CN1356668A - Rolling ball method for automatically generating digital road map - Google Patents
Rolling ball method for automatically generating digital road map Download PDFInfo
- Publication number
- CN1356668A CN1356668A CN 01145718 CN01145718A CN1356668A CN 1356668 A CN1356668 A CN 1356668A CN 01145718 CN01145718 CN 01145718 CN 01145718 A CN01145718 A CN 01145718A CN 1356668 A CN1356668 A CN 1356668A
- Authority
- CN
- China
- Prior art keywords
- road
- map
- intersection
- big ball
- digital
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Processing Or Creating Images (AREA)
- Image Processing (AREA)
- Instructional Devices (AREA)
Abstract
A ball rolling method for automatically generating digital road map is disclosed. The original binary image is pretreated to be transformed into map graphics. After primary curve is detected, it is input to digital road map generator to generate digital road map data by ball rolling method. In the ball rolling method, an assumed big ball with regulatable radius and central position is used to determine central line of road, node coordinates and map vectors. The road cross detection can provide precise cross position. The road network searching controller can use the road vectors in whole road network. Its advantages include high precision, speed and full automatization.
Description
Technical field:
The present invention relates to a kind of rolling ball method for automatically generating digital road map, be used for the digital road map exploitation, and the exploitation of GIS Geographic Information System, computerized information image processing technique field belonged to.
Background technology:
Along with developing rapidly of GIS, intelligent transportation system, Global Positioning System (GPS), computing machine and information science, numerical map obtains increasingly extensive application in a lot of fields, particularly in automobile navigation, vehicle location and the reallocation of land, numerical map plays key effect.In present numerical map production run, the interactive computer interfacial is extensively adopted in the input of map datum.Because this method is a kind of manual data input method, need large quantities of computer operators and considerable time to go to finish the data input service of map, cause numerical map cost of manufacture height, the production cycle is long.Adopt computer vision and mode identification technology, the automatic generation of numerical map can realize on computers.The automatic generation of numerical map can reduce the production cost of numerical map effectively, shortens the update cycle of map datum.Numerical map structurally comprises many levels, is the road data structural sheet in order to the core data level that navigates and locate.To the input of the data of this level, all in the world numerical map companies at present comprise Etak, the NaviTach of the U.S., and professional numerical map such as the silkworm woods of the Tele Atlas in Europe and Japan company all mainly adopts manual method to carry out the map datum input.Manual method is also all adopted in the exploitation of domestic digital road map.
The never interruption of research exploration of how understanding the geography information of map image and how therefrom extracting map datum.Aspect road-map identification, it is mainly studied and comprises D.Geman and B.Jedynak, (" AnActive Testing Model for Tracking Roads in Satellite Images ", IEEE Trans.PatternAnalysis and Machine Intelligence, vol.18, No.1, pp.1-14,1996.) and M.Barzoharand D.B.Cooper, (" Automatic Finding ofMain Roads in Aerial Images By UsingGeometric-Stochastic Models and Estimation ", IEEE Trans.Pattern Analysis andMachine Intelligence, vol.18, No.7, pp.707-721,1996.).These researchs have obtained success to a certain extent and have had certain theory and have been worth, but its achievement still can't be applied to the actual production process of numerical map.Their limitation is: 1) all the geographic model of realistic space has been made specific hypothesis, limited the range of application of its method.2) method of Geman and Baerohar all only detects the main stem road system of map, as highway.In fact, the meticulous road branch in the urban road network is only the geographical message part of high density of numerical map.
In order to adapt to the actual needs of map industry, it is the numerical map automatic creation system of raw data that some researchists begin to research and develop employing paper map.But because the complicacy of the scrambling of digital road map and automatic generation technique, the numerical map generation technique only limits to the house map automatically at present, and practical digital road map automatic creation system is not arranged in the world at present as yet.
Summary of the invention:
The objective of the invention is to above-mentioned deficiency and actual needs at prior art, provide a kind of new digital road map to generate method---rolling ball method automatically, for the performance history of numerical map provides high efficiency input tool, replace traditional manual data input method (as quantizing plate method or computer interface method) with data precision height, the automatic input method of the fireballing map datum of running, reduce the cost of development of numerical map, shorten the fabrication cycle of numerical map.
For realizing such purpose, digital road map proposed by the invention generates method automatically, it is rolling ball method, comprise three essential parts: spin, intersection detect, the road network search, adopt Flame Image Process and mode identification technology, the feature identification and the vector quantization of road-map are carried out in same process, whole road network is carried out the digitizing conversion, realize a full-automatic process from a seed points.
The avatars data that the desired input data of rolling ball method are original road-maps.The avatars data of road-map can be one of following two kinds:
1) (engineer's scale was less than 1: 1000 for the figure of paper map skeleton image; The skeleton image structural integrity).
2) (engineer's scale was less than 1: 1000 for the figure of gray scale map edge image; The edge image structural integrity).
The original binary image is removed noise through pre-service and is carried out refinement again, converts map graph then to.A pixel in the corresponding original image of the node of each figure.Subsequently, map graph carries out principal curve and detects, and eliminates the literal in the original map, the composition of symbol correspondence in map graph; And to the fracture road edge curve carry out match.Then, the map graph after the processing is admitted to the digital road map generator, generates the digital road map data by rolling ball method.
Rolling ball method comprises following step:
1, spin:
The main process that digital road map generates automatically is " spin ", i.e. identification and output vector numerical map.There is the adjustable big ball of a radius and center in imagination on road, when big ball rolled forward on road, the radius of the sphere center position of big ball and big ball was with the width and the right and wrong adjustment of road, so that the board forever surperficial and road of big ball is tangent.When big ball when an end of road is rolled into the other end, the track of the big ball centre of sphere is a center line of road; The position of the data point in the track then provides the node coordinate in the digital road map data; Connection between back end then is the digital road map vector.
2, intersection detects:
When big ball was rolled to intersection, the intersection detecting device was activated.Intersection detects two purposes: at first, because each branch of road crosses at intersection, intersection detects and makes the rolling of big ball expand to other each branches that are connected with the crossing from a branch, has kept rolling ball method operating continuity in road network.Secondly, intersection detects accurate intersection geographical position coordinates can be provided.The intersection coordinate is one of most important data of digital geographical map.
It is at the intersection place that intersection detects principle, and with the radius expansion of big ball and the shape of appropriate change spheroid, so that each distributor road of big ball and crossing intersects.The mid point of the cross surface of distributor road and big ball is decided to be the initial seed points of spin of this distributor road.During this distributor road digitizing, big ball will be pushed ahead from this seed points.
The center of intersection is defined as the intersection point of each distributor road center line.The center position coordinates of intersection is the least square solution of each distributor road center line equation.
3. road network Search Control:
Road network search controller coordination spin and intersection testing process interweave and carry out, and guarantee the arbitrary road inlet from road network, and rolling ball method road-map vector quantization process is through to whole map road network.The breadth-first search method (BFS) of standard has been used in the road network search; The distributor road seed points that intersection detection is soon obtained leaves in the queue, and the output of queue is next road vectors starting point; Big ball rolls forward from the road at this starting point place, and road is carried out vector quantization.
In the digital road map generative process of the present invention, the scanning of having only the original map image is manual process, exports to the digital road map data from the pre-service of map image, fully automatically running.Compare with traditional manual input method, the digital road map positional precision height that method of the present invention is developed, operating speed is fast, be about manual 10 times, shorten the map construction cycle and reduced cost of development greatly, can realize 24 hours fully automatic workings, and can be benefited along with development of computer.
Description of drawings and embodiment:
For understanding technical scheme of the present invention better, describe in further detail below in conjunction with drawings and Examples.
Fig. 1 is the FB(flow block) of digital road map automatic creation system.
As shown in the figure, original map data becomes digital image by scan conversion, remove noise through pre-service and carry out image thinning again, carry out the detection of figure principal curve after converting map graph then to, map graph after the processing is admitted to the digital road map generator, adopts core technology rolling ball method of the present invention to generate the digital road map data.
Fig. 2 is a rolling ball method schematic flow sheet of the present invention.
As shown in the figure, the spin process enters road and generates the digital road map data from initial seed points.When big ball was rolled to intersection, the intersection detecting device was activated, and detected the road branch that is connected with this crossing, and in each road branch the spin seed points was set.The seed points data are sent to the road network search controller, under the domination of controller, from seed points separately, respectively each road branch are carried out vector quantization with rolling ball method.Because each the road branch in the road network interconnects, thereby whole road network will be able to by vector quantization.
Fig. 3 is the prediction and adjustment synoptic diagram of spin parameter in the rolling ball method of the present invention.
As shown in Figure 3, the prediction of the spin parameter among the present invention and adjustment comprise the prediction and the adjustment of big ball center position and radius.Adjusted big ball center track is a center line of road.Rolling ball method allows the change of road turning and road width.Among the figure, O
i' and r
i' be respectively sphere center position and the radius of a ball length of predicting constantly at i; O
I-1' be i-1 centre of sphere predicted position constantly; O
I-1, r
I-1, O
I-2, r
I-2Be respectively that big ball is in i-1, i-2 reality (adjusted) sphere center position and radius of a ball length constantly.
Fig. 4 detects synoptic diagram for intersection in the rolling ball method of the present invention.
As shown in Figure 4, when big ball was rolled to intersection, the intersection detecting device was activated, and big ball is expanded.The position is at O
i' big ball from radius r
iExpand into R so that with each B of road branch of crossing
1, B
2, B
3Intersect.Sphere is at the i.e. initial seed points (Z of vector quantization of this road branch of the arcuate midway point of road branch inside
1, Z
3And Z
3).The vector quantization process of each road branch is from each self-corresponding seed points.
Fig. 5 finds the solution the center synoptic diagram of intersection with least square method for the present invention.
As shown in the figure, in rolling ball method, the center line intersection point of each road branch of crossing is defined as the center of this intersection.Because the center line of each road branch might not cross and a bit in practice, when determining the center of crossing, road, find the solution with least square method.Among the figure, B
0, B
1, B
2, B
3It is road branch sequence number.
Fig. 6 is the figure of original input map datum and result in the embodiment of the invention.
Wherein, Fig. 6 a is the digital image of original map after scanning; Fig. 6 b is the map graph through pre-service and refinement; Fig. 6 c is the road figure through principal curve detects and match obtains; Fig. 6 d has provided the spin seed points of spin process and road branch; The automatic generation digital road map of Fig. 6 e for obtaining at last.
Present embodiment is an original map data with the paper map.Instance system realizes that on the Sun170E workstation rolling ball method of the present invention is the core process of this system.The embodiment processing procedure is described as follows.
Fig. 6 a is that 1: 1000 original map converts binary picture to by the 300dpi scanner scanning.After process pre-service removal picture noise passes through refinement again, convert map graph to, see Fig. 6 b.The map graph of this moment and the map skeleton image after the refinement structurally there is no difference in essence; A node in the corresponding map graph of a pixel in the map skeleton image.Subsequently, map graph carries out principal curve and detects, and eliminates the literal in the original map, the composition of symbol correspondence in map graph; And the road edge curve of fracture carried out match, and what exported is the road edge curve of " totally ", output result at this moment is shown in Fig. 6 c.Then, the map graph after the processing is admitted to the digital road map generator, generates the digital road map data by rolling ball method.
Map among Fig. 6 d comprises two road networks.First network is the single road in the map upper right corner, and all the other roads belong to second road network.We discuss second road network numerical map now and generate automatically.The startup operation point of rolling ball method is from No. zero seed in the map upper left corner.Initial seed points can generate automatically, can also produce by hand.The rolling ball method operating procedure is as follows:
1) initial big ball forms: No. zero seed points expands and forms big ball, and is tangent until sphere and edge, the road left and right sides.
2) determine initial spin direction: big ball has two rotating directions along road: direction is rolled to left down, walks out map boundary line, abandons resulting data; To rolling successfully, this direction is the spin direction that initial road data generates to the upper right side.
3) initial road data generates: according to method of the present invention, carry out sphere center position and radius of a ball prediction and sphere center position and radius of a ball adjustment, make sphere and road boundary tangent.At this moment, sphere center position is exactly the node data of digital road map.Big ball generates direction along initial road data and rolls, and the centre of sphere, radius of a ball prediction adjustment process realize the road vectorsization of this road branch repeatedly.This road branch symbiosis becomes 8 node datas.
4) intersection detects: when big ball is rolled to intersection
During the place, the intersection detecting device is activated, and detects two road branches
With
These two seed datas are deposited in the seed formation reservoir.
5) the distributor road data generate: take out top seed points from the formation reservoir
Adopt rolling ball method to generate the vector data of this road branch.Because the road terminal point exceeds the border of map, there is not new road branch seed points to produce.
6) the distributor road data generate: take out top seed points from the formation reservoir
, adopt rolling ball method to generate the vector data of this road branch.When big ball is rolled to intersection
During the place, the intersection detecting device is activated, and detects two road branches
With
These two seed datas are deposited in the seed formation reservoir.
7) from the formation reservoir, take out top seed points, adopt rolling ball method to generate the vector data of road branch ..., so repeatedly, road branches all in the map road network are become numerical map by complete vector quantization.
Fig. 6 e is the digital road map that generates automatically.In numerical map, the distance between back end is half of road width, and the node location error is a pixel, and promptly the precision of scanner is 1/300 inch.
In the digital road map generative process, the scanning of having only the original map image is manual process, exports to the digital road map data from the pre-service of map image, fully automatically running.Needed first road data point in the rolling ball method, promptly initial seed points can be imported by hand, also can have system to produce automatically.In the running speed of system, most of running time is occupied by image thinning, and rolling ball method only operates the time with sub-fraction.In native system embodiment, the image thinning process adopts software to realize; If adopt hardware to carry out image thinning, the running time of total system can significantly reduce.
Claims (1)
1, a kind of rolling ball method for automatically generating digital road map, original map data becomes digital image by scan conversion, remove noise through pre-service, carry out image thinning, carry out the detection of figure principal curve after converting map graph then to, map graph after it is characterized in that handling adopts rolling ball method to generate the digital road map data again, divides following step to carry out:
A, " spin ": establish adjustable big ball of radius and center of existence on the road, the radius of the sphere center position of big ball and big ball is with the width and the right and wrong adjustment of road, make the board of the surface of big ball and road tangent, the track of the big ball centre of sphere is a center line of road, the position of the data point in the track then provides the node coordinate in the digital road map data, and the connection between back end then is the digital road map vector;
B, intersection detect: at the intersection place, the radius of big ball is enlarged and changes the shape of spheroid, each distributor road of big ball and crossing is intersected, the mid point of the cross surface of distributor road and big ball is decided to be the initial seed points of spin of this distributor road, the center of intersection is defined as the intersection point of each distributor road center line, and its coordinate is the minimum of each distributor road center line equation
Square law is separated;
C, road network Search Control: use the breadth-first search method, intersection is detected the distributor road seed points that is obtained to be left in the queue, the output of queue is next road vectors starting point, big ball rolls forward from the road at this starting point place, road is carried out vector quantization, and road network search controller coordination spin and intersection testing process interweave and carry out.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB01145718XA CN1141669C (en) | 2001-12-30 | 2001-12-30 | Rolling ball method for automatically generating digital road map |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB01145718XA CN1141669C (en) | 2001-12-30 | 2001-12-30 | Rolling ball method for automatically generating digital road map |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1356668A true CN1356668A (en) | 2002-07-03 |
| CN1141669C CN1141669C (en) | 2004-03-10 |
Family
ID=4678292
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB01145718XA Expired - Fee Related CN1141669C (en) | 2001-12-30 | 2001-12-30 | Rolling ball method for automatically generating digital road map |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1141669C (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1299220C (en) * | 2004-05-13 | 2007-02-07 | 上海交通大学 | Automatic splicing method for digital road map |
| CN100450277C (en) * | 2005-07-22 | 2009-01-07 | 中兴通讯股份有限公司 | Method for exchanging position information in mobile positioning service system |
| CN101630463B (en) * | 2009-07-30 | 2010-12-01 | 上海交通大学 | Method for automatic vectorization of road network digital raster map |
| CN101290725B (en) * | 2007-04-17 | 2012-11-28 | 株式会社日立制作所 | Digital road map generation system and map generation method |
| CN101014831B (en) * | 2004-03-23 | 2013-08-21 | 咕果公司 | A digital mapping system |
| CN109522903A (en) * | 2018-08-31 | 2019-03-26 | 董箭 | TIN_DDM Feature line extracting method based on the transformation of bidirectional rolling ball |
-
2001
- 2001-12-30 CN CNB01145718XA patent/CN1141669C/en not_active Expired - Fee Related
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101014831B (en) * | 2004-03-23 | 2013-08-21 | 咕果公司 | A digital mapping system |
| CN1299220C (en) * | 2004-05-13 | 2007-02-07 | 上海交通大学 | Automatic splicing method for digital road map |
| CN100450277C (en) * | 2005-07-22 | 2009-01-07 | 中兴通讯股份有限公司 | Method for exchanging position information in mobile positioning service system |
| CN101290725B (en) * | 2007-04-17 | 2012-11-28 | 株式会社日立制作所 | Digital road map generation system and map generation method |
| CN101630463B (en) * | 2009-07-30 | 2010-12-01 | 上海交通大学 | Method for automatic vectorization of road network digital raster map |
| CN109522903A (en) * | 2018-08-31 | 2019-03-26 | 董箭 | TIN_DDM Feature line extracting method based on the transformation of bidirectional rolling ball |
| CN109522903B (en) * | 2018-08-31 | 2022-10-25 | 中国人民解放军海军大连舰艇学院 | TIN _ DDM topographic feature line extraction method based on bidirectional rolling ball transformation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1141669C (en) | 2004-03-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109977945A (en) | Localization method and system based on deep learning | |
| CN111046768B (en) | Deep learning method for simultaneously extracting road pavement and center line of remote sensing image | |
| CN102289991B (en) | Visual-variable-based automatic classification and configuration method of map lettering | |
| Qin et al. | Automated reconstruction of parametric bim for bridge based on terrestrial laser scanning data | |
| CN114236552B (en) | Repositioning method and repositioning system based on laser radar | |
| CN112712138B (en) | Image processing method, device, equipment and storage medium | |
| CN109993064B (en) | Method for extracting connection path between road network nodes in picture | |
| CN114219819A (en) | Oblique photography model unitization method based on orthoscopic image boundary detection | |
| CN113505842A (en) | Automatic extraction method suitable for large-scale regional remote sensing image urban building | |
| CN114821492A (en) | YOLOv 4-based road vehicle detection system and method | |
| CN111090717A (en) | Automatic building model based on CGA rule | |
| CN1152341C (en) | Automatic generation method for digital house map | |
| CN115393745A (en) | Automatic bridge image progress identification method based on unmanned aerial vehicle and deep learning | |
| CN1141669C (en) | Rolling ball method for automatically generating digital road map | |
| CN114820668A (en) | End-to-end building regular outline automatic extraction method based on concentric ring convolution | |
| CN113704276A (en) | Map updating method and device, electronic equipment and computer readable storage medium | |
| CN1152343C (en) | Method for fast preparing vector map of house through man-machine interactive | |
| CN115082778B (en) | Multi-branch learning-based homestead identification method and system | |
| CN115661097A (en) | Object surface defect detection method and system | |
| CN1299220C (en) | Automatic splicing method for digital road map | |
| CN114547761A (en) | High-precision model-based prefabricated part steel bar optimization method and system | |
| CN114124104A (en) | Road positioning data compression storage method and terminal equipment | |
| CN113642313A (en) | Address text processing method, device, equipment, storage medium and program product | |
| Kolibarov et al. | Roof Segmentation Towards Digital Twin Generation in LoD2+ Using Deep Learning | |
| Ning et al. | SAMLoc: Structure-Aware Constraints With Multi-Task Distillation for Long-Term Visual Localization |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C06 | Publication | ||
| PB01 | Publication | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C19 | Lapse of patent right due to non-payment of the annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |