CN108036794A - A kind of high accuracy map generation system and generation method - Google Patents

A kind of high accuracy map generation system and generation method Download PDF

Info

Publication number
CN108036794A
CN108036794A CN201711187081.0A CN201711187081A CN108036794A CN 108036794 A CN108036794 A CN 108036794A CN 201711187081 A CN201711187081 A CN 201711187081A CN 108036794 A CN108036794 A CN 108036794A
Authority
CN
China
Prior art keywords
road
tracing point
data
coordinates
map
Prior art date
Application number
CN201711187081.0A
Other languages
Chinese (zh)
Inventor
郭剑鹰
陈华宇
刘炳祥
郑艳
Original Assignee
华域汽车系统股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 华域汽车系统股份有限公司 filed Critical 华域汽车系统股份有限公司
Priority to CN201711187081.0A priority Critical patent/CN108036794A/en
Publication of CN108036794A publication Critical patent/CN108036794A/en

Links

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/26Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
    • G01C21/28Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network with correlation of data from several navigational instruments
    • G01C21/30Map- or contour-matching
    • G01C21/32Structuring or formatting of map data
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/26Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
    • G01C21/34Route searching; Route guidance
    • G01C21/3407Route searching; Route guidance specially adapted for specific applications
    • G01C21/3415Dynamic re-routing, e.g. recalculating the route when the user deviates from calculated route or after detecting real-time traffic data or accidents
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01SRADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S19/00Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
    • G01S19/38Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system
    • G01S19/39Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system the satellite radio beacon positioning system transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
    • G01S19/40Correcting position, velocity or attitude
    • G01S19/41Differential correction, e.g. DGPS [differential GPS]
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01SRADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S19/00Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
    • G01S19/38Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system
    • G01S19/39Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system the satellite radio beacon positioning system transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
    • G01S19/42Determining position
    • G01S19/45Determining position by combining measurements of signals from the satellite radio beacon positioning system with a supplementary measurement

Abstract

The present invention relates to a kind of high-precision map generation system and method, wherein, system includes:One data acquisition module, the view data for the vehicle front road scene that its vehicle DGPS data provided according to DGPS systems and camera provide, obtains the running data of each tracing point and the plane xy coordinates for the tracing point being distributed on road axis on vehicle driving trace;One data processing module being connected with the data acquisition module, it is according to the running data of each tracing point and the plane xy Coordinate generation road axis of the tracing point being distributed on road axis, and road topology relation is generated according to the road axis, to obtain the structural information of whole map;And the map display module being connected with the data processing module.The present invention is by the way that the image information obtained above by camera and DGPS information are merged, it is achieved thereby that generating the purpose of high-precision map with lower cost solution.

Description

A kind of high accuracy map generation system and generation method
Technical field
The present invention relates to a kind of high-precision map generation system and generation method for garden pilotless automobile.
Background technology
It is well known that, it is necessary to which map making is automobile navigation in the R&D process of automatic driving vehicle.It is existing at present In technology, the method for map is generated mainly including following several:
1st, the method disclosed in patent document CN105579811A, utilizes external image, Inertial Measurement Unit and wireless E measurement technology GNSS (Global Navigation Satellite System, Global Satellite Navigation System) and other not wireless The data fusion of measurement sensor measurement is as a result, determine circuit track of the user in global coordinates, and combine this track and answer With vSLAM (Visual simultaneous localization and mapping, vision positioning immediately and map structure Build) image characteristic point of technology, obtain planar structure and (or) stereochemical structure map.
2nd, the method disclosed in patent document CN104573733A, using image taking module, horizontal laser light radar, GPS processing modules, inertial navigation module, rotary encoder and image pre-processing module and geographic information processing module next life Into map.
3rd, the method disclosed in patent document CN106441319A, utilizes satellite photo, onboard sensor (laser radar And camera), integrated positioning system (global position system and inertial navigation system) obtain original road data fusion generation ground Figure.
In the above method, the precision of map all is improved with relatively more sensors, but multisensor is brought The problem of be raising in cost, the especially cost of laser radar is differed from tens thousand of to hundreds thousand of.
The content of the invention
In order to solve the above-mentioned problems of the prior art, the present invention is intended to provide it is a kind of high accuracy map generation system and Generation method, while map generation precision is improved, effectively to reduce cost.
A kind of high-precision map generation system described in one of present invention, it includes:
Before the vehicle that one data acquisition module, its vehicle DGPS data provided according to DGPS systems and camera provide The view data of square road scene, obtains the running data of each tracing point on vehicle driving trace, each tracing point Running data includes:Plane xy coordinates, speed, course angle and tracing point of the tracing point to left and right sides lane line away from From;
One data processing module being connected with the data acquisition module, it is according to the running data of each tracing point Road axis is generated, and road topology relation is generated according to the road axis, to obtain the structural information of whole map; And
One map display module being connected with the data processing module, it is according to the structural information wash with watercolours of the whole map Dye roadway, and Real time displaying current vehicle position.
In above-mentioned high-precision map generation system, the data acquisition module includes:
The one DGPS information process units being connected with the DGPS systems, it parses the vehicle DGPS data, To extract latitude and longitude coordinates, speed and the course angle of each tracing point, and by gauss projection by each track The latitude and longitude coordinates of point are converted to the plane xy coordinates of each tracing point;And
The one camera information processing unit being connected with the camera, its image to the vehicle front road scene Data carry out image preprocessing, lane detection and lane line tracking, and according to default image coordinate and world coordinates it Between mapping relations, obtain each tracing point to the distance of left and right sides lane line.
In above-mentioned high-precision map generation system, the data processing module includes:
The one map vector generation unit being connected with the data acquisition module, it is according to the traveling of each tracing point Data, filter out bad point and shift point in the tracing point, and are fitted and are generated in the road after remaining tracing point is clustered Heart line;And
The one topological map generation unit being connected with the map vector generation unit, it is carried according to the road axis By way of crossing and dead end street information, to define road data, the road topology relation is generated.
In above-mentioned high-precision map generation system, the data processing module further includes one and is given birth to the topological map Into the characteristics map generation unit of unit connection, it is used to add map location mark and roadway characteristic mark on whole map Note.
In above-mentioned high-precision map generation system, further include one and be connected to the camera and the data acquisition module Camera information receiving module between block, and a DGPS being connected between the DPGS and the data acquisition module believe Cease receiving module.
In above-mentioned high-precision map generation system, further include one be connected with the data acquisition module be used for provide The camera calibration module of mapping relations between described image coordinate and world coordinates.
A kind of accurately drawing generating method described in the two of the present invention, it comprises the following steps:
The vehicle front road scene that step S1, the vehicle DGPS data provided according to DGPS systems and camera provide View data, obtain the running data of each tracing point on vehicle driving trace, the running data bag of each tracing point Include:Distance of plane xy coordinates, speed, course angle and the tracing point of the tracing point to left and right sides lane line;
Step S2, road axis is generated according to the running data of each tracing point, and according to the road-center Line generates road topology relation, to obtain the structural information of whole map;And
Step S3, roadway, and Real time displaying vehicle present bit are rendered to according to the structural information of the whole map Put.
In above-mentioned accurately drawing generating method, the step S1 includes:
Step S11, parses the vehicle DGPS data, extract each tracing point latitude and longitude coordinates, Speed and course angle, and the latitude and longitude coordinates of each tracing point are converted to by each tracing point by gauss projection Plane xy coordinates;And
Step S12, the view data of the vehicle front road scene is carried out image preprocessing, lane detection and Lane line tracks, and according to the mapping relations between default image coordinate and world coordinates, obtains each tracing point and arrive The distance of left and right sides lane line.
In above-mentioned accurately drawing generating method, the step S2 includes:
Step S21, according to the running data of each tracing point, filters out the bad point and shift point in the tracing point, And fitting generates the road axis after remaining tracing point is clustered;And
Step S22 is raw to define road data according to the road-center line drawing road mouth and dead end street information Into the road topology relation.
In above-mentioned accurately drawing generating method, the step S2 is further included:
Step S23, adds map location mark and roadway characteristic mark on whole map.
As a result of above-mentioned technical solution, the present invention (is used for FCW only with camera (or camera) with vehicle (Forward Collision Warning) forward direction collision warning, LDW (Lane departure warning) deviation report The forward sight camera multiplexing of the function such as alert) view data of road scene is obtained, cost is thus greatlyd save, and employ The double GPSs (that is, DGPS) different with existing GPS technology obtain vehicle DGPS data (absolute location information for including vehicle), Since DGPS contains master and slave two gps signals, the problems such as can be used for overcoming signal drift that single GPS may be brought from gps signal, Which thereby enhance signal accuracy.The present invention is by the way that the image information obtained above by camera and DGPS information are melted Close, it is achieved thereby that generating the purpose of high-precision map with lower cost solution.
Brief description of the drawings
Fig. 1 is a kind of structure diagram of high-precision map generation system of the present invention;
Fig. 2 is the structure diagram of data processing module in a kind of high-precision map generation system of the present invention;
Fig. 3 is the schematic diagram that the view data provided in the present invention according to camera corrects tracing point plane xy coordinates;
Fig. 4 is the schematic diagram that shift point is removed in the present invention;
Fig. 5 is the schematic diagram that vehicle is defined with road direction angle in the present invention.
Embodiment
Below in conjunction with the accompanying drawings, presently preferred embodiments of the present invention is provided, and is described in detail.
Please refer to Fig.1,2, one of present invention, i.e. it is a kind of high accuracy map generation system, including:Camera calibration module 1st, camera information receiving module 2, DGPS information receiving modules 3, data acquisition module 4, data processing module 5, data storage Module 6, map display module 7, central processing module 8 and driver module 9, wherein,
Camera calibration module 1 is connected with data acquisition module 4, camera 10 by camera information receiving module 2 with Data acquisition module 4 connects, and DGPS systems 20 are connected by DGPS information receiving modules 3 with data acquisition module 4, data acquisition Module 4 is connected by central processing module 8 with data processing module 5, and data processing module 5 passes through data memory module 6 and ground Figure display module 7 connects, and central processing module 8 is connected with map display module 7 and driver module 9 respectively, and display is driven Dynamic model block 9 is also connected with map display module 7.
Specifically, camera calibration module 1 is used to provide the mapping relations between image coordinate and world coordinates.
Data acquisition module 4 includes:The camera information processing unit 41 that is connected with camera information receiving module 2 and The DGPS information process units 42 being connected with DGPS information receiving modules 3, wherein,
The view data for the vehicle front road scene that camera information processing unit 41 provides camera 10 carries out figure As pretreatment, lane detection and lane line tracking etc. processing, and according to camera calibration module 1 provide image coordinate and Mapping relations between world coordinates, obtain on vehicle driving trace each tracing point to left and right sides lane line and road edge Distance;
DGPS information process units 42 parse the vehicle DGPS data that DGPS systems 20 provide, each to extract Latitude and longitude coordinates, speed and the course angle of tracing point, and be converted to the latitude and longitude coordinates of each tracing point by gauss projection The plane xy coordinates of each tracing point;
In addition, camera information processing unit 41 is also connected with DGPS information process units 42, so that according to each track Point is modified the plane xy coordinates of each tracing point to the distance of left and right sides lane line and road edge, to be distributed The plane xy coordinates of tracing point on heart line in the road.
Data processing module 5 includes:The map vector generation unit 51 and map vector being connected with data acquisition module 4 The topological map generation unit 52 and the characteristics map generation being connected with topological map generation unit 52 that generation unit 51 connects Unit 53, wherein,
Map vector generation unit 51 is according to the plane xy coordinates of the tracing point being distributed on road axis and each The running data of tracing point, speed, course angle and the tracing point of the tracing point being especially distributed across on road axis To the distance of left and right sides lane line and road edge etc., filter out the bad point in tracing point and shift point (in the present embodiment, may be used Realized by data prediction subelement 511), and generation road axis is fitted after remaining tracing point is clustered, i.e. generation The tracing point of orderly, equally distributed road-center (in the present embodiment, can be by connecting with data prediction subelement 511 The road axis generation subelement 512 connect is realized);
Topological map generation unit 52 is according to road-center line drawing road mouth and dead end street information (in the present embodiment In, can be realized by road Node extraction subelement 521), to define road data, the road topology relation is generated, that is, is carried Take generation road relation net (in the present embodiment, can be by the road topology relation that is connected with road Node extraction subelement 521 Generation subelement 522 is realized), to obtain the structural information of whole map, (including how many road, every road include in map Link relation of which tracing point, every road and other roads etc.);
Characteristics map generation unit 53 is used to add map location on whole map and mark (in the present embodiment, to lead to Map location mark subelement 531 is crossed to realize) and roadway characteristic mark (it is sub that roadway characteristic mark in the present embodiment, can be passed through Unit 532 is realized), wherein, map location mark includes fixed character position mark (such as position of gas station) and interim spy Sign mark (such as starting final position of navigation), the feature that roadway characteristic mark refers to add road according to traffic sign are believed Breath, for example, whether road restricts driving, road whether allow lane change etc. (in the present embodiment, roadway characteristic mark subelement 532 can use In definition road width and road).
Data memory module 6 is used for the cartographic information for storing the output of data processing module 5.
Map display module 7 is according to the structural information of the whole map stored in data memory module 6 (such as path locus Tracing point plane xy coordinates) be rendered to roadway, and Real time displaying current vehicle position.
Central processing module 8 is respectively that data acquisition module 4, data processing module 5, map display module 7 and display are driven Dynamic model block 9 provides command information, and (whether, for example, informing data acquisition module 4, whether its data format gathered is correct, stop Or continue the instruction such as gathered data), and the data that data acquisition module 4 is exported are supplied to data processing module 5.
Driver module 9 is used to provide driver to map denotation unit 7.
The two of the operation principle of said system, the i.e. present invention, i.e., a kind of accurately drawing generating method, including following step Suddenly:
The vehicle front road that step S1, the vehicle DGPS data provided according to DGPS systems 20 and camera 10 provide The view data of scene, obtains the running data of each tracing point on vehicle driving trace, the running data bag of each tracing point Include:Plane xy coordinates, speed, course angle and tracing point of the tracing point to left and right sides lane line and road edge away from From, and according to each tracing point to the distance of left and right sides lane line and road edge to the plane xy coordinates of each tracing point into Row is corrected, to obtain the plane xy coordinates for the tracing point being distributed on road axis;
Step S2, according to the running data of each tracing point and the plane xy for the tracing point being distributed on road axis Coordinate, generates road axis, and generates road topology relation according to road axis, to obtain the structure of whole map letter Breath;And
Step S3, roadway, and Real time displaying current vehicle position are rendered to according to the structural information of whole map.
In the present embodiment, above-mentioned steps S1 is specifically included:
Step S11, parses vehicle DGPS data, extracts the latitude and longitude coordinates, speed and boat of each tracing point To angle, and the latitude and longitude coordinates of each tracing point are converted to by gauss projection the plane xy coordinates of each tracing point;And
Step S12, image preprocessing, lane detection and track are carried out to the view data of vehicle front road scene Line tracks, and according to the mapping relations between default image coordinate and world coordinates, obtains each tracing point to the left and right sides The distance of lane line and road edge, and according to each tracing point to the distance of left and right sides lane line and road edge to each The plane xy coordinates of tracing point are modified, to obtain the plane xy coordinates for the tracing point being distributed on road axis.
Above-mentioned steps S2 is specifically included:
Step S21, according to the plane of the running data of each tracing point and the tracing point being distributed on road axis Xy coordinates, filter out bad point and shift point in tracing point, and generation road axis is fitted after remaining tracing point is clustered;With And
Step S22, according to road-center line drawing road mouth and dead end street information, to define road data, generates road Road topological relation;And
Step S23, adds map location mark and roadway characteristic mark on whole map.
In the present embodiment, camera information processing unit 41 is mainly used for the picture number according to vehicle front road scene Factually show lane detection and following function, which mainly includes image preprocessing, lane detection, lane line and track three Point, wherein, image preprocessing section point includes:Area-of-interest selection, gray processing, filtering, edge extracting and binary conversion treatment; Lane detection and tracking include:First it is detected that left and right two straight line nearest apart from selected tracing point, apart from this two Straight line of the air line distance in scope D is set as left and right straight line cluster, and it is (specific to be fitted left and right lane line according to the point in straight line cluster Reference can be made to the patent application of Application No. 201710568932.X《A kind of multilane detection method and tracking》, herein no longer It is described in detail).Camera information processing unit 41 is according to obtained above-mentioned left and right track line model and realizes to camera into rower Fixed result (mapping relations i.e. between image coordinate and world coordinates), you can obtain tracing point (i.e. vehicle body) and arrive the left and right sides The distance of lane line.
In the present embodiment, DGPS systems 20 send vehicle DGPS data with 5-20hz (frequency can set and fix), The valid data that DGPS information process units 42 therefrom extract include:(1) UTC time, hhmmss (Hour Minute Second);(2) shape is positioned State, A=are effectively positioned, the invalid positioning of V=;(3) latitude ddmm.mmmmm (degree point);(4) latitude hemisphere N (Northern Hemisphere) or S (south Hemisphere);(5) longitude dddmm.mmmmm (degree point);(6) longitude hemisphere E (east longitude) or W (west longitude);(7) ground speed (000.0 ~999.9 sections);(8) ground course (000.0~359.9 degree, with the positive north for reference data);(9) UTC dates, ddmmyy (day month year);(10) magnetic declination (000.0~180.0 degree, 0) then mend by leading digit deficiency;(11) magnetic biasing angular direction, E (east) or W (west);(12) pattern instruction (A=autonomous positionings, D=difference, E=estimations, N=data invalids).DGPS information process units 42 Above-mentioned latitude and longitude coordinates data are mainly converted to by the plane xy coordinate (meters on longitude and latitude in gauss projection by gauss projection Calculation method is known in the art general knowledge, and details are not described herein again).
In addition, in the present embodiment, during due to carrying out vehicle DGPS data acquisitions, DGPS systems 20 relative to road position Put (as shown in Figure 3) not being to determine.Therefore the view data according to camera is needed when generating road axis, to rail The plane xy coordinates (x, y) of mark point are modified, i.e. according to each tracing point to left and right sides lane line and road edge away from It is modified from the plane xy coordinates to each tracing point, is sat with obtaining the plane xy for the tracing point being distributed on road axis Mark, the coordinate for the tracing point being distributed in after amendment on road axis is (X, Y).As shown in figure 3, due to (its coordinate of tracing point 1 For (x1, y1)) and tracing point 2 (its coordinate is (x2, y2)) acquisition interval it is shorter, therefore be believed that in acquisition interval DGPS transport Dynamic rail mark is straight line, therefore the coordinate (X of tracing point on road axis can be obtained by equation below1, Y1):
Wherein, s1And l1It is distance of the tracing point 1 to both sides of the road edge respectively, s2And l2It is that tracing point 2 arrives road respectively The distance of both sides of the edge.
In the present embodiment, the data prediction subelement 511 in map vector generation unit 51 is mainly used for realizing bad Point filtering and shift point remove, wherein:
Bad point filtering refers to:If the velocity information that continuous moment vehicle DGPS data obtain is (since vehicle DGPS data are adopted Then divided by the sampling period collection is fixed frequency, therefore by calculating the distance between two tracing points, you can is obtained approximate Velocity amplitude) differ by more than threshold value, then illustrate the tracing point collection the very big probability of vehicle DGPS data be wrong, therefore based on Previous moment be valid data on the basis of, it is necessary to exclude the tracing point of later moment in time.
The principle that shift point removes can be such as Fig. 4:Pass through continuous three tracing points:1: 2: 3 is put to judge;Point 1 and point 3 The straight line being linked to be is denoted as L, if the length D of the projection of point 2 to straight line L have exceeded road and had a lot of social connections W, illustrates that a little 2 very big probabilities are to drift about Point, it is therefore desirable to remove point 2.
In the present embodiment, in map vector generation unit 51 road axis generation subelement 512 be mainly used for by Fitting generation road axis after remaining tracing point (being primarily referred to as being distributed in the tracing point on road axis) cluster, its In:
Remaining tracing point cluster is referred to:By clustering algorithm (due to the algorithm general knowledge known in this field, therefore below only Do simple introduction) carry out tracing point cluster along the storage order (i.e. the extending direction of road) of track.Specifically include with Lower step:
(1) density threshold P and cluster radius d is set, tracing point is not exactly at due to data error On traval trace, and it is distributed across on the position of twice of about road width and (regards shift point more than 2 times), d's Value can determine that the value of P is different under different pieces of information collection, should be bigger than normal under density data collection, sparse number according to this regularity of distribution According to collection under should be smaller;
(2) tracing point chain is stored in after all effectively vehicle DGPS data collected in road area being carried out Gaussian transformation In table, the attribute data of each tracing point includes its plane xy coordinates and a mark for whether having completed cluster;
(3) first track point data in chained list is chosen, if the data represent that the tracing point had been participating in cluster, Then continue to select next track point data in chained list, until finding the data not clustered;If tracing point did not participated in cluster, Then jump procedure (4);
(4) tracing point of cluster was not participated in using this as the center of circle, and distance d clusters for radius, judges the track in cluster circle Count out, if number has exceeded density threshold P, it is node (that is a, method by taking average, by model to be clustered Tracing point cluster in enclosing is a point), the plane xy coordinates of the node are respectively the equal of the plane xy coordinates of these tracing points Value, which is stored in cluster node chained list, performs step (5);
(5) it is all that these tracing points are poly- labeled as having participated in no matter track counts out whether exceed density threshold P in the circle Class, performs step (6);
(6) tracing point that is nearest and having neither part nor lot in cluster from the cluster node is found in tracing point chained list, if energy Find, then continuing step (4), (centre point in step (4) refers to a cluster node found in this step and upper most at this time Tracing point that is near and not participating in cluster), otherwise perform step (7);
(7) all cluster node set are exported, these cluster nodes are linked up and are regarded as path locus, that is, Say, a road can be indicated with limited cluster node in the form of a chained list, and adjacent cluster node exists It is continuous in chained list, such as in digital map navigation, automobile passes through per stretch, it is believed that each cluster node of approach.
The advantages of above-mentioned clustering method, is that obtained cluster node is that order is distributed, even if car during acquisition trajectories point There is repetition in the track opened or opens back and forth all without influence, and each cluster node distribution is than more uniform.
Road axis generates subelement 512 after above-mentioned tracing point cluster is completed, fitting generation road axis, wherein Including:
(1) road direction is calibrated
In general, it is inaccurate to calculate road direction by simple gps coordinate point, it is therefore desirable to which intelligent vehicle is more Real time calibration is carried out to road direction during secondary traveling, that is, judges whether current vehicle is in the state of smooth ride, That is, whether vehicle travel direction in one section of operating range tends towards stability, if so, then it will be understood that vehicle heading with it is current Road direction keeping parallelism, can make θ at this timeroadcarroadRepresent road direction, θcarRepresent vehicle heading).Specifically Simply it can judge whether the current travel direction of intelligent vehicle tends towards stability using the vehicle data in 10 sampling periods;Pass through 10 Vehicle traveling angle in a sampling period, its mean square deviation SD is calculated by equation below:
Wherein, θiRepresent the vehicle traveling angle (direction) in ith sample cycle;N indicates that N number of sampling period take part in Calculate;Represent the average of vehicle traveling angle (direction) in N number of sampling period;If SD < τ, then it is assumed that vehicle heading Tend towards stability, τ is the wealthy value of stability, is determined by actual conditions.
(2) lane line and path locus coordinate are calibrated
When running into camera or navigation equipment with the presence of side's invalid data, according to navigation GPS Status Flags and Camera reliability mark is judged.Since road has continuity in practice, therefore utilize this feature can be to the section " risk Road " carries out pre-generatmg.
If camera runs into invalid data, the road of road and last sampling period t tool residing for current vehicle are assumed There is identical extending direction, i.e. the orientation angle of present road remains unchanged within a sampling period.According to Fig. 5, by Vehicle sensors can read Vehicle Speed v, can be in the hope of:
L2=L1+vtsin (θ-α)
R2=R1-vtsin (θ-α)
Wherein, (x1, y1) is the coordinate of tracing point 1, and (x2, y2) is the coordinate of tracing point 2, and L1, R1 are respectively tracing point 1 To the distance at both sides of the road edge, L2, R2 are respectively distance of the tracing point 3 to both sides of the road edge, and α travels angle, θ for vehicle For road direction angle.
If DGPS systems are not at Differential positioning state, as shown in figure 5, can be according to current trace points and a upper track The information such as the distance of line, car speed, collection period calculate the GPS location of tracing point by equation below:
X2=x1+vtcos (α)
Y2=y1+vtsin (α)
Wherein, v is Vehicle Speed, and t is the sampling period, and (x1, y1) is the coordinate of tracing point 1, and (x2, y2) is track The coordinate of point 2, L1, R1 are respectively distance of the tracing point 1 to both sides of the road edge, and L2, R2 are respectively tracing point 2 to both sides of the road The distance at edge, α travel angle for vehicle, and θ is road direction angle.
Due to the uncertainty in risk section, therefore pass through minimum, it is necessary to taken multiple measurements to the road when generating map Square law obtains optimal path locus (i.e. the set of the tracing point of road axis coordinate composition) and road track line coordinates, leads to The road-center line coordinates obtained and the road width measured (the highway sideline distance obtained by camera obtains) are crossed, Road can be defined in more detail, and the relation of vehicle and road can be more accurately shown when being navigated.
In the present embodiment, the road Node extraction subelement 521 of topological map generation unit 52 can be used for extracting road Mouth and dead end street information, specifically:Since the cluster node order obtained in map vector generation unit 51 is distributed, and And label is all imparted, therefore road crossing tracing point can be extracted by following steps and (it is poly- to have been subjected to map vector generation unit 51 The node of class):
If the 1st, using cluster node as the center of circle, with 2 times of cluster radius for radius draw circle (this is because in cluster process, All be to be clustered with one times of cluster radius r, if therefore collection point it is more than enough, the cluster node after cluster is approximate In equally distributed, the distance of each node is in the range of r~2r), if the point in circle is more than 3, and the label put is not Continuously, then crossing is likely to (this is because the node after cluster is approaches uniformity, therefore if a crossing of not following the Way Road, within the cluster radius of 2r, points should no more than 3 for it;If more than 3, represent road and bifurcated road occur Mouthful), which is added into list L;
2nd, it can be found that road is more chaotic or road crossing near have more erroneous judgement point (this is because road bifurcated Two roads occur apart from closer situation in mouth, thus the point near real road interface port can be also included after clustering into Come), using distance as diversity factor, classify to the node in above-mentioned list L, cluster 1, cluster 2 etc. are defined as per class.
3rd, each point in cluster is judged, using each point as the center of circle, cluster radius draws circle for radius, finds out and is comprised in number in circle Most points, then it is crossing that the point, which has maximum probability,;And dead end street is associated because of other no roads, therefore it is cluster half The point that round dot is minimum is included in footpath;
When the cluster node between the crossing point and end point that have found road, each road cross point and end point belongs to This road, can thus generate the topological relation that subelement 522 defines road by road topology relation.It can finally lead to The attribute information that characteristics map generation unit 53 adds some roads to the cluster node of road is crossed, thus generates characteristics map.
In conclusion the present invention can be generated available in a short time by the DGPS and camera data fusion of low cost In the unmanned high-precision navigation map in garden.By the present invention, after only road in park information need to be collected by automobile, you can Map is quickly generated, and available for the optimum path planning to garden unmanned vehicle.Meanwhile adopted during automatic driving vehicle traveling The real time data of collection can be merged with historical data again, optimization, the precision for improving map, it is achieved thereby that sustainable low cost The generation method of high-precision two-dimensional map.
It is above-described, it is only presently preferred embodiments of the present invention, is not limited to the scope of the present invention, of the invention is upper Stating embodiment can also make a variety of changes.What i.e. every claims and description according to the present patent application were made Simply, equivalent changes and modifications, falls within the claims of patent of the present invention.The not detailed description of the present invention is Routine techniques content.

Claims (10)

1. a kind of high accuracy map generation system, it is characterised in that it includes:
The vehicle front road that one data acquisition module, its vehicle DGPS data provided according to DGPS systems and camera provide The view data of road scene, obtains the running data of each tracing point on vehicle driving trace, the traveling of each tracing point Data include:Plane xy coordinates, speed, course angle, tracing point to left and right sides lane line and road edge of the tracing point Distance;The data acquisition module is always according to each tracing point to the distance of left and right sides lane line and road edge to each The plane xy coordinates of tracing point are modified, to obtain the plane xy coordinates for the tracing point being distributed on road axis;
One data processing module being connected with the data acquisition module, its according to the running data of each tracing point and The plane xy coordinates of the tracing point being distributed on road axis, generate road axis, and according to the road-center Line generates road topology relation, to obtain the structural information of whole map;And
One map display module being connected with the data processing module, it is rendered to according to the structural information of the whole map Roadway, and Real time displaying current vehicle position.
2. high accuracy map generation system according to claim 1, it is characterised in that the data acquisition module includes:
The one DGPS information process units being connected with the DGPS systems, it parses the vehicle DGPS data, to carry Take out latitude and longitude coordinates, speed and the course angle of each tracing point, and by gauss projection by each tracing point Latitude and longitude coordinates are converted to the plane xy coordinates of each tracing point;And
The one camera information processing unit being connected with the camera and DGPS information process units, before it is to the vehicle The view data of square road scene carries out image preprocessing, lane detection and lane line tracking, and according to default image Mapping relations between coordinate and world coordinates, obtain each tracing point to left and right sides lane line and road edge away from From, and the plane according to each tracing point to the distance of left and right sides lane line and road edge to each tracing point Xy coordinates are modified, with the plane xy coordinates for the tracing point being distributed in described in acquisition on road axis.
3. high accuracy map generation system according to claim 1, it is characterised in that the data processing module includes:
The one map vector generation unit being connected with the data acquisition module, it is according to the running data of each tracing point And the plane xy coordinates of the tracing point being distributed on road axis, filter out the bad point in the tracing point and drift Point, and be fitted after remaining tracing point is clustered and generate the road axis;And
The one topological map generation unit being connected with the map vector generation unit, it is according to the road-center line drawing road Crossing and dead end street information, to define road data, generate the road topology relation.
4. high accuracy map generation system according to claim 3, it is characterised in that the data processing module further includes The one characteristics map generation unit being connected with the topological map generation unit, it is used to add map location on whole map Mark and roadway characteristic mark.
5. high accuracy map generation system according to claim 1, it is characterised in that the system also includes one to be connected to Camera information receiving module between the camera and the data acquisition module, and one be connected to the DPGS and institute State the DGPS information receiving modules between data acquisition module.
6. it is according to claim 1 high accuracy map generation system, it is characterised in that the system also includes one with it is described The camera calibration mould for being used to provide the mapping relations between described image coordinate and world coordinates of data acquisition module connection Block.
7. a kind of accurately drawing generating method, it is characterised in that the described method comprises the following steps:
The figure for the vehicle front road scene that step S1, the vehicle DGPS data provided according to DGPS systems and camera provide Picture data, obtain the running data of each tracing point on vehicle driving trace, and the running data of each tracing point includes:Should Plane xy coordinates, speed, course angle and tracing point of tracing point to left and right sides lane line and road edge distance, and The plane xy coordinates of each tracing point are repaiied according to each tracing point to the distance of left and right sides lane line and road edge Just, the plane xy coordinates for the tracing point being distributed in acquisition on road axis;
Step S2, according to the flat of the running data of each tracing point and the tracing point being distributed on road axis Face xy coordinates, generate road axis, and generate road topology relation according to the road axis, to obtain whole map Structural information;And
Step S3, roadway, and Real time displaying current vehicle position are rendered to according to the structural information of the whole map.
8. accurately drawing generating method according to claim 7, it is characterised in that the step S1 includes:
Step S11, parses the vehicle DGPS data, extracts latitude and longitude coordinates, the speed of each tracing point And course angle, and the latitude and longitude coordinates of each tracing point are converted to by gauss projection the plane of each tracing point Xy coordinates;And
Step S12, image preprocessing, lane detection and track are carried out to the view data of the vehicle front road scene Line tracks, and according to the mapping relations between default image coordinate and world coordinates, obtains each tracing point to left and right The distance of both sides lane line and road edge, and according to each tracing point to left and right sides lane line and road edge away from It is modified from the plane xy coordinates to each tracing point, with the tracing point that is distributed in described in acquisition on road axis Plane xy coordinates.
9. accurately drawing generating method according to claim 7, it is characterised in that the step S2 includes:
Step S21, according to the running data of each tracing point and the tracing point being distributed on road axis Plane xy coordinates, filter out bad point and shift point in the tracing point, and are fitted after remaining tracing point is clustered described in generation Road axis;And
Step S22, according to the road-center line drawing road mouth and dead end street information, to define road data, generates institute State road topology relation.
10. accurately drawing generating method according to claim 9, it is characterised in that the step S2 is further included:
Step S23, adds map location mark and roadway characteristic mark on whole map.
CN201711187081.0A 2017-11-24 2017-11-24 A kind of high accuracy map generation system and generation method CN108036794A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201711187081.0A CN108036794A (en) 2017-11-24 2017-11-24 A kind of high accuracy map generation system and generation method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201711187081.0A CN108036794A (en) 2017-11-24 2017-11-24 A kind of high accuracy map generation system and generation method

Publications (1)

Publication Number Publication Date
CN108036794A true CN108036794A (en) 2018-05-15

Family

ID=62094246

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201711187081.0A CN108036794A (en) 2017-11-24 2017-11-24 A kind of high accuracy map generation system and generation method

Country Status (1)

Country Link
CN (1) CN108036794A (en)

Cited By (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109059954A (en) * 2018-06-29 2018-12-21 广东星舆科技有限公司 The method and system for supporting high-precision map lane line real time fusion to update
CN109189872A (en) * 2018-08-13 2019-01-11 武汉中海庭数据技术有限公司 Accurately diagram data verifies device and method
CN109540157A (en) * 2018-11-12 2019-03-29 广东星舆科技有限公司 A kind of onboard navigation system and control method
CN109579856A (en) * 2018-10-31 2019-04-05 百度在线网络技术(北京)有限公司 Accurately drawing generating method, device, equipment and computer readable storage medium
CN109597862A (en) * 2018-10-31 2019-04-09 百度在线网络技术(北京)有限公司 Ground drawing generating method, device and computer readable storage medium based on puzzle type
CN109631873A (en) * 2018-11-01 2019-04-16 百度在线网络技术(北京)有限公司 Road generation method, device and the readable storage medium storing program for executing of high-precision map
CN109855637A (en) * 2018-12-24 2019-06-07 北京新能源汽车股份有限公司 A kind of automatic Pilot paths planning method, device and the equipment of vehicle
CN109859613A (en) * 2019-01-18 2019-06-07 驭势(上海)汽车科技有限公司 A kind of high-precision cartography method and mobile unit based on projected coordinate system
CN109974739A (en) * 2019-04-15 2019-07-05 西安交通大学 Global navigation system and guidance information generation method based on high-precision map
CN110047373A (en) * 2019-04-28 2019-07-23 北京超维度计算科技有限公司 A kind of high-precision map generation system based on common application processor
CN110047372A (en) * 2019-04-28 2019-07-23 北京超维度计算科技有限公司 A kind of high-precision map generation system based on Reconfigurable Computation
CN110082123A (en) * 2019-05-05 2019-08-02 中国汽车工程研究院股份有限公司 A kind of automatic Pilot automatization test system
CN110415330A (en) * 2019-04-29 2019-11-05 当家移动绿色互联网技术集团有限公司 Road generation method, device, storage medium and electronic equipment
WO2020029601A1 (en) * 2018-08-06 2020-02-13 武汉中海庭数据技术有限公司 Method and system for constructing transverse topological relationship of lanes in map, and memory
WO2020103532A1 (en) * 2018-11-20 2020-05-28 中车株洲电力机车有限公司 Multi-axis electric bus self-guiding method

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101290725A (en) * 2007-04-17 2008-10-22 株式会社日立制作所 Digital road map generation system and map generation method
KR20100036753A (en) * 2008-09-30 2010-04-08 신동주 System and method of construction and completion inspection using dgps in engineering works and construction
CN204881653U (en) * 2015-08-26 2015-12-16 莆田市云驰新能源汽车研究院有限公司 Outdoor scene video navigation of hi -Fix
CN106525057A (en) * 2016-10-26 2017-03-22 陈曦 Generation system for high-precision road map

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101290725A (en) * 2007-04-17 2008-10-22 株式会社日立制作所 Digital road map generation system and map generation method
KR20100036753A (en) * 2008-09-30 2010-04-08 신동주 System and method of construction and completion inspection using dgps in engineering works and construction
CN204881653U (en) * 2015-08-26 2015-12-16 莆田市云驰新能源汽车研究院有限公司 Outdoor scene video navigation of hi -Fix
CN106525057A (en) * 2016-10-26 2017-03-22 陈曦 Generation system for high-precision road map

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109059954A (en) * 2018-06-29 2018-12-21 广东星舆科技有限公司 The method and system for supporting high-precision map lane line real time fusion to update
WO2020029601A1 (en) * 2018-08-06 2020-02-13 武汉中海庭数据技术有限公司 Method and system for constructing transverse topological relationship of lanes in map, and memory
CN109189872A (en) * 2018-08-13 2019-01-11 武汉中海庭数据技术有限公司 Accurately diagram data verifies device and method
CN109579856A (en) * 2018-10-31 2019-04-05 百度在线网络技术(北京)有限公司 Accurately drawing generating method, device, equipment and computer readable storage medium
CN109597862A (en) * 2018-10-31 2019-04-09 百度在线网络技术(北京)有限公司 Ground drawing generating method, device and computer readable storage medium based on puzzle type
CN109631873A (en) * 2018-11-01 2019-04-16 百度在线网络技术(北京)有限公司 Road generation method, device and the readable storage medium storing program for executing of high-precision map
CN109540157A (en) * 2018-11-12 2019-03-29 广东星舆科技有限公司 A kind of onboard navigation system and control method
WO2020103532A1 (en) * 2018-11-20 2020-05-28 中车株洲电力机车有限公司 Multi-axis electric bus self-guiding method
CN109855637A (en) * 2018-12-24 2019-06-07 北京新能源汽车股份有限公司 A kind of automatic Pilot paths planning method, device and the equipment of vehicle
CN109859613A (en) * 2019-01-18 2019-06-07 驭势(上海)汽车科技有限公司 A kind of high-precision cartography method and mobile unit based on projected coordinate system
CN109974739A (en) * 2019-04-15 2019-07-05 西安交通大学 Global navigation system and guidance information generation method based on high-precision map
CN110047372A (en) * 2019-04-28 2019-07-23 北京超维度计算科技有限公司 A kind of high-precision map generation system based on Reconfigurable Computation
CN110047373A (en) * 2019-04-28 2019-07-23 北京超维度计算科技有限公司 A kind of high-precision map generation system based on common application processor
CN110415330A (en) * 2019-04-29 2019-11-05 当家移动绿色互联网技术集团有限公司 Road generation method, device, storage medium and electronic equipment
CN110415330B (en) * 2019-04-29 2020-05-29 当家移动绿色互联网技术集团有限公司 Road generation method, device, storage medium and electronic equipment
CN110082123A (en) * 2019-05-05 2019-08-02 中国汽车工程研究院股份有限公司 A kind of automatic Pilot automatization test system

Similar Documents

Publication Publication Date Title
US10197404B2 (en) Path curve confidence factors
Suhr et al. Sensor fusion-based low-cost vehicle localization system for complex urban environments
US10545029B2 (en) Lane network construction using high definition maps for autonomous vehicles
Hashemi et al. A critical review of real-time map-matching algorithms: Current issues and future directions
CN104854639B (en) Vehicle position estimation device and vehicle location method of estimation
JP2019527832A (en) System and method for accurate localization and mapping
Luettel et al. Autonomous ground vehicles—Concepts and a path to the future
US9946939B2 (en) Guided geometry extraction for localization of a device
Quddus et al. A high accuracy fuzzy logic based map matching algorithm for road transport
Levinson et al. Robust vehicle localization in urban environments using probabilistic maps
KR20200011566A (en) Sparse map for autonomous vehicle navigation
CN104573733B (en) A kind of fine map generation system and method based on high definition orthophotoquad
CN101975951B (en) Field environment barrier detection method fusing distance and image information
KR101534056B1 (en) Traffic signal mapping and detection
US20180025235A1 (en) Crowdsourcing the collection of road surface information
US20180285659A1 (en) Method, apparatus, and system for a parametric representation of lane lines
US7928905B2 (en) Method of using road signs to augment global positioning system (GPS) coordinate data for calculating a current position of a personal navigation device
US9377313B2 (en) Methods and systems for creating digital street network database
CN104508719B (en) Drive assist system and driving assistance method
CN102138163B (en) Bird's-eye image forming device, bird's-eye image forming method
US8791996B2 (en) Image processing system and position measurement system
JP5057184B2 (en) Image processing system and vehicle control system
US10074270B2 (en) Clustering observations of objects along roads for navigation-related operations
JP5505723B2 (en) Image processing system and positioning system
CN106199558A (en) Barrier method for quick

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