使用pgrouting和geotools实现最短路径,服务区分析
1本文主要讲解服务区分析的实现(最优路径已经有很多文章了)
设施服务范围指在一定限制条件下(如时间、费用或路程等)设施所能提供服务的最大空间领域, 在道路网络环境中,它通常由一系列结点及边组成。例如, 某救助站在接到求救电话后10 min 所能到达的区域;某物流公司在配送货物时500元花费所能到达的区域等。
(1)根据拓扑关系,计算地理网络的最大邻接结点数;
(2)构造邻接结点矩阵和初始判断矩阵描述地理网络结构;
(3)应用广度优先搜索算法确定地理网络中心服务范围。
本算法是对Dijkstra最短路径算法的改进(简称“最短路径算法”)。首先, 将网络中所有结点初始化为未标记结点。然后从起点(第一次搜索的起点为网络中心)开始搜索与其有路径连通的未标记结点, 计算阻值, 并将起点标记为已标记结点, 重复上述过程, 直到某结点的阻值超过网络中心的阻值。最后, 基于结点及边的阻值搜索并存储所有在中心阻值范围内的边, 这些边和结点的集合为网络中心的服务范围。
(但实际情况中可能需要内插一些点,直到找到阻值等于网络中心的阻值为止)
2实现过程:<1>数据读取:直接读取shp
//1读取shp文件,得到pgDatastore
public static void conShp(String path){
try {
File file=new File(path);
Map<String, Object> map = new HashMap<String, Object>();
map.put("url", file.toURI().toURL());
System.out.println(map);
pgDatastore = DataStoreFinder.getDataStore(map);
} catch (Exception e) {
e.printStackTrace();
}
}
从postgis中读取
首先读取postgis数据库得到DataStore对象,然后用getfeaturesource(LayerName)得到SimpleFeatureSource即可(注意:这里的LayerName即为表名)
//2读取postgis,得到pgDatastore
//链接postgis
public static void conPostGis(String dbtype, String host, String port,
String database, String userName, String password) {
Map<String, Object> params = new HashMap<String, Object>();
params.put(PostgisNGDataStoreFactory.DBTYPE.key, dbtype);
params.put(PostgisNGDataStoreFactory.HOST.key, host);
params.put(PostgisNGDataStoreFactory.PORT.key, new Integer(port));
params.put(PostgisNGDataStoreFactory.DATABASE.key, database);
params.put(PostgisNGDataStoreFactory.SCHEMA.key, "public");
params.put(PostgisNGDataStoreFactory.USER.key, userName);
params.put(PostgisNGDataStoreFactory.PASSWD.key, password);
try {
pgDatastore = DataStoreFinder.getDataStore(params);
if (pgDatastore != null) {
System.out.println("系统连接到位于:" + host + "的空间数据库" + database
+ "成功!");
} else {
System.out.println("系统连接到位于:" + host + "的空间数据库" + database
+ "失败!请检查相关参数");
}
} catch (IOException e) {
e.printStackTrace();
System.out.println("系统连接到位于:" + host + "的空间数据库" + database
+ "失败!请检查相关参数");
}
}
//3利用pgDatastore,得到featuresource(表)
public static SimpleFeatureSource getFeatureSource(String LayerName) throws IOException{
if(pgDatastore==null){
System.out.println("还未导入数据源,请导入pgDatastore");
return null;
}
featureSource = pgDatastore.getFeatureSource(LayerName);
System.out.println(featureSource.getCount(Query.ALL));
return featureSource;
}
注意事项:
读取postgis时,数据库里面的geom字段不能为二进制
读取文件时,文件中最好不要有中文
<2>进行拓扑将数据处理为Graph
(1)得到SimpleFeatureCollection
(2)创建一个FeatureGraphGenerator利用它添加SimpleFeature元素并调用其getGraph方法创建Graph
(3)创建出来的Graph中保存着V(节点)和E(边),这样就可以进行网络分析了
//创建graph
public static Graph getGraph(SimpleFeatureSource source) throws IOException{
if(source==null)
{
System.out.println("资源不存在,请先得到featureSource");
return null;
}
SimpleFeatureCollection fCollection = source.getFeatures();
//create a linear graph generate
//构图时也可以创建一个DirectedLineStringGraphGenerator构建有向图
LineStringGraphGenerator lineStringGen = new LineStringGraphGenerator();
//wrap it in a feature graph generator
FeatureGraphGenerator featureGen = new FeatureGraphGenerator( lineStringGen );
//throw all the features into the graph generator
FeatureIterator<SimpleFeature> iter = fCollection.features();
try {
while(iter.hasNext()){
Feature feature = iter.next();
featureGen.add(feature);
}
} finally {
iter.close();
}
graph = featureGen.getGraph();
return graph;
}
<3>最短路径
(1)最短路径:
使用Astar算法:
1.首先利用AstarFunctions设定权值(即通过此边的消耗)
2.然后设定start点(起点)和target点(终点)
3.调用AstarShortestFinder()来进行处理
具体代码如下:
设定权(成本):
public static double discost(Edge e ){
SimpleFeature feature = (SimpleFeature) e.getObject();
Geometry geom = (Geometry) feature.getDefaultGeometry();
//geom.convexHull()将其构成一个图形
if(Barriers!=null){
for(int i=0;i<Barriers.size();i++){
Geometry g=Barriers.get(i);
if(geom.intersects(g)){
return Double.POSITIVE_INFINITY;
}
}
}
return geom.getLength();
}
public static double discost(AStarNode n1, AStarNode n2){
Node nd1=n1.getNode();
Node nd2=n2.getNode();
Edge e=nd1.getEdge(nd2);
if(e!=null){
SimpleFeature feature=(SimpleFeature)e.getObject();
Geometry geom=(Geometry) feature.getDefaultGeometry();
if(Barriers!=null){
for(int i=0;i<Barriers.size();i++){
Geometry g=Barriers.get(i);
if(geom.intersects(g)){
return Double.POSITIVE_INFINITY;
}
}
}
return ((Point) n1.getNode().getObject())
.distance((Point)n2.getNode().getObject());
}else{
return ((Point) n1.getNode().getObject())
.distance((Point)n2.getNode().getObject());
}
}
//Astar方法的最短路径计算
public static Path searchRouteByAstar(Node source,Node destination) throws Exception{
if(graph==null){
System.out.println("graph不存在,请构建graph");
return null;
}
if(source.equals(destination)){
System.out.println("起点和终点相同,请重新选点");
return null;
}
Path path=null;
AStarFunctions afuncs=new AStarFunctions(destination) {
@Override
public double h(Node n) {
//结合Astar的算法可以知道这里的h指的是一个预估的距离destination的消耗值
//disPoint指的是预估的终点
Point disPoint=(Point)this.getDest().getObject();
return ((Point)n.getObject()).distance(disPoint);
}
@Override
public double cost(AStarNode n1, AStarNode n2) {
//注意矢量性和有向性
return discost(n1, n2);
}
};
AStarShortestPathFinder finder=new AStarShortestPathFinder(graph, source, destination, afuncs);
finder.calculate();
path=finder.getPath();
return path;
}
这里是可以看到传入的变量是node节点,但是我们实际中是要在地图上点击一个起点终点求出最优路径,因此还需要将鼠标点击的任意一点归算的graph的节点里去,这里最好使用数据库空间查询来算,本文只是用了最简单的遍历,算法如下:
//搜寻graph上最近节点的方法
//暂时先采用遍历的方法
//这里如果点隔的太远会直接把pointy输出,调用最短路径算法会抛出空指针异常
public static Node getNearestGraphNode(Point pointy){
if(graph==null){
System.out.println("graph不存在,请构建graph");
return null;
}
double dist=0;
Node nearestNode=null;
for(Object o:graph.getNodes()){
Node n=(Node)o;
Point gPoint=(Point)n.getObject();
double distance=gPoint.distance(pointy);
if(nearestNode==null||distance<dist){
dist=distance;
nearestNode=n;
}
}
return nearestNode;
}
归算到节点之后就可以改造下Astar算法了:
public static Path searchRouteByAstar(Point startPoint,Point endPoint) throws Exception{
if(graph==null){
System.out.println("graph不存在,请构建graph");
return null;
}
Node source=getNearestGraphNode(startPoint);
Node destination=getNearestGraphNode(endPoint);
if(source.equals(destination)){
System.out.println("起点和终点相同,请重新选点");
return null;
}
Path path=null;
AStarFunctions afuncs=new AStarFunctions(destination) {
@Override
public double h(Node n) {
//结合Astar的算法可以知道这里的h指的是一个预估的距离destination的消耗值
//disPoint指的是预估的终点
Point disPoint=(Point)this.getDest().getObject();
return ((Point)n.getObject()).distance(disPoint);
}
@Override
public double cost(AStarNode n1, AStarNode n2) {
//注意矢量性和有向性
return discost(n1, n2);
}
};
AStarShortestPathFinder finder=new AStarShortestPathFinder(graph, source, destination, afuncs);
finder.calculate();
path=finder.getPath();
return path;
}
这样看起来就挺完美了,但是如果要加入障碍点怎么办那?
其实我们在成本计算中已经考虑障碍物了,如果是个障碍范围就与当前的graph求交集,交集处的权设置成无穷就好了,这样就解决了障碍点的问题。
如果是停靠点那?
那就每段都计算一次最优路径加起来就行了。
使用Dijkstra算法:
1.首先利用Edgeweighter设定权值(即通过此边的消耗)
2.然后设定start点(起点)和target点(终点)
3.调用DijkstraShortestPathFinder()来进行处理
dijkstra算法大概差不多,直接贴代码:
//dijkstra方法
public static Path searchRouteByDijkstra(Node source,Node destination) throws Exception{
if(graph==null){
System.out.println("graph不存在,请构建graph");
return null;
}
Path path=null;
EdgeWeighter weighter = new EdgeWeighter() {
@Override
public double getWeight(Edge e) {
return discost(e);
}
};
// Create GraphWalker - in this case DijkstraShortestPathFinder
DijkstraShortestPathFinder pf = new DijkstraShortestPathFinder( graph, source, weighter );
pf.calculate();
path= pf.getPath(destination);
return path;
}
public static Path searchRouteByDijkstra(Point startPoint,Point endPoint) throws Exception{
if(graph==null){
System.out.println("graph不存在,请构建graph");
return null;
}
Node source=getNearestGraphNode(startPoint);
Node destination=getNearestGraphNode(endPoint);
Path path=null;
EdgeWeighter weighter = new EdgeWeighter() {
@Override
public double getWeight(Edge e) {
return discost(e);
}
};
// Create GraphWalker - in this case DijkstraShortestPathFinder
DijkstraShortestPathFinder pf = new DijkstraShortestPathFinder( graph, source, weighter );
pf.calculate();
path= pf.getPath(destination);
return path;
}
<4>服务区分析
改造DijkstraShortestPathFinder方法:
1.首先通过Edgeweighter设定权值(即通过此边的消耗)
2.然后设定start点(起点)
3.最后通过设置一个判定(该判定可能是根据距离也可能是根据时间)来终止该方法的搜索,然后得到该方法返回的所有边和节点。
public static List<Point> getAdjancyPoint(Node node){
if(graph==null){
System.out.println("graph不存在,请构建graph");
return null;
}
List<Point> points=new ArrayList<Point>();
Point pt=(Point)node.getObject();
System.out.println("传入的节点:"+pt);
List<Edge> edges=node.getEdges();
for(Edge e:edges){
Node nodeA=e.getNodeA();
Point pa=(Point)nodeA.getObject();
Node nodeB=e.getNodeB();
Point pb=(Point)nodeB.getObject();
if(!pt.equals(pa)){
points.add(pa);
}else if(!pt.equals(pb)){
points.add(pb);
}
}
List<Point>points1=(List<Point>) CollectionUtils.subtract(points,serviceAreaPoints);
System.out.println("加入的临近点:"+points1);
return points1;
}
//服务区范围,目前我只是把节点加入进去
public static void ServiceArea(Point startPoint, double cost) throws Exception{
if(graph==null){
System.out.println("graph不存在,请构建graph");
return;
}
Node source=getNearestGraphNode(startPoint);
Point pt=(Point)source.getObject();
serviceAreaPoints.add(pt);
//其实递归应该从这里开始,前面的不用递归
List<Point> pts=getAdjancyPoint(source);
for(Iterator<?>itr=pts.iterator();itr.hasNext();){
Point p=(Point)itr.next();
if(p!=null){
Geometry geo=iterRoute(searchRouteByAstar(serviceAreaPoints.get(0), p)).getRoutePath();
double len=geo.getLength();
if(len<=cost){
ServiceArea(p, cost);
System.out.println("点"+p+"加人serviceArea");
}
else{
System.out.println("点"+p+"不加人serviceArea");
}
}
}
}
//获得服务区点集合
public static Set<Point> getServiceAreaPoints() {
serviceAreaPoints1.clear();
serviceAreaPoints1.addAll(serviceAreaPoints);
return serviceAreaPoints1;
}
这样就完成了服务范围分析。
有什么问题欢迎大家评论与交流。
转载自:https://blog.csdn.net/zhgu1992/article/details/78883993