CN108492276B - Similarity measurement-based vector road change detection method and device - Google Patents

Similarity measurement-based vector road change detection method and device Download PDF

Info

Publication number
CN108492276B
CN108492276B CN201810083737.2A CN201810083737A CN108492276B CN 108492276 B CN108492276 B CN 108492276B CN 201810083737 A CN201810083737 A CN 201810083737A CN 108492276 B CN108492276 B CN 108492276B
Authority
CN
China
Prior art keywords
road
similarity
matching
link
sim
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.)
Active
Application number
CN201810083737.2A
Other languages
Chinese (zh)
Other versions
CN108492276A (en
Inventor
翟仁健
武芳
王昊
闫浩文
朱丽
巩现勇
行瑞星
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Information Engineering University of PLA Strategic Support Force
Original Assignee
Information Engineering University of PLA Strategic Support Force
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Information Engineering University of PLA Strategic Support Force filed Critical Information Engineering University of PLA Strategic Support Force
Priority to CN201810083737.2A priority Critical patent/CN108492276B/en
Publication of CN108492276A publication Critical patent/CN108492276A/en
Application granted granted Critical
Publication of CN108492276B publication Critical patent/CN108492276B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

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

Abstract

The invention relates to a method and a device for detecting vector road change based on similarity measurement, and belongs to the technical field of dynamic update of a vector map database. Firstly, reconstructing a topological relation of a road data set to be detected, extracting a road link, and determining a road radian contained in the link; then searching a matching candidate set of the road to be matched by adopting a buffer area method based on consistency constraint; then, establishing a similarity evaluation model according to the geometric characteristics of the road, and selecting a road object with the highest similarity from the matching candidate set by using the evaluation model as a matching object of the road to be matched; and finally, comparing the feature difference of the entity road with the same name and the road to be matched to determine the change condition of the road to be matched. The method accurately detects the change of the same-name road entity on which characteristics are changed by calculating the characteristic difference performance between roads, provides guarantee for the extraction of change information and the implementation of updating operation, and has high application value.

Description

Similarity measurement-based vector road change detection method and device
Technical Field
The invention relates to a method and a device for detecting vector road change based on similarity measurement, and belongs to the technical field of dynamic update of a vector map database.
Background
The road elements are main elements in the topographic map and are the most prominent elements, and in order to ensure the availability of the road data, the road data needs to be updated in real time. In the incremental updating of the road network, what changes happen to the road entities is a key problem in the updating of the road network how to detect and describe the changes, and the key problem directly influences the efficiency and level of storage organization, incremental information collection, updating processing, and analysis and distribution of the change information of the change entities.
Currently, researchers have conducted relevant research on the detection and expression of change information. The method comprises the steps that an incremental information classification and expression based on a geographic event and a target snapshot difference are provided by Zhuhuaji, Chenjun and the like, and a definition and expression model of change information based on the event and the snapshot difference is provided, but the method only classifies the change information from a single layer, neglects the diversity of the change information and does not consider the change condition of road network data under the complex condition; although it has been proposed that the difference in graphic data is used for detecting a change in a residential area element by calculating the difference in graphic data and determining the type of change, simple and complicated types of changes are considered, but the difference in graphic data is disadvantageous in application to a linear body such as a road.
Disclosure of Invention
The invention aims to provide a vector road change detection method based on similarity measurement, which aims to solve the problems of low accuracy and poor applicability of road change detection; the invention also provides a road change detection device based on the similarity measurement.
The invention provides a vector road change detection method based on similarity measurement for solving the technical problems, which comprises eight schemes, wherein the first scheme is as follows: the detection method comprises the following steps:
1) carrying out topology reconstruction on data to be detected, extracting a road link, and determining a road arc segment contained in the link;
2) searching a matching candidate set of a road chain to be matched by adopting a buffer area method based on consistency constraint;
3) establishing a spatial similarity evaluation model according to the geometric characteristics of roads, and selecting a road object with the highest similarity from the matching candidate set by using the model as a homonymous entity road of the road to be matched;
4) and comparing the characteristic difference of the road to be changed and the road with the same name entity road to determine the change condition of the road to be changed.
The second method comprises the following steps: on the basis of the first method scheme, the determination of the matching candidate set in the step 2) specifically comprises the following steps:
A. establishing a buffer area for each node in the road link to be matched according to the search radius, and taking the road nodes in the other data set in each buffer area as corresponding candidate matching nodes corresponding to each node in the road link to be matched;
B. performing path link consistency detection on candidate matching nodes of each node, taking all candidate matching nodes on the same path link as a group, sequencing according to the sequence of forming the path link, and adding the path link where each group of candidate matching nodes is located into a candidate matching set as a candidate matching path link;
C. and extracting all road section matching relation pairs to be evaluated according to the node corresponding relation between the road link to be matched and the candidate matching link, and putting the road section matching relation pairs to be evaluated into a linked list to obtain a candidate set matching set of the road link to be matched.
The third method scheme is as follows: on the basis of the first method scheme, the spatial similarity evaluation model established in the step 3) is as follows:
Sim＝ω1SimS+ω2SimD+ω3SimL+ω4SimA
wherein SimS、SimD、SimLAnd SimARespectively is the shape similarity SShapeDistance proximity SDistanceLength similarity SLengthSimilarity with direction SOrientationDimensionless normalized value, omega1、ω2、ω3And ω4Is the weight of the corresponding index, and ω1+ω2+ω3+ω4＝1。
The method scheme is as follows: on the basis of the third method proposal,the shape similarity SShapeThe distance of similarity of the shapes of the road line elements is represented and calculated by adopting a steering function, and the calculation formula is as follows:
Figure BDA0001561781210000031
Figure BDA0001561781210000032
Figure BDA0001561781210000033
the number P is 1, and the content of the compound,
wherein Dshape(L1,L2) Actually representing a polyline L1And L2Difference of steering function
Figure BDA0001561781210000034
Area of region enclosed by projection in horizontal direction, Dshape_toleranceEmpirical threshold for shape similarity distance, Dshape(L1,L2) The larger the value, the broken line L1And L2The smaller the similarity of the shapes.
The method scheme five: on the basis of the third method scheme, the distance proximity SDistanceRefers to the proximity between road line elements, and the distance between line elements is expressed by approximate polyline average distance, and the calculation formula is as follows:
Figure BDA0001561781210000035
Figure BDA0001561781210000036
wherein d isav(L1,L2) Represents a broken line L1And L2Approximate polyline average distance between,/k.i,i+1，k＝1 or 2, indicating that the vertex is from Lk.iTo Lk.i+1L ofk.i,i+1L represents the length of the line segment, lk.i,i′Representing vertices from Lk.iTo L'k.iLine segment of dtoleranceAnd taking the distance as a distance threshold value, wherein the value is the maximum value of the distance of the two broken line mapping nodes.
The method comprises the following steps: on the basis of the third method scheme, the length similarity SLengthRefers to the similarity in length of the roads to be matched,
Figure BDA0001561781210000037
Figure BDA0001561781210000038
wherein Δ ltoleranceIs the threshold value of the road arc length difference.
The method comprises the following steps: on the basis of the third method scheme, the direction similarity SOrientationRefers to the integral direction difference between road sections, the integral direction refers to the angle of the rotation of the connecting line of the first node and the last node of the road section relative to the horizontal axis,
Figure BDA0001561781210000041
where Δ θ is the overall directional difference between two road segments, Δ θtoleranceIs a direction difference threshold.
The method comprises the following steps: on the basis of the first method scheme, the road change detection method in the step 4) calculates the shape similarity S of the road to be changed and the road with the same name as the road to be changedShapeDistance proximity SDistanceLength similarity SLengthSimilarity with direction SOrientationAnd comparing the detected road to corresponding threshold value to judge whether the detected road to be changed has change on corresponding characteristic, if the similarity of some characteristic is greater than the threshold value, it indicates that the road to be changed has changeAnd finally, determining the type of the change according to the change condition of the road characteristics.
The invention also provides a vector road change detection device based on similarity measurement, which comprises the following four schemes: the detection device comprises a road link generation module, a matching candidate set determination module, a spatial similarity evaluation module and a change detection module;
the road link generation module is used for extracting road links in the road data set and determining road arc sections contained in the road links;
the matching candidate set determining module is used for determining a matching candidate set of a road to be matched by adopting a buffer area searching method based on consistency constraint;
the spatial similarity evaluation module is used for establishing a spatial similarity evaluation model according to the geometric characteristics of the roads, and selecting a road object with the highest similarity from the matching candidate set as a same-name entity road by using the model;
the change detection module is used for comparing the feature difference of the road to be detected and the road with the same name entity to determine the change condition of the road to be analyzed.
The device scheme II comprises the following steps: on the basis of the first device scheme, the process of determining the matching candidate set by the matching candidate set determining module is as follows:
A. establishing a buffer area for each node in the road link to be matched according to the search radius, and taking the road nodes in the other data set in each buffer area as corresponding candidate matching nodes corresponding to each node in the road link to be matched;
B. performing path link consistency detection on candidate matching nodes of each node, taking all candidate matching nodes on the same path link as a group, sequencing according to the sequence of forming the path link, and adding the path link where each group of candidate matching nodes is located into a candidate matching set as a candidate matching path link;
C. and extracting all road section matching relation pairs to be evaluated according to the node corresponding relation between the road link to be matched and the candidate matching link, and putting the road section matching relation pairs to be evaluated into a linked list to obtain a candidate set matching set of the road link to be matched.
The device scheme is as follows: on the basis of the first device scheme, a spatial similarity evaluation model established by the road similarity evaluation module is as follows:
Sim＝ω1SimS+ω2SimD+ω3SimL+ω4SimA
wherein SimS、SimD、SimLAnd SimARespectively is the shape similarity SShapeDistance proximity SDistanceLength similarity SLengthSimilarity with direction SOrientationDimensionless normalized value, omega1、ω2、ω3And ω4Is the weight of the corresponding index, and ω1+ω2+ω3+ω4＝1。
The device scheme is four: on the basis of the first device scheme, the change detection module calculates the shape similarity S of the road to be changed and the road with the same name as the road to be changedShapeDistance proximity SDistanceLength similarity SLengthSimilarity with direction SOrientationAnd comparing the feature with a corresponding threshold value, judging whether the road to be detected changes on the corresponding feature, if the similarity of the certain feature is greater than the threshold value, indicating that the road to be detected changes on the feature, otherwise, determining that the road to be detected changes, and finally determining the type of the change according to the change condition of the road feature.
The invention has the advantages that firstly, the road link is extracted, and the road radian included in the link is determined; then searching a matching candidate set of the road to be matched by adopting a buffer area method based on consistency constraint; then, establishing a similarity evaluation model according to the geometric characteristics of the road, and selecting a road object with the highest similarity from the matching candidate set by using the similarity evaluation model as a homonymous entity road of the road to be matched; and finally, comparing the characteristic difference of the road to be detected and the road with the same name entity to determine the change condition of the road to be detected. The method accurately detects the change of the road on which characteristics are changed by calculating the similarity between the roads, provides guarantee for the extraction of the change information and the implementation of the updating operation, and has high application value.
Drawings
FIG. 1 is a flow chart of road change detection identification;
FIG. 2-a Link S1And candidate matching link S2Schematic diagram of matching relationship possibly existing between the two;
FIG. 2-b is a link chain S2And link S1Forming a complete candidate matching relation schematic diagram;
FIG. 2-c is a link S2A part of (2) and a link chain S1Forming a candidate matching relation schematic diagram;
FIG. 2-d is a link chain S2And link S1A part of the candidate matching relationship diagram is formed;
FIG. 2-e is a link chain S2A part of (2) and a link chain S1Forming a candidate matching relation schematic diagram;
FIG. 2-f is a link S2And S1Forming a complete matching evaluation pair and a local matching evaluation pair schematic diagram;
FIG. 2-g is a link chain S2And S1Forming a complete matching evaluation pair and a local matching evaluation pair schematic diagram;
FIG. 2-h is a link S2And link S1A part of the candidate matching relationship diagram is formed;
FIG. 2-i is a link chain S2And S1Forming a complete matching evaluation pair and a local matching evaluation pair schematic diagram;
FIG. 2-j is a link S2And S1Forming a complete matching evaluation pair and a local matching evaluation pair schematic diagram;
FIG. 2-k is a link S2And S1Forming a complete matching evaluation pair and two partial matching evaluation pair schematic diagrams;
FIG. 3 is a schematic diagram depicting a dogleg shape based on a steering function;
FIG. 4 is a schematic diagram illustrating the principle of calculating the similarity distance between broken line shapes;
FIG. 5 is a schematic view of the overall direction of a broken line;
FIG. 6-a is a broken line L1And L2Schematic diagram of the calculation principle of the average distance between the two parts;
FIG. 6-b is a broken line L1Has a vertex at L2A schematic diagram of the relationship between corresponding points on the graph;
FIG. 6-c is a broken line L2Has a vertex at L1The relationship between the corresponding points on (a) and (b) is shown schematically.
Detailed Description
The following further describes embodiments of the present invention with reference to the drawings.
Embodiment of the vector road change detection method based on similarity measurement
The method comprises the steps of preprocessing new and old road network data, reconstructing a network topological relation, correcting topological errors and extracting road links (strokes); then searching a matching candidate set of the road to be matched by adopting a buffer area method based on consistency constraint; determining the entities of the roads with the same name by using a space similarity evaluation model; and finally, performing characteristic index difference analysis on the road to be changed and detected to determine whether the road has change.
1. Data pre-processing
And (4) carrying out quality inspection on the data, reconstructing a road topological relation and correcting a topological error. According to a link (stroke) generating principle, extracting a road link (stroke), recording road arc sections contained in the link (stroke), and actually representing a natural road by each link (stroke).
2. And searching a matching candidate set of the road to be matched by using a buffer area method based on consistency constraint.
The search process for a matching candidate set for a road is illustrated in fig. 2 as follows: suppose S1Is a road to be matched and comprises a node Pi(i＝0,2,…,n)，P0And PnRespectively, its head and end points, S1For searching matching candidate setsThe process is as follows:
(1) to S1In each node PiEstablishing a buffer area with a search radius R, and searching an end point in the buffer area of another data set, namely a node PiThe buffer search radius R is taken as
Figure BDA0001561781210000071
D1And D2The position accuracy of the two data sets respectively.
(2) For all nodes PiThe candidate matching nodes are subjected to link (stroke) consistency detection, candidate matching nodes on the same link (stroke) are detected, the candidate matching nodes on the same link are sorted according to the sequence of the links, fig. 2-a is taken as an example, and S is assumed to be2Is a road chain, T, obtained by road chain consistency testj、TkAnd TfAre respectively located in the link chain S2The head node, the intermediate nodes and the end node in the link chain S1The upper corresponding matched nodes are respectively Ph、PlAnd Pt。
(3) Extraction link S1The candidate matching object of (1).
Using the link S obtained in step (2)2For example, the candidate matching object extraction is divided into the following cases:
if Tj＝T0And Tf＝Tm、Ph＝P0And P ist＝PnRoad link S2As a link S1Is a matching evaluation pair of<S1-S2>(as shown in FIG. 2-b);
if TjAnd TfOne and only one is a link S2End point of, Ph＝P0And P ist＝PnRoad link S2A part of (2) and a link chain S1Forming a candidate matching relation, wherein the matching evaluation pair is<S1-TjTf>(as shown in FIG. 2-c);
(iii) if Tj＝T0And Tf＝Tm，PhAnd PtOne and only one is a link S1End point of, link S2And link S1Form a candidate matching relationship, a pair of matching relationships<PhPt-S2>(as shown in FIG. 2-d);
if TjAnd TfAre not links S2End point of, Ph＝P0And P ist＝PnRoad link S2A part of (2) and a link chain S1Forming a candidate matching relation, wherein the matching evaluation pair is<S1-TjTf>(as shown in fig. 2-e);
if TjAnd TfAre not links S2End point of, Ph＝P0，Pt≠PnRoad link S2And link S1Will form a perfect match evaluation pair<PhPt-TjTf>A local match evaluation pair<PtPn–TfTm>(as shown in FIG. 2-f);
if TjAnd TfAre not links S2End point of, Ph≠P0，Pt＝PnRoad link S2And link S1Will form a perfect match evaluation pair<PhPt-TjTf>A local match evaluation pair<P0Ph–T0Tj>(as shown in FIG. 2-g);
if PhAnd PtAre not links S1End point of, Tj＝T0，Tf＝TmRoad link S2And link S1Form a candidate matching relationship, match evaluation pair<PhPt–S2>(as shown in FIG. 2-h);
if PhAnd PtAre not links S1End point of, Tj＝T0，Tf≠TmRoad chainS2And link S1Will form a perfect match evaluation pair<PhPt–T0Tf>A local match evaluation pair<PtPn–TfTm>(as shown in FIG. 2-i);
ninthly if PhAnd PtAre not links S1End point of, Tj≠T0，Tf＝TmRoad link S2And link S1Will form a perfect match evaluation pair<PhPt–T0Tf>A local match evaluation pair<P0Ph–T0Tj>(as shown in FIG. 2-j);
if P inhAnd PtAre not links S1End point of, TjAnd TfNor is it a link S2End point of, link S2And link S1Will form a perfect match evaluation pair<PhPt–TjTf>Two partial match evaluation pairs<P0Ph–T0Tj>And<PtPn–TfTm>(as shown in fig. 2-k).
(4) And extracting candidate matching evaluation pairs for all the road links obtained after the road link consistency check by using the method, and putting all the candidate matching evaluation pairs into a linked list to complete the search of a road matching candidate set.
3. And selecting a road object with the highest similarity from the matching candidate set by using a spatial similarity evaluation model as the same-name entity road.
In the invention, the similarity evaluation model mainly considers the following geometric characteristic indexes: a shape feature, a distance feature, a length feature, and an orientation feature.
(1) Shape characteristics: shape is a common important geometric feature in road matching. The invention adopts a steering function method for describing the form of a linear road, and a steering function theta described by the form of a linear elementL(s) is represented in the form shown in FIG. 3: each vertex on the X-axis line corresponds toThe normalized distance of the reference point, Y-axis represents the angle between each line segment in the broken line corresponding to the line element and the horizontal direction (the counterclockwise direction is positive, the clockwise direction is negative). As can be seen from the figure, ΘL(s) the value between two consecutive vertices of the polyline is constant, the value at the vertex changes. The steering function is applied to shape matching, and the similarity of the shapes of elements to be matched is measured by calculating the similarity distance (or called matching distance) between the shapes. The calculation formula of the shape similarity distance is as follows:
Figure BDA0001561781210000091
Figure BDA0001561781210000092
p typically takes the value of 1.
In the above formula, the first and second carbon atoms are,
Figure BDA0001561781210000093
representation is used to describe the curve L1Function of shape, Dshape(L1,L2) Actually representing a polyline L1And L2Difference of steering function
Figure BDA0001561781210000101
The area of the enclosed region is projected in the horizontal direction (as shown in fig. 4). Dshape(L1,L2) The larger the value, the broken line L1And L2The smaller the similarity of the shapes. Equation (2) is a line element shape similarity evaluation function:
Figure BDA0001561781210000102
in the formula, Dshape_toleranceThe empirical threshold for the shape similarity distance, the shape similarity distance D between line elementsshape(L1,L2) Above this value, Dshape(L1,L2) The value is 0.
(2) The direction index is as follows: the road direction is expressed by the overall direction, which is approximately described by the angle of rotation of the connecting line of the head node and the tail node of the road relative to the horizontal axis, and alpha in fig. 5 is the overall direction of the connecting line. The integral direction difference delta theta between the two road arc sections to be matched is between [0 and pi ], when the delta theta is 0, the directions of the two arc sections are parallel along the consistent direction, and when the delta theta is pi, the directions of the two arc sections are parallel along the opposite direction. Equation (3) evaluates their directional similarity by the overall directional difference between the arc segments to be matched,
Figure BDA0001561781210000103
in the formula,. DELTA.theta.toleranceIs the arc segment direction difference threshold.
(3) Position index: the positional features are used to describe the proximity between the elements. Roads of the same-name entity should be very close in spatial position without considering systematic errors, and the possibility of whether they are the same-name entity is evaluated by comparing the degree of difference in position between spatial elements. The invention describes the proximity degree between roads by using a method of approximately calculating the average distance between folding lines.
According to a fold line L1＝＜L1.1,L1.2,…,L1.n-1,L1.n> and L2＝＜L2.1,L2.2,…,L2.n-1,L2.nFinding out the corresponding point of the vertex on another broken line L on the broken line1Has a vertex at L2Is recorded as L 'for the corresponding point set'1＝＜L′1.1,L′1.2,…,L′1.n-1,L′1.n> (as shown in FIG. 6-b), similarly to the broken line L2Has a vertex at L1Is recorded as L 'for the corresponding point set'2＝＜L′2.1,L′2.2,…,L′2.n-1,L2′.n> (as shown in FIG. 6-c), fold line L1And L2The average distance therebetween can be calculated by equation (4),
Figure BDA0001561781210000111
wherein lk.i,i+1(k is 1 or 2) denotes the vertex from Lk.iTo Lk.i+1L ofk.i,i+1L represents the length of the line segment, lk.i,i′Representing vertices from Lk.iTo L'k.iThe line segment of (2).
In road matching, the expression formula of the proximity between road objects is as follows:
Figure BDA0001561781210000112
wherein d istoleranceIs a distance threshold value, and takes the value of the maximum value of the distance of the two-fold line mapping node pair, davRepresenting the average distance between curves, and when the distance between road objects is greater than the threshold, it is assumed that they do not have the possibility of matching (S)Distance＝0)。
(4) Length characteristics: the road length is represented by a length of a broken line representing the road.
Figure BDA0001561781210000113
D represents the distance between two road curves, and the average distance D between the actual curvesavWhere l denotes the length of the curve, Δ l denotes the difference in length between the two curves, viNode, x, representing a curveiAnd yiRepresenting a node viThe coordinates of (a).
In order to evaluate the similarity of the road to be matched in length, a length similarity evaluation model is required to be established:
Figure BDA0001561781210000114
wherein Δ ltoleranceIs a threshold value of the difference of the lengths of the road arc sections, and two scores are usually takenMaximum value of the road length.
Each characteristic index only reflects one aspect of road characteristics, and in order to integrate the characteristic indexes and establish a comprehensive similarity evaluation model, the dimensional difference of the characteristic indexes needs to be eliminated, and dimensionless normalization processing is carried out. The evaluation model of the spatial similarity between roads is as follows:
Sim＝ω1Sims+ω2SimD+ω3SimL+ω4SimA (8)
wherein SimS、SimD、SimLAnd SimARespectively is the shape similarity SShapeDistance proximity SDistanceLength similarity SLengthSimilarity with direction SOrientationDimensionless normalized value, omega1、ω2、ω3And ω4Is the weight of the corresponding index, and ω1+ω2+ω3+ω4The value of the weight 1 is determined by an analytic hierarchy process.
4. And (4) carrying out feature difference analysis on the road, and detecting the change of the same-name entity road on which features.
Carrying out characteristic difference analysis on the same-name entity roads, wherein the road Change is expressed by a Change expression model of [ S, L, D, A]The expression is made for expressing the change of the road in the shape, length, distance and direction. The assignment of S, L, D and A requires calculation of their corresponding shape similarity simSSize similarity simLDistance proximity simDSimilarity with direction simAAnd compared to a threshold μ (sim)S、simL、simDAnd simAAll values after dimensionless normalization), if the similarity of a certain feature is similar>And mu, considering that the same-name entity road has no change on the characteristics, and correspondingly assigning a value of 0, otherwise, considering that the same-name entity road has a change and assigning a value of 1. When the road length characteristics change, the opposite conditions of extension and shortening exist, the extended change is assigned to +1, and the shortened change is assigned to-1. Through a large number of experiments, it is proved that,the threshold μ is set to 0.8, which is the most effective.
5. The type of road change is determined.
For the classification of road changes, the invention divides the road changes into simple changes and complex changes, wherein the simple changes mean that the road entity has one and only one characteristic to change. However, in reality, the change of the road entity is often not on more than one feature, but a combination of a plurality of feature changes, which is called a complex change. The specific classification of simple changes is shown in Table 1, and the classification of complex changes is shown in Table 2.
TABLE 1
Figure BDA0001561781210000121
Figure BDA0001561781210000131
TABLE 2
Figure BDA0001561781210000132
Embodiments of the similarity metric-based vector road change detection apparatus of the present invention
The detection device comprises a road link generation module, a matching candidate set determination module, a spatial similarity evaluation module and a change detection module; the road link generation module is used for extracting the road links in the road data set and determining road arc sections contained in the road links; the matching candidate set determining module is used for determining a matching candidate set of the road to be matched by adopting a buffer area searching method based on consistency constraint; the spatial similarity evaluation module is used for establishing a spatial similarity evaluation model according to the geometric characteristics of the roads, and selecting a road object with the highest similarity from the matching candidate set as a same-name entity road by using the model; the change detection module is used for comparing the feature difference of the road to be detected and the road with the same name entity to determine the change condition of the road to be analyzed. The specific implementation means of each module has been described in detail in the embodiment of the method, and is not described herein again.
Through the process, the method can accurately detect the characteristics of the same-name entity road, provides guarantee for the extraction of the change information and the implementation of the updating operation, and has high application value.

Claims (10)

1. A vector road change detection method based on similarity measurement is characterized by comprising the following steps:
1) carrying out topology reconstruction on data to be detected, extracting a road link, and determining a road arc segment contained in the link;
2) searching a matching candidate set of a road chain to be matched by adopting a buffer area method based on consistency constraint;
3) establishing a spatial similarity evaluation model according to the geometric characteristics of roads, and selecting a road object with the highest similarity from the matching candidate set by using the model as a homonymous entity road of the road to be matched;
4) comparing the characteristic difference of the road to be changed and the same-name entity road to determine the change condition of the road to be changed;
the determination of the matching candidate set in step 2) specifically comprises the following steps:
A. establishing a buffer area for each node in the road link to be matched according to the search radius, and taking the road nodes in the other data set in each buffer area as corresponding candidate matching nodes corresponding to each node in the road link to be matched;
B. performing path link consistency detection on candidate matching nodes of each node, taking all candidate matching nodes on the same path link as a group, sequencing according to the sequence of forming the path link, and adding the path link where each group of candidate matching nodes is located into a candidate matching set as a candidate matching path link;
C. and extracting all road section matching relation pairs to be evaluated according to the node corresponding relation between the road link to be matched and the candidate matching link, and putting the road section matching relation pairs to be evaluated into a linked list to obtain a candidate set matching set of the road link to be matched.
2. The method for detecting vector road change based on similarity measurement according to claim 1, wherein the spatial similarity evaluation model established in step 3) is as follows:
Sim＝ω1SimS+ω2SimD+ω3SimL+ω4SimA
wherein SimS、SimD、SimLAnd SimARespectively is the shape similarity SShapeDistance proximity SDistanceLength similarity SLengthSimilarity with direction SOrientationDimensionless normalized value, omega1、ω2、ω3And ω4Is the weight of the corresponding index, and ω1+ω2+ω3+ω4＝1。
3. The method according to claim 2, wherein the shape similarity S is a similarity measureShapeThe distance of similarity of the shapes of the road line elements is represented and calculated by adopting a steering function, and the calculation formula is as follows:
Figure FDA0002626286690000021
Figure FDA0002626286690000022
Figure FDA0002626286690000023
the number P is 1, and the content of the compound,
wherein Dshape(L1,L2) Actually representing a polyline L1And L2Difference of steering functionValue of
Figure FDA0002626286690000024
Area of region enclosed by projection in horizontal direction, Dshape_toleranceEmpirical threshold for shape similarity distance, Dshape(L1,L2) The larger the value, the broken line L1And L2The smaller the similarity of the shapes.
4. The method according to claim 2, wherein the distance proximity S is a distance between the two neighboring road segmentsDistanceRefers to the proximity between road line elements, and the distance between line elements is expressed by approximate polyline average distance, and the calculation formula is as follows:
Figure FDA0002626286690000025
Figure FDA0002626286690000026
wherein d isav(L1,L2) Represents a broken line L1And L2Approximate polyline average distance between,/k.i,i+1K is 1 or 2, and represents the vertex from Lk.iTo Lk.i+1L ofk.i,i+1L represents the length of the line segment, lk.i,i′Representing vertices from Lk.iTo L'k.iLine segment of dtoleranceAnd taking the distance as a distance threshold value, wherein the value is the maximum value of the distance of the two broken line mapping nodes.
5. The method according to claim 2, wherein the length similarity S is a measure of similarity between the road segmentsLengthRefers to the similarity in length of the roads to be matched,
Figure FDA0002626286690000031
Figure FDA0002626286690000032
wherein Δ ltoleranceIs the threshold value of the road arc length difference.
6. The method according to claim 2, wherein the direction similarity S is a vector road change detection method based on similarity measureOrientationRefers to the integral direction difference between road sections, the integral direction refers to the angle of the rotation of the connecting line of the first node and the last node of the road section relative to the horizontal axis,
Figure FDA0002626286690000033
where Δ θ is the overall directional difference between two road segments, Δ θtoleranceIs a direction difference threshold.
7. The method for detecting vector road change based on similarity measurement according to claim 1, wherein the method for detecting road change in step 4) is to calculate the similarity of shape S between the road to be detected and the road with the same name as the road to be detectedShapeDistance proximity SDistanceLength similarity SLengthSimilarity with direction SOrientationAnd comparing the feature with a corresponding threshold value, judging whether the road to be detected changes on the corresponding feature, if the similarity of the certain feature is greater than the threshold value, indicating that the road to be detected changes on the feature, otherwise, determining that the road to be detected changes, and finally determining the type of the change according to the change condition of the road feature.
8. A vector road change detection device based on similarity measurement is characterized by comprising a road link generation module, a matching candidate set determination module, a spatial similarity evaluation module and a change detection module;
the road link generation module is used for extracting road links in the road data set and determining road arc sections contained in the road links;
the matching candidate set determining module is used for determining a matching candidate set of a road to be matched by adopting a buffer area searching method based on consistency constraint;
the spatial similarity evaluation module is used for establishing a spatial similarity evaluation model according to the geometric characteristics of the roads, and selecting a road object with the highest similarity from the matching candidate set as a same-name entity road by using the model;
the change detection module is used for comparing the feature difference of the road to be detected and the road with the same name entity to determine the change condition of the road to be analyzed;
the process of the matching candidate set determination module determining the matching candidate set is as follows:
A. establishing a buffer area for each node in the road link to be matched according to the search radius, and taking the road nodes in the other data set in each buffer area as corresponding candidate matching nodes corresponding to each node in the road link to be matched;
B. performing path link consistency detection on candidate matching nodes of each node, taking all candidate matching nodes on the same path link as a group, sequencing according to the sequence of forming the path link, and adding the path link where each group of candidate matching nodes is located into a candidate matching set as a candidate matching path link;
C. and extracting all road section matching relation pairs to be evaluated according to the node corresponding relation between the road link to be matched and the candidate matching link, and putting the road section matching relation pairs to be evaluated into a linked list to obtain a candidate set matching set of the road link to be matched.
9. The device for detecting vector road change based on similarity measurement according to claim 8, wherein the spatial similarity evaluation model established by the road similarity evaluation module is as follows:
Sim＝ω1SimS+ω2SimD+ω3SimL+ω4SimA
wherein SimS、SimD、SimLAnd SimARespectively is the shape similarity SShapeDistance proximity SDistanceLength similarity SLengthSimilarity with direction SOrientationDimensionless normalized value, omega1、ω2、ω3And ω4Is the weight of the corresponding index, and ω1+ω2+ω3+ω4＝1。
10. The device according to claim 8, wherein the change detection module calculates the similarity of the shape S between the road to be detected and the road with the same name as the road to be detectedShapeDistance proximity SDistanceLength similarity SLengthSimilarity with direction SOrientationAnd comparing the feature with a corresponding threshold value, judging whether the road to be detected changes on the corresponding feature, if the similarity of the certain feature is greater than the threshold value, indicating that the road to be detected changes on the feature, otherwise, determining that the road to be detected changes, and finally determining the type of the change according to the change condition of the road feature.
CN201810083737.2A 2018-01-29 2018-01-29 Similarity measurement-based vector road change detection method and device Active CN108492276B (en)

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 CN108492276A (en) 2018-09-04
CN108492276B true 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)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109543712B (en) * 2018-10-16 2023-04-07 哈尔滨工业大学 Method for identifying entities on temporal data set
CN109949692B (en) * 2019-03-27 2021-03-26 腾讯大地通途（北京）科技有限公司 Road network matching method and device, computer equipment and storage medium
CN110750607B (en) * 2019-09-17 2023-03-14 西安工程大学 Road network matching method based on GNSS vehicle track data
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
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

Citations (5)

* Cited by examiner, † Cited by third party
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
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

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5686088B2 (en) * 2011-11-14 2015-03-18 アイシン・エィ・ダブリュ株式会社 Road data creation device, road data creation method and program
US10151592B2 (en) * 2016-04-28 2018-12-11 Here Global B.V. Map matching quality evaluation

Patent Citations (5)

* Cited by examiner, † Cited by third party
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
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

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
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;翟仁健;《中国博士学位论文全文数据库》;20120715;第29，34，46，57-59，63，97,106页 *

Also Published As

Publication number Publication date
CN108492276A (en) 2018-09-04

Similar Documents

Publication Publication Date Title
CN108492276B (en) Similarity measurement-based vector road change detection method and device
CN111028277B (en) SAR and optical remote sensing image registration method based on pseudo-twin convolution neural network
CN109145171B (en) Multi-scale map data updating method
CN108376408B (en) Three-dimensional point cloud data rapid weighting registration method based on curvature features
Chen et al. Automatically conflating road vector data with orthoimagery
CN101980250B (en) Method for identifying target based on dimension reduction local feature descriptor and hidden conditional random field
JP5251080B2 (en) Object recognition method
CN108492298B (en) Multispectral image change detection method based on generation countermeasure network
CN103400388A (en) Method for eliminating Brisk (binary robust invariant scale keypoint) error matching point pair by utilizing RANSAC (random sampling consensus)
CN110619258B (en) Road track checking method based on high-resolution remote sensing image
CN112330661A (en) Multi-period vehicle-mounted laser point cloud road change monitoring method
Andrášik et al. Efficient road geometry identification from digital vector data
CN112053622A (en) Multi-ring polygon self-intersection mode recognition and processing method
CN103854290A (en) Extended target tracking method based on combination of skeleton characteristic points and distribution field descriptors
CN113204871A (en) Aviation blade air film hole identification method, device and system
Schmidt et al. Forest point processes for the automatic extraction of networks in raster data
Xiong A three-stage computational approach to network matching
KR101667875B1 (en) Automatic matching device and method of building polygon on a plurality of maps
Seybold Robust map matching for heterogeneous data via dominance decompositions
CN112330604B (en) Method for generating vectorized road model from point cloud data
CN110310322A (en) Method for detecting assembly surface of 10-micron-level high-precision device
US6571173B1 (en) Three-dimensional space curve comparison using spatial angle variance metric and applications thereof
Huh et al. Line segment confidence region-based string matching method for map conflation
Liu et al. M: N Object matching on multiscale datasets based on MBR combinatorial optimization algorithm and spatial district
CN112766385B (en) Many-source vector line data geometric matching and attribute fusion method

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