使用底图和Matplotlib绘制时间序列轮廓数据

Question

I'm new to basemap and python but I'm trying to build a plotter for weather model for daily cron job. We plot about 1000 images a day.

I've wrote some script to achieve what i want. But it took very long time, because it re-draw the basemap for every time-step. It took 30 seconds to draw the basemap, and only took 4 seconds to draw the contourf().

I have some idea to speed up the process by pre-draw the basemap and update the contourf() in each iteration. But i don't understand how matplotlib objects work.

I have researched for the same question about this but haven't found any. But I found something similar from here from user3982706.

from matplotlib import animation
from matplotlib import pyplot as plt
import numpy as np
from mpl_toolkits.basemap import Basemap


fig, ax = plt.subplots()

# set up map projection
m = Basemap(projection='nsper',lon_0=-0,lat_0=90)
m.drawcoastlines()
m.drawparallels(np.arange(0.,180.,30.))
m.drawmeridians(np.arange(0.,360.,60.))

# some 2D geo arrays to plot (time,lat,lon)
data = np.random.random_sample((20,90,360))
lat = np.arange(len(data[0,:,0]))
lon = np.arange(len(data[0,0,:]))
lons,lats = np.meshgrid(lon,lat)

# ims is a list of lists, each row is a list of artists to draw in the
# current frame; here we are animating three artists, the contour and 2 
# annotatons (title), in each frame
ims = []
for i in range(len(data[:,0,0])):
    im = m.contourf(lons,lats,data[i,:,:],latlon=True)
    add_arts = im.collections
    text = 'title={0!r}'.format(i)
    te = ax.text(90, 90, text)
    an = ax.annotate(text, xy=(0.45, 1.05), xycoords='axes fraction')
    ims.append(add_arts + [te,an])

ani = animation.ArtistAnimation(fig, ims)
## If you have ffmpeg you can save the animation by uncommenting 
## the following 2 lines
# FFwriter = animation.FFMpegWriter()
# ani.save('basic_animation.mp4', writer = FFwriter)
plt.show()

That script save the contour data as a list of artist. I don't need animation. so I need to edit this script to save the figure inside the loop. So this script produce figure1.png, figure2.png, figure3.png, and so on.

Any suggestion?

Answer 1

(a) Using animation to save images

Because you already have the animation present in the code, you may directly save each image of the animation via the ani.save command.

ani.save("figure.png", writer="imagemagick")

This will create 20 figures called figure-0.png to figure-19.png . It requires imagemagick to be installed and available in your environment.

(b) Save individual figures

If the above fails for whatever reason, you may save the individual figures via plt.savefig() . At the end of each loop step, you would then remove the artists from the axes and delete them.

from matplotlib import pyplot as plt
import numpy as np
from mpl_toolkits.basemap import Basemap

fig, ax = plt.subplots()

m = Basemap(projection='nsper',lon_0=-0,lat_0=90)
m.drawcoastlines()
m.drawparallels(np.arange(0.,180.,30.))
m.drawmeridians(np.arange(0.,360.,60.))

data = np.random.random_sample((20,90,360))
lat = np.arange(len(data[0,:,0]))
lon = np.arange(len(data[0,0,:]))
lons,lats = np.meshgrid(lon,lat)

for i in range(len(data[:,0,0])):
    im = m.contourf(lons,lats,data[i,:,:],latlon=True)
    text = 'title={0!r}'.format(i)
    te = ax.text(90, 90, text)
    an = ax.annotate(text, xy=(0.45, 1.05), xycoords='axes fraction')
    # save the figure
    plt.savefig("figure_{}.png".format(i))
    # remove stuff from axes
    for c in im.collections:
        c.remove()
    te.remove()
    an.remove()
    del im; del te; del an

plt.show()