CN202177912U - Driver driving behavior monitoring device based on video detection - Google Patents
Driver driving behavior monitoring device based on video detection Download PDFInfo
- Publication number
- CN202177912U CN202177912U CN2011202666983U CN201120266698U CN202177912U CN 202177912 U CN202177912 U CN 202177912U CN 2011202666983 U CN2011202666983 U CN 2011202666983U CN 201120266698 U CN201120266698 U CN 201120266698U CN 202177912 U CN202177912 U CN 202177912U
- Authority
- CN
- China
- Prior art keywords
- processing unit
- central processing
- links
- driver
- module
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Traffic Control Systems (AREA)
Abstract
The utility model provides a driver driving behavior monitoring device based on the video detection. The monitoring device includes a central processor, and an image acquisition and pre-processing module, a function button module, an alarm, a display screen, an external memory card, and a power supply module which are connected with the central processor respectively. In monitoring, the device is self-checked, and the central processor arranges an internal register to the image data acquisition and pre-processing module; the image acquisition and pre-processing module is responsible for the image acquisition and the pre-processing to an image analog signal; the data after the pre-processing is converted in format via the central processor, and then the driver driving behavior identification is carried out based on the video detection; if the behavior is determined to be illegal, the alarm generates a corresponding alarm, and a corresponding illegal image is stored into the external memory card for the later checking and tracking. The driver driving behavior monitoring device based on video detection can effectively monitor whether or not a driver has an illegal behavior in the driving period, and generate an alarm to the illegal behavior, thus effectively preventing the traffic accident caused by the illegal operation of the driver.
Description
Technical field
The utility model relates to the automotive safety technical field, is specifically related to the driver's driving behavior supervising device based on Video Detection.
Background technology
In the control loop that people, car, road are formed, the driver is the maximum inducement of traffic hazard.2009, nearly 240,000 of China's road traffic accident, dead nearly 6.8 ten thousand people.Wherein, most accidents cause owing to driver's operation error and fatigue driving.Because the variation of age, physiology or aspects such as mental health state, mood keep its original good driving condition surely muchly even outstanding driver also differs, but the driver is difficult to recognize this gradual decay or disappears.Therefore; Monitoring driving person's driving behavior also gives alarm to unlawful practice, to the driving ability that improves the driver and reduce it and drive load, coordinates the relation between driver and vehicle and the traffic environment; From reducing the generation of traffic hazard situation in essence, significant.
At present, obtaining some achievements in research aspect monitoring driving person's driving behavior both at home and abroad, be broadly divided into two kinds: a kind of is to judge whether to drink according to the alcohol content in driver's the expiration; Judge whether fatigue driving of driver according to the relative principle of reflection of driver's eyelid and eyeball; Brain wave or cardiogram according to the driver judge that driver's device whether fatigue waits some monitoring driving persons on physiology, whether to be in normal condition comes driver's driving condition is estimated.Characteristics such as another kind of head movement situation to the driver, facial characteristics (like eyes, head, face) variation, utilization Computer Image Processing and mode identification technology are analyzed, to judge driver's the driving behavior and the state of mind.Yet these achievements in research all are indirectly monitoring to be judged in its driving behavior, driver's driving behavior itself are not directly studied, and have the more high restriction of measuring error and hardware cost.
The utility model content
In order to solve above-mentioned existing in prior technology problem; The utility model provides the driver's driving behavior supervising device based on Video Detection; The utility model comprises the image of driver's hand and bearing circle through camera collection; Again treatment of picture and identification are judged whether the driver has unlawful practice during driving, and make alarm prompting driver according to corresponding violation action.The utility model can be avoided the traffic hazard that causes because of driver's violation operation effectively.
The utility model is to realize through following technical scheme:
A kind of driver's driving behavior supervising device based on Video Detection comprises: power module, IMAQ and pre-processing module, function button module, warning horn, display screen, external memory card and central processing unit.Wherein: an output terminal of power module links to each other with the power input of pre-processing module with IMAQ, and another output terminal of power module links to each other with the power input of central processing unit; IMAQ links to each other with the image input interface of central processing unit with the image output interface of pre-processing module, and IMAQ links to each other with the EBI of central processing unit with the EBI of pre-processing module; The output terminal of function button module links to each other with the universal input/output interface of central processing unit; The input end of warning horn is connected with the PWM output interface of central processing unit; The input end of display screen links to each other with the demonstration output interface of central processing unit; External memory card links to each other through exterior storage card slot central processing unit.
