CN103600707B - A kind of parking position detection device and method of Intelligent parking system - Google Patents

A kind of parking position detection device and method of Intelligent parking system Download PDF

Info

Publication number
CN103600707B
CN103600707B CN201310545995.5A CN201310545995A CN103600707B CN 103600707 B CN103600707 B CN 103600707B CN 201310545995 A CN201310545995 A CN 201310545995A CN 103600707 B CN103600707 B CN 103600707B
Authority
CN
China
Prior art keywords
module
parking
parking position
obstacle
position
Prior art date
Application number
CN201310545995.5A
Other languages
Chinese (zh)
Other versions
CN103600707A (en
Inventor
陈慧
范梦阳
Original Assignee
同济大学
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 同济大学 filed Critical 同济大学
Priority to CN201310545995.5A priority Critical patent/CN103600707B/en
Publication of CN103600707A publication Critical patent/CN103600707A/en
Application granted granted Critical
Publication of CN103600707B publication Critical patent/CN103600707B/en

Links

Abstract

The present invention discloses the level detecting apparatus of parking of a kind of Intelligent parking system, including sensor unit: four wide-angle cameras and ultrasonic distance-measuring sensor;Signal processing unit: panoramic view generation module, parking position detection module;Human-machine interface unit: output module is in order to show that panoramic view and the warehouse compartment detected, input module are in order to accept the various instructions of driver's input.Wherein, parking position detection module includes warehouse compartment line detection sub-module, obstacle car detection sub-module and parking position output sub-module.Invention additionally discloses the parking position detection method of a kind of Intelligent parking system.The present invention uses low cost sensor plan, the image information that the distance value obtained by merging ultrasonic distance-measuring sensor arrives with camera collection, improve parking position detection accuracy and precision, and ground exist warehouse compartment line and exist obstacle car multiple in the case of parking position can effectively be detected.

Description

A kind of parking position detection device and method of Intelligent parking system

Technical field

The invention belongs to automobile technical field, relate to drive assist system, refer more particularly to the parking position detection of Intelligent parking system Device and method.

Background technology

Along with the increase of automobile pollution, parking stall growing tension, this problem of parking difficulty becomes more serious.For there is no warp For the new hand tested, vehicle safety ground is stopped into parking position extremely difficult;From the point of view of experienced driver, vehicle Stopping into narrow and small parking position is not an easy thing yet.Developing intellectual resource parking system, advantageously ensures that the peace of the process of parking Quan Xing, improves comfortableness and the convenience of process of parking simultaneously.

Intelligent parking system generally comprises following module: parking position detection module, path planning module, tracking module, hold Row module and human-machine interface module.Parking position detection module is the basis of Intelligent parking system, largely affects whole pool The performance of car system.

At present, the parking position detection method of Intelligent parking system mainly includes following three kinds: 1. utilize ultrasonic distance-measuring sensor to enter Row parking position detects.Owing to vehicle front or rear end exist fillet, limited to by ultrasonic distance-measuring sensor self character, front There is the phenomenon not receiving echo at the fillet of rear end, it is bigger than normal than actual that this will result in the parking position detected;Additionally, this side Method is not suitable for ground yet and there is warehouse compartment line but the situation that do not has front and back's obstacle car, such as parking lot etc..The auxiliary of parking of Volkswagen System utilizes the sonac being arranged on bumper both sides, is scanned vehicle both sides when vehicle travels, and then detection is parked Position.This system all has lift-launch in the vehicle such as sight, Magotan, CC of way under masses.2. utilize laser radar to carry out parking position inspection Survey, have an advantage in that accuracy of detection is high, but laser radar is with high costs.The patent of Honda Motor Co.'s application CN10187849 utilizes radar installations to send electromagnetic wave with predetermined time interval, detects electromagnetic wave according to the reception result of echo The pip of reflection on object, and judge to park sky according to the arrangement of the body dimensions data from car prestored and pip Between presence or absence.3. utilizing machine vision technique to carry out parking position detection, its difficult point is that image is vulnerable to the environment such as shade, illumination The impact of condition.Wang Xudong proposes in " designing with system based on the automatic parking technique study looked around " and becomes based on Radon The bright spot feature detection parking position changed and the method extracting empty parking position based on window of interest.Above scheme is respectively arranged with advantage, but Yet suffer from deficiency.

Summary of the invention

It is an object of the invention to provide the parking position detection device and method of a kind of Intelligent parking system, there is warehouse compartment line on ground With exist obstacle car multiple in the case of parking position can effectively be detected.

For reaching object above, solution of the present invention is:

A kind of level detecting apparatus of parking of Intelligent parking system, including sensor unit: the photographic head being positioned on automobile body and Distance measuring sensor;Signal processing unit: panoramic view generation module, parking position detection module;Man-machine interface: output module in order to Display panoramic view and the warehouse compartment detected, input module is in order to accept the various instructions of driver's input.

Described parking position detection module includes park bit line detection sub-module, obstacle car detection sub-module, parking position output sub-module; Bit line detection sub-module of parking utilize ground park bit line detection parking position;Obstacle car detection sub-module Use barriers car test is surveyed and is parked Position;Parking position output sub-module, according to bit line detection sub-module and the testing result of obstacle car detection sub-module of parking, finally determines Go out parking position.

Described sensor unit includes being positioned at vehicle front side, rear side, left side, four wide-angle cameras on right side and being positioned at vehicle two The ultrasonic distance-measuring sensor of side.

The installation site of described four photographic head should ensure that four camera collections to picture cover 360 degree of regions of vehicle's surroundings, And adjacent two camera collections to picture have overlapping region.

Described man-machine interface output module by the parking position Overlapping display that detects on panoramic view, if be detected that parking position not Correct or do not meet the wish of driver, driver can adjust position and (or) the direction of parking position by input module, After driver changes position and (or) the direction of parking position, new four angular coordinates of parking position are exported intelligent parking system The path planning module of system.

The method that the bit line detection sub-module of parking parked in level detecting apparatus of described Intelligent parking system determines parking position, bag Include following steps: (1) Image semantic classification;(2) from panoramic view, straight line is extracted;(3) straight line is carried out post processing, to retain Parking line segment corresponding to bit line in ground, deletes other mixed and disorderly line segments;(4) rectangle parking position is obtained.

The method that the obstacle car detection sub-module parked in level detecting apparatus of described Intelligent parking system determines parking position: according to Hold the position of profile to determine parking position before and after obstacle car.

First obstacle Herba Plantaginis end profile position in global coordinate system uses following methods to determine: (1) utilizes ultrasonic ranging Distance value that sensor obtains and vehicle posture information, simulate the straight line corresponding to obstacle car side profile;(2) when super During the distance value generation positive transition of sound ranging sensor, store the panoramic view in this moment;(3) in panoramic view, sense is set emerging Interest region (ROI), makes obstacle Herba Plantaginis end profile be positioned at area-of-interest;(4) to described area-of-interest image gray processing, Do rim detection, obtain the coordinate of obstacle Herba Plantaginis end profile protruding point (First Point);(5) by First Point to simulating Straight line make vertical line, determine obstacle Herba Plantaginis end profile position in global coordinate system;

If there are two obstacle cars, then second obstacle car rear end profile position in global coordinate system uses following methods to determine: (6) utilize distance value and vehicle posture information that ultrasonic distance-measuring sensor obtains, simulate corresponding to obstacle car side profile Straight line;(2) when the distance value of ultrasonic distance-measuring sensor occurs negative saltus step, the panoramic view in this moment is stored;(3) exist Panoramic view sets area-of-interest, makes obstacle car rear end profile be positioned at area-of-interest;(4) to described area-of-interest figure As gray processing, do rim detection, obtain the obstacle Herba Plantaginis end profile protruding point i.e. coordinate of rearmost point;(5) by rearmost point Make vertical line to the straight line simulated, determine obstacle car rear end profile position in global coordinate system.

Owing to have employed technique scheme, the method have the advantages that employing low cost sensor plan, by melting The image information that the distance value that conjunction ultrasonic distance-measuring sensor obtains arrives with camera collection, improves the accuracy of parking position detection And precision, and ground exist warehouse compartment line and exist obstacle car multiple in the case of parking position can effectively be detected.

Accompanying drawing explanation

Fig. 1 is the parking position detection apparatus function module diagram of Intelligent parking system of the present invention.

Fig. 2 is the flow chart of the parking position detection method of Intelligent parking system of the present invention.

Fig. 3 is sensor arrangement on vehicle in the embodiment of the present invention.

Fig. 4 is the panoramic view splicing generation in the embodiment of the present invention.

Fig. 5 is the algorithm flow chart of the bit line detection sub-module of parking of the present invention.

Fig. 6 is the flow chart of embodiment of the present invention cathetus post-processing approach.

Fig. 7 is the algorithm flow chart of obstacle car detection sub-module of the present invention.

Fig. 8 is global coordinate system and the schematic diagram of vehicle axis system in the present invention.

Fig. 9 is the flow chart that first obstacle Herba Plantaginis end profile of the present invention determines method.

Figure 10 is that in the embodiment of the present invention, first obstacle Herba Plantaginis end profile determines method schematic diagram.

Figure 11 is the parking position design sketch of Overlapping display in panoramic view that will detect in the embodiment of the present invention.

Detailed description of the invention

Below in conjunction with accompanying drawing illustrated embodiment, the present invention is further illustrated.

Refer to Fig. 1 and Fig. 2.

1, sensor unit:

Sensor unit includes: is positioned at vehicle front side, rear side, left side, four wide-angle cameras on right side and is positioned at vehicle both sides Ultrasonic distance-measuring sensor.Wherein, the installation site of four photographic head should ensure that four camera collections to picture cover 360 degree of regions of vehicle's surroundings, and adjacent two camera collections to picture have overlapping region.Ultrasonic distance-measuring sensor is permissible It is that left and right side respectively arranges one, it is also possible to it is multiple to be that left and right side is respectively arranged.In embodiment 1, front photographic head is arranged At vehicle front grid, left and right photographic head is arranged at the rearview mirror of left and right, and rear photographic head is arranged on vehicle back door;On a left side A ultrasonic distance-measuring sensor it is respectively mounted at right rear view mirror.Sensor arrangement on vehicle is as shown in Figure 3.

2, panoramic view generation module

(1) distortion is corrected

In order to obtain bigger field range, four photographic head being positioned at vehicle body surrounding in the present invention use flake wide-angle camera. The image that fish-eye camera collects also exists bigger distortion, it is necessary first to it is corrected distortion.

In example 2, the radial distortion of a consideration video camera and tangential distortion, the distortion of camera parameter obtained according to demarcation, Utilize following formula to four camera collections to image respectively be corrected distortion:

xcor=x+x(k1r2+k2r4+k3r6)+[2p1y+p2(r2+2x2)] formula (2-1)

ycor=y+y(k1r2+k2r4+k3r6)+[p1(r2+2y2)+2p2X] formula (2-2)

Wherein, (x y) is the original coordinates of a certain pixel;(xcor,ycor) be this pixel correcting distorted after coordinate;[k1,k2,k3] it is radial distortion parameter;[p1,p2] it is tangential distortion parameter.

(2) four width birds-eye views are generated

Four width images after correcting distortion are carried out Inverse projection respectively, is converted into the birds-eye view overlooking effect.

1) camera model

What camera imaging model described is that the imaging process of object, i.e. any point coordinate in three-dimensional world coordinate system are to being somebody's turn to do Mathe-matical map relation between the coordinate of some imaging.

1. world coordinates is tied to the transformation relation of camera coordinate system

X c Y c Z c 1 = R T 0 1 X w Y w Z w 1 Formula (2-3)

Wherein, R is the spin matrix of 3 × 3;T is the translation vector of 3 × 1; R T 0 1 For joining matrix, (X outside video cameraw,Yw,Zw) it is Certain some coordinate in world coordinate system in space;(Xc,Yc,Zc) it is this coordinate in camera coordinate system.

2. camera coordinates is tied to the transformation relation of image coordinate system

Z c x y 1 = f x 0 c x 0 0 f y c y 0 0 0 1 0 X c Y c Z c 1 = f x 0 c x 0 0 f y c y 0 0 0 1 0 R T 0 1 X w Y w Z w 1 Formula (2-4)

Wherein, f x 0 c x 0 0 f y c y 0 0 0 1 0 For video camera internal reference matrix;(x y) is this coordinate in image coordinate system.

2) inverse perspective projection transformation

From formula (2-4), according to certain some coordinate (X in three-dimensional world coordinate systemw,Yw,Zw) this point can be calculated Coordinate in image coordinate system (y), otherwise then can not by x.But it is if it is known that certain in certain some three-dimensional coordinate is one-dimensional, the most permissible According to the coordinate in this dot image coordinate system, (x y), calculates the another bidimensional of this three-dimensional coordinate.Inverse perspective mapping refers to: build Found the position corresponding relation of the point in image coordinate system and the point in known plane in three-dimensional world coordinate system.

Set up following coordinate system: choosing vehicle geometric center point and being projected in ground point straight down is zero Ow;YwAxle Being oriented parallel to vehicle rear axle direction, pointing to vehicle left side is just;XwIt is perpendicular to YwAxle, pointing to vehicle front is just;It is perpendicular to Ground is upwards ZwAxle positive direction.Using this coordinate system as world coordinate system.It is now assumed that Zw=0, i.e. suppose owning in image Point is all located on ground in three-dimensional world coordinate system, utilizes the internal reference matrix of four video cameras and outer ginseng matrix, to four shootings The image that machine collects carries out inverse perspective mapping respectively, obtains overlooking the birds-eye view of effect.

(3) splicing generates panoramic view

By inverse perspective mapping, obtaining four width and overlook the birds-eye view of effect, the birds-eye view that adjacent camera obtains has and partially overlaps Four width birds-eye views, by alignment overlapping region, can be spliced into panoramic view by region.

First, the field range of panoramic view is set.This has also determined that the zoom factor of birds-eye view;

Then, it is determined that piece.Choose four straight lines in four width birds-eye views overlapping region between any two as piece.

Finally, four width birds-eye views are cut out along the position of piece, spliced.

Fig. 4 is the panoramic view splicing generation in embodiment 2.

3, parking position detection module

Parking position detection module includes park bit line detection sub-module, obstacle car detection sub-module, parking position output sub-module.Pool Parking stall line detection sub-module utilize ground park bit line detection parking position;Parking position is surveyed in obstacle car detection sub-module Use barriers car test; Parking position output sub-module, according to bit line detection sub-module and the testing result of obstacle car detection sub-module of parking, finally determines pool Parking stall.

(1) park bit line detection sub-module

The algorithm flow chart of bit line detection sub-module of parking is as shown in Figure 5.

1) Image semantic classification

Panoramic view is carried out gray processing, cromogram is become gray-scale map.Gray-scale map is carried out medium filtering again, reduces in image Noise.

2) straight line is extracted

1. rim detection

In embodiment 3, use canny operator that gray-scale map is carried out rim detection.Canny edge detection operator uses two thresholds Value, if the gradient of a pixel is more than upper limit threshold, is then considered as edge pixel, if less than lower threshold, then thrown Abandon, if therebetween, only just can be accepted with when being connected higher than the pixel of upper limit threshold when it.

Otsu algorithm is used to calculate the high-low threshold value of Canny operator in embodiment 3.Otsu algorithm basic ideas are chosen Optimal threshold should make have best separation property between two classes obtained with this Threshold segmentation, and detailed process make use of the ash of image Degree rectangular histogram, so that the gray value variance of target and background is target to the maximum to determine the threshold value that image is split.Canny operator High threshold be:

T=ArgMax[w0(t)(u0(t)-u)2+w1(t)(u1(t)-u)2] formula (3-1)

Wherein, T is the high threshold of canny operator;T is the image segmentation threshold assumed;u0T () is the institute that brightness value is bigger than t There is the average brightness value of pixel;w0T () is the ratio that brightness value is more than shared by the pixel of t;u1T () is that brightness value is less than t The average brightness value of all pixels;w1T () is the ratio that brightness value is less than shared by the pixel of t;U is all pixels in image The average brightness value of point.

After being calculated high threshold T, the more conventional empirical equation using high threshold to be equal to Low threshold twice obtains Low threshold.

2. Hough transformation

Straight line on two dimensional surface can describe by equation below:

ρ=xcos (θ)+ysin (θ) formula (3-2)

Wherein, ρ is the initial point distance to straight line, and θ is the inclination angle of this straight line.

ρ and θ constitutes two-dimensional space H (ρ, θ), and any point in H (ρ, θ) all correspond to the straight line on a certain plane domain. Hough transform is exactly the point-line duality utilizing image space and Hough parameter space, the test problems in image space It is transformed into parameter space.By carrying out simple cumulative statistics in parameter space, then find counting at Hough parameter space The method detection of straight lines of device peak value.

3) straight line post processing

Owing to bit line of parking is with the presence of defect or ground shade etc., many mixed and disorderly line segments may be detected.It is thus desirable to utilize pool The constraints that parking stall line meets, processes the straight line detected, the line segment corresponding to retain warehouse compartment line, deletes other miscellaneous Random line segment.The constraints that bit line of parking meets includes: parking, bit line opposite side is parallel to each other, adjacent side is orthogonal, standard is parked The size of position, the live width etc. of bit line of parking, can choose one or several constraintss therein and process straight line.? In embodiment 3, the flow chart of straight line post-processing algorithm is as shown in Figure 6.Algorithm above provides a kind of method of straight line post processing, Different constraints can also be chosen in other embodiments.

(2) obstacle car detection sub-module

The algorithm flow chart of obstacle car detection sub-module is as shown in Figure 7.

1) position of first obstacle Herba Plantaginis end profile is calculated

Global coordinate system is defined as with starting point vehicle rear axle midpoint of parking for initial point OG;YGDirection of principal axis is along vehicle At the rear direction of principal axis of starting point of parking, pointing to vehicle left side is just;XGAxle is perpendicular to YGAxle, pointing to vehicle direction of advance is Just.Vehicle axis system is defined as with vehicle rear axle midpoint for initial point O;Y direction along the rear direction of principal axis of vehicle, Pointing to vehicle left side is just;X-axis is perpendicular to Y-axis, and pointing to vehicle direction of advance is just.Vehicle coordinate ties up to world coordinates System follows vehicle motion.Fig. 8 is the schematic diagram of the global coordinate system defined in the present invention and vehicle axis system.

Calculate the algorithm flow chart of first obstacle Herba Plantaginis end outline position as shown in Figure 9.

1. obstacle car side profile straight line

I calculates the coordinate in global coordinate system of the point on obstacle car side profile

Ultrasonic distance-measuring sensor can be that left and right side respectively arranges one, it is also possible to it is multiple to be that left and right side is respectively arranged.In reality Executing in example 4, vehicle left and right side respectively arranges a ultrasonic distance-measuring sensor.Forward travel, utilizes ultrasonic ranging to pass Sensor, obtains distance value between vehicle and obstacle car.There is not positive transition in the distance value returned when ultrasonic distance-measuring sensor Time, utilize the installation site parameter of distance value and ultrasonic distance-measuring sensor, obtain the point on side profile in vehicle axis system Coordinate.

X=XultFormula (3-1)

Y=dis+YultFormula (3-2)

Wherein, (X, Y) is the coordinate in vehicle axis system of the point on side profile;(Xult,Yult) it is that ultrasonic sensor exists The coordinate of installation site in vehicle axis system;The computational methods of dis are as follows: its value be left side ultrasonic distance-measuring sensor return away from Distance values, or the opposite number of the distance value of right-side ultrasonic-wave distance measuring sensor return.

Wherein, positive transition refers to: the increase of the absolute value of the dis of the current time absolute value relative to the dis of previous moment more than to Determine threshold value.

The vehicle posture information in recycling corresponding moment, the point being calculated on side profile coordinate in global coordinate system.

XG=(X+xv)cosθ-(Y+yv) sin θ formula (3-3)

YG=(X+xv)sinθ+(Y+yv) cos θ formula (3-4)

Wherein, (XG,YG) it is the coordinate in global coordinate system of the point on side profile;(xv,yv, θ) and it is the car in corresponding moment Posture information (xvAnd yvIt is respectively corresponding moment vehicle axis system initial point (i.e. vehicle rear axle midpoint) in global coordinate system X, y-coordinate;θ is the course angle of vehicle).

II simulates straight line with the point on side profile

In example 4, with method of least square, the point on side profile is done fitting a straight line.

Multiple ultrasonic distance-measuring sensor can also be respectively arranged in other embodiments at vehicle left and right side.Such as,

It is respectively arranged two ultrasonic distance-measuring sensors in arranged on left and right sides, then said method matching can be utilized to obtain two straight lines, This two straight lines equation in global coordinate system is respectively as follows: YG=k1XG+b1、YG=k2XG+b2.Take k=0.5* (k1+k2),b=0.5*(b1+b2), obtain straight line YG=kXG+ b, is the straight line that obstacle car side profile is corresponding.

2. obstacle car First Point

Adopting and define image coordinate system with the following method: initial point is defined on the left upper of image, level is to the right x-axis positive direction, It is y-axis positive direction straight down.

I sets ROI

When the distance value generation positive transition of ultrasonic distance-measuring sensor, store the panoramic view in this moment.Panoramic view is set ROI, Obstacle Herba Plantaginis end profile is made to be positioned at ROI region.

II gray processing, rim detection

First ROI image is converted to gray-scale map, and recycling canny edge detection operator carries out rim detection to gray-scale map.

III First Point coordinate in ROI image coordinate system

In ROI edge image, find the marginal point that y-coordinate value (in image coordinate system) is minimum, as obstacle Herba Plantaginis End profile protruding point (First Point).

IV First Point coordinate in global coordinate system

First, First Point coordinate in the image coordinate system of panoramic view is calculated:

XH=XR+XXFormula (3-5)

YH=YR+YYFormula (3-6)

Wherein, (XH,YH) it is First Point coordinate in the image coordinate system of panoramic view;(XR,YR) it is that First Point is schemed at ROI Coordinate in the image coordinate system of picture;(XX,YY) it is the upper left angle point of the ROI region coordinate in the image coordinate system of panoramic view.

Then, First Point coordinate in vehicle axis system is calculated:

XFV=(Ypix/2-YH)×(L/Ypix)+d formula (3-7)

YFV=(-XH+Xpix/2)×(L/Ypix) formula (3-8)

Wherein, (XFV,YFV) it is First Point coordinate in vehicle axis system, unit is centimetre;XpixSit at image for panoramic view Pixel number on mark system x direction, in example 4 its value 320;YpixFor panoramic view pixel on image coordinate system y direction Number, its value is 550 in the present embodiment;D is the geometric center distance to vehicle rear axle of vehicle, is 130(li in the present embodiment Rice);L is the actual distance value that scope that panoramic view can show in the longitudinal direction is corresponding, is 1550(centimetre in the present embodiment).

Finally, First Point coordinate in global coordinate system is calculated:

XF=(XFV+xv)cosθ-(YFV+yv) sin θ formula (3-9)

YF=(XFV+xv)sinθ+(YFV+yv) cos θ formula (3-10)

Wherein, (XF,YF) it is First Point coordinate in global coordinate system;(xv,yv, θ) and it is the distance value of ultrasonic sensor There is vehicle posture information (x during positive transitionvAnd yvIt is respectively corresponding moment vehicle axis system initial point (i.e. vehicle rear axle midpoint) X in global coordinate system, y-coordinate;θ is the course angle of vehicle).

3. vertical line is made to side profile straight line in First Point

Figure 10 is that in the present embodiment, first obstacle Herba Plantaginis end profile determines method schematic diagram.

2) second obstacle car is judged whether

After the distance value generation positive transition of ultrasonic distance-measuring sensor, vehicle continues to move forward, if moving forward to pre-spacing From, all there is not negative saltus step in the distance value of ultrasonic distance-measuring sensor, then is judged to there is not second obstacle car;If it is the most capable Sailing to preset distance, there is negative saltus step in the distance value of ultrasonic distance-measuring sensor, it is determined that for there is second obstacle car.Wherein, Negative saltus step refers to: the reduction amount of the absolute value of the dis of the current time absolute value relative to the dis of previous moment is more than given threshold value.dis Computational methods as follows: its value is the distance value that left side ultrasonic distance-measuring sensor returns, or right-side ultrasonic-wave distance measuring sensor returns The opposite number of the distance value returned.

3) rectangle parking position

If there is second obstacle car, then according to first obstacle Herba Plantaginis end profile and the position of second obstacle car rear end profile, really Make the rectangle parking position between two obstacle cars;If there is not second obstacle car, then according to first obstacle Herba Plantaginis end profile Or rear end profile determines rectangle parking position.

(3) parking position output sub-module

The testing result of park bit line detection sub-module and obstacle car detection sub-module is inputed to parking position output sub-module.Park The testing result of bit line detection sub-module and obstacle car detection sub-module includes following four situation, and parking position output sub-module is four The available parking position obtained in the case of Zhong is as shown in the table:

Wherein, √ represents that detecting that warehouse compartment, X represent is not detected by warehouse compartment.

For situation 1, if the parking position region of obstacle car detection sub-module output comprises bit line detection sub-module output of parking Parking position region, then export the parking position that bit line detection sub-module of parking detects;Otherwise, output obstacle car detection sub-module inspection The parking position measured.

For situation 3, the parking position that output obstacle car detection sub-module detects.

4, man-machine interface

Man-machine interface includes output module and input module.Output module is in order to show panoramic view and the warehouse compartment detected, Ke Yitong Cross touch screen, head up display (HUD) etc. to realize.Input module, in order to accept the various instructions of driver's input, can pass through In touch screen, phonetic order, steering indicating light, car, button etc. realize.

If the parking position that Overlapping display is on panoramic view is incorrect or does not meets the wish of driver, driver can pass through people The input module of machine interface adjusts position and (or) the direction of parking position.Position and (or) the side of parking position is changed driver Backward, new four angular coordinates of parking position are exported the path planning module of Intelligent parking system.

Figure 11 is the parking position design sketch of Overlapping display in panoramic view that will detect.

The above-mentioned description to embodiment is to be understood that for ease of those skilled in the art and apply the present invention.It is familiar with These embodiments obviously easily can be made various amendment by the personnel of art technology, and should General Principle described herein Use in other embodiments without through performing creative labour.Therefore, the invention is not restricted to above-described embodiment, art technology Personnel should be at the protection model of the present invention according to the announcement of the present invention, the improvement made without departing from scope of the invention and amendment Within enclosing.

Claims (5)

1. the level detecting apparatus of parking of an Intelligent parking system, it is characterised in that: including: sensor unit: be positioned at car Photographic head on vehicle body and distance measuring sensor;Signal processing unit: panoramic view generation module, parking position detect Module;Man-machine interface: output module is in order to show panoramic view and the warehouse compartment detected, input module is in order to accept The various instructions of driver's input;
Described parking position detection module includes park bit line detection sub-module, obstacle car detection sub-module, parking position output Submodule;Bit line detection sub-module of parking utilize ground park bit line detection parking position;Obstacle car detection sub-module Parking position is surveyed in Use barriers car test;Parking position output sub-module is according to bit line detection sub-module and the obstacle car test of parking Survey the testing result of submodule, finally determine parking position;
The method that described obstacle car detection sub-module determines parking position: the position according to holding profile before and after obstacle car is come really Determine parking position;
First obstacle Herba Plantaginis end profile position in global coordinate system uses following methods to determine: (1) utilizes ultrasonic Distance value that ripple distance measuring sensor obtains and vehicle posture information, simulate corresponding to obstacle car side profile Bar straight line;(2) when the distance value generation positive transition of ultrasonic distance-measuring sensor, looking around of this moment is stored Figure;(3) in panoramic view, set area-of-interest, make obstacle Herba Plantaginis end profile be positioned at area-of-interest;
(4) to described area-of-interest image gray processing, do rim detection, obtain obstacle Herba Plantaginis end profile The protruding point i.e. coordinate of First Point;(5) made vertical line by First Point to the straight line simulated, determine obstacle Herba Plantaginis End profile position in global coordinate system;
If there are two obstacle cars, then second obstacle car rear end profile position in global coordinate system uses with lower section Method determines: (6) utilize the distance value and vehicle posture information that ultrasonic distance-measuring sensor obtains, and simulate correspondence Straight line in obstacle car side profile;(7) when the distance value of ultrasonic distance-measuring sensor occurs negative saltus step, Store the panoramic view in this moment;(8) in panoramic view, set area-of-interest, make obstacle car rear end profile It is positioned at area-of-interest;(9) to described area-of-interest image gray processing, do rim detection, obtain obstacle The Herba Plantaginis end profile protruding point i.e. coordinate of rearmost point;(10) hung down to the straight line simulated by rearmost point Line, determines obstacle car rear end profile position in global coordinate system.
The level detecting apparatus of parking of Intelligent parking system the most according to claim 1, it is characterised in that: described sensing Device unit includes being positioned at vehicle front side, rear side, left side, four wide-angle cameras on right side and being positioned at vehicle both sides Ultrasonic distance-measuring sensor.
The level detecting apparatus of parking of Intelligent parking system the most according to claim 2, it is characterised in that: described four The installation site of photographic head should ensure that four camera collections to picture cover 360 degree of regions of vehicle's surroundings, And adjacent two camera collections to picture have overlapping region.
The level detecting apparatus of parking of Intelligent parking system the most according to claim 1, it is characterised in that: described man-machine Interface output module by the parking position Overlapping display that detects on panoramic view, if be detected that parking position the most just Really or do not meet the wish of driver, driver can adjust position and or the side of parking position by input module To, driver change parking position position and or direction after, new four angular coordinates of parking position are exported The path planning module of Intelligent parking system.
The level detecting apparatus of parking of Intelligent parking system the most according to claim 1, it is characterised in that park described in: Bit line detection sub-module determines that the method for parking position comprises the steps: (1) Image semantic classification;(2) from looking around Figure extracts straight line;(3) straight line is carried out post processing, the line segment corresponding to retain ground parking position line, delete Other mixed and disorderly line segments;(4) rectangle parking position is obtained.
CN201310545995.5A 2013-11-06 2013-11-06 A kind of parking position detection device and method of Intelligent parking system CN103600707B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201310545995.5A CN103600707B (en) 2013-11-06 2013-11-06 A kind of parking position detection device and method of Intelligent parking system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201310545995.5A CN103600707B (en) 2013-11-06 2013-11-06 A kind of parking position detection device and method of Intelligent parking system

Publications (2)

Publication Number Publication Date
CN103600707A CN103600707A (en) 2014-02-26
CN103600707B true CN103600707B (en) 2016-08-17

Family

ID=50119002

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201310545995.5A CN103600707B (en) 2013-11-06 2013-11-06 A kind of parking position detection device and method of Intelligent parking system

Country Status (1)

Country Link
CN (1) CN103600707B (en)

Families Citing this family (33)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104943686B (en) * 2014-03-28 2018-11-06 北京奇虎科技有限公司 Automatic parking implementation method and device based on wireless signal identification
CN104943687B (en) * 2014-03-28 2018-10-23 北京奇虎科技有限公司 Automatic parking implementation method and device based on wireless signal identification
CN103950448B (en) * 2014-04-21 2016-06-29 中国科学院深圳先进技术研究院 Park track guidance method and system, park orbit generation method and system
CN103950409B (en) * 2014-04-24 2016-04-27 中国科学院深圳先进技术研究院 parking assistance method and system
CN104608715B (en) * 2014-12-29 2016-11-30 江苏大学 A kind of parking position detection device and method of automated parking system
CN104637342B (en) * 2015-01-22 2017-01-04 江苏大学 A kind of narrow and small vertical parking stall scene intelligent identification and parking path planning system and method
CN104648243B (en) * 2015-01-27 2017-03-22 奇瑞汽车股份有限公司 Parking method and device
CN104951790B (en) * 2015-02-15 2018-02-02 北京联合大学 Based on multi-source against the seamless spliced Lane detection method of fluoroscopy images
US9582881B2 (en) * 2015-03-18 2017-02-28 Intel Corporation Machine vision image sensor calibration
CN104916162B (en) * 2015-05-28 2017-05-03 惠州华阳通用电子有限公司 Parking stall detection method and system
CN104933409B (en) * 2015-06-12 2018-04-03 北京理工大学 A kind of parking stall recognition methods based on panoramic picture dotted line feature
KR101705558B1 (en) * 2015-06-25 2017-02-13 (주)캠시스 Top view creating method for camera installed on vehicle and AVM system
CN104916163B (en) * 2015-06-29 2017-05-03 惠州华阳通用电子有限公司 Parking space detection method
CN105015418B (en) * 2015-07-17 2018-04-20 惠州华阳通用电子有限公司 Vehicle-carrying display screen video switching method
CN105139397B (en) * 2015-08-25 2017-12-19 广州视源电子科技股份有限公司 A kind of pcb board detection method and device
CN105513175B (en) * 2015-12-01 2017-10-24 盛视科技股份有限公司 Robot exit-entrance control method and system
CN105355083A (en) * 2015-12-14 2016-02-24 宁波裕兰信息科技有限公司 Vision-based 360-degree parking assist intelligent guiding system
CN106127736A (en) * 2016-06-15 2016-11-16 南京信必达智能技术有限公司 One is parked detection method and processor
CN107085402B (en) * 2017-01-21 2019-04-26 新昌县维利机械有限公司 The electric vehicle of Internet of Things monitoring management system is installed
CN107085393B (en) * 2017-01-22 2019-04-26 新昌县智创机械有限公司 The multifunctional electric of self-navigation
CN106971166A (en) * 2017-03-29 2017-07-21 纵目科技(上海)股份有限公司 The image pre-processing method and system of parking stall detection
CN106952308B (en) * 2017-04-01 2020-02-28 上海蔚来汽车有限公司 Method and system for determining position of moving object
CN107491738A (en) * 2017-07-25 2017-12-19 纵目科技(上海)股份有限公司 Parking space detection method and system, storage medium and electronic equipment
CN107527017A (en) * 2017-07-25 2017-12-29 纵目科技(上海)股份有限公司 Parking space detection method and system, storage medium and electronic equipment
CN107719361A (en) * 2017-10-10 2018-02-23 深圳市豪恩汽车电子装备股份有限公司 Automatic parking householder method and system based on image vision
CN110015288A (en) * 2018-01-09 2019-07-16 上海汽车集团股份有限公司 A kind of method, apparatus and electronic equipment detecting warehouse compartment
CN109720340A (en) * 2018-09-17 2019-05-07 魔门塔(苏州)科技有限公司 A kind of automated parking system and method for view-based access control model identification
CN109493633A (en) * 2018-12-20 2019-03-19 广州小鹏汽车科技有限公司 It is a kind of can parking stall detection method and device
CN109738900A (en) * 2019-01-02 2019-05-10 广州小鹏汽车科技有限公司 It is a kind of can parking stall detection method and device
CN109712427A (en) * 2019-01-03 2019-05-03 广州小鹏汽车科技有限公司 A kind of method for detecting parking stalls and device
CN109693666A (en) * 2019-02-02 2019-04-30 中国第一汽车股份有限公司 A kind of man-machine interactive system and method for parking for parking
CN110435638A (en) * 2019-06-28 2019-11-12 惠州市德赛西威汽车电子股份有限公司 A kind of parking position automatic tracking method
CN110415550A (en) * 2019-07-31 2019-11-05 北京智行者科技有限公司 The automatic parking method of view-based access control model

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1642808A1 (en) * 2004-09-30 2006-04-05 CLARION Co., Ltd. Parking-assist system using image information from an imaging camera and distance information from an infrared laser camera
CN1987357A (en) * 2006-12-26 2007-06-27 浙江工业大学 Intelligent parking auxiliary device based on omnibearing computer sight
CN101218127A (en) * 2005-07-08 2008-07-09 罗伯特·博世有限公司 Parking device
CN101426669A (en) * 2006-04-25 2009-05-06 丰田自动车株式会社 Parking assistance device and parking assistance method
CN102449673A (en) * 2009-06-03 2012-05-09 爱信精机株式会社 Method of monitoring the surroundings of a vehicle, and device for monitoring the surroundings of a vehicle
CN102474597A (en) * 2009-08-03 2012-05-23 爱信精机株式会社 Vehicle peripheral image generation device

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2005343417A (en) * 2004-06-07 2005-12-15 Auto Network Gijutsu Kenkyusho:Kk Parking assisting device

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1642808A1 (en) * 2004-09-30 2006-04-05 CLARION Co., Ltd. Parking-assist system using image information from an imaging camera and distance information from an infrared laser camera
CN101218127A (en) * 2005-07-08 2008-07-09 罗伯特·博世有限公司 Parking device
CN101426669A (en) * 2006-04-25 2009-05-06 丰田自动车株式会社 Parking assistance device and parking assistance method
CN1987357A (en) * 2006-12-26 2007-06-27 浙江工业大学 Intelligent parking auxiliary device based on omnibearing computer sight
CN102449673A (en) * 2009-06-03 2012-05-09 爱信精机株式会社 Method of monitoring the surroundings of a vehicle, and device for monitoring the surroundings of a vehicle
CN102474597A (en) * 2009-08-03 2012-05-23 爱信精机株式会社 Vehicle peripheral image generation device

Also Published As

Publication number Publication date
CN103600707A (en) 2014-02-26

Similar Documents

Publication Publication Date Title
US9916509B2 (en) Systems and methods for curb detection and pedestrian hazard assessment
EP3049285B1 (en) Driver assistance system for displaying surroundings of a vehicle
US10274598B2 (en) Navigation based on radar-cued visual imaging
JP5689907B2 (en) Method for improving the detection of a moving object in a vehicle
CN103770706B (en) Dynamic reversing mirror indicating characteristic
CN103021177B (en) Method and system for processing traffic monitoring video image in foggy day
Liu et al. Rear vehicle detection and tracking for lane change assist
Broggi Automatic vehicle guidance: the experience of the ARGO autonomous vehicle
Rasmussen Grouping dominant orientations for ill-structured road following
Bertozzi et al. Real-time lane and obstacle detection on the gold system
US6963661B1 (en) Obstacle detection system and method therefor
Suhr et al. Automatic free parking space detection by using motion stereo-based 3D reconstruction
Hautiére et al. Automatic fog detection and estimation of visibility distance through use of an onboard camera
CN101604448B (en) Method and system for measuring speed of moving targets
EP2769239B1 (en) Methods and systems for creating maps with radar-optical imaging fusion
Kühnl et al. Spatial ray features for real-time ego-lane extraction
CN101734214B (en) Intelligent vehicle device and method for preventing collision to passerby
Wu et al. Applying a functional neurofuzzy network to real-time lane detection and front-vehicle distance measurement
EP1267178B1 (en) Method for processing a high definition picture
López et al. Robust lane markings detection and road geometry computation
EP2237988B1 (en) Object detection and recognition system
JP3868876B2 (en) Obstacle detection apparatus and method
US6906620B2 (en) Obstacle detection device and method therefor
Hautière et al. Towards fog-free in-vehicle vision systems through contrast restoration
CN106796648A (en) For the system and method for detection object

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant