CN107491756B - Lane direction information recognition methods based on traffic sign and surface mark - Google Patents
Lane direction information recognition methods based on traffic sign and surface mark Download PDFInfo
- Publication number
- CN107491756B CN107491756B CN201710709060.4A CN201710709060A CN107491756B CN 107491756 B CN107491756 B CN 107491756B CN 201710709060 A CN201710709060 A CN 201710709060A CN 107491756 B CN107491756 B CN 107491756B
- Authority
- CN
- China
- Prior art keywords
- lane
- traffic sign
- surface mark
- recognition result
- direction information
- 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.)
- Active
Links
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
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/582—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of traffic signs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- 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)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
- Traffic Control Systems (AREA)
Abstract
The present invention relates to the lane direction information recognition methods based on traffic sign and surface mark, comprising: step 1, the image sequence data of real road is obtained using traverse measurement vehicle；Step 2, the detection for carrying out traffic sign to each image identifies, obtains the direction information of traffic sign in lane；Step 3, the detection for carrying out surface mark to each image identifies, obtains the direction information of surface mark in lane；Step 4, image sequence analysis is carried out to the direction information recognition result of step 2 and step 3, calculates separately out the confidence level of traffic sign Yu surface mark direction information, confidence level is highest as respective recognition result；Step 5, traffic sign turns recognition result and surface mark recognition result in comparison veritification lane, obtains the direction information of the roadway segment.The characteristics of present invention is distributed by surface mark in analysis real road and traffic sign, and mutual affirmation mechanism is used, realize the high-accuracy identification of lane direction information.
Description
Technical field
The invention mainly relates to intelligent transportation fields, especially for the direction information for extracting lane, by detecting traffic mark
Ground turn marking, the two are mutually authenticated in board and lane, Lai Tigao information accuracy.
Background technique
The rapid development of automobile provides convenience for the trip of people's daily life, but also brings various problems, traffic
Congestion, traffic accident occur again and again.Therefore, Traffic Sign Recognition System has obtained extensive concern, and system can be accurate in real time
Road information is passed to driver by ground, effectively driver is helped to make prediction possible danger, to realize that safety is driven
It sails.Since traffic sign type is more, what need to be handled contains much information, and there are a large amount of disturbing factors, such as illumination condition, road
Road situation complexity etc., so that Traffic Sign Images to be identified often have noise jamming, therefore it is required that Traffic Sign Recognition is calculated
Method has higher accuracy.
Currently, have some researchs to improving traffic sign identification accuracy both at home and abroad, but most of researchs all only for
Traffic sign identification, is continuously improved its arithmetic accuracy, also there is the identification of small part research surface mark.Jobson et al. is used
Retinex algorithm eliminates illumination effect, but algorithm is sufficiently complex and experiment effect is general；Teague is proposed using continuous
The orthogonal method away from construction not displacement carries out feature extraction, there are two main classes i.e. Zernike away from Legendre away from but calculating
There are discretization error when this kind of continuous not displacement, will lead to not displacement numerical value change is larger and orthogonal performance decline, cause to know
Other accuracy decline.It displays and proposes the pavement marker recognition methods based on Hu not bending moment and outline projection, Jie dazzles favour and devises base
In the crossing detection algorithm of bipolarity and Fusion Features, proposes the characteristic point searching algorithm based on convolutional filtering and stopped
Only line detects.Lower using the method false detection rate, processing speed is fast, but is directed to symbol algorithm for design, lacks versatility.
But from the point of view of existing algorithm, lack the mechanism to recognition result verifying confirmation.It is according to the actual situation it is found that more in the same direction
The highway in lane has been generally designed point to traveling lane, has its direction information in each lane, and will appear lane in the air
The case where driving direction mark, i.e., there are multiple marks instruction same information on real road.However existing research is not to containing same
Multiple marks of one information are utilized well.
Summary of the invention
For the accuracy for significantly improving traffic sign identification classification, multiple marks containing same information are made full use of, this
A kind of lane direction information recognition methods being mutually authenticated based on traffic sign and surface mark of disclosure of the invention.
The technical scheme is that a kind of lane direction information being mutually authenticated based on traffic sign and surface mark is known
Other method, includes the following steps,
Step 1, the image sequence data of real road are obtained；
Step 2, the detection for carrying out traffic sign to each image identifies, including following sub-step,
Step 2.1, the segmentation of hsv color capacity-threshold is carried out to each image, extracts the area for meeting traffic sign color
Domain；
Step 2.2, in the region for meeting traffic sign color extracted, SHAPE DETECTION is carried out, friendship will be met
The extracted region of logical label shape comes out；
Step 2.3, hough straight-line detection is carried out in the traffic sign regional scope of acquisition, is drawn according to obtained straight line
It is divided into different lanes, a turn marking is corresponding in each lane, extracts the feature vector V1 of each turn marking；
Step 2.4, feature vector V1 is input to support vector machines to identify, obtains the steering of traffic sign in lane
Information；
Step 3, the detection for carrying out surface mark to each image identifies, including following sub-step,
Step 3.1, the processing of gray processing, histogram equalization and scene rebuilding is carried out to each image；
Step 3.2, the lane line in image is extracted, obtains the regional scope in lane, and retain the surface mark between lane；
Step 3.3, within the scope of each lane that step 3.2 obtains, extract the feature of the turn marking in each lane to
Measure V2；
Step 3.4, feature vector V2 is input to support vector machines to identify, obtains the steering of surface mark in lane
Information；
Step 4, image sequence analysis is carried out to the direction information recognition result of step 2 and step 3, calculates separately out traffic
The confidence level of label and surface mark direction information is labeled as unreliable information if obtained confidence level is respectively less than specified value,
It remains to recheck, if obtained confidence level meets specified value, corresponding output has the recognition result of maximum confidence；
Step 5, traffic sign recognition result and surface mark recognition result in lane are veritified in comparison, obtain the roadway segment
Direction information, implementation is as follows,
First determine whether in the same lane whether and meanwhile include traffic sign and surface mark recognition result, if so,
According to priori knowledge, the weight of traffic sign and surface mark recognition result is determined respectively, then executes (1) and (2), if it is not,
Then execute (3)；
(1) if traffic sign recognition result is identical with surface mark recognition result, using any recognition result as
The direction information in the lane；
(2) if there are deviations for traffic sign recognition result and surface mark recognition result, using has greater weight
Direction information of the recognition result as the lane；
(3) if only existing traffic sign recognition result or surface mark recognition result in the same lane, this is identified
As a result the direction information as the lane.
Further, the image sequence data of real road are obtained in the step 1 by traverse measurement vehicle.
Further, SHAPE DETECTION is hough transform in the step 2.2, will meet the extracted region of traffic sign shape
Implementation out is as follows,
The length and width of its boundary rectangle can be indicated a figure in the hope of boundary rectangle, L, W, and Tag indicates rectangular degree Tag
After=S/ (L*W), S indicate segmentation, the number of the pixel of some figure, the i.e. actual size of figure；When the model of rectangular degree Tag
It is trapped among in [0.8,1.4], determines the figure for rectangle.
Further, in the step 2.3 using radial direction Tchebichef not bending moment extract the feature of each turn marking to
Measure V1.
Further, in the step 3.3 using normalized Fourier descriptor extract the feature of each turn marking to
Measure V2.
Further, the specified value in the step 4 is 80%.
The characteristics of present invention is distributed by surface mark in analysis real road and traffic sign, and using mutually confirmation machine
System realizes the high-accuracy identification of lane direction information, provides support for safe driving, while can serve high-precision map
Production, make it possible unmanned.
Detailed description of the invention
Fig. 1 is the flow chart of the embodiment of the present invention.
Fig. 2 is that the embodiment of the present invention mutually checks flow chart.
Specific embodiment
Below in conjunction with attached drawing and the embodiment of the present invention, detailed analysis explanation is carried out to technical solution of the present invention.Embodiment
Realization process can be summarized as following steps:
Step 1 obtains the image data of real road first, carries out data acquisition using traverse measurement vehicle, obtains corresponding
Image sequence data.
Traverse measurement vehicle is used herein, continuous acquisition includes traffic sign, the shadow of surface mark on urban road
As information.GPS system, therefore the trace information of available measurement vehicle are equipped on vehicle.It can be by road according to road junction
Road is divided into different road segment segments, and image can be classified according to road segment segment according to shooting time information, because being equipped on vehicle
Have GPS, so can have prompt when crossing, then can make marks to image at this time, occur at next crossing
Before, using the image of stretch photographs as a roadway segment.And traffic sign many places are turned near crossing, therefore can root
There is the larger image sequence for being likely to occur steering traffic sign according to this acquisition, for detection and identification later.And turn to letter
Breath is can to obtain number of track-lines, lane boundary with existing method for detecting lane lines here based on lane.
Step 2, the detection for carrying out traffic sign to each image identifies, including following sub-step,
Step 2.1, color segmentation is based on its color (usually red, blue, yellow etc.) feature, it would be possible to be traffic
The part of label extracts in whole picture figure；Hsv color system is than RGB system closer to the experience of people and to colour
Perception, is more suitable for the color segmentation of color image under natural scene.RGB color is transformed into HSV face by the embodiment of the present invention
Color threshold segmentation is carried out after the colour space, conversion formula is as follows,
Where MAX=MAX (R, G, B), MIN=MIN (R, G, B)
1 HSV Color Segmentation threshold value table of table
When it is implemented, those skilled in the art can sets itself hsv color segmentation threshold.After image segmentation, it can give birth to
At multiple connected regions, there is the color of oneself in each region, and the color of traffic sign is usually red, blue and yellow,
The extracting section for meeting traffic sign color to be come out.
Step 2.2, SHAPE DETECTION usually has specific shape (circle, triangle, rectangle) and ruler based on traffic sign
It is very little, in the region for meeting color extracted, region that those shapes are not inconsistent is filtered out (i.e. in addition to circle, triangle, square
Other shapes except shape), improve recognition efficiency；Hough transform is substantially carried out in the present embodiment, according to the connected domain detected
The number of pixel find out connected domain area, then find out its rectangular degree, be judged to rectangle if in suitable threshold value.
The length and width of its boundary rectangle can be indicated a figure in the hope of boundary rectangle, L, W, and Tag indicates rectangular degree Tag=
S/ (L*W), after S here indicates segmentation, the number of the pixel of some figure, the i.e. actual size of figure, those skilled in the art
Member can sets itself rectangular degree threshold value, rectangle traffic sign is extracted.
Step 2.3, feature extraction is divided into different zones according to the dotted line of label in the range that step 2.2 obtains,
Feature vector is extracted to each region respectively, and is used for the Classification and Identification of traffic sign；
Hough straight-line detection is carried out in the rectangular extent of acquisition, is divided different turn markings according to obtained straight line
It comes, for example has found four straight lines in rectangular extent, then it is (each that figure is just divided into three regions according to four straight lines
Region is a lane), every two adjacent straight lines determine a region, there is a turn marking in each region.It is extracting
When the feature vector respectively indicated, the embodiment of the present invention by the radial Tchebichef proposed using Mukundan et al. not bending moment,
For the image of N*N:
In formula, n is maximum pixel number on circumference, m=N/2,K=0,1,2...n-1.In this way, radial
Bending moment is not as follows by Tchebichef:
In formula,It is s respectivelypqReal part and imaginary part.
Use η11、η22、η42、η44Bending moment does not extract the feature vector of traffic sign image to four radial direction Tchebichef,
And it is used for the Classification and Identification of traffic sign.
Step 2.4, Classification and Identification, the embodiment of the present invention will use support vector machines (SVM) to identify, input in early period
Sample is trained, and obtains the classifier of better performances, and the feature vector in input step 2.3 exports generic.It has instructed
Sample has been perfected, has been tested then the direction information in each lane is put into trained classifier, correspondence can be exported
Direction information.According to dotted line division result and lane information, each region recognition result is assigned on lane.
Step 3: the detection for carrying out surface mark to each image identifies, specifically includes following sub-step,
Step 3.1, it pre-processes, including carries out gray processing to reduce operand, be filtered to eliminate noise, carry out histogram
Figure equalizes to further enhance mark and background contrast, carries out scene rebuilding to restore the shape feature of surface mark；
In order to reinforce the difference of mark with background, contrast will be enhanced using histogram equalization after image gray processing.
Simultaneously because the unknown and unstability and video camera of camera parameter itself speed in shooting process, shooting pitch angle etc.
Influence, making imaging process, there are biggish distortion, wherein most importantly perspective distortion.The distortion leads to road traffic label
Shape feature change, so as to cause the variation of target image characteristics, influence the accuracy of surface mark identification, because
This, it is essential for carrying out scene rebuilding to original image.Inverse perspective mapping is substantially exactly by the road under image coordinate system
Road image transforms in the plane under bodywork reference frame, after simple coordinate system transformation, so that it may obtain, image coordinate
It is the calculation formula that lower road image coordinate value is converted to actual physics distance under bodywork reference frame, by the transformation for mula, just
The coordinate value (developed width and height distance) of corresponding points on bodywork reference frame lower plane can be found out.
The model of inverse perspective mapping is as follows,
Wherein, c1=cos α, c2=cos β, s1=sin α, s2=sin β, α are pitch angles, and β is angle of drift, and (u, v) is shadow
As the coordinate in coordinate system, (X, Y, Z) is the coordinate in WGS-84 coordinate system, (fu, fv) be horizontal and vertical focal length length,
(cu, cv) it is principal point coordinate.H is height of the image center to ground.
When practical application, the method that can be demarcated by chessboard calibration method determines camera inside and outside parameter, and then completes coordinate and turn
It changes.Raw video can be carried out to inverse perspective and be converted to top view.
Step 3.2, lane line drawing and road area divide, including lane line is split from road image, with
The non-targeted object and interference information in road image are removed, the regional scope of specific lane is obtained.
Lane line is split from road image, the embodiment of the present invention carries out two using Otsu maximum variance between clusters
Value removes small connected domain (pixel value only have 1 to 20 can be seen as small connected domain) after binaryzation.Then ground is utilized
Face turn marking should position constraint between lane and area-constrained (turn marking is all to have fixed size), removal
Other objects in binary map, retain surface mark.
Step 3.3, feature extraction, including being divided into different zones according to lane line in the range that step 3.2 obtains,
Feature vector is extracted to each region respectively, and is used for the Classification and Identification of surface mark；
Step 3.2 obtain range in, different zones are divided into according to lane line, to each region extract respectively feature to
Amount, the embodiment of the present invention select Fourier descriptor to carry out feature extraction.The basic thought of Fourier descriptor: assuming that locate
The image outline of reason is a closed curve, and every bit P (t) has corresponding coordinate area (x (t), y (t)), t=on curve
0,1,2 ... N-1, N are total points on profile.Coordinate sequence is expressed as plural form x+yi, being considered as the period is that curve is total
Long periodic function obtains a series of coefficients by discrete Fourier transform, is Fourier descriptor, wherein high frequency coefficient is anti-
The detail section of profile is reflected, low frequency coefficient reflects the global shape of profile.The calculation formula of Fourier descriptor is as follows,
In:
According to the property of Fourier transformation, the scale of Fourier descriptor and shape, direction and curve initial point position
Related, i.e. the scale of figure is different, and the start position of profile is different, and obtained Fourier descriptor is also different.To guarantee
The scale of method, rotation and translation invariance, using method for normalizing.Normalized Fourier descriptor expression formula are as follows:
By seeking ratio with first order modulus value, normalized Fourier descriptor eliminates the influence of start position and ensures
Invariance calculates easy, control parameter that no setting is required, and the high stability of feature.Most of all, can be according to demand
The Fourier descriptor within the scope of certain frequency is selected to carry out shape recognition as shape feature.
Step 3.4, Classification and Identification, the embodiment of the present invention equally selects SVM to be identified herein, after classification
Identification information can be assigned on specific lane in conjunction with the specific lane of gained by obtaining surface mark recognition result.
Step 4: sequence analysis is carried out to the recognition result of step 2 and step 3, surface mark is calculated separately out and hands over
The confidence level of the direction information of logical label is marked if obtained confidence level is respectively less than specified value (usually taking 80% or higher)
It is denoted as unreliable information, remains to recheck, if obtained confidence level meets specified value, corresponding output has the knowledge of maximum confidence
Other result.
It include duplicate direction information in the image sequence extracted.For example, for the same right-turn lane,
May turn right mark actually on road there are three the ground being distributed at a certain distance, and the same mark here may
It was photographed by continuous 5 frame image.In this way, adding up to 3*5=15 frame image contains identical surface mark direction information.But in reality
It, may be not necessarily completely the same due to noise jamming or factors, the recognition results of this 15 frame such as shooting condition is bad when border identifies.
In order to utilize the information in this 15 frame image as far as possible, needs to carry out sequence analysis, it is cumulative that confidence level is employed herein, and selecting can
The highest strategy as a result of reliability.I.e. confidence level is directly proportional to there is frequency.Such as there is recognition result in this 15 frame
12 frames be turn right, 2 frames be straight trip, 1 frame be turn left, then select confidence level for 12/15 right-hand rotation as a result, being considered as ground at this time
Face turn marking is to turn right.Similarly, continuous if the same turn marking occurs 1 time on traffic sign on every image
5 frame images photographed, total 1*5=5 frame image contains traffic sign direction information, wherein 3 right-hand rotations, it is primary turn left, once
Straight trip, then 3/5 right-hand rotation is exactly the recognition result of traffic sign.Since the confidence level of surface mark recognition result is 12/15
=80% meets specified value, and traffic sign recognition result is that 3/5=60% is less than specified value, therefore only exports ground landmark identification
As a result.
Step 5, comparison veritify traffic sign and surface mark recognition result, obtain the direction information of the roadway segment.Including
First determine whether in the same lane whether and meanwhile include traffic sign and surface mark recognition result, if so, according to priori
Knowledge determines the weight of traffic sign and surface mark recognition result respectively, then (1) and (2) is executed, if it is not, then executing
(3)；In the present embodiment, if it is considered to surface mark recognition result is more acurrate, then the weight for assigning surface mark recognition result is
0.7, then the weight of traffic sign recognition result is 0.3,
(1) if traffic sign recognition result is identical with surface mark recognition result (as identical direction information), by it
In direction information of any recognition result as the lane；Such as traffic sign recognition result and surface mark recognition result are
It turns right, and confidence level is all satisfied specified value, then the direction information of the roadway segment is to turn right.
(2) if there are deviations for traffic sign recognition result and surface mark recognition result, using has greater weight
Direction information of the recognition result as the lane；Such as surface mark recognition result is to turn right (0.7), traffic sign recognition result
For left-hand rotation (0.3), then the direction information of the roadway segment is to turn right.
(3) if the same roadway segment only exists traffic sign recognition result or surface mark recognition result, this is identified
As a result the direction information as the lane；Such as surface mark recognition result is to turn right, traffic sign recognition result is sky, then
The direction information in the lane is to turn right.
Specific embodiment described herein is only an example for the spirit of the invention.The neck of technology belonging to the present invention
The technical staff in domain can make various modifications or additions to the described embodiments or replace by a similar method
In generation, however, it does not deviate from the spirit of the invention or beyond the scope of the appended claims.
Claims (6)
1. the lane direction information recognition methods based on traffic sign and surface mark, which comprises the following steps:
Step 1, the image sequence data of real road are obtained；
Step 2, the detection for carrying out traffic sign to each image identifies, including following sub-step,
Step 2.1, the segmentation of hsv color capacity-threshold is carried out to each image, extracts the region for meeting traffic sign color；
Step 2.2, in the region for meeting traffic sign color extracted, SHAPE DETECTION is carried out, traffic mark will be met
The extracted region of board shape comes out；
Step 2.3, hough straight-line detection is carried out in the traffic sign regional scope of acquisition, is divided into according to obtained straight line
Different lanes are corresponding with a turn marking in each lane, extract the feature vector V1 of each turn marking；
Step 2.4, feature vector V1 is input to support vector machines to identify, obtains the steering letter of traffic sign in lane
Breath；
Step 3, the detection for carrying out surface mark to each image identifies, including following sub-step,
Step 3.1, the processing of gray processing, histogram equalization and scene rebuilding is carried out to each image；
Step 3.2, the lane line in image is extracted, obtains the regional scope in lane, and retain the surface mark between lane；
Step 3.3, within the scope of each lane that step 3.2 obtains, the feature vector V2 of the turn marking in each lane is extracted；
Step 3.4, feature vector V2 is input to support vector machines to identify, obtains the steering letter of surface mark in lane
Breath；
Step 4, image sequence analysis is carried out to the direction information recognition result of step 2 and step 3, calculates separately out traffic sign
It is labeled as unreliable information, is remained if obtained confidence level is respectively less than specified value with the confidence level of surface mark direction information
Reinspection, if obtained confidence level meets specified value, corresponding output has the recognition result of maximum confidence；
Step 5, traffic sign recognition result and surface mark recognition result in lane are veritified in comparison, obtain the steering of the roadway segment
Information, implementation is as follows,
First determine whether in the same lane whether and meanwhile include traffic sign and surface mark recognition result, if so, according to
Priori knowledge determines the weight of traffic sign and surface mark recognition result respectively, then (1) and (2) is executed, if it is not, then holding
Row (3)；
(1) if traffic sign recognition result is identical with surface mark recognition result, using any recognition result as the vehicle
The direction information in road；
(2) if there are deviations for traffic sign recognition result and surface mark recognition result, using the identification with greater weight
As a result the direction information as the lane；
(3) if only existing traffic sign recognition result or surface mark recognition result in the same lane, by this recognition result
Direction information as the lane.
2. the lane direction information recognition methods based on traffic sign and surface mark, feature exist as described in claim 1
In: the image sequence data of real road are obtained in the step 1 by traverse measurement vehicle.
3. the lane direction information recognition methods based on traffic sign and surface mark, feature exist as claimed in claim 2
In: SHAPE DETECTION is hough transform in the step 2.2, the implementation that the extracted region for meeting traffic sign shape is come out
It is as follows,
The length and width of its boundary rectangle can be indicated a figure in the hope of boundary rectangle, L, W, and Tag indicates rectangular degree Tag=S/
(L*W), after S indicates segmentation, the number of the pixel of some figure, the i.e. actual size of figure；When the range of rectangular degree Tag exists
In [0.8,1.4], determine the figure for rectangle.
4. the lane direction information recognition methods based on traffic sign and surface mark, feature exist as claimed in claim 3
In: using radial direction Tchebichef, bending moment does not extract the feature vector V1 of each turn marking in the step 2.3.
5. the lane direction information recognition methods based on traffic sign and surface mark, feature exist as claimed in claim 4
In: the feature vector V2 of the turn marking in each lane is extracted in the step 3.3 using normalized Fourier descriptor.
6. the lane direction information recognition methods based on traffic sign and surface mark, feature exist as described in claim 1
In: the specified value in the step 4 is 80%.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710709060.4A CN107491756B (en) | 2017-08-17 | 2017-08-17 | Lane direction information recognition methods based on traffic sign and surface mark |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710709060.4A CN107491756B (en) | 2017-08-17 | 2017-08-17 | Lane direction information recognition methods based on traffic sign and surface mark |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107491756A CN107491756A (en) | 2017-12-19 |
| CN107491756B true CN107491756B (en) | 2019-10-11 |
Family
ID=60646414
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710709060.4A Active CN107491756B (en) | 2017-08-17 | 2017-08-17 | Lane direction information recognition methods based on traffic sign and surface mark |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107491756B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108492596B (en) * | 2018-03-16 | 2020-12-01 | 武汉理工大学 | Visual perception-based underground interchange rear-end collision prevention method |
| CN108520212A (en) * | 2018-03-27 | 2018-09-11 | 东华大学 | Method for traffic sign detection based on improved convolutional neural networks |
| CN109101924B (en) * | 2018-08-14 | 2021-05-04 | 武汉大学 | Machine learning-based road traffic sign identification method |
| CN111353351A (en) * | 2018-12-24 | 2020-06-30 | 观致汽车有限公司 | Method and device for determining whether a vehicle traveling on a main road enters a ramp |
| CN111460861B (en) * | 2019-01-21 | 2023-05-19 | 阿里巴巴集团控股有限公司 | Road traffic sign recognition method, device and recognition equipment |
| CN112052912A (en) * | 2020-09-23 | 2020-12-08 | 同济大学 | Intelligent flame combustion state identification method for fire-fighting robot |
| CN113221648B (en) * | 2021-04-08 | 2022-06-03 | 武汉大学 | Fusion point cloud sequence image guideboard detection method based on mobile measurement system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102609702A (en) * | 2012-02-03 | 2012-07-25 | 中山大学 | Method and system for rapidly positioning road guide sign |
| CN104361350A (en) * | 2014-10-28 | 2015-02-18 | 奇瑞汽车股份有限公司 | Traffic sign identification system |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1383098B1 (en) * | 2002-07-09 | 2006-05-17 | Accenture Global Services GmbH | System for automatic traffic sign recognition |
-
2017
- 2017-08-17 CN CN201710709060.4A patent/CN107491756B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102609702A (en) * | 2012-02-03 | 2012-07-25 | 中山大学 | Method and system for rapidly positioning road guide sign |
| CN104361350A (en) * | 2014-10-28 | 2015-02-18 | 奇瑞汽车股份有限公司 | Traffic sign identification system |
Non-Patent Citations (2)
| Title |
|---|
| Road sign detection in images:A case study;R.Belaroussi etal.;《2010 International Conference on Pattern Recognition》;20101231;第2010年卷;全文 * |
| 基于改进Hu不变距的路面交通标识识别;张明恒等;《大连理工大学学报》;20121130;第52卷(第6期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107491756A (en) | 2017-12-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107491756B (en) | Lane direction information recognition methods based on traffic sign and surface mark | |
| CN107729801B (en) | Vehicle color recognition system based on multitask deep convolution neural network | |
| CN110069986B (en) | Traffic signal lamp identification method and system based on hybrid model | |
| CN105160309B (en) | Three lanes detection method based on morphological image segmentation and region growing | |
| CN103927526B (en) | Vehicle detecting method based on Gauss difference multi-scale edge fusion | |
| CN107767383B (en) | Road image segmentation method based on superpixels | |
| CN105678285B (en) | A kind of adaptive road birds-eye view transform method and road track detection method | |
| CN102509098B (en) | Fisheye image vehicle identification method | |
| CN111814623A (en) | Vehicle lane departure visual detection method based on deep neural network | |
| KR20190036844A (en) | Method and apparatus of identifying driving lane | |
| CN107679531A (en) | Licence plate recognition method, device, equipment and storage medium based on deep learning | |
| Hu et al. | A multi-modal system for road detection and segmentation | |
| CN104778721A (en) | Distance measuring method of significant target in binocular image | |
| CN105654073B (en) | A kind of speed automatic control method of view-based access control model detection | |
| CN105005989B (en) | A kind of vehicle target dividing method under weak contrast | |
| CN105260699A (en) | Lane line data processing method and lane line data processing device | |
| CN107463890A (en) | A kind of Foregut fermenters and tracking based on monocular forward sight camera | |
| CN109948552B (en) | Method for detecting lane line in complex traffic environment | |
| CN105989334A (en) | Monocular vision-based road detection method | |
| CN107895375A (en) | The complicated Road extracting method of view-based access control model multiple features | |
| CN104200228A (en) | Recognizing method and system for safety belt | |
| CN111160328A (en) | Automatic traffic marking extraction method based on semantic segmentation technology | |
| CN112861748A (en) | Traffic light detection system and method in automatic driving | |
| CN112906616A (en) | Lane line extraction and generation method | |
| Kühnl et al. | Visual ego-vehicle lane assignment using spatial ray features |
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 |