In above-mentioned a kind of driver's driving behavior supervising device based on Video Detection, described power module comprises filtering circuit, transforming circuit and back-up source; Power module provides WV for respectively camera and central processing unit; In addition;, WV (is lower than normal voltage or outage) when being in ERST; The WV that back-up source provides a period of time for IMAQ and pre-processing module and central processing unit produces the phenomenon of loss of data with equipment under the situation that prevents abnormal voltage.
In above-mentioned a kind of driver's driving behavior supervising device based on Video Detection; Described IMAQ and pre-processing module comprise camera and video frequency processing chip; The output interface of camera links to each other through vision cable with the input end of video frequency processing chip, and the EBI of central processing unit links to each other with the EBI of video frequency processing chip in the pre-processing module with IMAQ; Driver's hand and bearing circle are positioned at the visual field of said camera, state driver's hand motion with observation post after an action of the bowels; Central processing unit is configured the video frequency processing chip internal register through EBI; Thereby IMAQ and pre-processing module have had the preprocessing function to the input simulating signal; Pre-service comprises: the control of colourity and brightness, the selection control of output data form and output image synchronizing signal etc.; Pretreated data are transferred to central processing unit through the image output interface of video frequency processing chip again.
In above-mentioned a kind of driver's driving behavior supervising device based on Video Detection, described function button module, the affirmation that can be used for manually setting up area-of-interest and set up area-of-interest automatically.
In above-mentioned a kind of driver's driving behavior supervising device based on Video Detection, described warning horn is mainly used in when different violation driving behaviors takes place the driver, can produce corresponding alarm and remind; Also can be used for equipment self-inspection.
In above-mentioned a kind of driver's driving behavior supervising device based on Video Detection, described display screen is mainly used in the image that shows after preview is handled.
In above-mentioned a kind of driver's driving behavior supervising device based on Video Detection; Described central processing unit comprises: image input interface, universal input/output interface, PWM output interface, display screen output interface, external memory area slot, power input interface and EBI.Wherein: the image input interface of central processing unit links to each other with the image output interface of pre-processing module with IMAQ; And the EBI of central processing unit links to each other with the EBI of pre-processing module with IMAQ; The central processing unit universal input/output interface links to each other with function button module output terminal; The PWM output interface of central processing unit is connected with the output terminal of warning horn; The demonstration output interface of central processing unit links to each other with the input end of display screen, and central processing unit links to each other with external memory card through the exterior storage card slot, and the power input interface of central processing unit links to each other with power module.Central processing unit mainly be responsible for image data format conversion, to based on the driving of identification, function button module and the warning horn of driver's driving behavior of Video Detection, transmit data to display screen and preserve data message in external memory card.
Above-mentioned central processing unit comprises the identification based on driver's driving behavior of Video Detection: read view data, orientation dish after format conversion, set up area-of-interest, extract hand-characteristic (area-of-interest to former figure carries out feature extraction), the hand-characteristic (0-1 eigenmatrix) that extracts is carried out Classification and Identification and sets up rule base in violation of rules and regulations according to the position relation of driver's hand and bearing circle, and judge whether violation of driver's operation according to rule base.
Above-mentioned a kind of driver's driving behavior supervising device based on Video Detection; Its method for supervising is: behind the power initiation; Equipment self-inspection and IMAQ and pre-processing module dispose internal register through the EBI of central processing unit, thereby have the input simulating signal through the pretreated function of row.Camera in IMAQ and the pre-processing module is responsible for acquisition of image data; Video frequency processing chip in IMAQ and the pre-processing module carries out pre-service to image analoging signal; Pre-service comprises: the control of colourity and brightness, the selection control of output data form and output image synchronizing signal; Pretreated data are through the central processing unit format conversion; Carry out a kind of driver's driving behavior identification again based on Video Detection; If be judged as unlawful practice, warning horn produces corresponding the warning, and is saved in corresponding violation picture in the external memory card; Write log file writing down in violation of rules and regulations, so that anaphase nucleus is to following the tracks of.The log file content comprises in violation of rules and regulations: in violation of rules and regulations the time, move in violation of rules and regulations, the duration, the numbering of corresponding image is in violation of rules and regulations formed.
Above-mentioned driver's driving behavior recognition methods based on Video Detection, concrete steps are following:
Step 1 reads the view data after format conversion, and the AP that is read should comprise bearing circle and driver's information (to call former figure in the following text) such as hand attitude;
Step 3 is set up area-of-interest.Set up area-of-interest at former figure according to oriented bearing circle, area-of-interest comprises the zone of bearing circle and driver's hand information;
Step 4 is carried out feature extraction to the area-of-interest of former figure, mainly is that hand-characteristic extracts.This step is divided into the M*N sub-areas with area-of-interest, adopts complexion model to judge that whether each subarea comprises hand-characteristic, sets up the 0-1 eigenmatrix.
Step 5; 0-1 eigenmatrix (representative hand-characteristic) to extracting carries out Classification and Identification according to the position relation of driver's hand and bearing circle, with driver's driver behavior divide into normal driving, both hands platen, both hands separation disc, right hand separation disc, left hand separation disc, both hands intersect the right hand at last and both hands intersection left hand last.
Step 6 is set up rule base in violation of rules and regulations based on the duration and the frequency of driver's driver behavior, and judges according to rule base whether in violation of rules and regulations driver's operation.
In the step 2, the bearing circle shape major part in the image is oval or circular, need carry out the ellipse detection to image bearing circle is positioned.Because the bearing circle profile possesses maximum circle or elliptical profile profile in the Vehicular video image; So, can adopt direct least square ellipse fitting algorithm that the profile of bearing circle is extracted and detect maximum elliptical shape and accomplish location to the direction disk area to the edge image that obtains through secondary treating.
In the step 3, behind the location of completion to the direction disk area, be the corresponding area-of-interest of center intercepting with the bearing circle.The intercepting of area-of-interest is mainly by the position corresponding relation decision of center, size and driver's contouring head characteristic of bearing circle.During concrete intercepting, through accomplishing intercepting from the bearing circle zone with the outward extending mode of preset proportion coefficient.For different vehicles, this scale-up factor is by experimental calibration or rule of thumb definite.
In the step 4, this step is divided into the M*N sub-areas with area-of-interest, adopts complexion model to judge that whether each subarea comprises hand-characteristic, sets up the 0-1 eigenmatrix.Selected simple Gauss model is discerned the hand colour of skin in each subarea as complexion model, sets up the 0-1 eigenmatrix, wherein 1 represents the skin pixel value, and 0 represents background pixel value.This method is proceeded in two phases, and at first selects suitable complexion model and confirms the parameter of model, and the parameter deterministic process is following: choose meet the human body complexion characteristic in a large number pixel as sample, add up its distribution and set up colour of skin Gauss model.Utilize this model to differentiate new pixel then or whether the zone is the colour of skin.The pixel that at first choose normal illumination, high light, meets human body features of skin colors under the different light in a large number under night (low light level) is added up its distribution and is set up colour of skin Gauss model as sample; Utilize model under the different light to differentiate new pixel then or whether the zone is the colour of skin; Wherein judge and be in which kind of illumination condition according to the gradation of image distribution.
In the step 5; After obtaining 0-1 eigenmatrix (representative hand-characteristic); Adopt neural network and BAYESIAN NETWORK CLASSIFIER that it is carried out Classification and Identification, concrete implementation procedure is following: 1) according to different vehicles, choose the picture of various driving behaviors by a certain percentage with quantity; 2) the utilization preceding method is handled the image of choosing and is obtained the motion state parameters corresponding with all kinds of driving behaviors, forms the training example set; 3) with the training example set neural network and BAYESIAN NETWORK CLASSIFIER are trained the driving behavior sorter that obtains being adapted to all kinds of vehicles.Can discern different hand elemental motions through sorter, set up basic driving behavior feature database in violation of rules and regulations, set up basic driving behavior feature database in violation of rules and regulations according to these basic actions in violation of rules and regulations again.
In the step 6, rule is moved by one or more basic violation in violation of rules and regulations, and the duration of being somebody's turn to do action, and occurrence frequency comes the common violation rule that determines; In conjunction with region of interest and violation rule base, judge whether in violation of rules and regulations behavior.
If judged result is unlawful practice as stated above, warning horn produces accordingly and reports to the police, and is saved in a corresponding violation picture in the external memory card, writes log file writing down in violation of rules and regulations, so that anaphase nucleus is to following the tracks of.The log file content comprises in violation of rules and regulations: in violation of rules and regulations the time, move in violation of rules and regulations, the duration, the numbering of corresponding image is in violation of rules and regulations formed.
Owing to adopted above scheme, made to the utlity model has following advantage and effect:
1, the utility model adopt to be followed the tracks of driver's hand and is moved whether in violation of rules and regulations to judge the driver, has opened up a new directly effective monitoring approach, to prevention since in violation of rules and regulations the traffic hazard that causes of driving behavior have great importance.
2, the utility model extracts hand-characteristic through central processing unit to skin detection; Select simple Gauss model as complexion model; Considered the distribution situation that falls into complexion model scope interior pixel point; The applied probability density formula judges that pixel belongs to the probability of the colour of skin, rather than directly all pixels that fall in the model scope is simply classified as colour of skin point, can better represent that with respect to regional model the colour of skin distributes; To also high many of Face Detection efficient, and the parameter of model also is easy to calculate.Therefore this device has the high and high advantage of reliability of accuracy of detection.
3, the utility model is through the driving behavior of central processing unit identification of driver; When there is the violation driver behavior in the driver; Just produce corresponding alarm and reminding driver, picture is saved in the external memory card in violation of rules and regulations simultaneously, so that the later stage tracking enquiry; There is figure that certificate is arranged, can effectively reduces driver's bad steering behavior.
4, high, little, the strong interference immunity of volume of the device degree of intelligence in the utility model is convenient to application.
Description of drawings
Fig. 1 is based on the structural representation of driver's driving behavior supervising device of Video Detection.
Fig. 2 is based on the process flow diagram of driver's violation driving behavior recognition methods of Video Detection.
Embodiment
Be described further below in conjunction with the practical implementation of accompanying drawing, but enforcement of the utility model and protection domain are not limited thereto the utility model.
In this embodiment, camera collection comprises the image of driver's hand and bearing circle, through treatment of picture and identification are judged whether the driver has unlawful practice during driving, and makes alarm prompting driver according to corresponding violation action.The utility model can be avoided the traffic hazard that causes because of driver's violation operation effectively.As shown in Figure 1, a kind of driver's driving behavior supervising device based on Video Detection comprises: power module 1, IMAQ and pre-processing module 2, function button module 3, warning horn 4, display screen 5, external memory card 6 and central processing unit 7.Wherein: an output terminal of power module 1 links to each other with the power input of IMAQ with pre-processing module 2, and another output terminal of power module 1 links to each other with the power input of central processing unit 7; IMAQ links to each other with the image input interface of central processing unit 7 with the image output interface of pre-processing module 2, and the EBI (I
C EBI) of the video frequency processing chip in IMAQ and the pre-processing module 2 links to each other with the EBI (I
C EBI) of central processing unit 7; The output terminal of function button module 3 links to each other with the universal input/output interface of central processing unit 7; The input end of warning horn 4 is connected with the PWM output interface of central processing unit 7; The input end of display screen 5 links to each other with the demonstration output interface of central processing unit 7; External memory card 6 links to each other through exterior storage card slot central processing unit 7.
Described power module 1 comprises filtering circuit, transforming circuit and back-up source; Power module 1 provides WV for respectively camera and central processing unit; In addition;, WV (is lower than normal voltage or outage) when being in ERST; The WV that back-up source provides a period of time for IMAQ and pre-processing module 2 and central processing unit 7 produces the phenomenon of loss of data with equipment under the situation that prevents abnormal voltage.
Described IMAQ and pre-processing module 2 comprise camera and video frequency processing chip, and video frequency processing chip is selected the SAA7113 chip for use among the embodiment, and the SAA7113 chip is supported the input and the data output format of various video signal; The output interface of camera links to each other through vision cable with the input end of video frequency processing chip, and central processing unit 7 links to each other through I
the C EBI of the video frequency processing chip in EBI (I
C EBI) and the pre-processing module 2; Driver's hand and bearing circle are positioned at the visual field of said camera, so that observe said driver's hand motion; Central processing unit 7 passes through EBI (I
C EBI) to configuration SAA7113 chip internal corresponding registers; Thereby IMAQ and pre-processing module 2 have had the preprocessing function to the input simulating signal; Pre-service comprises: the control of colourity and brightness, the selection control of output data form and output image synchronizing signal etc.; The principle of work of IMAQ and pre-processing module 2 is following: the camera induced environment changes; The pal mode simulating signal of output is transferred to image data acquiring and pre-processing module 2 through vision cable; SAA7113 chip in image data acquiring and the pre-processing module 2 (the SAA7113 chip is supported the input and the data output format of various video signal) begins to gather pal mode simulating signal (only one tunnel composite video signal of input being sampled); The video simulation of the SAA7113 chip in image data acquiring and the pre-processing module 2 output signal after pre-service with the numeral output of the 4:2:2 of the standard of ITU656 agreement, and as the input of image input and output (camera interface) interface of central processing unit 7 (S3C2440).
Described function button module 3 can be used for manually setting up area-of-interest and to setting up the affirmation of area-of-interest automatically.
Described warning horn 4 is mainly used in when different violation driving takes place the driver, can produce corresponding warning horn and remind; Also can be used for equipment self-inspection.
Described display screen 5 is mainly used in the image that shows after preview is handled.
Described central processing unit 7 is also claimed MCU, selects the ARM9 chip of Samsung S3C2440 microprocessor among this embodiment for use, dominant frequency 400MHz, 133MHz bus frequency.CPU module 7 comprises: image input interface, universal input/output interface, PWM output interface, demonstration output interface, exterior storage card slot, power input interface and EBI (I
C EBI).Wherein, The image input interface of central processing unit 7 links to each other with the image output interface of pre-processing module 2 with IMAQ, and the EBI of central processing unit 7 (I
C EBI) links to each other with the EBI (I
C EBI) of pre-processing module 2 with IMAQ; The universal input/output interface of central processing unit 7 links to each other with function button module 3 output terminals; The PWM output interface of central processing unit 7 is connected with the input end of warning horn 4; The demonstration output interface of central processing unit 7 links to each other with the input end of display screen 5; Central processing unit 7 links to each other with slot external memory card 6 through external memory card; The power input interface of central processing unit 7 links to each other with power module 1.Central processing unit 7 is main be responsible for image data formats conversion, to based on the identification of driver's driving behavior of Video Detection, drive function button module 3 and warning horn 4, transmit data to display screen 5 and data message is saved in the external memory card.
Above-mentioned central processing unit comprises the identification based on driver's driving behavior of Video Detection: read view data, orientation dish after format conversion, set up area-of-interest, hand-characteristic extracts (area-of-interest to former figure carries out feature extraction), carries out Classification and Identification and set up rule base in violation of rules and regulations according to the position relation of driver's hand and bearing circle extracting hand-characteristic (0-1 eigenmatrix), and judge whether violation of driver's operation according to rule base.
Above-mentioned a kind of driver's driving behavior supervising device based on Video Detection; Its method for supervising is roughly following: after power module 1 starts; Equipment self-inspection and IMAQ and pre-processing module 2 are through the EBI configuration internal register of central processing unit 7, thereby IMAQ and pre-processing module 2 have had the preprocessing function to the input simulating signal.Camera in IMAQ and the pre-processing module 2 is responsible for acquisition of image data; Video frequency processing chip in IMAQ and the pre-processing module 2 carries out pre-service to image analoging signal; Pre-service comprises: the control of colourity and brightness, the selection control of output data form and output image synchronizing signal; Pretreated data are carried out a kind of driver's driving behavior identification based on Video Detection again through the conversion of central processing unit 7 image data formats; If be judged as unlawful practice, warning horn produces accordingly and reports to the police, and is saved in a corresponding violation picture in the external memory card 6, writes log file writing down in violation of rules and regulations, so that anaphase nucleus is to following the tracks of.The log file content comprises in violation of rules and regulations: in violation of rules and regulations the time, move in violation of rules and regulations, the duration, the numbering of corresponding image is in violation of rules and regulations formed.
Above-mentioned a kind of driver's driving behavior recognition methods based on Video Detection is as shown in Figure 2, and concrete steps are following:
Step 1 reads the view data after format conversion, and the AP that is read should comprise bearing circle and driver's information (to call former figure in the following text) such as hand attitude;
Step 3 is set up area-of-interest.Set up area-of-interest at former figure according to oriented bearing circle, area-of-interest comprises the zone of bearing circle and driver's hand information;
Step 4 is carried out feature extraction to the area-of-interest of institute's reading images in the step 1, mainly is that hand-characteristic extracts.This step is divided into the M*N sub-areas with area-of-interest, adopts complexion model to judge that whether each subarea comprises hand-characteristic, sets up the 0-1 eigenmatrix.
Step 5; 0-1 eigenmatrix (representative hand-characteristic) to extracting carries out Classification and Identification according to the position relation of driver's hand and bearing circle, with driver's driver behavior divide into normal driving, both hands platen, both hands separation disc, right hand separation disc, left hand separation disc, both hands intersect the right hand at last and both hands intersection left hand last.
Step 6 is set up rule base in violation of rules and regulations based on the duration and the frequency of driver's driver behavior, and judges according to rule base whether in violation of rules and regulations driver's operation.
In the step 2, the bearing circle shape major part in the image is oval or circular, need carry out the ellipse detection to image bearing circle is positioned.Ellipse commonly used detects and can be divided into based on ballot and optimized two big class methods.The representative algorithm of ballot class methods comprises Hough conversion and RANSAC scheduling algorithm.Optimization method then comprises least square method and genetic algorithm etc.Because elliptic parameter is more, the emphasis of ballot type research generally all is the screening of data point and the utilization of elliptic geometry character.The Hough conversion, RANSAC is the method that adopts mapping, and sample point is projected to parameter space, the method for perhaps birdsing of the same feather flock together with totalizer detects ellipse.This type algorithm has good robustness, a plurality of ellipses of ability one-time detection, but need complex calculations and a large amount of storage spaces.Another kind of method comprises the least square fitting algorithm, genetic algorithm and other optimization ellipse fitting method.The principal feature of these class methods is that accuracy is high, but can't directly be used for the detection of a plurality of ellipses, and the sensitivity of noise is higher than last class methods.Because the bearing circle profile is to possess maximum circle or elliptical profile profile in the Vehicular video image; Promptly only need detect the elliptical shape of a maximum; So this embodiment has adopted direct least square ellipse fitting algorithm to extract the bearing circle profile and has detected maximum elliptical shape; Detect the circumscribed rectangle of this ellipse again, accomplish location the direction disk area.
In the step 3, the intercepting of area-of-interest is mainly by the position corresponding relation decision of center, size and driver's contouring head characteristic of bearing circle.During concrete intercepting, accomplish behind the direction disk area location, through accomplishing intercepting from the bearing circle zone with the outward extending mode of certain proportion coefficient.For different vehicles, this scale-up factor is by experimental calibration or can rule of thumb confirm.
In the step 4, feature extraction being carried out in the zone of intercepting, mainly is hand-characteristic, can extract hand-characteristic through the detection to skin.This step is divided into the M*N sub-areas with area-of-interest, adopts complexion model to judge that whether each subarea comprises hand-characteristic, sets up the 0-1 eigenmatrix.
In order to detect skin, need selected suitable complexion model, the hand colour of skin in intercepting zone to be discerned, complexion model commonly used roughly is divided into two types in the Flame Image Process: simple threshold values is cut apart and probability model.Wherein probability model has histogram model, simple Gauss model and mixed Gauss model.Simple Gauss model is a kind of model that the hypothesis colour of skin is distributed as uni-modal gaussian.Simple Gauss model has been considered the distribution situation that falls into complexion model scope interior pixel point; The applied probability density formula judges that pixel belongs to the probability of the colour of skin; Rather than directly all pixels that fall in the model scope are simply classified as colour of skin point; Can better represent that with respect to regional model the colour of skin distributes, therefore also high many of its Face Detection efficient comparatively speaking, and the parameter of model also is easy to calculate.This embodiment adopts simple Gauss model as complexion model, and the hand colour of skin in each subarea is discerned.Set up the 0-1 eigenmatrix, wherein 1 represent the skin pixel value, 0 represents background pixel value.This method is proceeded in two phases, and at first selects suitable complexion model and confirms the parameter of model, and the parameter deterministic process is following: choose meet the human body complexion characteristic in a large number pixel as sample, add up its distribution and set up colour of skin Gauss model.Utilize this model to differentiate new pixel then or whether the zone is the colour of skin.The pixel that at first choose normal illumination, high light, meets human body features of skin colors under the different light in a large number under night (low light level) is added up its distribution and is set up colour of skin Gauss model as sample; Utilize model under the different light to differentiate new pixel then or whether the zone is the colour of skin, thereby set up the 0-1 eigenmatrix.Wherein judge and be in which kind of illumination condition according to the gradation of image distribution.
In the step 5, the 0-1 eigenmatrix (representative hand-characteristic) that extracts is carried out Classification and Identification.What the utility model method adopted is that neural network and BAYESIAN NETWORK CLASSIFIER carry out Classification and Identification to the 0-1 eigenmatrix, through identification different character matrix, distinguishes whether belong to driving behavior in violation of rules and regulations.After obtaining the 0-1 eigenmatrix, adopt neural network and BAYESIAN NETWORK CLASSIFIER that it is carried out Classification and Identification, concrete implementation procedure is following: 1) according to different vehicles, choose the picture of various driving behaviors by a certain percentage with quantity; 2) the utilization preceding method is handled the image of choosing and is obtained and all kinds of driving behavior corresponding states parameters, forms the training example set; 3) with the training example set neural network and BAYESIAN NETWORK CLASSIFIER are trained the driving behavior sorter that obtains being adapted to all kinds of vehicles.Can discern different hand elemental motion through sorter; Set up basic driving behavior feature database in violation of rules and regulations, can discern seven kinds of elemental motions through sorter among this embodiment: normal driving, both hands platen, both hands separation disc, right hand separation disc, left hand separation disc, the both hands intersection right hand intersect left hand last at last, both hands.
Step 6 is set up rule base in violation of rules and regulations based on the duration and the frequency of driver's driver behavior, and judges according to rule base whether in violation of rules and regulations driver's operation.
If judged result is unlawful practice as stated above, warning horn produces accordingly and reports to the police, and is saved in a corresponding violation picture in the external memory card 6, writes log file writing down in violation of rules and regulations, so that anaphase nucleus is to following the tracks of.The log file content comprises in violation of rules and regulations: in violation of rules and regulations the time, move in violation of rules and regulations, the duration, the numbering of corresponding image is in violation of rules and regulations formed.
Whether present embodiment identification of driver automatically has hand to move in violation of rules and regulations during driving, and to the driver alarm is provided, and writes log file writing down in violation of rules and regulations, so that anaphase nucleus is to following the tracks of.The accuracy of present embodiment is high, can effectively avoid the generation of the traffic hazard that the driver causes because of violation operation.
Claims (4)
1. based on driver's driving behavior supervising device of Video Detection; It is characterized in that comprising power module, IMAQ and pre-processing module, function button module, warning horn, display screen, external memory card and central processing unit; Wherein: an output terminal of power module links to each other with the power input of pre-processing module with IMAQ, and another output terminal of power module links to each other with the power input of central processing unit; IMAQ links to each other with the image input interface of central processing unit with the image output interface of pre-processing module, and IMAQ links to each other with the EBI of central processing unit with the EBI of pre-processing module; The output terminal of function button module links to each other with the universal input/output interface of central processing unit; The input end of warning horn is connected with the PWM output interface of central processing unit; The input end of display screen links to each other with the demonstration output interface of central processing unit; External memory card links to each other through exterior storage card slot central processing unit.
2. supervising device according to claim 1 is characterized in that said power module comprises filtering circuit, transforming circuit and back-up source.
3. supervising device according to claim 1; It is characterized in that said IMAQ and pre-processing module comprise camera and video frequency processing chip; The output interface of camera links to each other through vision cable with the input end of video frequency processing chip, and the EBI of central processing unit links to each other with the EBI of video frequency processing chip in the pre-processing module with IMAQ; Driver's hand and bearing circle are positioned at the visual field of said camera.
4. according to each described supervising device of claim 1～3; It is characterized in that said central processing unit comprises image input interface, universal input/output interface, PWM output interface, display screen output interface, external memory area slot, power input interface and EBI; The image input interface of central processing unit links to each other with the image output interface of pre-processing module with IMAQ; And the EBI of central processing unit links to each other with the EBI of pre-processing module with IMAQ; The central processing unit universal input/output interface links to each other with function button module output terminal; The PWM output interface of central processing unit is connected with the output terminal of warning horn; The demonstration output interface of central processing unit links to each other with the input end of display screen, and central processing unit links to each other with external memory card through the exterior storage card slot, and the power input interface of central processing unit links to each other with power module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011202666983U CN202177912U (en) | 2011-07-26 | 2011-07-26 | Driver driving behavior monitoring device based on video detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011202666983U CN202177912U (en) | 2011-07-26 | 2011-07-26 | Driver driving behavior monitoring device based on video detection |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN202177912U true CN202177912U (en) | 2012-03-28 |
Family
ID=45867784
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2011202666983U Expired - Fee Related CN202177912U (en) | 2011-07-26 | 2011-07-26 | Driver driving behavior monitoring device based on video detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN202177912U (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104751663A (en) * | 2015-02-28 | 2015-07-01 | 北京壹卡行科技有限公司 | Safe driving auxiliary system and safe driving auxiliary method for driver |
| CN111439269A (en) * | 2020-04-13 | 2020-07-24 | 重庆车辆检测研究院有限公司 | Evaluation method, device and system of driver monitoring device and evaluator |
-
2011
- 2011-07-26 CN CN2011202666983U patent/CN202177912U/en not_active Expired - Fee Related
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104751663A (en) * | 2015-02-28 | 2015-07-01 | 北京壹卡行科技有限公司 | Safe driving auxiliary system and safe driving auxiliary method for driver |
| CN111439269A (en) * | 2020-04-13 | 2020-07-24 | 重庆车辆检测研究院有限公司 | Evaluation method, device and system of driver monitoring device and evaluator |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102263937B (en) | Driver's driving behavior monitoring device and monitoring method based on video detection | |
| CN101593425B (en) | Machine vision based fatigue driving monitoring method and system | |
| Ji et al. | Fatigue state detection based on multi-index fusion and state recognition network | |
| CN107704805B (en) | Method for detecting fatigue driving, automobile data recorder and storage device | |
| Gupta et al. | Implementation of motorist weariness detection system using a conventional object recognition technique | |
| CN101587544B (en) | Based on the carried on vehicle antitracking device of computer vision | |
| CN101599207A (en) | A kind of fatigue driving detection device and automobile | |
| Junaedi et al. | Driver drowsiness detection based on face feature and PERCLOS | |
| CN101877051A (en) | Driver attention state monitoring method and device | |
| CN104881956A (en) | Fatigue driving early warning system | |
| CN107977607A (en) | A kind of fatigue driving monitoring method based on machine vision | |
| Yimyam et al. | The System for Driver Fatigue Monitoring Using Decision Tree via Wireless Sensor Network for Intelligent Transport System. | |
| He et al. | A real-time driver fatigue detection method based on two-stage convolutional neural network | |
| CN103247150A (en) | Fatigue driving preventing system | |
| CN202177912U (en) | Driver driving behavior monitoring device based on video detection | |
| CN114771545A (en) | Intelligent safe driving system | |
| Stan et al. | Eye-gaze tracking method driven by raspberry PI applicable in automotive traffic safety | |
| Suryawanshi et al. | Driver drowsiness detection system based on lbp and haar algorithm | |
| CN108304764B (en) | Fatigue state detection device and detection method in simulated flight driving process | |
| Cristiani et al. | Driver drowsiness identification by means of passive techniques for eye detection and tracking | |
| Kavitha et al. | Implementing OpenCV and Dlib Open-Source library for detection of driver’s fatigue | |
| Reddy et al. | Soft Computing Techniques for Driver Alertness | |
| Liu et al. | Design and implementation of multimodal fatigue detection system combining eye and yawn information | |
| CN215526691U (en) | Multifunctional safety face recognition system | |
| Sharran et al. | Drowsy Driver Detection System |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20120328 Termination date: 20140726 |
|
| EXPY | Termination of patent right or utility model |