CN109191531A - A kind of automatic outer ginseng scaling method of the rear in-vehicle camera based on lane detection - Google Patents
A kind of automatic outer ginseng scaling method of the rear in-vehicle camera based on lane detection Download PDFInfo
- Publication number
- CN109191531A CN109191531A CN201810851930.6A CN201810851930A CN109191531A CN 109191531 A CN109191531 A CN 109191531A CN 201810851930 A CN201810851930 A CN 201810851930A CN 109191531 A CN109191531 A CN 109191531A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- view
- parameter
- frame
- calculated
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/588—Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30248—Vehicle exterior or interior
- G06T2207/30252—Vehicle exterior; Vicinity of vehicle
- G06T2207/30256—Lane; Road marking
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
- Traffic Control Systems (AREA)
Abstract
The automatic outer ginseng scaling method of the invention discloses a kind of rear in-vehicle camera based on lane detection, the following steps are included: obtained from vehicle steering wheel angle information when the absolute value of steering wheel angle φ is less than 2 °, the interframe of φ changes less than 1 °, carries out proving operation；At least 2 frame images are shot, T is denoted as0Frame and T1Frame, wherein T0And T1The acquisition time of two frames is respectively indicated, then the time interval between two frames is (T1‑T0), facilitate agreement T for statement0<T0；The intersection point P for calculating L1F and L2F, is calculated Rx, Ry according to position of the P point in F figure, if the width of lane line is W, W ∈ Sw, tx is calculated according to parameter current, defines matrixVehicle velocity V parameter is obtained, matrix is definedCalculating SSD parameter, the Rx being calculated, Ry, Rz, tx, ty are just intended to the optimized parameter asked.
Description
Technical field
The present invention relates to the automatic of camera calibration field more particularly to a kind of rear in-vehicle camera based on lane detection
Outer ginseng scaling method.
Background technique
The calibration of video camera is essential technology in field of machine vision.Especially in the advanced driving auxiliary of automotive field
System, including vehicle detection, pedestrian detection, blind area detection, lane detection etc. require calibration and are just able to achieve.Calibration is general again
Internal reference calibration can be divided into and outer ginseng demarcates two classes.Outer ginseng calibration mainly relates to following parameter: the spatially position in 3 directions
Move freedom degree, tx, ty, tz, and freedom degree Rx, Ry, Rz around the rotation of 3 reference axis.
The in-vehicle camera calibration technique of mainstream is divided into two classes at present: one kind is the producing line calibration of depot, builds a calibration
Place, accurately graphing installs camera by precision instrument and makes phase by the position of limiter constraint vehicle on the ground
The installation error very little of machine.
This technology has a shortcoming to be: calibration site requirements is high, and calibration place construction cost is very high.
Another shortcoming of this technology is: the assembly precision of camera requires high, it is necessary to have high-precision measurement and peace
The standby auxiliary of installing.
Another shortcoming of this technology is: if emblem mark must be returned by being changed in user using the posture of middle camera
Fixed, at high cost, the time is long.
In addition one kind is demarcated under line, be normally applied with the shop 4s and repacking shop, around vehicle be laid with calibration cloth, use ruler
The mounting height of son measurement camera and the data of installation site.Then it is demarcated using special calibration tool.Advantage is:
1, stated accuracy is higher by 2, Calibration Field flexible 3, it is of less demanding to the installation progress of camera.4, camera posture can after changing
With artificial calibration again.
This technology has a shortcoming to be: calibration process technical requirements are higher to need professional
Another shortcoming of this technology is: calibration process complicated and time consumption is laborious.
Become skyline there are also calibration under a kind of simple line to demarcate, operation is very simple, exactly adjusts the pitching of camera
Angle rx screen online to be overlapped with skyline (line of demarcation of sky and ground).
This technology has a shortcoming to be: stated accuracy is low to cause functional performance to decline.
Summary of the invention
In view of the deficiencies of the prior art, the present invention intends to provide a kind of rear vehicle based on lane detection
The automatic outer ginseng scaling method of camera is carried, to improve the requirement that the convenience of calibration meets precision simultaneously.This method not to tz into
Rower is fixed, and index determines Rx, Ry, Rz, tx, ty.
A target being realized of the invention be to provide it is a kind of without specific calibration place or scaling board, demarcate cloth
Scaling method.
Another that be realized target of the invention is to provide a kind of calibration without special manual measurement camera installation site
Method.
Another that be realized target of the invention is to provide that a kind of computation complexity is lower can be real in embedded platform
The scaling method of Shi Yunhang.
To achieve the above object, the invention provides the following technical scheme:
A kind of automatic outer ginseng scaling method of the rear in-vehicle camera based on lane detection, comprising the following steps:
Step 1: starting calibration by key or in a manner of touching screen when user thinks that scene meets calibration request
Program；
Step 2: obtained from vehicle steering wheel angle information；
Step 3: the interframe of φ changes less than 1 °, and this state continues when the absolute value of steering wheel angle φ is less than 2 °
Think that vehicle is in when more than 30 frames and stablize driving status, proving operation can be carried out；
Step 4: shooting at least 2 frame images, are denoted as T0Frame and T1Frame, wherein T0And T1When respectively indicating the acquisition of two frames
Between, then the time interval between two frames is (T1-T0), facilitate agreement T for statement0<T1；
Step 5: setting in original image is a little indicated with Pr, a little indicated with Pv on virtual camera image, in world coordinate system
A little indicated with Pw；It is respectively front view, left view and right view, hereinafter referred to as F figure, L figure on 3 views that original image, which projects to,
Scheme with R；
Step 6: being applied on F figure but being not limited in step 1 method detection from the lane line L1F- of vehicle two sides from a vehicle left side
Side lane line, L2F- are from vehicle right-hand lane line；
Step 7: the intersection point P of L1F and L2F is calculated, if
P=(Xp,Yp),Ps1=(Xs1,Ys1),Pe1=(Xe1,Ye1),Ps2=(Xs2,Ys2),Pe2=(Xe2,Ye2)
Yp=k1Xp+b1
Whereinb1=Ys1-k1×Xs1,b2=Ys2-k2×Xs2；
Step 8: Rx, Ry are calculated according to position of the P point in F figure,
Internal reference optical center is CxF,CyFFocal length is fF, the focal length of virtual camera F is P in the coordinate of x-axisx, it is in the coordinate of y-axis
Py
Step 9: set the width of lane line as W, W ∈ Sw,SwIndicate that the width set of standard vehicle diatom is such as, but not limited to
Sw={ 2.5,3,3.5 }, SwThe number of middle element is Nw, taking Wi to calculate corresponding its meaning of ty is that camera heights hereafter use HcTable
Show；
Step 10: calculating tx according to parameter current, tx can be solved by the geometrical relationship of the following figure:
Step 11: Rz∈SrzSrzThe quantity for indicating element in the set set of Rz value is Nrz；
Step 12: defining matrix
Step 13: obtaining vehicle velocity V parameter, the displacement of vehicle is calculated according to speed
Δ Z=V (t1-t0)；
Step 14: defining matrix
Step 15: calculating SSD parameter, SSD is Sum of Squared Differences；
Step 16: repeating step 11 to 15, all N are traversedwA possible W value, each W value correspond to NrzIt is a can
The Rz value of energy, obtains altogether Nw×NrzA D value；
Step 17: from Nw×NrzTake one the smallest in a D value, W and Rz parameter corresponding to the minimum value is denoted as W '
With R 'z, by W ' and R 'zThe Rx being calculated, Ry, Rz, tx, ty are just intended to the optimized parameter asked；
Step 18: calibration is completed, parameter Rx, Ry, Rz, tx, ty obtained in step 10 seven are exported.
As a further solution of the present invention, in step 1, calibration needs, and vehicle is even in the straight road for have lane line
Fast traveling placed in the middle.
As a further solution of the present invention, in step 5, Pr is mapped to Pw, then Pw is mapped to Pv, by above
Two steps have just obtained the mapping relations from original image to virtual camera image.
As a further solution of the present invention, in step 6, the information of a lane line includes starting point Ps and terminal Pe.
As a further solution of the present invention, in step 9, lane line L1L, L1R are detected on L, R view；Pl(x0,
y0),Pl(x1,y1) respectively indicate the beginning endpoint and end of a period endpoint of lane line L1L
It is the projection plane that A-A is virtual camera, fiFor the focal length of virtual camera L
The camera heights calculated on --- --- L view
H can similarly be calculatedrSeveral camera heights on --- --- R view
As a further solution of the present invention, in step 10, Hc: to calculate gained in step 9；
fl: the focal length of left side virtual camera；
fr: the focal length of left side virtual camera；
Yld: the y-coordinate of the lane line of detection position in L view；
Yrd: the y-coordinate of the lane line of detection position in R view；
Cyl: the y location of left side virtual camera optical center on the image；
Cyr: the y location of left side virtual camera optical center on the image；
Step 9 and step 10 calculate gained as tx and ty corresponding to a W, each lane line width all corresponds to
This 3 parameters are referred to generally at lane line relevant parameter: Lpara together with W by this 2 parameters.
As a further solution of the present invention, in step 11, { -2.0, -1.5, -1, -0.5,0,0.5,1,1.5,2 },
An element is taken to be counted as Rzj plus Lpara, Rx, Ry, (tz is considered vehicle rear origin so tz=0) along with internal reference K is with regard to group
At the parameter of complete set；Using this set parameter by the lane line L1F endpoint detected in F view be P1, P2 indicate, L2F
Endpoint is that P3, P4 are mapped in TopView figure that constitute L1T endpoint be P1 ' P2 ', and L2T endpoint is P3 ' P4 '；Mapping process is as follows
If coordinate of the P1 point under F view are as follows:
(u, v) homogeneous form isIf its corresponding world3D coordinate is that (X, Y, Z) homogeneous form is
If mapping point of the x in TopView figure
Known x is by formula (2) (3)It can be calculated X (the mapping matrix matrix that P is virtual camera F), acquire X
Afterwards by formula (1) (3) x′=PtX can be calculated P1 '；
P2 ', P3 ', P4 ' can similarly be calculated.
As a further solution of the present invention, in step 12, matrix is definedMethod it is as follows:
Rect1 is enabled to indicate a rectangular area on TopView view, diagonal line coordinates is (P1x-20, P1y), (P2x+
20,P2y)；
Rect2 is enabled to indicate a rectangular area on TopView view, diagonal line coordinates is (P1x-20, P1y), (P2x+
20,P2y)；
The image that the region Rect1 is extracted on the Top view of T0 frame, uses matrixIt indicates,
The image that the region Rect2 is extracted on the Top view of T0 frame, uses matrixIt indicates.
As a further solution of the present invention, the definition matrixMethod it is as follows:
A displacement has occurred in T1 frame camera compared with T0 frame, and motion vector is
It substitutes into formula (3), calculates separately to obtain T0、T1The TopView view homography matrix of frame
Pt0,Pt1；
P1 on the TopView view of T1 frame, P2, P3, the subpoint of P4 can be obtained by the process in step 11；
The image for repeating the procedure extraction T1 frame of step 12, obtains matrix
As a further solution of the present invention, the method for calculating SSD parameter is as follows:
It calculatesWithSSD be D1, calculateWithSSD be D2；
Calculate summation D=D1+D2.
As a further solution of the present invention, a kind of method calculating SSD is as follows:
R indicates the set of the position at image array midpoint,
The gray value of I expression image.
In order to explain the structural features and functions of the invention more clearly, come with reference to the accompanying drawing with specific embodiment to this hair
It is bright to be described in detail.
Detailed description of the invention
Fig. 1 is camera model coordinate system schematic diagram.
Fig. 2 is the main view of camera model coordinate system in Fig. 1.
Fig. 3 is image coordinate system and plane coordinate system schematic diagram in the present invention.
Fig. 4 is that the rotation of coordinate system in the present invention translates schematic diagram.
Fig. 5 is the structural schematic diagram of camera installation site in the present invention.
Fig. 6 is the dead astern view of vehicle in the present invention.
Fig. 7 is the left view of vehicle in the present invention.
Fig. 8 is the right view of vehicle in the present invention.
Fig. 9 is the top view of vehicle in the present invention.
Figure 10 is lane line schematic diagram in the present invention.
Figure 11 is Pl perspective view on L, R view in the present invention.
Figure 12 is left view perspective view in the present invention.
Figure 13 is left view projection plane expanded view in the present invention.
Figure 14 is flow chart of the method for the present invention.
Specific embodiment
The following further describes the technical solution of the present invention in the following with reference to the drawings and specific embodiments.
Referring to Fig. 1-14, a kind of automatic outer ginseng scaling method of the rear in-vehicle camera based on lane detection, including with
Lower step:
1, when user thinks that scene meets calibration request, start calibrating procedure by key or in a manner of touching screen.
Note: calibration needs, this vehicle traveling at the uniform velocity placed in the middle in the straight road for have lane line.
2, obtained from vehicle steering wheel angle information.
3, when the absolute value of steering wheel angle φ is less than 2 °, the interframe of φ changes less than 1 °, and this state continues 30 frames
Think that vehicle is in when above and stablize driving status, proving operation can be carried out.
4, at least 2 frame images are shot, T is denoted as0Frame and T1Frame, wherein T0And T1Respectively indicate the acquisition time of two frames, then two
Time interval between frame is (T1-T0).Facilitate agreement T for statement0<T1。
5, setting in original image is a little indicated with Pr, is a little indicated with Pv on virtual camera image, in world coordinate system a bit
It is indicated with Pw.Pr is mapped to Pw using formula (1), formula (2) is reapplied by Pw and is mapped to Pv, just obtained by above two step
From original image to the mapping relations of virtual camera image.Original image project on 3 views be respectively front view, left view and
Right view, hereinafter referred to as F figure, L figure and R figure.
6, applied on F figure but be not limited in [1] method detection from the lane line L1F- of vehicle two sides from vehicle left-hand lane line,
L2F- is from vehicle right-hand lane line.The information of one lane line includes starting point Ps and terminal Pe.
7, the intersection point P of L1F and L2F is calculated.If
P=(Xp,Yp),Ps1=(Xs1,Ys1),Pe1=(Xe1,Ye1),Ps2=(Xs2,Ys2),Pe2=(Xe2,Ye2)
Yp=k1Xp+b1
Whereinb1=Ys1-k1×Xs1,b2=Ys2-k2×Xs2
8, Rx, Ry are calculated according to position of the P point in F figure.
Internal reference optical center is CxF,CyFFocal length is fF(focal length of virtual camera F) is P in the coordinate of x-axisx, in the coordinate of y-axis
For Py
9, the width of lane line is set as W, W ∈ Sw,SwIndicate that the width set of standard vehicle diatom is such as, but not limited to Sw=
{ 2.5,3,3.5 }, SwThe number of middle element is Nw, taking Wi to calculate corresponding its meaning of ty is that camera heights hereafter use HcIt indicates.
Lane line L1L, L1R are detected on L, R view.Pl(x0,y0),Pl(x1,y1) respectively indicate opening for lane line L1L
Beginning point and end of a period endpoint
A-A is the projection plane of virtual camera, flFor the focal length of virtual camera L
The camera heights calculated on --- --- L view
H can similarly be calculatedrSeveral camera heights on --- --- R view
10, tx is calculated according to parameter current, tx can be solved by the geometrical relationship of the following figure.
Referring to shown in Figure 12,13, A-A is image projection plane in figure, and A-A plane is amplified.
Gained is calculated in Hc:9
fl: the focal length of left side virtual camera
fr: the focal length of left side virtual camera
Yld: the y-coordinate of the lane line of detection position in L view
Yrd: the y-coordinate of the lane line of detection position in R view
Cyl: the y location of left side virtual camera optical center on the image
Cyr: the y location of left side virtual camera optical center on the image
9-10 calculates gained as tx and ty corresponding to a W, each lane line width all corresponds to this 2 ginsengs
This 3 parameters are referred to generally at lane line relevant parameter: Lpara together with W by number.
11、Rz∈SrzSrzThe quantity for indicating element in the set set of Rz value is Nrz, such as -2.0, -1.5, -1, -
}, 0.5,0,0.5,1,1.5,2 take an element be counted as Rzj plus Lpara, Rx, Ry, (tz is considered vehicle rear origin so tz
=0) along with internal reference K just constitutes the parameter of complete set.The lane line L1F that will be detected in F view using this set parameter
Endpoint is P1, and P2 is indicated, L2F endpoint is that P3, P4 are mapped in TopView figure that constitute L1T endpoint be P1 ' P2 ', and L2T endpoint is
P3'P4'.Mapping process such as divides into coordinate of the P1 point under F view.
(u, v) homogeneous form isIf its corresponding world3D coordinate is that (X, Y, Z) homogeneous form isIf mapping point of the x in TopView figure
Known x is by formula (2) (3)It can be calculated X (the mapping matrix matrix that P is virtual camera F), acquire X
Afterwards by formula (1) (3) x '=PtX can be calculated P1 '
P2 ' P3 ' P4 ' can similarly be calculated.
12, matrix is definedMethod is as follows:
Rect1 is enabled to indicate a rectangular area on TopView view, diagonal line coordinates is (P1x-20, P1y), (P2x+
20, P2y),
Rect2 is enabled to indicate a rectangular area on TopView view, diagonal line coordinates is (P1x-20, P1y), (P2x+
20,P2y)
The image that the region Rect1 is extracted on the Top view of T0 frame, uses matrixIt indicates
The image that the region Rect2 is extracted on the Top view of T0 frame, uses matrixIt indicates.
13, vehicle velocity V parameter is obtained, the displacement of vehicle is calculated according to speed
Δ Z=V (t1-t0)。
14, matrix is definedMethod is as follows:
A displacement has occurred in T1 frame camera compared with T0 frame, and motion vector is
It substitutes into formula (3), calculates separately to obtain T0、T1The TopView view homography matrix of frame
Pt0,Pt1。
P1 on the TopView view of T1 frame, P2, P3, the subpoint of P4 can be obtained by the process in step 11.
The image for repeating the procedure extraction T1 frame of step 12, obtains matrix
15, SSD parameter (Sum of Squared Differences) is calculated, the method is as follows:
It calculatesWithSSD be D1, calculateWithSSD be D2.
Calculate summation D=D1+D2.
Preferably, a kind of method calculating SSD is as follows:
R indicates the set of the position at image array midpoint.
The gray value of I expression image.
16, step (11)-(15) are repeated and traverses all NwA possible W value, each W value correspond to NrzA possible Rz
Value, obtains altogether Nw×NrzA D value.
17, from Nw×NrzTake one the smallest in a D value, W and Rz parameter corresponding to the minimum value is denoted as W ' and R 'z,
By W ' and R 'zThe Rx being calculated, Ry, Rz, tx, ty are just intended to the optimized parameter asked.
18, calibration is completed, and exports parameter Rx, Ry, Rz, tx, ty obtained in step 17.
The technical principle of the invention is described above in combination with a specific embodiment, is only the preferred embodiment of the present invention.This
The protection scope of invention is not limited merely to above-described embodiment, and all technical solutions belonged under thinking of the present invention belong to the present invention
Protection scope.Those skilled in the art, which does not need to pay for creative labor, can associate other specific realities of the invention
Mode is applied, these modes will fall within the scope of protection of the present invention.
Claims (11)
1. a kind of automatic outer ginseng scaling method of rear in-vehicle camera based on lane detection, which is characterized in that including following
Step:
Step 1: starting calibrating procedure by key or in a manner of touching screen when user thinks that scene meets calibration request；
Step 2: obtained from vehicle steering wheel angle information；
Step 3: the interframe of φ changes less than 1 °, and this state continues 30 frames when the absolute value of steering wheel angle φ is less than 2 °
Think that vehicle is in when above and stablize driving status, carries out proving operation；
Step 4: shooting at least 2 frame images, are denoted as T0Frame and T1Frame, wherein T0And T1Respectively indicate the acquisition time of two frames, then two
Time interval between frame is (T1-T0), facilitate agreement T for statement0< T1；
Step 5: setting in original image is a little indicated with Pr, a little indicated with Pv on virtual camera image, in world coordinate system a bit
It is indicated with Pw；It is respectively front view, left view and right view, hereinafter referred to as F figure, L figure and R on 3 views that original image, which projects to,
Figure；
Step 6: applied on F figure but be not limited to the detection of method in step 1 from the lane line L1F- of vehicle two sides the vehicle from the left of vehicle
Diatom, L2F- are from vehicle right-hand lane line；
Step 7: the intersection point P of L1F and L2F is calculated, if
P=(Xp, Yp), Ps1=(Xs1, Ys1), Pe1=(Xe1, Ye1), Ps2=(Xs2, Ys2), Pe2=(Xe2, Ye2)
Yp=k1Xp+b1
Whereinb1=Ys1-k1×Xs1, b2=Ys2-k2×Xs2；
Step 8: Rx, Ry are calculated according to position of the P point in F figure,
Internal reference optical center is CxF, CyFFocal length is fF, the focal length of virtual camera F is P in the coordinate of x-axisx, it is P in the coordinate of y-axisy
Step 9: set the width of lane line as W, W ∈ Sw, SwIndicate that the width set of standard vehicle diatom is such as, but not limited to Sw=
{ 2.5,3,3.5 }, SwThe number of middle element is Nw, taking Wi to calculate corresponding its meaning of ty is that camera heights hereafter use HcIt indicates；
Step 10: calculating tx according to parameter current, tx is solved by the geometrical relationship of the following figure:
Step 11: Rz∈Srz SrzThe quantity for indicating element in the set set of Rz value is Nrz；
Step 12: defining matrix
Step 13: obtaining vehicle velocity V parameter, the displacement of vehicle is calculated according to speed
Δ Z=V (t1-t0)；
Step 14: defining matrix
Step 15: calculating SSD parameter, SSD is Sum of Squared Differences；
Step 16: repeating step 11 to 15, all N are traversedwA possible W value, each W value correspond to NrzIt is a possible
Rz value, obtains altogether Nw×NrzA D value；
Step 17: from Nw×NrzTake one the smallest in a D value, W and Rz parameter corresponding to the minimum value is denoted as W ' and R 'z,
By W ' and R 'zThe Rx being calculated, Ry, Rz, tx, ty are just intended to the optimized parameter asked；
Step 18: calibration is completed, parameter Rx, Ry, Rz, tx, ty obtained in step 10 seven are exported.
2. a kind of automatic outer ginseng scaling method of rear in-vehicle camera based on lane detection according to claim 1,
It is characterized in that, calibration needs in step 1, vehicle traveling at the uniform velocity placed in the middle in the straight road for have lane line.
3. a kind of automatic outer ginseng scaling method of rear in-vehicle camera based on lane detection according to claim 2,
It is characterized in that, Pr is mapped to Pw, then Pw is mapped to Pv in step 5, just obtained by above two step from original image
To the mapping relations of virtual camera image.
4. a kind of automatic outer ginseng scaling method of rear in-vehicle camera based on lane detection according to claim 3,
It is characterized in that, the information of a lane line includes starting point Ps and terminal Pe in step 6.
5. a kind of automatic outer ginseng scaling method of rear in-vehicle camera based on lane detection according to claim 4,
It is characterized in that, detecting lane line L1L, L1R on L, R view in step 9；Pl(x0, y0), Pl (x1, y1) respectively indicate vehicle
The beginning endpoint and end of a period endpoint of diatom L1L
It is the projection plane that A-A is virtual camera, flFor the focal length of virtual camera L
The camera heights calculated on --- --- L view
H can similarly be calculatedrSeveral camera heights on --- --- R view
6. a kind of automatic outer ginseng scaling method of rear in-vehicle camera based on lane detection according to claim 5,
It is characterized in that, in step 10, Hc: to calculate gained in step 9；
fl: the focal length of left side virtual camera；
fr: the focal length of left side virtual camera；
Yld: the y-coordinate of the lane line of detection position in L view；
Yrd: the y-coordinate of the lane line of detection position in R view；
Cyl: the y location of left side virtual camera optical center on the image；
Cyr: the y location of left side virtual camera optical center on the image；
Step 9 and step 10 calculate gained as tx and ty corresponding to a W, each lane line width all correspond to this 2
This 3 parameters are referred to generally at lane line relevant parameter: Lpara together with W by a parameter.
7. a kind of automatic outer ginseng scaling method of rear in-vehicle camera based on lane detection according to claim 6,
It is characterized in that, in step 11, { -2.0, -1.5, -1, -0.5,0,0.5,1,1.5,2 } takes an element to be counted as Rzj and adds
Lpara, Rx, Ry, (tz is considered vehicle rear origin so tz=0) add internal reference K and just constitute the parameter of complete set；Make
It by the lane line L1F endpoint detected in F view is P1 with this set parameter, P2 is indicated, L2F endpoint is that P3, P4 are mapped to
It is P1 ' P2 ' that L1T endpoint is constituted in TopView figure, and L2T endpoint is P3 ' P4 '；Mapping process such as divides into P1 point under F view
Coordinate are as follows:
(u, v) homogeneous form isIf its corresponding world3D coordinate is that (X, Y, Z) homogeneous form isIf x
Mapping point in TopView figure
Known x is by formula (2) (3)Can be calculated X (the mapping matrix matrix that P is virtual camera F), acquire after X by
Formula (1) (3) x '=PtX can be calculated P1 '；
P2 ', P3 ', P4 ' can similarly be calculated.
8. a kind of automatic outer ginseng scaling method of rear in-vehicle camera based on lane detection according to claim 7,
It is characterized in that, defining matrix in step 12Method it is as follows:
Enable Rectl indicate TopView view on a rectangular area, diagonal line coordinates be (P1x-20, P1y), (P2x+20,
P2y)；
Enable Rect2 indicate TopView view on a rectangular area, diagonal line coordinates be (P1x-20, P1y), (P2x+20,
P2y)；
The image that the region Rectl is extracted on the Top view of T0 frame, uses matrixIt indicates,
The image that the region Rect2 is extracted on the Top view of T0 frame, uses matrixIt indicates.
9. a kind of automatic outer ginseng scaling method of rear in-vehicle camera based on lane detection according to claim 8,
It is characterized in that, the definition matrixMethod it is as follows:
A displacement has occurred in T1 frame camera compared with T0 frame, and motion vector is
It substitutes into formula (3), calculates separately to obtain T0、T1The TopView view homography matrix P of framet0, Pt1；
P1 on the TopView view of T1 frame, P2, P3, the subpoint of P4 can be obtained by the process in step 11；
The image for repeating the procedure extraction T1 frame of step 12, obtains matrix
10. a kind of automatic outer ginseng scaling method of rear in-vehicle camera based on lane detection according to claim 9,
It is characterized in that, the method for calculating SSD parameter is as follows:
It calculatesWithSSD be D1, calculateWithSSD be D2；
Calculate summation D=D1+D2.
11. a kind of automatic outer ginseng calibration side of rear in-vehicle camera based on lane detection according to claim 10
Method, which is characterized in that a kind of method for calculating SSD is as follows:
R indicates the set of the position at image array midpoint,
The gray value of I expression image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810851930.6A CN109191531A (en) | 2018-07-30 | 2018-07-30 | A kind of automatic outer ginseng scaling method of the rear in-vehicle camera based on lane detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810851930.6A CN109191531A (en) | 2018-07-30 | 2018-07-30 | A kind of automatic outer ginseng scaling method of the rear in-vehicle camera based on lane detection |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109191531A true CN109191531A (en) | 2019-01-11 |
Family
ID=64937859
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810851930.6A Pending CN109191531A (en) | 2018-07-30 | 2018-07-30 | A kind of automatic outer ginseng scaling method of the rear in-vehicle camera based on lane detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109191531A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109859278A (en) * | 2019-01-24 | 2019-06-07 | 惠州市德赛西威汽车电子股份有限公司 | The scaling method and calibration system joined outside in-vehicle camera system camera |
| CN110413942A (en) * | 2019-06-04 | 2019-11-05 | 联创汽车电子有限公司 | Lane line equation screening technique and its screening module |
| CN110466453A (en) * | 2019-08-28 | 2019-11-19 | 安徽江淮汽车集团股份有限公司 | Blind monitoring system lane width thresholding value adjustment method |
| CN111627066A (en) * | 2019-02-27 | 2020-09-04 | 南京地平线机器人技术有限公司 | Method and device for adjusting external parameters of camera |
| CN112509054A (en) * | 2020-07-20 | 2021-03-16 | 北京智行者科技有限公司 | Dynamic calibration method for external parameters of camera |
| CN112862899A (en) * | 2021-02-07 | 2021-05-28 | 黑芝麻智能科技(重庆)有限公司 | External parameter calibration method, device and system for image acquisition equipment |
| CN114633692A (en) * | 2022-03-14 | 2022-06-17 | 深圳市艾为智能有限公司 | Application method of eccentric lens in CMS system |
| WO2022134518A1 (en) * | 2020-12-22 | 2022-06-30 | 上海商汤临港智能科技有限公司 | Method and apparatus for calibrating camera device, and electronic device and storage medium |
-
2018
- 2018-07-30 CN CN201810851930.6A patent/CN109191531A/en active Pending
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109859278B (en) * | 2019-01-24 | 2023-09-01 | 惠州市德赛西威汽车电子股份有限公司 | Calibration method and calibration system for camera external parameters of vehicle-mounted camera system |
| CN109859278A (en) * | 2019-01-24 | 2019-06-07 | 惠州市德赛西威汽车电子股份有限公司 | The scaling method and calibration system joined outside in-vehicle camera system camera |
| CN111627066A (en) * | 2019-02-27 | 2020-09-04 | 南京地平线机器人技术有限公司 | Method and device for adjusting external parameters of camera |
| CN110413942B (en) * | 2019-06-04 | 2023-08-08 | 上海汽车工业(集团)总公司 | Lane line equation screening method and screening module thereof |
| CN110413942A (en) * | 2019-06-04 | 2019-11-05 | 联创汽车电子有限公司 | Lane line equation screening technique and its screening module |
| CN110466453A (en) * | 2019-08-28 | 2019-11-19 | 安徽江淮汽车集团股份有限公司 | Blind monitoring system lane width thresholding value adjustment method |
| CN112509054A (en) * | 2020-07-20 | 2021-03-16 | 北京智行者科技有限公司 | Dynamic calibration method for external parameters of camera |
| CN112509054B (en) * | 2020-07-20 | 2024-05-17 | 重庆兰德适普信息科技有限公司 | Camera external parameter dynamic calibration method |
| WO2022134518A1 (en) * | 2020-12-22 | 2022-06-30 | 上海商汤临港智能科技有限公司 | Method and apparatus for calibrating camera device, and electronic device and storage medium |
| CN112862899A (en) * | 2021-02-07 | 2021-05-28 | 黑芝麻智能科技(重庆)有限公司 | External parameter calibration method, device and system for image acquisition equipment |
| CN112862899B (en) * | 2021-02-07 | 2023-02-28 | 黑芝麻智能科技(重庆)有限公司 | External parameter calibration method, device and system for image acquisition equipment |
| CN114633692A (en) * | 2022-03-14 | 2022-06-17 | 深圳市艾为智能有限公司 | Application method of eccentric lens in CMS system |
| CN114633692B (en) * | 2022-03-14 | 2023-10-03 | 深圳市艾为智能有限公司 | Application method of eccentric lens in CMS system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109191531A (en) | A kind of automatic outer ginseng scaling method of the rear in-vehicle camera based on lane detection | |
| CN110567469B (en) | Visual positioning method and device, electronic equipment and system | |
| CN103578109B (en) | A kind of CCTV camera distance-finding method and device | |
| CN103871071B (en) | Join scaling method outside a kind of camera for panoramic parking system | |
| CN108805934A (en) | A kind of method for calibrating external parameters and device of vehicle-mounted vidicon | |
| CN106256606A (en) | A kind of lane departure warning method based on vehicle-mounted binocular camera | |
| CN110203210A (en) | A kind of lane departure warning method, terminal device and storage medium | |
| CN105091750A (en) | Projector calibration method based on double four-step phase shift | |
| CN112529966B (en) | On-line calibration method of vehicle-mounted looking-around system and vehicle-mounted looking-around system thereof | |
| CN103294886A (en) | System for reproducing virtual objects | |
| CN107492123B (en) | Road monitoring camera self-calibration method using road surface information | |
| CN103679729A (en) | Full-automatic camera parameter calibration method based on colored calibration board | |
| EP3816663A2 (en) | Method, device, equipment, and storage medium for determining sensor solution | |
| CN112967344B (en) | Method, device, storage medium and program product for calibrating camera external parameters | |
| EP3998580B1 (en) | Camera calibration method and apparatus, electronic device, storage medium, program product, and road side device | |
| JP2013037394A (en) | Vehicle detection device | |
| CN105551020A (en) | Method and device for detecting dimensions of target object | |
| CN107145825A (en) | Ground level fitting, camera calibration method and system, car-mounted terminal | |
| CN108121941A (en) | A kind of object speed calculation method based on monitoring device | |
| CN111435540A (en) | Annular view splicing method of vehicle-mounted annular view system | |
| CN100416466C (en) | Single-eye vision semi-matter simulating system and method | |
| CN110415299B (en) | Vehicle position estimation method based on set guideboard under motion constraint | |
| CN112419423A (en) | Calibration method, calibration device, electronic equipment and storage medium | |
| JP2013024712A (en) | Method and system for calibrating multiple camera | |
| CN103139532A (en) | Vehicle periphery monitor |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |