CN103714538A - Road edge detection method, device and vehicle - Google Patents
Road edge detection method, device and vehicle Download PDFInfo
- Publication number
- CN103714538A CN103714538A CN201310711070.3A CN201310711070A CN103714538A CN 103714538 A CN103714538 A CN 103714538A CN 201310711070 A CN201310711070 A CN 201310711070A CN 103714538 A CN103714538 A CN 103714538A
- Authority
- CN
- China
- Prior art keywords
- straight line
- line segment
- road
- curb
- pixel
- 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
Abstract
The invention discloses a road edge detection method comprising obtaining an image frame including road edge information of a current road where a vehicle drives, performing edge detection on the image frame to obtain a plurality of edge points, utilizing the plurality of edge points to extract a plurality of linear line sections, and extracting the road edge line sections from the plurality of linear line sections based on the road edge structure features of the current road. The invention also discloses a road edge detection device and vehicle. Through the above method, the road edge line sections of the current road where the vehicle drives can be automatically detected, operation complexity for a machine operator is reduced and the detection precision is high.
Description
Technical field
The present invention relates to field of information processing, particularly relate to a kind of road-edge detection method, device and vehicle.
Background technology
The vehicles such as someone steering vehicle or automatic driving vehicle are expert in the process of sailing, and often need to detect the curb line segment of the current road that vehicle travels, so that the actual range between subsequent calculations vehicle and curb line segment guarantees the safety traffic of vehicle.In prior art, conventionally adopt following two kinds of methods to carry out the detection of road edge: a kind ofly for tractor driver, by the reflective mirror on vehicle, to detect the curb line segment of current road; Another kind of for camera to be installed in vehicle, after collecting road edge image, transmit in real time this image to vehicle, for tractor driver, carry out manual detection road edge.
Present inventor finds in long-term R & D, and two kinds of road-edge detection methods of prior art are comparatively complicated for tractor driver's operation requirements, and tractor driver's labour intensity is larger; In the situation that the light environments such as night are darker, tractor driver is difficult to see clearly road edge, and accuracy of detection is lower.
Summary of the invention
The technical matters that the present invention mainly solves is to provide a kind of road-edge detection method, device and vehicle, can realize the curb line segment of the current road that automatic detection vehicle travels, and the operation complexity and the accuracy of detection that reduce tractor driver are higher.
For solving the problems of the technologies described above, the technical scheme that the present invention adopts is: a kind of road-edge detection method is provided, comprises: the picture frame that obtains the road edge information that comprises the current road that vehicle travels; Picture frame is carried out to rim detection, to obtain a plurality of marginal points; Utilize a plurality of marginal points to extract a plurality of straight line line segments; According to the curb architectural characteristic of current road, from a plurality of straight line line segments, extract curb line segment.
The step of wherein, picture frame being carried out to rim detection further comprises: from picture frame, obtain the predefined calibration point topography in presumptive area around; In topography, carry out rim detection.
The step of wherein, carrying out rim detection in topography further comprises: the gray average that calculates the pixel in topography; According to the gray average of the pixel in topography, set low threshold parameter and the high threshold parameter of canny edge detection algorithm, utilize canny edge detection algorithm in topography, to carry out rim detection.
Wherein, the step of extracting curb line segment according to the curb architectural characteristic of current road from a plurality of straight line line segments comprises: according to the comparing result of the pixel distance between the actual range between the road edge of current road and straight line line segment and/or according to the comparing result of the pixel color difference of the actual color difference of the road edge both sides of current road and straight line line segment both sides, from a plurality of straight line line segments, extract curb line segment.
Wherein, according to the comparing result of the pixel distance between the actual range between the road edge of current road and straight line line segment and/or according to the comparing result of the pixel color difference of the actual color difference of the road edge both sides of current road and straight line line segment both sides, from a plurality of straight line line segments, extract before curb line segment and further comprise: from a plurality of straight line line segments, delete the straight line line segment that slope does not meet predetermined slope requirement.
Wherein, the step of extracting curb line segment according to the comparing result of the pixel distance between the actual range between the road edge of current road and straight line line segment from a plurality of straight line line segments comprises: utilize calibration coefficient that the pixel distance between the actual range between the road edge of the current road obtaining under space coordinates and the straight line line segment that obtains under image coordinate system is transformed into the same coordinate system, wherein calibration coefficient by predefined calibration point the actual coordinate under space coordinates and the image coordinate of calibration point under image coordinate system calculate and obtain; Under the same coordinate system, the pixel distance between the actual range between the road edge of current road and straight line line segment is carried out to difference computing, and therefrom select difference to be less than the straight line line segment of redundancy error.
Wherein, the step of extracting curb line segment according to the comparing result of the pixel color difference of the actual color difference of the road edge both sides of current road and straight line line segment both sides from a plurality of straight line line segments comprises: calculate the gray average of the pixel between each straight line line segment and adjacent straight line line segment or in the predetermined lateral width range of each straight line line segment both sides, and according to gray average, determine the pixel color difference of straight line line segment both sides; From a plurality of straight line line segments, extract the consistent or straight line line segment in error allowed band of the actual color difference of the pixel color difference of straight line line segment both sides and the road edge both sides of current road.
Wherein, according to the comparing result of the pixel distance between the actual range between the road edge of current road and straight line line segment with according to the step that the comparing result of the pixel color difference of the actual color difference of the road edge both sides of current road and straight line line segment both sides extracts curb line segment from a plurality of straight line line segments, comprise: utilize calibration coefficient that the pixel distance between the actual range between the road edge of the current road obtaining under space coordinates and the straight line line segment that obtains under image coordinate system is transformed into the same coordinate system, the wherein calibration coefficient actual coordinate under space coordinates and calibration point image coordinate calculating acquisition under image coordinate system by predefined calibration point, under the same coordinate system, the pixel distance between the actual range between the road edge of current road and straight line line segment is carried out to difference computing, and therefrom select difference to be less than a plurality of alternative straight line segments of redundancy error, calculate the gray average of the pixel between each alternative straight line segment and adjacent alternative straight line segment or in the predetermined lateral width range of each alternative straight line segment both sides, and according to gray average, determine the pixel color difference of alternative straight line segment both sides, from a plurality of alternative straight line segments, extract the consistent or alternative straight line segment in error allowed band of the actual color difference of the pixel color difference of alternative straight line segment both sides and the road edge both sides of current road.
Wherein, method further comprises: utilize acquired curb line segment to follow the tracks of a plurality of straight line line segments of the follow-up subsequent image frames of obtaining, and then extract curb line segment from a plurality of straight line line segments of subsequent image frames.
Wherein, method further comprises: the pixel coordinate according to curb line segment under image coordinate system calculates curb line segment actual range with respect to vehicle under space coordinates.
For solving the problems of the technologies described above, another technical scheme that the present invention adopts is: a kind of road-edge detection device is provided, comprises: picture frame acquisition module, for obtaining the picture frame of the road edge information that comprises the current road that vehicle travels; Rim detection module, for picture frame is carried out to rim detection, to obtain a plurality of marginal points; Straight line line segments extraction module, for utilizing a plurality of marginal points to extract a plurality of straight line line segments; Curb line segment extraction module extracts curb line segment according to the curb architectural characteristic of current road from a plurality of straight line line segments.
Wherein, rim detection module is further used for obtaining the predefined calibration point topography in presumptive area around from picture frame, carries out rim detection in Bing topography.
Wherein, rim detection module is further used for calculating the gray average of the pixel in topography, and according to low threshold parameter and the high threshold parameter of the gray average setting canny edge detection algorithm of the pixel in topography, and utilize canny edge detection algorithm in topography, to carry out rim detection.
Wherein, curb line segment extraction module is further used for from a plurality of straight line line segments, extracting curb line segment according to the comparing result of the pixel distance between the actual range between the road edge of current road and straight line line segment and/or according to the comparing result of the pixel color difference of the actual color difference of the road edge both sides of current road and straight line line segment both sides.
Wherein, straight line line segments extraction module is further used for, before curb line segment extraction module extracts curb line segment according to the comparing result of the pixel distance between the actual range between the road edge of current road and straight line line segment and/or according to the comparing result of the pixel color difference of the actual color difference of the road edge both sides of current road and straight line line segment both sides from a plurality of straight line line segments, deleting the straight line line segment that slope does not meet predetermined slope requirement from a plurality of straight line line segments.
Wherein, curb line segment extraction module is further used for utilizing calibration coefficient that the pixel distance between the actual range between the road edge of the current road obtaining under space coordinates and the straight line line segment that obtains under image coordinate system is transformed into the same coordinate system, and under the same coordinate system, the pixel distance between the actual range between the road edge of current road and straight line line segment is carried out to difference computing, and therefrom select difference to be less than the straight line line segment of redundancy error, the wherein calibration coefficient actual coordinate under space coordinates and calibration point image coordinate calculating acquisition under image coordinate system by predefined calibration point.
Wherein, curb line segment extraction module is further used for calculating the gray average of the pixel between each straight line line segment and adjacent straight line line segment or in the predetermined lateral width range of each straight line line segment both sides, and according to gray average, determine the pixel color difference of straight line line segment both sides, and then from a plurality of straight line line segments, extract the consistent or straight line line segment in error allowed band of the actual color difference of the pixel color difference of straight line line segment both sides and the road edge both sides of current road.
Wherein, curb line segment extraction module is further used for utilizing calibration coefficient that the pixel distance between the actual range between the road edge of the current road obtaining under space coordinates and the straight line line segment that obtains under image coordinate system is transformed into the same coordinate system, and under the same coordinate system, the pixel distance between the actual range between the road edge of current road and straight line line segment is carried out to difference computing, and therefrom select difference to be less than a plurality of alternative straight line segments of redundancy error, the wherein calibration coefficient actual coordinate under space coordinates and calibration point image coordinate calculating acquisition under image coordinate system by predefined calibration point, curb line segment extraction module is further used for calculating the gray average of the pixel between each alternative straight line segment and adjacent alternative straight line segment or in the predetermined lateral width range of each alternative straight line segment both sides, and according to gray average, determine the pixel color difference of alternative straight line segment both sides, and then from a plurality of alternative straight line segments, extract the consistent or alternative straight line segment in error allowed band of the actual color difference of the pixel color difference of alternative straight line segment both sides and the road edge both sides of current road.
Wherein, curb line segment extraction module is further used for utilizing acquired curb line segment to follow the tracks of a plurality of straight line line segments of the follow-up subsequent image frames of obtaining, and then extracts curb line segment from a plurality of straight line line segments of subsequent image frames.
Wherein, device further comprises: actual distance calculation module, and for according to curb line segment, the pixel coordinate under image coordinate system calculates curb line segment actual range with respect to vehicle under space coordinates.
For solving the problems of the technologies described above, the another technical scheme that the present invention adopts is: a kind of vehicle is provided, and this vehicle comprises the road-edge detection device of a technical scheme.
The invention has the beneficial effects as follows: be different from the situation of prior art, the present invention is by obtaining the picture frame of the road edge information that comprises the current road that vehicle travels; Picture frame is carried out to rim detection to obtain a plurality of marginal points; Further utilize a plurality of marginal points to extract a plurality of straight line line segments; Finally according to the curb architectural characteristic of current road, from a plurality of straight line line segments, extract curb line segment; Can realize the curb line segment of the current road that automatic detection vehicle travels, the operation complexity and the accuracy of detection that reduce tractor driver are higher.
Accompanying drawing explanation
Fig. 1 is the process flow diagram of road-edge detection method the first embodiment of the present invention;
Fig. 2 is the process flow diagram of road-edge detection method the second embodiment of the present invention;
Fig. 3 is the schematic diagram of vehicle and road edge in road-edge detection method the second embodiment of the present invention;
Fig. 4 is the schematic diagram of the picture frame that comprises road edge information in road-edge detection method the second embodiment of the present invention;
Fig. 5 is the schematic diagram of pixel distance and actual range in road-edge detection method the second embodiment of the present invention;
Fig. 6 is the theory diagram of road-edge detection device one embodiment of the present invention.
Embodiment
Below in conjunction with the accompanying drawing in embodiment of the present invention, the technical scheme in embodiment of the present invention is clearly and completely described, obviously, described embodiment is only the present invention's part embodiment, rather than whole embodiments.Embodiment based in the present invention, those of ordinary skills, not making the every other embodiment obtaining under creative work prerequisite, all belong to the scope of protection of the invention.
Refer to Fig. 1, road-edge detection method the first embodiment of the present invention comprises:
Step S101: the picture frame that obtains the road edge information that comprises the current road that vehicle travels;
In this step, the picture frame of the road edge information that comprises the current road that vehicle travels specifically can gather by being installed on the image capture devices such as video camera on vehicle or camera, and video camera can be digital camera or analog video camera.The camera of video camera can be infrared camera, so that all can carry out the collection of image with night by day.In other embodiments, video camera and camera also can be other kinds, do not make too many restrictions herein.Vehicle can be in there being the vehicle of people's driving condition or automatic Pilot state.Above-mentioned picture frame, except comprising road edge information, also can comprise that the trees of planting in road side, the street lamp that is installed on road side are, the information on other roads such as buildings in roadside.
Step S102: picture frame is carried out to rim detection, to obtain a plurality of marginal points;
In this step, rim detection is the common technique means in image processing and computer vision, and the object of rim detection is that in discriminating digit image, brightness changes obvious point, i.e. marginal point.The algorithm of rim detection can be canny edge detection algorithm or other algorithms well known in the art, below by take canny edge detection algorithm, concrete rim detection mode is described in detail as example.
Step S103: utilize a plurality of marginal points to extract a plurality of straight line line segments;
A plurality of marginal points that in this step, can utilize Hough conversion or other algorithms well known in the art to obtain from step S102 extract a plurality of straight line line segments.In these straight line line segments, comprised all possible marginal information that comprises road edge information.
Step S104: extract curb line segment according to the curb architectural characteristic of current road from a plurality of straight line line segments.
In this step, curb structure comprises all curb structures of the structures such as lane line (track white line or track yellow line etc.) and/or paving stone and/or kerbstone.Curb line segment correspondence comprises left side line, the right-hand line of lane line, and/or left side line, the right-hand line of paving stone, and/or left side line, the right-hand line of kerbstone.Curb architectural characteristic can arrange acquisition by the priori data obtaining in advance for various different highway layouts, mainly can comprise the information such as change color rule of width, paving stone width, sidewalk width and the above-mentioned zones of different of the lane line of different roads.
Be appreciated that road-edge detection method the first embodiment of the present invention is by obtaining the picture frame of the road edge information that comprises the current road that vehicle travels; Picture frame is carried out to rim detection to obtain a plurality of marginal points; Further utilize a plurality of marginal points to extract a plurality of straight line line segments; Finally according to the curb architectural characteristic of current road, from a plurality of straight line line segments, extract curb line segment, can realize the curb line segment of the current road that automatic detection vehicle travels, the operation complexity and the accuracy of detection that reduce tractor driver are higher.
See also Fig. 2-5, road-edge detection method the second embodiment of the present invention comprises:
Step S201: the picture frame that obtains the road edge information that comprises the current road that vehicle travels;
In this step, this picture frame specifically gathers by being installed on the image capture devices such as video camera on vehicle or camera.Specifically, as shown in Figure 3, the image capture device of present embodiment comprises preposition image capture device 3 and/or side image capture device 4, preposition image capture device 3 is specifically installed on pilothouse 1 the place ahead of vehicle to gather the picture frame of the relevant information of the road edge 5 that comprises vehicle front, straight line 31 is the axis, visual angle of preposition image capture device 3, side image capture device 4 is specifically installed on vehicle body 2 sides of vehicle to gather the picture frame of the relevant information of the road edge 5 that comprises automobile side, straight line 41 is the axis, visual angle of side image capture device 4.Below by take the captured picture frame of preposition image capture device 3, as example, present invention is described, and the captured picture frame of the concrete processing mode of the picture frame that side image capture device 4 is captured and preposition image capture device 3 is similar, does not repeat them here.
Step S202: obtain the predefined calibration point topography in presumptive area around from picture frame;
In this step, from picture frame, obtain the predefined calibration point P topography in presumptive area around, described in its process is specific as follows:
Picture frame is carried out to greyscale transformation, and this picture frame is generally coloured image, by greyscale transformation, RGB coloured image is converted to gray level image, to promote the speed that picture frame is processed.
Obtain and take centered by calibration point P and the R of topography that area size is s * t
st, image-region R
stbig or small selection lane line 51, paving stone 52 and the kerbstone 53 of take in can coverage diagram 4 as good.Calibration point P can detect and need to choose setting according to reality, for preposition image capture device 3, as need detect the road edge far away with vehicle distances, chooses from the more longer-distance calibration point as front end of vehicle; For side image capture device 4, can on the axis, visual angle 41 of side image capture device 4, choose any as the calibration point of side.The calibration point of front end and the calibration point of side, after setting, are set without repeating.Certainly, in other embodiments, the above-mentioned R of topography can chosen
stcarry out again afterwards grey scale change, or also can not carry out greyscale transformation in the situation of follow-up edge detection algorithm and Straight Line Extraction permission.
In addition, after setting calibration point P, further comprise and obtain calibration coefficient λ, calibration coefficient λ can be by predefined calibration point P the actual coordinate under space coordinates and the image coordinate of calibration point P under image coordinate system calculate and obtain, the picture frame that the preposition image capture device 3 of take gathers is example, and the horizontal ordinate of calibration point P under space coordinates is x
actual, the horizontal ordinate of calibration point P under image coordinate system is x
image, the axis, visual angle 31 of preposition image capture device 3 is with through calibration point P and perpendicular to the intersection point of the calibration point straight line 6 of y axle, the horizontal ordinate under space coordinates and image coordinate system is respectively x
actual', x
image', calibration coefficient
for guaranteeing the precision of road-edge detection, calibration coefficient λ should be less than 1, i.e. at least corresponding 1 the image pixel distance of effective unit distance.
Step S203: carry out rim detection in topography;
The step of in this step, carrying out rim detection in topography comprises:
Calculate the gray average of the pixel in topography, shown in formula specific as follows (1):
Wherein, the gray average of the pixel in BLWei topography, f (x, y) is the gray-scale value of pixel in topography.
According to the gray average BL of the pixel in topography, set low threshold parameter and the high threshold parameter of canny edge detection algorithm, utilize canny edge detection algorithm in topography, to carry out rim detection.Canny edge detection algorithm is the multistage edge detection algorithm that John F.Canny developed in 1986, and it is as described below that present embodiment utilizes canny edge detection algorithm in topography, to carry out the detailed process of rim detection:
Utilize the level and smooth above-mentioned topography of Gaussian filter to remove picture noise, improve the precision of road-edge detection.
Obtain Grad and the direction value of each pixel in topography, shown in formula specific as follows (2), (3):
Wherein, the Grad that M (x, y) is pixel, the direction value that α (x, y) is pixel, g
x, g
ybe respectively the local derviation of pixel on the x of image coordinate system direction of principal axis and y direction of principal axis.G
x, g
ycan be tried to achieve by Sobel template correspondence, Sobel template adopts Sobel operator, and two group of 3 * 3 matrix of this operator inclusion is respectively horizontal and longitudinal, and two groups of matrixes are respectively:
Above-mentioned two groups of matrixes and image are made to planar convolution, can draw respectively laterally and the approximate value of brightness difference longitudinally, i.e. g
x, g
y.
Utilize Grad M (x, y) and the direction value α (x, y) of pixel to carry out non-maximal value inhibition to obtain candidate pixel point, in candidate pixel point, comprised all marginal points and the non-marginal point of part in topography.
According to the gray average BL of the pixel in topography, set low threshold parameter and the high threshold parameter of canny edge detection algorithm, shown in formula specific as follows (4), (5):
T
L＝BL×γ （4）
T
H＝3×T
L （5）
Wherein, T
lfor low threshold parameter, T
hfor high threshold parameter, γ is light diversity factor coefficient, and γ can measure by experiment or ask optimal value by EM algorithm.EM(Expectation-maximization algorithm) algorithm is greatest hope algorithm, EM algorithm is the algorithm of finding parameter maximal possibility estimation or maximum a posteriori estimation in probability model, and wherein probability model depends on the hidden variable that cannot observe.
Utilize low threshold parameter T
l, high threshold parameter T
hfurther in candidate pixel point, obtain a plurality of marginal points.Wherein, be less than low threshold parameter T
lcandidate pixel point be non-marginal point, be greater than high threshold parameter T
hcandidate pixel point be marginal point, at T
l-T
hbetween candidate pixel point may be marginal point, present embodiment is at T
l-T
hbetween choose contiguous high threshold parameter T
hthe candidate pixel point of certain limit and choose and be greater than high threshold parameter T
hcandidate pixel point be marginal point.
Step S204: utilize a plurality of marginal points to extract a plurality of straight line line segments;
In this step, specifically by Hough mapping mode, utilize a plurality of marginal points to extract a plurality of straight line line segments, Hough conversion is a kind of parameter estimation techniques of using voting principle, it utilizes the duality of image space and Hough parameter space, and the test problems in image space is transformed into parameter space.The first step of extracting straight line line segment by Hough mapping mode is to obtain the polar equation of the straight line line segment that a plurality of marginal points are corresponding, shown in formula specific as follows (6):
ρ＝xcosθ+ysinθ （6）
Wherein, formula (6) is the polar equation of straight line line segment, and x, y are respectively horizontal ordinate, the ordinate of marginal point in image coordinate system, the utmost point footpath that ρ is marginal point, the polar angle that θ is marginal point.
Second step is for to be converted to corresponding rectangular equation by the polar equation of straight line line segment.
Step S205: delete the straight line line segment that slope does not meet predetermined slope requirement from a plurality of straight line line segments;
In this step, from image-forming principle, when vehicle is during along straight line road driving, the road edge of its both sides becomes certain slope at captured picture frame.For example, the slope for the right-side course Road Edge of vehicle in picture frame is born, and the slope in picture frame is positive for the left-side course Road Edge of vehicle, and edge line between paving stone 52 slope on picture frame is substantially close to zero.Therefore, different according to the chosen position of calibration point P, and according to the prior imformation of the curb structure of the current road to obtain, by image-forming principle, can calculate the theoretical slope of different road edges in picture frame, consider certain redundancy error simultaneously, and then can determine the slope requirement of different road edges, and then eliminating not obviously the straight line line segment of road edge line, the straight line line segment of the edge line between paving stone 52 for example, can improve the efficiency of follow-up road-edge detection.
Further, in other embodiments, GPS function that can be by vehicle or angle inductor etc. detect vehicle heading, thereby change above-mentioned slope requirement according to vehicle heading, for example, when vehicle turns to the right, the slope of the right-side course Road Edge of vehicle in picture frame levels off to zero, therefore need to adjust above-mentioned slope requirement according to actual conditions.
Therefore before the above-mentioned second step by Hough mapping mode extraction straight line line segment, also comprise: the span that limits polar angle θ is calculated corresponding utmost point footpath ρ, and by corresponding ρ, θ parameter matrix unit accumulated counts, select cumulative unit larger in ρ, θ parameter matrix unit and then determine shown in the polar equation formula specific as follows (7) that meets the straight line line segment that predetermined slope requires:
ρ
i＝xcosθ
i+ysinθ
i （7）
In the present embodiment, θ
ispan be [90,0].
Further the polar equation (7) that meets the straight line line segment of predetermined slope requirement is converted to rectangular equation, shown in formula specific as follows (8):
y＝f
i(x) （8）
Wherein, formula (8) is for meeting the rectangular equation of the straight line line segment of predetermined slope requirement.
Step S206: extract curb line segment according to the curb architectural characteristic of current road from a plurality of straight line line segments;
In this step, can from a plurality of straight line line segments, extract curb line segment according to the comparing result of the pixel distance between the actual range between the road edge of current road and straight line line segment and/or according to the comparing result of the pixel color difference of the actual color difference of the road edge both sides of current road and straight line line segment both sides.
When extracting curb line segment according to the comparing result of the pixel distance between the actual range between the road edge of current road and straight line line segment from a plurality of straight line line segments, it specifically comprises:
Utilize calibration coefficient λ that the pixel distance between the actual range between the road edge of the current road obtaining under space coordinates and the straight line line segment that obtains under image coordinate system is transformed into the same coordinate system.Actual range between the road edge of current road is the prior imformation of road edge, specifically comprises the left side line 531 of kerbstone 53 and the actual range S of right-hand line 532
curb, the left side line 521 of paving stone 52 and the actual range S of right-hand line 522
roundand the left side line 511 of lane line 51 and the actual range S of right-hand line 512
white, S
whitenumerical range be generally respectively 12-15cm, 40cm, 10-12cm, the numerical range of above-mentioned prior imformation also can be other sizes, does not make too many restrictions herein.Obtain the process of the pixel distance between straight line line segment specific as follows described in:
Obtain and meet the straight line line segment of predetermined slope requirement and the coordinate of the intersection point of calibration point straight line 6 in image coordinate system, be specially: when picture frame is gathered by preposition image capture device 3, calibration point straight line 6 is for through calibration point P and perpendicular to the straight line of y axle, the horizontal ordinate of the intersection point that obtains straight line line segment and calibration point straight line 6 in image coordinate system, shown in formula specific as follows (9):
x
i=f
i'(y
P) （9）
Wherein, x
ifor the intersection point of straight line line segment and calibration point straight line 6 horizontal ordinate in image coordinate system, y
pfor the ordinate of calibration point P in image coordinate system.
Horizontal ordinate x to the intersection point of straight line line segment and calibration point straight line 6 in image coordinate system
isort by size, further utilize the horizontal ordinate x of intersection point
iobtain the pixel distance between straight line line segment, shown in formula specific as follows (10):
d
k=|x
i-x
j| （10）
Wherein, d
kfor the pixel distance between straight line line segment, i < j.
In addition, when picture frame is gathered by side image capture device 4, calibration point straight line is for through calibration point P and perpendicular to the straight line of x axle, in like manner the ordinate of the intersection point by obtaining straight line line segment and calibration point straight line in image coordinate system and then obtain the pixel distance between straight line line segment now.
After the pixel distance obtaining between straight line line segment, utilize calibration coefficient λ that the pixel distance between the actual range between road edge and straight line line segment is transformed into the same coordinate system, it comprises and actual range is converted to corresponding pixel distance so that actual range and pixel distance are in together under an image coordinate system, or the pixel distance between straight line line segment is converted to corresponding actual range so that actual range and pixel distance are in together under space coordinates.For example, for actual range is transformed under image coordinate system, above-mentioned each actual range S
curb, S
round, S
whitecorresponding pixel distance is d
curb, d
ground, d
white,
Be appreciated that pixel distance is converted to corresponding actual range is: by pixel distance d
kbe multiplied by calibration coefficient λ and can obtain corresponding actual range.
Further under the same coordinate system, the pixel distance between the actual range between the road edge of current road and straight line line segment is carried out to difference computing, and therefrom select difference to be less than the straight line line segment of redundancy error.Take under image coordinate system is example, and difference is less than shown in pixel distance formula specific as follows (11) between the straight line line segment of redundancy error, (12), (13):
D
curb={d
k:|d
curb-d
k|<e} （11）
D
ground={d
k:|d
ground-d
k|<e} （12）
D
white={d
k:|d
white-d
k|<e} （13）
Wherein, ε is redundancy error, 0< ε <0.5*min{d
curb, d
ground, d
white.D
curb, D
ground, D
whitefor difference corresponding to kerbstone, paving stone and lane line is less than the pixel distance between the straight line line segment of redundancy error ε, further obtain D
curb, D
ground, D
whitecorresponding difference is less than the straight line line segment of redundancy error ε, the straight line line segment that difference is less than redundancy error ε is curb line segment, and curb line segment comprises the left side line 511 and right-hand line 512 of the left side line 531 of kerbstone 53 and the left side line 521 of right-hand line 532, paving stone 52 and right-hand line 522 and lane line 51.
When extracting curb line segment according to the comparing result of the pixel color difference of the actual color difference of the road edge both sides of current road and straight line line segment both sides from a plurality of straight line line segments, it specifically comprises:
Calculate the gray average of the pixel between each straight line line segment and adjacent straight line line segment or in the predetermined lateral width range of each straight line line segment both sides.When picture frame is gathered by preposition image capture device 3, shown in gray average formula specific as follows (14):
Wherein,
for the gray average of the pixel between each straight line line segment and adjacent straight line line segment or in predetermined lateral width range, type ∈ curb, and ground, white},
represent respectively Dt
ypein t candidate left and right sides straight line line segment with through calibration point P and perpendicular to the horizontal ordinate of the intersection point of the calibration point straight line 6 of y axle, or form the left and right horizontal ordinate of predetermined lateral width range.Predetermined lateral width range can be set as the width range such as 1/2,1/3 of pixel distance between adjacent straight line line segment, and predetermined lateral width range also can be set to such as fixed values such as 2 pixel unit distances; For example, for straight line line segment a, the predetermined lateral width range of its both sides can be set to respectively 1/4 of straight line line segment a and adjacent left side, the pixel distance between the straight line line segment of right side.Be appreciated that when picture frame is gathered by side image capture device 4, correspondence utilize straight line line segment with through calibration point P and perpendicular to the ordinate of the intersection point of the calibration point straight line of x axle, and the horizontal ordinate of calibration point P obtains gray average.
Further according to gray average
determine the pixel color difference of straight line line segment both sides, pixel color difference is the difference of the gray average of straight line line segment both sides; From a plurality of straight line line segments, extract the consistent or straight line line segment in error allowed band of the actual color difference of the pixel color difference of straight line line segment both sides and the road edge both sides of current road, the straight line line segment extracting is curb line segment.Curb line segment comprises the left side line 511 and right-hand line 512 of the left side line 531 of kerbstone 53 and the left side line 521 of right-hand line 532, paving stone 52 and right-hand line 522 and lane line 51.Wherein, the actual grey average between the left and right sides straight line of kerbstone 53 is V
curb, the actual grey average between the left and right sides straight line of paving stone 52 is V
ground, the actual grey average between the left and right sides straight line of lane line 51 is V
white, the magnitude relationship of three kinds of actual grey averages is: V
white>V
curb>V
ground.The pixel color difference that for example obtains straight line line segment b both sides is c, and the pixel color difference of left side line 531 both sides of the kerbstone 53 of current road is also c, therefore can determine that straight line line segment b corresponds to the left side line 531 of kerbstone 53.
Further, in the present embodiment, adopt two kinds of modes of distance calculating and color contrast curb line segment to be extracted simultaneously, according to the comparing result of the pixel distance between the actual range between the road edge of current road and straight line line segment with according to the comparing result of the pixel color difference of the actual color difference of the road edge both sides of current road and straight line line segment both sides, from a plurality of straight line line segments, extract curb line segment, it specifically comprises:
Utilize calibration coefficient that the pixel distance between the actual range between the road edge of the current road obtaining under space coordinates and the straight line line segment that obtains under image coordinate system is transformed into the same coordinate system; Under the same coordinate system, the pixel distance between the actual range between the road edge of current road and straight line line segment is carried out to difference computing, and therefrom select difference to be less than a plurality of alternative straight line segments of redundancy error.
Calculate the gray average of the pixel between each alternative straight line segment and adjacent alternative straight line segment or in the predetermined lateral width range of each alternative straight line segment both sides, and according to gray average, determine the pixel color difference of alternative straight line segment both sides; From a plurality of alternative straight line segments, extract the consistent or alternative straight line segment in error allowed band of the actual color difference of the pixel color difference of alternative straight line segment both sides and the road edge both sides of current road.
Step S207: utilize acquired curb line segment to follow the tracks of a plurality of straight line line segments of the follow-up subsequent image frames of obtaining, and then extract curb line segment from a plurality of straight line line segments of subsequent image frames;
Utilize acquired curb line segment to follow the tracks of a plurality of straight line line segments of the follow-up subsequent image frames of obtaining, and then extract curb line segment from a plurality of straight line line segments of subsequent image frames.Specifically can utilize nearest neighbour method or Kalman filter method to follow the tracks of a plurality of straight line line segments of subsequent image frames.
The process of utilizing nearest neighbour method to follow the tracks of straight line line segment specifically comprises: the coordinate of a plurality of straight line line segments that obtain subsequent image frames under image coordinate system, and then this coordinate is deducted to the coordinate of each curb line segment of a upper picture frame under image coordinate system, when two error of coordinates are less than redundancy error ε, determine that the straight line line segment of subsequent image frames and the curb line segment of a upper picture frame are same straight line.For example utilize straight line line segment that above-mentioned formula (9) obtains subsequent image frames with through calibration point P and perpendicular to the intersection point of the calibration point straight line 6 of y axle the horizontal ordinate under image coordinate system; The horizontal ordinate of intersection point is deducted to the horizontal ordinate of the intersection point of each curb line segment of a upper picture frame and calibration point straight line 6, if the difference of this straight line line segment and a certain curb line segment horizontal ordinate is less than redundancy error ε and is defined as same straight line: | x " x ' | <e; wherein " be the horizontal ordinate of the straight line line segment of subsequent image frames and the intersection point of calibration point straight line 6, x ' is the horizontal ordinate of a certain curb line segment of a upper picture frame and the intersection point of calibration point straight line 6 to x.
Kalman filter is a kind of regressive filter for time-varying linear systems being proposed by Kalman (Kalman), this time-varying linear systems available packages is described containing the Differential Equation Model of quadrature variable, and Kalman filter is the measurement evaluated error in past to be merged to new measuring error estimate error in the future.Kalman filter method, by the coordinate points data of straight line line segment, curb line segment are expressed as to Kalman filter, is utilized the principle of Kalman filter, and each straight line line segment is followed the tracks of.
After the tracking of each straight line line segment that completes a picture frame, repeat to upgrade the coordinate information of current each curb line segment.At water stain, shade, block etc. under complex working condition, the true curb of some picture frame may be difficult to detect, by above two kinds of trackings, utilize the historical information data of road edge structure can realize under complex working condition straight line line segment is followed the tracks of, and then extract curb line segment.
Step S208: the pixel coordinate according to curb line segment under image coordinate system calculates curb line segment actual range with respect to vehicle under space coordinates.
Pixel coordinate according to curb line segment under image coordinate system calculates curb line segment actual range with respect to vehicle under space coordinates.In present embodiment for pixel coordinate under image coordinate system of the right-hand line 522 of utilizing the paving stone 52 in curb line segment and then calculate actual range: when picture frame is gathered by preposition image capture device 3, as shown in Figure 5, obtain the right-hand line 522 of the paving stone 52 in curb line segment and the horizontal ordinate x of the intersection point of calibration point straight line 6
1, and the axis, visual angle 31 of preposition image capture device 3 and the horizontal ordinate x of the intersection point of calibration point straight line 6
2, by two horizontal ordinate x
1, x
2subtract each other and obtain the pixel distance L between right-hand line 522 and axis, visual angle 31
1=| x
1-x
2|, further utilize calibration coefficient λ and pixel distance L
1calculate curb line segment actual range S with respect to vehicle under space coordinates
1=λ * L
1; When picture frame is gathered by side image capture device 4, calibration point is on the axis, visual angle 41 of side image capture device 4, now calibration point straight line is for through calibration point P and perpendicular to the straight line of x axle, calibration point straight line is the axis, visual angle 41 of side image capture device 4, now obtains the ordinate L of the intersection point of right-hand line 522 and calibration point straight line
2, and then utilize calibration coefficient λ and this ordinate L
2calculate curb line segment actual range S with respect to vehicle under space coordinates
2=λ * L
2.
In other embodiments, also can utilize the pixel coordinate under image coordinate system such as other curb line segments such as left side line 531 grades such as kerbstone 53 calculate curb line segment under space coordinates with respect to the actual range of vehicle, do not make too many restrictions herein.
For utilizing the picture frame of side image capture device collection, correspondence is calculated current time curb line segment with respect to the actual range of vehicle, can further to the actual range of current time, judge, if the actual range of current time is less than a certain predeterminable range threshold value, by sending the modes such as warning sound or word demonstration, point out the actual range of current time vehicle and curb line segment to exceed safe distance scope, tractor driver or automated driving system can be adjusted according to this information the actual range of vehicle and curb line segment.For utilizing the picture frame of preposition image capture device collection, correspondence obtains curb line segment with respect to the actual range of vehicle, this actual range is the anticipation of the actual range between a certain moment in future and curb line segment to vehicle, can judge this actual range equally, if it surpasses a certain predeterminable range threshold value, send equally information and soon exceed safe distance scope with the actual range of curb line segment with prompting vehicle, tractor driver or automated driving system can be adjusted the actual range of vehicle and curb line segment in advance.
In addition, calculating curb line segment under space coordinates after the actual range with respect to vehicle, further on vehicle, show in real time the actual range of curb line segment and vehicle, specifically can carry out the real-time demonstration of actual range by the display screen being installed on vehicle.Above-mentioned information can show equally on vehicle.
Be appreciated that road-edge detection method the second embodiment of the present invention is by obtaining the picture frame of the road edge information that comprises the current road that vehicle travels, from picture frame, obtain the predefined calibration point topography in presumptive area around; To carrying out rim detection in topography to obtain a plurality of marginal points; Further utilize a plurality of marginal points to extract a plurality of straight line line segments; From a plurality of straight line line segments, delete the straight line line segment that slope does not meet predetermined slope requirement; According to the comparing result of the pixel distance between the actual range between the road edge of current road and straight line line segment and/or according to the comparing result of the pixel color difference of the actual color difference of the road edge both sides of current road and straight line line segment both sides, from a plurality of straight line line segments, extract curb line segment; Utilize acquired curb line segment a plurality of straight line line segments of subsequent image frames are followed the tracks of and then extracted curb line segment from a plurality of straight line line segments of subsequent image frames; It is last that according to curb line segment, the pixel coordinate under image coordinate system calculates curb line segment actual range with respect to vehicle under space coordinates.
By the way, can realize the curb line segment of the current road that automatic detection vehicle travels, the operation complexity and the accuracy of detection that reduce tractor driver are higher.In addition by carrying out rim detection and delete the straight line line segment that slope does not meet predetermined slope requirement in topography, can improve the efficiency of road-edge detection; By straight line line segment being followed the tracks of and can be realized equally rapid extraction curb line segment in subsequent image frames; It is last that according to curb line segment, the pixel coordinate under image coordinate system calculates curb line segment actual range with respect to vehicle under space coordinates, can realize the actual range between the following a certain moment vehicle of actual range between automatic measurement current time vehicle and curb line segment and anticipation and curb line segment, reduce tractor driver's operation complexity and the range accuracy of acquisition higher, realize the safe driving of vehicle.
Refer to Fig. 6, road-edge detection device one embodiment of the present invention comprises:
Picture frame acquisition module 71, rim detection module 72, straight line line segments extraction module 73 and curb line segment extraction module 74.
Picture frame acquisition module 71 is for obtaining the picture frame of the road edge information that comprises the current road that vehicle travels.Picture frame acquisition module 71 is specially preposition image capture device or the side image capture device described in the respective embodiments described above.
In topography, carry out rim detection module 72 in the process of rim detection and be further used for calculating the gray average of the pixel in topography, and according to low threshold parameter and the high threshold parameter of the gray average setting canny edge detection algorithm of the pixel in topography, and utilize canny edge detection algorithm in topography, to carry out rim detection.
Straight line line segments extraction module 73 is for utilizing a plurality of marginal points to extract a plurality of straight line line segments.
Curb line segment extraction module 74 is for extracting curb line segment according to the curb architectural characteristic of current road from a plurality of straight line line segments.Curb line segment extraction module 74 is further used for from a plurality of straight line line segments, extracting curb line segment according to the comparing result of the pixel distance between the actual range between the road edge of current road and straight line line segment and/or according to the comparing result of the pixel color difference of the actual color difference of the road edge both sides of current road and straight line line segment both sides.
Straight line line segments extraction module 73 is further used for, before curb line segment extraction module 74 extracts curb line segment according to the comparing result of the pixel distance between the actual range between the road edge of current road and straight line line segment and/or according to the comparing result of the pixel color difference of the actual color difference of the road edge both sides of current road and straight line line segment both sides from a plurality of straight line line segments, deleting the straight line line segment that slope does not meet predetermined slope requirement from a plurality of straight line line segments.
When curb line segment extraction module 74 is when extracting curb line segment according to the comparing result of the pixel distance between the actual range between the road edge of current road and straight line line segment from a plurality of straight line line segments, it is further used for:
Utilize calibration coefficient that the pixel distance between the actual range between the road edge of the current road obtaining under space coordinates and the straight line line segment that obtains under image coordinate system is transformed into the same coordinate system; And under the same coordinate system, the pixel distance between the actual range between the road edge of current road and straight line line segment is carried out to difference computing, and therefrom select difference to be less than the straight line line segment of redundancy error.The wherein calibration coefficient actual coordinate under space coordinates and calibration point image coordinate calculating acquisition under image coordinate system by predefined calibration point.
When curb line segment extraction module 74 is when extracting curb line segment according to the comparing result of the pixel color difference of the actual color difference of the road edge both sides of current road and straight line line segment both sides from a plurality of straight line line segments, it is further used for:
Calculate the gray average of the pixel between each straight line line segment and adjacent straight line line segment or in the predetermined lateral width range of each straight line line segment both sides, and according to gray average, determine the pixel color difference of straight line line segment both sides; And then the actual color difference of extracting the pixel color difference of straight line line segment both sides and the road edge both sides of current road from a plurality of straight line line segments is consistent or the straight line line segment in error allowed band.
When curb line segment extraction module 74 is when extracting curb line segment according to the comparing result of the pixel distance between the actual range between the road edge of current road and straight line line segment with according to the comparing result of the pixel color difference of the actual color difference of the road edge both sides of current road and straight line line segment both sides from a plurality of straight line line segments, it is further used for:
Utilize calibration coefficient that the pixel distance between the actual range between the road edge of the current road obtaining under space coordinates and the straight line line segment that obtains under image coordinate system is transformed into the same coordinate system; And under the same coordinate system, the pixel distance between the actual range between the road edge of current road and straight line line segment is carried out to difference computing, and therefrom select difference to be less than a plurality of alternative straight line segments of redundancy error.The wherein calibration coefficient actual coordinate under space coordinates and calibration point image coordinate calculating acquisition under image coordinate system by predefined calibration point.
Curb line segment extraction module 74 is further used for calculating the gray average of the pixel between each alternative straight line segment and adjacent alternative straight line segment or in the predetermined lateral width range of each alternative straight line segment both sides, and according to gray average, determines the pixel color difference of alternative straight line segment both sides; And then the actual color difference of extracting the pixel color difference of alternative straight line segment both sides and the road edge both sides of current road from a plurality of alternative straight line segments is consistent or the alternative straight line segment in error allowed band.
After extracting curb line segment, curb line segment extraction module 74 is further used for utilizing acquired curb line segment to follow the tracks of a plurality of straight line line segments of the follow-up subsequent image frames of obtaining, and then extracts curb line segment from a plurality of straight line line segments of subsequent image frames.
In addition, road-edge detection device further comprises: actual distance calculation module, and for according to curb line segment, the pixel coordinate under image coordinate system calculates curb line segment actual range with respect to vehicle under space coordinates.
The present invention also provides a kind of vehicle, and this vehicle comprises the road-edge detection device described in above-mentioned embodiment, by this road-edge detection device, realizes the vehicle road edge of the current road of real time automatic detection in the process of moving.
The foregoing is only embodiments of the present invention; not thereby limit the scope of the claims of the present invention; every equivalent structure or conversion of equivalent flow process that utilizes instructions of the present invention and accompanying drawing content to do; or be directly or indirectly used in other relevant technical fields, be all in like manner included in scope of patent protection of the present invention.
Claims (21)
1. a road-edge detection method, is characterized in that, comprising:
Obtain the picture frame of the road edge information that comprises the current road that vehicle travels;
Described picture frame is carried out to rim detection, to obtain a plurality of marginal points;
Utilize described a plurality of marginal point to extract a plurality of straight line line segments;
According to the curb architectural characteristic of described current road, from described a plurality of straight line line segments, extract curb line segment.
2. method according to claim 1, is characterized in that, the described step that described picture frame is carried out to rim detection further comprises:
From described picture frame, obtain the predefined calibration point topography in presumptive area around;
In described topography, carry out rim detection.
3. method according to claim 2, is characterized in that, described step of carrying out rim detection in described topography further comprises:
Calculate the gray average of the pixel in described topography;
According to the gray average of the pixel in described topography, set low threshold parameter and the high threshold parameter of canny edge detection algorithm, utilize described canny edge detection algorithm in described topography, to carry out rim detection.
4. method according to claim 1, is characterized in that, the described step of extracting curb line segment according to the curb architectural characteristic of described current road from described a plurality of straight line line segments comprises:
According to the comparing result of the pixel distance between the actual range between the road edge of described current road and described straight line line segment and/or according to the comparing result of the pixel color difference of the actual color difference of the road edge both sides of described current road and described straight line line segment both sides, from described a plurality of straight line line segments, extract described curb line segment.
5. method according to claim 4, it is characterized in that, describedly according to the comparing result of the pixel distance between the actual range between the road edge of described current road and described straight line line segment and/or according to the comparing result of the pixel color difference of the actual color difference of the road edge both sides of described current road and described straight line line segment both sides, from described a plurality of straight line line segments, extract before described curb line segment and further comprise:
From described a plurality of straight line line segments, delete the described straight line line segment that slope does not meet predetermined slope requirement.
6. method according to claim 4, it is characterized in that, the described step of extracting described curb line segment according to the comparing result of the pixel distance between the actual range between the road edge of described current road and described straight line line segment from described a plurality of straight line line segments comprises:
Utilize calibration coefficient that the pixel distance between the actual range between the road edge of the described current road obtaining under space coordinates and the described straight line line segment that obtains under image coordinate system is transformed into the same coordinate system, wherein said calibration coefficient by predefined calibration point the actual coordinate under described space coordinates and the image coordinate of described calibration point under described image coordinate system calculate and obtain;
Under described the same coordinate system, the pixel distance between the actual range between the road edge of described current road and described straight line line segment is carried out to difference computing, and therefrom select difference to be less than the described straight line line segment of redundancy error.
7. method according to claim 4, it is characterized in that, the described step of extracting described curb line segment according to the comparing result of the pixel color difference of the actual color difference of the road edge both sides of described current road and described straight line line segment both sides from described a plurality of straight line line segments comprises:
Calculate the gray average of the pixel in the predetermined lateral width range of straight line line segment both sides between straight line line segment described in each and adjacent described straight line line segment or described in each, and according to described gray average, determine the pixel color difference of described straight line line segment both sides;
From described a plurality of straight line line segments, extract the pixel color difference of described straight line line segment both sides and the actual color difference of the road edge both sides of described current road is consistent or the straight line line segment in error allowed band.
8. method according to claim 4, it is characterized in that, describedly according to the comparing result of the pixel distance between the actual range between the road edge of described current road and described straight line line segment with according to the step that the comparing result of the pixel color difference of the actual color difference of the road edge both sides of described current road and described straight line line segment both sides extracts described curb line segment from described a plurality of straight line line segments, comprise:
Utilize calibration coefficient that the pixel distance between the actual range between the road edge of the described current road obtaining under space coordinates and the described straight line line segment that obtains under image coordinate system is transformed into the same coordinate system, wherein said calibration coefficient by predefined calibration point the actual coordinate under described space coordinates and the image coordinate of described calibration point under described image coordinate system calculate and obtain;
Under described the same coordinate system, the pixel distance between the actual range between the road edge of described current road and described straight line line segment is carried out to difference computing, and therefrom select difference to be less than a plurality of alternative straight line segments of redundancy error;
Calculate the gray average of the pixel in the predetermined lateral width range of alternative straight line segment both sides between alternative straight line segment described in each and adjacent described alternative straight line segment or described in each, and according to described gray average, determine the pixel color difference of described alternative straight line segment both sides;
From described a plurality of alternative straight line segments, extract the pixel color difference of described alternative straight line segment both sides and the actual color difference of the road edge both sides of described current road is consistent or the alternative straight line segment in error allowed band.
9. method according to claim 1, is characterized in that, described method further comprises:
Utilize acquired described curb line segment to follow the tracks of a plurality of described straight line line segment of the follow-up subsequent image frames of obtaining, and then extract described curb line segment from a plurality of described straight line line segment of subsequent image frames.
10. according to the method described in claim 1-9 any one, it is characterized in that, described method further comprises:
According to described curb line segment, at the pixel coordinate under image coordinate system, calculate described curb line segment actual range with respect to described vehicle under space coordinates.
11. 1 kinds of road-edge detection devices, is characterized in that, comprising:
Picture frame acquisition module (71), for obtaining the picture frame of the road edge information that comprises the current road that vehicle travels;
Rim detection module (72), for described picture frame is carried out to rim detection, to obtain a plurality of marginal points;
Straight line line segments extraction module (73), for utilizing described a plurality of marginal point to extract a plurality of straight line line segments;
Curb line segment extraction module (74) extracts curb line segment according to the curb architectural characteristic of described current road from described a plurality of straight line line segments.
12. road-edge detection devices according to claim 11, is characterized in that,
Described rim detection module (72) is further used for obtaining the predefined calibration point topography in presumptive area around from described picture frame, and in described topography, carries out rim detection.
13. road-edge detection devices according to claim 12, is characterized in that,
Described rim detection module (72) is further used for calculating the gray average of the pixel in described topography, and according to low threshold parameter and the high threshold parameter of the gray average setting canny edge detection algorithm of the pixel in described topography, and utilize described canny edge detection algorithm in described topography, to carry out rim detection.
14. road-edge detection devices according to claim 11, is characterized in that,
Described curb line segment extraction module (74) is further used for from described a plurality of straight line line segments, extracting described curb line segment according to the comparing result of the pixel distance between the actual range between the road edge of described current road and described straight line line segment and/or according to the comparing result of the pixel color difference of the actual color difference of the road edge both sides of described current road and described straight line line segment both sides.
15. road-edge detection devices according to claim 14, is characterized in that,
Described straight line line segments extraction module (73) is further used for, before described curb line segment extraction module (74) extracts described curb line segment according to the comparing result of the pixel distance between the actual range between the road edge of described current road and described straight line line segment and/or according to the comparing result of the pixel color difference of the actual color difference of the road edge both sides of described current road and described straight line line segment both sides from described a plurality of straight line line segments, deleting the described straight line line segment that slope does not meet predetermined slope requirement from described a plurality of straight line line segments.
16. road-edge detection devices according to claim 14, is characterized in that,
Described curb line segment extraction module (74) is further used for utilizing calibration coefficient that the pixel distance between the actual range between the road edge of the described current road obtaining under space coordinates and the described straight line line segment that obtains under image coordinate system is transformed into the same coordinate system, and under described the same coordinate system, the pixel distance between the actual range between the road edge of described current road and described straight line line segment is carried out to difference computing, and therefrom select difference to be less than the described straight line line segment of redundancy error, wherein said calibration coefficient is the actual coordinate under described space coordinates and the image coordinate calculating acquisition of described calibration point under described image coordinate system by predefined calibration point.
17. road-edge detection devices according to claim 14, is characterized in that,
Described curb line segment extraction module (74) is further used for calculating the gray average of the pixel in the predetermined lateral width range of straight line line segment both sides between straight line line segment described in each and adjacent described straight line line segment or described in each, and according to described gray average, determine the pixel color difference of described straight line line segment both sides, and then from described a plurality of straight line line segments, extract the pixel color difference of described straight line line segment both sides and the actual color difference of the road edge both sides of described current road is consistent or the straight line line segment in error allowed band.
18. road-edge detection devices according to claim 14, is characterized in that,
Described curb line segment extraction module (74) is further used for utilizing calibration coefficient that the pixel distance between the actual range between the road edge of the described current road obtaining under space coordinates and the described straight line line segment that obtains under image coordinate system is transformed into the same coordinate system, and under described the same coordinate system, the pixel distance between the actual range between the road edge of described current road and described straight line line segment is carried out to difference computing, and therefrom select difference to be less than a plurality of alternative straight line segments of redundancy error, wherein said calibration coefficient is the actual coordinate under described space coordinates and the image coordinate calculating acquisition of described calibration point under described image coordinate system by predefined calibration point,
Described curb line segment extraction module (74) is further used for calculating the gray average of the pixel in the predetermined lateral width range of alternative straight line segment both sides between alternative straight line segment described in each and adjacent described alternative straight line segment or described in each, and according to described gray average, determine the pixel color difference of described alternative straight line segment both sides, and then from described a plurality of alternative straight line segments, extract the pixel color difference of described alternative straight line segment both sides and the actual color difference of the road edge both sides of described current road is consistent or the alternative straight line segment in error allowed band.
19. road-edge detection devices according to claim 11, is characterized in that,
Described curb line segment extraction module (74) is further used for utilizing acquired described curb line segment to follow the tracks of a plurality of described straight line line segment of the follow-up subsequent image frames of obtaining, and then extracts described curb line segment from a plurality of described straight line line segment of subsequent image frames.
20. according to the road-edge detection device described in claim 11-19 any one, it is characterized in that, described device further comprises:
Actual distance calculation module, for calculating described curb line segment actual range with respect to described vehicle under space coordinates according to described curb line segment at the pixel coordinate under image coordinate system.
21. 1 kinds of vehicles, is characterized in that, described vehicle comprises the road-edge detection device as described in claim 11-20 any one.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310711070.3A CN103714538B (en) | 2013-12-20 | 2013-12-20 | road edge detection method, device and vehicle |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310711070.3A CN103714538B (en) | 2013-12-20 | 2013-12-20 | road edge detection method, device and vehicle |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103714538A true CN103714538A (en) | 2014-04-09 |
| CN103714538B CN103714538B (en) | 2016-12-28 |
Family
ID=50407481
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310711070.3A Active CN103714538B (en) | 2013-12-20 | 2013-12-20 | road edge detection method, device and vehicle |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103714538B (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103927526A (en) * | 2014-04-30 | 2014-07-16 | 长安大学 | Vehicle detecting method based on Gauss difference multi-scale edge fusion |
| CN105224908A (en) * | 2014-07-01 | 2016-01-06 | 北京四维图新科技股份有限公司 | A kind of roadmarking acquisition method based on orthogonal projection and device |
| CN106017313A (en) * | 2015-03-31 | 2016-10-12 | 株式会社三丰 | Edge detection bias correction value calculation method, edge detection bias correction method, and equipment |
| CN106491040A (en) * | 2016-11-14 | 2017-03-15 | 浙江工业大学 | A kind of swing type window wiping robot glass edge touches detection process method |
| CN106491039A (en) * | 2016-11-14 | 2017-03-15 | 浙江工业大学 | A kind of swing type window wiping robot glass edge detection method |
| CN106491042A (en) * | 2016-11-14 | 2017-03-15 | 浙江工业大学 | Processing method is moved down at a kind of swing type window wiping robot glass edge |
| CN106546185A (en) * | 2016-10-18 | 2017-03-29 | 福州觉感视觉软件科技有限公司 | A kind of profile quality determining method based on Machine Vision Detection |
| CN106663193A (en) * | 2014-05-14 | 2017-05-10 | 御眼视觉技术有限公司 | Systems and methods for curb detection and pedestrian hazard assessment |
| CN106650726A (en) * | 2016-12-05 | 2017-05-10 | 渤海大学 | License plate recognition method |
| CN107133625A (en) * | 2017-03-31 | 2017-09-05 | 纵目科技（上海）股份有限公司 | The extracting method and system, car-mounted terminal of a kind of yellow lane line |
| CN107194940A (en) * | 2017-05-23 | 2017-09-22 | 北京计算机技术及应用研究所 | A kind of coloured image contour extraction method based on color space and line segment |
| CN107392080A (en) * | 2016-05-17 | 2017-11-24 | 纬创资通股份有限公司 | Image evaluation method and electronic device thereof |
| CN107862699A (en) * | 2017-09-22 | 2018-03-30 | 中国科学院深圳先进技术研究院 | Bone edges extracting method, device, equipment and the storage medium of Bone CT image |
| CN108416306A (en) * | 2018-03-12 | 2018-08-17 | 海信集团有限公司 | Continuous type obstacle detection method, device, equipment and storage medium |
| CN109740469A (en) * | 2018-12-24 | 2019-05-10 | 百度在线网络技术（北京）有限公司 | Method for detecting lane lines, device, computer equipment and storage medium |
| WO2023124221A1 (en) * | 2021-12-31 | 2023-07-06 | 中国第一汽车股份有限公司 | Area edge detection method and apparatus for vehicle drivable area, and storage medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020061123A1 (en) * | 2000-11-22 | 2002-05-23 | Nissan Motor Co., Ltd. | Apparatus and method for detecting road white line for automotive vehicle |
| CN101567086A (en) * | 2009-06-03 | 2009-10-28 | 北京中星微电子有限公司 | Method of lane line detection and equipment thereof |
| CN101620732A (en) * | 2009-07-17 | 2010-01-06 | 南京航空航天大学 | Visual detection method of road driving line |
| CN102314599A (en) * | 2011-10-11 | 2012-01-11 | 东华大学 | Identification and deviation-detection method for lane |
| CN102436644A (en) * | 2011-11-02 | 2012-05-02 | 南京物联网研究院发展有限公司 | Unstructured road detection method based on adaptive edge registration |
| CN102521589A (en) * | 2011-11-18 | 2012-06-27 | 深圳市宝捷信科技有限公司 | Method and system for detecting lane marked lines |
| CN102609938A (en) * | 2012-01-16 | 2012-07-25 | 浙江大学 | Method for detecting vanishing points of road based on single image |
| CN102682292A (en) * | 2012-05-10 | 2012-09-19 | 清华大学 | Method based on monocular vision for detecting and roughly positioning edge of road |
| CN103400150A (en) * | 2013-08-14 | 2013-11-20 | 浙江大学 | Method and device for road edge recognition based on mobile platform |
-
2013
- 2013-12-20 CN CN201310711070.3A patent/CN103714538B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020061123A1 (en) * | 2000-11-22 | 2002-05-23 | Nissan Motor Co., Ltd. | Apparatus and method for detecting road white line for automotive vehicle |
| CN101567086A (en) * | 2009-06-03 | 2009-10-28 | 北京中星微电子有限公司 | Method of lane line detection and equipment thereof |
| CN101620732A (en) * | 2009-07-17 | 2010-01-06 | 南京航空航天大学 | Visual detection method of road driving line |
| CN102314599A (en) * | 2011-10-11 | 2012-01-11 | 东华大学 | Identification and deviation-detection method for lane |
| CN102436644A (en) * | 2011-11-02 | 2012-05-02 | 南京物联网研究院发展有限公司 | Unstructured road detection method based on adaptive edge registration |
| CN102521589A (en) * | 2011-11-18 | 2012-06-27 | 深圳市宝捷信科技有限公司 | Method and system for detecting lane marked lines |
| CN102609938A (en) * | 2012-01-16 | 2012-07-25 | 浙江大学 | Method for detecting vanishing points of road based on single image |
| CN102682292A (en) * | 2012-05-10 | 2012-09-19 | 清华大学 | Method based on monocular vision for detecting and roughly positioning edge of road |
| CN103400150A (en) * | 2013-08-14 | 2013-11-20 | 浙江大学 | Method and device for road edge recognition based on mobile platform |
Cited By (27)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103927526B (en) * | 2014-04-30 | 2017-02-15 | 长安大学 | Vehicle detecting method based on Gauss difference multi-scale edge fusion |
| CN103927526A (en) * | 2014-04-30 | 2014-07-16 | 长安大学 | Vehicle detecting method based on Gauss difference multi-scale edge fusion |
| US10984261B2 (en) | 2014-05-14 | 2021-04-20 | Mobileye Vision Technologies Ltd. | Systems and methods for curb detection and pedestrian hazard assessment |
| CN106663193B (en) * | 2014-05-14 | 2020-12-25 | 御眼视觉技术有限公司 | System and method for curb detection and pedestrian hazard assessment |
| CN106663193A (en) * | 2014-05-14 | 2017-05-10 | 御眼视觉技术有限公司 | Systems and methods for curb detection and pedestrian hazard assessment |
| US10685246B2 (en) | 2014-05-14 | 2020-06-16 | Mobileye Vision Technologies Ltd. | Systems and methods for curb detection and pedestrian hazard assessment |
| CN105224908A (en) * | 2014-07-01 | 2016-01-06 | 北京四维图新科技股份有限公司 | A kind of roadmarking acquisition method based on orthogonal projection and device |
| CN106017313A (en) * | 2015-03-31 | 2016-10-12 | 株式会社三丰 | Edge detection bias correction value calculation method, edge detection bias correction method, and equipment |
| CN107392080A (en) * | 2016-05-17 | 2017-11-24 | 纬创资通股份有限公司 | Image evaluation method and electronic device thereof |
| CN106546185A (en) * | 2016-10-18 | 2017-03-29 | 福州觉感视觉软件科技有限公司 | A kind of profile quality determining method based on Machine Vision Detection |
| CN106546185B (en) * | 2016-10-18 | 2019-09-20 | 福州觉感视觉软件科技有限公司 | A kind of profile quality determining method based on Machine Vision Detection |
| CN106491042A (en) * | 2016-11-14 | 2017-03-15 | 浙江工业大学 | Processing method is moved down at a kind of swing type window wiping robot glass edge |
| CN106491040B (en) * | 2016-11-14 | 2019-04-09 | 浙江工业大学 | A kind of swing type window wiping robot glass edge touching detection processing method |
| CN106491040A (en) * | 2016-11-14 | 2017-03-15 | 浙江工业大学 | A kind of swing type window wiping robot glass edge touches detection process method |
| CN106491039A (en) * | 2016-11-14 | 2017-03-15 | 浙江工业大学 | A kind of swing type window wiping robot glass edge detection method |
| CN106491039B (en) * | 2016-11-14 | 2019-01-08 | 浙江工业大学 | A kind of swing type window wiping robot glass edge detection method |
| CN106491042B (en) * | 2016-11-14 | 2019-01-08 | 浙江工业大学 | Processing method is moved down at a kind of swing type window wiping robot glass edge |
| CN106650726A (en) * | 2016-12-05 | 2017-05-10 | 渤海大学 | License plate recognition method |
| CN107133625A (en) * | 2017-03-31 | 2017-09-05 | 纵目科技（上海）股份有限公司 | The extracting method and system, car-mounted terminal of a kind of yellow lane line |
| CN107133625B (en) * | 2017-03-31 | 2020-05-22 | 纵目科技（上海）股份有限公司 | Yellow lane line extraction method and system, and vehicle-mounted terminal |
| CN107194940A (en) * | 2017-05-23 | 2017-09-22 | 北京计算机技术及应用研究所 | A kind of coloured image contour extraction method based on color space and line segment |
| CN107862699B (en) * | 2017-09-22 | 2020-11-10 | 中国科学院深圳先进技术研究院 | Bone edge extraction method, device, equipment and storage medium of bone CT image |
| CN107862699A (en) * | 2017-09-22 | 2018-03-30 | 中国科学院深圳先进技术研究院 | Bone edges extracting method, device, equipment and the storage medium of Bone CT image |
| CN108416306B (en) * | 2018-03-12 | 2020-12-25 | 海信集团有限公司 | Continuous obstacle detection method, device, equipment and storage medium |
| CN108416306A (en) * | 2018-03-12 | 2018-08-17 | 海信集团有限公司 | Continuous type obstacle detection method, device, equipment and storage medium |
| CN109740469A (en) * | 2018-12-24 | 2019-05-10 | 百度在线网络技术（北京）有限公司 | Method for detecting lane lines, device, computer equipment and storage medium |
| WO2023124221A1 (en) * | 2021-12-31 | 2023-07-06 | 中国第一汽车股份有限公司 | Area edge detection method and apparatus for vehicle drivable area, and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103714538B (en) | 2016-12-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103714538A (en) | Road edge detection method, device and vehicle | |
| US20210208587A1 (en) | All weather autonomously driven vehicles | |
| Yang et al. | 3D local feature BKD to extract road information from mobile laser scanning point clouds | |
| CN105922991B (en) | Based on the lane departure warning method and system for generating virtual lane line | |
| CN102682292B (en) | Method based on monocular vision for detecting and roughly positioning edge of road | |
| CN105700532B (en) | The Intelligent Mobile Robot navigator fix control method of view-based access control model | |
| Berriel et al. | Ego-lane analysis system (elas): Dataset and algorithms | |
| CN104008377A (en) | Ground traffic sign real-time detection and recognition method based on space-time correlation | |
| CN109085823A (en) | The inexpensive automatic tracking running method of view-based access control model under a kind of garden scene | |
| CN105260699A (en) | Lane line data processing method and lane line data processing device | |
| CN105654073B (en) | A kind of speed automatic control method of view-based access control model detection | |
| CN104899554A (en) | Vehicle ranging method based on monocular vision | |
| CN102982304B (en) | Utilize polarized light image to detect the method and system of vehicle location | |
| CN104700414A (en) | Rapid distance-measuring method for pedestrian on road ahead on the basis of on-board binocular camera | |
| CN110531376A (en) | Detection of obstacles and tracking for harbour automatic driving vehicle | |
| CN108280450A (en) | A kind of express highway pavement detection method based on lane line | |
| CN104657727A (en) | Lane line detection method | |
| CN202134079U (en) | Unmanned vehicle lane marker line identification and alarm device | |
| CN103577809A (en) | Ground traffic sign real-time detection method based on intelligent driving | |
| CN102201054A (en) | Method for detecting street lines based on robust statistics | |
| KR20110001427A (en) | High speed road lane detection method based on extraction of roi-lb | |
| Li et al. | Inverse perspective mapping based urban road markings detection | |
| CN110379168A (en) | A kind of vehicular traffic information acquisition method based on Mask R-CNN | |
| CN105426868A (en) | Lane detection method based on adaptive region of interest | |
| CN115774444B (en) | Path planning optimization method based on sparse navigation map |
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 |