CN108492276A - A kind of vector link change detection method and device based on similarity measurement - Google Patents
A kind of vector link change detection method and device based on similarity measurement Download PDFInfo
- Publication number
- CN108492276A CN108492276A CN201810083737.2A CN201810083737A CN108492276A CN 108492276 A CN108492276 A CN 108492276A CN 201810083737 A CN201810083737 A CN 201810083737A CN 108492276 A CN108492276 A CN 108492276A
- Authority
- CN
- China
- Prior art keywords
- road
- chain
- similarity
- node
- matched
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30248—Vehicle exterior or interior
- G06T2207/30252—Vehicle exterior; Vicinity of vehicle
- G06T2207/30256—Lane; Road marking
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Geometry (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
- Traffic Control Systems (AREA)
Abstract
The vector link change detection method and device based on similarity measurement that the present invention relates to a kind of, belong to map vector data library dynamic update method field.The present invention reconstructs topological relation to road data collection to be detected first, extracts Road chain, and determine the road radian for including in the chain of road；Then the matching candidate collection of road to be matched is searched for using the buffer way based on consistency constraint；Similarity evaluation model is established further according to the geometric properties of road, selects the highest road object of a similarity as the matching object of road to be matched from matching candidate concentration using the evaluation model；Feature difference comparison finally is carried out to physical road of the same name and road to be matched, with the situation of change of determination road to be matched.The present invention accurately detects road entity of the same name changes in which feature by the feature difference performance between calculating road, and the implementation of extraction and update operation for change information provides guarantee, has very high application value.
Description
Technical field
The vector link change detection method and device based on similarity measurement that the present invention relates to a kind of, belong to map vector
Dynamic data base updating technical field.
Background technology
Road element is the trunk element in topographic map, and variation a kind of element the most outstanding, in order to ensure road
The Up-to-date state of data, it is necessary to which real-time update is carried out to road data.In road network incremental update, which occurs actually for road entity
A little variations are that how to detect and how to describe these variations is the critical issue in road network update, they are directly affected
Storage organization, increment information acquisition, update processing, change information analysis and the efficiency of publication and level of change information.
Currently, to the detection of change information and expression, scholar carries out correlative study.Zhu Huaji, Chen Jun etc. are proposed
Increment information classification based on geographic event and target snapshot difference and expression, and give the change based on event and snapshot difference
Change definition and the expression model of information, but such method is only classified from simple layer in face of change information, and change is ignored
The diversity for changing information, there is no consider the situation of change of road network data in complex situations；Ji Cunwei, which is proposed, to be directed to
Residential feature is changed detection by calculating graph data difference, and differentiates its type, it is contemplated that simply with the variation of complexity
Type, however graph data difference is defective in the application of the Linear Entities such as road.
Invention content
The vector link change detection method based on similarity measurement that the object of the present invention is to provide a kind of, with solving road
The variation problem that detection accuracy is low, applicability is not strong；The present invention also provides a kind of link changes based on similarity measurement
Detection device.
The vector link change detection based on similarity measurement that In order to solve the above technical problems, the present invention provides a kind of
Method, including eight schemes, method scheme one：Detection method includes the following steps for this：
1) topology reconstruction is carried out to data to be tested, extracts Road chain, and determine the road segmental arc for including in the chain of road；
2) the matching candidate collection of Road chain to be matched is searched for using the buffer way based on consistency constraint；
3) spatial simlanty evaluation model is established according to the geometric properties of road, is concentrated and is selected from matching candidate using the model
Select of the same name physical road of the highest road object of a similarity as road to be matched；
4) feature difference comparison is carried out to detection road to be changed and its physical road of the same name, with determination detection road to be changed
The situation of change on road.
Method scheme two：On the basis of method scheme one, the determination of the matching candidate collection in the step 2), specific mistake
Journey is as follows：
A. buffering area is established according to search radius to each node in Road chain to be matched, another data is centrally located at
Correspondence candidate matches node of the road node as each node in corresponding Road chain to be matched in each buffering area；
B. to the candidate matches node of each node into walking along the street chain consistency detection, by all times on the chain of same road
It selects matched nodes as one group, and is ranked up according to the precedence for constituting road chain, it will be where each group of candidate matches node
Road chain as candidate matches road chain be added to candidate matches concentrate；
C. according to the node correspondence of Road chain to be matched and candidate matches road chain, all roads to be evaluated are extracted
Section matching relationship pair, by the section matching relationship of these matching evaluations to be evaluated to being put into chained list, to obtain road to be matched
The Candidate Set set of matches of Lu Lulian.
Method scheme three：On the basis of method scheme one, the spatial simlanty evaluation model of foundation in the step 3)
It is as follows：
Sim=ω1SimS+ω2SimD+ω3SimL+ω4SimA
Wherein SimS、SimD、SimLAnd SimAIt is shape similarity S respectivelyShape, apart from proximity SDistance, length it is similar
Spend SLengthWith direction similarity SOrientationIt is dimensionless normalized treated value, ω1、ω2、ω3And ω4For corresponding index
Weight, and ω1+ω2+ω3+ω4=1.
Method scheme four：On the basis of method scheme three, the shape similarity SShapeIndicate road line feature it
Between shape similarity distance, using turn to function be calculated, calculation formula is as follows：
P is 1,
Wherein Dshape(L1,L2) reality expression broken line L1And L2Turn to the difference of functionTo horizontal direction
Project the area of area defined, Dshape_toleranceFor the empirical value of shape similarity distance, Dshape(L1,L2) value gets over
Greatly, broken line L1And L2The similarity of shape is with regard to smaller.
Method scheme five：It is described apart from proximity S on the basis of method scheme threeDistanceRefer to that Road is wanted
Proximity between element, the distance between line feature indicate that calculation formula is as follows using approximate broken line average distance：
Wherein, dav(L1,L2) indicate broken line L1And L2Between approximate broken line average distance, lk.i,i+1, k=1 or 2, expression
Vertex is from Lk.iTo Lk.i+1Line segment, | lk.i,i+1| indicate the length of the line segment, lk.i,i′Indicate vertex from Lk.iTo L 'k.iLine
Section, dtoleranceFor distance threshold, value is the maximum value that two broken line mapping nodes are adjusted the distance.
Method scheme six：On the basis of method scheme three, the length similarity SLengthRefer to road to be matched
Similitude in length,
Wherein Δ ltoleranceIt is the threshold value of road segmental arc difference in length.
Method scheme seven：On the basis of method scheme three, the direction similarity SOrientationRefer to Road
General direction difference between section, general direction refers to the angle that the line of road section first and last node is rotated relative to trunnion axis
Degree,
General direction differences of the wherein Δ θ between the section of two road, Δ θtoleranceFor direction discrepancy threshold.
Method scheme eight：On the basis of method scheme one, link change detection method in the step 4) is to pass through meter
Calculate the shape similarity S of detection road and its physical road of the same name to be changedShape, apart from proximity SDistance, length similarity
SLengthWith direction similarity SOrientation, and be compared with corresponding threshold value, judge detection road to be changed in individual features
On have it is unchanged, if the similarity of some above-mentioned feature be more than threshold value, illustrate that detection road to be changed does not have in this feature
It changes, otherwise regards as changing, finally according to the situation of change of roadway characteristic, determine the type of variation.
The vector link change detection device based on similarity measurement that the present invention also provides a kind of, including following four side
Case, device scheme one：The detection device includes that Road chain generation module, matching candidate collection determining module, space are similar
Property evaluation module and change detection module；
The Road chain generation module is used to extract the Road chain of road data concentration, and determines and wrapped in Road chain
The road segmental arc contained；
The matching candidate collection determining module is used to wait for using the buffering area searching method determination based on consistency constraint
Matching candidate collection with road；
The spatial simlanty evaluation module is used to establish spatial simlanty evaluation model according to the geometric properties of road, profit
The model is used to select the highest road object of a similarity as physical road of the same name from matching candidate concentration；
The change detection module is used to carry out feature difference ratio to detection road to be changed and its physical road of the same name
Compared with the situation of change of determination road to be analyzed.
Device scheme two：On the basis of device scheme one, the matching candidate collection determining module determines matching candidate collection
Process it is as follows：
A. buffering area is established according to search radius to each node in Road chain to be matched, another data is centrally located at
Correspondence candidate matches node of the road node as each node in corresponding Road chain to be matched in each buffering area；
B. to the candidate matches node of each node into walking along the street chain consistency detection, by all times on the chain of same road
It selects matched nodes as one group, and is ranked up according to the precedence for constituting road chain, it will be where each group of candidate matches node
Road chain as candidate matches road chain be added to candidate matches concentrate；
C. according to the node correspondence of Road chain to be matched and candidate matches road chain, all roads to be evaluated are extracted
Section matching relationship pair, by the section matching relationship of these matching evaluations to be evaluated to being put into chained list, to obtain road to be matched
The Candidate Set set of matches of Lu Lulian.
Device scheme three：On the basis of device scheme one, the space phase of road similarity evaluation module foundation
It is as follows like property evaluation model：
Sim=ω1SimS+ω2SimD+ω3SimL+ω4SimA
Wherein SimS、SimD、SimLAnd SimAIt is shape similarity S respectivelyShape, apart from proximity SDistance, length it is similar
Spend SLengthWith direction similarity SOrientationIt is dimensionless normalized treated value, ω1、ω2、ω3And ω4For corresponding index
Weight, and ω1+ω2+ω3+ω4=1.
Device scheme four：On the basis of device scheme one, the change detection module is by calculating detection to be changed
The shape similarity S of road and its physical road of the same nameShape, apart from proximity SDistance, length similarity SLengthThe direction and
Similarity SOrientation, and be compared with corresponding threshold value, it is unchanged to judge that detection road to be changed has in individual features,
If the similarity of some above-mentioned feature is more than threshold value, illustrate that detection road to be changed does not change in this feature,
Otherwise it regards as changing, finally according to the situation of change of roadway characteristic, determines the type of variation.
The beneficial effects of the invention are as follows:The present invention extracts Road chain first, and determines the road radian for including in the chain of road；
Then the matching candidate collection of road to be matched is searched for using the buffer way based on consistency constraint；It is special further according to the geometry of road
Sign establishes similarity evaluation model, and one highest road of similarity of selection is concentrated from matching candidate using the similarity evaluation model
Of the same name physical road of the road object as road to be matched；Finally detection road to be changed and its physical road of the same name are carried out special
Comparison in difference is levied, with the situation of change of determination detection road to be changed.The present invention is accurate by the similar performance between calculating road
Which really detect road to change in feature, the implementation of extraction and update operation for change information provides guarantor
Card has very high application value.
Description of the drawings
Fig. 1 is link change detection identification process figure；
The roads Fig. 2-a chain S1With candidate matches road chain S2Between matching relationship schematic diagram that may be present；
Fig. 2-b are road chain S2With road chain S1Constitute complete candidate matches relation schematic diagram；
Fig. 2-c are road chain S2A part and road chain S1Constitute candidate matches relation schematic diagram；
Fig. 2-d are road chain S2With road chain S1A part constitute candidate matches relation schematic diagram；
Fig. 2-e are road chain S2A part and road chain S1Constitute candidate matches relation schematic diagram；
Fig. 2-f are road chain S2With S1It constitutes one and exactly matches evaluation pair and a local matching evaluation to schematic diagram；
Fig. 2-g are road chain S2With S1It constitutes one and exactly matches evaluation pair and a local matching evaluation to schematic diagram；
Fig. 2-h are road chain S2With road chain S1A part constitute candidate matches relation schematic diagram；
Fig. 2-i are road chain S2With S1It constitutes one and exactly matches evaluation pair and a local matching evaluation to schematic diagram；
Fig. 2-j are road chain S2With S1It constitutes one and exactly matches evaluation pair and a local matching evaluation to schematic diagram；
Fig. 2-k are road chain S2With S1It constitutes one and exactly matches evaluation pair and two local matching evaluations to schematic diagram；
Fig. 3 is to describe schematic diagram based on the broken line shape for turning to function；
Fig. 4 is to calculate broken line shape similarity distance principle schematic；
Fig. 5 is broken line general direction schematic diagram；
Fig. 6-a are broken line L1And L2Between average distance Computing Principle schematic diagram；
Fig. 6-b are broken line L1On vertex in L2On corresponding points between relation schematic diagram；
Fig. 6-c are broken line L2On vertex in L1On corresponding points between relation schematic diagram.
Specific implementation mode
The specific implementation mode of the present invention is described further below in conjunction with the accompanying drawings.
The present invention is based on the embodiments of the vector link change detection method of similarity measure
The present invention first pre-processes new and old road network data, reconstructed network topological relation, corrects Topology Error, carries
Take Road chain (stroke)；Then the matching candidate of road to be matched is searched for using the buffer way based on consistency constraint
Collection；The similar evaluation model in space is recycled to determine road entity of the same name；It is poor that characteristic index finally is carried out to detection road to be changed
Specific analysis determines road with the presence or absence of variation.
1. data prediction
Quality examination is carried out to data, reconstructs road topology relationship, corrects Topology Error.It is generated according to road chain (stroke)
Principle, extraction Road chain (stroke), the road segmental arc that record road chain (stroke) includes, each road chain (stroke) are real
Border represents a natural road.
2. searching for the matching candidate collection of road to be matched using the buffer way based on consistency constraint.
The search process of the matching candidate collection of one road is described as follows with Fig. 2：Assuming that S1For a road to be matched,
Including node be Pi(i=0,2 ..., n), P0And PnIt is its first and last endpoint, S respectively1The process of matching candidate collection search is as follows：
(1) to S1In each node PiThe buffering area that search radius is R is established, searching for another data, to be centrally located at it slow
The endpoint rushed in area is node PiCandidate matches node, buffering area search radius R values areD1And D2Respectively
For the positional precision of two datasets.
(2) to all node PiCandidate matches node into walking along the street chain (stroke) consistency detection, detect to be located at same
Candidate matches node on chain (stroke) all the way, will be according to composition road chain by the candidate matches node on same chain all the way
Precedence is ranked up, by taking Fig. 2-a as an example, it is assumed that S2It is the road chain obtained through chain consistency check of passing by one's way, Tj、TkWith
TfIt is to be located at road chain S respectively2On candidate matches node in first node, intermediate node and last node, they are in road chain S1Shang pair
The matched nodes answered are respectively Ph、PlAnd Pt。
(3) road chain S is extracted1Candidate matches object.
With the road chain S obtained in step (2)2For, candidate matches object extraction is divided into following several situations：
1. if Tj=T0And Tf=Tm、Ph=P0And Pt=Pn, road chain S2As road chain S1Complete candidate matches object,
I.e. matching evaluation to for<S1-S2>(as shown in Fig. 2-b)；
2. if TjAnd TfAnd only there are one be road chain S2Endpoint, Ph=P0And Pt=Pn, road chain S2A part and road
Chain S1Constitute candidate matches relationship, matching evaluation to for<S1-TjTf>(as shown in fig. 2-c)；
3. if Tj=T0And Tf=Tm, PhAnd PtAnd only there are one be road chain S1Endpoint, road chain S2With road chain S1One
Part constitutes candidate matches relationship, matching relationship pair<PhPt-S2>(as shown in Fig. 2-d)；
4. if TjAnd TfThe roads Dou Bushi chain S2Endpoint, Ph=P0And Pt=Pn, road chain S2A part and road chain S1It constitutes
Candidate matches relationship, matching evaluation to for<S1-TjTf>(as shown in Fig. 2-e)；
5. if TjAnd TfThe roads Dou Bushi chain S2Endpoint, Ph=P0, Pt≠Pn, road chain S2With road chain S1To constitute one it is complete
Full matching evaluation pair<PhPt-TjTf>, a local matching evaluation pair<PtPn–TfTm>(as shown in Fig. 2-f)；
6. if TjAnd TfThe roads Dou Bushi chain S2Endpoint, Ph≠P0, Pt=Pn, road chain S2With road chain S1To constitute one it is complete
Full matching evaluation pair<PhPt-TjTf>, a local matching evaluation pair<P0Ph–T0Tj>(as shown in Fig. 2-g)；
7. if PhAnd PtThe roads Dou Bushi chain S1Endpoint, Tj=T0, Tf=Tm, road chain S2With road chain S1A part constitute
Candidate matches relationship, matching evaluation pair<PhPt–S2>(as shown in Fig. 2-h)；
8. if PhAnd PtThe roads Dou Bushi chain S1Endpoint, Tj=T0, Tf≠Tm, road chain S2With road chain S1To constitute one it is complete
Full matching evaluation pair<PhPt–T0Tf>, a local matching evaluation pair<PtPn–TfTm>(as shown in Fig. 2-i)；
9. if PhAnd PtThe roads Dou Bushi chain S1Endpoint, Tj≠T0, Tf=Tm, road chain S2With road chain S1To constitute one it is complete
Full matching evaluation pair<PhPt–T0Tf>, a local matching evaluation pair<P0Ph–T0Tj>(as shown in Fig. 2-j)；
10. if PhAnd PtThe roads Dou Bushi chain S1Endpoint, TjAnd TfThe roads Ye Doubushi chain S2Endpoint, road chain S2With road chain S1
One will be constituted and exactly match evaluation pair<PhPt–TjTf>, two local matching evaluations pair<P0Ph–T0Tj>With<PtPn–TfTm>(such as
Shown in Fig. 2-k).
(4) it to all road chains obtained after chain consistency check of passing by one's way, is evaluated using above method extraction candidate matches
It is right, by the evaluation of all candidate matches to being put into chained list, complete the search of a path adaptation Candidate Set.
3. concentrating the highest road object of one similarity of selection to make from matching candidate using spatial simlanty evaluation model
For physical road of the same name.
In the present invention, similarity evaluation model mainly considers following geometric properties index：Shape feature, distance feature, length
Spend feature and direction character.
(1) shape feature：Shape is a common important geometric properties in path adaptation.The present invention is using steering function
Method is for describing linear road form, the steering function Θ of line feature shape descriptionL(s) it is expressed as form shown in Fig. 3：X-axis
Indicate that normalized cumulant of each vertex relative to reference point on line, Y-axis indicate each line segment and level in broken line corresponding to line feature
The angle (counterclockwise for just, be negative clockwise) in direction.It can be seen from the figure that ΘL(s) continuous the two of broken line
Value between a vertex is constant, is changed in the value of apex.Function will be turned to match applied to shape, waited for by calculating
The shape similarity degree for matching the shape similarity distance (or being matching distance) between line feature to weigh them.Shape is similar
The calculation formula of distance is：
The usual values of P are 1.
In above formula,It indicates for describing curve L1The function of shape, Dshape(L1,L2) reality expression broken line L1And L2
Turn to the difference of functionThe area (as shown in Figure 4) of area defined is projected to horizontal direction.Dshape
(L1,L2) value is bigger, broken line L1And L2The similarity of shape is with regard to smaller.Formula (2) is line feature shape similarity evaluation function：
In formula, Dshape_toleranceFor the empirical value of shape similarity distance, when the shape similarity distance between line feature
Dshape(L1,L2) be more than the value when, Dshape(L1,L2) value be 0.
(2) cardinal direction marker：Road direction indicates that general direction refers to the company with road first and last node using general direction
Line carrys out approximate description relative to the angle that trunnion axis rotates, such as the general direction that the α in Fig. 5 is the broken line.Two roads to be matched
General direction discrepancy delta θ between the segmental arc of road is between [0, π], when Δ θ is 0, indicates the direction of two segmental arcs along consistent
Direction is parallel, when Δ θ is π, indicates that the direction of two segmental arcs is parallel in opposite direction.Formula (3) is between segmental arc to be matched
General direction difference evaluate their directional similarity,
Δ θ in formulatoleranceIt is segmental arc direction discrepancy threshold.
(3) positioning index：Position feature is used to describe the proximity between element.The case where not considering systematic error
Under, physical road of the same name should be very close on spatial position, by comparing the difference between Space Elements in position
Off course degree come assess they whether be entity of the same name possibility.The present invention utilizes average distance between a kind of approximate calculation broken line
Method describes the proximity between road.
According to broken line L1=＜ L1.1,L1.2,…,L1.n-1,L1.n＞ and L2=＜ L2.1,L2.2,…,L2.n-1,L2.nThe vertex ＞
Information finds out corresponding points of the vertex on another broken line, L on broken line1Vertex in L2On correspondence point set be denoted as L '1=＜
L′1.1,L′1.2,…,L′1.n-1,L′1.n＞ (as shown in Fig. 6-b), similarly broken line L2On vertex in L1On correspondence point set note
For L '2=＜ L '2.1,L′2.2,…,L′2.n-1,L2′.n＞ (as shown in Fig. 6-c), broken line L1With L2Between average distance can be by formula
(4) it is calculated,
Wherein, lk.i,i+1(k=1 or 2) indicates vertex from Lk.iTo Lk.i+1Line segment, | lk.i,i+1| indicate the length of the line segment
Degree, lk.i,i′Indicate vertex from Lk.iTo L 'k.iLine segment.
In path adaptation, the expression formula of proximity is between road object：
Wherein dtoleranceFor distance threshold, value is the maximum value that two broken line mapping nodes are adjusted the distance, davIndicate curve it
Between average distance, when the distance between road object is more than the threshold value, that is, think they do not have it is matched may (SDistance
=0).
(4) length characteristic：Link length is with indicating that the broken line length of road indicates.
D indicates the distance between two road curves, the average distance d between reality curveavIt indicates, l indicates curve
Length, Δ l indicate the length difference between two curves, viIndicate node of a curve, xiAnd yiIndicate node viCoordinate.
In order to evaluate similitude of the road to be matched in length, length similarity evaluation model need to be established：
Wherein Δ ltoleranceIt is the threshold value of road segmental arc difference in length, usually takes the maximum value of two evaluation path length.
Above-mentioned each characteristic index is to reflect some aspect of roadway characteristic, in order to integrate features described above index and build
Comprehensive similarity evaluation model is found, needs to eliminate their differences in dimension, carries out dimensionless normalized processing.Road it
Between spatial simlanty evaluation model be：
Sim=ω1Sims+ω2SimD+ω3SimL+ω4SimA (8)
Wherein SimS、SimD、SimLAnd SimAIt is shape similarity S respectivelyShape, apart from proximity SDistance, length it is similar
Spend SLengthWith direction similarity SOrientationIt is dimensionless normalized treated value, ω1、ω2、ω3And ω4For corresponding index
Weight, and ω1+ω2+ω3+ω4The value of=1 weight is determined by analytic hierarchy process (AHP).
4. pair road carries out feature difference analysis, detects physical road of the same name and changed in which feature.
Physical road of the same name carries out feature difference analysis, link change by change expression model Change=[S, L,
D, A] it is expressed, for expressing situation of change of the road on shape, length, distance and direction.The assignment of S, L, D and A need
Calculate their corresponding shape similarity simS, size similarity simL, apart from proximity simDWith direction similarity simA,
And (sim is compared with threshold value μS、simL、simDAnd simAAll it is that have passed through dimensionless normalized treated value), if certain
The similarity of a feature>μ, then it is assumed that physical road of the same name does not change in this feature, is assigned a value of 0 accordingly, on the contrary
Then think to change, is assigned a value of 1.It is opposite with shortening two kinds there is also extending when changing for link length feature
The case where, extended variation is assigned a value of+1, the variation of shortening is assigned a value of -1.By many experiments, threshold value μ is set as 0.8 and is
Best results.
5. determining the type of link change.
Classification for link change, link change is divided into simple change and complicated two kinds of variation by the present invention, simple to become
Change refers to that road entity one and only one feature changes.However, the variation that road entity occurs in reality is often more than
In one feature, but the combination of multiple changing features, referred to as complicated variation.The specific classification of simple change is shown in Table 1, compound change
The classification of change is shown in Table 2.
Table 1
Table 2
The embodiment of the vector link change detection device based on similarity measurement of the present invention
The detection device of the present invention includes that Road chain generation module, matching candidate collection determining module, spatial simlanty are commented
Valence module and change detection module；Road chain generation module is used to extract the Road chain of road data concentration, and determines road
The road segmental arc for including in Lu Lulian；Matching candidate collection determining module is used to use the buffering area searcher based on consistency constraint
Method determines the matching candidate collection of road to be matched；Spatial simlanty evaluation module is used to establish space according to the geometric properties of road
Similarity evaluation model selects the highest road object of a similarity as reality of the same name using the model from matching candidate concentration
Body road；Compared with change detection module is used to carry out feature difference with its physical road of the same name to detection road to be changed, with
Determine the situation of change of road to be analyzed.The specific implementation means of a module have carried out specifically in the embodiment of method
Bright, which is not described herein again.
By the above process, the present invention can accurately detect physical road of the same name and change in which feature, for
The implementation of extraction and the update operation of change information provides guarantee, has very high application value.
Claims (10)
1. a kind of vector link change detection method based on similarity measurement, which is characterized in that the detection method includes following
Step：
1) topology reconstruction is carried out to data to be tested, extracts Road chain, and determine the road segmental arc for including in the chain of road；
2) the matching candidate collection of Road chain to be matched is searched for using the buffer way based on consistency constraint；
3) spatial simlanty evaluation model is established according to the geometric properties of road, selection one is concentrated from matching candidate using the model
Of the same name physical road of the highest road object of similarity as road to be matched；
4) feature difference comparison is carried out to detection road to be changed and its physical road of the same name, with determination detection road to be changed
Situation of change.
2. the vector link change detection method according to claim 1 based on similarity measurement, which is characterized in that described
The determination of matching candidate collection in step 2), detailed process are as follows：
A. buffering area is established according to search radius to each node in Road chain to be matched, another data is centrally located at each slow
Rush correspondence candidate matches node of the road node in area as each node in corresponding Road chain to be matched；
B. to the candidate matches node of each node into walking along the street chain consistency detection, by all candidates on the chain of same road
It with node as one group, and is ranked up according to the precedence for constituting road chain, by the road where each group of candidate matches node
Chain is added to candidate matches as candidate matches road chain and concentrates；
C. according to the node correspondence of Road chain to be matched and candidate matches road chain, all sections to be evaluated are extracted
With relationship pair, by the section matching relationship of these matching evaluations to be evaluated to being put into chained list, to obtain Road to be matched
The Candidate Set set of matches of chain.
3. the vector link change detection method according to claim 1 based on similarity measurement, which is characterized in that described
The spatial simlanty evaluation model established in step 3) is as follows：
Sim=ω1SimS+ω2SimD+ω3SimL+ω4SimA
Wherein SimS、SimD、SimLAnd SimAIt is shape similarity S respectivelyShape, apart from proximity SDistance, length similarity
SLengthWith direction similarity SOrientationIt is dimensionless normalized treated value, ω1、ω2、ω3And ω4For the power of corresponding index
Weight, and ω1+ω2+ω3+ω4=1.
4. the vector link change detection method according to claim 3 based on similarity measurement, which is characterized in that described
Shape similarity SShapeIt indicates the shape similarity distance between road line feature, is calculated using function is turned to, calculated public
Formula is as follows：
P is 1,
Wherein Dshape(L1,L2) reality expression broken line L1And L2Turn to the difference of functionInstitute is projected to horizontal direction
The area in the region surrounded, Dshape_toleranceFor the empirical value of shape similarity distance, Dshape(L1,L2) value is bigger, broken line
L1And L2The similarity of shape is with regard to smaller.
5. the vector link change detection method according to claim 3 based on similarity measurement, which is characterized in that described
Apart from proximity SDistanceRefer to the proximity between road line feature, the distance between line feature is using approximate broken line
Average distance indicates that calculation formula is as follows：
Wherein, dav(L1,L2) indicate broken line L1And L2Between approximate broken line average distance, lk.i,i+1, k=1 or 2, expression vertex
From Lk.iTo Lk.i+1Line segment, | lk.i,i+1| indicate the length of the line segment, lk.i,i′Indicate vertex from Lk.iTo L 'k.iLine segment,
dtoleranceFor distance threshold, value is the maximum value that two broken line mapping nodes are adjusted the distance.
6. the vector link change detection method according to claim 3 based on similarity measurement, which is characterized in that described
Length similarity SLengthRefer to similitude of the road to be matched in length,
Wherein Δ ltoleranceIt is the threshold value of road segmental arc difference in length.
7. the vector link change detection method according to claim 3 based on similarity measurement, which is characterized in that described
Direction similarity SOrientationRefer to that the general direction difference between road section, general direction refer to road section first and last
The angle that the line of node is rotated relative to trunnion axis,
General direction differences of the wherein Δ θ between the section of two road, Δ θtoleranceFor direction discrepancy threshold.
8. the vector link change detection method according to claim 1 based on similarity measurement, which is characterized in that described
Link change detection method in step 4) is the shape similarity by calculating detection road and its physical road of the same name to be changed
SShape, apart from proximity SDistance, length similarity SLengthWith direction similarity SOrientation, and carried out with corresponding threshold value
Compare, it is unchanged to judge that detection road to be changed has in individual features, if the similarity of some above-mentioned feature is more than threshold value,
Then illustrate that detection road to be changed does not change in this feature, otherwise regard as changing, finally according to road spy
The situation of change of sign determines the type of variation.
9. a kind of vector link change detection device based on similarity measurement, which is characterized in that the detection device includes
Road chain generation module, matching candidate collection determining module, spatial simlanty evaluation module and change detection module；
The Road chain generation module is used to extract the Road chain of road data concentration, and determines in Road chain and include
Road segmental arc；
The matching candidate collection determining module is used to determine road to be matched using the buffering area searching method based on consistency constraint
The matching candidate collection on road；
The spatial simlanty evaluation module is used to establish spatial simlanty evaluation model according to the geometric properties of road, utilizes this
Model selects the highest road object of a similarity as physical road of the same name from matching candidate concentration；
Compared with the change detection module is used to carry out feature difference with its physical road of the same name to detection road to be changed, with
Determine the situation of change of road to be analyzed.
10. the vector link change detection device according to claim 9 based on similarity measurement, which is characterized in that institute
It states matching candidate collection determining module and determines that the process of matching candidate collection is as follows：
A. buffering area is established according to search radius to each node in Road chain to be matched, another data is centrally located at each slow
Rush correspondence candidate matches node of the road node in area as each node in corresponding Road chain to be matched；
B. to the candidate matches node of each node into walking along the street chain consistency detection, by all candidates on the chain of same road
It with node as one group, and is ranked up according to the precedence for constituting road chain, by the road where each group of candidate matches node
Chain is added to candidate matches as candidate matches road chain and concentrates；
C. according to the node correspondence of Road chain to be matched and candidate matches road chain, all sections to be evaluated are extracted
With relationship pair, by the section matching relationship of these matching evaluations to be evaluated to being put into chained list, to obtain Road to be matched
The Candidate Set set of matches of chain.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810083737.2A CN108492276B (en) | 2018-01-29 | 2018-01-29 | Similarity measurement-based vector road change detection method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810083737.2A CN108492276B (en) | 2018-01-29 | 2018-01-29 | Similarity measurement-based vector road change detection method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108492276A true CN108492276A (en) | 2018-09-04 |
| CN108492276B CN108492276B (en) | 2021-03-19 |
Family
ID=63343829
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810083737.2A Active CN108492276B (en) | 2018-01-29 | 2018-01-29 | Similarity measurement-based vector road change detection method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108492276B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109543712A (en) * | 2018-10-16 | 2019-03-29 | 哈尔滨工业大学 | Entity recognition method on temporal dataset |
| CN109949692A (en) * | 2019-03-27 | 2019-06-28 | 腾讯大地通途（北京）科技有限公司 | Road network method, apparatus, computer equipment and storage medium |
| CN110750607A (en) * | 2019-09-17 | 2020-02-04 | 西安工程大学 | Road network matching method based on GNSS vehicle track data |
| CN111291790A (en) * | 2020-01-19 | 2020-06-16 | 华东师范大学 | Turning path extraction and road network topology change detection framework method based on track similarity |
| CN112559660A (en) * | 2020-12-11 | 2021-03-26 | 腾讯科技（深圳）有限公司 | Road data processing method and device, electronic equipment and storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1653505A (en) * | 2002-03-29 | 2005-08-10 | 松下电器产业株式会社 | Map matching method, map matching device, database for shape matching, and shape matching device |
| CN101324440A (en) * | 2008-07-29 | 2008-12-17 | 光庭导航数据（武汉）有限公司 | Map-matching method based on forecast ideology |
| US20130238648A1 (en) * | 2011-11-14 | 2013-09-12 | Aisin Aw Co., Ltd. | Road data creating device, road data creating method, and program |
| CN104361142A (en) * | 2014-12-12 | 2015-02-18 | 华北水利水电大学 | Detection method for rapid change in multi-source navigation electronic map vector road network |
| CN105825510A (en) * | 2016-03-17 | 2016-08-03 | 中南大学 | Automatic matching method between point of interest and road network |
| CN105956542A (en) * | 2016-04-28 | 2016-09-21 | 武汉大学 | Structure wiring harness counting and matching high-resolution remote-sensing image road extraction method |
| US20170314934A1 (en) * | 2016-04-28 | 2017-11-02 | Here Global B.V. | Map matching quality evaluation |
-
2018
- 2018-01-29 CN CN201810083737.2A patent/CN108492276B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1653505A (en) * | 2002-03-29 | 2005-08-10 | 松下电器产业株式会社 | Map matching method, map matching device, database for shape matching, and shape matching device |
| CN101324440A (en) * | 2008-07-29 | 2008-12-17 | 光庭导航数据（武汉）有限公司 | Map-matching method based on forecast ideology |
| US20130238648A1 (en) * | 2011-11-14 | 2013-09-12 | Aisin Aw Co., Ltd. | Road data creating device, road data creating method, and program |
| CN104361142A (en) * | 2014-12-12 | 2015-02-18 | 华北水利水电大学 | Detection method for rapid change in multi-source navigation electronic map vector road network |
| CN105825510A (en) * | 2016-03-17 | 2016-08-03 | 中南大学 | Automatic matching method between point of interest and road network |
| CN105956542A (en) * | 2016-04-28 | 2016-09-21 | 武汉大学 | Structure wiring harness counting and matching high-resolution remote-sensing image road extraction method |
| US20170314934A1 (en) * | 2016-04-28 | 2017-11-02 | Here Global B.V. | Map matching quality evaluation |
Non-Patent Citations (1)
| Title |
|---|
| 翟仁健: "基于全局一致性评价的多尺度矢量空间数据匹配方法研究", 《中国博士学位论文全文数据库》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109543712A (en) * | 2018-10-16 | 2019-03-29 | 哈尔滨工业大学 | Entity recognition method on temporal dataset |
| CN109949692A (en) * | 2019-03-27 | 2019-06-28 | 腾讯大地通途（北京）科技有限公司 | Road network method, apparatus, computer equipment and storage medium |
| CN110750607A (en) * | 2019-09-17 | 2020-02-04 | 西安工程大学 | Road network matching method based on GNSS vehicle track data |
| CN111291790A (en) * | 2020-01-19 | 2020-06-16 | 华东师范大学 | Turning path extraction and road network topology change detection framework method based on track similarity |
| CN111291790B (en) * | 2020-01-19 | 2021-03-26 | 华东师范大学 | Turning path extraction and road network topology change detection framework method based on track similarity |
| CN112559660A (en) * | 2020-12-11 | 2021-03-26 | 腾讯科技（深圳）有限公司 | Road data processing method and device, electronic equipment and storage medium |
| CN112559660B (en) * | 2020-12-11 | 2022-06-17 | 腾讯科技（深圳）有限公司 | Road data processing method and device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108492276B (en) | 2021-03-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108492276A (en) | A kind of vector link change detection method and device based on similarity measurement | |
| Yang et al. | MGraphDTA: deep multiscale graph neural network for explainable drug–target binding affinity prediction | |
| Gálvez-López et al. | Bags of binary words for fast place recognition in image sequences | |
| CN106485740B (en) | A kind of multidate SAR image registration method of combination stable point and characteristic point | |
| CN109145173B (en) | Similarity-based vector element change comparison method | |
| CN108492298B (en) | Multispectral image change detection method based on generation countermeasure network | |
| CN108801171A (en) | A kind of tunnel cross-section deformation analytical method and device | |
| Hönigschmid et al. | Accurate prediction of helix interactions and residue contacts in membrane proteins | |
| CN103854290A (en) | Extended target tracking method based on combination of skeleton characteristic points and distribution field descriptors | |
| Gong et al. | A two-level framework for place recognition with 3D LiDAR based on spatial relation graph | |
| Buhmann et al. | Synaptic partner prediction from point annotations in insect brains | |
| He et al. | Online semantic-assisted topological map building with LiDAR in large-scale outdoor environments: Toward robust place recognition | |
| Seybold | Robust map matching for heterogeneous data via dominance decompositions | |
| Ballardini et al. | MassUntangler: A novel alignment tool for label-free liquid chromatography–mass spectrometry proteomic data | |
| Bai et al. | Fast exact fingerprint indexing based on compact binary minutia cylinder codes | |
| Tian et al. | Discriminative and semantic feature selection for place recognition towards dynamic environments | |
| Rehrl et al. | Optimization and evaluation of a high-performance open-source map-matching implementation | |
| Huh et al. | Line segment confidence region-based string matching method for map conflation | |
| Shi et al. | Structured deep learning based object-specific distance estimation from a monocular image | |
| Lin et al. | A probabilistic embedding clustering method for urban structure detection | |
| Wang et al. | Hierarchical stroke mesh: A new progressive matching method for detecting multi-scale road network changes using OpenStreetMap | |
| Li et al. | Wimage: Crowd sensing based heterogeneous information fusion for indoor localization | |
| Li et al. | Contrastive meta-learning for drug-target binding affinity prediction | |
| Satpute et al. | Decision tree classifier for classification of proteins using the Protein Data Bank | |
| CN114679683A (en) | Indoor intelligent positioning method based on derivative fingerprint migration |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |