CN103832357B - A kind of lane-departure warning system and method based on machine vision - Google Patents
A kind of lane-departure warning system and method based on machine vision Download PDFInfo
- Publication number
- CN103832357B CN103832357B CN201210486538.9A CN201210486538A CN103832357B CN 103832357 B CN103832357 B CN 103832357B CN 201210486538 A CN201210486538 A CN 201210486538A CN 103832357 B CN103832357 B CN 103832357B
- Authority
- CN
- China
- Prior art keywords
- processing unit
- image
- lane
- digital signal
- vehicle
- 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.)
- Active
Links
Landscapes
- Traffic Control Systems (AREA)
- Image Processing (AREA)
Abstract
A kind of lane-departure warning system based on machine vision, including interface processing unit, forward sight camera, digital signal processing unit and power module, forward sight camera does not block driver's sight and the cleanable position arrived of windscreen wiper installed in shield glass middle and upper part, gather vehicle front image, and by image transmitting to digital signal processing unit, whether digital signal processing unit intelligent decision vehicle will run-off-road, image information after being handled through digital signal processing unit is transferred to interface processing unit, interface processing unit control warning device work, power module gives other module for power supply.It uses advanced recognition in drosophila algorithm, vehicle-mounted camera driving road-condition is analyzed with reference to high speed digital signal processor, can predict vehicle will run-off-road, but the actual warning having no in the case that consciousness is so done in terms of driver sends vision, the sense of hearing or tactile of driver, to point out driver to take care traveling, effectively accident is reduced.
Description
Technical field
The present invention discloses a kind of vehicle warning system, particularly a kind of lane-departure warning system based on machine vision and
Method.
Background technology
With the gradually development of automobile industry, car ownership increasingly increases, automobile give people to live the convenience brought with
Quick self-evident, flourishing for automobile industry carrys out very big contribution to economy-zone, but also produces a series of societies simultaneously and ask
Topic, the problem of wherein traffic accident is most serious.According to statistics, 210812, death toll occur for Kuomintang-Communist in 2011：62387
People.The cause of wherein most accidents is that driver drives rashly, caused by not observing traffic rules and regulations, is greatly department in this
What machine was caused in driving procedure across two tracks, although highway is provided with track warning line, but in driving procedure, car
Whether line ball, completely by driver driving experience judge, it is extremely inaccurate.
The content of the invention
For easy line ball during car steering of the prior art mentioned above, the shortcoming of traffic accident is caused,
The present invention provides a kind of new lane-departure warning system and method based on machine vision, and it uses advanced computer vision
Algorithm for pattern recognition, is analyzed vehicle-mounted camera driving road-condition in real time with reference to high speed digital signal processor, can be with
Predict vehicle will run-off-road, but actual have no in the case that consciousness is so done of driver send vision to driver, listen
Feel or tactile in terms of warning, to point out driver to take care traveling, effectively reduce accident.
The present invention solve its technical problem use technical scheme be：A kind of lane departur warning system based on machine vision
System, system includes interface processing unit, forward sight camera, digital signal processing unit and power module, described forward sight shooting
Head does not block driver's sight and the cleanable position arrived of windscreen wiper installed in shield glass middle and upper part, gathers vehicle front figure
Picture, and by image transmitting to digital signal processing unit, whether digital signal processing unit intelligent decision vehicle will deviate car
Road, the image information after being handled through digital signal processing unit is transferred to interface processing unit, interface processing unit control alarm
Device works, and power module gives other module for power supply.
A kind of to realize blind spot vehicle detection and method for warming based on machine vision using above-mentioned system, this method includes
Following step：
A, igniting：System electrification；
B, before automobile is not walked or the too low system of speed enters idle pulley, can't now start lane departur warning
Function；
After C, disengaging idle condition, system detectio side-marker lamp on off state, if side-marker lamp is not opened, camera enters mould in the daytime
Formula；
If D, side-marker lamp are opened, into night vision mode；
E, system open camera collection real-time to road conditions and carry out AGC according to light intensity, and the image of distortion is entered
Row correction, the information collected is per second or 30 frames are per second is sent to digital signal processing unit with 25 frames；
F, digital signal processing unit are converted into the vision signal received the colour picture of rgb format；
G, YUV are encoded:The image of rgb format is converted into yuv format by matrixer：Obtain luminance signal Y
With two colour difference signal R-Y, i.e. U, B-Y, i.e. V, luminance signal Y and carrier chrominance signal U, V are separated；
H, picture breakdown are two parts up and down, and the latter half is only retained when processing and is used as calculating；
I, pattern analysis：Confirm weather conditions；
If J, wiper system or fog lamp system are opened, it is rain, mist, snowy day gas that system, which will be analyzed, now using numeral filter
Ripple device carries out image noise reduction；If wiper system or fog lamp system are directly entered K flows without unlatching；
K, lane identification analysis：Using Hough transform algorithm, rim detection identification is carried out to track：If having detected car
Road then enters deviation identification process L；If otherwise without track, system enters idle condition；
L, deviation identification：
The track identified is analyzed, in units of 1 pixel, works as X1>α and when θ 1 spends more than β, deviation is left
Track；
The deviation right lane when X2 is more than β ' less than α ', and θ 2；
Wherein α, β, α ', and β ' value according to different automobile types, the setting of different installation site timings；
If vehicle does not have run-off-road, this returns to idle condition and detected again；
If detecting deviation, step M is entered；
M, when track is that will deviate from, detect turning indicator control state：
If the steering indicating light of corresponding orientation is opened, then it is assumed that is that driver actively switches circuit, is then returned to idle condition；
If conversely, without unlatching corresponding orientation steering indicating light, then it is assumed that the unconscious tangent line of driver；
N, steering wheel inclination angle detection：Whether assistant analysis vehicle has steering behavior；
If O, network analysis result are unconscious tangent line, sound, image or vibrations are sent according to different danger classes
Caution signal.
The technical scheme that the present invention solves the use of its technical problem further comprises：
The beneficial effects of the invention are as follows：The present invention uses advanced computer vision algorithm for pattern recognition, with reference to high speed number
Word signal processor is analyzed vehicle-mounted camera driving road-condition in real time, can will deviate car predicting vehicle
Road, but the actual warning having no in terms of sending vision, the sense of hearing or tactile to driver in the case that consciousness is so done of driver, to carry
Show that driver takes care traveling, effectively reduces accident.
Below in conjunction with the drawings and specific embodiments, the present invention will be further described.
Brief description of the drawings
Fig. 1 is present system block diagram.
Fig. 2 is mounting structure schematic diagram of the present invention.
Fig. 3 is computational methods schematic diagram of the present invention.
Fig. 4 is present system flow chart.
In figure, 1- car bodies, 2- forward sight cameras, 3- warning panels, 4- roads, the road route mark after 5-Hough conversion
Will.
Embodiment
The present embodiment is the preferred embodiment for the present invention, and other all its principle and basic structure are identical or near with the present embodiment
As, within the scope of the present invention.
It refer to the lane-departure warning system based on machine vision in accompanying drawing 1 and accompanying drawing 2, the present invention, including interface
Processing unit, forward sight camera 2, digital signal processing unit and power module, forward sight camera 2 are arranged on before car body 1
Windshield middle and upper part does not block driver's sight and the cleanable position arrived of windscreen wiper, gathers vehicle front image, and image is passed
Be defeated by digital signal processing unit, digital signal processing unit intelligent decision vehicle whether will run-off-road, through data signal
Image information after processing unit processes is transferred to interface processing unit, interface processing unit control warning device work, power supply
Module is given in other module for power supply, the present embodiment, and power module is power interface, and power interface is connected with automobile power source, is passed through
Automobile power source (i.e. automobile batteries, 12V or 24V) give system power supply.In the present embodiment, interface microcontroller processing unit passes through I/O
Mouth or CAN bus buses are connected with the sensor in automobile, are obtained in information of vehicles, the present embodiment, it is necessary to the signal bag obtained
Include：(1) turning indicator control：For analysing whether as driver's active steering；(2) GES：For decision system activation or
Sleep pattern；(3) steering wheel angle：Whether assistant analysis vehicle turns to；(4) side-marker lamp signal：For recognizing that day night pattern is cut
Change；(5) fog lamp signal：For recognizing mist synoptic model；(6) windscreen wiper signal：For recognizing rainy day gas and switch mode；(7) point
Fiery signal：For recognizing whether vehicle launch lights a fire.In the present embodiment, digital signal processing unit is believed using DSP high-speed figures
Number processor, DSP high speed digital signal processors are regarded by not blocking driver installed in the front windshield middle and upper part of car body 1
Line and windscreen wiper it is cleanable to the forward sight camera 2 of position obtain road conditions image in driving, carried out with reference to speed information complicated
Machine Vision Recognition, whether intelligent decision vehicle will run-off-road.DSP high speed digital signal processors can be inclined according to vehicle
Condition of shifting one's love makes interface microcontroller processing unit send different grades of warning message to panel 3 is alerted by CAN bus buses, solely
Vertical warning panel 3 will show different warning messages on panel after warning message is received to driver.
The present invention is connected to power supply and warning panel 3 by wire harness and connector.User need to only plug connector, i.e.,
Plug-and-play is convenient and swift without carrying out complicated setting.
Accompanying drawing 4 is refer to, the method for the lane departur warning of the invention based on machine vision comprises the steps：
A, igniting：System electrification；
Do not moved due to car after B, upper electricity, or speed is too low need not start lane departur warning function, so that system enters
Idle condition；When speed, higher than 10KPH, (i.e. 10 kilometer per hours, be defined as the dynamic and motionless boundary of car in the present invention
Line) when system depart from idle condition；
After C, disengaging idle condition, system detectio side-marker lamp on off state, if side-marker lamp is not opened, camera enters in the daytime
Pattern；
If D, side-marker lamp are opened, camera enters night vision mode；
E, system open camera collection real-time to road conditions and carry out AGC (i.e. brake gain is controlled) according to light intensity,
And the image of distortion is corrected (in the present embodiment, AGC and image flame detection are automatically performed by camera module), and collection
To information with 25 frames (pal mode) per second or 30 frames (TSC-system formula) per second be sent to DSP Processor；
F, DSP Processor are converted into the vision signal received the colour picture of rgb format；
G, YUV are encoded：The image of rgb format is converted into yuv format by matrixer：Obtain luminance signal Y
With two colour difference signal R-Y (i.e. U), B-Y (i.e. V), luminance signal Y and carrier chrominance signal U, V are separated, if only Y-signal
Component is without U, V signal component, then the image then represented is exactly black and white gray level image；
H, picture breakdown are two parts：Because the picture useful information that camera is collected is：Track in image the latter half,
To reduce the amount of calculation of arithmetic unit, track analysis (i.e. step K) the erroneous judgement interference that reduction top half data are produced, the present embodiment
In, picture be divided into above and below two parts, the latter half is only retained when processing and is used as calculating；
I, pattern analysis：Confirm weather conditions：Fine, rain, mist, snow, etc.；
If J, wiper system or fog lamp system are opened, it is rain, mist, snowy day gas that system, which will be analyzed, now because misty rain is avenged
Block vision, causes drawing unintelligible, it is necessary to carry out image noise reduction using digital filter, if without unlatching, this is directly entered K streams
Journey；
K, lane identification analysis：Using Hough transform algorithm, rim detection identification is carried out to track, if having detected car
Road then enters deviation identification process L；If otherwise without track, system enters idle condition；
L, deviation identification：
It refer to accompanying drawing 3, the track identified analyzed in the present invention, in units of 1 pixel, works as X1>α and θ 1
When being spent more than β, then illustrate deviation left-lane,
The deviation right lane when X2 is more than β ' less than α ', and θ 2；
In the present embodiment, X1 is the extended line of left-lane and the intersection point of image bottom line to the distance of image left side edge, X2
The distance for image left side edge that to be right lane extended line arrive with the intersection point of image bottom line, θ 1 is extended line and the image bottom of left-lane
The angle of line, θ 2 is the angle of right lane extended line and image bottom line, when α is normal vehicle operation, the extended line of left-lane with
The intersection point of image bottom line to the distance of image left side edge maximum limit, when β is normal vehicle operation, the extension of left-lane
The maximum of line and the angle of image bottom line, when α ' is normal vehicle operation, the extended line of right lane and the intersection point of image bottom line
When to the minimum limit value and β ' of the distance of image left side edge being normal vehicle operation, extended line and the image bottom line of right lane
Angle maximum, α, β, α ' and β ' value need to according to actual installation timing setting (according to different automobile types, different installation positions
Put, its numeral is different)；
If vehicle does not have run-off-road, return to idle condition and detect again；
If detecting deviation, step M is entered；
M, when track is that will deviate from, detect turning indicator control state：
If the steering indicating light of corresponding orientation is opened, then it is assumed that is that driver actively switches circuit, is then returned to idle condition；
If conversely, without unlatching corresponding orientation steering indicating light, then it is assumed that the unconscious tangent line of driver；
N, steering wheel inclination angle detection：Whether assistant analysis vehicle has steering behavior；
If O, network analysis result are unconscious tangent line, sound, image, vibrations etc. are sent according to different danger classes
Caution signal, or three kinds of signals are simultaneously emitted by.
The present invention use advanced computer vision algorithm for pattern recognition, with reference to high speed digital signal processor to installed in
Camera driving road-condition on car is analyzed in real time, can predict vehicle will run-off-road, but driver is actual has no consciousness
Warning in terms of sending vision, the sense of hearing or tactile to driver in the case of so doing, to point out driver to take care traveling, has
Reduction accident is imitated to occur.
Claims (4)
1. a kind of blind spot vehicle detection and method for warming based on machine vision, using the lane departur warning based on machine vision
System realizes that described system includes interface processing unit, forward sight camera, digital signal processing unit and power module, institute
The forward sight camera stated does not block driver's sight and the cleanable position arrived of windscreen wiper installed in shield glass middle and upper part, adopts
Collect vehicle front image, and by image transmitting to digital signal processing unit, digital signal processing unit intelligent decision vehicle is
It is no will run-off-road, the image information after being handled through digital signal processing unit is transferred to interface processing unit, interface processing
Unit control warning device work, power module gives other module for power supply, it is characterized in that：Described method comprises the steps：
A, igniting：System electrification；
B, before automobile is not walked or the too low system of speed enters idle pulley, can't now start lane departur warning work(
Energy；
After C, disengaging idle condition, system detectio side-marker lamp on off state, if side-marker lamp is not opened, camera enters day mode；
If D, side-marker lamp are opened, into night vision mode；
E, system open camera collection real-time to road conditions and carry out AGC according to light intensity, and the image of distortion is rectified
Just, it is the information collected is per second or 30 frames are per second is sent to digital signal processing unit with 25 frames；
F, digital signal processing unit are converted into the vision signal received the colour picture of rgb format；
G, YUV are encoded:The image of rgb format is converted into yuv format by matrixer：Obtain luminance signal Y and two
Individual colour difference signal R-Y, i.e. U, B-Y, i.e. V, separate luminance signal Y and carrier chrominance signal U, V；
H, picture breakdown are two parts up and down, and the latter half is only retained when processing and is used as calculating；
I, pattern analysis：Confirm weather conditions；
If J, wiper system or fog lamp system are opened, it is rain, mist, snowy day gas that system, which will be analyzed, now using digital filter
Carry out image noise reduction；If wiper system or fog lamp system are directly entered K flows without unlatching；
K, lane identification analysis：Using Hough transform algorithm, rim detection identification is carried out to track：If track has been detected
Into deviation identification process L；If otherwise without track, system enters idle condition；
L, deviation identification：
The track identified is analyzed, in units of 1 pixel, works as X1>α and when θ 1 spends more than β, the left car of deviation
Road；
The deviation right lane when X2 is more than β ' less than α ', and θ 2；
Wherein, X1 is the extended line of left-lane with the intersection point of image bottom line to the distance of image left side edge, and X2 is that right lane prolongs
The intersection point of long line and image bottom line is to the distance of image left side edge, and θ 1 is angle of the extended line with image bottom line of left-lane, θ
2 be the angle of right lane extended line and image bottom line, when α is normal vehicle operation, extended line and the image bottom line of left-lane
Intersection point to the distance of image left side edge maximum limit, when β is normal vehicle operation, the extended line of left-lane and image bottom
The maximum of the angle of line, when α ' is normal vehicle operation, the extended line of right lane and the intersection point of image bottom line to image left side
When the minimum limit value and β ' of the distance at edge are normal vehicle operations, the extended line of right lane and the angle of image bottom line are most
Big value, α, β, α ', and β ' value according to different automobile types, different installation site timings settings；
If vehicle does not have run-off-road, this returns to idle condition and detected again；
If detecting deviation, step M is entered；
M, when track is that will deviate from, detect turning indicator control state：
If the steering indicating light of corresponding orientation is opened, then it is assumed that is that driver actively switches circuit, is then returned to idle condition；
If conversely, without unlatching corresponding orientation steering indicating light, then it is assumed that the unconscious tangent line of driver；
N, steering wheel inclination angle detection：Whether assistant analysis vehicle has steering behavior；
If O, network analysis result are unconscious tangent line, sound, image or vibrations are sent according to different danger classes and alerted
Signal.
2. blind spot vehicle detection and method for warming according to claim 1 based on machine vision, it is characterized in that：Described
Interface processing unit is connected by I/O mouthfuls or CAN bus buses with the sensor in automobile.
3. blind spot vehicle detection and method for warming according to claim 1 based on machine vision, it is characterized in that：Described
Power module is power interface, and power interface is connected with automobile power source, by automobile power source to system power supply.
4. blind spot vehicle detection and method for warming according to claim 1 based on machine vision, it is characterized in that：Described
Car speed is too low to be less than 10KPH for automobile speed per hour.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210486538.9A CN103832357B (en) | 2012-11-23 | 2012-11-23 | A kind of lane-departure warning system and method based on machine vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210486538.9A CN103832357B (en) | 2012-11-23 | 2012-11-23 | A kind of lane-departure warning system and method based on machine vision |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103832357A CN103832357A (en) | 2014-06-04 |
| CN103832357B true CN103832357B (en) | 2017-10-13 |
Family
ID=50796357
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210486538.9A Active CN103832357B (en) | 2012-11-23 | 2012-11-23 | A kind of lane-departure warning system and method based on machine vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103832357B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9764689B2 (en) * | 2014-10-08 | 2017-09-19 | Livio, Inc. | System and method for monitoring driving behavior |
| CN104691333B (en) * | 2015-03-30 | 2017-03-15 | 吉林大学 | Coach driver fatigue state evaluation method based on the deviation frequency |
| CN105539293B (en) * | 2016-02-03 | 2018-01-09 | 北京中科慧眼科技有限公司 | Lane departure warning method and device and car steering accessory system |
| CN106558248A (en) * | 2016-12-13 | 2017-04-05 | 天津泓耘财科技发展有限公司 | A kind of panorama sees deviation prewarning monitoring system |
| CN106828308A (en) * | 2017-01-24 | 2017-06-13 | 桂林师范高等专科学校 | Lane departure warning device |
| CN108630014A (en) * | 2018-05-10 | 2018-10-09 | 苏州天瞳威视电子科技有限公司 | A kind of lane deviates early warning system and method |
| CN108973855B (en) * | 2018-07-19 | 2020-11-24 | 南京地平线机器人技术有限公司 | Method and device for lane departure warning |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101264755A (en) * | 2008-03-06 | 2008-09-17 | 上海交通大学 | Vehicle running safety intelligence monitoring and controlling device |
| CN102303609A (en) * | 2011-06-16 | 2012-01-04 | 广东铁将军防盗设备有限公司 | System and method for prewarning lane deviation |
| CN202413787U (en) * | 2011-12-12 | 2012-09-05 | 浙江吉利汽车研究院有限公司 | Prewarning control system for automobile lane deviation |
| CN203293984U (en) * | 2012-11-23 | 2013-11-20 | 深圳华一汽车科技有限公司 | Lane departure warning system based on machine vision |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE10307607A1 (en) * | 2003-02-22 | 2004-09-09 | Daimlerchrysler Ag | Distributed image processing system for motor vehicles |
-
2012
- 2012-11-23 CN CN201210486538.9A patent/CN103832357B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101264755A (en) * | 2008-03-06 | 2008-09-17 | 上海交通大学 | Vehicle running safety intelligence monitoring and controlling device |
| CN102303609A (en) * | 2011-06-16 | 2012-01-04 | 广东铁将军防盗设备有限公司 | System and method for prewarning lane deviation |
| CN202413787U (en) * | 2011-12-12 | 2012-09-05 | 浙江吉利汽车研究院有限公司 | Prewarning control system for automobile lane deviation |
| CN203293984U (en) * | 2012-11-23 | 2013-11-20 | 深圳华一汽车科技有限公司 | Lane departure warning system based on machine vision |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103832357A (en) | 2014-06-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103832357B (en) | A kind of lane-departure warning system and method based on machine vision | |
| CN105799574B (en) | Night intelligence DAS (Driver Assistant System) | |
| CN106652558B (en) | Vehicle-road cooperative intelligent traffic control system | |
| CN103236178B (en) | Signal lamp pattern-recognition system for prompting and method | |
| WO2018058958A1 (en) | Road vehicle traffic alarm system and method therefor | |
| CN101984478B (en) | Abnormal S-type driving warning method based on binocular vision lane marking detection | |
| CN201325462Y (en) | Train intelligent traffic monitoring system based on driven thermal infrared imager | |
| CN103723073B (en) | A kind of danger of Fusion is overtaken other vehicles prior-warning device and method for early warning thereof | |
| CN108074410A (en) | Traffic cloud warns platform | |
| CN205768952U (en) | Night intelligence auxiliary driving device | |
| CN203332106U (en) | Automobile warning anti-collision system based on monocular vision | |
| CN110782673A (en) | Vehicle violation identification and detection system based on unmanned aerial vehicle shooting cloud computing | |
| CN104527427A (en) | Velocity-limiting mark recognizing and vehicle-velocity control method and system based on vehicle traveling recording instrument | |
| CN206805801U (en) | Hazardous vehicles cross the street caution system | |
| CN113487874A (en) | System and method for collecting, identifying and classifying following behavior scene data | |
| CN103832380A (en) | Dead zone automobile detection and warning system and method based on machine vision | |
| CN110838231A (en) | Pedestrian crossing intelligent detection system and method | |
| CN104228670B (en) | The implementation method of a kind of drive recorder for oversize vehicle | |
| CN104361648A (en) | Event data recorder provided with signal lamps for reminding other vehicles and control method of event data recorder | |
| CN203293984U (en) | Lane departure warning system based on machine vision | |
| CN206524010U (en) | A kind of motor line is through automatic reminding system during crossing | |
| CN206541437U (en) | A kind of traffic detecting system | |
| CN108621924A (en) | A kind of control system for rear light and its control method with prompt facility | |
| CN114954307A (en) | Driving assistance system based on artificial intelligence | |
| CN108010324A (en) | Highway and Emergency Vehicle Lane traffic alarm system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |