CN105654073A - Automatic speed control method based on visual detection - Google Patents
Automatic speed control method based on visual detection Download PDFInfo
- Publication number
- CN105654073A CN105654073A CN201610176917.6A CN201610176917A CN105654073A CN 105654073 A CN105654073 A CN 105654073A CN 201610176917 A CN201610176917 A CN 201610176917A CN 105654073 A CN105654073 A CN 105654073A
- Authority
- CN
- China
- Prior art keywords
- speed
- vehicle
- image
- limit
- motor 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 46
- 238000001514 detection method Methods 0.000 title claims abstract description 38
- 230000000007 visual effect Effects 0.000 title abstract description 3
- 238000009826 distribution Methods 0.000 claims description 7
- 238000005070 sampling Methods 0.000 claims description 7
- 238000012549 training Methods 0.000 claims description 5
- 238000003708 edge detection Methods 0.000 claims description 3
- 238000011478 gradient descent method Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000004519 manufacturing process Methods 0.000 claims 1
- 230000006870 function Effects 0.000 description 11
- 238000012545 processing Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 3
- 238000005286 illumination Methods 0.000 description 3
- 230000033001 locomotion Effects 0.000 description 3
- 230000000903 blocking effect Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000001537 neural effect Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000001149 cognitive effect Effects 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 108010052322 limitin Proteins 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000013011 mating Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000011897 real-time detection Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000035807 sensation Effects 0.000 description 1
- 230000009466 transformation Effects 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
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/582—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of traffic signs
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W30/00—Purposes of road vehicle drive control systems not related to the control of a particular sub-unit, e.g. of systems using conjoint control of vehicle sub-units
- B60W30/18—Propelling the vehicle
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W40/00—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
- B60W40/10—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to vehicle motion
- B60W40/105—Speed
-
- 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
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/588—Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/09—Recognition of logos
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Transportation (AREA)
- Mechanical Engineering (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Traffic Control Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an automatic speed control method based on visual detection. The automatic speed control method comprises the steps that 1, statistics is conducted on the positions of traffic signs in an image under the situation that the distance between a vehicle and a right side lane and the distance between the vehicle and a left side lane are different to obtain a table, a function f1 of the geometric deformation degree epsilon of the traffic signs and road corners theta and a function f2 of the critical speeds vcurlim of curves and the theta are established; 2, the edge line position L1 of the right side lane of the vehicle to be controlled and the position L2 away from a left-side vehicle are detected; 3, image areas including the traffic signs are selected from the image according to the L1, L2 and the table; 4, the road corners theta in the image are calculated by integrating two methods; 5, the geometric deformation degree epsilon of the traffic signs in the image are calculated according to the f1 and theta, the image is restored to detect speed limit signs and identify limit speed values v1; 6, the speed of the vehicle is controlled according to the relative speed of the vehicle in front, the speeds of vehicles on both sides in front, the v1 and the vcurlim. By the adoption of the automatic speed control method, the cost is saved while the system reliability is improved.
Description
Technical field
The invention belongs to technical field of computer vision and sensation target detection field, relate to a kind of based on computer vision and imageThe speed of a motor vehicle autocontrol method of processing.
Background technology
Along with intelligent transportation, unmanned and drive the development of ancillary technique, the visual capacity that improves automobile seems particularly important.Most achievements in research concentrate on the detection and Identification to traffic sign now, and then realize speed of a motor vehicle adjusting, but in complicated actual fieldUnder scape, be still difficult to the reliability that reaches very high.
Most traffic sign detects recognition system and made Preliminary detection based on color before this, then does further detection based on shapeLocation.
Preliminary detection main method based on color has based on color cluster, multichannel Threshold segmentation (RGB+HIS, the method applicationThe most extensive), color index (mating with ATL by histogram), is subject to illumination and around but these methods detect performancesAmbient influnence is larger, and under complex scene, accuracy rate is not high; Or employing neural metwork training (to the color classification of traffic sign,Shades of colour is converted into several basic colors, illumination and noise is had to certain robustness), but this method processing time isLong, be difficult to meet the requirement of real-time detecting.
Detecting main method based on shape has template matches (amount of calculation is large, and is subject to the influence of crust deformation huge), and Hough changes(consumption storage consuming time), corner detection (robustness is low, detects effect poor under complex scene), based on neural metwork training (consumptionTime), mathematical Morphology Algorithm (opening and closing operation, extracts skeleton function, and computing time is long).
Except traffic sign is detected identification, road scene information detects also very important for speed of a motor vehicle control, and the one, due to notUnder same road surface (pavement grade, wet and slippery degree), the criticality safety speed that different bend curvature is corresponding different and traffic routeSelect, common algorithm is that lane line is detected, but this method is mainly for open section, for urban trafficLarge section, is affected by building trees shade and nearby vehicle, and lane line is subject to coverage extent serious, must design new detection method;The 2nd, in bend section, city, due to spatial limitation, radius of turn is generally less, causes traffic sign deformation degree in frame of videoHave a strong impact on Mark Detection; The 3rd, under actual scene, road scene complexity, maximum speed limit many factors, only examines according to labelSurvey is difficult to meet requirements for automatic control, and automotive multipurpose GPS carries out speed detection in addition, (as in tunnel, vertical under a lot of situationsHand under bridge, when both sides building is higher), can affect gps signal and receive, be difficult to carry out real-time speed detection, and gather around in trafficStifled section, automobile many places are at a slow speed or idling mode, and now absolute velocity Clinical significance of detecting is little.
The defect of prior art
1. under outdoor environment, affected by road actual scene and self travel conditions, the traffic sign collecting easily deforms,Make accuracy rate can not reach very high;
The whole detection identifying processing time longer, be difficult to realize real-time detection;
3. there is blocking of a large amount of natural scenes and culture in actual scene, lacks effective detection method.
Summary of the invention
For the technical problem existing in prior art, the object of the present invention is to provide a kind of speed based on vision-based detection automaticControl method.
Technical scheme of the present invention is:
Based on a speed automatic control method for vision-based detection, the steps include:
1) under calculating vehicle and right-hand lane line and left side vehicle different distance, the position of traffic sign in image, obtains a tableLattice; And based on sampling training, set up the function f between traffic sign geometry deformation degree ε and road corner θ1, curvedRoad critical speed vcurlimAnd the function f between road corner θ2;
2) from the image of vehicle collection to be controlled, detect the right-hand lane edge line position L of this vehicle to be controlled1To be controlled with thisThe position L of vehicle distances processed left side vehicle2;
3) according to the L detecting1、L2From this image, select the image-region that comprises traffic sign with this form;
4) this image-region is carried out to many color spaces and cut apart, remove that color characteristic in this image-region does not mate, Color-Connected districtRegion and Color-Connected region that territory is greater than capping are less than the region of setting lower limit;
5) according to getting location parameter L in this image of frame per second q, front vehicles and the velocity information v of front vehiclesr1CalculateRoad corner θ in this image1; Adopt edge detection method to obtain the road corner in this image according to lane line positionθ2; By θ1And θ2Be weighted and obtain final road corner θ;
6) according to function f1Calculate the traffic sign geometry deformation degree ε in this image with road corner θ, then according to this ε valueThis image is recovered; Then from the image recovering, detect speed(-)limit sign and identify speed limit velocity amplitude v1;
7) according to the relative speed of a motor vehicle v of front vehiclesf, both sides, front vehicle speed of a motor vehicle v2And the speed limit velocity amplitude v identifying1, curvedRoad critical speed vcurlimControl together the speed of this vehicle to be controlled.
Further, the method that detects speed(-)limit sign is:
21) in this image, get 2 points, determine this gradient vector of 2 according to sobel operator, using gradient vector intersection point as circleHeart position, calculate this center of circle respectively with the distance of 2, be less than setting value if this two segment distance differs, get radius rFor the average of this two segment distance; Otherwise again choose 2 definite centers of circle and radius r;
22) do circle with the radius of determining, when pixel in its arc-shaped edges setting range outnumber threshold value η time, judge shouldIn circle, there is circle, then by detecting whether have a red oblique line and letter or number height, to prohibitory sign, limitHeight, limit for width and speed(-)limit sign carry out Fast Classification.
Further, the establishing method of described threshold value η is: road is divided into polytype, every kind of road type blocked journeyDegree γ Gaussian distributed; For every kind of road type Si, one group of number { a is seti,bi,δi, then according to formulaDetermine road type SiCorresponding threshold value η; Wherein,For confidence level,γ~N(air+bi,r2δi),Calculate according to standard Gaussian distribution inverse function, N represents its Gaussian distributed.
Further, described relative speed of a motor vehicle vfDetection method be: first sampling statistics obtain before and after vehicle relative velocity distributeThen identify the car plate of front vehicles and deconvolute and obtain corresponding original image Represent vrCross stream component,Represent vrLongitudinal component; Then try to achieve optimization aim function with gradient descent methodThe ν that maximum is correspondingrBe the relative speed of a motor vehicle v of front vehiclesf。
Further, both sides, described front car speed v2Detection method be: first adopt template matching method to obtain in imageThe left and right sides, front vehicle location, then adopts frame differential method to process both sides, the front vehicle of consecutive frame image, obtainsThe relative speed of a motor vehicle v of both sides, front vehicle2。
Further, described step 7) in, according to the relative speed of a motor vehicle v of front vehiclesfSpeed of a motor vehicle v with both sides, front vehicle2, with knowledgeThe speed limit velocity amplitude v not going out1, bend critical speed vcurlimThe method of controlling together the speed of this vehicle to be controlled is: according to phaseTo speed of a motor vehicle vf, regulation speed and front vehicles keep setting spacing; As speed of a motor vehicle v2While being greater than setting threshold ε 2, vehicle is subtractedSpeed.
Compared with prior art, good effect of the present invention is:
1. the understanding based on to road scene, selects processing region, promotes the detection efficiency of traffic sign;
2. research corner traffic sign geometric shape changes, and promotes accuracy in detection;
3. propose improved Hough and change, realize fast the detection to speed(-)limit sign circular edge, blocking under environment, also haveGood detection speed and accuracy rate;
4. carry out speed control in conjunction with peripheral path vehicle, cost-saving in Hoisting System reliability.
Brief description of the drawings
Fig. 1 is the anti-hypervelocity autocontrol method flow chart based on vision-based detection of the present invention;
Fig. 2 is speed control flow chart of the present invention;
Fig. 3 is that traffic sign of the present invention detects identification process figure;
Fig. 4 is road scene detection flow for the automobile figure of the present invention.
Detailed description of the invention
Below in conjunction with accompanying drawing, the present invention is explained in further detail, method flow of the present invention as shown in Figure 1, concrete speedDegree control flow as shown in Figure 2.
1, as shown in Figure 3, traffic sign detection and Identification:
The first step: affected by the factor such as weather, real road environment, get after frame picture and easily produce the fuzzy and distortion of color, firstIt is carried out to pretreatment, eliminate illumination, motion blur and influence of noise (Wiener filtering, histogram equalization), then to collectingImage based on front vehicles and self driving situation, detect right-hand lane edge line position L by edge detection operator1(wagon flowSmaller part is very effective), and a simple template matches roughly finds the approximate location L of left side vehicle2(mainly for vehicle flowrateLarger section). And for different L1,L2, in frame of video traffic sign position have different distributions (for label position,Generally be positioned at picture upper right side; And when vehicle distances right-hand lane edge line is near or time far away apart from left side vehicle distances, traffic markWill position taking directly over as main), according to L1,L2, select priority treatment region (in distances different from left side vehicle lane lineFrom under, the position of statistics traffic sign in image is tabulating; Again according to current L1, L2 and this form from present imageChoose priority treatment region, comprise the image-region of traffic sign), then cut apart through too much color space, further remove this districtIn territory color characteristic do not mate the part excessive or too small with area (Color-Connected cluster exceed picture in its entirety area more than 1/4 orBelow 1/80, directly remove, stay and be communicated with the region of area between two values), cut apart as for many color spaces, be mainlyAccording to the RGB of the each pixel of image and HSV numerical value, judge compared with pre-determined threshold value (specifically with reference to W.G.Shaded,D.I.Abu-Al-Nadi,M.J.Mismar,RoadTrafficSignDetectioninColorImages,Proceedingsofthe10thIEEEInternationalConferenceonElectronics,CircuitsandSystems,2003(2):890-893;X.W.Gao,L.Podladchikova,D.Shaposhnikov,K.Hong,N.Shevtso.RecognitionofTrafficSignsBasedontheirColourandShapeFeaturesExtractedUsingHumanVisionModels,JournalofVisualCommunicationandImageRepresentation,2006(17):675-685;M. Benallal,J.Meunier.Real-timeColorSegmentationofRoadSigns,CCECE,2003(3):1823-1826;)。
Under actual scene, traffic sign presents geometry deformation in various degree in frame of video, and straight line section, becomes hardlyShape; Turn round section especially at bend place, city, and due to spatial limitation, radius of turn is less, presents certain on widthDimensional variation. Based on actual scene sampling training, obtain traffic sign geometry deformation degree (aspect ratio threshold according to River Bend ModelsValue scope ε) and bend critical speed vcurlimAnd the function between road corner θ is respectively f1,f2:
ε=f1(θ);vcurlim=f2(θ);(1)
Detect for the road corner relating to above, calculate (every kind of method has himself pluses and minuses) finally according to two kinds of methodsBy its combination:
One: carry out front vehicles detection for picture region remaining after above-mentioned pretreatment; The section place of turning round, front vehiclesGenerally be positioned at about picture near middle, position coordinates L according to front vehicles in picture, get frame per second q and (get frame per second qBe a determined value) and the velocity information v of front vehiclesr1, be different from straight line section, be positioned at the front vehicles at road corner placeIts speed vr1Generally be difficult to direct-detection, first obtain non-intersection speed by the method for statistics and distribute, simulate by a Gaussian distribution,Be designated as P (vr1), the road corner detected value calculating by the method is (v represents self current vehicle speed):
θ1=∫f3(L,qv,vr1)P(vr1)dvr1(2)
Two: by rim detection, obtain turning information θ according to lane line position2(with reference to Wang Jianwen, An Xiangjing. a kind of distributionFormula bend detection algorithm. national pattern-recognition academic conference, 2010; OthmanO.Khalifa, ImranMoezKhan,AbdulhakamA.M.Assidiq,AHyperbola-pairBasedLaneDetectionSystemforVehicleGuidance,WorldCongressonEngineeringandComputerScience,Sanfrancisco,USA,2010:585-588)。
Finally by θ1,θ2Combine and determine that turning, section value (concrete combination coefficient is added up and obtained by actual samples) is accurately:
θ=a·θ1+β·θ2(3)
Calculate the aspect ratio threshold range ε of speed(-)limit sign in image according to formula (1) by this turning, more each logical to pictureIts aspect ratio (region boundary rectangle length-width-ratio) is calculated in region, and both compare, and judges whether this aspect ratio drops on speed limitIn the threshold range of mark aspect ratio, can further eliminate most of irrelevant region.
Second step: because front portion has been removed most of non-circle diagram shape, do after affine transformation by aspect ratio threshold value above,Speed(-)limit sign can be reverted to actual circle, adopt following improved Hough conversion (in noise distortion and subregion incompletenessIn situation, still can obtain desirable effect) can realize location fast and accurately:
According to travel situations, speed(-)limit sign radius is gathered in variation within the scope of very little, therefore can first get 2 points, getsCertain radius determines that (determine this gradient vector of 2 according to sobel operator, gradient vector intersection point, is the center of circle to home positionPosition, calculates the center of circle and 2 distances, is less than 5 pixels if this two segment distance differs, and getting radius is that this two segment distance is equalValue; Otherwise think that on same circle, now reselects at 2 and calculates at these 2), then do circle with this radius, calculate circleWhether the arc limit among a small circle number of interior pixel exceedes certain threshold value η (consideration street periphery object circumstance of occlusion, concrete computational processIn, according to pre-determined numerical value, when a side occurs that one section when discontinuous, then detect the symmetric position point in its relative center of circle,If these both sides still occur one section discontinuous, be non-circle), judge whether get rid of remaining area behind irrelevant region exists circleShape mark, the method can be dwindled detection time greatly, Hoisting System real-time.
For determining of threshold value η above, because circle is larger, its respective threshold should be also larger, but simultaneously its blocked canEnergy property is also just larger. Block possibility relevant with current road surrounding enviroment, be divided into general urban road, highway, crowdedBlocks etc. are several, are designated as { S1,S2,...Sn, for every kind of situation, establish and be subject to coverage extent γ Gaussian distributed (coverage extentγ and threshold value η inverse function each other), for every kind of scene Si, actual samples can be calculated one group of corresponding number { ai,bi,δi, establish circleRadius is r,
γ~N(air+bi,r2δi)(4)
Wherein, N represents its Normal Distribution, { ai,bi,δiSampling to add up obtains.
Reliability (reliability requirement) is set isUnder this scene, there is threshold value
Calculate according to standard Gaussian distribution inverse function, it is surrounding pixel point number to be greater than this threshold value if detectA circle, otherwise be not.
The 3rd step: cognitive phase: prohibitory sign, freight weight limit, limit for width and speed(-)limit sign are all Hong Quan edge, for prohibitory signHave one from the fixing red line of angle left to bottom right, can detect red line, determine whether it is rapidly a prohibitory sign;For limit for height limit for width mark, compared with speed(-)limit sign, there is the much smaller letter of a short transverse " t " or " m " in digital lower right corner place,Therefore can detect the black connected region of label, whether more each black region height is consistent, and consistent words are speed(-)limit sign,Otherwise be not.
Determine behind speed(-)limit sign position because velocity amplitude is not a lot, can adopt the method for template matches (database storage in advance)Determine mark numerical value, be convenient to strengthen the degree of accuracy and the stability of identification.
2, as shown in Figure 4, vehicle detection in road scene:
For the above-mentioned GPS defect that tests the speed, especially, under the congestion of city, detect surrounding vehicles travel situations more meaningful: pass throughDetect the relative speed of a motor vehicle v with front vehiclesf(comparing with threshold epsilon 1), with the certain spacing of its maintenance, detect in addition a left side, front, crossingRight both sides automobile speed v2, in the time that the both sides speed of a motor vehicle is larger, (compare with threshold epsilon 2), automobile is slowed down. For cost-saving,In the testing process of previous step speed(-)limit sign, the relative velocity of detection simultaneously and front and the left and right sides, front vehicle.
For with the relative velocity of front vehicles, select front vehicles car plate to detect, first know that it is mainly distributed in a left side(right side) below (relevant with camera installation site), yellow end surplus rectangle, after image pretreatment, cuts apart according to many color spacesCan get rid of most irrelevant parts, then get suitable mask corner detection is carried out to (with reference to Arturodela in this regionEscalera,MiguelAngelSalichs,RoadTrafficSignDetectionandClassification,IEEETransactionOnIndustrialElectronics, 1997 (44): 848-859; ), can effectively determine licence plate position. Certainly have big difference in the speed of a motor vehicle, distanceIn the situation close to, this region can produce motion blur, and knows that its point spread function is relevant with speed of related movement, is to saveIn the processing time, only, to longitudinally carrying out deblurring processing, can accurately draw relative velocity according to adjacent two frames, and concrete steps are as follows:
(1) in advance sampling statistics obtain before and after vehicle relative velocity be distributed as
For saving time and ensureing recognition accuracy, get vr∈(-3δr,3δr). N represents Gaussian distributed, δrRepresent its sidePoor, learn by prior sampling statistics.
(2) because the longitudinal majority around of front vehicles car plate is pure color, therefore, in the time that deblurring is processed, there is not ringing,Can directly deconvolute and obtain possible original image(Represent vrCross stream component;Represent vrLongitudinal component)Optimization aim function is
Try to achieve ν corresponding to maximum with gradient descent methodr, be the relative speed of a motor vehicle v of front vehiclesf。
For the left and right sides, front, crossing vehicle, know that it generally appears at top (right side) to the left and locates. Because only needing to judge car speedWhether larger, without obtaining its accurate numerical value, carry out can specifying the left and right sides, front car after simple template matches with auto modelPosition; Then can directly adopt frame differential method to carry out simple process and obtain fast both sides speed of a motor vehicle v2, this method is to rush horse by mistakeThe pedestrian on road also has good detection effect, can effectively avoid accident to occur.
(3) according to the relative speed of a motor vehicle v of front vehiclesfSpeed of a motor vehicle v with both sides, front vehicle2, with the speed limit velocity amplitude v identifying1、Bend critical speed vcurlimControl together the speed of this vehicle to be controlled, concrete control flow as shown in Figure 2.
Claims (6)
1. the speed automatic control method based on vision-based detection, the steps include:
1) under calculating vehicle and right-hand lane line and left side vehicle different distance, the position of traffic sign in image, obtains a tableLattice; And based on sampling training, set up the function f between traffic sign geometry deformation degree ε and road corner θ1, curvedRoad critical speed vcurlimAnd the function f between road corner θ2;
2) from the image of vehicle collection to be controlled, detect the right-hand lane edge line position L of this vehicle to be controlled1To be controlled with thisThe position L of vehicle distances processed left side vehicle2;
3) according to the L detecting1、L2From this image, select the image-region that comprises traffic sign with this form;
4) this image-region is carried out to many color spaces and cut apart, remove that color characteristic in this image-region does not mate, Color-Connected districtRegion and Color-Connected region that territory is greater than capping are less than the region of setting lower limit;
5) according to getting location parameter L in this image of frame per second q, front vehicles and the velocity information v of front vehiclesr1CalculateRoad corner θ in this image1; Adopt edge detection method to obtain the road corner in this image according to lane line positionθ2; By θ1And θ2Be weighted and obtain final road corner θ;
6) according to function f1Calculate the traffic sign geometry deformation degree ε in this image with road corner θ, then according to this ε valueThis image is recovered; Then from the image recovering, detect speed(-)limit sign and identify speed limit velocity amplitude v1;
7) according to the relative speed of a motor vehicle v of front vehiclesf, both sides, front vehicle speed of a motor vehicle v2And the speed limit velocity amplitude v identifying1, curvedRoad critical speed vcurlimControl together the speed of this vehicle to be controlled.
2. the method for claim 1, is characterized in that, the method that detects speed(-)limit sign is:
21) in this image, get 2 points, determine this gradient vector of 2 according to sobel operator, using gradient vector intersection point as circleHeart position, calculate this center of circle respectively with the distance of 2, be less than setting value if this two segment distance differs, get radius rFor the average of this two segment distance; Otherwise again choose 2 definite centers of circle and radius r;
22) do circle with the radius of determining, when pixel in its arc-shaped edges setting range outnumber threshold value η time, judge shouldIn circle, there is circle, more whether exist a red oblique line whether consistent with letter or number height by detecting, to forbiddingMark, limit for height, limit for width and speed(-)limit sign carry out Fast Classification.
3. method as claimed in claim 2, is characterized in that, the establishing method of described threshold value η is: road is divided into multiple typesType, every kind of road type be subject to coverage extent γ Gaussian distributed; For every kind of road type Si, one group of number is set{ai,bi,δi, then according to formulaDetermine road type SiCorresponding threshold value η;Wherein,For confidence level, γ~N (air+bi,r2δi),Calculate N representative according to standard Gaussian distribution inverse functionIt is a Gaussian distribution.
4. the method as described in claim 1 or 2 or 3, is characterized in that, described relative speed of a motor vehicle vfDetection method be: first adoptBefore and after sample statistics obtains, vehicle relative velocity distributesThen identify the car plate of front vehicles and remove volumeAmass and obtain corresponding original image Represent vrCross stream component,Represent vrLongitudinal component; ThenTry to achieve optimization aim function with gradient descent methodThe v that maximum is correspondingrBe front vehicles relativeSpeed of a motor vehicle vf。
5. the method as described in claim 1 or 2 or 3, is characterized in that, both sides, described front car speed v2Detection method be:First adopt template matching method to obtain the left and right sides, front vehicle location in image, then adopt frame differential method to consecutive frameBoth sides, the front vehicle of image is processed, and obtains the relative speed of a motor vehicle v of both sides, front vehicle2。
6. the method as described in claim 1 or 2 or 3, is characterized in that described step 7) in, according to front vehicles relativelySpeed of a motor vehicle vfSpeed of a motor vehicle v with both sides, front vehicle2, with the speed limit velocity amplitude v identifying1, bend critical speed vcurlimOne starts to controlThe method of making the speed of this vehicle to be controlled is: according to relative speed of a motor vehicle vf, regulation speed and front vehicles keep setting spacing;As speed of a motor vehicle v2While being greater than setting threshold ε 2, vehicle is slowed down.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610176917.6A CN105654073B (en) | 2016-03-25 | 2016-03-25 | A kind of speed automatic control method of view-based access control model detection |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610176917.6A CN105654073B (en) | 2016-03-25 | 2016-03-25 | A kind of speed automatic control method of view-based access control model detection |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105654073A true CN105654073A (en) | 2016-06-08 |
CN105654073B CN105654073B (en) | 2019-01-04 |
Family
ID=56494411
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610176917.6A Active CN105654073B (en) | 2016-03-25 | 2016-03-25 | A kind of speed automatic control method of view-based access control model detection |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105654073B (en) |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106228849A (en) * | 2016-08-19 | 2016-12-14 | 深圳市元征科技股份有限公司 | A kind of vehicle early warning method and mobile terminal |
CN106815558A (en) * | 2016-12-21 | 2017-06-09 | 上海智驾电子科技有限公司 | Car based on image recognition enters tunnel headlight and shifts to an earlier date open method automatically |
CN107066933A (en) * | 2017-01-25 | 2017-08-18 | 武汉极目智能技术有限公司 | A kind of road sign recognition methods and system |
CN108646718A (en) * | 2018-07-03 | 2018-10-12 | 吉林大学 | The detection device of automobile intelligent speed limiting control system |
CN108898857A (en) * | 2018-06-13 | 2018-11-27 | 同济大学 | A kind of intersection motor vehicle green light interval setting method considering security reliability |
CN109325390A (en) * | 2017-08-01 | 2019-02-12 | 郑州宇通客车股份有限公司 | A kind of localization method combined based on map with FUSION WITH MULTISENSOR DETECTION and system |
CN109733196A (en) * | 2019-01-30 | 2019-05-10 | 深圳鸿鹏新能源科技有限公司 | Vehicle and its active safety control method and apparatus |
CN110414385A (en) * | 2019-07-12 | 2019-11-05 | 淮阴工学院 | A kind of method for detecting lane lines and system based on homography conversion and characteristic window |
CN110462437A (en) * | 2017-03-29 | 2019-11-15 | 株式会社电装 | Photodetector |
CN112639808A (en) * | 2018-07-31 | 2021-04-09 | 法雷奥开关和传感器有限责任公司 | Driving assistance for longitudinal and/or lateral control of a motor vehicle |
CN113074749A (en) * | 2021-06-07 | 2021-07-06 | 湖北亿咖通科技有限公司 | Road condition detection and update method, electronic equipment and computer-readable storage medium |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102156977A (en) * | 2010-12-22 | 2011-08-17 | 浙江大学 | Vision-based road detection method |
CN103544484A (en) * | 2013-10-30 | 2014-01-29 | 广东工业大学 | Traffic sign identification method and system based on SURF |
CN104386063A (en) * | 2014-09-19 | 2015-03-04 | 奇瑞汽车股份有限公司 | Driving assistance system based on artificial intelligence |
CN105261224A (en) * | 2015-09-02 | 2016-01-20 | 奇瑞汽车股份有限公司 | Intelligent vehicle control method and apparatus |
US20160034769A1 (en) * | 2014-07-29 | 2016-02-04 | Magna Electronics Inc. | Vehicle vision system with traffic sign recognition |
-
2016
- 2016-03-25 CN CN201610176917.6A patent/CN105654073B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102156977A (en) * | 2010-12-22 | 2011-08-17 | 浙江大学 | Vision-based road detection method |
CN103544484A (en) * | 2013-10-30 | 2014-01-29 | 广东工业大学 | Traffic sign identification method and system based on SURF |
US20160034769A1 (en) * | 2014-07-29 | 2016-02-04 | Magna Electronics Inc. | Vehicle vision system with traffic sign recognition |
CN104386063A (en) * | 2014-09-19 | 2015-03-04 | 奇瑞汽车股份有限公司 | Driving assistance system based on artificial intelligence |
CN105261224A (en) * | 2015-09-02 | 2016-01-20 | 奇瑞汽车股份有限公司 | Intelligent vehicle control method and apparatus |
Non-Patent Citations (2)
Title |
---|
丁淑艳等: ""基于模糊形状判别的鲁棒交通标志检测算法"", 《机电一体化》 * |
左丹丹等: ""仿射变换在交通标志检测中的应用"", 《宁波大学学报(理工版)》 * |
Cited By (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106228849A (en) * | 2016-08-19 | 2016-12-14 | 深圳市元征科技股份有限公司 | A kind of vehicle early warning method and mobile terminal |
CN106815558A (en) * | 2016-12-21 | 2017-06-09 | 上海智驾电子科技有限公司 | Car based on image recognition enters tunnel headlight and shifts to an earlier date open method automatically |
CN106815558B (en) * | 2016-12-21 | 2020-06-30 | 上海智驾汽车科技有限公司 | Automatic early starting method for tunnel headlamp of vehicle based on image recognition |
CN107066933A (en) * | 2017-01-25 | 2017-08-18 | 武汉极目智能技术有限公司 | A kind of road sign recognition methods and system |
CN107066933B (en) * | 2017-01-25 | 2020-06-05 | 武汉极目智能技术有限公司 | Road sign identification method and system |
CN110462437A (en) * | 2017-03-29 | 2019-11-15 | 株式会社电装 | Photodetector |
CN109325390A (en) * | 2017-08-01 | 2019-02-12 | 郑州宇通客车股份有限公司 | A kind of localization method combined based on map with FUSION WITH MULTISENSOR DETECTION and system |
CN108898857A (en) * | 2018-06-13 | 2018-11-27 | 同济大学 | A kind of intersection motor vehicle green light interval setting method considering security reliability |
CN108646718A (en) * | 2018-07-03 | 2018-10-12 | 吉林大学 | The detection device of automobile intelligent speed limiting control system |
CN108646718B (en) * | 2018-07-03 | 2023-11-10 | 吉林大学 | Detection equipment of intelligent speed limiting control system of automobile |
CN112639808A (en) * | 2018-07-31 | 2021-04-09 | 法雷奥开关和传感器有限责任公司 | Driving assistance for longitudinal and/or lateral control of a motor vehicle |
CN112639808B (en) * | 2018-07-31 | 2023-12-22 | 法雷奥开关和传感器有限责任公司 | Driving assistance for longitudinal and/or transverse control of a motor vehicle |
CN109733196A (en) * | 2019-01-30 | 2019-05-10 | 深圳鸿鹏新能源科技有限公司 | Vehicle and its active safety control method and apparatus |
CN110414385A (en) * | 2019-07-12 | 2019-11-05 | 淮阴工学院 | A kind of method for detecting lane lines and system based on homography conversion and characteristic window |
CN110414385B (en) * | 2019-07-12 | 2021-06-25 | 淮阴工学院 | Lane line detection method and system based on homography transformation and characteristic window |
CN113074749A (en) * | 2021-06-07 | 2021-07-06 | 湖北亿咖通科技有限公司 | Road condition detection and update method, electronic equipment and computer-readable storage medium |
CN113074749B (en) * | 2021-06-07 | 2021-08-20 | 湖北亿咖通科技有限公司 | Road condition detection and update method, electronic equipment and computer-readable storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN105654073B (en) | 2019-01-04 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105654073A (en) | Automatic speed control method based on visual detection | |
CN109299674B (en) | Tunnel illegal lane change detection method based on car lamp | |
CN107766821B (en) | Method and system for detecting and tracking full-time vehicle in video based on Kalman filtering and deep learning | |
CN111563412B (en) | Rapid lane line detection method based on parameter space voting and Bessel fitting | |
CN104008645B (en) | One is applicable to the prediction of urban road lane line and method for early warning | |
CN105913041B (en) | It is a kind of based on the signal lamp recognition methods demarcated in advance | |
CN109670376B (en) | Lane line identification method and system | |
CN103984950B (en) | A kind of moving vehicle brake light status recognition methods for adapting to detection on daytime | |
CN101783076B (en) | Method for quick vehicle type recognition under video monitoring mode | |
CN109190523B (en) | Vehicle detection tracking early warning method based on vision | |
US8670592B2 (en) | Clear path detection using segmentation-based method | |
CN104657727B (en) | A kind of detection method of lane line | |
CN105488454A (en) | Monocular vision based front vehicle detection and ranging method | |
CN104899554A (en) | Vehicle ranging method based on monocular vision | |
CN104392212A (en) | Method for detecting road information and identifying forward vehicles based on vision | |
CN107316486A (en) | Pilotless automobile visual identifying system based on dual camera | |
CN110298216A (en) | Vehicle deviation warning method based on lane line gradient image adaptive threshold fuzziness | |
CN105005771A (en) | Method for detecting full line of lane based on optical flow point locus statistics | |
CN104008377A (en) | Ground traffic sign real-time detection and recognition method based on space-time correlation | |
CN105678285A (en) | Adaptive road aerial view transformation method and road lane detection method | |
CN109190483B (en) | Lane line detection method based on vision | |
CN109948552B (en) | Method for detecting lane line in complex traffic environment | |
CN106887004A (en) | A kind of method for detecting lane lines based on Block- matching | |
CN106022243A (en) | Method for recognizing converse vehicle driving in vehicle lanes on the basis of image processing | |
CN104966049A (en) | Lorry detection method based on images |
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 |