CN106323301A

CN106323301A - Road information obtaining method and device

Info

Publication number: CN106323301A
Application number: CN201510369833.XA
Authority: CN
Inventors: 王淼; 石清华; 潘惠丹
Original assignee: Navinfo Co Ltd
Current assignee: Navinfo Co Ltd
Priority date: 2015-06-29
Filing date: 2015-06-29
Publication date: 2017-01-11
Anticipated expiration: 2035-06-29
Also published as: CN106323301B

Abstract

The invention provides a road information obtaining method and device. The method comprises the following steps: obtaining the track point data and road data in a preset area in real time; obtaining the information type of road information needed to be extracted, according to the information type, subjecting the track point data and road data to differential operation so as to obtain the differential track points that are matched with the information type and are not matched with the road data; subjecting the differential tract points to a noise reducing treatment and clustering calculation to obtain a clustering point set comprising multiple differential track points; subjecting the clustering point set to road shape characteristic tests to obtain a clustering point set that can label a road; and according to the clustering point set that can label a road, obtaining the road information of the preset area. Through realtime differentiation a large amount of track point data and current data, clustering algorithm, and noise reducing algorithm, the information of increased roads or big road changes can be extracted.

Description

The acquisition methods of a kind of road informations and device

Technical field

The present invention relates to field of navigation technology, particularly relate to acquisition methods and the device of a kind of road informations.

Background technology

The most a large amount of cell phone map products and its spin-off such as navigation software, call a taxi software etc., substantial amounts of location information can be accumulated during operation, some map producing business or service provider utilize this kind of data to carry out data mining work, one of them free-revving engine be by with existing map data base differential comparison, find " new added road " or " road generation great change " information, and respond rapidly to this type of information, launch the collection in worksite for this type of road data in time, editor's warehouse-in etc. updates work, and push in time in the client of all kinds of graphical user.

Wherein, the attribute information that every road informations should possess includes: type, grade, exact position and scope, recommended value and reliability.In order to solve the problems referred to above, it is desirable to data processing technique scheme need to meet following condition:

1. can adapt to mass data, the most each provinces and regions terabyte every day (TB) DBMS amount；

2. can adapt to the location sensors of multiple quality, possess rejecting wrong data, adapt to the ability of low quality data；

3. can quickly process data, can obtain data processed result within much smaller than the time of data accumulation, the data collected for such as a week should so it is possible to update " new added road " data of one week in one week less than in one day；

4. the result of data processing software system output, i.e. information set, need to meet available requirement, i.e. next link of map producing obtains the amount of information and quality is all controlled, and after importance and reliability screening, the number obtained is very limited, could carry out follow-up work within the limited time.

In existing technical scheme, the most typical two class solutions are as follows:

Scheme one: trajectory is mated with Road

The first step, sets up trajectory (from a series of orderly coordinate points with timestamp of same alignment sensor) according to user tracing point set in a special time period；

Second step, carries out " line-line " by trajectory with existing road data and mates；

3rd step, the trajectory do not mated, it is potential " new added road line "；

4th step, owing to, on same section, may produce a lot of bar trajectory, therefore the trajectory also needing to obtain the 3rd step carries out " polymerization ", the trajectory that will be likely located at same section merges into a trajectory, this result merged, it is possible to as " new added road line ".

But, scheme one is the scheme that a kind of theoretical model is the most rigorous, global positioning system (Global Positioning System for a small amount of high-quality, GPS) positioning result data have well adapting to property, as high-end bus location track data possesses high accuracy, such as in plane, error less than 10 meters and has inertial navigation module and calibrates)；High sample frequency, such as 1 hertz；Behavioral pattern is fixed, and the such as overwhelming majority time travels at a relatively high speed and mating formation on highway or street.But for magnanimity and alignment sensor originate not unique internet, applications location data and inapplicable, its reason is as follows:

On the one hand, scheme one cannot support mass data.Such as, typical internet, applications location data volume, only such a city in Beijing, i.e. produce 1,000,000,000 location data every day, " setting up trajectory " needs travel through all data and rearrange storage, time-consuming and to consume disk huge, only it is possible to consume one day Time Calculation in this step；On the other hand, scheme one cannot support low sample frequency data, even if such as solving the efficiency of " setting up trajectory ", quality for " trajectory " cannot guarantee that, and owing to mobile Internet application location data sampling frequency is low, the sampling interval major part of same user is more than 10 seconds, and therefore its trajectory greatly differs from each other with real road shape, therefore, the subsequent step of scheme one is the most meaningless.

Scheme two: Grid Method

Grid Method vector locus data will be converted to raster data, i.e. digital picture, and each pixel represents the small grid on map, takes advantage of 10 meters for such as 10 meters.Each pixel is scanned, if this pixel is included in has learned that route, just the tracing point in this pixel is deleted.The collection of pixels remained is the alternative of " new added road ", then this image is carried out " binaryzation ", the image recognition processing such as " line detection ", extracts " line feature " i.e. " new added road line ".

It can thus be appreciated that, the advantage of this this scheme is: on the one hand, efficiently and mass data can be supported: the number of pixels owing to scanning is constant, and judge that the computing whether pixel is included on known road is the simplest, therefore efficiency height optimization, and its process performance can be improved further by the strategy of data piecemeal in advance is carried out parallel distributed calculating；On the other hand, scheme two can carry out line drawing by ripe image recognition algorithm, and owing to being not required to foundation " trajectory ", is therefore the most not strict with the sampling interval of data.

But, the shortcoming of scheme two is the most fatal: on the one hand, theoretical model is simple and crude, do not support to there is clover leaf downtown area, owing to judging whether certain tracing point belongs to known road, if therefore " new added road " intersects with " known road " is intensive in a large number only by plan-position is the most overlapping, the tracing point belonging to this " new added road " is the most deleted, and in center, large size city, the intensive intersection of road stereoscopic, it is virtually impossible to extract any useful " new added road " line.And for navigation map, most important urban area exactly；On the other hand, it is applicable only to " new added road detection " scene, deficiency is supported in complex properties change detection.

Summary of the invention

In order to overcome the above-mentioned problems in the prior art, aspects of which provide acquisition methods and the device of a kind of road informations, by the difference to magnanimity tracing point data Yu road data, and by cluster and noise reduction algorithm, find the road data disappearance in existing navigation map data storehouse or defect.

In order to solve above-mentioned technical problem, the present invention adopts the following technical scheme that

According to one aspect of the present invention, it is provided that the acquisition methods of a kind of road informations, the method includes:

Obtain the tracing point data in predeterminable area and road data in real time；

Obtain the information type of the required road informations extracted, and according to described information type, described tracing point data and road data are carried out calculus of differences, it is thus achieved that mate with described information type and difference tracing point unmatched with described road data；

Described difference tracing point is carried out noise reduction process and cluster calculation, it is thus achieved that include the cluster point group of multiple difference tracing point；

Described cluster point group is carried out road shape characteristic test, it is thus achieved that the cluster point group of a road can be identified；

Cluster point group according to identifying a road obtains the road informations in described predeterminable area.

Wherein, described according to described information type, described tracing point data and road data are carried out calculus of differences, it is thus achieved that mate with described information type and difference tracing point unmatched with described road data, specifically include:

From described tracing point data, obtain error in the plane of tracing point, and delete in plane error more than the tracing point data of the first predetermined threshold value, and tracing point data unmatched with described information type, it is thus achieved that the tracing point data after filtration；

Tracing point data after described filtration are carried out piecemeal according to geographical network, it is thus achieved that the one-level piecemeal tracing point group in multiple one-level segmented areas and each one-level segmented areas；

Tracing point data in one-level piecemeal tracing point group each described and described road data are carried out calculus of differences, it is thus achieved that the difference tracing point under one-level piecemeal tracing point group.

Wherein, when described information type is new added road, described tracing point data in one-level piecemeal tracing point group each described and described road data are carried out calculus of differences, it is thus achieved that the difference tracing point under one-level piecemeal tracing point group, specifically include:

The road being positioned at each described one-level segmented areas is obtained according to described road data；

Judge tracing point in each one-level piecemeal tracing point group minimum vertical line distance to each road belonged to described tracing point in same one-level segmented areas whether more than or equal to the second predetermined threshold value, if the determination result is YES, the most described tracing point is difference tracing point；Or

Judge the wheeled direction to the minimum vertical line distance of each road belonged in same one-level segmented areas of the tracing point in each one-level piecemeal tracing point group with described tracing point, whether it is more than or equal to the 3rd predetermined threshold value with the absolute value of the difference in the current course of described tracing point, if the determination result is YES, the most described tracing point is difference tracing point.

Wherein, described described difference tracing point is carried out noise reduction process and cluster calculation, it is thus achieved that include a step for the cluster point group of multiple difference tracing point, specifically include:

To the difference tracing point under each one-level piecemeal, carry out piecemeal according to geographic grid, it is thus achieved that two grades of piecemeal difference tracing point groups；

Judge that whether the distance between any two difference tracing point under each two grades of difference block tracing point groups is less than the 4th predetermined threshold value；

If the determination result is YES, then judge that whether the absolute value of difference of the course value of said two difference tracing point is less than the 5th predetermined threshold value, if the determination result is YES, then the immediate neighbor point of said two difference tracing point the other side each other；

Travel through described difference tracing point successively, it is judged that whether the immediate neighbor point number of each described difference tracing point is less than the 6th predetermined threshold value；

If the determination result is YES, then the immediate neighbor point number difference tracing point less than the 6th predetermined threshold value is deleted, it is thus achieved that the difference tracing point after noise reduction under two grades of piecemeals；

With the difference tracing point after the noise reduction under each described two grades of piecemeals as seed points, carry out cluster calculation according to seed points, it is thus achieved that one includes the cluster point group of multiple difference tracing point；

Judge whether the difference tracing point after the noise reduction under the described two grades of piecemeals not carrying out clustering, if existing, with the difference tracing point after the noise reduction under each described two grades of piecemeals not carrying out and clustering as seed points, cluster calculation is carried out according to seed points, obtain a cluster point group including multiple difference tracing point, if not existing, stop cluster calculation.

Wherein, cluster calculation is carried out according to seed points, it is thus achieved that one includes the step of the cluster point group of multiple difference tracing point, specifically includes:

A cluster point group is set up according to described seed points；

Search for the immediate neighbor point of described seed points, and judge that whether the distance between described immediate neighbor point and described seed points is more than presetting tolerance；

If the determination result is YES, then stop described immediate neighbor point is carried out next stage immediate neighbor point search；

If judged result is no, then the immediate neighbor point of described seed points is included in described cluster point group, and judge whether the distance between immediate neighbor point and the described seed points of described seed points meets the distance condition of preset search, when meeting, the immediate neighbor point of described seed points is carried out next stage immediate neighbor point search, until heading crossing angle maximum between two difference tracing points after any noise reduction in described cluster point group more than the described default tolerance clustering some group, it is thus achieved that a cluster point group.

Wherein, described to described cluster point a group carry out road shape characteristic test, it is thus achieved that can identify a road cluster point group a step, specifically include:

Judge whether the number of the difference tracing point in each cluster point group is more than the 7th predetermined threshold value；

If the determination result is YES, with the difference tracing point in cluster point group each described as the center of circle, described 4th predetermined threshold value is that radius is justified, and merges described circle, it is thus achieved that the relief area corresponding with each cluster point group；

Calculate the area of described relief area, and judge that whether described area is more than preset area threshold value；

If the determination result is YES, then the average heading of difference tracing point in each described cluster point group is calculated, and using described average heading as the direction, major axis place of described relief area；

Obtain the short axle perpendicular with described major axis according to described major axis, and judge whether described major axis is more than the 8th predetermined threshold value with the length ratio of described short axle,

If the determination result is YES, the shape of the most described relief area visually approximates wire；

When the shape of described relief area visually approximates wire, it is judged that whether the length of the major axis of described relief area is more than the 9th predetermined threshold value；

If the determination result is YES, the most described cluster point group can identify a road.

Wherein, when the key element of described road informations includes type, grade, position, recommended value and reliability value, the cluster point group according to identifying a road obtains the road informations in described predeterminable area, specifically includes:

Number according to the difference tracing point in described cluster point group and the area of the relief area corresponding with described cluster point group, obtain the grade of described road informations；

Obtain the geometric center of the relief area corresponding with described cluster point group, and using described geometric center as the position obtaining described road informations；

The data that described cluster is put the difference tracing point in group are averaged and are obtained the recommended value of described road informations；

The standard deviation of described road informations is obtained according to described recommended value, and using the inverse ratio of described standard deviation as the reliability value of described road informations.

According to another aspect of the present invention, additionally provide the acquisition device of a kind of road informations, including:

Data input module, obtains the tracing point data in predeterminable area and road data in real time；

Described tracing point data and road data for obtaining the information type of the road informations of required extraction, and are carried out calculus of differences according to described information type, it is thus achieved that mate with described information type and difference tracing point unmatched with described road data by difference block；

Cluster module, carries out noise reduction process and cluster calculation to described difference tracing point, it is thus achieved that include the cluster point group of multiple difference tracing point；

Test module, for carrying out road shape characteristic test to described cluster point group, it is thus achieved that can identify the cluster point group of a road；

Information builds module, for obtaining the road informations in described predeterminable area according to the cluster point group that can identify a road.

Wherein, described difference block includes:

Data filtering units, error in the plane obtaining tracing point from described tracing point data, and delete the tracing point data more than the first predetermined threshold value of error in plane, and tracing point data unmatched with described information type, it is thus achieved that the tracing point data after filtration；

One-level blocking unit, for carrying out piecemeal by the tracing point data after described filtration according to geographical network, it is thus achieved that the one-level piecemeal tracing point group in multiple one-level segmented areas and each one-level segmented areas；

Difference unit, for carrying out calculus of differences to the tracing point data in one-level piecemeal tracing point group each described and described road data, it is thus achieved that the difference tracing point under one-level piecemeal tracing point group.

Wherein, when described information type is new added road, described difference unit includes:

First obtains subelement, for obtaining the road being positioned at each described one-level segmented areas according to described road data；

First judgment sub-unit, for judging that whether tracing point in each one-level piecemeal tracing point group minimum vertical line distance to each road belonged to described tracing point in same one-level segmented areas is more than or equal to the second predetermined threshold value, if the determination result is YES, the most described tracing point is difference tracing point；Or

Second judgment sub-unit, for judging the wheeled direction to the minimum vertical line distance of each road belonged in same one-level segmented areas of the tracing point in each one-level piecemeal tracing point group with described tracing point, whether it is more than or equal to the 3rd predetermined threshold value with the absolute value of the difference in the current course of described tracing point, if the determination result is YES, the most described tracing point is difference tracing point.

Wherein, described cluster module includes:

Two grades of blocking unit, for the difference tracing point under each one-level piecemeal, carry out piecemeal according to geographic grid, it is thus achieved that two grades of piecemeal difference tracing point groups；

First judging unit, for judging whether the distance between any two difference tracing point under each two grades of difference block tracing point groups is less than the 4th predetermined threshold value；

Second judging unit, for if the determination result is YES, then judges that whether the absolute value of difference of the course value of said two difference tracing point is less than the 5th predetermined threshold value, if the determination result is YES, then the immediate neighbor point of said two difference tracing point the other side each other；

3rd interpretation unit, for traveling through described difference tracing point successively, it is judged that whether the immediate neighbor point number of each described difference tracing point is less than the 6th predetermined threshold value；

Delete unit, for if the determination result is YES, then delete the immediate neighbor point number difference tracing point less than the 6th predetermined threshold value, it is thus achieved that the difference tracing point after noise reduction under two grades of piecemeals；

First cluster cell, for the difference tracing point after the noise reduction under each described two grades of piecemeals as seed points, carries out cluster calculation according to seed points, it is thus achieved that one includes the cluster point group of multiple difference tracing point；

Second cluster cell, difference tracing point after judging whether the noise reduction under the described two grades of piecemeals not carrying out clustering, if existing, with the difference tracing point after the noise reduction under each described two grades of piecemeals not carrying out and clustering as seed points, cluster calculation is carried out according to seed points, obtaining a cluster point group including multiple difference tracing point, if not existing, stopping cluster calculation.

Wherein, described first cluster cell includes:

Create subelement, for setting up a cluster point group according to described seed points；

First search subelement, for searching for the immediate neighbor point of described seed points, and judges that whether the distance between described immediate neighbor point and described seed points is more than presetting tolerance；

First processes subelement, for if the determination result is YES, then stops described immediate neighbor point is carried out next stage immediate neighbor point search；

Second processes subelement, if being no for judged result, then the immediate neighbor point of described seed points is included in described cluster point group, and judge whether the distance between immediate neighbor point and the described seed points of described seed points meets the distance condition of preset search, when meeting, and the immediate neighbor point of described seed points is carried out next stage immediate neighbor point search, until heading crossing angle maximum between two difference tracing points after any noise reduction in described cluster point group more than the described default tolerance clustering some group, it is thus achieved that a cluster point group.

Wherein, described test module includes:

Number judgment sub-unit, whether the number of the difference tracing point in judging each cluster point group is more than the 7th predetermined threshold value；

Relief area obtains subelement, and for if the determination result is YES, with the difference tracing point in cluster point group each described as the center of circle, described 4th predetermined threshold value is that radius is justified, and merges described circle, it is thus achieved that the relief area corresponding with each cluster point group；

Area judgment sub-unit, for calculating the area of described relief area, and judges that whether described area is more than preset area threshold value；

Major axis computation subunit, for if the determination result is YES, then calculates the average heading of difference tracing point in each described cluster point group, and using described average heading as the direction, major axis place of described relief area；

Wire interpretation subelement, for obtaining the short axle perpendicular with described major axis according to described major axis, and judge whether described major axis is more than the 8th predetermined threshold value with the length ratio of described short axle, if the determination result is YES, the shape of the most described relief area visually approximates wire；

Major axis judgment sub-unit, for when the shape of described relief area visually approximates wire, it is judged that whether the length of the major axis of described relief area is more than the 9th predetermined threshold value；If the determination result is YES, the most described cluster point group cluster point group can identify a road.

Wherein, described information structure module includes:

Grade acquiring unit, for the number according to the difference tracing point in described cluster point group and the area of the relief area corresponding with described cluster point group, obtains the grade of described road informations；

Position acquisition unit, for obtaining the geometric center of the relief area corresponding with described cluster point group, and using described geometric center as obtaining described road informations exact position；

Recommended value acquiring unit, obtains the recommended value of described road informations for averaging the data of the difference tracing point in described cluster point group；

Reliability acquiring unit, for obtaining the standard deviation of described road informations, and using the inverse ratio of described standard deviation as the reliability value of described road informations according to described recommended value.

The invention has the beneficial effects as follows:

The acquisition methods of the road informations of the present invention, by obtaining magnanimity tracing point data and road data in real time, tracing point data and road data are carried out calculus of differences, and is obtained the cluster point group that can identify a road by noise reduction and clustering algorithm, and then obtain road informations.Therefore, the acquisition methods of the road informations of the present invention, it is adaptable to the internet, applications location data differed in magnanimity and alignment sensor source, and can find to exist the road informations of the center in clover leaf city equally.Additionally, due to have employed the thinking of tracing point and Road vector matching, so need not set up so-called trajectory, therefore positioning the sample frequency not requirement of data, and then improving the response speed of navigation map more new production.

Accompanying drawing explanation

Fig. 1 represents the flow chart of the acquisition methods of the road informations of the embodiment of the present invention；

Fig. 2 represents the acquisition device structured flowchart of the road informations of the embodiment of the present invention；

Fig. 3 represents the structured flowchart of the difference block of the embodiment of the present invention；

Fig. 4 represents the structured flowchart of the difference unit of the embodiment of the present invention；

Fig. 5 represents the structured flowchart of the cluster module of the embodiment of the present invention；

Fig. 6 represents the structured flowchart of the first cluster cell of the embodiment of the present invention；

Fig. 7 represents the structured flowchart of the test module of the embodiment of the present invention；

Fig. 8 represents that the information of the embodiment of the present invention builds the structured flowchart of module；

Fig. 9 represents the principle schematic of the cluster calculation of the embodiment of the present invention；

Figure 10 represents the acquisition methods entirety principle schematic of the road informations of the embodiment of the present invention.

Detailed description of the invention

It is more fully described the exemplary embodiment of the disclosure below with reference to accompanying drawings.Although accompanying drawing showing the exemplary embodiment of the disclosure, it being understood, however, that may be realized in various forms the disclosure and should not limited by embodiments set forth here.On the contrary, it is provided that these embodiments are able to be best understood from the disclosure, and complete for the scope of the present disclosure can be conveyed to those skilled in the art.

Embodiment one

An aspect according to the embodiment of the present invention, it is provided that the acquisition methods of a kind of road informations, first the method, obtains the tracing point data in predeterminable area and road data in real time；Then, obtain the information type of the required road informations extracted, and according to described information type, described tracing point data and road data are carried out calculus of differences, it is thus achieved that mate with described information type and difference tracing point unmatched with described road data；Again, described difference tracing point is carried out noise reduction process and cluster calculation, it is thus achieved that include the cluster point group of multiple difference tracing point；Again, described cluster point group is carried out road shape characteristic test, it is thus achieved that the cluster point group of a road can be identified；Finally, the road informations in described predeterminable area is obtained according to the cluster point group that can identify a road.

Therefore, the acquisition methods of the road informations of the embodiment of the present invention, by a large amount of tracing point data and existing road data real time differential, and extracts " new added road " or " road generation great change " information by noise reduction and clustering algorithm.

As it is shown in figure 1, the method includes:

Tracing point data in step S11, in real time acquisition predeterminable area and road data.

Wherein, user's positioning track data that tracing point data are applied from mobile Internet, and every tracing point data generally comprise following key element:

1, User Identity number (ID), i.e. anonymous identification, is used for identifying tracing point produced by same user)；

2, timestamp, is typically accurate to the second；

3, current plane positioning result, including longitude and latitude；

4, error in the plane of location；

5, present speed；

6, current course.

Additionally, in the common application scenarios of the present invention, be generally at least all user's positioning track data to one province in January on the 1st, common data scale is 1,000,000,000～30,000,000,000 track datas.

For road data, obtained by the up-to-date transportation database safeguarded from navigation map production unit, and this data base is generally large commercial relevant database." road informations " described in the acquisition methods of the road informations of the embodiment of the present invention i.e. in order to point out the difference of reality change and known navigation transportation database, by extracting " road informations ", responds " information ", and purpose is also to update this data base rapidly.

In addition, the key element of the acquisition methods of the road informations that road key element is the embodiment of the present invention of road data, only road factor data is the newest, the guiding that just can be given closest to reality by the acquisition methods of the road informations of the embodiment of the present invention is calculated, and the main information of road key element is stored in link table, its key concept model is that every road key element is expressed as the limit of a directed graph, connecting two road junctions, its important attribute comprises:

1, geometric representation, many some strings that shape point represents in order；

2, current direction；

3, grade, such as national highway, provincial highway, county road；

4, form, such as through street, bypass, service area, ring road, internal passageway, tunnel；

5, passage rate or speed limit.

Magnanimity tracing point data are carried out calculus of differences by road important attribute described above, for find magnanimity tracing point in statistical significance with the significant difference of known road attribute.

Step S13, the information type of the required road informations extracted of acquisition, and according to described information type, described tracing point data and road data are carried out calculus of differences, it is thus achieved that mate with described information type and difference tracing point unmatched with described road data.

Wherein, step S13 specifically includes:

Owing to needs obtain different types of road informations, therefore track point data and road data are carried out calculus of differences when, also should have different difference strategies or difference method, such as, when information type is new added road, described tracing point data in one-level piecemeal tracing point group each described and described road data are carried out calculus of differences, it is thus achieved that the difference tracing point under one-level piecemeal tracing point group, specifically include:

Wherein, if a certain road is neighbouring around the corner there is a tracing point, then this tracing point exists multiple to the vertical line distance of this road, therefore, in the acquisition methods of the road informations of the embodiment of the present invention, that need acquisition is minimum in multiple vertical line distance one, when this minimum vertical line distance is more than or equal to the second predetermined threshold value, this tracing point does not mates with this road, when this tracing point with belong to this tracing point the road in same segmented areas all cannot mate time, this tracing point is difference tracing point.

Or, also by judging that tracing point pair and this tracing point belong to the absolute value of the difference in the wheeled direction of the minimum vertical line distance of each road in same segmented areas whether more than or equal to the 3rd predetermined threshold value, judge whether this tracing point is difference tracing point.

In the acquisition methods of the road informations of the embodiment of the present invention, owing to tracing point data volume is big and quality is uncontrollable, therefore in the calculus of differences to track point data and road data, " filter " firstly the need of track point data is carried out data, only extract the tracing point data matched with information type, filter the undesirable data of obvious quality, and data volume can be greatly decreased, be more beneficial for later stage treatment effeciency and promote.

For " filter method " therein, on the one hand, error in tracing point data midplane is rejected more than the tracing point of the first predetermined threshold value, the width of such as Ordinary Rd is 10 meters～50 meters, if the middle error of the most a certain tracing point is more than 20 meters, i.e. it is believed that the positioning precision of this point is poor, can reject；On the other hand, need according to information type, track point data to be screened, it is " newly-increased highway " that such as certain secondary data processes the information type needing to extract, i.e. it is indifferent to internal passageway, so " speed " attribute of tracing point can be filtered, less than certain threshold value, such as 15,000 ms/h (km/h), think that it is positioned at internal passageway in statistical significance, this type of tracing point should be rejected.

But, after data filtering, although data volume is greatly decreased, but still fall within mass data, i.e. cannot directly read in calculator memory and disposably analyze and process.Therefore must be by magnanimity tracing point piecemeal, the data volume size of piecemeal should be as the criterion disposably can be loaded into calculator memory.Wherein, each one-level segmented areas comprises tracing point data volume and has learned that circuit-switched data.

Owing to track data point distance each other is the nearest, relation is the strongest, and distance is the most remote, and relation is the most weak.Therefore, conventional partition strategy is to carry out piecemeal according to geographic grid.Piecemeal is the least, and the effect of the data volume reducing single treatment is the most obvious, if but piecemeal is too small, then and fragmentation is serious, can affect the effect of subsequent treatment.Through practice summary, it is normally applied the piecemeal size of 10km*10km to 30km*30km.

After tracing point data carry out one-level piecemeal, the relation between block and block can be ignored, individually process, say, that the big data frameworks such as mapping/reduction (Map/Reduce) can be utilized to carry out distributed process, from framework, ensure the quick process of mass data.

For the one-level piecemeal tracing point group in each one-level segmented areas, needing to carry out calculus of differences with the road data in this segmented areas, the difference algorithm wherein applied is the reverse application of " map match " device common in navigation application.

Wherein, " map match " is described below:

Assume that A is a tracing point, its position be (x, y), and has L1 in current map piecemeal, L2 ..., n the road segment segment such as Ln, wherein, x and y represents the position coordinate value of tracing point A, L1, L2 ..., Ln represents different road segment segment.General, if A point meets:

1, (x, y) distance is to the vertical dimension of certain road segment segment Li less than the second predetermined threshold value, and wherein, the second predetermined threshold value is usually 1/2nd of road width, and i takes positive integer in the position of A；

2, the current course of road segment segment Li representated by the closest approach of A Yu Li, i.e. intersection point, the difference absolute value being worth with the current course of A point, should be less than the 3rd predetermined threshold value, wherein, the 3rd predetermined threshold value is typically set to 30 degree；

I.e. it is believed that A point the match is successful with Li.Assume that A point all cannot mate with when all roads in previous stage segmented areas, then it is assumed that A point is " difference tracing point ".

Because carrying out between calculus of differences, track point data is screened by the acquisition methods at the road informations of the embodiment of the present invention according to information type, and the method that therefore every secondary data process performs the embodiment of the present invention can only obtain a certain types of class information.Such as, information type is " road speed significantly changes ", then the tracing point of the maximum possible passage rate that present speed should be not more than road segment segment is identified as difference tracing point, therefore the some set being combined into by this class tracing point, can be used to extract the information of " road speed significantly changes ".

Step S15, described difference tracing point is carried out noise reduction process and cluster calculation, it is thus achieved that include the cluster point group of multiple difference tracing point.

Wherein, assume that initial trace point data quality is the most reliable, positioning precision is in 10 meters, so " difference tracing point " all will be located in " new added road " or the roadside of " road attribute generation great change ", but the multiformity due to mobile Internet subscriber equipment, system cannot its quality of data of accurate evaluation, such as " error in plane positioning " inherently statistical estimation, also there is uncertainty, therefore " difference tracing point " set obtained through difference processing contains substantial amounts of noise, i.e. cannot successfully and the tracing point of path adaptation owing to the quality of data is low.Therefore, the acquisition methods of the road informations of the embodiment of the present invention, need the difference tracing point to obtaining to carry out noise reduction process and cluster calculation.

Wherein, step S15 specifically includes:

A cluster point group is set up according to described seed points；

In the acquisition methods of the road informations of the embodiment of the present invention, after tracing point data in each one-level segmented areas and road data are carried out calculus of differences, the data scale of the difference tracing point in each the one-level segmented areas obtained is about the 10% of original data volume, but it is for cluster seeking, the most excessive.Analyze the feasibility of plane piecemeal when " one-level piecemeal " is described, therefore, piecemeal can have been carried out further herein, reduce the operand of cluster seeking, be referred to as " two grades of piecemeals ".If the 1/t of the size one-level piecemeal of two grades of piecemeals, then search arithmetic amount can be reduced to { (n/3 (t^2)) }！, wherein, n represents the number of the difference tracing point in one-level segmented areas, and t is positive integer.But, two grades of piecemeals can not be too small, it is assumed that the road informations being set smaller than 100 meters that system data processes does not processes, then the size of piecemeal cannot be less than 100 meters.Practice summary, be typically set to 500 meters relatively reasonable to 1 km.

In the acquisition methods of the road informations of the embodiment of the present invention, after two grades of piecemeals, noise reduction process and cluster calculation for the difference tracing point in each two grades of segmented areas are all parallel processings, and equal non-interference between each two grades of segmented areas.

In addition, between multiple difference tracing points in two grades of segmented areas, there is certain relation, such as, distance and the absolute value of the difference of course value between any two difference tracing point meet pre-conditioned, then have immediate neighbor relation between two difference tracing points.Therefore, the difference tracing point not having " neighbours " or " neighbours " little near can be referred to as " vision noise ".Vision noise is easy to disallowable, therefore the acquisition methods of the road informations of the embodiment of the present invention, by judging " immediate neighbor " number of each difference tracing point, if less than the 6th predetermined threshold value, then it is assumed that be vision noise, directly reject, if exceeding this threshold value, then stop traversal.Practice summary, rejecting the difference tracing point quantity after vision noise is normally about 1/3 before rejecting, and further reduces the operand of the acquisition methods of the road informations of the embodiment of the present invention, and then improving operational speed.

In addition, in the search procedure of immediate neighbor point, also can preset a distance condition, such as assume the preferential recursive search distance point more than 5 meters, if the size of two grades of piecemeals is 1 km, the number of times then searched for should be square (1 km/5 meter=200) of no more than 200, due in search procedure, it is constrained to course difference, therefore the direction searching for extension should be fixed on a scope, do not have repeatedly, it is understood that square (1 km/5 meter=200) that time scale is not over 200.Mathematic(al) representation can be summarized as (s/d) ^2.Wherein s is two fraction block sizes, and d is depth-first distance during neighbor seaching, further reduces operand.

In order to acquisition road informations, need the difference tracing point after the noise reduction under each two grades of piecemeal is carried out cluster calculation, the concrete principle of cluster calculation is as shown in Figure 9, the difference tracing point included after assuming two grades of piecemeal noise reductions is A, B, C, D, B1, B2, C1, C2, D1, D2, wherein, line segment between letter represents that between two letters be immediate neighbor relation, and the letter being positioned at same circle represents and belongs to same cluster point group.

In the acquisition methods of the road informations of the embodiment of the present invention, multiple difference tracing points as shown in Figure 9 are carried out specifically comprising the following steps that of cluster calculation

Step 801: with each point of A～D2 as seed, creates 10 cluster point groups successively；

Step 802: the immediate neighbor point of search A, is found to be B and C, and judges that B, C distance respectively and between A, whether more than presetting tolerance, finds not larger than, then to be included in by B and C in the first cluster point group；

Step 803: the immediate neighbor point of search B and C, is found to be B1, B2, C1, C2, and judges whether the distance between B1, B2, C1, C2 and A has exceeded default tolerance successively；Find C2 exceed, and other not less than, then B1, B2, C1 are included in the first cluster point group, and stop search, it is thus achieved that one cluster point a group；

Step 804: be then seed points to B point, search for its immediate neighbor, finding the A searched, B1, B2 are in step 802, the when of 803, recorded in " the immediate neighbor list " of B point, therefore, do not find new immediate neighbor, then B is that the search of seed points terminates, and namely this some group is invalid；

Step 805: then to B1, B2, C1, C2 successively as seed points, repeat step 804, find with B1, B2 and C1 to be that the cluster point group that seed points creates is the most invalid；

Step 806: scan for the immediate neighbor of D point, D1 and D2 not less than tolerance, is then included in the second cluster point group by the distance being found to be between D1 and D2, and D1 and D2 and A；

Step 807: find that institute had the most been searched for as seed points, then circulate stopping, finally giving A, D two cluster point group, the first cluster point group the most as shown in Figure 8 and the second cluster point group.

Step S17, to described cluster point a group carry out road shape characteristic test, it is thus achieved that the cluster point group of a road can be identified.

The cluster point group obtained in step S15 can't clearly represent a disappearance section, therefore, it is necessary to cluster point group is carried out road shape characteristic test, filters out the cluster point group really being able to represent " disappearance road shape ".

Specifically, step S17 includes:

Obtaining the short axle perpendicular with described major axis according to described major axis, and judge whether described major axis is more than the 8th predetermined threshold value with the length ratio of described short axle, if the determination result is YES, the shape of the most described relief area visually approximates wire；

After the multiple cluster point groups obtained are completed, filter out multiple cluster point group really being able to represent " road shape ".For multiple qualified cluster point groups, the form of difference tracing point Clustering table can be used to store, specifically, as shown in the table:

Table 1: difference tracing point Clustering table

Therefore, it can in difference tracing point Clustering table to every record, carry out the corresponding road informations of structure according to information type.Owing to this table is to design according to the standard of geographic information data, therefore can be visualized by general cartographic software.

Step S19, basis can identify the cluster point group of a road and obtain the road informations in described predeterminable area.

Wherein, step S19 specifically includes:

The qualified cluster point group obtained according to step S17, builds and exports " navigation map road informations table " the i.e. road informations of respective type, and wherein, every information is from a record of " difference tracing point Clustering table ", and needs to comprise following key element:

1, type, such as new added road, road driving direction change etc.；

2, grade, i.e. importance: can be in conjunction with the difference tracing point number in cluster point group, the size of relief area, and the long axis length of relief area, provide a rating calculation function, with these three variable positive correlations: the most proofs of counting this to build the source data amount of this information big, the area coverage of relief area area this information of the biggest proof is big, and the link length that buffer length this information of the longest proof is corresponding is longer.Therefore, the design rule of rating calculation function is: grade and PointsOfDiff, the Buffer area in difference tracing point Clustering table, TrackLength is all directly proportional.

3, position, such as exact position and scope: i.e. Buffer field in difference tracing point Clustering table, difference tracing point Clustering buffering polygon facet.Can be by asking geometric center, external contact zone etc. to set up spatial index on Buffer polygon, it is simple to space querying.

4, recommended value: can obtain by the tracing point correspondence attribute in packet is averaged；Such as: when information type is " travel direction renewal ", " travel direction " of " difference tracing point " in Clustering is averaged, is " recommended value " of information.

5, reliability: reliability is a statistic concept, is similar to " confidence level ".Standard deviation can be used to combine " recommended value " calculating estimate, standard deviation is the least, then reliability is the highest.

So far, the acquisition methods of the road informations of the embodiment of the present invention can obtain the road informations of respective type.According to each key element of information, can inquire about, sequence, business personnel according to service needed, can carry out the scheduling of resource of information processing: such as, is distributed to not commensurate or department according to type, it is dispatched to relevant treatment personnel according to position, by grade, reliably carries out prioritization.In the generation of whole " information " and processing procedure, quantity, quality, resource, the time is the most controlled.

The arrangement principle schematic of the acquisition methods of the road informations of the embodiment of the present invention in sum, as shown in Figure 10, with it, the magnanimity tracing point different types of road informations of data quick obtaining gathered can be passed through, the navigation system so that upgrading in time, it is simple to user uses.

Embodiment two

According to another aspect of the present invention, additionally provide the acquisition device of a kind of road informations, as in figure 2 it is shown, this device 200 includes:

Data input module 201, obtains the tracing point data in predeterminable area and road data in real time；

Difference block 203, for obtaining the information type of the road informations of required extraction, and according to described information type, described tracing point data and road data are carried out calculus of differences, it is thus achieved that mate with described information type and difference tracing point unmatched with described road data；

Cluster module 205, carries out noise reduction process and cluster calculation to described difference tracing point, it is thus achieved that include the cluster point group of multiple difference tracing point；

Test module 207, for carrying out road shape characteristic test to described cluster point group, it is thus achieved that can identify the cluster point group of a road；

Information builds module 209, for obtaining the road informations in described predeterminable area according to the cluster point group that can identify a road.

Alternatively, as it is shown on figure 3, described difference block 203 includes:

Data filtering units 2031, error in the plane obtaining tracing point from described tracing point data, and delete the tracing point data more than the first predetermined threshold value of error in plane, and tracing point data unmatched with described information type, it is thus achieved that the tracing point data after filtration；

One-level blocking unit 2032, for carrying out piecemeal by the tracing point data after described filtration according to geographical network, it is thus achieved that the one-level piecemeal tracing point group in multiple one-level segmented areas and each one-level segmented areas；

Difference unit 2033, for carrying out calculus of differences to the tracing point data in one-level piecemeal tracing point group each described and described road data, it is thus achieved that the difference tracing point under one-level piecemeal tracing point group.

Alternatively, as shown in Figure 4, when described information type is new added road, described difference unit 2033 includes:

First obtains subelement 20331, for obtaining the road being positioned at each described one-level segmented areas according to described road data；

First judgment sub-unit 20332, for judging that whether tracing point in each one-level piecemeal tracing point group minimum vertical line distance to each road belonged to described tracing point in same one-level segmented areas is more than or equal to the second predetermined threshold value, if the determination result is YES, the most described tracing point is difference tracing point；Or

Second judgment sub-unit 20333, for judging the wheeled direction to the minimum vertical line distance of each road belonged in same one-level segmented areas of the tracing point in each one-level piecemeal tracing point group with described tracing point, whether it is more than or equal to the 3rd predetermined threshold value with the absolute value of the difference in the current course of described tracing point, if the determination result is YES, the most described tracing point is difference tracing point.

Alternatively, as it is shown in figure 5, described cluster module 205 includes:

Two grades of blocking unit 2051, for the difference tracing point under each one-level piecemeal, carry out piecemeal according to geographic grid, it is thus achieved that two grades of piecemeal difference tracing point groups；

First judging unit 2052, for judging whether the distance between any two difference tracing point under each two grades of difference block tracing point groups is less than the 4th predetermined threshold value；

Second judging unit 2053, for if the determination result is YES, then judges that whether the absolute value of difference of the course value of said two difference tracing point is less than the 5th predetermined threshold value, if the determination result is YES, then the immediate neighbor point of said two difference tracing point the other side each other；

3rd interpretation unit 2054, for traveling through described difference tracing point successively, it is judged that whether the immediate neighbor point number of each described difference tracing point is less than the 6th predetermined threshold value；

Delete unit 2055, for if the determination result is YES, then delete the immediate neighbor point number difference tracing point less than the 6th predetermined threshold value, it is thus achieved that the difference tracing point after noise reduction under two grades of piecemeals；

First cluster cell 2056, for the difference tracing point after the noise reduction under each described two grades of piecemeals as seed points, carries out cluster calculation according to seed points, it is thus achieved that one includes the cluster point group of multiple difference tracing point；

Second cluster cell 2057, difference tracing point after judging whether the noise reduction under the described two grades of piecemeals not carrying out clustering, if existing, with the difference tracing point after the noise reduction under each described two grades of piecemeals not carrying out and clustering as seed points, cluster calculation is carried out according to seed points, obtaining a cluster point group including multiple difference tracing point, if not existing, stopping cluster calculation.

Alternatively, as shown in Figure 6, described first cluster cell 2056 includes:

Create subelement 20561, for setting up a cluster point group according to described seed points；

First search subelement 20562, for searching for the immediate neighbor point of described seed points, and judges that whether the distance between described immediate neighbor point and described seed points is more than presetting tolerance；

First processes subelement 20563, for if the determination result is YES, then stops described immediate neighbor point is carried out next stage immediate neighbor point search；

Second processes subelement 20564, if being no for judged result, then the immediate neighbor point of described seed points is included in described cluster point group, and judge whether the distance between immediate neighbor point and the described seed points of described seed points meets the distance condition of preset search, when meeting, the immediate neighbor point of described seed points is carried out next stage immediate neighbor point search, until heading crossing angle maximum between two difference tracing points after any noise reduction in described cluster point group more than the described default tolerance clustering some group, it is thus achieved that a cluster point group.

Alternatively, as shown in Figure 7, it is characterised in that, described test module 207 includes:

Number judgment sub-unit 2071, whether the number of the difference tracing point in judging each cluster point group is more than the 7th predetermined threshold value；

Relief area obtains subelement 2072, and for if the determination result is YES, with the difference tracing point in cluster point group each described as the center of circle, described 4th predetermined threshold value is that radius is justified, and merges described circle, it is thus achieved that the relief area corresponding with each cluster point group；

Area judgment sub-unit 2073, for calculating the area of described relief area, and judges that whether described area is more than preset area threshold value；

Major axis computation subunit 2074, for if the determination result is YES, then calculates the average heading of difference tracing point in each described cluster point group, and using described average heading as the direction, major axis place of described relief area；

Wire interpretation subelement 2075, for obtaining the short axle perpendicular with described major axis according to described major axis, and judge whether described major axis is more than the 8th predetermined threshold value with the length ratio of described short axle, if the determination result is YES, the shape of the most described relief area visually approximates wire；

Major axis judgment sub-unit 2076, for when the shape of described relief area visually approximates wire, it is judged that whether the length of the major axis of described relief area is more than the 9th predetermined threshold value；If the determination result is YES, the most described cluster point group cluster point group can identify a road.

Alternatively, as shown in Figure 8, described information structure module 209 includes:

Grade acquiring unit 2091, for the number according to the difference tracing point in described cluster point group and the area of the relief area corresponding with described cluster point group, obtains the grade of described road informations；

Position acquisition unit 2092, for obtaining the geometric center of the relief area corresponding with described cluster point group, and using described geometric center as obtaining described road informations exact position；

Recommended value acquiring unit 2093, obtains the recommended value of described road informations for averaging the data of the difference tracing point in described cluster point group；

Reliability acquiring unit 2094, for obtaining the standard deviation of described road informations, and using the inverse ratio of described standard deviation as the reliability value of described road informations according to described recommended value.

Above-described is the preferred embodiment of the present invention; should be understood that for the ordinary person of the art; can also make some improvements and modifications under without departing from principle premise of the present invention, these improvements and modifications are the most within the scope of the present invention.

Claims

1. the acquisition methods of a road informations, it is characterised in that including:

2. the method for claim 1, it is characterized in that, described according to described information type, described tracing point data and road data are carried out calculus of differences, it is thus achieved that mate with described information type and difference tracing point unmatched with described road data, specifically include:

3. method as claimed in claim 2, it is characterized in that, when described information type is new added road, described tracing point data in one-level piecemeal tracing point group each described and described road data are carried out calculus of differences, obtain the difference tracing point under one-level piecemeal tracing point group, specifically include:

4. method as claimed in claim 2, it is characterised in that described described difference tracing point is carried out noise reduction process and cluster calculation, it is thus achieved that include a step for the cluster point group of multiple difference tracing point, specifically include:

5. method as claimed in claim 4, it is characterised in that carry out cluster calculation according to seed points, it is thus achieved that includes the step of the cluster point group of multiple difference tracing point, specifically includes:

A cluster point group is set up according to described seed points；

6. method as claimed in claim 4, it is characterised in that described a described cluster point group is carried out road shape characteristic test a, it is thus achieved that step for the cluster point group of a road can be identified, specifically include:

7. method as claimed in claim 6, it is characterized in that, when the key element of described road informations includes type, grade, position, recommended value and reliability value, the cluster point group according to identifying a road obtains the road informations in described predeterminable area, specifically includes:

8. the acquisition device of a road informations, it is characterised in that including:

9. device as claimed in claim 8, it is characterised in that described difference block includes:

10. device as claimed in claim 9, it is characterised in that when described information type is new added road, described difference unit includes:

11. devices as claimed in claim 9, it is characterised in that described cluster module includes:

12. devices as claimed in claim 11, it is characterised in that described first cluster cell includes:

Second processes subelement, if being no for judged result, then the immediate neighbor point of described seed points is included in described cluster point group, and judge whether the distance between immediate neighbor point and the described seed points of described seed points meets the distance condition of preset search, when meeting, the immediate neighbor point of described seed points is carried out next stage immediate neighbor point search, until heading crossing angle maximum between two difference tracing points after any noise reduction in described cluster point group more than the described default tolerance clustering some group, it is thus achieved that a cluster point group.

13. devices as claimed in claim 11, it is characterised in that described test module includes:

14. devices as claimed in claim 13, it is characterised in that described information builds module and includes: