi>s</mi>
<mi>k</mi>
<mi>i</mi>
<mi>n</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>4</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein：M be Gaussian component number, πiIt is the proportion coefficient of Gaussian component, meets：πi>=0, i=1,2 ... m；Pi(c | skin) is the Gaussian density function of ith pixel point in step 2；
Step 5: using step 1 to step 4, the mixed Gaussian statistical model built by step 4 carries out area of skin color
Extraction；
Wherein, the process for carrying out skin cluster is：Driver's image is inputted in mixed Gaussian statistical model, first by color
Spatial transformation is to YCbCrIn color space, the C of each pixel is obtainedbCrValue, formula (4) is recycled to calculate each pixel
CbCrIt is worth the probable value in mixed Gauss model, the pixel that probable value is more than to training data statistical threshold is labeled as the colour of skin
Point, and then the region that all colour of skin points are formed retains area of skin color, wipes non-colour of skin area as the area of skin color of image
Domain；
Step 6: the extraction image that step 5 is obtained, human face region is identified in the face monitor trained；
Step 7: the human face region that removal step six obtains in the area of skin color obtained from step 5, and other regions are determined
Justice is human hand region, according to the position of human hand region in the picture, defines left-hand area and right-hand area, then according to left hand area
The position relationship in domain, right-hand area and human face region, judge driver whether abnormal driving.
2. commerial vehicle abnormal driving behavioral value method as claimed in claim 1, it is characterised in that：Instructed in the step 6
Practicing face monitor is trained using the 500 driver's photos collected, and specific training process is：
Initial thick training is carried out in whole 180 degree scope to [- 90 °, the 90 °] visual angle rotated horizontally outside face plane first；It is right
[- 30 °, the 30 °] visual angle to be turned clockwise in face plane carries out initial thick training in whole 60 degree of scopes；Outside to face plane
[- 20 °, the 20 °] visual angle from top to bottom of rotation carries out initial thick training in whole 40 degree of scopes；
Secondly [- 90 °, the 90 °] visual angle rotated horizontally outside face plane is divided into [- 90 °, -30 °], [- 30 °, 30 °],
[30 °, 90 °] three subintervals are finely divided training；[- 30 °, the 30 °] visual angle to be turned clockwise in face plane is divided into
[- 30 °, -10 °], [- 10 °, 10 °], [10 °, 30 °] three subintervals are finely divided training；It will be rotated outside face plane from upper
Arrive down [- 20 °, 20 °] visual angle and be divided into [- 20 °, 0 °], [0 °, 20 °] two subintervals are finely divided training；
[- 90 °, the 90 °] visual angle rotated horizontally outside face plane is divided into [- 90 °, -60 °] again, [- 60 °, -30 °], [-
30 °, 0 °], [0 °, 30 °], [30 °, 60 °], [60 °, 90 °] six subintervals are further finely divided training；By in face plane
[- 30 °, the 30 °] visual angle to turn clockwise is divided into [- 30 °, -20 °], [- 20 °, -10 °], [- 10 °, 0 °], [0 °, 10 °],
[10 °, 20 °], [20 °, 30 °] six subintervals are further finely divided training；By rotate outside face plane from top to bottom [-
20 °, 20 °] visual angle is divided into [- 20 °, -10 °], and [- 10 °, 0 °], [0 °, 10 °], [10 °, 20 °] four subintervals are further entered
Row subdivision training；
The detector that every class visual angle is trained on different sections is finally layered to integrated, initially thick training according to above-mentioned partition order
Detector upper, segment the detector of training under, the top-down multi-view face detection device for forming a layering cascade.
A kind of 3. commerial vehicle abnormal driving behavioral value system, using the commerial vehicle abnormal driving shown in claim 1 or 2
Behavioral value method, it is characterised in that including：
Camera：Immediately ahead of commerial vehicle, for gathering the realtime graphic of driver；
Car-mounted terminal：It is connected with camera output end, handles, detect for receiving realtime graphic, and to realtime graphic
After driver has abnormal driving behavior, the behavioural information of abnormal driving is sent to Surveillance center；
Monitoring unit：For receiving the behavioural information of car-mounted terminal transmission, and the behavioural information of abnormal driving is recorded；
The input of the output end of the camera and car-mounted terminal connects, the output end of the car-mounted terminal and monitoring unit
Input connects.
4. commerial vehicle abnormal driving behavioral value system as claimed in claim 3, it is characterised in that on the car-mounted terminal
Prompting module is provided with, the prompting module is used to remind driver safety to drive.
5. the commerial vehicle abnormal driving behavioral value system as described in claim 3 or 4, it is characterised in that the vehicle-mounted end
End is provided with wireless communication module and microprocessor, and the output end of the input connection camera of the microprocessor is described micro-
The output end of processor and the input of wireless communication module connect, and the wireless communication module carries out channel radio with radio network gateway
Letter；
The monitoring unit includes sending the radio network gateway of abnormal driving behavioural information, monitor for receiving car-mounted terminal and deposited
Module is stored up, the input of the monitor and the output end of radio network gateway connect, the output end and memory module of the monitor
Input connection；
The monitor is used for the behavioural information for handling abnormal driving, to realize the behavioural information of abnormal driving and the driver's
Personal information matches；
The memory module is used for the abnormal driving behavioural information for recording the driver.
6. commerial vehicle abnormal driving behavioral value system as claimed in claim 5, it is characterised in that the radio communication mold
The transmission means of block be 2G networks, the one or more in 3G network, 4G networks.
7. commerial vehicle abnormal driving behavioral value system as claimed in claim 5, it is characterised in that the microprocessor
Model ARM7TDMI-S.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711277950.9A CN107844783A (en) | 2017-12-06 | 2017-12-06 | A kind of commerial vehicle abnormal driving behavioral value method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711277950.9A CN107844783A (en) | 2017-12-06 | 2017-12-06 | A kind of commerial vehicle abnormal driving behavioral value method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107844783A true CN107844783A (en) | 2018-03-27 |
Family
ID=61664486
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711277950.9A Pending CN107844783A (en) | 2017-12-06 | 2017-12-06 | A kind of commerial vehicle abnormal driving behavioral value method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107844783A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109167874A (en) * | 2018-07-18 | 2019-01-08 | 奇酷互联网络科技（深圳）有限公司 | Mobile terminal and monitoring drive the method, apparatus of user action state |
| CN109684990A (en) * | 2018-12-20 | 2019-04-26 | 天津天地伟业信息系统集成有限公司 | A kind of behavioral value method of making a phone call based on video |
| 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 |
| CN110154893A (en) * | 2019-05-29 | 2019-08-23 | 辽宁工业大学 | A kind of vehicle security drive method for early warning based on driver characteristics |
| CN110334592A (en) * | 2019-05-27 | 2019-10-15 | 天津科技大学 | A kind of monitoring of driver's abnormal behaviour and safety control system and safety control method |
| CN112991718A (en) * | 2019-12-13 | 2021-06-18 | 丰田自动车株式会社 | Driver assistance device, non-transitory storage medium storing driver assistance program, and driver assistance system |
| CN113291311A (en) * | 2021-05-28 | 2021-08-24 | 同济大学 | Method for detecting abnormal behavior of driver under emergency collision avoidance working condition and storage medium |
| CN114466333A (en) * | 2022-04-11 | 2022-05-10 | 北京烽火万家科技有限公司 | Internet of vehicles control system based on eSIM and 5G |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102073841A (en) * | 2009-11-20 | 2011-05-25 | 中国移动通信集团广东有限公司 | Poor video detection method and device |
| CN103748985B (en) * | 2010-06-09 | 2012-02-22 | 北京理工大学 | A kind of expectation maximization method for parameter estimation for gauss hybrid models |
| CN103824419A (en) * | 2012-11-16 | 2014-05-28 | 西安众智惠泽光电科技有限公司 | Automobile fatigue driving remote monitoring and real-time prompting system |
| CN106022242A (en) * | 2016-05-13 | 2016-10-12 | 哈尔滨工业大学（威海） | Driver call making identification method in intelligent transportation system |
-
2017
- 2017-12-06 CN CN201711277950.9A patent/CN107844783A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102073841A (en) * | 2009-11-20 | 2011-05-25 | 中国移动通信集团广东有限公司 | Poor video detection method and device |
| CN103748985B (en) * | 2010-06-09 | 2012-02-22 | 北京理工大学 | A kind of expectation maximization method for parameter estimation for gauss hybrid models |
| CN103824419A (en) * | 2012-11-16 | 2014-05-28 | 西安众智惠泽光电科技有限公司 | Automobile fatigue driving remote monitoring and real-time prompting system |
| CN106022242A (en) * | 2016-05-13 | 2016-10-12 | 哈尔滨工业大学（威海） | Driver call making identification method in intelligent transportation system |
Non-Patent Citations (1)
| Title |
|---|
| 邢圆: "针对司机违章通话的视频检测算法研究" * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109167874A (en) * | 2018-07-18 | 2019-01-08 | 奇酷互联网络科技（深圳）有限公司 | Mobile terminal and monitoring drive the method, apparatus of user action state |
| CN109684990A (en) * | 2018-12-20 | 2019-04-26 | 天津天地伟业信息系统集成有限公司 | A kind of behavioral value method of making a phone call based on video |
| 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 |
| CN110334592A (en) * | 2019-05-27 | 2019-10-15 | 天津科技大学 | A kind of monitoring of driver's abnormal behaviour and safety control system and safety control method |
| CN110154893A (en) * | 2019-05-29 | 2019-08-23 | 辽宁工业大学 | A kind of vehicle security drive method for early warning based on driver characteristics |
| CN110154893B (en) * | 2019-05-29 | 2020-12-11 | 辽宁工业大学 | Automobile safe driving early warning method based on driver characteristics |
| CN112991718A (en) * | 2019-12-13 | 2021-06-18 | 丰田自动车株式会社 | Driver assistance device, non-transitory storage medium storing driver assistance program, and driver assistance system |
| CN113291311A (en) * | 2021-05-28 | 2021-08-24 | 同济大学 | Method for detecting abnormal behavior of driver under emergency collision avoidance working condition and storage medium |
| CN113291311B (en) * | 2021-05-28 | 2022-06-17 | 同济大学 | Method for detecting abnormal behavior of driver under emergency collision avoidance working condition and storage medium |
| CN114466333A (en) * | 2022-04-11 | 2022-05-10 | 北京烽火万家科技有限公司 | Internet of vehicles control system based on eSIM and 5G |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107844783A (en) | A kind of commerial vehicle abnormal driving behavioral value method and system | |
| CN104013414B (en) | A kind of Study in Driver Fatigue State Surveillance System based on intelligent movable mobile phone | |
| CN105769120B (en) | Method for detecting fatigue driving and device | |
| Rongben et al. | Monitoring mouth movement for driver fatigue or distraction with one camera | |
| CN103956028B (en) | The polynary driving safety means of defence of a kind of automobile | |
| CN103770733B (en) | Method and device for detecting safety driving states of driver | |
| CN103824420B (en) | Fatigue driving identification system based on heart rate variability non-contact measurement | |
| CN105151049B (en) | The early warning system detected based on driver's face feature and deviation | |
| CN103767715B (en) | A kind of driver safety driving state detection device | |
| CN102436715B (en) | Detection method for fatigue driving | |
| CN103871242B (en) | A kind of driving behavior overall evaluation system and method | |
| CN105354988B (en) | A kind of driver tired driving detecting system and detection method based on machine vision | |
| CN104123549B (en) | Eye positioning method for real-time monitoring of fatigue driving | |
| CN108474866A (en) | A kind of manhole cover loss detecting system and method based on deep learning | |
| CN108596087B (en) | Driving fatigue degree detection regression model based on double-network result | |
| CN109367539A (en) | A kind of intelligence system detecting fatigue driving | |
| CN103164711A (en) | Regional people stream density estimation method based on pixels and support vector machine (SVM) | |
| CN111753674A (en) | Fatigue driving detection and identification method based on deep learning | |
| CN103839055B (en) | A kind of detection method in pilot's line of vision direction | |
| CN109685026A (en) | A kind of driver holds the method for real-time of mobile phone communication | |
| CN107563346A (en) | One kind realizes that driver fatigue sentences method for distinguishing based on eye image processing | |
| CN202257872U (en) | Monitoring and alarming device for preventing losing of objects of passenger in taxi | |
| CN104881956A (en) | Fatigue driving early warning system | |
| CN103700220A (en) | Fatigue driving monitoring device | |
| CN112071084A (en) | Method and system for judging illegal parking by utilizing deep learning |
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: 20180327 |
|
| RJ01 | Rejection of invention patent application after publication |