Merge or remove neighbouring polygons#

Based on a set of conditions.

import geopandas
from shapely.geometry import box, Polygon
from geoplanar import merge_touching

Merge or remove polygons based on a set of conditions.
If polygon does not share any boundary with another polygon, remove. If it shares some boundary with
a neighbouring polygon, join to that polygon. If ``largest=None`` it picks one randomly, otherwise it 
picks the polygon with which it shares the largest (True) or the smallest (False) boundary.

Example 1: remove or merge polygons under a certain area threshold#

p1 = box(0, 0, 1, 1)
p2 = box(1, 0, 11, 10)
p3 = box(15, 0, 25, 10)
p4 = box(0, 15, 1, 16)
gdf = geopandas.GeoDataFrame(geometry=[p1, p2, p3, p4])
index = [0, 3]

gdf
geometry
0 POLYGON ((1.00000 0.00000, 1.00000 1.00000, 0....
1 POLYGON ((11.00000 0.00000, 11.00000 10.00000,...
2 POLYGON ((25.00000 0.00000, 25.00000 10.00000,...
3 POLYGON ((1.00000 15.00000, 1.00000 16.00000, ... 
gdf.plot(edgecolor='k')

<Axes: >
_images/ea271aeb117a6c314223abb9e9ae9169a717e6f7e58e58b1ef4079fb025c07ac.png

Find index of polygons with an area of less than 5#

gdf[gdf.area < 5]
geometry
0 POLYGON ((1.00000 0.00000, 1.00000 1.00000, 0....
3 POLYGON ((1.00000 15.00000, 1.00000 16.00000, ... 
index = gdf[gdf.area < 5].index

index

gdf1 = merge_touching(gdf,index)

gdf1.plot(edgecolor='k')

<Axes: >
_images/1ecfa714965c9d4467c85239f71338193ee4c9cd055441cf29d0618e6fdeb9e3.png

Example 2: remove or merge polygons of shape triangle#

p1 = Polygon( [[0, 0], [10,0], [10,10], [0,10] ] )
p2 = Polygon( [(10, -0), (10,-7), (14,-10), (14,0)] )
p3 = Polygon( [(10, 0), (10,5), (15,0)] )
gdf=geopandas.GeoDataFrame(geometry=[p1,p2,p3])

gdf.plot(edgecolor='k')

<Axes: >
_images/cc3fc34fab77ba6d8353bcfea4a1825a2f0d631bae2ae64d0537f9162bc42798.png 
# triangles have 4 points in their exterior (first and last are the same so that the polygon is closed)
gdf['geometry'].apply(lambda x: len(x.exterior.coords))

0    5
1    5
2    4
Name: geometry, dtype: int64

gdf[gdf['geometry'].apply(lambda x: len(x.exterior.coords)) == 4]
geometry
2 POLYGON ((10.00000 0.00000, 10.00000 5.00000, ... 
index = gdf[gdf['geometry'].apply(lambda x: len(x.exterior.coords)) == 4].index

index

Index([2], dtype='int64')

gdf1 = merge_touching(gdf,index)

gdf1.plot(edgecolor='k')

<Axes: >
_images/1fa91fcb6d9e23b6dede4ff2fd8c17dbfdc92369d364add3a10eab8064ed8bda.png

Set largest = True to merge with the neighbour with the longest shared boundary

gdf1 = merge_touching(gdf,index,largest=True)

gdf1.plot(edgecolor='k')

<Axes: >
_images/f0775e145d557361b08fd4e5635932d3625110ecd95754159eed4aa06c81cd6d.png 
gdf1 = merge_touching(gdf,index,largest=False)

gdf1.plot(edgecolor='k')

<Axes: >
_images/1fa91fcb6d9e23b6dede4ff2fd8c17dbfdc92369d364add3a10eab8064ed8bda.png