CN103116748A - Method and system for identifying illegal driving behavior based on road signs - Google Patents
Method and system for identifying illegal driving behavior based on road signs Download PDFInfo
- Publication number
- CN103116748A CN103116748A CN2013100770488A CN201310077048A CN103116748A CN 103116748 A CN103116748 A CN 103116748A CN 2013100770488 A CN2013100770488 A CN 2013100770488A CN 201310077048 A CN201310077048 A CN 201310077048A CN 103116748 A CN103116748 A CN 103116748A
- Authority
- CN
- China
- Prior art keywords
- road
- image
- guiding
- sign
- identification
- 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
Images
Landscapes
- Traffic Control Systems (AREA)
- Image Processing (AREA)
Abstract
The invention provides a method and a system for identifying an illegal driving behavior based on road signs. The method comprises the following steps of: acquiring a vehicle driving road image, and preprocessing the road image to generate a binary image of the road image; determining the boundaries of a road according to white pixel points in the binary image to generate road boundary lines; detecting whether guide signs exist between road lines corresponding to the road boundary lines or not from the road image, wherein the guide signs comprise a bus-only sign and direction signs; and when the guide signs are the direction signs, judging whether a vehicle is illegally driven or not according to the direction signs and vehicle state information. According to the method in the embodiment of the invention, the illegal driving behavior is identified by generating the binary image of the road image and distinguishing the guide signs in the binary image, so that traffic accidents are prevented in advance, and driving safety is improved.
Description
Technical field
The present invention relates to the recognition technology field of driving behavior, particularly a kind of method and system based on road surface identification identification violation driving behavior.
Background technology
Increase rapidly along with vehicle fleet size has brought problems to road safety, and for example, traffic congestion, traffic hazard take place frequently etc.Caused by driver's driving lack of standardization greatly in traffic hazard.
The problem of road surface identification identification has the following aspects:
The road surface situation is complicated: the track situation mainly can be divided into two classes, and the one, the road surface situation when travelling in the urban district, this moment, front vehicles was more, and road surface, the place ahead may be blocked, the changeable complexity of travel situations.The 2nd, the road surface situation when highway driving, this moment, the road surface was spacious, and vehicle is fewer, is beneficial to the detection of lane line and the identification of sign.
Illumination effect: during due to Vehicle Driving Cycle, illumination changes for the pavement image impact that collects greatly.Main manifestations is to gather the uneven illumination that gathers image not fogging clear, daytime and all angles may occur, tunnel and road both sides trees night for the impact that gathers image etc.Therefore need design for the algorithm of illumination variation robust.
The algorithm travelling speed: speed during Vehicle Driving Cycle, and need to carry out lane detection and sign identification.Therefore, the requirement of real-time the when lane line that we study and the recognizer of sign must be able to satisfy high vehicle speeds is to provide the safe guarantee of travelling.
The method that generally adopts at present is to detect the method for roadside Sign Board and car status information.The method is determined current track to the restriction situation of Vehicle Driving Cycle parameter by the detection to the roadside Sign Board with identification, and car status information obtains by onboard sensor.Detect the roadside Sign Board and judge in conjunction with car status information whether current Vehicle Driving Cycle situation is unlawful practice.
The shortcoming of prior art:
(1) the roadside Sign Board is often blocked by vehicle and trees etc., and accuracy of detection is not high, does not detect road surface identification reliable and stable.
(2) detect with the decision logic of car status information judgement violation driving behavior perfect not based on the roadside Sign Board.
(3) detection algorithm is subjected to the impact of illumination condition larger.
Summary of the invention
Purpose of the present invention is intended to solve at least one of above-mentioned technological deficiency.
For achieving the above object, the embodiment of one aspect of the present invention proposes a kind of method based on road surface identification identification violation driving behavior, comprise the following steps: obtain the road image of Vehicle Driving Cycle, and described road image is carried out pre-service to generate the binary image of described road image; Determine that according to the white pixel point in described binary image the border of road is to generate lane boundary line; Detect from described road image whether the guiding sign is arranged between the Road corresponding with described lane boundary line, wherein, described guiding sign comprises public transport specific identity and direction signs; And when described guiding is designated direction signs, judge according to described direction signs and car status information whether described vehicle travels in violation of rules and regulations.
According to the method for the embodiment of the present invention, by generating the binary image of road image, and differentiate the identification of guiding sign and whether drive standard in this binary image, prevent in advance the generation of traffic hazard, and then improved the security of driving.
In one embodiment of the present of invention, when guiding is designated bus dedicated Lanes sign, send information to the driver.
In one embodiment of the present of invention, the characteristic similarity of described guiding sign and guiding sign template is during less than threshold value, and to identify template identical for described guiding sign and corresponding guiding.
In one embodiment of the present of invention, described pre-service comprises linear filtering and interframe stack.
For achieving the above object, embodiments of the invention propose a kind of system based on road surface identification identification violation driving behavior on the other hand, comprise: pretreatment module, be used for obtaining the road image of Vehicle Driving Cycle, and described road image is carried out pre-service to generate the binary image of described road image; Generation module is used for determining that according to the white pixel point of described binary image the border of road is to generate lane boundary line; Detection module is used for detecting from described road image whether the guiding sign is arranged between the Road corresponding with described lane boundary line, and wherein, described guiding sign comprises public transport specific identity and direction signs; And processing module, be used for when described guiding is designated direction signs, judge according to described direction signs and car status information whether described vehicle travels in violation of rules and regulations.
According to the system of the embodiment of the present invention, by generating the binary image of road image, and differentiate the identification of guiding sign and whether drive standard in this binary image, prevent in advance the generation of traffic hazard, and then improved the security of driving.
In one embodiment of the present of invention, described processing module also is used for sending information to the driver when guiding is designated bus dedicated Lanes sign.
In one embodiment of the present of invention, the characteristic similarity of described guiding sign and guiding sign template is during less than threshold value, and to identify template identical for described guiding sign and corresponding guiding.
In one embodiment of the present of invention, described pre-service comprises linear filtering and interframe stack.
The aspect that the present invention adds and advantage part in the following description provide, and part will become obviously from the following description, or recognize by practice of the present invention.
Description of drawings
Above-mentioned and/or the additional aspect of the present invention and advantage will become from the following description of the accompanying drawings of embodiments and obviously and easily understand, wherein:
Fig. 1 identifies the process flow diagram of the method for driving behavior in violation of rules and regulations based on road surface identification according to an embodiment of the invention;
Fig. 2 is real road image according to an embodiment of the invention;
Fig. 3 is carnival hat arithmograph picture according to an embodiment of the invention;
Fig. 4 is binary image according to an embodiment of the invention;
Fig. 5 is for carrying out according to an embodiment of the invention the binary image of interframe stack;
Fig. 6 is the search schematic diagram of lane boundary line according to an embodiment of the invention;
Fig. 7 is the template of guiding sign according to an embodiment of the invention figure;
Fig. 8 is upper left for the line of keeping straight on according to an embodiment of the invention, the schematic diagram of right-hand bend;
Fig. 9 takies schematic diagram for bus zone according to an embodiment of the invention;
The schematic diagram of Figure 10 for keeping straight on left and right according to an embodiment of the invention turning roadway; And
Figure 11 be the embodiment of the present invention based on the road surface identification identification structured flowchart of the system of driving behavior in violation of rules and regulations.
Embodiment
The below describes embodiments of the invention in detail, and the example of embodiment is shown in the drawings, and wherein same or similar label represents same or similar element or the element with identical or similar functions from start to finish.Be exemplary below by the embodiment that is described with reference to the drawings, only be used for explaining the present invention, and can not be interpreted as limitation of the present invention.
Fig. 1 be the embodiment of the present invention based on the road surface identification identification process flow diagram of the method for driving behavior in violation of rules and regulations.As shown in Figure 1, the method based on road surface identification identification violation driving behavior according to the embodiment of the present invention comprises the following steps:
Step S101 obtains the road image of Vehicle Driving Cycle, and road image is carried out pre-service to generate the binary image of road image.Wherein, pre-service comprises linear filtering and interframe stack.
Particularly, this project adopts so-called " carnival hat " computing to carry out morphologic filtering." carnival hat " operation definition be between this image and its opening operation by the pixel difference, and opening operation is defined as image is first corroded, then expands.The corrosion here and dilation operation are all the computings on real number field.
Corrosion: dst (x, y)=min (x', y'); Element (x', y') ≠ 0src (x+x', y+y'), wherein, src is source images, and dst is target image, and element is structural element, x', y' and x, y represent respectively the coordinate of structural element and image pixel.
Expand: dst (x, y)=max (x', y'); Element (x', y') ≠ 0src (x+x', y+y'), wherein, src is source images, and dst is target image, and element is structural element, x', y' and x, y represent respectively the coordinate of structural element and image pixel.
So " carnival hat " computing can be defined as: dst=src-open (src, element)=src-dilate (erode (src, element)), wherein, open (src, dlement) represents opening operation, dilate (erode (src, element)) expression is first corroded, then expands.The point brighter than surrounding pixel given prominence in " carnival hat " computing, and Fig. 2 and Fig. 3 are respectively real road image and carnival hat arithmograph picture.As can be seen from Figure 3 the result of carnival hat computing is not subjected to the impact of global illumination, in other words, result only with the magnitude relationship decision of pixel and neighboring pixel, this advantage comes from the difference link in computing, this advantage is consistent with the Gradient Features of picture, but Gradient Features is more responsive for the severe jamming that the center inverted image brings.
Directly carry out thresholding by " carnival hat " computing and can access result preferably, but result depends on the selection of threshold value, therefore utilize template filtering to realize.The feature of noticing Road is not only has obvious edge, Road also has obvious width simultaneously, therefore consider that template as shown in table 1 carries out convolution to image, again convolved image is carried out thresholding afterwards, take the threshold value (gray-scale value 220) of fixing, obtain binary image, as shown in Figure 4.But its Road is discontinuous lines, therefore carries out the interframe stack again and makes Road become continuous line.
Table 1
| -1 | -1 | -1 | -1 | 1 | 1 | 1 | 1 |
| -1 | -1 | -1 | -1 | 1 | 1 | 1 | 1 |
Road is always not continuous, and is for example shown in Figure 4, and the Road left side is continuous, and the dotted line of right side for being interrupted.And on highway, common dotted line type has two kinds of long dotted line (as Fig. 4) and short dash lines, can process respectively three kinds of dotted line types, but the interframe stack to inter-frame information can make the Road that originally is interrupted become continuous straight line, thereby obtains final binary image, as shown in Figure 5.
Step S102 determines that according to the white pixel point in binary image the border of road is to generate lane boundary line.
particularly, mathematical model for road has multiple choices, the match of straight line model is the simplest, also have and comprise more higher order polynomial model and the hyperbolic model of multiparameter, native system has been selected straight line model, its main cause also is except calculating simple consideration: only limit to a bit of distance before car for the sensing range of road surface symbol, the road surface that curvature is very large even (being similar to the annular bend on viaduct), also can roughly be approximated to straight line in a bit of distance of vehicle front, there is no help about road tendency information in the distance for nearby road surface symbol detection and curve model brings.
In one embodiment of the invention, obtaining on the basis of binary image, need to obtain the accurate parameter of Road, namely determining the boundary line of road.By adopting the searching method line by line based on empirical rule, this method has calculates advantage fast, although this method easily is subject to the interference of noise and road surface symbol, if suitably utilize inter-frame information, the impact of above-mentioned interference can significantly be subdued.
In one embodiment of the invention, search element by row in binary image, therefrom the alignment left and right sides begins search, first white point that every row memory search obtains, this point is used as possible Road point, often add a new point in road point set, hunting zone next time is just near a upper point.Until think that search finishes when continuously appearance is repeatedly searched for failure or reached the manual border of setting, existing result is carried out fitting a straight line, if related coefficient is accepted Search Results greater than certain threshold value (for example 0.99), otherwise just from the set first point near choose a new starting point, proceed search procedure.
Successful in typical case search as shown in Figure 6, the track of the point of circle representative search, during search left-side course route, at first begin to search for left from center line, first white pixel that will run into adds the Road set, each search afterwards is all at before lastrow, and volumes of searches has been reduced in the hunting zone on horizontal direction so to a great extent near before.May run in search procedure is not the pixel on Road but its gray-scale value is 255, and due to a plurality of white pixel that can't find take such pixel as the follow-up search of beginning on same straight line, so wrong initial point can be abandoned.Can obtain in this way the white pixel point set of a line-like type.
After each point is positioned at position on the road inside edge from the set of resulting white pixel point, adopt least square fitting to generate the boundary line of road.Least square fitting is a kind of linear optimal technology, known one group of point (x
i, y
i) be positioned on straight line x=ky+b, known to y
iMeasurement be accurately, whole errors comes from x coordinate (i.e. row position be that error is arranged), least square method has provided the estimation of minimum error square k, b,
Wherein, S
x=∑ x
iBe white pixel point x coordinate sum, x
iBe white pixel point x coordinate, S
y=∑ y
iBe white pixel point y coordinate sum, y
iBe white pixel point y coordinate, S
xy=∑ x
iy
i,
Be the corresponding computing of white pixel point coordinate,
Be the estimated value of parameter k,
Estimated value for parameter b.
In one embodiment of the invention, really comprised the resistivity for noise in the Road searching algorithm, but only limit to the not noise on same straight line, and the symbol between road has comprised the white pixel point of One's name is legion, be easy to find out several pixels on same straight line, thereby cause the searching algorithm cisco unity malfunction.Suppose that resulting previous frame image is correct, this information just provides important foundation for eliminating the middle symbol of road, because meet at end point as long as two hypotenuses are extended, article two, the bottom line of oblique line and level has formed leg-of-mutton roadway area, roadway area is outward Road, and be the symbol that possible cause between the road of interference in roadway area, therefore removing the method for disturbing is exactly that the pixel in roadway area all is set to black, can draw accurately thus the boundary line of road.
Step S103 detects from road image whether the guiding sign is arranged between the Road corresponding with lane boundary line, and wherein, the guiding sign comprises public transport specific identity and direction signs.
Particularly, the guiding sign has guiding sign template, can identify template by guiding because it is simple in structure and compare with processing the guiding sign that obtains, thereby determine it is which kind of guiding sign.The guiding sign comprises public transport specific identity and direction signs, as shown in Figure 7.
Step S104 when guiding is designated direction signs, judges according to direction signs and car status information whether vehicle travels in violation of rules and regulations.
In one embodiment of the invention, after determining Road, in order more stably to carry out pattern classification, need to be plan view with the view transformation of looking squarely of video camera, such conversion is extremely important for coupling, because the horizontal shift meeting of video camera causes the shape of symbol to change in looking squarely view, and in plan view, as long as correctly determined Road, the shape of symbol is exactly roughly to fix.After conversion, overlook that in piece, there is a strong possibility and exist sign, determine the kind of sign below by the method for template matches.Namely define related coefficient C (I, I between picture and template
i), I wherein
iBe the template of guiding sign i, the classification of image is determined in the following manner: x=argmax
iC (I, I
i), wherein, x represents the final guiding sign classification of determining.
Namely seek a width the highest with the picture degree of correlation in all templates.Simultaneously in order to process the sign that is not included in template base and to strengthen resistivity to noise, still hour think recognition failures when maximum correlation coefficient, abandon the identification to this symbol, so classification foundation can intactly be expressed as:
Wherein, TH represents the manually correlation coefficient threshold of setting.Classification is output as-1 presentation class failure.
The selection of related coefficient has important impact for the effect of classification, has proposed the computing method of multiple related coefficient with regard to this Study on Problems person, need to preferentially choose in different occasions.
In one embodiment of the invention, the similarity measure of the most frequently used real vector comprises that inner product, cosine related coefficient, Pearson correlation coefficient and Hamming estimate.
(1) inner product
Be defined as C (x, y)=x
TY=∑ x
iy
i, wherein, x
iWith y
iThe coordinate that represents respectively vector x and y.In most of the cases, when vector x and y were had identical length a by normalization, the upper and lower bound of C was respectively+a
2With-a
2
(2) cosine related coefficient
The cosine related coefficient is relevant to inner product larger, is defined as:
Wherein, x and y represent respectively vector.This related coefficient has the geometric meaning of vector angle cosine, is also the related coefficient of normalization between-1 to+1, and is invariable rotary.
(3) Pearson correlation coefficient.Its expression formula is:
Wherein, x
d=(x
1-m
x, x
2-m
x..., x
n-m
x) the difference coordinate of expression vector x and average, y
d=(y
1-m
y, y
2-m
y..., y
n-m
y) the difference coordinate of expression vectorial y and average,
The average of expression vector x,
The average that represents vectorial y, x
iWith y
iBe respectively the coordinate of vector x and y.
(4) Hamming is estimated, and establishes
Wherein, x
iWith y
iBe respectively the coordinate of vector x and y.
Obtain the characteristic similarity of guiding sign and guiding sign template according to C (x, y), when similarity is related coefficient during less than threshold value, guiding sign is identical with corresponding guiding sign template, thereby determines the classification that guiding identifies.
In one embodiment of the invention, because the cosine related coefficient can obtain classifying quality preferably, therefore adopt the cosine related coefficient as the method for measuring similarity between guiding sign and guiding sign template.
Violation that the present invention will identify is driven and is comprised that the craspedodrome line is left and right and turn, and turns and does not open steering indicating light, takies bus zone and left and right turning roadway and keeps straight on.
