CN106157289A - Line detecting method and equipment - Google Patents
Line detecting method and equipment Download PDFInfo
- Publication number
- CN106157289A CN106157289A CN201510163192.2A CN201510163192A CN106157289A CN 106157289 A CN106157289 A CN 106157289A CN 201510163192 A CN201510163192 A CN 201510163192A CN 106157289 A CN106157289 A CN 106157289A
- Authority
- CN
- China
- Prior art keywords
- line
- model
- line segment
- segment
- detection
- 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
Landscapes
- Image Analysis (AREA)
Abstract
Provide a kind of line detecting method and equipment.Described method includes: extract the line segment of the feature of the line conforming to detection in current frame image;Obtain the line model pre-building;Update described line model based on the line segment being extracted;According to the line model after renewal, determine line to be detected.Can be greatly reduced by described line detecting method and equipment and block the impact updating line model with noise, thus improve the accuracy of line detection.
Description
Technical field
This patent disclosure relates generally to image procossing, be specifically related to line detecting method and equipment.
Background technology
Line detection technique has a wide range of applications in image processing field, and for example, lane detection is line inspection
One important application of survey technology.Existing lane detection technology can be divided into two classes: feature based
Method and the method based on model.
The method of feature based carrys out positioning car diatom by the combination of low-level image feature.But, the method for
The edge shape of lane line does not has any global restriction, is therefore affected bigger by blocking with noise.
Compared to the method for feature based, some can be used based on the method for model to join when detecting lane line
Number represents lane line, and therefore the method is for blocking and noise more robust.Method based on model
General principle is to update lane line model according to the feature of input, and then obtains lane line to be detected.
Generally, the feature of input is based on the characteristic point representing that the parameter detecting of lane line arrives, such as marginal point or mistake
Parallax point etc. after filter.When model modification, it will choose for each bar lane line in lane line model
Characteristic point, and utilize the characteristic point chosen to enter line matching to update lane line model.Selected characteristic point
Principle be normally based on the distance of each bar lane line in characteristic point and lane line model.But, detection
To characteristic point be inaccurate, wherein may include noise spot.Such as Fig. 1 is exemplified with the spy detecting
The illustrative case of noise spot is included in levying a little.Therefore, when the characteristic point chosen is too much, particularly elected
When containing noise spot in the characteristic point taking, the renewal of lane line model will be affected thus cause not
Lane detection result accurately.
Content of the invention
The purpose of the disclosure is at least to solve the problems referred to above.Specifically, the purpose of the disclosure is to carry
Go out a kind of line detection technique, can be greatly reduced by this technology and block the shadow updating line model with noise
Ring, thus improve the accuracy of line detection.
According to embodiments of the invention, provide a kind of line detecting method, comprising: in current frame image
Extract the line segment of the feature of the line conforming to detection;Obtain the line model pre-building;Based on extracted
Line segment updates described line model;According to the line model after renewal, determine line to be detected.
According to another embodiment of the present invention, a kind of line detection equipment is provided, comprising: line segment extraction part
Part, is configured in current frame image extract the line segment of the feature of the line conforming to detection;Model acquisition unit
Part, is configured to obtain the line model pre-building;Update parts, be configured to extracted line segment more
New described line model;Detection part, the line model after being configured to according to renewal, determine line to be detected.
Line detection technique according to embodiments of the present invention, utilizes Eigenvector rather than characteristic point to carry out mould
Type matching is to update line model.Owing to considering feature when selected characteristic line segment is to carry out models fitting simultaneously
In line segment and line model, the positional distance of each bar line and direction distance, therefore substantially reduce and choose noise
Line segment updates the possibility of line model, thus reduces the impact that line model is updated by noise, improves
The accuracy of line detection.
Brief description
By combining accompanying drawing, disclosure embodiment is described in more detail, the disclosure above-mentioned and its
Its purpose, feature and advantage will be apparent from.Accompanying drawing is used for providing entering one to disclosure embodiment
Step understands, and constitutes a part for specification, is used for explaining the disclosure together with disclosure embodiment,
It is not intended that restriction of this disclosure.In the accompanying drawings, identical reference number typically represent same parts or
Step.
Fig. 1 is exemplified with the illustrative case including noise spot in the characteristic point detecting.
Fig. 2 shows the flow chart of line detecting method according to embodiments of the present invention.
Fig. 3 shows and examines at line according to embodiments of the present invention in the case that line to be detected is lane line
Survey method is extracted the example of the line segment obtaining.
Fig. 4 shows the example of a lane line model.
Fig. 5 is exemplified with the process updating line model according to an embodiment of the invention based on the line segment extracting
Flow chart.
Fig. 6 is exemplified with the functional configuration block diagram of line detection equipment according to embodiments of the present invention.
Fig. 7 is exemplified with the general hardware block diagram of line detecting system according to embodiments of the present invention.
Detailed description of the invention
Below in conjunction with the accompanying drawing in disclosure embodiment, the technical scheme in disclosure embodiment is carried out
Clearly and completely describe, it is clear that described embodiment is only a part of embodiment of the disclosure, and
It is not all, of embodiment.Based on the embodiment in the disclosure, those of ordinary skill in the art are not doing
The every other embodiment being obtained under the premise of going out creative work, broadly falls into the scope of disclosure protection.
Fig. 2 shows the flow chart of line detecting method according to embodiments of the present invention.For convenience of description,
Hereinafter, illustrate as a example by line to be detected is lane line.
As in figure 2 it is shown, in step S201, current frame image extracts the feature of the line conforming to detection
Line segment.
The feature of described line to be detected can be any feature that can characterize this line.For example, it is permissible
It is the features such as the color of (but not limited to) line, gray scale, shape, edge, parallax, or these features
Any combination.
In this step, the line of the feature of the line conforming to detection will be extracted in any suitable manner
Section.As an example, can be in the current frame image that shooting obtains, by Hough transformation etc.
Line detection method directly detects the line segment of the feature of the line conforming to detection.As another example, can
To be primarily based on the feature of line to be detected, by detection method corresponding with this feature at current frame image
Middle detection characteristic point (for example, if the feature of line to be detected is edge feature, then can pass through edge
Detection method detects characteristic point), the characteristic point matching then utilizing detection to obtain obtains described line segment.Figure
3 show the example being extracted the line segment obtaining in the case that line to be detected is lane line by this step.
Runic white line segment table in this figure shows the line segment that extraction obtains.
In step S202, obtain the line model pre-building.
As a example by line to be detected is lane line, correspondingly, described line model is lane line model.This car
Road line model can apply any existing mode in this area to pre-build.How to set up lane line
Model is not the key point of the present invention, and be merely to illustrate that is complete herein, exemplary to one
Lane line model set up mode and simply describe.
Lane line model to be set up, first has to carry out lane detection.Herein, can use any existing
Mode detect lane line.For example, as a kind of conventional method, can use at the beginning of B-Snake model
CHEVP algorithm in beginningization detection lane line in the picture frame that shooting obtains.For another example, as one
More basic method, can determine lane line by manually marking in picture frame.
After lane line being detected, lane line model can be set up according to lane line.Existing lane line mould
Type has a variety of, including linear model, isolated point model, parabola model and extension thereof, hyperbola mould
Type, clothoid model, Spline Model, Snake model, 3D model etc..Those skilled in the art are permissible
Set up the lane line model of any one type according to real needs.In this example, multinomial is used to represent
Lane line model.
Concrete, as shown in Figure 4, it is assumed that the center line in a certain track is Lmid, the left-hand lane line in this track
It is L-1, right-hand lane line is L1, the left-hand lane line in this n-th track, left side, track is L-n, the right n-th
The left-hand lane line in individual track is Ln;The lane width in this certain track is w0, its n-th track, left side
Lane width is w-n, on the right of it, the lane width in n-th track is wn;Vp represents the position that lane line disappears
The ordinate put;F is the focal length of camera lens, and H is the height of camera position.
LmidPolynomial repressentation as shown in equation (1) for the employing:
Then lane line model is represented by:
Wherein,
As it was previously stated, first carried out lane detection before setting up lane line model, therefore can pass through
This lane detection result is substituted into above-mentioned expression formula (1), (2), in (3), is calculated aiWith
wi(i=0,1 ..., value n).Thus, the foundation of lane line model is completed.
Lane line model shown in above-mentioned expression formula (2) is only an example, it would however also be possible to employ other shapes
The expression formula of formula.For example, it is possible to do not use track center line, but directly utilize left side and the right side in track
Lane line sets up lane line model expression.More particularly, the multinomial shown in expression formula (1) can
Being used directly to represent the left side in a certain track or right-hand lane line, rather than the center line in this track, and phase
Ground is answered to adjust the expression way in each track in expression formula (2).
Below mode is set up to exemplary lane line model and carried out simple description.In this step
In S202, obtain the lane line model being pre-build by this way of example or any other mode.Can
Understand, as an example, when detecting lane line continuously in a series of frame of video, carry out
In the case that the current frame image of detection is other two field pictures in addition to the first two field picture, described build in advance
Vertical line model can be to detect the lane line model that lane line is used in previous frame image.Need explanation
, in practical operation, when detecting lane line continuously in a series of frame of video, generally at this
Front N frame in a series of frame of video is set up line model and carries out verifying to it and (for example build in the first frame
Vertical line model simultaneously verifies this model in N-1 frame subsequently, and N is greater than the integer of 1, and its value can root
Set according to concrete needs), rather than set up just directly use after line model only in the first frame.Therefore,
As another example, the feelings of the image of the frame after the current frame image to detect is nth frame
Under condition, the described line model pre-building can take and detect the car that lane line is used in its previous frame image
Road line model.
In step S203, update described line model based on the line segment being extracted.
It is understood that the line segment owing to extracting in step s 201 is the feature of the line conforming to detection,
Therefore these line segments are the candidate line sections forming line to be detected, i.e. can obtain based on these line segments and want
The line of detection.Concrete, in this step, obtain using extracted line segment to update in step S202
The line model taking.
As a kind of basic update method, each line segment being extracted can be directly utilized and carry out model plan
Close, thus obtain the line model updating.
Conventional least square method can be used to carry out described models fitting.For example, for such as expression formula
(2) the lane line model described in, can choose so that following formula (4) has the matching of minimum of a value
Mode, and determine a in this fit approachiAnd wiValue, thus obtain update lane line model:
Wherein, DIFFjiFor line segment SjWith the line L in line modeliDifference, this difference can be passed through similar
The any-modes such as degree, distance represent.Certainly, least square method is only a kind of example, it is possible to use
Under such as gradient, other existing modes of degradation carry out models fitting.
Above-mentioned line model update method effect in some cases is not likely to be fine.For example, such as Fig. 3 institute
Showing, lane line to be detected has 3, i.e. comprises 3 lane lines in lane line model, and extracts and obtain
Line segment be dispersed in current frame image, therefore extraction is obtained line segment carry out models fitting when, it is impossible to
Determine which bar lane line each bar line segment should correspond to respectively, thus the renewal of lane line model can be affected.
For another example, it can be seen that being noise with the line segment that circle outlines in such as Fig. 3, it shall not be applied to track
The renewal of line model, and according to above-mentioned line model update method, this line segment also will be used for model to be intended
Close, thus the renewal of lane line model will be affected.For this problem, as example, in step S203
In can use the line model update method being described below.Below with reference to Fig. 5, the method is retouched
State.
Fig. 5 is exemplified with the process updating line model according to an embodiment of the invention based on the line segment extracting
Flow chart.
As it is shown in figure 5, in step S2031, for each described line segment, determine in line model with its
Corresponding line.
In this step, as example, for each described line segment, it may be considered that during it is with line model
The positional distance of every line and direction distance both determine corresponding line.Show below for this
Example method is described in detail.
Assume DijFor line segment SjLine L in line modeliPositional distance, then
Wherein, njFor line segment SjOn the number of pixel, dkiFor line segment SjOn some k to line model
In line LiPositional distance.
Positional distance d in expression formula (5)kiCan by various suitable by way of calculate.For example, make
For example, it is possible to use Euclidean distance represents this positional distance dki, i.e. assume described line LiExpression formula
For aix+biy+ci=0 (as a example by straight line), then
It is understood that as the line L in line modeliWhen being curve, equally utilize in this areas such as derivation
Known method calculates line segment SjOn point to this line LiEuclidean distance, here is omitted.
On the other hand, Euclidean distance is utilized to carry out the positional distance d in expression (5)kiIt is only one
Example, it is also possible to represent this positional distance d by such as Pasteur distance, mahalanobis distance etc.ki。
Line segment can be calculated by various suitable methods with the direction distance of the line in line model.For example,
When the line in line model is straight line, can directly calculate between the direction of this straight line and the direction of line segment
Angular separation is as described direction distance.In the disclosure, as a kind of example, following expression is passed through
(7) the direction distance of line segment and the line in line model is calculated.Concrete, it is assumed that TijFor line segment SjTo line
Line L in modeli(this line LiBoth can be straight line, it is possible to be curve) direction distance, then
Wherein,For line segment SjOn some k at gradient (line segment SjOn each point at gradient all identical,
The i.e. gradient of this line segment),For the line L in line modeliAt a gradient direction for the tangent line at k ' place, this k '
It is line LiGo up and line segment SjOn some k there are identical abscissa or the corresponding points of identical ordinate.
It is being calculated its direction distance with every line in line model as described for each line segment
After positional distance, can using the line minimum with the direction distance and positional distance of this line segment as with this line
The corresponding line of section.As example, can by following process determine direction distance with described line segment and
The minimum line of positional distance: for every line in line model, calculates described line as shown in expression formula (8)
The ranking operation result of section and the direction distance of this line and positional distance, and select to transport with minimum weighting
Calculate the corresponding line of result as the line minimum with the direction distance and positional distance of described line segment:
w1*Dij+w2*Tij (8)
Wherein, DijAnd TijIt is respectively line segment SjLine L in line modeliPositional distance and direction distance,
W1 represents the weight of positional distance, and w2 represents the weight of direction distance, and described weight can be appointed as required
Meaning sets, usual w1+w2=1.
In the above example, based on ranking operation result determine the direction distance with line segment and position away from
From minimum line, this is only a kind of example, it is also possible to for example square based on direction distance and positional distance
Difference etc. determines the direction distance with line segment and the line of positional distance minimum.
In step S2032, according to determined by the corresponding relation of each line segment and the line in line model, profit
Carry out models fitting with each line segment described, thus obtain the line model updating.
In above-mentioned steps S2031, line corresponding in line model is determined for each line segment, i.e.
Determine each bar line segment and should be used for the renewal of any bar lane line respectively.In this step S2032, will be based on
The corresponding relation of this determination, utilizes each line segment described to carry out models fitting.
As previously mentioned, it is possible to use under such as least square method, gradient, the various mode of degradation carries out model
Matching.For example, as a example by using least square method to be fitted, for as illustrated in expression formula (2)
Lane line model, can choose so that following formula (9) has the fit approach of minimum of a value, and determine
A in this fit approachiAnd wiValue, thus obtain update lane line model:
Wherein, DijAnd TijIt is respectively line segment SjLine L in line modeliPositional distance and direction distance,
N is the quantity of model center line, miRepresent and the quantity of the corresponding line segment of i-th line, w1 represent position away from
From weight, w2 represents the weight of direction distance, and described weight can arbitrarily set, generally as required
W1+w2=1.
Describe the example updating line model based on the line segment extracting above in association with Fig. 5.Optionally, in step
In rapid S2031, if for a certain line segment, this line segment and every line in line model are calculated
The ranking operation result of direction distance and positional distance (or variance etc.), and wherein ranking operation result (or
Variance etc.) minimum of a value also greater than a threshold value set in advance, then it is assumed that this line segment is noise, line model
In the not corresponding line with this line segment.So, step S2032 subsequently will not utilize this line segment
Carry out models fitting, such that it is able to avoid the impact that line model is updated by noise.
Return to Fig. 2, in step S204, according to the line model after renewal, determine line to be detected.
Described each bar line in current frame image by the line model after the renewal that above-mentioned process obtains, because of
This can directly obtain the line to be detected current frame image from the line model after this renewal.
Line detecting method according to embodiments of the present invention described in detail above, utilizes feature in the method
Line segment carries out models fitting to update line model.In the above-mentioned line model update method of the present embodiment,
Not only consider the position of Eigenvector and each bar line in line model when selected characteristic line segment carries out models fitting
Distance further contemplates the direction distance between them, only considers feature when being therefore fitted with selected characteristic point
Point, compared with the scheme of the positional distance of line, substantially reduces and chooses noise line segment to update line model
Possibility, thus reduce the impact that line model is updated by noise, improve the accuracy of line detection.
Below with reference to Fig. 6, the line detection equipment according to disclosure embodiment is described.Fig. 6 is exemplified with according to this
The functional configuration block diagram of the line detection equipment 600 of inventive embodiments.As shown in Figure 6, line detection equipment 600
Can include that the 602nd, line segments extraction parts the 601st, model obtaining widget updates parts 603 and detection part
604, described all parts can perform each step of the line detecting method describing above in conjunction with Fig. 2 respectively
/ function.Therefore, only the major function of each parts of this line detection equipment 600 is described below, and
Omit the detail content having been described above.
Line segments extraction parts 601 are configured in current frame image to extract the feature of the line conforming to detection
Line segment.The feature of described line to be detected can be any feature that can characterize this line.For example, it can
Being the features such as the color of (but not limited to) line, gray scale, shape, edge, parallax, or these features
Any combination.Line segments extraction parts 601 conform to detection by extracting in any suitable manner
The line segment of the feature of line.As an example, can be in the current frame image that shooting obtains, by all
As Hough transformation isoline detection method directly detects the line segment of the feature of the line conforming to detection.As separately
One example, can be primarily based on the feature of line to be detected, by detection method corresponding with this feature
Detecting characteristic point in current frame image, the characteristic point matching then utilizing detection to obtain obtains described line segment.
Model obtaining widget 602 is configured to obtain the line model pre-building.As described above, this line mould
Type can apply any existing mode in this area to pre-build, and how to set up line model simultaneously
Non-is the key point of the present invention, is therefore not described in detail herein.It is understood that as an example,
It when detecting lane line continuously in a series of frame of video, is to remove at the current frame image to detect
In the case of other two field pictures outside first two field picture, the described line model pre-building can be previous
Two field picture detects the lane line model that lane line is used.As another example, when regarding a series of
When frequently frame detecting lane line continuously, if the first frame wherein being set up line model and subsequently
It is verified by N-1 frame, then the frame after the current frame image to detect is nth frame
In the case of image, the described line model pre-building can take detection lane line institute in its previous frame image
The lane line model using.
Update parts 603 and be configured to the line segment more new model acquisition unit that line segments extraction parts 601 extract
The line model that part 602 obtains.
As an example, update parts 603 and utilize each line segment being extracted to carry out models fitting, from
And obtain the line model updating.Concrete, update parts 603 and can utilize such as least square method, ladder
The lower various mode of degradation of degree carries out models fitting.
As another example, update parts 603 be configured to for each line segment determine in line model with
Its corresponding line, and according to determined by the corresponding relation of line in each line segment and line model, utilize institute
State each line segment and carry out models fitting, thus obtain the line model updating.Wherein, for each line
When section determines line corresponding in line model, update parts 603 and be configurable to for described in each
Line segment, calculate its with line model in the direction distance of every line and positional distance, and by with this line segment
The line of direction distance and positional distance minimum is as line corresponding with this line segment.
Line model after detection part 604 is configured to according to renewal determines line to be detected.
Fig. 7 is exemplified with the general hardware block diagram of line detecting system 700 according to embodiments of the present invention.Such as figure
Shown in 7, line detecting system 700 may include that input equipment 710, for from the relevant image of outside input
Or information, the depth map of such as video camera shooting, gray-scale map (cromogram) etc., this input equipment is for example
Can be keyboard, mouse, video camera etc.;Processing equipment 720, above-mentioned according to these public affairs for implementing
Opening the line detecting method of embodiment, or being embodied as above-mentioned line detection equipment, this processing equipment can be
Being capable of any device with disposal ability of above-mentioned functions, for example it can be to be designed for carrying out
The general processor of function described herein, digital signal processor (DSP), ASIC, field programmable door
Array signal (FPGA) or other PLDs (PLD), discrete gate or transistor logic,
Discrete nextport hardware component NextPort or its any combination;Output equipment 730, implements above-mentioned for the output to outside
Result obtained by line detection process, the line for example detecting, this output equipment can be for example display,
Printer etc.;And storage device 740, for storing the inspection of above-mentioned line in the way of volatile and nonvolatile
Such as depth map, gray-scale map (cromogram) involved by survey process, various threshold value, the line mould pre-building
Type, the line segment extracting, the line model etc. of renewal, this storage device can be for example that arbitrary access is deposited
Reservoir (RAM), read-only storage (ROM), hard disk or semiconductor memory etc. various volatile
Or nonvolatile memory.
Carry out as a example by line detection technique according to embodiments of the present invention is applied to lane detection above
Describe, it is to be understood that line detection technique according to embodiments of the present invention also can apply to need to use line
In other lines various detection situation of model.
Describe the general principle of the present invention above in association with specific embodiment, however, it is desirable to it is noted that
The advantage mentioned in the disclosure, advantage, effect etc. are only exemplary rather than limiting, it is impossible to think that these are excellent
Point, advantage, effect etc. are that each embodiment of the present invention is prerequisite.In addition, tool disclosed above
Body details is merely to the effect of example and the effect readily appreciating, and unrestricted, and above-mentioned details does not limit
The present invention processed is for using above-mentioned concrete details to realize.
The device that relates in the disclosure, device, equipment, the block diagram of system are only used as exemplary example
And it is not intended to requirement or hint must be attached according to the mode shown in block diagram, arrange, configure.
As the skilled person will recognize, can connect by any-mode, arrange, configure these devices,
Device, equipment, system.Such as " include ", "comprising", the word of " having " etc. are open vocabulary,
Refer to " including but not limited to ", and use can be exchanged with it.Vocabulary "or" used herein above and " with " refer to vocabulary
"and/or", and use can be exchanged with it, unless it is not such that context is explicitly indicated.Used herein above
Vocabulary " such as " refer to phrase " such as, but not limited to ", and use can be exchanged with it.
Flow chart of steps in the disclosure and above method describe and are only used as exemplary example and unawareness
The step that figure requires or hint must carry out each embodiment according to the order providing, some step can be simultaneously
Row, independently of one another or perform according to other suitable orders.In addition, such as " thereafter ", " then ", " connect
Get off " etc. word be not intended to limit the order of step;These words are only used for guiding reader to read over these
The description of method.
In addition, as used herein, the "or" enumerating middle use at the item starting with " at least one " refers to
Show enumerating of separation, in order to enumerating of such as " at least one of A, B or C " means A or B or C,
Or AB or AC or BC, or ABC (i.e. A and B and C).Additionally, wording " example " it is not intended that
It is preferred or more preferable than other examples for the example describing.
It may also be noted that in apparatus and method of the present invention, each parts or each step are to divide
Solve and/or reconfigure.These decompose and/or reconfigure the equivalents that should be regarded as the present invention.
For those of ordinary skill in the art, it is to be understood that disclosed method and device whole or
Person's any part, can be at any computing device (including processor, storage medium etc.) or calculating dress
In the network put, realized with hardware, firmware, software or combinations thereof.Described hardware is permissible
It is general processor, the digital signal processor utilizing and being designed to carry out function described herein
(DSP), ASIC, field programmable gate array signal (FPGA) or other PLDs (PLD),
Discrete gate or transistor logic, discrete nextport hardware component NextPort or its any combination.General processor can be
Microprocessor, but as replacing, this processor can be any commercially available processor, control
Device processed, microcontroller or state machine.Processor is also implemented as the combination of computing device, such as DSP
One or more microprocessor of cooperating with DSP core with the combination of microprocessor, multi-microprocessor or
Any other such configuration.Described software may reside in any type of computer-readable tangible deposit
In storage media.By example rather than restriction, such computer-readable tangible media can wrap
Include RAM, ROM, EEPROM, CD-ROM or other optical disc storage, disk storage or other magnetic
Memory device or may be used for carrying or store the desired program code of instruction or data structure form simultaneously
And any other tangible medium that can be accessed by computer.As used herein, dish includes compact disk
(CD), laser disk, CD, digital universal disc (DVD), floppy disk and Blu-ray disc.
Intelligent control technology disclosed by the invention can also be by running a program in any computing device
Or batch processing realizes.Described computing device can be known fexible unit.Presently disclosed
Intellectual technology also can be only by the journey providing the program code comprising to realize described method or device
Sequence product realizes, or is realized by any storage medium of such program product that is stored with.
Can carry out to technology described herein without departing from the technology instructed defined by the appended claims
Various changes, replacement and change.Additionally, the scope of the claim of the disclosure is not limited to the above
Process, machine, manufacture, the composition of event, means, the specific aspect of method and action.Can profit
With to corresponding aspect described herein carry out essentially identical function or realize essentially identical result work as
Front existence or to be developed after a while process, machine, manufacture, the composition of event, means, method or
Action.Thus, claims include the such process in the range of it, machine, manufacture, thing
The composition of part, means, method or action.
The above description of disclosed aspect is provided so that any person skilled in the art can make or
Use the present invention.It is to show very much and easy to those skilled in the art to the various modifications in terms of these
See, and General Principle defined herein can apply to other aspects without deviating from the scope of the present invention.
Therefore, the present invention is not intended to be limited to the aspect being shown in which, but according to principle disclosed herein
The widest range consistent with novel feature.
In order to purpose of illustration and description has been presented for above description.It is not intended to this additionally, this describes
Inventive embodiment is restricted to form disclosed herein.Although multiple exemplary aspect already discussed above and
Embodiment, but it would be recognized by those skilled in the art that its some modification, modification, change, add and son
Combination.
Claims (10)
1. a line detecting method, comprising:
The line segment of the feature of the line conforming to detection is extracted in current frame image;
Obtain the line model pre-building;
Update described line model based on the line segment being extracted;And
According to the line model after renewal, determine line to be detected.
2. line detecting method as claimed in claim 1, the feature of wherein said line to be detected includes line
Color, gray scale, shape, edge, at least one in parallax feature.
3. line detecting method as claimed in claim 1, wherein said extraction in current frame image meets
The line segment of the feature of line to be detected farther includes:
Based on the feature of line to be detected, current frame image detects characteristic point;
The characteristic point matching utilizing detection to obtain obtains described line segment.
4. line detecting method as claimed in claim 1, wherein said extraction in current frame image meets
The line segment of the feature of line to be detected farther includes:
Detection in current frame image conforms to the line segment of the feature of the line of detection.
5. line detecting method as claimed in claim 1, is wherein frame of video sequence at described current frame image
In row in the case of the image of the frame after nth frame image, the described line model pre-building is that this is current
The line model using in the previous frame image of two field picture, wherein N is greater than the arbitrary integer equal to 1.
6. line detecting method as claimed in claim 1, wherein said updates institute based on the line segment being extracted
State line model to farther include:
Utilize each line segment being extracted to carry out models fitting, thus obtain the line model updating.
7. line detecting method as claimed in claim 1, wherein said updates institute based on the line segment being extracted
State line model to farther include:
For each described line segment, determine line corresponding in line model;
Each line segment determined by according to and the corresponding relation of the line in line model, utilize each line segment described
Carry out models fitting, thus obtain the line model updating.
8. line detecting method as claimed in claim 7, wherein determines line mould for each described line segment
Line corresponding in type farther includes:
For each described line segment, calculate its direction distance with every line in line model and position away from
From;
Using the line minimum with the direction distance and positional distance of this line segment as line corresponding with this line segment.
9. line detecting method as claimed in claim 8, wherein by the direction distance and position with this line segment
The minimum line of distance farther includes as line corresponding with this line segment:
For every line in line model, calculate direction distance and the positional distance of described line segment and this line
Ranking operation result;
If the minimum of a value in each described ranking operation result is less than predetermined threshold value, then will be with this minimum
It is worth corresponding line as line corresponding with this line segment.
10. a line detection equipment, comprising:
Line segments extraction parts, are configured in current frame image extract the line of the feature of the line conforming to detection
Section;
Model obtaining widget, is configured to obtain the line model pre-building;
Update parts, be configured to extracted line segment and update described line model;
Detection part, the line model after being configured to according to renewal, determine line to be detected.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510163192.2A CN106157289B (en) | 2015-04-08 | 2015-04-08 | Line detecting method and equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510163192.2A CN106157289B (en) | 2015-04-08 | 2015-04-08 | Line detecting method and equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106157289A true CN106157289A (en) | 2016-11-23 |
CN106157289B CN106157289B (en) | 2019-08-09 |
Family
ID=57336812
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510163192.2A Expired - Fee Related CN106157289B (en) | 2015-04-08 | 2015-04-08 | Line detecting method and equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106157289B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109931884A (en) * | 2019-01-31 | 2019-06-25 | 上海市质量监督检验技术研究院 | A kind of strip water nozzle rotation angle non-contact measurement method |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102133876A (en) * | 2011-02-23 | 2011-07-27 | 郭长有 | Anti-deviation alarm device for automobile |
CN102529975A (en) * | 2010-12-13 | 2012-07-04 | 通用汽车环球科技运作有限责任公司 | Systems and methods for precise sub-lane vehicle positioning |
CN202345671U (en) * | 2011-11-11 | 2012-07-25 | 长安大学 | Video camera-based device for vehicle lane departure detection and early warning |
CN103605362A (en) * | 2013-09-11 | 2014-02-26 | 天津工业大学 | Learning and anomaly detection method based on multi-feature motion modes of vehicle traces |
CN103942533A (en) * | 2014-03-24 | 2014-07-23 | 河海大学常州校区 | Urban traffic illegal behavior detection method based on video monitoring system |
CN103971081A (en) * | 2013-01-25 | 2014-08-06 | 株式会社理光 | Multi-lane detection method and system |
CN104008387A (en) * | 2014-05-19 | 2014-08-27 | 山东科技大学 | Lane line detection method based on feature point piecewise linear fitting |
-
2015
- 2015-04-08 CN CN201510163192.2A patent/CN106157289B/en not_active Expired - Fee Related
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102529975A (en) * | 2010-12-13 | 2012-07-04 | 通用汽车环球科技运作有限责任公司 | Systems and methods for precise sub-lane vehicle positioning |
CN102133876A (en) * | 2011-02-23 | 2011-07-27 | 郭长有 | Anti-deviation alarm device for automobile |
CN202345671U (en) * | 2011-11-11 | 2012-07-25 | 长安大学 | Video camera-based device for vehicle lane departure detection and early warning |
CN103971081A (en) * | 2013-01-25 | 2014-08-06 | 株式会社理光 | Multi-lane detection method and system |
CN103605362A (en) * | 2013-09-11 | 2014-02-26 | 天津工业大学 | Learning and anomaly detection method based on multi-feature motion modes of vehicle traces |
CN103942533A (en) * | 2014-03-24 | 2014-07-23 | 河海大学常州校区 | Urban traffic illegal behavior detection method based on video monitoring system |
CN104008387A (en) * | 2014-05-19 | 2014-08-27 | 山东科技大学 | Lane line detection method based on feature point piecewise linear fitting |
Non-Patent Citations (6)
Title |
---|
JIANG RUYI 等: "Lane detection and tracking using a new lane model and distance transform", 《MACHINE VISION AND APPLICATIONS》 * |
JIN-WOOK LEE 等: "Effective lane detection and tracking method using statistical modeling of color and lane edge-orientation", 《2009 FOURTH INTERNATIONAL CONFERENCE ON COMPUTER SCIENCES AND CONVERGENCE INFORMATION TECHNOLOGY》 * |
余厚云 等: "直线模型下的车道线跟踪与车道偏离检测", 《自动化仪表》 * |
刘献如 等: "结构化道路车道线的鲁棒检测与跟踪", 《光电子·激光》 * |
张翀 等: "基于直线模型的车道线实时检测方法", 《计算机工程与设计》 * |
王晓云 等: "基于线性双曲线模型的车道线检测算法", 《杭州电子科技大学学报》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109931884A (en) * | 2019-01-31 | 2019-06-25 | 上海市质量监督检验技术研究院 | A kind of strip water nozzle rotation angle non-contact measurement method |
Also Published As
Publication number | Publication date |
---|---|
CN106157289B (en) | 2019-08-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Cheng et al. | Curb detection for road and sidewalk detection | |
EP3321887A1 (en) | Urban road recognition method, apparatus, storage medium and device based on laser point cloud | |
CN104050654A (en) | Road edge detection method and device | |
CN105740802A (en) | Disparity map-based obstacle detection method and device as well as automobile driving assistance system | |
Sumer et al. | An adaptive fuzzy-genetic algorithm approach for building detection using high-resolution satellite images | |
WO2016210227A1 (en) | Aligning 3d point clouds using loop closures | |
CN115540896B (en) | Path planning method and device, electronic equipment and computer readable medium | |
CN104424649A (en) | Method and system for detecting moving object | |
CN110895819B (en) | Target tracking method, target tracking device, computer-readable storage medium and computer equipment | |
Tapia-Espinoza et al. | Robust lane sensing and departure warning under shadows and occlusions | |
WO2016066042A1 (en) | Segmentation method for commodity picture and device thereof | |
CN101582171B (en) | Method and device for creating depth maps | |
EP3182369A1 (en) | Stereo matching method, controller and system | |
CN105741312A (en) | Target object tracking method and device | |
CN107808524B (en) | Road intersection vehicle detection method based on unmanned aerial vehicle | |
Lu et al. | Lane marking based vehicle localization using particle filter and multi-kernel estimation | |
CN106157283A (en) | The detection method of lane segmentation thing and device | |
KR101507732B1 (en) | Method for segmenting aerial images based region and Computer readable storage medium for storing program code executing the same | |
KR20180078115A (en) | Method and apparatus for processing image | |
WO2017200889A1 (en) | Classifying entities in digital maps using discrete non-trace positioning data | |
JP6561605B2 (en) | Target analysis method and target analysis system | |
CN106157289A (en) | Line detecting method and equipment | |
Saleem et al. | Improved stixel estimation based on transitivity analysis in disparity space | |
Huang et al. | Probabilistic Lane Estimation using Basis Curves. | |
CN108090401B (en) | Line detection method and line detection apparatus |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20190809 |
|
CF01 | Termination of patent right due to non-payment of annual fee |