CN108090401A - Line detecting method and line detection device - Google Patents
Line detecting method and line detection device Download PDFInfo
- Publication number
- CN108090401A CN108090401A CN201611037142.0A CN201611037142A CN108090401A CN 108090401 A CN108090401 A CN 108090401A CN 201611037142 A CN201611037142 A CN 201611037142A CN 108090401 A CN108090401 A CN 108090401A
- Authority
- CN
- China
- Prior art keywords
- line
- model
- feature
- extraction
- initialization
- 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
- 238000000034 method Methods 0.000 title claims abstract description 74
- 238000001514 detection method Methods 0.000 title claims abstract description 63
- 238000000605 extraction Methods 0.000 claims abstract description 50
- 238000013507 mapping Methods 0.000 claims description 14
- 230000006870 function Effects 0.000 claims description 13
- 238000003860 storage Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 4
- 230000007423 decrease Effects 0.000 claims description 4
- 241000208340 Araliaceae Species 0.000 claims 2
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims 2
- 235000003140 Panax quinquefolius Nutrition 0.000 claims 2
- 235000008434 ginseng Nutrition 0.000 claims 2
- 238000012545 processing Methods 0.000 abstract description 36
- 239000003550 marker Substances 0.000 abstract description 22
- 238000010586 diagram Methods 0.000 description 32
- 230000008569 process Effects 0.000 description 21
- 230000008901 benefit Effects 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 6
- 241000270295 Serpentes Species 0.000 description 3
- 230000009471 action Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- IYLGZMTXKJYONK-ACLXAEORSA-N (12s,15r)-15-hydroxy-11,16-dioxo-15,20-dihydrosenecionan-12-yl acetate Chemical compound O1C(=O)[C@](CC)(O)C[C@@H](C)[C@](C)(OC(C)=O)C(=O)OCC2=CCN3[C@H]2[C@H]1CC3 IYLGZMTXKJYONK-ACLXAEORSA-N 0.000 description 2
- 101100517651 Caenorhabditis elegans num-1 gene Proteins 0.000 description 2
- 238000003708 edge detection Methods 0.000 description 2
- 238000012805 post-processing Methods 0.000 description 2
- IYLGZMTXKJYONK-UHFFFAOYSA-N ruwenine Natural products O1C(=O)C(CC)(O)CC(C)C(C)(OC(C)=O)C(=O)OCC2=CCN3C2C1CC3 IYLGZMTXKJYONK-UHFFFAOYSA-N 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 241000283070 Equus zebra Species 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/588—Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Traffic Control Systems (AREA)
Abstract
This disclosure relates to a kind of line detecting method and line detection device based on model.The line detecting method includes:Line feature is extracted from the current frame image of input;The line feature based on extraction, performs the initialization of line model;The line model of the line feature and initialization based on extraction, updates the line model;And according to the newer line model, determine the line detected.According to the line detecting method of the disclosure and line detection device, influence of the noise for lane detection can be overcome, and time and processing expense are significantly saved compared with the traditional detection method for detecting lane line and pavement marker respectively.
Description
Technical field
This disclosure relates to image processing field, more specifically, this disclosure relates to a kind of line detecting method based on model and
Line detection device.
Background technology
Line detection technique has a wide range of applications in image processing field, for example, lane detection is line detection technique
One important application.
When performing lane detection, the pavement marker (arrow, zebra stripes etc.) on road surface will influence lane line inspection
Survey accuracy.In order to overcome influence of the pavement marker for lane detection, current solution include pretreatment mode and
Post processing mode.In pretreatment mode, pavement marker detection and identification be with the completely self-contained process of lane detection, so
Two independent image processing procedures cause the time and handle expense increase.In post processing mode, based on pavement marker with
Difference between lane line removes pavement marker from image to be detected, such image processing process it is relatively difficult and easily by
The influence of noise in image.
Accordingly, it is desirable to provide a kind of more robust and efficient line detecting method and line detection device, can overcome and make an uproar
Influence of the sound for lane detection, and compared with the traditional detection method of detection lane line and pavement marker respectively significantly
Save time and processing expense.
The content of the invention
In view of the above problems, the disclosure provides a kind of line directly detected based on the track line model for removing pavement marker
Detection method and line detection device.
According to one embodiment of the disclosure, a kind of line detecting method is provided, including:It is carried from the current frame image of input
Line taking feature;The line feature based on extraction, performs the initialization of line model;The line feature based on extraction and initial
The line model changed, updates the line model;And according to the newer line model, determine the line detected.
In addition, according to the line detecting method of one embodiment of the disclosure, wherein the current frame image of the input includes
Anaglyph and gray level image, the line feature includes characteristic point and Eigenvector, and the Eigenvector is by the feature
Point fitting obtains.
In addition, according to the line detecting method of one embodiment of the disclosure, wherein the line feature based on extraction,
Performing the initialization of line model includes:Obtain the line feature of the extraction;Obtain predetermined line model;It randomly selects described
Line feature calculates the model parameter, support points and cost function of the predetermined line model;And it determines to meet predetermined condition
And with the line model for making support points maximum and the model parameter of Least-cost, the line model as initialization.
In addition, according to the line detecting method of one embodiment of the disclosure, wherein described determine to meet the predetermined condition
Including:Determine that the every line based on the line model met different zones is mutually matched condition;And it determines based on the line
Each line of model meets width restrictive condition.
In addition, according to the line detecting method of one embodiment of the disclosure, wherein described determine based on the line model
The condition that is mutually matched that every line meets different zones includes:The every line based on the line model is divided into multiple regions;Meter
The feature that calculating has in each region of the every line in the multiple region is counted out;For the every line, determine
The feature count out more than predetermined characteristic count out threshold value region number of regions;And determine that the number of regions is more than
What the line of presumptive area quantity threshold met different zones is mutually matched condition.
In addition, according to the line detecting method of one embodiment of the disclosure, wherein described determine based on the line model
Each line meets width restrictive condition and includes:Inverse perspective mapping is carried out to the characteristic image being made of the line feature extracted,
To obtain inverse perspective mapping image;For the width of every line computation in the inverse perspective mapping image to predetermined characteristic point,
To obtain width histogram;Determine that the line with the peak value more than predetermined peak value threshold value meets width restrictive condition.
In addition, according to the line detecting method of one embodiment of the disclosure, wherein the line feature based on extraction
And the line model of initialization, updating the line model includes:The line feature based on extraction, perform gradient decline or
Person's gauss-newton method updates the model parameter of the line model.
According to another embodiment of the disclosure, a kind of line detection device is provided, including:Feature extraction unit, configuration
To extract line feature from the current frame image of input;Initialization unit is configured to the line feature of extraction, performs line mould
The initialization of type;Updating block is configured to the line feature of extraction and the line model of initialization, described in update
Line model;And detection unit, it is configured to, according to the newer line model, determine the line detected.
In addition, according to the line detection device of another embodiment of the disclosure, wherein the initialization unit is further matched somebody with somebody
It is set to the line feature for obtaining the extraction;Obtain predetermined line model;The line feature is randomly selected, is calculated described predetermined
Line model model parameter, support points and cost function;And it determines to meet predetermined condition and having makes support count
Maximum and the model parameter of Least-cost line model, the line model as initialization.
In addition, according to the line detection device of another embodiment of the disclosure, wherein the initialization unit is further matched somebody with somebody
It is set to and determines that the every line based on the line model met different zones is mutually matched condition;And it determines based on the line mould
Each line of type meets width restrictive condition.
Line detecting method and line detection device in accordance with an embodiment of the present disclosure, based on the lane line for removing pavement marker
Model, which performs, directly to be detected.More specifically, in the initialization procedure for the track line model of detection, by using true vehicle
The specified conditions that diatom should meet calculate lane line model parameter, and are performed using the track line model obtained and directly examined
It surveys, so as to remove influence of the noise for lane detection of such as pavement marker, and compared with detecting lane line respectively
Time and processing expense are significantly saved with the traditional detection method of pavement marker.
It is to be understood that foregoing general description and following detailed description are both illustrative, and it is intended to
In the further explanation for providing claimed technology.
Description of the drawings
The embodiment of the present invention is described in more detail in conjunction with the accompanying drawings, the above and other purposes of the present invention,
Feature and advantage will be apparent.Attached drawing is used for providing further understanding the embodiment of the present invention, and forms explanation
A part for book for explaining the present invention together with the embodiment of the present invention, is not construed as limiting the invention.In the accompanying drawings,
Identical reference number typically represents same parts or step.
Fig. 1 is the flow chart for illustrating line detecting method in accordance with an embodiment of the present disclosure;
Fig. 2 is the flow of the line feature extraction processing in the line detecting method of further diagram in accordance with an embodiment of the present disclosure
Figure;
Fig. 3 is the schematic diagram of the line feature of diagram extraction;
Fig. 4 is the stream of the line model initialization process in the line detecting method of further diagram in accordance with an embodiment of the present disclosure
Cheng Tu;
Fig. 5 is the predetermined one exemplary schematic diagram of line model of diagram;
Fig. 6 is in the line model initialization process of further diagram in accordance with an embodiment of the present disclosure at far and near Region Matching
The flow chart of reason;
Fig. 7 is the schematic diagram of the far and near Region Matching subprocessing of diagram;
Fig. 8 is that the line model initialization process center line tolerance system of further diagram in accordance with an embodiment of the present disclosure is handled
Flow chart;
Fig. 9 A and 9B are the schematic diagrames for illustrating inverse perspective mapping in line width limitation subprocessing;
Figure 10 is to illustrate the schematic diagram that width histogram filters in line width limitation subprocessing;
Figure 11 is the functional configuration block diagram for illustrating line detection device in accordance with an embodiment of the present disclosure;
Figure 12 is the general hardware block diagram for illustrating line detecting system in accordance with an embodiment of the present disclosure;And
Figure 13 is the configuration block diagram for illustrating line detection device in accordance with an embodiment of the present disclosure.
Specific embodiment
In order to enable the object, technical solutions and advantages of the present invention become apparent, root is described in detail below with reference to accompanying drawings
According to example embodiments of the present invention.Obviously, described embodiment is only the part of the embodiment rather than this hair of the present invention
Bright whole embodiments, it should be appreciated that the limitation of the invention from example embodiment described herein.Described in the disclosure
The embodiment of the present invention, those skilled in the art's obtained all other embodiment in the case where not making the creative labor
It should all fall under the scope of the present invention.
Hereinafter, preferred embodiment of the present disclosure will be described in detail with reference to the attached drawings.
Fig. 1 is the flow chart for illustrating line detecting method in accordance with an embodiment of the present disclosure.In accordance with an embodiment of the present disclosure
Line detecting method comprises the following steps.
In step S101, line feature is extracted from the current frame image of input.
Specifically, in one embodiment of the disclosure, the line feature of the line to be detected can any can characterize this
The feature of line.For example, it can include but is not limited to the features such as the color of line, gray scale, shape, edge, parallax or these features
Any combination.Any appropriate mode may be employed to extract the line feature for the line for conforming to detection.Hereinafter, will further join
Line feature extraction processing is described according to Fig. 2 and Fig. 3.Hereafter, processing enters step S102.
In step s 102, the line feature based on extraction, performs the initialization of line model.
Specifically, in one embodiment of the disclosure, by taking the line to be detected is lane line as an example, correspondingly, the line
Model is track line model.There are a variety of existing track line models, including linear model, isolated point model, parabola model and
Its extension, hyperbolic model, convolution line model, Spline Model, Snake models, 3D models etc..In an implementation of the disclosure
In example, one of a variety of existing track line models are selected, perform the model initialization based on consistent (RANSAC) algorithm of random sampling.
Specifically, for the predetermined track line model of selection, the line feature is randomly selected, and calculates the mould of the predetermined line model
Shape parameter, support points and cost function, finally meeting predetermined condition surely and having makes support count maximum and cost most
The line model of small model parameter, the line model as initialization.In the specific embodiment of the disclosure, the predetermined item
Part includes but not limited to far and near Region Matching condition and line width restrictive condition, and such as road is eliminated by predetermined condition limitation
Influence of the noise of mark for line model.Hereinafter, will be referred to further Fig. 4 to Figure 10 describe the initialization process of line model with
And far and near Region Matching subprocessing and line width limitation subprocessing in line model initialization process.Hereafter, processing enters step S103.
In step s 103, the line feature based on extraction and the line model of initialization, update the line mould
Type.
Specifically, in one embodiment of the disclosure, the line feature based on extraction performs gradient decline or high
This-Newton method updates the model parameter of the line model.Hereafter, processing enters step S104.
In step S104, according to the newer line model, the line detected is determined.
Specifically, the updated line model obtained by above-mentioned processing describes each line in current frame image, because
This can directly obtain the line to be detected in current frame image from the updated line model.
More than, outline line detecting method in accordance with an embodiment of the present disclosure with reference to Fig. 1.With reference to Fig. 1 descriptions according to this
The line detecting method of disclosed embodiment realizes the line model based on the noise for eliminating such as pavement marker and directly detects.With
Under, each processing step in the line detecting method of each attached drawing detailed description in accordance with an embodiment of the present disclosure will be referred to further.
Fig. 2 is the flow of the line feature extraction processing in the line detecting method of further diagram in accordance with an embodiment of the present disclosure
Figure;Fig. 3 is the schematic diagram of the line feature for the line feature extraction processing extraction that diagram passes through Fig. 2.
As shown in Fig. 2, the line feature extraction processing in line detecting method in accordance with an embodiment of the present disclosure includes following step
Suddenly.
In step s 201, the current frame image of input is obtained.In one embodiment of the disclosure, the present frame of input
Image includes anaglyph and gray level image.Hereafter, processing enters step S202.
In step S202, characteristic point and Eigenvector are extracted.In one embodiment of the disclosure, feature can be used
Point or Eigenvector are exported as line feature.As an example, in the current frame image that can be obtained in shooting, by such as
The directly detection of Hough transformation isoline detection method conforms to the line segment of the line feature of the line of detection.As another example, may be used
Be primarily based on the line to be detected line feature, feature detected in current frame image by detection method corresponding with this feature
Point (if for example, the feature for the line to be detected be edge feature, can characteristic point be detected by edge detection method), so
The characteristic point obtained afterwards using detection is fitted to obtain the line segment.Hereafter, processing enters step S203.
In step S203, the line feature of extraction is exported.In one embodiment of the disclosure, the characteristic curve of extraction is exported
Initialization and update for the line model that then will be described in.
As shown in figure 3, in the line feature 301 of extraction is handled by line feature extraction shown in Fig. 2, including representing track
The feature of line and the noise characteristic for representing such as pavement marker.
Hereinafter, the line feature 301 that extraction is handled by line feature extraction shown in Fig. 2 is utilized with reference to Fig. 4 to Figure 10 descriptions
Line model initialization process.
Fig. 4 is the stream of the line model initialization process in the line detecting method of further diagram in accordance with an embodiment of the present disclosure
Cheng Tu.As shown in figure 4, line model initialization process in line detecting method in accordance with an embodiment of the present disclosure includes following step
Suddenly.
In step S121, the line feature of the extraction is obtained.That is, obtain through line feature extraction shown in Fig. 2
Handle the line feature 301 of extraction.Hereafter, processing enters step S122.
In step S122, predetermined line model is obtained.As described above, predetermined line model can be existing lane line mould
Type, including linear model, isolated point model, parabola model and its extension, hyperbolic model, convolution line model, Spline Model,
Snake models, 3D models etc..
Fig. 5 is the predetermined one exemplary schematic diagram of line model of diagram.Predetermined line model shown in Fig. 5 is polynomial module
Formula.
Specifically, as shown in Figure 5, it is assumed that the center line of current lane is Lmid, the left-hand lane line of current lane is L-1, it is right
Side lane line is L1, the left-hand lane line in n-th of left side track is L-n, the left-hand lane line in n-th of right side track is Ln;Deserve
The lane width in preceding track is w0, the lane width in n-th of the track in left side is w-n, the lane width in n-th of the track in right side
It is wn;Vp represents the ordinate of the position of track heading line off;F is the focal length of camera lens, and H is the height of camera position.
LmidUsing the polynomial repressentation as shown in equation (1):
Then track line model is represented by:
Wherein,
Referring back to Fig. 4, after predetermined line model is obtained, processing enters step S123.
In step S123, the line feature is randomly selected, calculates model parameter, the supporting point of the predetermined line model
Number and cost function.That is, in one embodiment of the disclosure, perform based on consistent (RANSAC) algorithm of random sampling
Model initialization.Hereafter, processing enters step S124.
In step S124, determine to meet predetermined condition and with the model for making support points maximum and Least-cost
The line model of parameter, the line model as initialization.In one embodiment of the disclosure, the predetermined condition include but
Far and near Region Matching condition and line width restrictive condition are not limited to, making an uproar for such as pavement marker is eliminated by predetermined condition limitation
Influence of the sound for line model.Hereinafter, will be described with reference to Fig. 6 and Fig. 7 in line model initialization process at far and near Region Matching
Reason, and the system processing of line model initialization process center line tolerance is described with reference to Fig. 8 to Figure 10.
Fig. 6 is in the line model initialization process of further diagram in accordance with an embodiment of the present disclosure at far and near Region Matching
The flow chart of reason;Fig. 7 is the schematic diagram of the far and near Region Matching subprocessing of diagram.
As shown in fig. 6, far and near Region Matching subprocessing bag in line model initialization process in accordance with an embodiment of the present disclosure
Include following steps.
In step S1241, the every line based on the line model is divided into multiple regions.
Specifically, as shown in fig. 7, dividing 3 in vertical scope from as far as near order with range image capture device
Region:R1, R2 and R3.It is easily understood that vertical scope can also be divided into other several purpose regions.The one of the disclosure
In a embodiment, the principle of region division is that the regional extent of distant place is smaller and regional extent nearby is relatively large.
Referring back to Fig. 6, after multiple regions are obtained, processing enters step S1242.
In step S1242, the characteristic point having in each region of the every line in the multiple region is calculated
Number (for example, n1, n2 ...).In one embodiment of the disclosure, depending on the type of line feature extracted, the spy
Sign point number can be the number of real characteristic point or the length of Eigenvector.Hereafter, processing enters step
S1243。
In step S1243, for the every line, determine that the feature is counted out and count out threshold value more than predetermined characteristic
Region number of regions.In one embodiment of the disclosure, for each region predetermined characteristic is set to count out threshold value (example
Such as, t1, t2 ...), if a line in a region characteristic point number (for example, n1) be more than the region predetermined characteristic
Threshold value of counting out (for example, t1), that is, meet n1>T1, then the region meet predetermined characteristic for one of this line and count out threshold value
Pass through region.Hereafter, number of regions s by region of the every line in all areas is counted.Hereafter, processing enters step
S1244。
In step S1244, determine that the number of regions meets different zones more than the line of presumptive area quantity threshold
It is mutually matched condition.In one embodiment of the disclosure, presumptive area quantity threshold f (n) is pre-set, such as f (n) can be with
It is set to the 1/2 of region total number.For every line, if its number of regions by region in all areas is pre- more than this
Determine number of regions threshold value f (n), that is, meet s>F (n), it is determined that this bar line met different zones is mutually matched condition.
As shown in fig. 7, in the case where region total number is 3, presumptive area quantity threshold f (n) is 3/2.The leftmost side and
The line (real solid line lane line) of the rightmost side is since all continued presence, i.e. three regions belong to the logical of its in three regions
Cross region, then it is 3 by the number of regions s in region, meets s>f(n).That is, the line of the leftmost side and the rightmost side is confirmed as
Meet different zones is mutually matched condition.
Similarly, intermediate line (real dotted line lane line) is since there are two pass through region (R2 to tool in three regions
And R3), then it is 2 by the number of regions s in region, equally meets s>f(n).That is, intermediate line is also determined as meeting not
Condition is mutually matched with region.
In addition, the line (pavement marker) of intermediate line both sides is since only there are one pass through region to tool in three regions
(R3), then it is 1 by the number of regions s in region, is unsatisfactory for s>f(n).That is, the line (pavement marker) of intermediate line both sides
Be confirmed as being unsatisfactory for different zones is mutually matched condition.
As described above, far and near Region Matching subprocessing in line model initialization process is described by referring to Fig. 6 and Fig. 7, into
Work(is mutually matched condition using be determined as being unsatisfactory for different zones as the pavement marker of noise.
Fig. 8 is that the line model initialization process center line tolerance system of further diagram in accordance with an embodiment of the present disclosure is handled
Flow chart;Fig. 9 A and 9B are the schematic diagrames for illustrating inverse perspective mapping in line width limitation subprocessing;Figure 10 is diagram line width siding stopping
The schematic diagram that width histogram filters in processing.
As shown in figure 8, in accordance with an embodiment of the present disclosure line model initialization process center line tolerance system processing include with
Lower step.
In step S1245, inverse perspective mapping is carried out to the characteristic image being made of the line feature extracted, to obtain
Inverse perspective mapping image.
Specifically, Fig. 9 A show original image, wherein the line that extraction is handled by line feature extraction shown in Fig. 2 is special
Sign.Fig. 9 B, which are shown, performs original image the inverse perspective mapping image that inverse perspective mapping obtains.
Referring back to Fig. 8, after inverse perspective mapping image is obtained, processing enters step S1246.
In step S1246, for the width of every line computation in the inverse perspective mapping image to predetermined characteristic point,
To obtain width histogram.
Specifically, Figure 10 shows the width histogram.In histogram shown in Fig. 10, abscissa represents width,
Ordinate represents the existing feature under specific width and counts out.
Referring back to Fig. 8, after width histogram is obtained, processing enters step S1247.
In step S1427, determine that the line with the peak value more than predetermined peak value threshold value meets width restrictive condition.
Specifically, as shown in Figure 10, the peak-peak in width histogram is MaxNum, other peak values are Num1, Num2
And Num3.For example, it is that f (MaxNum) is the 1/3 of peak-peak to set predetermined peak value threshold value, i.e. f (MaxNum)=MaxNum/3,
Num1/Num2 is determined>F (MaxNum) meets width restrictive condition, and Num3<F (MaxNum) is unsatisfactory for width limitation item
Part.
As described above, it is handled by referring to the line model initialization process center line tolerance system of Fig. 8 to Figure 10 descriptions, equally
Success will be determined as being unsatisfactory for line width restrictive condition as the pavement marker of noise.
More than, it describes and exemplary line detecting method is implemented according to the disclosure.Hereinafter, profit is further described with reference to the accompanying drawings
With the line detection device of the line detecting method and line detecting system.
Figure 11 is the functional configuration block diagram for illustrating line detection device in accordance with an embodiment of the present disclosure.As shown in figure 11, line
Detection device 10 can include feature extraction unit 101, initialization unit 102, updating block 103 and detection unit 104, described
Unit can perform each step/function of the line detecting method above in conjunction with Fig. 1 descriptions respectively.Therefore, it is only right below
The major function of each unit of the line detection device 10 is described, and omits the detail content having been described above.
The feature extraction unit 101 is configured to extract line feature from the current frame image of input.Specifically, to be detected
The line feature of line can be any feature that can characterize the line.For example, its can include but is not limited to the color of line, gray scale,
Any combination of the features such as shape, edge, parallax or these features.Any appropriate mode may be employed to conform to extract
The line feature of the line of detection.As an example, the feature extraction unit 101 can be in the obtained current frame image of shooting
In, pass through the line segment of the direct line feature for detecting the line for conforming to detection of such as Hough transformation isoline detection method.As another
One example, the feature extraction unit 101 can be primarily based on the line feature for the line to be detected, by corresponding with this feature
Detection method detected in current frame image characteristic point (if for example, the feature for the line to be detected be edge feature, can lead to
Edge detection method is crossed to detect characteristic point), the characteristic point then obtained using detection is fitted to obtain the line segment.
The initialization unit 102 is configured to the line feature of extraction, performs the initialization of line model.Specifically
Ground, the initialization unit 102 for selection predetermined track line model (include but not limited to linear model, isolated point model,
Parabola model and its extension, hyperbolic model, convolution line model, Spline Model, Snake models, 3D models etc.), at random
The line feature is chosen, and calculates the model parameter of the predetermined line model, support points and cost function, final fixed satisfaction
Predetermined condition and with making support points maximum and the line model of the model parameter of Least-cost, as described in initialization
Line model.In the specific embodiment of the disclosure, the predetermined condition that the initialization unit 102 utilizes includes but not limited to
Far and near Region Matching condition and line width restrictive condition, by predetermined condition limitation eliminate the noise of such as pavement marker for
The influence of line model.
The updating block 103 is configured to the line feature of extraction and the line model of initialization, update
The line model.Specifically, the line feature of the updating block 103 based on extraction performs gradient decline or Gauss-ox
Method of pausing updates the model parameter of the line model.
The detection unit 104 is configured to, according to the newer line model, determine the line detected.
Figure 12 is the general hardware block diagram for illustrating line detecting system in accordance with an embodiment of the present disclosure.As shown in figure 12, line
Detecting system 20 can include:Input equipment 201, for from the related image of external input or information, such as video camera shooting
Depth map, gray-scale map (cromogram) etc., the input equipment 201 for example can be keyboard, mouse, video camera etc.;Processing equipment
202, it, should for implementing the above-mentioned line detecting method according to the embodiment of the present disclosure or being embodied as above-mentioned line detection device
Processing equipment can be that by any device with processing capacity of above-mentioned function, for example, its can be designed for into
General processor, digital signal processor (DSP), ASIC, the field programmable gate array signal of row function described herein
(FPGA) or other programmable logic device (PLD), discrete gate or transistor logics, discrete nextport hardware component NextPort or its arbitrary group
It closes;Storage device 203, for stored in a manner of volatile and nonvolatile such as depth map involved by above-mentioned line detection process,
Gray-scale map (cromogram), various threshold values, the line model pre-established, the line segment extracted, newer line model etc., the storage
Equipment for example can be each of random access memory (RAM), read-only memory (ROM), hard disk or semiconductor memory etc.
The volatile and nonvolatile property memory of kind;And output equipment 204, obtained by implementing above-mentioned line detection process to outside output
As a result, the line for example detected, the output equipment for example can be display, printer etc..
Figure 13 is the configuration block diagram for illustrating line detection device in accordance with an embodiment of the present disclosure.As shown in figure 13, according to this
The line detection device 130 of disclosed embodiment includes memory 1301 and processor 1302.It is stored on the memory 1301
Computer program instructions, the computer program instructions are performed when being run by processor 1302 above with reference to Fig. 1 to Figure 10
The line detecting method of description.
More than, describe line detecting method in accordance with an embodiment of the present disclosure and line detection device with reference to the accompanying drawings.The line
Detection method and line detection device are performed based on the track line model for removing pavement marker directly to be detected.More specifically, for
In the initialization procedure of the track line model of detection, the specified conditions that should meet by using true lane line calculate lane line
Model parameter, and performed using the track line model obtained and directly detected, so as to remove the noise pair of such as pavement marker
It is significantly saved in the influence of lane detection, and compared with the traditional detection method for detecting lane line and pavement marker respectively
Time and processing expense.
The basic principle of the present invention is described above in association with specific embodiment, however, it is desirable to, it is noted that in the disclosure
The advantages of referring to, advantage, effect etc. are only exemplary rather than limiting, it is impossible to which it is of the invention to think these advantages, advantage, effect etc.
Each embodiment is prerequisite.In addition, detail disclosed above is merely to exemplary effect and the work readily appreciated
With, and it is unrestricted, above-mentioned details is not intended to limit the present invention as that must be realized using above-mentioned concrete details.
Device, device, equipment, the block diagram of system involved in the disclosure only as illustrative example and are not intended to
It is required that or hint must be attached in a manner that box illustrates, arrange, configure.As those skilled in the art will appreciate that
, it can connect, arrange by any way, configuring these devices, device, equipment, system.Such as " comprising ", "comprising", " tool
" etc. word be open vocabulary, refer to " including but not limited to ", and can be used interchangeably with it.Vocabulary used herein above
"or" and " and " refer to vocabulary "and/or", and can be used interchangeably with it, unless it is not such that context, which is explicitly indicated,.Here made
Vocabulary " such as " refers to phrase " such as, but not limited to ", and can be used interchangeably with it.
Step flow chart and above method in the disclosure describe only as illustrative example and are not intended to require
Or imply the step of must carrying out each embodiment according to the order that provides, some steps can parallel, it is independent of one another or according to
Other appropriate orders perform.In addition, such as " thereafter ", " then ", the word of " following " etc. be not intended to limit step
Sequentially;These words are only used for the description that guiding reader reads over these methods.
In addition, as used herein, with the item of " at least one " beginnings enumerate the middle "or" used indicate it is separated
It enumerates, so that enumerating for such as " A, B or C's being at least one " means A or B or C or AB or AC or BC or ABC (i.e. A and B
And C).In addition, wording " exemplary " does not mean that the example of description is preferred or more preferable than other examples.
It may also be noted that in apparatus and method of the present invention, each component or each step are can to decompose and/or again
Combination nova.These decompose and/or reconfigure the equivalent scheme that should be regarded as the present invention.
For those of ordinary skill in the art, it is to be understood that whole or any portions of disclosed method and device
Point, can in any computing device (including processor, storage medium etc.) or the network of computing device, with hardware, firmware,
Software or combination thereof are realized.The hardware can be using being designed to carry out the logical of function described herein
With processor, digital signal processor (DSP), ASIC, field programmable gate array signal (FPGA) or other programmable logic devices
Part (PLD), discrete gate or transistor logic, discrete nextport hardware component NextPort or its any combination.General processor can be micro- place
Device is managed, but as an alternative, the processor can be any commercially available processor, controller, microcontroller or shape
State machine.Processor is also implemented as the combination of computing device, such as the combination of DSP and microprocessor, multi-microprocessor, with
The one or more microprocessors of DSP core cooperation or any other such configuration.The software can reside in any form
Computer-readable tangible media in.By example rather than limitation, such computer-readable tangible storage is situated between
Matter can include RAM, ROM, EEPROM, CD-ROM or other optical disc storages, disk storage or other magnetic memory devices or can
It can be accessed for carrying or the desired program code of store instruction or data structure form and by computer any
Other tangible mediums.As used herein, disk include compact disk (CD), laser disk, CD, digital versatile disc (DVD), floppy disk and
Blu-ray disc.
Intelligent control technology disclosed by the invention can also be by running a program or one on any computing device
Program is organized to realize.The computing device can be well known fexible unit.Intellectual technology disclosed in this invention can also be only
Only realized or comprising the program product for realizing the method either program code of device by being stored with this by providing
The arbitrary storage medium of the program product of sample is realized.
The technology instructed defined by the appended claims can not departed from and carried out to the various of technology described herein
Change, replace and change.In addition, the scope of the claim of the disclosure is not limited to process described above, machine, manufacture, thing
Composition, means, method and the specific aspect of action of part.It can be essentially identical using being carried out to corresponding aspect described herein
Function either realize essentially identical result there is currently or to be developed later processing, machine, manufacture, event group
Into, means, method or action.Thus, appended claims include such processing within its scope, machine, manufacture, event
Composition, means, method or action.
The above description of disclosed aspect is provided so that any person skilled in the art can make or use this
Invention.Various modifications in terms of these are readily apparent to those skilled in the art, and are defined herein
General Principle can be applied to other aspect without departing from the scope of the present invention.Therefore, the present invention is not intended to be limited to
Aspect shown in this, but according to the widest range consistent with principle disclosed herein and novel feature.
In order to which purpose of illustration and description has been presented for above description.In addition, this description is not intended to the reality of the present invention
It applies example and is restricted to form disclosed herein.Although already discussed above multiple exemplary aspects and embodiment, this field skill
Art personnel will be recognized that its some modifications, modification, change, addition and sub-portfolio.
Claims (11)
1. a kind of line detecting method, including:
Line feature is extracted from the current frame image of input;
The line feature based on extraction, performs the initialization of line model;
The line model of the line feature and initialization based on extraction, updates the line model;And
According to the newer line model, the line detected is determined.
2. line detecting method as described in claim 1, wherein the current frame image of the input includes anaglyph and gray scale
Image, the line feature includes characteristic point and Eigenvector, and the Eigenvector is fitted by the characteristic point and obtained.
3. line detecting method as claimed in claim 2, wherein the line feature based on extraction, performs the first of line model
Beginningization includes:
Obtain the line feature of the extraction;
Obtain predetermined line model;
The line feature is randomly selected, calculates the model parameter, support points and cost function of the predetermined line model;And
It determines to meet predetermined condition and there is the line model for making support points maximum and the model parameter of Least-cost, as
The line model of initialization.
4. line detecting method as claimed in claim 3, wherein described determine that meeting the predetermined condition includes:
Determine that the every line based on the line model met different zones is mutually matched condition;And
Determine that each line based on the line model meets width restrictive condition.
5. line detecting method as claimed in claim 4, wherein described determine that the every line based on the line model meets difference
The condition that is mutually matched in region includes:
The every line based on the line model is divided into multiple regions;
The feature that calculating has in each region of the every line in the multiple region is counted out;
For the every line, determine the feature count out more than predetermined characteristic count out threshold value region number of regions;
And
Determine that the number of regions was more than that the line of presumptive area quantity threshold meets different zones is mutually matched condition.
6. line detecting method as claimed in claim 4, wherein described determine that each line based on the line model meets width
Restrictive condition includes:
Inverse perspective mapping is carried out to the characteristic image being made of the line feature extracted, to obtain inverse perspective mapping image;
For the width of every line computation in the inverse perspective mapping image to predetermined characteristic point, to obtain width histogram;
Determine that the line with the peak value more than predetermined peak value threshold value meets width restrictive condition.
7. line detecting method as described in claim 1, wherein the institute of the line feature and initialization based on extraction
Line model is stated, updating the line model includes:
The line feature based on extraction, performs gradient decline or gauss-newton method updates the model ginseng of the line model
Number.
8. a kind of line detection device, including:
Feature extraction unit is configured to extract line feature from the current frame image of input;
Initialization unit is configured to the line feature of extraction, performs the initialization of line model;
Updating block is configured to the line feature of extraction and the line model of initialization, updates the line model;
And
Detection unit is configured to, according to the newer line model, determine the line detected.
9. line detection device as claimed in claim 8, wherein the initialization unit is further configured to obtain the extraction
The line feature;Obtain predetermined line model;The line feature is randomly selected, calculates the model ginseng of the predetermined line model
Number, support points and cost function;And it determines to meet predetermined condition and have to make support points maximum and Least-cost
Model parameter line model, the line model as initialization.
10. line detection device as claimed in claim 9, wherein the initialization unit is further configured to determine based on described
What every line of line model met different zones is mutually matched condition;And determine that each line based on the line model meets width
Spend restrictive condition.
11. a kind of line detection device, including:
Processor;And
Memory is configured to storage computer program instructions;
Wherein, when the computer program instructions are run by the processor, perform such as any one institute of claim 1 to 7
The line detecting method stated.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611037142.0A CN108090401B (en) | 2016-11-23 | 2016-11-23 | Line detection method and line detection apparatus |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611037142.0A CN108090401B (en) | 2016-11-23 | 2016-11-23 | Line detection method and line detection apparatus |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108090401A true CN108090401A (en) | 2018-05-29 |
CN108090401B CN108090401B (en) | 2021-12-14 |
Family
ID=62168609
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611037142.0A Expired - Fee Related CN108090401B (en) | 2016-11-23 | 2016-11-23 | Line detection method and line detection apparatus |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108090401B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110210451A (en) * | 2019-06-13 | 2019-09-06 | 重庆邮电大学 | A kind of zebra line detecting method |
WO2021063228A1 (en) * | 2019-09-30 | 2021-04-08 | 上海商汤临港智能科技有限公司 | Dashed lane line detection method and device, and electronic apparatus |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102592114A (en) * | 2011-12-26 | 2012-07-18 | 河南工业大学 | Method for extracting and recognizing lane line features of complex road conditions |
CN102663744A (en) * | 2012-03-22 | 2012-09-12 | 杭州电子科技大学 | Complex road detection method under gradient point pair constraint |
CN103617412A (en) * | 2013-10-31 | 2014-03-05 | 电子科技大学 | Real-time lane line detection method |
CN104408460A (en) * | 2014-09-17 | 2015-03-11 | 电子科技大学 | A lane line detecting and tracking and detecting method |
CN104599502A (en) * | 2015-02-13 | 2015-05-06 | 重庆邮电大学 | Method for traffic flow statistics based on video monitoring |
CN105224909A (en) * | 2015-08-19 | 2016-01-06 | 奇瑞汽车股份有限公司 | Lane line confirmation method in lane detection system |
CN105930800A (en) * | 2016-04-21 | 2016-09-07 | 北京智芯原动科技有限公司 | Lane line detection method and device |
-
2016
- 2016-11-23 CN CN201611037142.0A patent/CN108090401B/en not_active Expired - Fee Related
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102592114A (en) * | 2011-12-26 | 2012-07-18 | 河南工业大学 | Method for extracting and recognizing lane line features of complex road conditions |
CN102663744A (en) * | 2012-03-22 | 2012-09-12 | 杭州电子科技大学 | Complex road detection method under gradient point pair constraint |
CN103617412A (en) * | 2013-10-31 | 2014-03-05 | 电子科技大学 | Real-time lane line detection method |
CN104408460A (en) * | 2014-09-17 | 2015-03-11 | 电子科技大学 | A lane line detecting and tracking and detecting method |
CN104599502A (en) * | 2015-02-13 | 2015-05-06 | 重庆邮电大学 | Method for traffic flow statistics based on video monitoring |
CN105224909A (en) * | 2015-08-19 | 2016-01-06 | 奇瑞汽车股份有限公司 | Lane line confirmation method in lane detection system |
CN105930800A (en) * | 2016-04-21 | 2016-09-07 | 北京智芯原动科技有限公司 | Lane line detection method and device |
Non-Patent Citations (3)
Title |
---|
MRINAL HALOI ET AL: "A robust lane detection and departure warning system", 《2015 IEEE INTELLIGENT VEHICLES SYMPOSIUM(IV)》 * |
高志峰 等: "一种鲁棒的非平坦路面车道线检测算法", 《北京理工大学学报》 * |
鲁曼 等: "道路区域分割的车道线检测方法", 《智能***学报》 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110210451A (en) * | 2019-06-13 | 2019-09-06 | 重庆邮电大学 | A kind of zebra line detecting method |
CN110210451B (en) * | 2019-06-13 | 2022-07-08 | 重庆邮电大学 | Zebra crossing detection method |
WO2021063228A1 (en) * | 2019-09-30 | 2021-04-08 | 上海商汤临港智能科技有限公司 | Dashed lane line detection method and device, and electronic apparatus |
Also Published As
Publication number | Publication date |
---|---|
CN108090401B (en) | 2021-12-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110348445B (en) | Instance segmentation method fusing void convolution and edge information | |
CN107358596B (en) | Vehicle loss assessment method and device based on image, electronic equipment and system | |
Nurunnabi et al. | Outlier detection and robust normal-curvature estimation in mobile laser scanning 3D point cloud data | |
Ahmadi et al. | Automatic urban building boundary extraction from high resolution aerial images using an innovative model of active contours | |
US9053540B2 (en) | Stereo matching by census transform and support weight cost aggregation | |
WO2017020466A1 (en) | Urban road recognition method, apparatus, storage medium and device based on laser point cloud | |
CN102855459B (en) | For the method and system of the detection validation of particular prospect object | |
CN106887018B (en) | Stereo matching method, controller and system | |
CN105981041A (en) | Facial landmark localization using coarse-to-fine cascaded neural networks | |
CN110659658B (en) | Target detection method and device | |
CN108550166B (en) | Spatial target image matching method | |
CN105374035A (en) | Star sensor star point extraction method under stray light interference | |
JP6700373B2 (en) | Apparatus and method for learning object image packaging for artificial intelligence of video animation | |
CN104021552A (en) | Multi-objective particle swarm parameter optimization method based on graph segmentation process | |
JP2016058085A (en) | Method and device for detecting shielding of object | |
CN116452644A (en) | Three-dimensional point cloud registration method and device based on feature descriptors and storage medium | |
CN108090401A (en) | Line detecting method and line detection device | |
CN110413942B (en) | Lane line equation screening method and screening module thereof | |
JP6561605B2 (en) | Target analysis method and target analysis system | |
CN104680493A (en) | Digital image repairing method based on scale optimization | |
CN104700384A (en) | Display system and display method based on augmented reality technology | |
CN107305688B (en) | Method, device and system for detecting road vanishing point | |
Ananth et al. | A geodesic active contour level set method for image segmentation | |
CN102901490B (en) | Image matching method based on dynamic threshold, and system | |
CN110705330A (en) | Lane line detection method, lane line detection apparatus, and computer-readable storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20211214 |
|
CF01 | Termination of patent right due to non-payment of annual fee |