CN109086671A - One kind being suitable for unpiloted night traffic lane line video detecting method - Google Patents
One kind being suitable for unpiloted night traffic lane line video detecting method Download PDFInfo
- Publication number
- CN109086671A CN109086671A CN201810723815.0A CN201810723815A CN109086671A CN 109086671 A CN109086671 A CN 109086671A CN 201810723815 A CN201810723815 A CN 201810723815A CN 109086671 A CN109086671 A CN 109086671A
- Authority
- CN
- China
- Prior art keywords
- area
- road
- interest
- lane line
- lane
- 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
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/24—Aligning, centring, orientation detection or correction of the image
- G06V10/245—Aligning, centring, orientation detection or correction of the image by locating a pattern; Special marks for positioning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/48—Extraction of image or video features by mapping characteristic values of the pattern into a parameter space, e.g. Hough transformation
Abstract
The present invention relates to one kind to be suitable for unpiloted night traffic lane line video detecting method, the following steps are included: 1) obtain road at night time image, and it is pre-processed, including inhibiting picture noise using median filtering and carrying out road edge enhancing using Sobel operator, the garbage in image is eliminated;2) the adaptive area-of-interest of "eight" shape is generated according to pretreated road at night time image;3) road boundary characteristic point is carried out in the adaptive area-of-interest of "eight" shape to classify, and obtain left-lane quadrilateral area interested and right lane quadrilateral area interested respectively according to classification results;4) using Improved Hough Transform fitting identification traffic lane line.Compared with prior art, the present invention has many advantages, such as to reduce screening road boundary point range, the classification of lane boundary point, the detection time for reducing lane line.
Description
Technical field
The present invention relates to intelligent transportation active safety fields, are suitable for unpiloted night lane more particularly, to one kind and mark
Will line video detecting method.
Background technique
With the fast development of China's urban economy, Car holding amount is risen year by year, and traffic safety problem is also gradually shown
It is existing.According to the official of State Statistics Bureau statistics indicate that the traffic accident more than 180,000 that China in 2015 occurs rises, wherein the number of casualties has been
30% is up to through the death rate more than 250,000 people, accident.Wherein since driver attention does not concentrate, vehicle is caused to deviate normal
Traveling lane and the phenomenon that leading to major motor vehicle casualty accident it is commonplace.Therefore, the active peace of research road mark line detection
Full technology shows important especially.
Lane detection technology can be used for driving assistance system, while be also the key in automatic driving vehicle R&D process
Technology.It is American-European-Japanese at present to have put into research and development department's divided lane detection system, wherein representative system such as RALPH system,
Start system, AURORA system and ALVINN system.In above system, different road models and different boundaries are utilized
Extractive technique carries out the detection of road.But above-mentioned detection algorithm is mainly applicable in uniform illumination on daytime scene, algorithm shortage pair
The adaptability of light variation.
And there has been no the products that Road Detection is carried out specifically for night scenes in China at present.About grinding for Road Detection
Study carefully and predominantly stays in uniform illumination scene on daytime.Due to night scenes complexity, uneven illumination is even, and road surface shade is many and diverse, rare
Research about night complex scene.And automatic driving vehicle inevitably needs to solve the demand in night running.Therefore
Need to study reliable road at night time detection technique, to improve unpiloted safety and practicability.
And the area-of-interest established in current Road Detection algorithm mainly uses changeless area-of-interest
Or adaptive area-of-interest is established using Kalman filter method, and area-of-interest is generally rectangular.Actually lane line
It is distributed in the picture in "eight" shape, there are 95% or more useless areas in rectangle area-of-interest, while being easy by noise
Interference.In the recent period, the external area-of-interest for having research to establish is " Λ " type.But night Lane Mark is due to luminaire unilateral side cloth
It sets, left and right lane is distributed after edge enhances to be had differences, and area-of-interest still has optimization space.It is emerging for the sense of nighttime image
The optimization in interesting region, can be improved the speed of algorithm, to meet the requirement of real-time of automatic driving vehicle.
Generally speaking, China lacks the Related product that Road Detection is carried out for night scenes, and existing research at present
It is insufficient for the development of unmanned technology, lacks reliability and real-time.To avoid safety accident caused by deviation,
The invention proposes one kind to be suitable for unpiloted night traffic lane line video detection technology.
Summary of the invention
It is an object of the present invention to overcome the above-mentioned drawbacks of the prior art and provide one kind suitable for unmanned
Night traffic lane line video detecting method.
The purpose of the present invention can be achieved through the following technical solutions:
One kind being suitable for unpiloted night traffic lane line video detecting method, comprising the following steps:
1) road at night time image is obtained, and is pre-processed, including picture noise is inhibited using median filtering and is adopted
Road edge enhancing is carried out with Sobel operator, eliminates the garbage in image;
2) the adaptive area-of-interest of "eight" shape is generated according to pretreated road at night time image;
3) road boundary characteristic point is carried out in the adaptive area-of-interest of "eight" shape to classify, and tied according to classification
Fruit obtains left-lane quadrilateral area interested and right lane quadrilateral area interested respectively;
4) using Improved Hough Transform fitting identification traffic lane line.
The step 2) specifically includes the following steps:
21) initialization "eight" shape area-of-interest is established:
The initialization area B of rectangle is set in the enhanced first width image in edgei, by the way that coefficient of variation f is arrangedlExpand
Initialization area BiWidth, the initialization area B after expansioniThe middle endpoint for carrying out multi-direction search and obtaining road boundary point,
Initialization "eight" shape area-of-interest is formed in first width image;
22) the adaptive area-of-interest of "eight" shape is established:
According to the area-of-interest R of the enhanced road at night time image of upper breadths edgea, expanded using horizontal expansion coefficient
RaRegion R after being expandedb, and in region RbThe middle end for carrying out multi-direction search and obtaining road boundary point in present image
Point.
The multi-direction search obtains the endpoint specific steps of road boundary point are as follows:
For any group of laterally adjacent marginal point a and b in initialization area, if yb-ya≥dmin∩yb-ya≤dmax, then
The two o'clock for the endpoint of road boundary point and is divided into road boundary point set B, wherein xa,ya、xb,ybRespectively marginal point a
With the pixel coordinate of b, dminAnd dmaxFor the minimum value and maximum value of neighbor distance.
In the step 3), carries out road boundary characteristic point and classifies specifically:
For the road boundary point in left plane:
F if (x, y) ∈ B ∩ f (x, y) ∈ Rcl∩Sy> 0, then f (x, y) ∈ Blo;
F if (x, y) ∈ B ∩ f (x, y) ∈ Rcl∩Sy< 0, then f (x, y) ∈ Bli;
For the road boundary point in right plane:
F if (x, y) ∈ B ∩ f (x, y) ∈ Rcr∩Sy> 0, then f (x, y) ∈ Bro;
F if (x, y) ∈ B ∩ f (x, y) ∈ Rcr∩Sy< 0, then f (x, y) ∈ Bri;
Wherein, f (x, y) is the pixel position in the adaptive area-of-interest of "eight" shape, and B is road boundary point set,
BloFor left outside boundary set, BroFor right outer boundary collection, BliFor left inside boundary set, BriFor right inner boundary collection, RclCertainly for "eight" shape
Adapt to the left part of area-of-interest, RcrFor the right part of the adaptive area-of-interest of "eight" shape, SyFor pixel point value
The value being calculated by 3 × 3 longitudinal Sobel operator template.
In the step 4), two parts in left and right are divided the image into, traffic lane line is identified respectively, obtains a left side for road
Right margin.
In the step 4), left fluctuation angle A is setflWith right fluctuation angle Afr, then the boundary of left-lane and x
Line brcrNormal angle.
Compared with prior art, the invention has the following advantages that
(1) adaptive "eight" shape area-of-interest is established
Quadrangle area-of-interest is independently established according to the distribution of left and right lane boundary point, is suitable for road at night time brightness point
The unbalanced scene of cloth can effectively evade the interference of noise, while substantially reduce the range of screening road boundary point, make to handle
The time of image is reduced.
(2) based on road boundary textural characteristics classification lane boundary point
Have found road boundary textural characteristics, and on the "eight" shape area-of-interest of foundation according to this feature by lane
Boundary point is divided into left outside boundary set, left inside boundary set, the right inner boundary collection of right outer boundary set, facilitates subsequent progress lane line knowledge
Not.
(3) road mark line is identified based on improved Hough transform
During carrying out Hough transform detection road mark line, in the "eight" shape area-of-interest based on foundation
The angle of two quadrangle sides controls the search angle of left and right lane line, to reduce the detection time of lane line.
Detailed description of the invention
Fig. 1 is suitable for unpiloted night traffic lane line Video Detection Algorithm flow chart.
Fig. 2 is the road at night time original image of acquisition.
Fig. 3 is the road at night time image after median filtering.
Fig. 4 is Sobel operator edge enhancing figure.
Fig. 5 is initialization area-of-interest flow chart.
Fig. 6 is multi-direction search road boundary point endpoint schematic diagram.White box inner region is rectangle initialization area in figure
Bi, the white arrow expression direction of search, intensive white point is the lane boundary point that screening obtains, and white dot expression is searched for
The endpoint of the road boundary point arrived.
Fig. 7 is initialization area-of-interest figure.White box inner region according to figure 5 establish to obtain by method in figure
Initialize area-of-interest.
Fig. 8 is the adaptive area-of-interest flow chart of "eight" shape.
Fig. 9 is multi-direction search road boundary point endpoint schematic diagram.White box inner region is by a upper picture in figure
R after area-of-interest expansionb.White arrow indicates the direction of search, and intensive white point is the lane boundary point that screening obtains,
White dot indicates the endpoint for the road boundary point that search obtains.
Figure 10 is the adaptive area-of-interest figure of "eight" shape.White box inner region is method according to figure 8 in figure
Establish the obtained adaptive area-of-interest of "eight" shape.
Figure 11 is Hough transform coordinate system diagram.Quadrangle a in figurelblcldlWith quadrangle arbrcrdrThe left side respectively formed
The area-of-interest of right lane.α and β be respectively left and right lane boundary and x-axis formed by angle.
Figure 12 is road at night time markings testing result figure.
Specific embodiment
The present invention is described in detail with specific embodiment below in conjunction with the accompanying drawings.
Embodiment
The present invention provide it is a kind of be suitable for unpiloted night traffic lane line video detecting method, by median filtering and
Sobel operator pre-processes image, then establishes adaptive "eight" shape area-of-interest based on lane line geometrical characteristic
And classified according to the textural characteristics of road boundary to boundary point, then, roadside is detected using improved Hough transform
Boundary.The present invention can carry out road at night time detection, avoid occur as divert one's attention drive and traffic thing caused by deviation occurs
Therefore while can be automatic driving vehicle service, ensure the safety of traveler, specific steps are as follows:
Step 1;Carry out road at night time and detect refractory gold ores: road at night time is detected than the road under uniform illumination scene on daytime
It detects increasingly complex, is mainly reflected in:
1, road at night time image grayscale is integrally lower, and it is big to extract lane boundary point difficulty.
2, there is alternate hot spot in road surface, and there are still edge missings after edge enhancing.
3, nighttime image is influenced by car light, and the brightness at close shot is high, and brightness is low at distant view, and the lane line edge at distant view is easy
Missing.
4, vehicle windscreen is easy reflective, forms luminance area in the picture, expands the gray value of pixel in region.
Step 2;Road at night time image preprocessing: since the complexity of nighttime image increases the difficulty of Road Detection, make
Its processing method is different from the processing that image is acquired under uniform illumination scene on daytime.Image is inhibited to make an uproar using median filtering first
Sound can overcome the brings image detail mould such as linear filter such as mean filter, least means square under certain condition
Paste problem, effective protection edge and profile information.Its statistical nature that image is not needed in actual operation simultaneously, therefore handle
Speed is relatively fast, and the road image suitable for automatic driving vehicle pre-processes.It is then based on Sobel operator and carries out road edge
Enhancing.Sobel operator is a kind of first order differential operator, can effectively eliminate most of garbage in image.
Step 3;Construct the adaptive area-of-interest of "eight" shape:
Initially set up initialization area-of-interest: input is by the enhanced image in Sobel operator edge and rectangle is arranged
Initialization area Bi, it is assumed that BiThe pixel of interior laterally adjacent marginal point is f (xa,ya) and f (xb,yb), road boundary collection B is adjacent
Minimum range dmin, adjacent maximum distance dmax, road boundary point is extracted according to inside and outside road boundary point distance feature.If yb-ya
≥dmin∩yb-ya≤dmax, then f (xa,ya)∈B,f(xb,yb)∈B.In BiThe endpoint of interior multi-direction search road boundary point.For
Area-of-interest coefficient of variation f is arranged in consideration road boundary point as much as possiblel, expand area-of-interest width, formed just
Beginningization "eight" shape area-of-interest.
Then establish the adaptive area-of-interest of "eight" shape: the road boundary endpoint based on a upper picture calculates its sense
Interest region RaThe linear relationship on boundary, and according to horizontal expansion coefficient ehExpand RaObtain Rb.Sobel operator side is passed through in input
The enhanced new images of edge, in RbIt is middle that road boundary point is extracted according to distance feature.In RbInterior multi-direction search road boundary point
Endpoint.Area-of-interest coefficient of variation f is setl, expand area-of-interest width, form the adaptive region of interest of "eight" shape
Domain.
Step 4;The classification of road boundary point: the road boundary point set defined in the region of interest is B, " eight " word of foundation
The left part of shape area-of-interest is Rcl, right part Rcr, left outside boundary set Blo, left inside boundary set Bli, right outer boundary collection
Bro, right inner boundary collection Bri.It is assumed that current pixel f, position in the picture is (x, y), according to the textural characteristics of road boundary
Following classification is carried out to each boundary point: for the road boundary point in left plane.1. if f (x, y) ∈ B ∩ f (x, y) ∈ Rcl
∩Sy> 0, then f (x, y) ∈ Blo.2. if f (x, y) ∈ B ∩ f (x, y) ∈ Rcl∩Sy< 0, then f (x, y) ∈ Bli.For on the right side
The road boundary point of plane.3. if f (x, y) ∈ B ∩ f (x, y) ∈ Rcr∩Sy> 0, then f (x, y) ∈ Bro.4. if f (x, y)
∈B∩f(x,y)∈Rcr∩Sy< 0, then f (x, y) ∈ Bri.Above-mentioned classification method is in the adaptive area-of-interest of "eight" shape
It is interior to the progress of road boundary characteristic point, it can effectively evade noise jamming, meet the requirement of accuracy.
Step 5;Road mark line identification: the lane model of this paper is selected as straight line model, and improved Hough transform is selected to know
Other lane boundary.The quadrangle a in Hough transform coordinate system diagramlblcldlWith quadrangle arbrcrdrThe left and right vehicle respectively formed
The area-of-interest in road.It is assumed that straight line a in left arealdlIt is y=k in the linear relationship of rectangular coordinate systemlox+blo, normal
Angle beStraight line blclLinear relationship be y=klix+bli, the angle of normal isAnd 0 <
klo≤kli;Straight line a in right areardrLinear relationship be y=krix+bri, the angle of normal isStraight line
ardrLinear relationship be y=krox+bri, the angle of normal isAnd kri≤kro< 0.
Efficiency is calculated to obtain in order to improve, propose following search premise: 1. left-lane is located at the left side plane of image, right lane
Positioned at image by half-plane.Therefore during lane detection, left and right two parts are divided the image into, identify a left side for road respectively
Right margin;2. assuming that left and right search fluctuation angle is AflAnd Afr, angle formed by the boundary in left and right lane and x-axis is respectively α
And β, the computer capacity control of α and β is existed3. in quadrangle
alblcldlLeft-lane line is searched in area-of-interest, in quadrangle arbrcrdrRight-lane line is searched in area-of-interest, reduces vehicle
The time of diatom detection.
Step 6;Algorithm optimization: Road Detection practicability and safety in order to meet automatic driving vehicle, to above-mentioned road
It is as follows that road detection algorithm proposes Optimized Measures: 1. as the k of the "eight" shape area-of-interest of foundationlo、kli、kriAnd kroWith it is upper
One time Road Detection respective value difference is greater than threshold value cf, then it is assumed that area-of-interest establishes error.The k of error is set as last
The respective value of Road Detection, and expand area-of-interest using linear relationship, continue subsequent operation.2. the sense when foundation is emerging
Interesting region area is less than threshold value sfWhen, expand area-of-interest using linear relationship, continues subsequent operation.3. when direction is searched
When rope road boundary endpoint malfunctions, then it is assumed that since brightness is low, road boundary point quantity is very few after the enhancing of Sobel operator edge is led
It causes.1.2.1 is changed to the enhancing of the edge Canny, re-starts operation.4. going out when changing identification road mark line using Hough
It staggers the time, it is believed that be to be unable to satisfy Hough transform fitting since road boundary point quantity is very few and require.Processing method is the same as 3..
Claims (6)
1. one kind is suitable for unpiloted night traffic lane line video detecting method, which comprises the following steps:
1) road at night time image is obtained, and is pre-processed, including picture noise and use are inhibited using median filtering
Sobel operator carries out road edge enhancing, eliminates the garbage in image;
2) the adaptive area-of-interest of "eight" shape is generated according to pretreated road at night time image;
3) it carries out road boundary characteristic point in the adaptive area-of-interest of "eight" shape to classify, and according to classification results point
It Huo get not left-lane quadrilateral area interested and right lane quadrilateral area interested;
4) using Improved Hough Transform fitting identification traffic lane line.
2. according to claim 1 a kind of suitable for unpiloted night traffic lane line video detecting method, feature
Be, the step 2) specifically includes the following steps:
21) initialization "eight" shape area-of-interest is established:
The initialization area B of rectangle is set in the enhanced first width image in edgei, by the way that coefficient of variation f is arrangedlExpand initial
Change region BiWidth, the initialization area B after expansioniThe middle endpoint for carrying out multi-direction search and obtaining road boundary point, in head
Initialization "eight" shape area-of-interest is formed in width image;
22) the adaptive area-of-interest of "eight" shape is established:
According to the area-of-interest R of the enhanced road at night time image of upper breadths edgea, R is expanded using horizontal expansion coefficienta?
Region R after to expansionb, and in region RbThe middle endpoint for carrying out multi-direction search and obtaining road boundary point in present image.
3. according to claim 2 a kind of suitable for unpiloted night traffic lane line video detecting method, feature
It is, the multi-direction search obtains the endpoint specific steps of road boundary point are as follows:
For any group of laterally adjacent marginal point a and b in initialization area, if yb-ya≥dmin∩yb-ya≤dmax, then should
Two o'clock is the endpoint of road boundary point and is divided into road boundary point set B, wherein xa,ya、xb,ybRespectively marginal point a's and b
Pixel coordinate, dminAnd dmaxFor the minimum value and maximum value of neighbor distance.
4. according to claim 2 a kind of suitable for unpiloted night traffic lane line video detecting method, feature
It is, in the step 3), carries out road boundary characteristic point and classify specifically:
For the road boundary point in left plane:
F if (x, y) ∈ B ∩ f (x, y) ∈ Rcl∩Sy> 0, then f (x, y) ∈ Blo;
F if (x, y) ∈ B ∩ f (x, y) ∈ Rcl∩Sy< 0, then f (x, y) ∈ Bli;
For the road boundary point in right plane:
F if (x, y) ∈ B ∩ f (x, y) ∈ Rcr∩Sy> 0, then f (x, y) ∈ Bro;
F if (x, y) ∈ B ∩ f (x, y) ∈ Rcr∩Sy< 0, then f (x, y) ∈ Bri;
Wherein, f (x, y) is the pixel position in the adaptive area-of-interest of "eight" shape, and B is road boundary point set, BloFor
Left outside boundary set, BroFor right outer boundary collection, BliFor left inside boundary set, BriFor right inner boundary collection, RclAdaptively feel for "eight" shape
The left part in interest region, RcrFor the right part of the adaptive area-of-interest of "eight" shape, SyFor pixel point value by 3 ×
The value that 3 longitudinal Sobel operator template is calculated.
5. according to claim 1 a kind of suitable for unpiloted night traffic lane line video detecting method, feature
It is, in the step 4), divides the image into two parts in left and right, identify traffic lane line respectively, obtain the left and right of road
Boundary.
6. according to claim 1 a kind of suitable for unpiloted night traffic lane line video detecting method, feature
It is, in the step 4), sets left fluctuation angle AflWith right fluctuation angle Afr, then formed by the boundary of left-lane and x-axis
The scope control of angle α isAngle β formed by the boundary of right lane and x-axis
Scope control isWherein,For left-lane area-of-interest
alblcldlLeft side bearing aldlNormal angle,For left-lane quadrilateral area a interestedlblcldlRight side bearing blcl
Normal angle,For right lane quadrangle quadrilateral area a interestedrbrcrdrLeft side bearing ardrNormal angle,For right lane quadrilateral area a interestedrbrcrdrRight side bearing brcrNormal angle.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810723815.0A CN109086671B (en) | 2018-07-04 | 2018-07-04 | Night lane marking line video detection method suitable for unmanned driving |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810723815.0A CN109086671B (en) | 2018-07-04 | 2018-07-04 | Night lane marking line video detection method suitable for unmanned driving |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109086671A true CN109086671A (en) | 2018-12-25 |
CN109086671B CN109086671B (en) | 2021-05-11 |
Family
ID=64837270
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810723815.0A Active CN109086671B (en) | 2018-07-04 | 2018-07-04 | Night lane marking line video detection method suitable for unmanned driving |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109086671B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110163166A (en) * | 2019-05-27 | 2019-08-23 | 北京工业大学 | A kind of Robust Detection Method of vcehicular tunnel LED illumination lamp |
CN111931560A (en) * | 2020-06-23 | 2020-11-13 | 东南大学 | Linear acceleration lane marking line detection method suitable for formula-free racing car |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101382997A (en) * | 2008-06-13 | 2009-03-11 | 青岛海信电子产业控股股份有限公司 | Vehicle detecting and tracking method and device at night |
CN103297754A (en) * | 2013-05-02 | 2013-09-11 | 上海交通大学 | Monitoring video self-adaption interesting area coding system |
EP2813073A1 (en) * | 2012-02-10 | 2014-12-17 | Google, Inc. | Adaptive region of interest |
CN107895151A (en) * | 2017-11-23 | 2018-04-10 | 长安大学 | Method for detecting lane lines based on machine vision under a kind of high light conditions |
-
2018
- 2018-07-04 CN CN201810723815.0A patent/CN109086671B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101382997A (en) * | 2008-06-13 | 2009-03-11 | 青岛海信电子产业控股股份有限公司 | Vehicle detecting and tracking method and device at night |
EP2813073A1 (en) * | 2012-02-10 | 2014-12-17 | Google, Inc. | Adaptive region of interest |
CN103297754A (en) * | 2013-05-02 | 2013-09-11 | 上海交通大学 | Monitoring video self-adaption interesting area coding system |
CN107895151A (en) * | 2017-11-23 | 2018-04-10 | 长安大学 | Method for detecting lane lines based on machine vision under a kind of high light conditions |
Non-Patent Citations (2)
Title |
---|
游峰等: "基于边界点分布特征的夜问道路检测算法研究", 《交通信息与安全》 * |
高建明: "自适应感兴趣区域车道检测算法", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110163166A (en) * | 2019-05-27 | 2019-08-23 | 北京工业大学 | A kind of Robust Detection Method of vcehicular tunnel LED illumination lamp |
CN110163166B (en) * | 2019-05-27 | 2021-06-25 | 北京工业大学 | Robust detection method for LED lighting lamp of highway tunnel |
CN111931560A (en) * | 2020-06-23 | 2020-11-13 | 东南大学 | Linear acceleration lane marking line detection method suitable for formula-free racing car |
Also Published As
Publication number | Publication date |
---|---|
CN109086671B (en) | 2021-05-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107330376B (en) | Lane line identification method and system | |
CN105893949B (en) | A kind of method for detecting lane lines under complex road condition scene | |
CN100414560C (en) | Vehicle plate extracting method based on small wave conversion and Redon transform | |
KR101094752B1 (en) | Lane Classification Method Using Statistical Model of HSI Color Information | |
CN104182756B (en) | Method for detecting barriers in front of vehicles on basis of monocular vision | |
TWI401473B (en) | Night time pedestrian detection system and method | |
CN108805065A (en) | One kind being based on the improved method for detecting lane lines of geometric properties | |
CN105528609A (en) | Vehicle license plate location method based on character position | |
CN101334836A (en) | License plate positioning method incorporating color, size and texture characteristic | |
CN107316486A (en) | Pilotless automobile visual identifying system based on dual camera | |
CN108932472A (en) | A kind of automatic Pilot running region method of discrimination based on lane detection | |
CN102999753A (en) | License plate locating method | |
CN101916383A (en) | Vehicle detecting, tracking and identifying system based on multi-camera | |
CN102938057B (en) | A kind of method for eliminating vehicle shadow and device | |
CN103902985B (en) | High-robustness real-time lane detection algorithm based on ROI | |
CN106887004A (en) | A kind of method for detecting lane lines based on Block- matching | |
CN107895151A (en) | Method for detecting lane lines based on machine vision under a kind of high light conditions | |
CN105718916A (en) | Lane line detection method based on Hough transform | |
CN109886168B (en) | Ground traffic sign identification method based on hierarchy | |
Saha et al. | Automated road lane detection for intelligent vehicles | |
Lin et al. | Adaptive IPM-based lane filtering for night forward vehicle detection | |
CN107563331A (en) | A kind of road sign line detecting method and system based on geometrical relationship | |
CN105139011A (en) | Method and apparatus for identifying vehicle based on identification marker image | |
CN104537348A (en) | Special vehicle recognition method and system | |
CN109086671A (en) | One kind being suitable for unpiloted night traffic lane line video detecting method |
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 |