我一直在尝试调试此问题,但无法执行此操作。我试图找到两个Polygon
对象的交集。它大部分时间都有效,但对于以下情况,它引发了以下异常:
P1 area: 13.125721955
P2 area: 1.0
Traceback (most recent call last):
File "geom2d.py", line 235, in <module>
print p1.intersection(p2)
File "/usr/local/lib/python2.7/dist-packages/shapely/geometry/base.py", line 334, in intersection
return geom_factory(self.impl['intersection'](self, other))
File "/usr/local/lib/python2.7/dist-packages/shapely/topology.py", line 47, in __call__
"The operation '%s' produced a null geometry. Likely cause is invalidity of the geometry %s" % (self.fn.__name__, repr(this)))
shapely.geos.TopologicalError: The operation 'GEOSIntersection_r' produced a null geometry. Likely cause is invalidity of the geometry <shapely.geometry.polygon.Polygon object at 0x8e5ad6c>
代码如下。
from shapely.geometry import Point,Polygon,MultiPolygon
poly1 = [(35.0041000000000011, -88.1954999999999956), (34.9917999999999978, -85.6068000000000069), (32.8404000000000025, -85.1756000000000029), (32.2593000000000032, -84.8927000000000049), (32.1535000000000011, -85.0341999999999985), (31.7946999999999989, -85.1358000000000033), (31.5199999999999996, -85.0438000000000045), (31.3384000000000000, -85.0836000000000041), (31.2092999999999989, -85.1069999999999993), (31.0023000000000017, -84.9943999999999988), (30.9953000000000003, -87.6008999999999958), (30.9422999999999995, -87.5926000000000045), (30.8538999999999994, -87.6256000000000057), (30.6744999999999983, -87.4072000000000031), (30.4404000000000003, -87.3687999999999931), (30.1463000000000001, -87.5240000000000009), (30.1545999999999985, -88.3863999999999947), (31.8938999999999986, -88.4742999999999995), (34.8937999999999988, -88.1020999999999930), (34.9478999999999971, -88.1721000000000004), (34.9106999999999985, -88.1461000000000041)]
poly2 = [(34.7998910000000024, -88.2202139999999986), (34.7998910000000024, -87.2202139999999986), (35.7998910000000024, -87.2202139999999986), (35.7998910000000024, -88.2202139999999986)]
p1 = Polygon(poly1)
p2 = Polygon(poly2)
print 'P1 area:',p1.area
print 'P2 area:',p2.area
print p1.intersection(p2)
由于它打印了两个多边形的区域,我假设多边形是正确形成的。我也(以某种方式)打印了第一个多边形,以确保它确实是一个简单的多边形。
有人可以解释为什么我会收到这个例外吗?
编辑:我打印了p1.is_valid,结果证明是假的。有一些解释here。搜索字符串is_valid
。它说
有效的Polygon可能没有任何重叠的外部或内部环。
有人可以解释这意味着什么,以及是否有可能的解决方法?
顺便说一句,我也注意到,如果我从poly1
删除最后一个坐标,整个过程就可以了。也许整个坐标列表使多边形复杂。
答案 0 :(得分:14)
如前所述,p1
无效。在绘制它时,我注意到右下角有一个“领结”。我假设你的多边形不需要这个;如果没有,你可以尝试使用Shapely的buffer(0)
技巧(在Shapely手册中记录)来解决这个问题:
In [382]: p1.is_valid
Out[382]: False
In [383]: p1 = p1.buffer(0)
In [384]: p1.is_valid
Out[384]: True
buffer(0)
具有以下效果:
在:
后:
现在你可以这样做:
print p1.intersection(p2)
POLYGON ((34.9396324323625151 -88.1614025927056559, 34.8937999999999988 -88.1020999999999930, 34.7998910000000024 -88.1137513649788247, 34.7998910000000024 -87.2202139999999986, 34.9994660069532983 -87.2202139999999986, 35.0041000000000011 -88.1954999999999956, 34.9396324323625151 -88.1614025927056559))
请注意,我有一些情况(区域看起来更像是“鸟巢”,而不是简单的鞠躬),这种情况不起作用;检查以确保您返回单个Polygon
对象,而不是MultiPolygon
对象。
答案 1 :(得分:6)
您收到此异常是因为p1
不是有效的多边形。
>>> p1.is_valid
False
>>> p2.is_valid
True
有效的Polygon可能不具有任何重叠的外部或内部 环。
请记住,由于多边形的第一个和最后一个点的形状不同,因此将第一个点附加到列表的末尾。
>>> list(p1.exterior.coords)
[(35.004100000000001, -88.195499999999996), (34.991799999999998, -85.606800000000007), (32.840400000000002, -85.175600000000003), (32.259300000000003, -84.892700000000005), (32.153500000000001, -85.034199999999998), (31.794699999999999, -85.135800000000003), (31.52, -85.043800000000005), (31.3384, -85.083600000000004), (31.209299999999999, -85.106999999999999), (31.002300000000002, -84.994399999999999), (30.9953, -87.600899999999996), (30.942299999999999, -87.592600000000004), (30.853899999999999, -87.625600000000006), (30.674499999999998, -87.407200000000003), (30.4404, -87.368799999999993), (30.1463, -87.524000000000001), (30.154599999999999, -88.386399999999995), (31.893899999999999, -88.474299999999999), (34.893799999999999, -88.102099999999993), (34.947899999999997, -88.1721), (34.910699999999999, -88.146100000000004), (35.004100000000001, -88.195499999999996)]
多边形的外部是一个线性环,它似乎也是无效的:
>>> p1.exterior.type
'LinearRing'
>>> p1.exterior.is_valid
False
您还可以看到,如果您要将多边形的外部转换为线串,那将很复杂:
>>> l1 = LineString(p1.exterior.coords)
>>> l1.is_simple
False
多边形的外部以某种方式交叉或接触自己。
在数据中挖掘更多内容,我们可以在地图上将其可视化:
>>> import cgpolyencode
>>> encoder = cgpolyencode.GPolyEncoder()
>>> encoder.encode((y, x) for x, y in p1.exterior.coords)
{'points': 'svstEzthyOzkAkrxNfecL_fsAznpBcgv@ftSjsZnaeA~yRzst@_~P~mb@vwFzeXfqCvlg@w~Tvj@ra|NfjI{r@ngPfmEf`b@_ti@bvl@_oFbmx@~h]{r@~lgDsurIjdPk|hQgugAaqIntLlgFoaDwfQvsH', 'numLevels': 18, 'zoomFactor': 2, 'levels': 'PPLMKMKGKPNIKLMNPLLKJP'}
如果你将其插入Google's Polyline Encoder,你可以看到它是阿拉巴马州。如果放大到阿拉巴马州的左上角,你可以看到两条线相互交叉。要解决此问题,您需要删除最后一个点或将最后一个点与第二个点交换到最后一个点。