python - Cartopy 绘图点与正交投影不正确

Question

我正在尝试使用 Cartopy 和 Anaconda Python 绘制 map 点，但在转换时遇到了一些奇怪的失败。在我的简单示例中，我试图绘制 3 个点，但它们正在加倍。

import matplotlib.pyplot as plt
import cartopy.crs as ccrs

lons = [214.5, 2.7, 197.5]                                                                                               
lats = [35, 36, 37.]

ax = plt.axes(projection=ccrs.Orthographic(
    central_longitude=0,
central_latitude=90))
# plot lat/lon points                                                                                                         
ax.plot(lons, lats, 'ro',
        transform=ccrs.Geodetic())
# plot north pole for reference                                                                                               
ax.plot([0], [90], 'b^',
    transform=ccrs.Geodetic())
# add coastlines for reference                                                                                                
ax.coastlines(resolution='50m')
ax.set_global()

plt.show()

测试:

cartopy==0.16.0 和 Cartopy-0.16.1.dev179-

proj4==4.9.3, proj4==5.0.1, proj4==5.0.2

我唯一的提示是，使用 Cartopy-0.16.1.dev179- 和 proj4==5.0.1，我得到了这个 UserWarning :

/Users/***/anaconda3/lib/python3.6/site-packages/cartopy/crs.py:1476: UserWarning: The Orthographic projection in Proj between 5.0.0 and 5.1.0 incorrectly transforms points. Use this projection with caution.

我在 https://github.com/SciTools/cartopy/issues/1172 上开了一个问题但问题已关闭。任何人都知道如何让 cartopy 与正交投影一起正常工作？

Answer 1

据我所知，您可以使用多种方法来获得预期的结果。

首先，显式转换要在原生投影中的点...

import matplotlib.pyplot as plt
import cartopy.crs as ccrs

# create the lat/lon points
lons = np.array([214.5, 2.7, 197.5])
lats = np.array([35, 36, 37.])

# create the projections
ortho = ccrs.Orthographic(central_longitude=0, central_latitude=90)
geo = ccrs.Geodetic()

# create the geoaxes for an orthographic projection
ax = plt.axes(projection=ortho)

# transform lat/lons points to othographic points
points = ortho.transform_points(geo, lons, lats)

# plot native orthographic points                                                                                
ax.plot(points[:, 0], points[:, 1], 'ro')

# plot north pole for reference (with a projection transform)                                                                                           
ax.plot([0], [90], 'b^', transform=geo)

# add coastlines for reference                                                                                                
ax.coastlines(resolution='50m')
ax.set_global()

这个情节符合预期......

Expected Orthographic projection plot

您看到的最初问题是 cartopy 试图将点序列解释为有界几何(或路径)，但有点困惑。将纬度/经度点显式转换为原生正交点可避免此子弹。

了解这些信息后，我们可以选择调用适当的方法，将点列表视为单个点(并避免 cartopy 做出不符合我们期望的假设)，方法是使用 scatter 而不是 plot...

import matplotlib.pyplot as plt
import cartopy.crs as ccrs

# create the lat/lon points
lons = np.array([214.5, 2.7, 197.5])
lats = np.array([35, 36, 37.])

# create the projections
ortho = ccrs.Orthographic(central_longitude=0, central_latitude=90)
geo = ccrs.Geodetic()

# create the geoaxes for an orthographic projection
ax = plt.axes(projection=ortho)

# plot native orthographic scatter points                                                                                
ax.scatter(lons, lats, marker='o', c='r', transform=geo)

# plot north pole for reference                                                                                               
ax.plot([0], [90], 'b^', transform=geo)

# add coastlines for reference                                                                                                
ax.coastlines(resolution='50m')
ax.set_global()

这也适用于我。

HTH