CN108197134A - Grouped point object automatic Synthesis algorithm under big data support - Google Patents
Grouped point object automatic Synthesis algorithm under big data support Download PDFInfo
- Publication number
- CN108197134A CN108197134A CN201711048489.XA CN201711048489A CN108197134A CN 108197134 A CN108197134 A CN 108197134A CN 201711048489 A CN201711048489 A CN 201711048489A CN 108197134 A CN108197134 A CN 108197134A
- Authority
- CN
- China
- Prior art keywords
- point
- coverage
- crowd
- polygon
- value
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Remote Sensing (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Grouped point object automatic Synthesis algorithm under big data support, the weight of point is the important parameter of point group automatic Synthesis, but has theoretical foundation of the algorithm to its determining shortage abundance.Big data can provide abundant, real-time data reference for the weight of point.It is proposed to this end that a kind of Grouped point object automatic Synthesis algorithm under big data support.Basic principle is as follows：First, coverage is introduced with influencing weight index of the two kinds of factors of crowd as point, is obtained related data and is handled and visualized；Secondly, on the basis of the basic principle of Map Generalization is followed, using coverage area of a polygon and crowd's quantity is influenced as foundation, the choice realized a little using normalization and " concentric circles " method is operated；Finally, synthesis result and visualized graphs comparative analysis and experimental verification have been subjected to.It is demonstrated experimentally that the algorithm not only inherits existing algorithm advantage, but also scientifically calculates the weight information of point group and applied to during point group automatic Synthesis, obtained result is more rationally and with stronger real-time.
Description
Technical field
The invention belongs to cartography and Geographical Information Sciences technical field, are that a kind of Grouped point object based on big data is automatic
Integration algorithm.
Background technology
Point group synthesis is the important component of Map Generalization, and the purpose is to correct as possible in the case of reduction of counting out
The Global Information of point group is expressed, i.e., is extracted from original point group containing certain amount point using certain model or algorithm
Subclass.The weight of point reflects the individual significance level a little in space group entirety, is the important parameter of point group synthesis.And
In existing point group integration algorithm, there are a kind of algorithm, such as algorithm based on Voronoi diagram etc., do not cared in combined process
And the weight information to point；Although another kind of algorithm has following two defects with respect to the weight information of point：
（1）Determining inaccurate and lacking sufficient calculation basis, such as the algorithm based on weighted Voronoi diagrams figure for weighted value, is mesh
Before until fairly perfect point group integration algorithm, but the weighted value at algorithm midpoint is given by expertise, such as by first
Grade hospital weight setting is 2, and second class hospital is set as 1, and irrigation is set as 2 with well weights, other well weights are set as 1；（2）Weighted value is
It is previously given can't be changed.But in actual geographic space, over time, the weighted value of point is often
It changes, such as certain the second class hospital represented in map with point is directed to the present situation that cardiovascular and cerebrovascular case largely increases, and introduces people
Just greatly develop corresponding profession so that its scale and medical treatment number have surmounted many first class hospitals, its weighted value also should at this time
With variation.
Big data can provide abundant real time information for goal object, can be not only that the determining of weighted value of point provides
Scientific and effective reference also makes the real-time of weighted value of invocation point be changed as possibility.
Invention content
For the above situation, big data is introduced herein, as the foundation for calculating point group weight, proposes one on this basis
Kind Grouped point object automatic Synthesis algorithm.
The weighted data selection of point
Point group is the important component of space and geographical target, and when map scale reduces, many space atural objects are all shown as
Point group, such as settlement place, hospital, school, supermarket, restaurant, trees.For more than point group, mostly with respective coverage
Etc. features, the point groups such as hospital, school, supermarket have respective coverage and influence crowd.In consideration of it, herein by point
The index of coverage and influence crowd as the weight for weighing point.
Choose weight index of this two kinds of factors as point, be because point coverage and influence crowd and the two it
Between relationship reflect two aspect property of point group：
(1) height of point object grade.For a point object, around usually have the service for spreading apart and centered on it
Range, this range can reflect the hierarchy level of point object.If this range it is relatively large and meanwhile influence crowd's quantity compared with
Mostly just illustrate that this point is higher ranked, such as hospital's coverage is wide, influence crowd quantity can mostly illustrate that this hospital advises
Mould is big, Medical Devices are advanced, the horizontal height of Health Services, and the corresponding grade of such hospital is also higher in actual life；If on the contrary,
The corresponding coverage of point is smaller while influences crowd's negligible amounts, then illustrates that this grade is relatively low；
(2) point group local density size.If the coverage of a point object is wide and the crowd quantity of influence can illustrate less
Similar point group number is less near the point, and local density is smaller.
In the combined process of point, higher ranked point, importance degree is also relatively high, should be retained；Part is close
Smaller point is spent, in order to keep the structure feature of space group as far as possible, should also be retained.
Therefore in the algorithm, the data for obtaining and handling are mainly the coverage of space group and influence crowd.
The acquisition and processing of big data
Two weight measurement indexs are selected in this algorithm, acquisition methods are as follows：
（1）Coverage：Obtain coverage data the first step be influence crowd to be obtained source place information, these positions
Information can be extracted or by cellular network mobile phone track following data acquisition, using web crawlers mode based on network data
Stream etc. is obtained, naturally it is also possible to be obtained by the internal information of respective point.Such as the coverage for hospital, it can be with
Hospital carries out track following for starting point, inquires final on trajectory and is denoted as influence crowd source place, can also registering by hospital
Acquisition of information.
（2）Influence crowd's quantity：With popularizing for network and social software etc., the crowd incremental data of influence can be by rising
News position data or microblogging data of registering crawl, and the internal information of point can also be utilized to obtain.Such as the influence for hospital
Crowd's incremental data can utilize Tencent's position big data Approximate Calculation, can also utilize in certain period of time around some point
Microblogging is registered data and approximate is obtained in specified range.
More accurate effective data in order to obtain, it is desirable that cleaned to initial data.This paper algorithms are mainly from following two
A aspect extracts valid data：(1) validity is examined.The validity of data is checked, if in zone of reasonableness
Within, such as it cannot be negative to influence crowd's quantity；(2) applicability is examined.For collected coverage data point, according to it
Longitude and latitude imports ArcGIS analysis platforms, is stacked with research area's range, will study the point deletion of area periphery.
The expression of coverage：Coverage is by a coverage Polygons Representation for surrounding.This algorithm provides, influences model
Polygon is enclosed by distributing edge polygon that the point that projects on map of source place for influencing crowd is formed to represent[.Influence model
The building process for enclosing polygon is as follows：
Step1：Influence crowd source place has very big contingency, in order to avoid influence of this contingency to range, record
The frequency of occurrences of collected data point, and by its descending sort, give up 10% point of frequency minimum, remaining collection point is thrown
Shadow is to plane coordinates as original point group.
Step2：Original point group is scanned, building it using the Delaunay Triangulation of belt restraining constrains Delaunay triangles
Net is re-introduced into dynamic threshold " peeling " method and builds its distributing edge polygon.
Original point group is stored in array pointArray, acquires the distributing edge area of a polygon that each pair of point in array is answered
And it is sequentially stored into array areaArray.
(2) expression of crowd's quantity is influenced：The size of influence crowd is by accessing the people of the region in a period of time
Number determines.For the ease of representing and reducing error, this algorithm is carried out the influence demographic data of each point using hierarchical clustering algorithm
Cluster so that influence crowd's quantity variance of the point in same cluster is smaller, and the influence people of the point between different clusters
Group's quantity variance is larger.It is represented on map, shows as the coverage polygon Fill Color of gradual change, color deeper generation
Influence crowd's quantity of table respective point is bigger, and on the contrary then corresponding points influence crowd's quantity are smaller.
Equally, the influence crowd quantity deposit array numArray each pair of point in array pointArray answered.
3 points of choice
The weight at algorithm midpoint by coverage area of a polygon and influence two factors of crowd's quantity determine, on this basis,
Point is divided into three types：It is high-grade must retain (I type), inferior grade directly gives up that (II type), fall between participation
Choose competition (III type).Define two kinds of selection constraintss：
（1）Grade constraints according to coverage and influences crowd's quantity differentiation three types point, retains I type during choosing
Point directly deletes 1 type point.Wherein：
TypeⅠ={P i |P i The big or of its influence crowd's quantity of the corresponding big or of coverage（Its influence of the big and of its coverage
Crowd's quantity is big）}
TypeⅡ={P i |P i Its influence crowd's quantity of the corresponding small and of coverage is small }
（2）Proximity relations constraints, for 1 type point, by it as the following formula（1）The unit coverage area effect people acquired
It is deleted one by one after numerical value ascending sort, it, will be corresponding with the polygon that its coverage polygon is neighbouring if certain point is deleted
Point carry out " curing ", find and handle next non-" curing " point, until satisfaction deletion termination condition；If until institute in range
It is a little deleted termination condition by " curing " and is not met yet, then thaw " curing " point, carries out the 2nd time, the 3rd time successively and deletes behaviour
Make.
The choice algorithm of point
According to constraints above condition, 1 type point is coverage in point group and influences all smaller point of crowd's quantity, its choosing
Take simple compared with 1 type point and III type point, therefore this algorithm takes following three kinds of strategies during the choice of point：
（1）It according to root law, chooses 1 type point of specified quantity and is deleted, left point then forms the result after synthesis；
（2）In order to realize the comparison between different dimension data, respectively by coverage area of a polygon and influence crowd's quantity
Value is converted into dimensionless scalar using normalized method；
（3）Normalized result is represented to be abscissa, coverage area of a polygon for ordinate to influence crowd's quantity
Plane coordinate system in, the smaller point of weighted value at this time（1 type point）Closer to coordinate origin, therefore use and done by the center of circle of origin
The method of 1/4 concentric circles, the successively smaller point of weight selection value judge it and delete operation, behind be called " with one heart
Circle " method.
The specific steps of the deletion of point are described as follows：
（1）The number deleted in advance in combined process a little is acquired according to root lawn；
（2）It is that the longitudinal axis establishes plane right-angle coordinate to influence crowd's quantity as horizontal axis, coverage area of a polygon, will influences
Element in crowd quantity array numArray and coverage area of a polygon array areaArray is normalized respectively,
And its result is corresponded and is represented in a coordinate system, point in corresponding original point group of each point in coordinate system at this time
Weight properties value；
（3）Using coordinate origin as origin, the minimum value using weight properties value point and coordinate origin is initial radium in coordinate
1/4 circle is drawn in system, and " flag=' D ' are deposited into " II type " point array by the point addition deletion label on circular arc
SecondType, point addition corresponding to its coverage polygon neighbours' polygon retain label " flag=' R ' "；
（4）Minimum planes distance between weight properties value point in plane coordinate system as increment, update radius value, using origin as
1/4 concentric circles is done in the center of circle, and the point without any label in the annulus formed on its circular arc and its with previous concentric circles is added
Deletion is added to mark, is still deposited into array secondType, while to its corresponding point of coverage polygon neighbours' polygon
Addition retains label；
（5）ComparenWith the size of element number value in secondType arrays, if n values are big, return (4)；If two values are identical,
Then turn (6)；Otherwise, the point of last round of addition array secondType is deleted, and remove the deletion mark of corresponding points from array
Note, the unit area of these points is acquired by equation 2 above influences numbern p ‘ , and by its ascending sort, from front to back successively in sequencen p ‘ Label is deleted in the minimum point addition without any label of value, is stored in array secondType, and polygon to its coverage
Shape neighbours, which add, retains label, until element number value is identical with n values in secondType arrays, then turns (6)；
（6）The corresponding point of all elements in secondType arrays is deleted in original point group, residual point-group is then synthesis result,
Algorithm terminates.
Description of the drawings
Fig. 1 is coverage polygon structure
Fig. 2 is the cluster of influence crowd
Fig. 3 is algorithm flow chart
Fig. 4 is the combined process of supermarket's point group
Fig. 5 is the deletion process of point
Fig. 6 is the combined process of somewhere Partial Hospitals
Fig. 7 is the combined process of somewhere gas station
Fig. 8 is the combined process of somewhere traffic lights.
Claims (1)
1. the Grouped point object automatic Synthesis algorithm under a kind of big data support, feature include the following steps：
（1）The acquisition and processing of weighted data
Coverage and the acquisition for influencing demographic data
The coverage of supermarket is using supermarket as starting point, by crawling cell phone track data terminal identification letter in one month
Breath and cell phone signal framing information and obtain；It influences crowd's quantity information this certain model of Ge Yuenei supermarkets corresponding points
Microblogging in enclosing is registered the expression of the sum of data；
Data processing
The information of acquisition is cleaned, deletes other than survey region and wrong data point；It is original for treated
The influence crowd quantitative value of all the points carries out Hierarchical clustering analysis in point group, will influence point similar in crowd's quantitative value and is classified as one
Class, and the data difference between class and class is increased as far as possible；
The point where supermarket is marked on ArcGIS platforms, builds the coverage polygon of these points；It will influence crowd's quantity
Value is according to cluster result descending sort and chooses gradient color by being deep to the shallow coverage polygon for filling corresponding points；
（2）The choice of point
The definition and expression of constraints
The weight at algorithm midpoint by coverage area of a polygon and influence two factors of crowd's quantity determine, on this basis,
Point is divided into three types：It is high-grade must retain (I type), inferior grade directly gives up that (II type), fall between participation
Choose competition (III type)；
Define two kinds of selection constraintss：
Grade constraints according to coverage and influences crowd's quantity differentiation three types point, retains 1 type point during choosing,
Directly delete 1 type point；Wherein：TypeⅠ={P i |P i The big or of its influence crowd's quantity of the corresponding big or of coverage（It influences model
Enclose big and its influence crowd's quantity it is big）}
TypeⅡ={P i |P i Its influence crowd's quantity of the corresponding small and of coverage is small }
Proximity relations constraints, for 1 type point, by its by after unit coverage area effect number value ascending sort by
Point corresponding with the polygon that its coverage polygon is neighbouring if certain point is deleted, is carried out " curing ", sought by a deletion
It looks for and handles next non-" curing " point, termination condition is deleted until meeting；If until all the points are deleted by " curing " in range
Except termination condition does not meet yet, then thaw " curing " point, carries out the 2nd time, the 3rd time delete operation successively；
The choice algorithm of point
According to constraints above condition, 1 type point is coverage in point group and influences all smaller point of crowd's quantity, its choosing
Take simple compared with 1 type point and III type point, therefore this algorithm takes following three kinds of strategies during the choice of point：
A. it according to root law, chooses 1 type point of specified quantity and is deleted, left point then forms the knot after synthesis
Fruit；
B. in order to realize the comparison between different dimension data, respectively by coverage area of a polygon and influence crowd's quantitative value
Dimensionless scalar is converted into using normalized method；
C. normalized result is represented to be abscissa, coverage area of a polygon for ordinate to influence crowd's quantity
In plane coordinate system, the smaller point of weighted value at this time（1 type point）Closer to coordinate origin, therefore use and do 1/4 by the center of circle of origin
The method of concentric circles, the successively smaller point of weight selection value judge it and delete operation, behind be called " concentric circles "
Method；
The specific steps of the deletion of point are described as follows：
A. the number deleted in advance in combined process a little is acquired according to root lawn；
B. it is that the longitudinal axis establishes plane right-angle coordinate to influence crowd's quantity as horizontal axis, coverage area of a polygon,
By influence crowd quantity array numArray and coverage area of a polygon array areaArray in element respectively into
Row normalization, and its result is corresponded and is represented in a coordinate system, each in coordinate system puts corresponding original point group at this time
In a point weight properties value；
C. it using coordinate origin as origin, is drawn in a coordinate system as initial radium using the minimum value of weight properties value point and coordinate origin
1/4 circle, and by circular arc point addition delete label " flag=' D ', be deposited into " II type " point array secondType,
Point addition corresponding to its coverage polygon neighbours' polygon retains label " flag=' R ' "；
D. the minimum planes distance between weight properties value point in plane coordinate system is increment, updates radius value, using origin as
1/4 concentric circles is done in the center of circle, and the point without any label in the annulus formed on its circular arc and its with previous concentric circles is added
Deletion is added to mark, is still deposited into array secondType, while to its corresponding point of coverage polygon neighbours' polygon
Addition retains label, goes to (5)；
E. comparenWith the size of element number value in secondType arrays, if n values are big, return (d)；If two values are identical,
Turn (f)；Otherwise, the point of last round of addition array secondType from array is deleted, and removes the deletion label of corresponding points,
Acquiring the unit area of these points influences numbern p ‘ , and by its ascending sort, from front to back successively in sequencen p ‘ Value minimum
Label is deleted in the point addition for not having any label, is stored in array secondType, and its coverage polygon neighbours are added
Retain label, until element number value is identical with n values in secondType arrays, then turn (6)；
F. the corresponding point of all elements in secondType arrays is deleted in original point group, residual point-group is then synthesis result,
Algorithm terminates.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711048489.XA CN108197134B (en) | 2017-10-31 | 2017-10-31 | Automatic point group target synthesis algorithm supported by big data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711048489.XA CN108197134B (en) | 2017-10-31 | 2017-10-31 | Automatic point group target synthesis algorithm supported by big data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108197134A true CN108197134A (en) | 2018-06-22 |
| CN108197134B CN108197134B (en) | 2021-09-07 |
Family
ID=62572953
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711048489.XA Active CN108197134B (en) | 2017-10-31 | 2017-10-31 | Automatic point group target synthesis algorithm supported by big data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108197134B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111047019A (en) * | 2018-10-12 | 2020-04-21 | 兰州交通大学 | Construction algorithm of network weighted Voronoi graph |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102073981A (en) * | 2011-02-23 | 2011-05-25 | 中国测绘科学研究院 | Point group geographic entity selection method under the restriction of correlated elements |
| CN105740381A (en) * | 2016-01-27 | 2016-07-06 | 北京工业大学 | User interest mining method based on complex network characteristics and neural network clustering |
| US20170234693A1 (en) * | 2016-02-17 | 2017-08-17 | International Business Machines Corporation | Method, apparatus and product for efficient solution of nearest object problems |
-
2017
- 2017-10-31 CN CN201711048489.XA patent/CN108197134B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102073981A (en) * | 2011-02-23 | 2011-05-25 | 中国测绘科学研究院 | Point group geographic entity selection method under the restriction of correlated elements |
| CN105740381A (en) * | 2016-01-27 | 2016-07-06 | 北京工业大学 | User interest mining method based on complex network characteristics and neural network clustering |
| US20170234693A1 (en) * | 2016-02-17 | 2017-08-17 | International Business Machines Corporation | Method, apparatus and product for efficient solution of nearest object problems |
Non-Patent Citations (4)
| Title |
|---|
| AHUJA N: "Dot Pattern Processing Using Voronoi Neighborhoods", 《IEEE》 * |
| 杨敏: "顾及道路目标stroke特征保持的路网自动综合方法", 《测绘学报》 * |
| 禄小敏: "基于约束Delaunay三角网的线、面群目标分布边界计算", 《测绘工程》 * |
| 闫浩文: "地图点群综合的加权Voronoi算法", 《武汉大学学报（信息科技版）》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111047019A (en) * | 2018-10-12 | 2020-04-21 | 兰州交通大学 | Construction algorithm of network weighted Voronoi graph |
| CN111047019B (en) * | 2018-10-12 | 2022-09-23 | 兰州交通大学 | Network weighting Voronoi diagram construction method based on improved pulse coupling neural network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108197134B (en) | 2021-09-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108876027B (en) | GIS-based rural residential point centralized residential area site selection and optimization method | |
| CN109813273B (en) | Agricultural machinery repeated operation area judgment method based on spatial analysis | |
| CN114610829B (en) | Land information management method based on intelligent terminal and remote sensing intelligent identification | |
| CN109840260A (en) | A kind of extensive real-time rainfall automatic Observation station ranked data processing method based on dynamic interpolation | |
| CN106790468A (en) | A kind of distributed implementation method for analyzing user's WiFi event trace rules | |
| Karimi et al. | Urban expansion modeling using an enhanced decision tree algorithm | |
| CN106326923B (en) | A kind of position data clustering method of registering taking position into account and repeating with density peaks point | |
| CN113505510B (en) | Ecological safety pattern recognition method fusing landscape index and random walk model | |
| CN107330734A (en) | Business address system of selection based on Co location patterns and body | |
| CN107977673A (en) | A kind of economically active population's recognition methods based on big data | |
| CN109284773A (en) | Traffic trip endpoint recognition methods based on multilayer Agglomerative Hierarchical Clustering algorithm | |
| CN112954623B (en) | Resident occupancy rate estimation method based on mobile phone signaling big data | |
| CN110059881A (en) | The method of novel tea garden soil soil moisture content sensor optimization layout based on clustering algorithm | |
| CN110596008A (en) | Plot-based soil nutrient digital mapping method for agricultural region of Chinese Hongsheng plain | |
| CN111475746B (en) | Point-of-interest mining method, device, computer equipment and storage medium | |
| CN114881430A (en) | Community life convenience evaluation method based on network map service | |
| CN115311343A (en) | Grassland pest occurrence area investigation method based on multi-stage sampling | |
| CN114820532A (en) | Locust area analysis method and device | |
| CN108197134A (en) | Grouped point object automatic Synthesis algorithm under big data support | |
| CN112613530A (en) | Cell resident identification method and system based on adaptive density clustering algorithm | |
| CN113298301A (en) | Urban service facility evaluation and optimization method based on crowd digital portrait | |
| CN116976568A (en) | Data sharing method and system for assisting urban and rural planning and compiling | |
| Girardello et al. | Identifying important areas for butterfly conservation in Italy | |
| CN115661670B (en) | Method and device for identifying transition region on urban and rural natural landscape gradient | |
| CN116862258A (en) | Rural aggregation space reconstruction method considering aggregation and three-generation functions |
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 |