CN108920481A - A kind of road network method for reconstructing and system based on mobile phone location data - Google Patents

A kind of road network method for reconstructing and system based on mobile phone location data Download PDF

Info

Publication number
CN108920481A
CN108920481A CN201810361223.9A CN201810361223A CN108920481A CN 108920481 A CN108920481 A CN 108920481A CN 201810361223 A CN201810361223 A CN 201810361223A CN 108920481 A CN108920481 A CN 108920481A
Authority
CN
China
Prior art keywords
road
data set
mobile phone
road network
triangles
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
Application number
CN201810361223.9A
Other languages
Chinese (zh)
Other versions
CN108920481B (en
Inventor
杨林
王帅鑫
张志勇
左泽均
李圣文
叶亚琴
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.)
China University of Geosciences
Original Assignee
China University of Geosciences
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 China University of Geosciences filed Critical China University of Geosciences
Priority to CN201810361223.9A priority Critical patent/CN108920481B/en
Publication of CN108920481A publication Critical patent/CN108920481A/en
Application granted granted Critical
Publication of CN108920481B publication Critical patent/CN108920481B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W4/00Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02Services making use of location information
    • H04W4/029Location-based management or tracking services
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/10Services
    • G06Q50/26Government or public services

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Tourism & Hospitality (AREA)
  • Economics (AREA)
  • Human Resources & Organizations (AREA)
  • Educational Administration (AREA)
  • Signal Processing (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • General Health & Medical Sciences (AREA)
  • Development Economics (AREA)
  • Marketing (AREA)
  • Primary Health Care (AREA)
  • Strategic Management (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Traffic Control Systems (AREA)

Abstract

一种基于手机定位数据的道路网重建方法及系统,包括:获取原始数据集,原始数据集包括训练数据集和预测数据集;原始数据集为根据手机定位数据得到轨迹点;对训练数据集和预测数据集分别利用时间和密度两个维度进行特征构建与组合,并随之生成训练集和预测集;使用生成的训练集,基于机器学习中的算法得到采样点分类模型;将学习到的分类模型应用于预测集,将轨迹点划分为道路点和非道路点;基于约束Delaunay三角网对已识别出的道路点进行道路构建;按照拓扑关系,修改道路中心线,从而完成道路网的重建。本发明提出了一种高质量的精细化道路网生产和更新的新方式,将为无缝导航技术尤其是精细级道路网方面提供技术支撑,扩大路径导航服务的覆盖范围。

A road network reconstruction method and system based on mobile phone positioning data, comprising: obtaining an original data set, the original data set includes a training data set and a prediction data set; the original data set is to obtain trajectory points according to the mobile phone positioning data; the training data set and The prediction data set uses the two dimensions of time and density to construct and combine features, and then generates a training set and a prediction set; using the generated training set, the sampling point classification model is obtained based on the algorithm in machine learning; the learned classification The model is applied to the prediction set, and the trajectory points are divided into road points and non-road points; road construction is carried out on the identified road points based on the constrained Delaunay triangulation; according to the topological relationship, the road centerline is modified to complete the reconstruction of the road network. The invention proposes a new way of producing and updating high-quality refined road networks, which will provide technical support for seamless navigation technology, especially fine-level road networks, and expand the coverage of route navigation services.

Description

A kind of road network method for reconstructing and system based on mobile phone location data
Technical field
The present invention relates to space-time trajectory data digging technology, especially a kind of road network based on mobile phone location data is rebuild Method and system belong to GIS-Geographic Information System and intelligent transportation research field.
Background technique
Road is the important infrastructure in city, during rapid urban, how to be replenished in time and more new town Road data is a problem urgently to be resolved to keep its accuracy, timeliness and degree of perfection.Residential communities are advised in city Mould constantly expands, and internal passageway net is increasingly fine and complicated, but existing cell-level road information fine degree is not enough even Missing causes outdoor thin portion navigation in residential block to there is " blind area ", causes very big obstacle to the development of outdoor seamless navigation technology. At the same time, road network structure information is the core content of electronic map basic data, and even more electronic map realizes navigation feature, intelligence The data basis of energy traffic system, therefore, accurately and in time more new town fine-grade road network is an important problem (city City's fining road network refers to the cell-level branch that Intra-cell is respectively closed in city).
The convenience obtained with crowd-sourced track data and its real-time updated are greatly improved, academia and work Industry can be readily available a large amount of track datas comprising space time information, so that the road network based on space-time trajectory data mentions Method is taken to be possibly realized.Currently, there are mainly two types of Automatic Road Extraction methods:One is mention from vehicle GPS track data Take road information;Another kind is to extract road network information based on high-resolution remote sensing image.But existing road network extracting method master It concentrates in the research of each major trunk roads in city, and the fine road network of cell-level is usually ignored in city.Wherein, remote sensing image because Remote for image-forming range, coverage area is big, while portraying the scarce capacity of different spaces scale fine information, based on such data Processing method is only capable of preferably solving the information extraction of general major trunk roads, can not finally solve fining road network building and ask Topic;In addition, the acquisition of GPS track data is limited inside the building range such as general school, cell to a certain extent, and pedestrian Mobile trajectory data randomness is strong, and sample frequency is low, comprising various types of ground objects such as road, building and squares, so that being based on Mobile trajectory data extracts road network, and there are numerous difficulties.
Summary of the invention
The technical problems to be solved by the invention are for road fine degree in existing cell-level road network not enough and part The problem of missing, provides a kind of fining Road network extraction method, can be convenient the building work for accurately completing fine-grade road network Make.This method will be converted into two classes of waypoint Yu non-rice habitats point the problem of carrying out road network reconstruction based on extensive pedestrian's data Classification problem carries out the building of feature set using the various combination of two dimensions of time and density first, generates training set, then Learn sampled point disaggregated model using random forests algorithm, application class model divides data set later, obtains road Point set finally borrows constraint Delaunay triangulation network to having identified that road point set carries out road building and mention with road axis It takes, to complete the building for refining road network in closed area.It includes the following steps:
Step 1) obtains raw data set, and raw data set includes training dataset and predictive data set;Wherein original number It is that tracing point is obtained according to mobile phone location data according to collection;
Step 2) is utilized respectively two dimensions of time and density to training dataset and predictive data set and carries out feature construction With combine, and generate training set and forecast set therewith;
Step 3) learns to obtain sampled point disaggregated model based on the algorithm in machine learning using the training set generated;
The disaggregated model learnt is applied to forecast set by step 4), and tracing point is divided into waypoint and non-rice habitats point;
Step 5) clicks through the building of trade road to the road identified based on constraint Delaunay triangulation network;
Step 6) modifies road axis, to complete the reconstruction of road network according to topological relation.
Further, in the road network method for reconstructing of the invention based on mobile phone location data, training in the step 1 Data set is obtained according to following manner:The distribution pattern for obtaining multiple closing Intra-cell tracing points, selects tracing point distribution The different multiple nonoverlapping small-scale regions of mode form training dataset.
Further, in the road network method for reconstructing of the invention based on mobile phone location data, density in the step 2 Characteristic value includes:Binding time dimension carries out Delaunay Triangulation, the institute that each tracing point is connected to training dataset There is the average length of adjacent triangle edges.
Further, in the road network method for reconstructing of the invention based on mobile phone location data, density in the step 2 Characteristic value includes:Binding time dimension constructs Voronoi diagram, Voronoi diagram corresponding to each tracing point to training dataset The inverse of middle area of a polygon.
Further, in the road network method for reconstructing of the invention based on mobile phone location data, the time in the step 2 Specific step is as follows with the feature construction of two latitudes of density and combination:
By analyzing the behavior pattern of people, based on the space-time Density Distribution difference structure for closing small internal road and non-rice habitats Feature space is built, that is, passes through track points amount, the adjoining triangle edges length based on constraint Delaunay triangulation network in buffer area It is poor with the aggregation of three kinds of geometrical characteristic joint time dimension expression tracing point spatial and temporal distributions of area of polygon in Voronoi diagram It is different;The identification of road waypoint is carried out based on the feature set building classifier that above-mentioned three kinds different geometry dimension control conditions are established.
Further, in the road network method for reconstructing of the invention based on mobile phone location data, learn in the step 3 Algorithm includes random forest, support vector machines.
Further, in the road network method for reconstructing of the invention based on mobile phone location data, for training dataset In track point data generated based on road net data incomplete in OSM according to certain radius using corresponding OSM data Line buffer area falls in the tracing point in buffer area labeled as 1, falls in the tracing point except buffer area labeled as 0, is based on this, will The tracing point that training data is concentrated is divided into two class of waypoint and non-rice habitats point, label when as subsequent training machine learning model Data.
Further, in the road network method for reconstructing of the invention based on mobile phone location data, the step 5 is specifically wrapped It includes:
The road point set building constraint Delaunay triangulation network formed using the road point extracted in step 4, root According to adjacent triangle number, the triangle in the triangulation network is divided into 3 classes, respectively correspond the entrance of road, the trunk of road and Intersection is respectively handled it, road axis is extracted, the specific implementation process is as follows described:
Building constraint Delaunay triangulation network, the triangle in the triangulation network is divided into the road point set identified 3 classes:Only having adjacent triangle on one side is A class, and it is B class that, which there is adjacent triangle on both sides, and there is adjacent triangle on three sides For C class, A class triangle connects the vertex of the midpoint on unique proximity side corresponding thereto, and B class triangle connects in two neighbouring sides Point, C class triangle connect the midpoint of center of gravity and three sides, in search connection procedure, first search for, terminate since C class triangle In A class triangle or C class triangle, network a line is obtained, until having handled C class triangle;It uses and is opened from A class triangle again Begin to search for, terminates at A class triangle, obtain all sides, to form road axis network.
Further, in the road network method for reconstructing of the invention based on mobile phone location data, the step 6 is specifically wrapped It includes:
Optimization and perfect road center line, carry out the road axis near intersection processing is straightened, to complete to seal Closed region refines the reconstruction of road network, and specific processing mode is as described below:
For the arc road network near intersection, since crosspoint, straight line is cropped with certain length, then compare sanction The angle of rear remaining line is cut, when the angle of two lines meets a certain range, two lines of connection are into a line, when being crossroad, Node of the intersection point as net is acquired, road network is ultimately formed, to complete the reconstruction of closed area road network.
The present invention is to solve its technical problem, additionally provides a kind of road network reconstructing system based on mobile phone location data, Road network reconstruction is carried out using the road network method for reconstructing based on mobile phone location data of any of the above-described.
Implement the road network method for reconstructing and system of the invention based on mobile phone location data, the present invention is according to the row of people Mining analysis is carried out to track data according to space-time density feature for mode, using included in a large amount of sparse mobile phone location data In semantic information, propose it is a kind of constructed using mobile phone location data fining road network new method, biography can be improved Unite road network update efficiency and reduce maintenance cost, substantially reduce road network extraction time, can be make up cell-level road The not perfect and missing problem of information provides a kind of new mode.
Detailed description of the invention
Present invention will be further explained below with reference to the attached drawings and examples, in attached drawing:
Fig. 1 is flow chart of the method for the present invention;
Fig. 2 is the schematic diagram that the present invention calculates each area of a polygon in Voronoi diagram;
Fig. 3 is the schematic diagram for the average length that the present invention calculates the adjacent triangle edges of tracing point in Delaunay triangulation network.
Specific embodiment
For a clearer understanding of the technical characteristics, objects and effects of the present invention, now control attached drawing is described in detail A specific embodiment of the invention.
The present invention relates to a kind of road network method for reconstructing based on mobile phone location data, carries out at fining road network Reason.The present invention is analyzed by the distribution pattern to multiple closing Intra-cell tracing points first, then selects different community The small range regions of interior multiple representative distribution patterns constructs training dataset, and remainder incorporates prediction data into Collection for the track point data that training data is concentrated, carries out manual sort, is divided into waypoint and non-rice habitats point two at the same time Class, the label value as training data;Using two dimensions of time and density carry out feature constructions with combine, wherein time dimension It can be by being divided into 7 different periods (6 for one day:00-8:00,08:00-12:00,12:00-14:00,14:00-17: 00,17:00-20:00,20:00-24:00,24:00-06:00) it realizes, in terms of density, can be constructed from three angles Feature:Points amount in track in buffer area, the area of each polygon based on Voronoi diagram and based on constraint Delaunay tri- The average length of the adjacent triangle edges of each tracing point is netted at angle, combines all of above feature and corresponding with each tracing point Label value forms training set;Using the training set of generation, based on the random forests algorithm study sampled point classification in sorting algorithm Model;Later, the disaggregated model acquired is applied in predictive data set, for the road waypoint in the point data of identification prediction track With non-rice habitats point, road point set is obtained;Finally in road point set, using constraint Delaunay triangulation network, road is extracted Center line completes the building of fining road network in closing cell.
In order to make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with attached drawing and specific implementation Example, the present invention will be described in further detail.It should be appreciated that specific embodiment described herein is only to explain this hair It is bright, it is not intended to limit the present invention.
As shown in Figure 1, Fig. 1 is the side provided by the invention for carrying out fining road network based on mobile phone location data and rebuilding Method flow chart, the described method comprises the following steps:
Step 1, raw data set is obtained, raw data set includes training dataset and predictive data set;Wherein original number It is that tracing point is obtained according to mobile phone location data according to collection.Preferably, concentrate selection tracing point distribution pattern different from initial data Multiple nonoverlapping small-scale regions form training datasets, remaining track point data is divided into predictive data set;Meanwhile it is right In the track point data that training data is concentrated, needs to carry out manual sort, be divided into two class of waypoint and non-rice habitats point, as subsequent Label data when training machine learning model.
Embodiment is implemented as follows:
For raw data set, tracing point distribution situation therein is analyzed first, it is multiple from different area of space selections Different distribution patterns, such as:Center Assembled distribution, ribbon distribution, random discrete distribution etc., by a variety of distribution patterns Track point data is merged into as training dataset, remaining track point data is divided into predictive data set;Meanwhile for training data Corresponding OSM data can be used in the track point data of concentration, based on incomplete road net data in OSM according to certain half Diameter generates line buffer area, falls in the tracing point in buffer area labeled as 1, falls in the tracing point except buffer area labeled as 0, is based on This, is divided into two class of waypoint and non-rice habitats point for the tracing point that training data is concentrated, when as subsequent training machine learning model Label data.
Wherein, the radius of line buffer area can be determined according to the mean radius of cell-level branch in city, Intra-cell road The radius on road is generally within 8 meters.
Step 2, two dimensions of time and density are utilized respectively to training dataset and predictive data set to complete feature set Building with combine, and generate training set and forecast set therewith.Wherein time dimension can be by being divided into 7 differences for one day Period can carry out construction feature from three angles in terms of density:Points amount in track in buffer area, based on Voronoi diagram The area of each polygon and the average length that triangle edges are abutted based on each tracing point of constraint Delaunay triangulation network, group All of above feature and label value corresponding with each tracing point are closed, training set is formed.
Specific building process is as follows:
According to the trip of people rule, on time dimension, 7 periods was divided by one day, are 06 respectively:00-08: 00,8:00-12:00,12:00-14:00,14:00-17:00,17:00-20:00,20:00-24:00,24:00-06:00;? In terms of density, buffer area radius takes 3 different values according to road conditions respectively:8m, 12m, 16m, by 3 buffer area radiuses and 7 A period carries out feature two-by-two and combines, and constructs 21 features, for indicating some tracing point in special time period, radius is Density value in 8/12/16 meter, quantitative formula are as follows:
Wherein, { 6,8,12 } Radius=, Point={ P1,P2,....,Pn, Time=(6,8), (8,12), (12, 14), (14,17), (17,20), (20,24), (0,6) },It indicates in period TaUnder, tracing point PiIt is R in radiusbModel The number of interior included tracing point is enclosed, the combination of different time sections and different radii feature is as shown in the table:
Later according to 7 different periods, training dataset is divided into 7 parts, construct respectively corresponding Voronoi diagram and Delaunay triangulation network is constrained, for the Voronoi diagram that building obtains, the area of each polygon in figure is calculated, calculates therewith The inverse of area, if Fig. 2 is a part of a Voronoi diagram, the present invention needs to calculate polygon R corresponding to tracing point P Area, and the inverse and time dimension of area are combined, as 7 features of expression track vertex neighborhood density, are calculated public Formula is as follows:
In above-mentioned formula,Indicate point PiThe area of corresponding polygon;Then for constraint Delaunay triangle Net calculates the average length of the adjacent triangle edges of each tracing point, if Fig. 3 is a part of a Delaunay triangulation network, this Invention needs to calculate the average length of all adjacent edges of tracing point P, and by it in conjunction with time dimension, adjacent as expression tracing point Other 7 features of domain density, calculation formula are as follows:
In above-mentioned formula, Neighbor_P indicates point PiThe set of all consecutive points, L indicate the distance between consecutive points, such as A in Fig. 3, b, c, d, e, f, g, h;Finally, by above-described 35 features and corresponding label described in step 1 Value merges, and generates training set.
Step 3, using the training set of generation, based in Python machine learning packet scikit-learn RandomForestClassifier module, on the basis of partial parameters in adjustment algorithm, sampling that learning data is concentrated Point disaggregated model.
Step 4, the sampled point disaggregated model learnt based on training set is applied in predictive data set, it is pre- for identifying The road waypoint and non-rice habitats point in the point data of track are surveyed, road point set therein is extracted.
Step 5, using the road point set building constraint Delaunay triangulation network extracted in step 4, according to adjoining Triangle in the triangulation network is divided into 3 classes by triangle number, is respectively corresponded the entrance of road, the trunk of road and road and is handed over Prong is respectively handled it, road axis is extracted, the specific implementation process is as follows described:
Building constraint Delaunay triangulation network, the triangle in the triangulation network is divided into the road point set identified 3 classes:Only having adjacent triangle on one side is A class, and it is B class that, which there is adjacent triangle on both sides, and there is adjacent triangle on three sides For C class, A class triangle connects the vertex of the midpoint on unique proximity side corresponding thereto, and B class triangle connects in two neighbouring sides Point, C class triangle connect the midpoint of center of gravity and three sides, in search connection procedure, first search for, terminate since C class triangle In A class triangle or C class triangle, network a line is obtained, until having handled C class triangle;It uses and is opened from A class triangle again Begin to search for, terminates at A class triangle, obtain all sides, to form road axis network.
Step 6, optimization and perfect road center line, carry out the road axis near intersection processing is straightened, thus The reconstruction of closed area fining road network is completed, specific processing mode is as described below:
For the arc road network near intersection, since crosspoint, straight line is cropped with certain length, then compare sanction The angle of rear remaining line is cut, when the angle of two lines meets a certain range, two lines of connection are into a line, when being crossroad, Node of the intersection point as net is acquired, road network is ultimately formed, to complete the reconstruction of closed area fining road network.
The embodiment of the present invention is described with above attached drawing, but the invention is not limited to above-mentioned specific Embodiment, the above mentioned embodiment is only schematical, rather than restrictive, those skilled in the art Under the inspiration of the present invention, without breaking away from the scope protected by the purposes and claims of the present invention, it can also make very much Form, all of these belong to the protection of the present invention.

Claims (10)

1.一种基于手机定位数据的道路网重建方法,其特征在于,包括以下步骤:1. A road network reconstruction method based on mobile phone location data, is characterized in that, comprises the following steps: 步骤1)获取原始数据集,原始数据集包括训练数据集和预测数据集;其中原始数据集为根据手机定位数据得到轨迹点;Step 1) obtain original data set, original data set comprises training data set and prediction data set; Wherein original data set is to obtain track point according to mobile phone positioning data; 步骤2)对训练数据集和预测数据集分别利用时间和密度两个维度进行特征构建与组合,并随之生成训练集和预测集;Step 2) construct and combine the features of the training data set and the prediction data set using the time and density dimensions respectively, and then generate the training set and the prediction set; 步骤3)使用生成的训练集,基于机器学习中的算法学习得到采样点分类模型;Step 3) use the generated training set to obtain the sampling point classification model based on algorithm learning in machine learning; 步骤4)将学习到的分类模型应用于预测集,将轨迹点划分为道路点和非道路点;Step 4) apply the learned classification model to the prediction set, and divide the track points into road points and non-road points; 步骤5)基于约束Delaunay三角网对已识别出的道路点进行道路构建;Step 5) Carry out road construction to the identified road points based on the constrained Delaunay triangulation; 步骤6)按照拓扑关系,修改道路中心线,从而完成道路网的重建。Step 6) Modify the road centerline according to the topological relationship, so as to complete the reconstruction of the road network. 2.根据权利要求1所述的基于手机定位数据的道路网重建方法,其特征在于,2. the road network reconstruction method based on mobile phone positioning data according to claim 1, is characterized in that, 所述步骤1中训练数据集是根据下述方式得到:获取多个封闭小区内部轨迹点的分布模式,选择轨迹点分布模式不同的多个不重叠的小规模区域组成训练数据集。The training data set in the step 1 is obtained according to the following method: the distribution pattern of trajectory points inside a plurality of closed cells is obtained, and a plurality of non-overlapping small-scale areas with different distribution patterns of trajectory points are selected to form the training data set. 3.根据权利要求1所述的基于手机定位数据的道路网重建方法,其特征在于,所述步骤2中密度特征值包括:结合时间维度对训练数据集进行Delaunay三角剖分,每个轨迹点所连接的所有邻接三角形边的平均长度。3. the road network reconstruction method based on mobile phone location data according to claim 1, is characterized in that, in described step 2, density characteristic value comprises: in conjunction with time dimension, training data set is carried out Delaunay triangulation, each track point The average length of all adjacent triangle sides connected. 4.根据权利要求1所述的基于手机定位数据的道路网重建方法,其特征在于,所述步骤2中密度特征值包括:结合时间维度对训练数据集构建Voronoi图,每个轨迹点所对应的Voronoi图中多边形面积的倒数。4. the road network reconstruction method based on mobile phone positioning data according to claim 1, is characterized in that, in described step 2, density feature value comprises: combine time dimension to training data set construction Voronoi diagram, each locus point corresponds The reciprocal of the area of a polygon in a Voronoi diagram. 5.根据权利要求1所述的基于手机定位数据的道路网重建方法,其特征在于,所述步骤2中时间和密度两个纬度的特征构建和组合的具体步骤如下:5. the road network rebuilding method based on mobile phone positioning data according to claim 1, is characterized in that, in described step 2, the specific steps of the feature construction and combination of two latitudes of time and density are as follows: 通过分析人们的行为模式,基于封闭小区内道路和非道路的时空密度分布差异构建特征空间,即通过缓冲区内轨迹点数量、基于约束Delaunay三角网的邻接三角形边长度和Voronoi图中多边形的面积三种几何特征联合时间维度表达轨迹点时空分布的聚集性差异;基于上述三种不同几何维度控制条件所建立的特征集构建分类器进行道路点识别。By analyzing people's behavior patterns, the feature space is constructed based on the difference in the spatial-temporal density distribution of roads and non-roads in the closed area, that is, through the number of trajectory points in the buffer zone, the length of the adjacent triangle side based on the constrained Delaunay triangulation and the area of the polygon in the Voronoi diagram The three geometric features combined with the time dimension express the aggregation difference of the temporal and spatial distribution of track points; based on the feature sets established by the above three different geometric dimension control conditions, a classifier is constructed for road point recognition. 6.根据权利要求1所述的基于手机定位数据的道路网重建方法,其特征在于,所述步骤3中学习算法包括随机森林、支持向量机。6. The road network reconstruction method based on mobile phone positioning data according to claim 1, characterized in that, in the step 3, the learning algorithm comprises random forest and support vector machine. 7.根据权利要求1所述的基于手机定位数据的道路网重建方法,其特征在于,对于训练数据集中的轨迹点数据,使用相对应的OSM数据,基于OSM中不完善的路网数据按照一定半径生成线缓冲区,落在缓冲区内的轨迹点标记为1,落在缓冲区之外的轨迹点标记为0,基于此,将训练数据集中的轨迹点分为道路点和非道路点两类,作为后续训练机器学习模型时的标签数据。7. The road network reconstruction method based on mobile phone positioning data according to claim 1, characterized in that, for the track point data in the training data set, use corresponding OSM data, based on the imperfect road network data in the OSM according to a certain The radius generates a line buffer. The track points falling in the buffer zone are marked as 1, and the track points falling outside the buffer zone are marked as 0. Based on this, the track points in the training data set are divided into road points and non-road points. class, as the label data for the subsequent training of the machine learning model. 8.根据权利要求1所述的基于手机定位数据的道路网重建方法,其特征在于,所述步骤5具体包括:8. The road network reconstruction method based on mobile phone positioning data according to claim 1, wherein said step 5 specifically comprises: 使用步骤4中提取得到的道路点形成的道路点集合构建约束Delaunay三角网,根据邻接三角形个数,将三角网中的三角形分为3类,分别对应道路的出入口、道路的主干和道路交叉口,分别对其进行处理,提取道路中心线,具体实现过程如下所述:Use the road point collection formed by the road points extracted in step 4 to construct a constrained Delaunay triangular network, and divide the triangles in the triangular network into three categories according to the number of adjacent triangles, corresponding to the entrance and exit of the road, the trunk of the road and the intersection of the road. , process them separately, and extract the road centerline. The specific implementation process is as follows: 在已识别出的道路点集合上构建约束Delaunay三角网,将三角网中的三角形分为3类:只有一边有邻近三角形的为A类,两边有邻近三角形的为B类,三边均有邻近三角形的为C类,A类三角形连接唯一邻近边的中点与其相对的顶点,B类三角形连接两条邻近边的中点,C类三角形连接重心与三边的中点,在搜索连接过程中,先从C类三角形开始搜索,终止于A类三角形或C类三角形,得到网络一条边,直到处理完C类三角形;再采用从A类三角形开始搜索,终止于A类三角形,得到所有边,从而形成道路中心线网络。Construct a constrained Delaunay triangulation network on the identified road point set, divide the triangles in the triangulation network into three categories: only one side has adjacent triangles as A category, two sides have adjacent triangles as B category, and all three sides have adjacent triangles. The triangles are of class C, the triangles of class A connect the midpoint of the only adjacent side and its opposite vertex, the triangles of class B connect the midpoints of two adjacent sides, and the triangles of class C connect the center of gravity and the midpoints of three sides. During the search connection process , first start searching from type C triangles, end at type A triangles or C type triangles, get one edge of the network, until processing type C triangles; then start searching from type A triangles, end at type A triangles, get all edges, Thus a road centerline network is formed. 9.根据权利要求1所述的基于手机定位数据的道路网重建方法,其特征在于,所述步骤6具体包括:9. The road network reconstruction method based on mobile phone positioning data according to claim 1, wherein said step 6 specifically comprises: 优化和完善道路中心线,对交叉口附近的道路中心线进行拉直处理,从而完成封闭区域精细化道路网的重建,具体处理方式如下所述:Optimize and improve the road centerline, and straighten the road centerline near the intersection, so as to complete the reconstruction of the refined road network in the closed area. The specific processing methods are as follows: 对于交叉口附近的弧形路网,从交叉点开始,以一定的长度裁剪掉直线,再比较裁剪后剩余线的角度,当两线的角度满足一定范围时,连接两线成一条线,当是十字路口时,求得交点作为网的结点,最终形成道路网,从而完成封闭区域道路网的重建。For the arc-shaped road network near the intersection, start from the intersection, cut off the straight line with a certain length, and then compare the angle of the remaining line after cutting. When the angle of the two lines meets a certain range, connect the two lines to form a line. When it is a crossroad, the intersection point is obtained as the node of the network, and finally the road network is formed, so as to complete the reconstruction of the road network in the closed area. 10.一种基于手机定位数据的道路网重建系统,其特征在于,采用权利要求1-9任一项所述的基于手机定位数据的道路网重建方法进行道路网重建。10. A road network reconstruction system based on mobile phone positioning data, characterized in that the road network reconstruction is carried out by using the road network reconstruction method based on mobile phone positioning data according to any one of claims 1-9.
CN201810361223.9A 2018-04-20 2018-04-20 A method and system for road network reconstruction based on mobile phone positioning data Expired - Fee Related CN108920481B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810361223.9A CN108920481B (en) 2018-04-20 2018-04-20 A method and system for road network reconstruction based on mobile phone positioning data

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810361223.9A CN108920481B (en) 2018-04-20 2018-04-20 A method and system for road network reconstruction based on mobile phone positioning data

Publications (2)

Publication Number Publication Date
CN108920481A true CN108920481A (en) 2018-11-30
CN108920481B CN108920481B (en) 2020-11-27

Family

ID=64404350

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810361223.9A Expired - Fee Related CN108920481B (en) 2018-04-20 2018-04-20 A method and system for road network reconstruction based on mobile phone positioning data

Country Status (1)

Country Link
CN (1) CN108920481B (en)

Cited By (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109684384A (en) * 2018-12-29 2019-04-26 太原科技大学 A kind of track data space-time density Analysis System and its analysis method
CN109934865A (en) * 2019-03-12 2019-06-25 重庆市勘测院 Extraction method of road network centerline
CN110008872A (en) * 2019-03-25 2019-07-12 浙江大学 A road network extraction method combining vehicle trajectories and remote sensing images
CN110096535A (en) * 2019-05-16 2019-08-06 北京航空航天大学 A kind of traffic elasticity communication mode method for digging based on elastic triangle theory
CN110580388A (en) * 2019-08-28 2019-12-17 北方工业大学 A Channel Network Extraction Method Based on Crowd-source Trajectory Data
CN110633342A (en) * 2019-07-29 2019-12-31 武汉光庭信息技术股份有限公司 Lane topology network generation method
CN110688958A (en) * 2019-09-28 2020-01-14 中国测绘科学研究院 Intersection Recognition Method Based on GoogLeNet Neural Network
CN110988880A (en) * 2019-12-12 2020-04-10 南京莱斯电子设备有限公司 Geographic information extraction and target tracking method based on SMR target track
CN111426329A (en) * 2020-03-12 2020-07-17 深圳小库科技有限公司 Road generation method and device, computer equipment and storage medium
CN111833630A (en) * 2019-12-31 2020-10-27 北京嘀嘀无限科技发展有限公司 Method and device for determining data release position, storage medium and electronic equipment
CN112013862A (en) * 2020-07-31 2020-12-01 深圳大学 Pedestrian network extraction and updating method based on crowdsourcing trajectory
CN112566025A (en) * 2020-12-10 2021-03-26 南京市城市与交通规划设计研究院股份有限公司 Bus passenger flow source destination identification method and device, electronic equipment and storage medium
CN112598724A (en) * 2021-03-01 2021-04-02 武大吉奥信息技术有限公司 Improved TIN-based planar element center line extraction method
CN113008251A (en) * 2021-02-22 2021-06-22 湖南大学 Digital map updating method for unstructured roads in closed area
CN113052084A (en) * 2021-03-26 2021-06-29 中国地质大学(武汉) Community-level vector road network extraction method based on mobile phone positioning data
WO2022031222A1 (en) * 2020-08-01 2022-02-10 Grabtaxi Holdings Pte. Ltd. Processing appratus and method for generating route navigation data
CN117891892A (en) * 2024-03-14 2024-04-16 南京师范大学 A ranger trajectory optimization algorithm based on historical patrol data

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8396257B2 (en) * 2008-07-25 2013-03-12 Navteq B.V. End user image open area maps
CN105069824A (en) * 2015-08-11 2015-11-18 中南大学 GPS data based automatic construction method and system for open-pit mine road network
CN106528740A (en) * 2016-11-04 2017-03-22 中科宇图科技股份有限公司 Delaunay triangular net-based road center line extraction method
CN106778605A (en) * 2016-12-14 2017-05-31 武汉大学 Remote sensing image road net extraction method under navigation data auxiliary

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8396257B2 (en) * 2008-07-25 2013-03-12 Navteq B.V. End user image open area maps
CN105069824A (en) * 2015-08-11 2015-11-18 中南大学 GPS data based automatic construction method and system for open-pit mine road network
CN106528740A (en) * 2016-11-04 2017-03-22 中科宇图科技股份有限公司 Delaunay triangular net-based road center line extraction method
CN106778605A (en) * 2016-12-14 2017-05-31 武汉大学 Remote sensing image road net extraction method under navigation data auxiliary

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
杨伟 等: "基于众源轨迹数据的道路中心线提取", 《地理与地理信息科学》 *
杨伟 等: "运用约束Delaunay三角网从众源轨迹线提取道路边界", 《测绘学报》 *

Cited By (27)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109684384A (en) * 2018-12-29 2019-04-26 太原科技大学 A kind of track data space-time density Analysis System and its analysis method
CN109934865B (en) * 2019-03-12 2020-11-27 重庆市勘测院 Extraction method of road network centerline
CN109934865A (en) * 2019-03-12 2019-06-25 重庆市勘测院 Extraction method of road network centerline
CN110008872A (en) * 2019-03-25 2019-07-12 浙江大学 A road network extraction method combining vehicle trajectories and remote sensing images
CN110008872B (en) * 2019-03-25 2021-04-06 浙江大学 Road network extraction method combining vehicle track and remote sensing image
CN110096535A (en) * 2019-05-16 2019-08-06 北京航空航天大学 A kind of traffic elasticity communication mode method for digging based on elastic triangle theory
CN110096535B (en) * 2019-05-16 2021-06-29 北京航空航天大学 A traffic elastic propagation mode mining method based on elastic triangle theory
CN110633342A (en) * 2019-07-29 2019-12-31 武汉光庭信息技术股份有限公司 Lane topology network generation method
CN110580388A (en) * 2019-08-28 2019-12-17 北方工业大学 A Channel Network Extraction Method Based on Crowd-source Trajectory Data
CN110580388B (en) * 2019-08-28 2023-04-07 北方工业大学 Channel network extraction method based on crowd-sourced trajectory data
CN110688958A (en) * 2019-09-28 2020-01-14 中国测绘科学研究院 Intersection Recognition Method Based on GoogLeNet Neural Network
CN110988880A (en) * 2019-12-12 2020-04-10 南京莱斯电子设备有限公司 Geographic information extraction and target tracking method based on SMR target track
CN111833630A (en) * 2019-12-31 2020-10-27 北京嘀嘀无限科技发展有限公司 Method and device for determining data release position, storage medium and electronic equipment
CN111426329A (en) * 2020-03-12 2020-07-17 深圳小库科技有限公司 Road generation method and device, computer equipment and storage medium
CN111426329B (en) * 2020-03-12 2021-10-22 深圳小库科技有限公司 Road generation method and device, computer equipment and storage medium
CN112013862A (en) * 2020-07-31 2020-12-01 深圳大学 Pedestrian network extraction and updating method based on crowdsourcing trajectory
CN112013862B (en) * 2020-07-31 2022-06-17 深圳大学 A Crowdsourcing Trajectory-Based Pedestrian Network Extraction and Update Method
US11869348B2 (en) 2020-08-01 2024-01-09 Grabtaxi Holdings Pte. Ltd. Processing apparatus and method for generating route navigation data
WO2022031222A1 (en) * 2020-08-01 2022-02-10 Grabtaxi Holdings Pte. Ltd. Processing appratus and method for generating route navigation data
CN112566025A (en) * 2020-12-10 2021-03-26 南京市城市与交通规划设计研究院股份有限公司 Bus passenger flow source destination identification method and device, electronic equipment and storage medium
CN112566025B (en) * 2020-12-10 2021-10-01 南京市城市与交通规划设计研究院股份有限公司 Bus passenger flow source destination identification method and device, electronic equipment and storage medium
CN113008251B (en) * 2021-02-22 2022-11-25 湖南大学 Digital map updating method for unstructured roads in closed area
CN113008251A (en) * 2021-02-22 2021-06-22 湖南大学 Digital map updating method for unstructured roads in closed area
CN112598724A (en) * 2021-03-01 2021-04-02 武大吉奥信息技术有限公司 Improved TIN-based planar element center line extraction method
CN113052084A (en) * 2021-03-26 2021-06-29 中国地质大学(武汉) Community-level vector road network extraction method based on mobile phone positioning data
CN117891892A (en) * 2024-03-14 2024-04-16 南京师范大学 A ranger trajectory optimization algorithm based on historical patrol data
CN117891892B (en) * 2024-03-14 2024-05-31 南京师范大学 Forest guard track optimization method based on historical patrol data

Also Published As

Publication number Publication date
CN108920481B (en) 2020-11-27

Similar Documents

Publication Publication Date Title
CN108920481A (en) A kind of road network method for reconstructing and system based on mobile phone location data
CN107918753B (en) Point cloud data processing method and device
CN109635748B (en) Method for extracting road characteristics in high-resolution image
CN109189949B (en) A kind of population distribution calculation method
Tang et al. Exploring urban travel patterns using density-based clustering with multi-attributes from large-scaled vehicle trajectories
WO2023109720A1 (en) Urban public transportation planning method
CN109272170A (en) A kind of traffic zone dividing system based on Louvain algorithm
CN117173895B (en) Management monitoring system for automatic illumination adjustment of urban road
CN109447160A (en) A kind of method of image and vector road junction Auto-matching
CN113932821B (en) Track map matching method based on continuous window average direction characteristics
CN108921173A (en) A kind of deep learning method of combination OSM and remote sensing image extraction overpass
CN109657614B (en) Automatic road identification method in aerial photography road traffic accident scene investigation
CN111539432A (en) Method for extracting urban road by using multi-source data to assist remote sensing image
CN103699654A (en) Cross-measuring-scale vector map water network same name target matching method
CN106971535A (en) A kind of urban traffic blocking index calculating platform based on Floating Car GPS real time datas
CN114861277A (en) Long-time-sequence national soil space function and structure simulation method
CN116663724A (en) A Method and Device for Predicting and Visualizing Carbon Emissions of Motor Vehicles on Urban Streets
CN109165245A (en) The motion track of multisource data fusion generates the spatiotemporal mode method for digging of model
CN111627103A (en) Smart city CIM imaging method based on pedestrian activity and density perception
CN106504219A (en) Constrained path morphology high-resolution remote sensing image road Enhancement Method
CN109800903A (en) A kind of profit route planning method based on taxi track data
CN112905856A (en) Method for constructing high-speed traffic data set with space-time dependence
CN115130263B (en) Multi-traffic facility equal time circle calculation method oriented to homeland space planning
CN114996544B (en) Travel path extraction method based on mobile phone signaling data and Mix-Markov model
CN111931968B (en) A method of applying mobile phone GPS data to optimize the layout of public bicycle stations

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
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20201127

CF01 Termination of patent right due to non-payment of annual fee