Fig. 8 is upper left for the line of keeping straight on according to an embodiment of the invention, the schematic diagram of right-hand bend.As shown in Figure 8, should not there be the driving behavior of left and right turning in vehicle on Through Lane, should change on the left/right rotation sweep if the driver thinks left and right turning.Turn on the craspedodrome line, even opened Lamp for indicating change of drive direction, it is the thread-changing behavior that the rear driver also can take for front vehicles, causes the rear driver to fail efficient solution and reads the place ahead driver's operation intention, there is no the preparation on certain Psychology and behaviors, cause potential safety hazard.
Road surface identification detects when keeping straight on sign, detects the bearing circle angular signal, if steering wheel angle changes in certain threshold value, illustrates that under current state, vehicle does not have turning behavior; If steering wheel angle has surpassed certain threshold value, illustrate that vehicle has turning behavior.If vehicle detected, turning behavior is arranged, program is exported early warning information in violation of rules and regulations.
In one embodiment of the invention, when turn inside diameter travels, generally be accompanied by the driving behavior that reduces the speed of a motor vehicle, if do not open steering indicating light this moment, follow vehicle and fail effectively to pick out the driving intention of front truck, straight line fast running still, just can cause serious traffic hazard this moment.Turn and do not open the steering indicating light behavior, endanger huge.
Road surface identification detects or turns right when identifying, and illustrates that Vehicle Driving Cycle is on added turning lane.Detect the bearing circle angular signal this moment, if steering wheel angle in certain corner threshold value, human pilot is opened steering indicating light, illustrates that the driver has produced the consciousness that turns to, and has opened Lamp for indicating change of drive direction.If surpass certain corner threshold value, human pilot is not opened Lamp for indicating change of drive direction yet, and program is exported early warning information in violation of rules and regulations, notifies the driver should open steering indicating light.Road is established dedicated Lanes, in dedicated Lanes, only permits the vehicle pass-through of regulation, and other vehicles must not enter in dedicated Lanes and travel.
Fig. 9 takies schematic diagram for bus zone according to an embodiment of the invention.As shown in Figure 9, in the serious rare reality of path resource, it is in order to solve the trip problem of most people that public transportation lane is set.The setting of public transportation lane has promoted the enthusiasm that citizen take pubic transport, and dedicated bus lanes is used for reminding citizen should change trip mode, reduces drive private car, selects public transport, thereby alleviates traffic congestion.In communications and transportation peak time, take bus zone, certainly will cause traffic congestion, affect the traffic transport power.When road surface identification detects " public affairs ", " friendship ", " car ", " road " sign, illustrate that under present case, Vehicle Driving Cycle is on dedicated bus lanes, taken bus zone, program is sent early warning information in violation of rules and regulations, notifies the driver change lane, changes to other tracks from current track.
The schematic diagram of Figure 10 for keeping straight on left and right according to an embodiment of the invention turning roadway.As shown in figure 10, setting up left and right changing trains or buses is to travel for the ease of turn inside diameter.When turn inside diameter travels, generally be accompanied by the deceleration behavior.If on left and right changing trains or buses, front vehicles has turning behavior, it reduces speed now, and unlatching Lamp for indicating change of drive direction, if still run at high speed on left and right change trains or buses from car this moment, fail useful consciousness to the turning behavior of front truck, and very possible and front truck bumps, and causes serious traffic accidents.
Road surface identification detects or turns right when identifying, and illustrates that Vehicle Driving Cycle is on added turning lane.This moment detect vehicle speed signal, with the speed of a motor vehicle to time integral, if in certain distance range, human pilot is opened steering indicating light, the driving intention that the driver is described is Turning travel, and has recognized from car and travelled on added turning lane, has opened Lamp for indicating change of drive direction and has prepared to change.If surpass certain distance threshold, human pilot is not opened Lamp for indicating change of drive direction yet, still straight-line travelling, violation program output early warning information.
According to the method for the embodiment of the present invention, by generating the binary image of road image, and differentiate the identification of guiding sign and whether drive standard in this binary image, prevent in advance the generation of traffic hazard, and then improved the security of driving.
Figure 11 be the embodiment of the present invention based on the road surface identification identification structured flowchart of the system of driving behavior in violation of rules and regulations, as shown in figure 11, the system state evaluating system based on road surface identification identification violation driving behavior according to the embodiment of the present invention comprises pretreatment module 100, generation module 200, detection module 300 and processing module 400.
Pretreatment module 100 is used for obtaining the road image of Vehicle Driving Cycle, and road image is carried out pre-service to generate the binary image of road image.
Particularly, this project adopts so-called " carnival hat " computing to carry out morphologic filtering." carnival hat " operation definition be between this image and its opening operation by the pixel difference, and opening operation is defined as image is first corroded, then expands.The corrosion here and dilation operation are all the computings on real number field.
Corrosion: dst (x, y)=min (x', y'); Element (x', y') ≠ 0src (x+x', y+y'), wherein, src is source images, and dst is target image, and element is structural element, x', y' and x, y represent respectively the coordinate of structural element and image pixel.
Expand: dst (x, y)=max (x', y'); Element (x', y') ≠ 0src (x+x', y+y'), wherein, src is source images, and dst is target image, and element is structural element, x', y' and x, y represent respectively the coordinate of structural element and image pixel.
So " carnival hat " computing can be defined as: dst=src-open (src, element)=src-dilate (erode (src, element)), wherein, open (src, dlement) represents opening operation, dilate (erode (src, element)) expression is first corroded, then expands.The point brighter than surrounding pixel given prominence in " carnival hat " computing, and Fig. 2 and Fig. 3 are respectively real road image and carnival hat arithmograph picture.As can be seen from Figure 3 the result of carnival hat computing is not subjected to the impact of global illumination, in other words, result only with the magnitude relationship decision of pixel and neighboring pixel, this advantage comes from the difference link in computing, this advantage is consistent with the Gradient Features of picture, but Gradient Features is more responsive for the severe jamming that the center inverted image brings.
Directly carry out thresholding by " carnival hat " computing and can access result preferably, but result depends on the selection of threshold value, therefore utilize template filtering to realize.The feature of noticing Road is not only has obvious edge, Road also has obvious width simultaneously, therefore consider that template as shown in table 1 carries out convolution to image, again convolved image is carried out thresholding afterwards, take the threshold value (gray-scale value 220) of fixing, obtain binary image, as shown in Figure 4.But its Road is discontinuous lines, therefore carries out the interframe stack again and makes Road become continuous line.
Table 1
| -1 | -1 | -1 | -1 | 1 | 1 | 1 | 1 |
| -1 | -1 | -1 | -1 | 1 | 1 | 1 | 1 |
Road is always not continuous, and is for example shown in Figure 4, and the Road left side is continuous, and the dotted line of right side for being interrupted.And on highway, common dotted line type has two kinds of long dotted line (as Fig. 4) and short dash lines, can process respectively three kinds of dotted line types, but the interframe stack to inter-frame information can make the Road that originally is interrupted become continuous straight line, thereby obtains final binary image, as shown in Figure 5.
Generation module 200 is used for determining that according to the white pixel point of binary image the border of road is to generate lane boundary line.
particularly, mathematical model for road has multiple choices, the match of straight line model is the simplest, also have and comprise more higher order polynomial model and the hyperbolic model of multiparameter, native system has been selected straight line model, its main cause also is except calculating simple consideration: only limit to a bit of distance before car for the sensing range of road surface symbol, the road surface that curvature is very large even (being similar to the annular bend on viaduct), also can roughly be approximated to straight line in a bit of distance of vehicle front, there is no help about road tendency information in the distance for nearby road surface symbol detection and curve model brings.
In one embodiment of the invention, obtaining on the basis of binary image, need to obtain the accurate parameter of Road, namely determining the boundary line of road.By adopting the searching method line by line based on empirical rule, this method has calculates advantage fast, although this method easily is subject to the interference of noise and road surface symbol, if suitably utilize inter-frame information, the impact of above-mentioned interference can significantly be subdued.
In one embodiment of the invention, search element by row in binary image, therefrom the alignment left and right sides begins search, first white point that every row memory search obtains, this point is used as possible Road point, often add a new point in road point set, hunting zone next time is just near a upper point.Until think that search finishes when continuously appearance is repeatedly searched for failure or reached the manual border of setting, existing result is carried out fitting a straight line, if related coefficient is accepted Search Results greater than certain threshold value (for example 0.99), otherwise just from the set first point near choose a new starting point, proceed search procedure.
Successful in typical case search as shown in Figure 6, the track of the point of circle representative search, during search left-side course route, at first begin to search for left from center line, first white pixel that will run into adds the Road set, each search afterwards is all at before lastrow, and volumes of searches has been reduced in the hunting zone on horizontal direction so to a great extent near before.May run in search procedure is not the pixel on Road but its gray-scale value is 255, and due to a plurality of white pixel that can't find take such pixel as the follow-up search of beginning on same straight line, so wrong initial point can be abandoned.Can obtain in this way the white pixel point set of a line-like type.
After each point is positioned at position on the road inside edge from the set of resulting white pixel point, adopt least square fitting to generate the boundary line of road.Least square fitting is a kind of linear optimal technology, known one group of point (x
i, y
i) be positioned on straight line x=ky+b, known to y
iMeasurement be accurately, whole errors comes from x coordinate (i.e. row position be that error is arranged), least square method has provided the estimation of minimum error square k, b,
Wherein, S
x=∑ x
iBe white pixel point x coordinate sum, x
iBe white pixel point x coordinate, S
y=∑ y
iBe white pixel point y coordinate sum, y
iBe white pixel point y coordinate, S
xy=∑ x
iy
i,
Be the corresponding computing of white pixel point coordinate,
Be the estimated value of parameter k,
Estimated value for parameter b.
In one embodiment of the invention, really comprised the resistivity for noise in the Road searching algorithm, but only limit to the not noise on same straight line, and the symbol between road has comprised the white pixel point of One's name is legion, be easy to find out several pixels on same straight line, thereby cause the searching algorithm cisco unity malfunction.Suppose that resulting previous frame image is correct, this information just provides important foundation for eliminating the middle symbol of road, because meet at end point as long as two hypotenuses are extended, article two, the bottom line of oblique line and level has formed leg-of-mutton roadway area, roadway area is outward Road, and be the symbol that possible cause between the road of interference in roadway area, therefore removing the method for disturbing is exactly that the pixel in roadway area all is set to black, can draw accurately thus the boundary line of road.
Detection module 300 is used for detecting from road image whether the guiding sign is arranged between the Road corresponding with lane boundary line, and wherein, the guiding sign comprises public transport specific identity and direction signs.
Particularly, the guiding sign has guiding sign template, can identify template by guiding because it is simple in structure and compare with processing the guiding sign that obtains, thereby determine it is which kind of guiding sign.The guiding sign comprises public transport specific identity and direction signs, as shown in Figure 7.
Processing module 400 is used for when guiding is designated direction signs, judges according to direction signs and car status information information whether vehicle travels in violation of rules and regulations.
In one embodiment of the invention, after determining Road, in order more stably to carry out pattern classification, need to be plan view with the view transformation of looking squarely of video camera, such conversion is extremely important for coupling, because the horizontal shift meeting of video camera causes the shape of symbol to change in looking squarely view, and in plan view, as long as correctly determined Road, the shape of symbol is exactly roughly to fix.After conversion, overlook that in piece, there is a strong possibility and exist sign, determine the kind of sign below by the method for template matches.Namely define related coefficient C (I, I between picture and template
i), I wherein
iBe the template of guiding sign i, the classification of image is determined in the following manner: x=argmax
iC (I, I
i), wherein, x represents the final guiding sign classification of determining.
Namely seek a width the highest with the picture degree of correlation in all templates.Simultaneously in order to process the sign that is not included in template base and to strengthen resistivity to noise, still hour think recognition failures when maximum correlation coefficient, abandon the identification to this symbol, so classification foundation can intactly be expressed as:
Wherein, TH represents the manually correlation coefficient threshold of setting.Classification is output as-1 presentation class failure.
The selection of related coefficient has important impact for the effect of classification, has proposed the computing method of multiple related coefficient with regard to this Study on Problems person, need to preferentially choose in different occasions.
In one embodiment of the invention, the similarity measure of the most frequently used real vector comprises that inner product, cosine related coefficient, Pearson correlation coefficient and Hamming estimate.
(1) inner product
Be defined as C (x, y)=x
TY=∑ x
iy
i, wherein, x
iWith y
iThe coordinate that represents respectively vector x and y.In most of the cases, when vector x and y were had identical length a by normalization, the upper and lower bound of C was respectively+a
2With-a
2
(2) cosine related coefficient
The cosine related coefficient is relevant to inner product larger, is defined as:
Wherein, x and y represent respectively vector.This related coefficient has the geometric meaning of vector angle cosine, is also the related coefficient of normalization between-1 to+1, and is invariable rotary.
(3) Pearson correlation coefficient.Its expression formula is:
Wherein, x
d=(x
1-m
x, x
2-m
x..., x
n-m
x) the difference coordinate of expression vector x and average, y
d=(y
1-m
y, y
2-m
y..., y
n-m
y) the difference coordinate of expression vectorial y and average,
The average of expression vector x,
The average that represents vectorial y, x
iWith y
iBe respectively the coordinate of vector x and y.
(4) Hamming is estimated, and establishes
Wherein, x
iWith y
iBe respectively the coordinate of vector x and y.
Obtain the characteristic similarity of guiding sign and guiding sign template according to C (x, y), when similarity is related coefficient during less than threshold value, guiding sign is identical with corresponding guiding sign template, thereby determines the classification that guiding identifies.In one embodiment of the invention, because the cosine related coefficient can obtain classifying quality preferably, therefore adopt the cosine related coefficient as the method for measuring similarity between guiding sign and guiding sign template.
Violation that the present invention will identify is driven and is comprised that the craspedodrome line is left and right and turn, and turns and does not open steering indicating light, takies bus zone and left and right turning roadway and keeps straight on.
Fig. 8 is upper left for the line of keeping straight on according to an embodiment of the invention, the schematic diagram of right-hand bend.As shown in Figure 8, should not there be the driving behavior of left and right turning in vehicle on Through Lane, should change on the left/right rotation sweep if the driver thinks left and right turning.Turn on the craspedodrome line, even opened Lamp for indicating change of drive direction, it is the thread-changing behavior that the rear driver also can take for front vehicles, causes the rear driver to fail efficient solution and reads the place ahead driver's operation intention, there is no the preparation on certain Psychology and behaviors, cause potential safety hazard.
Road surface identification detects when keeping straight on sign, detects the bearing circle angular signal, if steering wheel angle changes in certain threshold value, illustrates that under current state, vehicle does not have turning behavior; If steering wheel angle has surpassed certain threshold value, illustrate that vehicle has turning behavior.If vehicle detected, turning behavior is arranged, program is exported early warning information in violation of rules and regulations.
In one embodiment of the invention, when turn inside diameter travels, generally be accompanied by the driving behavior that reduces the speed of a motor vehicle, if do not open steering indicating light this moment, follow vehicle and fail effectively to pick out the driving intention of front truck, straight line fast running still, just can cause serious traffic hazard this moment.Turn and do not open the steering indicating light behavior, endanger huge.
Road surface identification detects or turns right when identifying, and illustrates that Vehicle Driving Cycle is on added turning lane.Detect the bearing circle angular signal this moment, if steering wheel angle in certain corner threshold value, human pilot is opened steering indicating light, illustrates that the driver has produced the consciousness that turns to, and has opened Lamp for indicating change of drive direction.If surpass certain corner threshold value, human pilot is not opened Lamp for indicating change of drive direction yet, and program is exported early warning information in violation of rules and regulations, notifies the driver should open steering indicating light.Road is established dedicated Lanes, in dedicated Lanes, only permits the vehicle pass-through of regulation, and other vehicles must not enter in dedicated Lanes and travel.
Fig. 9 takies schematic diagram for bus zone according to an embodiment of the invention.As shown in Figure 9, in the serious rare reality of path resource, it is in order to solve the trip problem of most people that public transportation lane is set.The setting of public transportation lane has promoted the enthusiasm that citizen take pubic transport, and dedicated bus lanes is used for reminding citizen should change trip mode, reduces drive private car, selects public transport, thereby alleviates traffic congestion.In communications and transportation peak time, take bus zone, certainly will cause traffic congestion, affect the traffic transport power.When road surface identification detects " public affairs ", " friendship ", " car ", " road " sign, illustrate that under present case, Vehicle Driving Cycle is on dedicated bus lanes, taken bus zone, program is sent early warning information in violation of rules and regulations, notifies the driver change lane, changes to other tracks from current track.
The schematic diagram of Figure 10 for keeping straight on left and right according to an embodiment of the invention turning roadway.As shown in figure 10, setting up left and right changing trains or buses is to travel for the ease of turn inside diameter.When turn inside diameter travels, generally be accompanied by the deceleration behavior.If on left and right changing trains or buses, front vehicles has turning behavior, it reduces speed now, and unlatching Lamp for indicating change of drive direction, if still run at high speed on left and right change trains or buses from car this moment, fail useful consciousness to the turning behavior of front truck, and very possible and front truck bumps, and causes serious traffic accidents.
Road surface identification detects or turns right when identifying, and illustrates that Vehicle Driving Cycle is on added turning lane.This moment detect vehicle speed signal, with the speed of a motor vehicle to time integral, if in certain distance range, human pilot is opened steering indicating light, the driving intention that the driver is described is Turning travel, and has recognized from car and travelled on added turning lane, has opened Lamp for indicating change of drive direction and has prepared to change.If surpass certain distance threshold, human pilot is not opened Lamp for indicating change of drive direction yet, still straight-line travelling, violation program output early warning information.
According to the system of the embodiment of the present invention, by generating the binary image of road image, and differentiate the identification of guiding sign and whether drive standard in this binary image, prevent in advance the generation of traffic hazard, and then improved the security of driving.
Although the above has illustrated and has described embodiments of the invention, be understandable that, above-described embodiment is exemplary, can not be interpreted as limitation of the present invention, those of ordinary skill in the art is not in the situation that break away from principle of the present invention and aim can change above-described embodiment within the scope of the invention, modification, replacement and modification.
Claims (8)
1. the method based on road surface identification identification violation driving behavior, is characterized in that, comprises the following steps:
Obtain the road image of Vehicle Driving Cycle, and described road image is carried out pre-service to generate the binary image of described road image;
Determine that according to the white pixel point in described binary image the border of road is to generate lane boundary line;
Detect from described road image whether the guiding sign is arranged between the Road corresponding with described lane boundary line, wherein, described guiding sign comprises public transport specific identity and direction signs; And
When described guiding is designated direction signs, judge according to described direction signs and car status information whether described vehicle travels in violation of rules and regulations.
2. the method based on road surface identification identification violation driving behavior as claimed in claim 1, is characterized in that, when guiding is designated bus dedicated Lanes sign, sends information to the driver.
3. as claimed in claim 1 or 2ly it is characterized in that based on the road surface identification identification method of driving behavior in violation of rules and regulations, the characteristic similarity of described guiding sign and guiding sign template is during less than threshold value, and to identify template identical for described guiding sign and corresponding guiding.
4. the method based on road surface identification identification violation driving behavior as claimed in claim 1, is characterized in that, described pre-service comprises linear filtering and interframe stack.
5. the system based on road surface identification identification violation driving behavior, is characterized in that, comprising:
Pretreatment module is used for obtaining the road image of Vehicle Driving Cycle, and described road image is carried out pre-service to generate the binary image of described road image;
Generation module is used for determining that according to the white pixel point of described binary image the border of road is to generate lane boundary line;
Detection module is used for detecting from described road image whether the guiding sign is arranged between the Road corresponding with described lane boundary line, and wherein, described guiding sign comprises public transport specific identity and direction signs; And
Processing module is used for when described guiding is designated direction signs, judges according to described direction signs and car status information whether described vehicle travels in violation of rules and regulations.
6. the system based on road surface identification identification violation driving behavior as claimed in claim 5, is characterized in that, described processing module also is used for sending information to the driver when guiding is designated bus dedicated Lanes sign.
7. as describedly in claim 5 or 6 it is characterized in that based on the road surface identification identification system of driving behavior in violation of rules and regulations, the characteristic similarity of described guiding sign and guiding sign template is during less than threshold value, and to identify template identical for described guiding sign and corresponding guiding.
8. the system based on road surface identification identification violation driving behavior as claimed in claim 5, is characterized in that, described pre-service comprises linear filtering and interframe stack.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310077048.8A CN103116748B (en) | 2013-03-11 | 2013-03-11 | Based on the method and system of road surface identification identification violation driving behavior |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310077048.8A CN103116748B (en) | 2013-03-11 | 2013-03-11 | Based on the method and system of road surface identification identification violation driving behavior |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103116748A true CN103116748A (en) | 2013-05-22 |
| CN103116748B CN103116748B (en) | 2016-03-23 |
Family
ID=48415120
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310077048.8A Active CN103116748B (en) | 2013-03-11 | 2013-03-11 | Based on the method and system of road surface identification identification violation driving behavior |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103116748B (en) |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103886745A (en) * | 2013-11-25 | 2014-06-25 | 天津思博科科技发展有限公司 | Automobile-dedicated lane early warning system |
| CN105679092A (en) * | 2016-01-29 | 2016-06-15 | 深圳市美好幸福生活安全系统有限公司 | Driving behavior analysis system and method |
| CN105761525A (en) * | 2016-05-11 | 2016-07-13 | 南京信息职业技术学院 | System device for warning car to enter bus lane |
| CN106004883A (en) * | 2016-05-24 | 2016-10-12 | 北京小米移动软件有限公司 | Vehicle violation reminding method and device |
| CN109740424A (en) * | 2018-11-23 | 2019-05-10 | 深圳市华尊科技股份有限公司 | Traffic violations recognition methods and Related product |
| CN109987090A (en) * | 2018-01-03 | 2019-07-09 | 奥迪股份公司 | Driving assistance system and method |
| CN110455302A (en) * | 2018-05-08 | 2019-11-15 | 奥迪股份公司 | Navigation system control method, device, computer equipment and storage medium |
| CN111409649A (en) * | 2019-01-04 | 2020-07-14 | 奥迪股份公司 | Early warning method and device for vehicle lane change, computer equipment and storage medium |
| CN111815959A (en) * | 2020-06-19 | 2020-10-23 | 浙江大华技术股份有限公司 | Vehicle violation detection method and device and computer readable storage medium |
| CN111833629A (en) * | 2019-04-23 | 2020-10-27 | 上海博泰悦臻网络技术服务有限公司 | Special lane prompting method and system and vehicle |
| CN112699825A (en) * | 2021-01-05 | 2021-04-23 | 上海博泰悦臻网络技术服务有限公司 | Lane line identification method and device |
| WO2021077760A1 (en) * | 2019-10-23 | 2021-04-29 | 江苏智通交通科技有限公司 | Abnormal driving early warning method on basis of reasonable driving range of vehicle at intersection |
| CN112949470A (en) * | 2021-02-26 | 2021-06-11 | 上海商汤智能科技有限公司 | Method, device and equipment for identifying lane-changing steering lamp of vehicle and storage medium |
| CN114485671A (en) * | 2022-01-24 | 2022-05-13 | 轮趣科技(东莞)有限公司 | Automatic turning method and device for mobile equipment |
| CN114619950A (en) * | 2020-12-14 | 2022-06-14 | 宝能汽车集团有限公司 | Control method of vehicle steering lamp, storage medium, electronic device and vehicle |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201149729Y (en) * | 2007-12-28 | 2008-11-12 | 四川川大智胜软件股份有限公司 | Video traffic sign alarm |
| CN101470807A (en) * | 2007-12-26 | 2009-07-01 | 河海大学常州校区 | Accurate detection method for highroad lane marker line |
-
2013
- 2013-03-11 CN CN201310077048.8A patent/CN103116748B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101470807A (en) * | 2007-12-26 | 2009-07-01 | 河海大学常州校区 | Accurate detection method for highroad lane marker line |
| CN201149729Y (en) * | 2007-12-28 | 2008-11-12 | 四川川大智胜软件股份有限公司 | Video traffic sign alarm |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103886745A (en) * | 2013-11-25 | 2014-06-25 | 天津思博科科技发展有限公司 | Automobile-dedicated lane early warning system |
| CN105679092A (en) * | 2016-01-29 | 2016-06-15 | 深圳市美好幸福生活安全系统有限公司 | Driving behavior analysis system and method |
| CN105679092B (en) * | 2016-01-29 | 2018-05-04 | 深圳市美好幸福生活安全系统有限公司 | A kind of driving behavior analysis system and method |
| CN105761525A (en) * | 2016-05-11 | 2016-07-13 | 南京信息职业技术学院 | System device for warning car to enter bus lane |
| CN106004883A (en) * | 2016-05-24 | 2016-10-12 | 北京小米移动软件有限公司 | Vehicle violation reminding method and device |
| CN106004883B (en) * | 2016-05-24 | 2018-12-11 | 北京小米移动软件有限公司 | The method and device that rule-breaking vehicle is reminded |
| CN109987090A (en) * | 2018-01-03 | 2019-07-09 | 奥迪股份公司 | Driving assistance system and method |
| CN110455302B (en) * | 2018-05-08 | 2023-10-20 | 奥迪股份公司 | Navigation system control method, device, computer equipment and storage medium |
| CN110455302A (en) * | 2018-05-08 | 2019-11-15 | 奥迪股份公司 | Navigation system control method, device, computer equipment and storage medium |
| CN109740424A (en) * | 2018-11-23 | 2019-05-10 | 深圳市华尊科技股份有限公司 | Traffic violations recognition methods and Related product |
| CN111409649A (en) * | 2019-01-04 | 2020-07-14 | 奥迪股份公司 | Early warning method and device for vehicle lane change, computer equipment and storage medium |
| CN111409649B (en) * | 2019-01-04 | 2023-10-20 | 奥迪股份公司 | Early warning method and device for lane change of vehicle, computer equipment and storage medium |
| CN111833629A (en) * | 2019-04-23 | 2020-10-27 | 上海博泰悦臻网络技术服务有限公司 | Special lane prompting method and system and vehicle |
| WO2021077760A1 (en) * | 2019-10-23 | 2021-04-29 | 江苏智通交通科技有限公司 | Abnormal driving early warning method on basis of reasonable driving range of vehicle at intersection |
| CN111815959A (en) * | 2020-06-19 | 2020-10-23 | 浙江大华技术股份有限公司 | Vehicle violation detection method and device and computer readable storage medium |
| CN114619950A (en) * | 2020-12-14 | 2022-06-14 | 宝能汽车集团有限公司 | Control method of vehicle steering lamp, storage medium, electronic device and vehicle |
| CN112699825A (en) * | 2021-01-05 | 2021-04-23 | 上海博泰悦臻网络技术服务有限公司 | Lane line identification method and device |
| CN112949470A (en) * | 2021-02-26 | 2021-06-11 | 上海商汤智能科技有限公司 | Method, device and equipment for identifying lane-changing steering lamp of vehicle and storage medium |
| CN114485671A (en) * | 2022-01-24 | 2022-05-13 | 轮趣科技(东莞)有限公司 | Automatic turning method and device for mobile equipment |
| CN114485671B (en) * | 2022-01-24 | 2024-06-07 | 轮趣科技(东莞)有限公司 | Automatic turning method and device for mobile equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103116748B (en) | 2016-03-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103116748A (en) | Method and system for identifying illegal driving behavior based on road signs | |
| US10885777B2 (en) | Multiple exposure event determination | |
| CN106203398B (en) | A kind of method, apparatus and equipment detecting lane boundary | |
| US8385601B2 (en) | In-vehicle white line recognition apparatus | |
| US9633267B2 (en) | Robust windshield detection via landmark localization | |
| KR100969995B1 (en) | System of traffic conflict decision for signalized intersections using image processing technique | |
| Cualain et al. | Automotive standards-grade lane departure warning system | |
| US20100259617A1 (en) | Boundary line recognition apparatus | |
| US9470537B2 (en) | Accurate position determination near exit lanes | |
| TWI493513B (en) | Lane departure warning system and lane identification apparatus and related method thereof | |
| CN102682304A (en) | Multi-feature integrated passer-by detection method and device | |
| KR20220040473A (en) | detection of emergency vehicles | |
| Wei et al. | Research on lane detection and tracking algorithm based on improved hough transform | |
| US9835468B2 (en) | Entry detection apparatus and entry detection method | |
| Liu et al. | ROI perspective transform based road marking detection and recognition | |
| Kamble et al. | Lane departure warning system for advanced drivers assistance | |
| WO2018030103A1 (en) | Displayed content recognition device and vehicle control device | |
| JP5204144B2 (en) | In-vehicle white line recognition device | |
| Coronado et al. | Detection and classification of road signs for automatic inventory systems using computer vision | |
| CN112270258B (en) | Method and device for acquiring violation information of non-motor vehicle | |
| JP2014026519A (en) | On-vehicle lane marker recognition device | |
| Boumediene et al. | Vehicle detection algorithm based on horizontal/vertical edges | |
| KR101458387B1 (en) | Method and apparatus for preventing drunken driving | |
| CN103935293A (en) | LDW system, lane recognition device and related method | |
| JP4972596B2 (en) | Traffic flow measuring device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |