如何在 Shapely 多边形中舍入多边形坐标？答案

【问题标题】：How to round off polygon coordinates in Shapely polygons?如何在 Shapely 多边形中舍入多边形坐标？
【发布时间】：2020-12-14 17:11:23
【问题描述】：

我有一个使用以下代码的复杂用例：

from shapely import wkt

result_polygons = [wkt.loads('POLYGON ((-45.30 5.99, -2.63 5.99, -2.63 1.1, -45.30 1.1, -45.30 5.99))'),
                   wkt.loads('POLYGON ((-2.63 5.99, -2.63 1.1, -57.63 1.1, -57.63 5.99, -2.63 5.99))'),
                   wkt.loads('POLYGON ((-4.74 1.1, -37.94 1.1, -45.30 1.10, -45.30 5.99, -12.00 5.99, -4.74 1.1))')]
intersect = wkt.loads('POLYGON ((-4.74 1.1, -57.63 1.1, -57.63 5.99, -12.00 5.99, -4.74 1.1))')
tmp_intersect = [wkt.loads('POLYGON ((-45.30 1.1, -45.30 5.99, -12.00 5.99, -4.74 1.1, -45.30 1.1))')]

for poly in result_polygons:             # result_polygons has some polygons
    inter = intersect.intersection(poly) # inter has intersection of poly and intersect
    same_intersection = False            
    # avoid adding to list intersections that already exist
    for tmp in tmp_intersect:
        if tmp.equals(inter):
            same_intersection = True
            break
    if not same_intersection:
        tmp_intersect.append(inter)

for i in range(len(tmp_intersect)):
    print(tmp_intersect[i].wkt)

这很好用。但问题是我在多边形中没有四舍五入的点。我有以下导致错误结果的数据（由于多边形中浮点值的微小变化，相同的多边形区域存储在 tmp_intersect 值中）。

result_polygons = [wkt.loads('POLYGON ((-45.3015 5.999999999999997, -2.631 5.999999999999999, -2.631 1.1, -45.3015 1.100000000000006, -45.3015 5.999999999999997))'),
                   wkt.loads('POLYGON ((-2.631 5.999999999999999, -2.631 1.1, -57.631 1.100000000000007, -57.631 5.999999999999996, -2.631 5.999999999999999))'),
                   wkt.loads('POLYGON ((-4.740353646471659 1.1, -37.94482693640743 1.100000000000005, -45.3015 1.100000000000006, -45.3015 5.999999999999997, -12.00490239218409 5.999999999999998, -4.740353646471659 1.1))')]
intersect = wkt.loads('POLYGON ((-4.740353646471659 1.1, -57.631 1.100000000000007, -57.631 5.999999999999996, -12.00490239218409 5.999999999999998, -4.740353646471659 1.1))')
tmp_intersect = [wkt.loads('POLYGON ((-45.3015 1.100000000000006, -45.3015 5.999999999999997, -12.00490239218409 5.999999999999998, -4.740353646471659 1.1, -45.3015 1.100000000000006))')]

有没有办法在我提取多边形点时对其进行四舍五入？

【问题讨论】：

标签： python-2.7 polygon shapely

【解决方案1】：

我想，我已经找到了解决方案。可以使用https://gis.stackexchange.com/a/276512/173433中提到的以下代码来完成：

shapely.wkt.loads(shapely.wkt.dumps(geom, rounding_precision=n))

其中 rounding_precision=n 是您希望坐标四舍五入的第 n 个小数点。

所以在下面的代码中使用这个代码对我来说是诀窍：

from shapely import wkt

result_polygons = [wkt.loads('POLYGON ((-45.3015 5.999999999999997, -2.631 5.999999999999999, -2.631 1.1, -45.3015 1.100000000000006, -45.3015 5.999999999999997))'),
                   wkt.loads('POLYGON ((-2.631 5.999999999999999, -2.631 1.1, -57.631 1.100000000000007, -57.631 5.999999999999996, -2.631 5.999999999999999))'),
                   wkt.loads('POLYGON ((-4.740353646471659 1.1, -37.94482693640743 1.100000000000005, -45.3015 1.100000000000006, -45.3015 5.999999999999997, -12.00490239218409 5.999999999999998, -4.740353646471659 1.1))')]
intersect = wkt.loads('POLYGON ((-4.740353646471659 1.1, -57.631 1.100000000000007, -57.631 5.999999999999996, -12.00490239218409 5.999999999999998, -4.740353646471659 1.1))')
tmp_intersect = [wkt.loads('POLYGON ((-45.3015 1.100000000000006, -45.3015 5.999999999999997, -12.00490239218409 5.999999999999998, -4.740353646471659 1.1, -45.3015 1.100000000000006))')]


round_rp = []
for temp in result_polygons:
    round_rp.append(wkt.loads(wkt.dumps(temp, rounding_precision=5)))


round_ti = []
for temp in tmp_intersect:
    round_ti.append(wkt.loads(wkt.dumps(temp, rounding_precision=5)))

for poly in round_rp:
    inter = shape(intersect).intersection(poly)
    same_intersection = False
    # avoid adding to list intersections that already exist
    for tmp in round_ti:
        if tmp.equals(inter):
            same_intersection = True
            break
    if not same_intersection:
        round_ti.append(inter)

for i in range(len(round_ti)):
    print(round_ti[i].wkt)

【讨论】：