CN108985230A - Method for detecting lane lines, device and computer readable storage medium - Google Patents
Method for detecting lane lines, device and computer readable storage medium Download PDFInfo
- Publication number
- CN108985230A CN108985230A CN201810787347.3A CN201810787347A CN108985230A CN 108985230 A CN108985230 A CN 108985230A CN 201810787347 A CN201810787347 A CN 201810787347A CN 108985230 A CN108985230 A CN 108985230A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- cloud data
- multiframe
- data
- 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.)
- Pending
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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of method for detecting lane lines, device and computer readable storage medium, method for detecting lane lines includes: to obtain multiframe point cloud data, obtains the position of all frame point clouds after carrying out pose optimization to the multiframe point cloud data;Figure, which is locally built, based on multiframe point cloud data progress cloud obtains local map;The format of the local map is converted into picture format;Lane line is detected in the local map of picture format, obtains lane line data.The present invention realizes detection lane line due to obtaining multiframe point cloud data using laser measurement principle, and blind area when can be to avoid single-frame images or single frames point cloud detection lane line detects larger range of lane line data, and algorithm comparison is simple.
Description
Technical field
The present invention relates to lane detection technical field more particularly to method for detecting lane lines, device and computer-readable
Storage medium.
Background technique
Unmanned intelligent driving automobile is a complicated intelligence control system, is determined comprising Mechanical course, path planning, path
Multiple modules such as plan and environment sensing.Lane detection is the pith in environmental perception module, can be vehicle driving
Deviate and carries out early warning, provides decision information for lane.
Unmanned recent years are quickly grown.If unmanned vehicle is wanted to realize automatic Pilot, from the angle of vision its
First to learn observation road and specifically exactly detect lane line.Positional relationship including identifying lane line and vehicle, is solid line
Or dotted line etc..
Traditional unmanned vehicle automatic Pilot is when driving, it is often necessary to detect lane line and be fitted to determine travelable area
Domain.However, current lane detection is often detected with single-frame images or single frames point cloud, detection range is smaller, and false detection rate compared with
It is high.
Summary of the invention
The main purpose of the present invention is to provide a kind of method for detecting lane lines, device and computer readable storage medium,
Aim to solve the problem that the technical problem that method for detecting lane lines detection range is small in the prior art.
To achieve the above object, the present invention provides a kind of method for detecting lane lines, and the method for detecting lane lines includes:
Multiframe point cloud data is obtained, obtains the position of all frame point clouds after carrying out pose optimization to the multiframe point cloud data
It sets;
Figure, which is locally built, based on multiframe point cloud data progress cloud obtains local map;
The format of the local map is converted into picture format;
Lane line is detected in the local map of picture format, obtains lane line data.
Preferably, the step of acquisition multiframe point cloud data includes:
Obtain the history multiframe laser radar data of buffer area storage;
Multiframe point cloud data is obtained based on the history multiframe laser radar data.
Preferably, it is described pose carried out to the multiframe point cloud data optimize the step of obtaining optimum results include:
By laser radar SLAM algorithm to the location information and attitude angle of frame point cloud each in the multiframe point cloud data
It optimizes, obtains the position of all frame point clouds.
Preferably, described cloud is carried out based on the multiframe point cloud data locally to build the step of figure obtains local map and include:
It filters the multiframe point cloud data and obtains the first multiframe point cloud data, wherein the first multiframe point cloud data is
The point cloud data that the optimum results decline on the ground;
The first multiframe point cloud data is projected on X/Y plane, then carries out rasterizing and handles to obtain the anti-of each grid
Penetrate strength mean value;
Local map is obtained according to the reflected intensity mean value.
Preferably, the step of filtering multiframe point cloud data obtains the first multiframe point cloud data include:
Reject the second multiframe point cloud data in the multiframe point cloud data, wherein the second multiframe point cloud data is
The point cloud data on ground is not fallen in the multiframe point cloud data.
Preferably, described the step of obtaining local map according to the reflected intensity mean value, includes:
The reflected intensity mean value of each grid is weighted and averaged to obtain reflection intensity values, is based on the reflection intensity values
Obtain local map.
Preferably, described the step of detecting lane line in the local map of picture format, obtaining lane line data, includes:
Picture noise is carried out to the local map of described image format, map image is obtained by filtration;
Histogram image is obtained after carrying out binaryzation and histogram equalization to the map image;
Lane detection is carried out to the histogram image using corner feature and edge feature, obtains lane line data.
Preferably, after described the step of detecting lane line in the local map of picture format, obtaining lane line data,
The method for detecting lane lines further include:
Based on the lane line data, the fitting of straight line and curve is carried out by Hough transformation, obtains fitting result;
Travelable region is determined by fitting result.
In addition, to achieve the above object, the present invention also provides a kind of lane detection device, the lane detection device
Include: memory, processor and be stored in the lane detection program that can be run on the memory and on the processor,
The step of lane detection program realizes method for detecting lane lines as described above when being executed by the processor.
In addition, to achieve the above object, it is described computer-readable the present invention also provides a kind of computer readable storage medium
It is stored with lane detection program on storage medium, is realized when the lane detection program is executed by processor as described above
The step of method for detecting lane lines.
In the present invention, multiframe point cloud data is obtained, obtains all frame point clouds after carrying out pose optimization to multiframe point cloud data
Position, the corresponding figure of obtained point cloud data is more acurrate;Figure is locally built based on multiframe point cloud data progress cloud to obtain locally
Figure realizes the filtering to point cloud data, the format of local map is converted to picture format, convenient for carrying out histogram to local map
Figure equalization processing, so that the detection to lane line is realized, due to obtaining multiframe point cloud data using laser measurement principle
Realize that detection lane line, blind area when can be to avoid single-frame images or single frames point cloud detection lane line detect bigger model
The lane line data enclosed, and algorithm comparison is simple.Through the invention, lane line is detected using multiframe point cloud data, is obtained
The lane line data bigger to detection range, lane line is continuous, reduces false detection rate.
Detailed description of the invention
Fig. 1 is the lane detection apparatus structure schematic diagram for the hardware running environment that the embodiment of the present invention is related to;
Fig. 2 is the flow diagram of method for detecting lane lines first embodiment of the present invention;
Fig. 3 is the flow diagram of method for detecting lane lines second embodiment of the present invention;
Fig. 4 is the flow diagram of method for detecting lane lines 3rd embodiment of the present invention.
The embodiments will be further described with reference to the accompanying drawings for the realization, the function and the advantages of the object of the present invention.
Specific embodiment
It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not intended to limit the present invention.
As shown in Figure 1, Fig. 1 is the lane detection apparatus structure for the hardware running environment that the embodiment of the present invention is related to
Schematic diagram.
Lane detection device of the embodiment of the present invention can be PC, be also possible to smart phone, tablet computer, portable calculating
Machine etc. has the terminal device of certain data-handling capacity.
As shown in Figure 1, the lane detection device may include: processor 1001, such as CPU, network interface 1004 is used
Family interface 1003, memory 1005, communication bus 1002.Wherein, communication bus 1002 is for realizing the company between these components
Connect letter.User interface 1003 may include display screen (Display), input unit such as keyboard (Keyboard), can be selected
Family interface 1003 can also include standard wireline interface and wireless interface.Network interface 1004 optionally may include standard
Wireline interface, wireless interface (such as WI-FI interface).Memory 1005 can be high speed RAM memory, be also possible to stable deposit
Reservoir (non-volatile memory), such as magnetic disk storage.Memory 1005 optionally can also be independently of aforementioned place
Manage the storage device of device 1001.
It will be understood by those skilled in the art that lane detection apparatus structure shown in Fig. 1 is not constituted to lane line
The restriction of detection device may include perhaps combining certain components or different components than illustrating more or fewer components
Arrangement.
As shown in Figure 1, as may include that operating system, network are logical in a kind of memory 1005 of computer storage medium
Believe module, Subscriber Interface Module SIM and lane detection program.
In lane detection device shown in Fig. 1, network interface 1004 is mainly used for connecting background server, with backstage
Server carries out data communication;User interface 1003 is mainly used for connecting client (user terminal), and it is logical to carry out data with client
Letter;And processor 1001 can be used for calling the lane detection program stored in memory 1005, and execute following operation:
Multiframe point cloud data is obtained, obtains the position of all frame point clouds after carrying out pose optimization to the multiframe point cloud data
It sets;
Figure, which is locally built, based on multiframe point cloud data progress cloud obtains local map;
The format of the local map is converted into picture format;
Lane line is detected in the local map of picture format, obtains lane line data.
Further, processor 1001 can call the lane detection program stored in memory 1005, also execute with
Lower operation:
Obtain the history multiframe laser radar data of buffer area storage;
Multiframe point cloud data is obtained based on the history multiframe laser radar data.
Further, processor 1001 can call the lane detection program stored in memory 1005, also execute with
Lower operation:
By laser radar SLAM algorithm to the location information and attitude angle of frame point cloud each in the multiframe point cloud data
It optimizes, obtains the position of all frame point clouds.
Further, processor 1001 can call the lane detection program stored in memory 1005, also execute with
Lower operation:
It filters the multiframe point cloud data and obtains the first multiframe point cloud data, wherein the first multiframe point cloud data is
The point cloud data that the optimum results decline on the ground;
The first multiframe point cloud data is projected on X/Y plane, then carries out rasterizing and handles to obtain the anti-of each grid
Penetrate strength mean value;
Local map is obtained according to the reflected intensity mean value.
Further, processor 1001 can call the lane detection program stored in memory 1005, also execute with
Lower operation:
Reject the second multiframe point cloud data in the multiframe point data, wherein second multiframe point cloud data be
The point cloud data on ground is not fallen in the optimum results.
Further, processor 1001 can call the lane detection program stored in memory 1005, also execute with
Lower operation:
The reflected intensity mean value of each grid is weighted and averaged to obtain reflection intensity values, is based on the reflection intensity values
Obtain local map.
Further, processor 1001 can call the lane detection program stored in memory 1005, also execute with
Lower operation:
Picture noise is carried out to the local map of described image format, map image is obtained by filtration;
Histogram image is obtained after carrying out binaryzation and histogram equalization to the map image;
Lane detection is carried out to the histogram image using corner feature and edge feature, obtains lane line data.
Further, processor 1001 can call the lane detection program stored in memory 1005, also execute with
Lower operation:
Based on the lane line data, the fitting of straight line and curve is carried out by Hough transformation, obtains fitting result;
Travelable region is determined by fitting result.
It is the flow diagram of method for detecting lane lines first embodiment of the present invention referring to Fig. 2, Fig. 2.
In one embodiment, method for detecting lane lines includes:
Step S10 obtains multiframe point cloud data, obtains all frame points after carrying out pose optimization to the multiframe point cloud data
The position of cloud.
In the present embodiment, multiframe point cloud data is obtained, pose is carried out to the multiframe point cloud data and optimizes to obtain all frames
The position of point cloud.Obtain the mode of multiframe point cloud data are as follows: history multiframe laser radar data, base are retained by a buffer area
Multiframe point cloud data is obtained in history multiframe laser radar data, it is excellent to carry out pose for the multiframe laser radar data got
The position of all frame point clouds is obtained after change.
In the present embodiment, laser radar is an optical remote sensing technology, it intensively adopts earth surface using laser
Sample, to generate high-precision x, y, z measured value.Laser radar be mainly used for airborne laser drawing application program in, just increasingly at
There is cost-benefit new technology for substitution conventional measurement techniques (such as photogrammetric).Laser radar, which can generate, to be passed through
The discrete multiple spot cloud data set that ArcGIS is managed, shows, analyzes and shares, from the laser thunder for spatially carrying out tissue
It is referred to as point cloud data up to data.Initial point cloud is the big collection of 3D elevational point comprising x value, y value, z value and GPS time
Other attributes such as stamp.Initial laser radar points cloud after post treatment after, the particular surface element that laser encounters can be divided
Class, the object that ground, building, Forest Canopy, highway and any laser beam encounter in measurement process constitute a little
Cloud data.
In the present embodiment, the point cloud obtained according to laser measurement principle, including three-dimensional coordinate (XYZ) and laser reflection intensity
(Intensity), it carries out pose to multiframe point cloud data to optimize to obtain optimum results, pose optimization, which refers to, passes through laser radar
SLAM algorithm such as LOAM carries out pose optimization to the history multiframe laser radar data, and the whole concept of LOAM is exactly will be complicated
SLAM problem be divided into: 1, the estimation of high frequency;2, the environment of low frequency builds figure;Laser radar receives data, carries out first a little
Cloud registration, laser radar range carries out estimation with the frequency of 10Hz and coordinate is converted, and laser radar charts with the frequency of 1Hz
Three-dimensional map is constructed, transform integrals completes the optimization of pose, and structure parallel in this way ensure that the real-time of system.
Step S20 locally builds figure based on multiframe point cloud data progress cloud and obtains local map.
In the present embodiment, optimum results are obtained after carrying out pose optimization to multiframe point cloud data, and be based on optimum results
Progress cloud locally builds figure and obtains local map.The detailed process of figure is built in part are as follows: seeks multiframe point cloud first with laser SLAM technology
Pose, then be filtered road surface, only retain the point fallen on the ground.It will be fallen other than point cloud data on the ground when filtering
Point cloud data all weeds out, and retains the point cloud data fallen on the ground, such as: it is 1.7 that laser radar, which is mounted on apart from ground level,
The position of rice, in scanning, point cloud data can be fallen in the three-dimensional space between -1.7 meters or so laser radar, it would be desirable to
In addition to the point cloud data fallen on road surface remains, in reprojection to X/Y plane, then carries out rasterizing and handle to obtain each grid
The reflected intensity mean value of lattice is weighted and averaged the reflected intensity mean value of each grid, obtains a local map.
The format of the local map is converted to picture format by step S30.
In the present embodiment, by the way that local map is converted to picture format, the local map of picture format is obtained, it is available
The histogram equalizing method of OpenCV enhances contrast, enhances the contrast of lane line and road surface, convenient for detection lane line.
OpenCV be one based on BSD license (open source) issue cross-platform computer vision library, may operate in Linux,
In Windows, Android and MacOS operating system.Histogram equalization is the gray-scale distribution by adjusting image, so that 0
Distribution in~255 grayscale is more balanced, improves the contrast of image, achievees the purpose that improve image subjective vision effect,
The lower image of contrast is suitble to enhance image detail using histogram equalization method.
Step S40 detects lane line in the local map of picture format, obtains lane line data.
In the present embodiment, after obtaining local map, lane line is detected in local map.Local map is carried out
Filtering filtering, binaryzation and histogram equalization, enhance contrast, then corner feature and local edge are utilized in local map
Lane line is detected, to obtain lane line data.
In the present invention, multiframe point cloud data is obtained, obtains multiframe point cloud after carrying out pose optimization to multiframe point cloud data
Position, the corresponding figure of obtained point cloud data are more acurrate;Figure is locally built based on optimum results progress cloud and obtains local map, is realized
The format of local map is converted to picture format by the filtering to point cloud data, convenient for carrying out histogram equalization to local map
Change processing realizes inspection due to obtaining multiframe point cloud data using laser measurement principle to realize the detection to lane line
Measuring car diatom, blind area when can be to avoid single-frame images or single frames point cloud detection lane line, detects larger range of vehicle
Diatom data, and algorithm comparison is simple.Through the invention, lane line is detected using multiframe point cloud data, is detected
The bigger lane line data of range, lane line is continuous, reduces false detection rate.
Further, in one embodiment of method for detecting lane lines of the present invention, the step S10 includes:
Obtain the history multiframe laser radar data of buffer area storage;
Multiframe point cloud data is obtained based on the history multiframe laser radar data.
In the present embodiment, history multiframe laser radar data is stored using buffer area, is based on history multiframe radar data
Obtain multiframe point cloud data.Since radar laser is scanned apart from certain pavement-height position, when radar laser scans
Real time data be unable to satisfy the acquisition of multiframe point cloud data, a buffer area is set, for storing history multiframe laser radar
Data enable and obtain multiframe laser point cloud data based on history multiframe laser radar data.
Further, in one embodiment of method for detecting lane lines of the present invention, the step S10 further includes including:
By laser radar SLAM algorithm to the location information and attitude angle of frame point cloud each in the multiframe point cloud data
It optimizes, obtains the position of all frame point clouds.
In the present embodiment, multiframe point cloud data pose is optimized specifically: by laser radar SLAM algorithm to multiframe point
The location information of each frame point cloud and attitude angle optimize in cloud data, obtain the position of all frame point clouds.ORB-SLAM makees
For monocular SLAM, precision is largely determined by the whether accurate of the optimization of the pose between frame and frame.Therefore optimization exists
Critically important role is played inside ORB-SLAM.Because camera calibration and the precision of tracking are inadequate.The mistake of camera calibration
Difference can embody in the reconstruction (such as when trigonometry reconstruction), and the error tracked can then be embodied in the pose between different key frames
In, and in reconstruction (monocular).The pose that the continuous accumulation of error will lead to subsequent frames is more and more remoter from attained pose, eventually limits
The precision of system entirety processed.No matter in monocular, binocular or RGBD, the pose tracked all has error.Monocular
In SLAM, if there are enough characteristic points between two frames, the pose between two frames can be both directly obtained, can also have been passed through
An optimization problem is solved to obtain.Due to the uncertainty of monocular mesoscale, the error of scale can be also introduced.Since tracking obtains
Always relative pose, before the error of a certain frame can pass up to and go below, leading to tracking, have can for position and attitude error to the end
It can be very big.In order to improve the precision of tracking, closed loop detection can also can use in part and global optimization pose to optimize
Pose.
Further, it is based on first embodiment, the second embodiment of method for detecting lane lines of the present invention is proposed, such as Fig. 3 institute
Show, step S20 includes:
S21 filters the multiframe point cloud data and obtains the first multiframe point cloud data, wherein the first multiframe point cloud number
According to fall point cloud data on the ground;
S22 projects on X/Y plane the first multiframe point cloud data, then carries out rasterizing and handle to obtain each grid
Reflected intensity mean value;
S23 obtains local map according to the reflected intensity mean value.
It in the present embodiment, filters the optimum results and obtains the first multiframe point cloud data, wherein the first multiframe point cloud
Data are the point cloud data fallen on the ground;The first multiframe point cloud data is projected on X/Y plane, then carries out rasterizing
Processing obtains the reflected intensity mean value of each grid;Local map is obtained according to the reflected intensity mean value.Specifically: it will fall in
Point cloud data on ground retains, and the point cloud data not fallen on ground is weeded out;For the point cloud data fallen on the ground,
Rasterizing processing is carried out in reprojection to ground.Since the mass cloud data that laser radar obtains is irregular discrete points
According to collection, the geometrical relationship between data point is irregular, during the digital product that rasterizing is generated using these discrete points datas
Need to carry out the neighborhood search and sequence of grid point.Assuming that the fluctuation of the Z-direction of ground relatively flat, i.e. ground based scanning point compared with
It is small, by the way that scanning area is carried out grid division, scanning element cloud is projected into x/y plane, in terms of z-axis in statistics grid
The difference of highs and lows judges whether the point in grid is ground point or obstacle object point;Obstacle quality testing is mapped in grid
After survey, barrier is clustered, to realize that barrier is divided, obtains the ground separated with barrier.First by point in this case
After cloud data filtering, until only retaining the point cloud data fallen on the ground, the point cloud data for projecting to x/y plane is subjected to grid
Change processing, directly obtains the reflected intensity mean value of each grid, effectively avoids the processing to aerial point cloud data, so that processed
Cheng Gengwei is simple, also, obtains local map according to the reflected intensity mean value, completes part and builds figure, also, the local map
For the local map of a reflection intensity values.
In the present embodiment, in such a way that rasterizing handles point cloud data, processing mode is simple, and is conducive on detection ground
Lane line.
Further, in the second embodiment of method for detecting lane lines of the present invention, step S21 includes:
The second multiframe point cloud data in the multiframe point cloud data is rejected, to obtain the first multiframe point cloud data,
Wherein, the second multiframe point cloud data is the point cloud data not fallen on ground.
In the present embodiment, the second multiframe point cloud data is extra multiframe point cloud data, needs to weed out, due to being only
To the Point Cloud Processing fallen to the ground, so that treatment process is simpler.
Further, step S23 includes: in the second embodiment of method for detecting lane lines of the present invention
The reflected intensity mean value of each grid is weighted and averaged to obtain reflection intensity values, is based on the reflection intensity values
Obtain local map.
In the present embodiment, local map is obtained according to reflected intensity mean value specifically: equal to the reflected intensity of each grid
Value is weighted and averaged to obtain reflection intensity values, and obtains local map according to reflection intensity values.Due to being only to falling on ground
Point Cloud Processing on face, so that treatment process is simpler.Therefore, the reflected intensity mean value of each grid is added
It is simpler and be easily achieved that weight average obtains the processing modes of reflection intensity values.
Further, it is based on first embodiment, the 3rd embodiment of method for detecting lane lines of the present invention is proposed, such as Fig. 4 institute
Show, step S40 includes:
S41 carries out picture noise to the local map of described image format and map image is obtained by filtration;
S42 obtains histogram image after carrying out binaryzation and histogram equalization to the map image;
S43 carries out lane detection to the histogram image using corner feature and edge feature, obtains lane line number
According to.
In the present embodiment, picture noise is carried out to the local map of described image format, map image is obtained by filtration;To institute
It states after map image carries out binaryzation and histogram equalization and obtains histogram image;Using corner feature and edge feature to described
Histogram image carries out lane detection, obtains lane line data.
Picture noise can be mainly divided into impulsive noise and Gaussian noise, and the image that we see in reality generally all can
Containing noise, when doing subsequent processing to image, denoising, so also referred to as image filtering, image are carried out to image
Filtering can be divided into filter in spatial domain, frequency filtering two major classes.It is obtained after being filtered by the local map to picture format
Map image, and obtained map image edge and profile are more obvious;To map image progress binary conversion treatment, the two of image
Value exactly sets 0 or 255 for the gray value of the pixel on image, that is, whole image is showed apparent
There is black and white visual effect, since the map image being obtained by filtration is gray level image, the processing of to map image binaryzation will
Greyscale image transitions are bianry image, the pixel grey scale for being greater than some threshold grey scale value are set as gray scale maximum, less than this
The pixel grey scale of a value is set as gray scale minimum, to realize binaryzation.By binary conversion treatment map image, lane line is obtained
General location.
Since the effect that our picture seems is not so clear, image can at this time be performed some processing
Carry out the range that enlarged image pixel value is shown.Such as some image whole pixel values are relatively low, some features in image are seen not
It is to be apparent, only indistinctly sees some profile traces, at this moment image can be made to see after image histogram equalization
Get up brighter, is also convenient for subsequent processing.Histogram equalization is an important application of greyscale transformation, it is efficient and is easy to
It realizes, is widely used in image enhancement processing.The pixel grey scale variation of image be it is random, the figure of histogram height is not
Together, histogram equalization is exactly the method for keeping histogram substantially gentle with certain algorithm.Histogram in the opencv that this case uses
The input picture of figure equalization algorithm need to be eight single channel images, that is, gray level image.If wanting place calculates color image
Equalization figure, image first can be subjected to channel separation with split function, handle the image in each channel respectively, with
Merge function merges, and obtains histogram image.
The edge detection of image refers to the set of those of its surrounding pixel gray scale change dramatically pixel, it is image most base
This feature, and the edge detection of image is the marginal point of first detection image, and marginal point is connected into wheel according still further to certain strategy
Exterior feature, to constitute cut zone;Angle point is image important feature, and the understanding and analysis to image graphics have critically important work
With.Angle point can be effectively reduced the data volume of information, make the content of its information while retaining image graphics important feature
It is very high, the speed of calculating is effectively improved, the reliable matching of image is conducive to, so that being treated as possibility in real time.Angle point exists
The computer vision fields such as 3 D scene rebuilding, estimation, target following, target identification, image registration and matching play non-
Normal important role.Current Corner Detection Algorithm can be summarized as 3 classes: Corner Detection based on gray level image is based on binary map
The Corner Detection of picture, the Corner Detection based on contour curve.Angle point is image important feature, to the understanding of image graphics and
Analysis plays a very important role.The Corner Detection Algorithm of gray level image, bianry image, edge contour curve is summarized, point
Relevant algorithm has been analysed, and evaluation is given to various detection algorithms.This case utilizes lane in Edge Gradient Feature histogram image
The edge of line obtains lane line data in conjunction with corner feature, and lane line is continuously uninterrupted.
Further, after step 40, method for detecting lane lines further include:
Based on the lane line data, the fitting of straight line and curve is carried out by Hough transformation, fitting result is obtained, passes through
Fitting result determines travelable region.
In the present embodiment, Hough transformation (Hough) is the method for a very important detection discontinuous point boundary shape, it
By the way that image coordinate space is transformed to parameter space, to realize the fitting of straight line and curve.By Hough transformation to lane line
Data realize the fitting of straight line and curve, realize the determination that can travel region.
In addition, the embodiment of the present invention also proposes a kind of computer readable storage medium, the computer readable storage medium
On be stored with lane detection program, the lane detection program realizes lane line inspection as described above when being executed by processor
The step of survey method.
Each embodiment of the specific embodiment of computer readable storage medium of the present invention and above-mentioned method for detecting lane lines
Essentially identical, this will not be repeated here.
It should be noted that, in this document, the terms "include", "comprise" or its any other variant are intended to non-row
His property includes, so that the process, method, article or the system that include a series of elements not only include those elements, and
And further include other elements that are not explicitly listed, or further include for this process, method, article or system institute it is intrinsic
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including being somebody's turn to do
There is also other identical elements in the process, method of element, article or system.
The serial number of the above embodiments of the invention is only for description, does not represent the advantages or disadvantages of the embodiments.
Through the above description of the embodiments, those skilled in the art can be understood that above-described embodiment side
Method can be realized by means of software and necessary general hardware platform, naturally it is also possible to by hardware, but in many cases
The former is more preferably embodiment.Based on this understanding, technical solution of the present invention substantially in other words does the prior art
The part contributed out can be embodied in the form of software products, which is stored in one as described above
In storage medium (such as ROM/RAM, magnetic disk, CD), including some instructions are used so that terminal device (it can be mobile phone,
Computer, server, air conditioner or network equipment etc.) execute method described in each embodiment of the present invention.
The above is only a preferred embodiment of the present invention, is not intended to limit the scope of the invention, all to utilize this hair
Equivalent structure or equivalent flow shift made by bright specification and accompanying drawing content is applied directly or indirectly in other relevant skills
Art field, is included within the scope of the present invention.
Claims (10)
1. a kind of method for detecting lane lines, which is characterized in that the method for detecting lane lines includes:
Multiframe point cloud data is obtained, obtains the position of all frame point clouds after carrying out pose optimization to the multiframe point cloud data;
Figure, which is locally built, based on multiframe point cloud data progress cloud obtains local map;
The format of the local map is converted into picture format;
Lane line is detected in the local map of picture format, obtains lane line data.
2. method for detecting lane lines as described in claim 1, which is characterized in that the step of acquisition multiframe point cloud data wraps
It includes:
Obtain the history multiframe laser radar data of buffer area storage;
Multiframe point cloud data is obtained based on the history multiframe laser radar data.
3. method for detecting lane lines as described in claim 1, which is characterized in that described to carry out position to the multiframe point cloud data
Appearance optimizes the step of obtaining optimum results and includes:
It is carried out by location information and attitude angle of the laser radar SLAM algorithm to frame point cloud each in the multiframe point cloud data
Optimization, obtains the position of all frame point clouds.
4. method for detecting lane lines as described in claim 1, which is characterized in that described to be carried out based on the multiframe point cloud data
Cloud locally builds the step of figure obtains local map
It filters the multiframe point cloud data and obtains the first multiframe point cloud data, wherein the first multiframe point cloud data is described
Fall point cloud data on the ground;
The first multiframe point cloud data is projected on X/Y plane, then carry out rasterizing handle to obtain each grid reflection it is strong
Spend mean value;
Local map is obtained according to the reflected intensity mean value.
5. method for detecting lane lines as claimed in claim 4, which is characterized in that obtain the first multiframe point cloud described in the filtering
The step of data includes:
The second multiframe point cloud data of the multiframe point cloud data is rejected, to obtain the first multiframe point cloud data, wherein institute
Stating the second multiframe point cloud data is the point cloud data not fallen in the multiframe point cloud data on the ground.
6. method for detecting lane lines as claimed in claim 4, which is characterized in that described to be obtained according to the reflected intensity mean value
The step of local map includes:
The reflected intensity mean value of each grid is weighted and averaged to obtain reflection intensity values, is obtained based on the reflection intensity values
Local map.
7. method for detecting lane lines as described in claim 1, which is characterized in that described to be examined in the local map of picture format
Measuring car diatom, the step of obtaining lane line data include:
Picture noise is carried out to the local map of picture format, map image is obtained by filtration;
Histogram image is obtained after carrying out binaryzation and histogram equalization to the map image;
Lane detection is carried out to the histogram image using corner feature and edge feature, obtains lane line data.
8. the method for detecting lane lines as described in any one of claims 1 to 7, which is characterized in that described in picture format
After the step of detecting lane line in local map, obtaining lane line data, the method for detecting lane lines further include:
Based on the lane line data, the fitting of straight line and curve is carried out by Hough transformation, obtains fitting result;
Travelable region is determined by fitting result.
9. a kind of lane detection device, which is characterized in that the lane detection device includes: memory, processor and deposits
The lane detection program that can be run on the memory and on the processor is stored up, the lane detection program is by institute
State the step of realizing the method for detecting lane lines as described in claim 1 to 7 when processor executes.
10. a kind of computer readable storage medium, which is characterized in that be stored with lane line on the computer readable storage medium
Program is detected, such as lane described in any item of the claim 1 to 8 is realized when the lane detection program is executed by processor
The step of line detecting method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810787347.3A CN108985230A (en) | 2018-07-17 | 2018-07-17 | Method for detecting lane lines, device and computer readable storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810787347.3A CN108985230A (en) | 2018-07-17 | 2018-07-17 | Method for detecting lane lines, device and computer readable storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108985230A true CN108985230A (en) | 2018-12-11 |
Family
ID=64549565
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810787347.3A Pending CN108985230A (en) | 2018-07-17 | 2018-07-17 | Method for detecting lane lines, device and computer readable storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108985230A (en) |
Cited By (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109766404A (en) * | 2019-02-12 | 2019-05-17 | 湖北亿咖通科技有限公司 | Points cloud processing method, apparatus and computer readable storage medium |
CN109870706A (en) * | 2019-01-31 | 2019-06-11 | 深兰科技(上海)有限公司 | A kind of detection method of road surface identification, device, equipment and medium |
CN110008851A (en) * | 2019-03-15 | 2019-07-12 | 深兰科技(上海)有限公司 | A kind of method and apparatus of lane detection |
CN110097025A (en) * | 2019-05-13 | 2019-08-06 | 奇瑞汽车股份有限公司 | Detection method, device and the storage medium of lane line |
CN110675436A (en) * | 2019-09-09 | 2020-01-10 | 中国科学院微小卫星创新研究院 | Laser radar and stereoscopic vision registration method based on 3D feature points |
CN111080662A (en) * | 2019-12-11 | 2020-04-28 | 北京建筑大学 | Lane line extraction method and device and computer equipment |
CN111316284A (en) * | 2019-02-13 | 2020-06-19 | 深圳市大疆创新科技有限公司 | Lane line detection method, device and system, vehicle and storage medium |
CN111316328A (en) * | 2019-04-24 | 2020-06-19 | 深圳市大疆创新科技有限公司 | Method for maintaining lane line map, electronic device and storage medium |
CN111508021A (en) * | 2020-03-24 | 2020-08-07 | 广州视源电子科技股份有限公司 | Pose determination method and device, storage medium and electronic equipment |
CN111598034A (en) * | 2020-05-22 | 2020-08-28 | 知行汽车科技(苏州)有限公司 | Obstacle detection method, obstacle detection device and storage medium |
CN112215144A (en) * | 2020-10-12 | 2021-01-12 | 北京四维智联科技有限公司 | Method and system for processing lane line |
WO2021051344A1 (en) * | 2019-09-19 | 2021-03-25 | 深圳市大疆创新科技有限公司 | Method and apparatus for determining lane lines in high-precision map |
CN112578404A (en) * | 2019-09-27 | 2021-03-30 | 北京地平线机器人技术研发有限公司 | Method and device for determining driving path |
CN112988922A (en) * | 2019-12-16 | 2021-06-18 | 长沙智能驾驶研究院有限公司 | Perception map construction method and device, computer equipment and storage medium |
US20210270612A1 (en) * | 2020-03-02 | 2021-09-02 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method, apparatus, computing device and computer-readable storage medium for positioning |
CN113628335A (en) * | 2021-07-28 | 2021-11-09 | 深圳优艾智合机器人科技有限公司 | Point cloud map construction method and device and computer readable storage medium |
CN113888621A (en) * | 2021-09-29 | 2022-01-04 | 中科海微(北京)科技有限公司 | Loading rate determining method, loading rate determining device, edge computing server and storage medium |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106127113A (en) * | 2016-06-15 | 2016-11-16 | 北京联合大学 | A kind of road track line detecting method based on three-dimensional laser radar |
CN106908775A (en) * | 2017-03-08 | 2017-06-30 | 同济大学 | A kind of unmanned vehicle real-time location method based on laser reflection intensity |
CN107316328A (en) * | 2017-05-11 | 2017-11-03 | 浙江大学 | A kind of closed loop detection method based on two dimensional laser scanning instrument Corner Feature |
CN108171131A (en) * | 2017-12-15 | 2018-06-15 | 湖北大学 | Based on the Lidar point cloud data road marking line extracting methods and system for improving MeanShift |
CN108280866A (en) * | 2016-12-30 | 2018-07-13 | 乐视汽车(北京)有限公司 | Road Processing Method of Point-clouds and system |
-
2018
- 2018-07-17 CN CN201810787347.3A patent/CN108985230A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106127113A (en) * | 2016-06-15 | 2016-11-16 | 北京联合大学 | A kind of road track line detecting method based on three-dimensional laser radar |
CN108280866A (en) * | 2016-12-30 | 2018-07-13 | 乐视汽车(北京)有限公司 | Road Processing Method of Point-clouds and system |
CN106908775A (en) * | 2017-03-08 | 2017-06-30 | 同济大学 | A kind of unmanned vehicle real-time location method based on laser reflection intensity |
CN107316328A (en) * | 2017-05-11 | 2017-11-03 | 浙江大学 | A kind of closed loop detection method based on two dimensional laser scanning instrument Corner Feature |
CN108171131A (en) * | 2017-12-15 | 2018-06-15 | 湖北大学 | Based on the Lidar point cloud data road marking line extracting methods and system for improving MeanShift |
Non-Patent Citations (5)
Title |
---|
ALBERTO HATA ET AL.: "Road Marking Detection Using LIDAR Reflective Intensity Data and its Application to Vehicle Localization", 《2014 IEEE 17TH INTERNATIONAL CONFERENCE ON INTELLIGENT TRANSPORTATION SYSTEMS》 * |
HAIYAN GUAN ET AL.: "Using mobile laser scanning data for automated extraction of road markings", 《ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING》 * |
JI ZHANG ET AL.: "LOAM: Lidar Odometry and Mapping in Real-time", 《RESEARCHGATE》 * |
于宁波等: "一种基于RGB-D 的移动机器人未知室内环境自主探索与地图构建方法", 《机器人ROBOT》 * |
杨必胜,方莉娜: "车载激光扫描点云中道路标识线提取", 《第一届全国激光雷达对地观测高级学术研讨会》 * |
Cited By (24)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109870706A (en) * | 2019-01-31 | 2019-06-11 | 深兰科技(上海)有限公司 | A kind of detection method of road surface identification, device, equipment and medium |
CN109766404A (en) * | 2019-02-12 | 2019-05-17 | 湖北亿咖通科技有限公司 | Points cloud processing method, apparatus and computer readable storage medium |
CN111316284A (en) * | 2019-02-13 | 2020-06-19 | 深圳市大疆创新科技有限公司 | Lane line detection method, device and system, vehicle and storage medium |
CN110008851A (en) * | 2019-03-15 | 2019-07-12 | 深兰科技(上海)有限公司 | A kind of method and apparatus of lane detection |
CN110008851B (en) * | 2019-03-15 | 2021-11-19 | 深兰科技(上海)有限公司 | Method and equipment for detecting lane line |
CN111316328A (en) * | 2019-04-24 | 2020-06-19 | 深圳市大疆创新科技有限公司 | Method for maintaining lane line map, electronic device and storage medium |
CN110097025A (en) * | 2019-05-13 | 2019-08-06 | 奇瑞汽车股份有限公司 | Detection method, device and the storage medium of lane line |
CN110097025B (en) * | 2019-05-13 | 2023-08-04 | 奇瑞汽车股份有限公司 | Lane line detection method, device and storage medium |
CN110675436A (en) * | 2019-09-09 | 2020-01-10 | 中国科学院微小卫星创新研究院 | Laser radar and stereoscopic vision registration method based on 3D feature points |
WO2021051344A1 (en) * | 2019-09-19 | 2021-03-25 | 深圳市大疆创新科技有限公司 | Method and apparatus for determining lane lines in high-precision map |
CN112578404A (en) * | 2019-09-27 | 2021-03-30 | 北京地平线机器人技术研发有限公司 | Method and device for determining driving path |
CN112578404B (en) * | 2019-09-27 | 2022-10-04 | 北京地平线机器人技术研发有限公司 | Method and device for determining driving path |
CN111080662A (en) * | 2019-12-11 | 2020-04-28 | 北京建筑大学 | Lane line extraction method and device and computer equipment |
CN112988922A (en) * | 2019-12-16 | 2021-06-18 | 长沙智能驾驶研究院有限公司 | Perception map construction method and device, computer equipment and storage medium |
US11852751B2 (en) * | 2020-03-02 | 2023-12-26 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method, apparatus, computing device and computer-readable storage medium for positioning |
US20210270612A1 (en) * | 2020-03-02 | 2021-09-02 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method, apparatus, computing device and computer-readable storage medium for positioning |
CN111508021A (en) * | 2020-03-24 | 2020-08-07 | 广州视源电子科技股份有限公司 | Pose determination method and device, storage medium and electronic equipment |
CN111508021B (en) * | 2020-03-24 | 2023-08-18 | 广州视源电子科技股份有限公司 | Pose determining method and device, storage medium and electronic equipment |
CN111598034B (en) * | 2020-05-22 | 2021-07-23 | 知行汽车科技(苏州)有限公司 | Obstacle detection method, obstacle detection device and storage medium |
CN111598034A (en) * | 2020-05-22 | 2020-08-28 | 知行汽车科技(苏州)有限公司 | Obstacle detection method, obstacle detection device and storage medium |
CN112215144A (en) * | 2020-10-12 | 2021-01-12 | 北京四维智联科技有限公司 | Method and system for processing lane line |
CN112215144B (en) * | 2020-10-12 | 2024-05-14 | 北京四维智联科技有限公司 | Method and system for processing lane lines |
CN113628335A (en) * | 2021-07-28 | 2021-11-09 | 深圳优艾智合机器人科技有限公司 | Point cloud map construction method and device and computer readable storage medium |
CN113888621A (en) * | 2021-09-29 | 2022-01-04 | 中科海微(北京)科技有限公司 | Loading rate determining method, loading rate determining device, edge computing server and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108985230A (en) | Method for detecting lane lines, device and computer readable storage medium | |
CN105825173B (en) | General road and lane detection system and method | |
US10049492B2 (en) | Method and apparatus for rendering facades of objects of interest from three-dimensional point clouds | |
Vandapel et al. | Natural terrain classification using 3-d ladar data | |
CN108052624A (en) | Processing Method of Point-clouds, device and computer readable storage medium | |
US20160154999A1 (en) | Objection recognition in a 3d scene | |
US8086046B2 (en) | Image analysis device and image analysis method | |
Negru et al. | Image based fog detection and visibility estimation for driving assistance systems | |
CN107767400B (en) | Remote sensing image sequence moving target detection method based on hierarchical significance analysis | |
CN108960183A (en) | A kind of bend target identification system and method based on Multi-sensor Fusion | |
CN109738910A (en) | A kind of curb detection method based on three-dimensional laser radar | |
CN107330378A (en) | A kind of driving behavior detecting system based on embedded image processing | |
JP2021536071A (en) | Obstacle detection method, intelligent driving control method, device, medium, and equipment | |
CN109492714A (en) | Image processing apparatus and its method | |
CN110232389A (en) | A kind of stereoscopic vision air navigation aid based on green crop feature extraction invariance | |
CN113378760A (en) | Training target detection model and method and device for detecting target | |
Zhang | Rapid inspection of pavement markings using mobile laser scanning point clouds | |
Zelener et al. | Cnn-based object segmentation in urban lidar with missing points | |
CN111524113A (en) | Lifting chain abnormity identification method, system, equipment and medium | |
CN105893970A (en) | Nighttime road vehicle detection method based on luminance variance characteristics | |
CN109063669A (en) | A kind of bridge zone ship's navigation Situation analysis method and device based on image recognition | |
Kusetogullari et al. | Unsupervised change detection in landsat images with atmospheric artifacts: a fuzzy multiobjective approach | |
CN106169086B (en) | High-resolution optical image under navigation data auxiliary damages method for extracting roads | |
CN108388854A (en) | A kind of localization method based on improvement FAST-SURF algorithms | |
Ballinas-Hernández et al. | Marked and unmarked speed bump detection for autonomous vehicles using stereo vision |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181211 |