動畫Python Matplotlib圖中的更新速度較慢。我如何讓它更快？

Question

我正在使用matplotlib來運行實時船模擬器。下面的代碼是一個非常短的代碼來說明我使用matplotlib創建動畫圖的方法。代碼在特定座標上繪製一個簡單的船並旋轉它。

我應該使用什麼方法，如果我想這個繪圖渲染速度比14fps更快？例如，在matplotlib中有沒有獲取GPU渲染的方法？運行該腳本

import time 
import math 

from matplotlib import pyplot as plt 
import matplotlib as mpl 
from matplotlib.patches import Polygon 

# boat dimensions 
l = 10.0 #m 
w = 3.0 #m 
b = 2.0 #m 

fig = plt.figure(figsize=(3,3)) 
subplot_def = 111 
ax = fig.add_subplot(subplot_def) 
plt.ion() # set plot to animated 
fig.canvas.draw() 

plt.show(block=False) 

prev_time = time.time() # for getting fps 
simulation_start_time = time.time() # how long the simulation has been running 
while True: 
    time_debug = time.time() 
    current_time = time.time() - simulation_start_time 

    # set boat coordinates 
    boat_x = 100*math.sin(current_time/100.0) 
    boat_y = 10*math.sin(current_time/100.0) 
    boat_z = current_time/2*math.pi 

    # rotate the boat 
    ts = ax.transData 
    tr = mpl.transforms.Affine2D().rotate_around(boat_x, boat_y, boat_z) 
    t = tr + ts 
    print("transform: {:.3f} ms".format((time.time() - time_debug)*1000), end=', ') 
    time_debug = time.time() 

    # clear the previous plot 
    ax.clear() 
    print("clear: {:.3f} ms".format((time.time() - time_debug)*1000), end=', ') 
    time_debug = time.time() 

    # add boat 
    boat1 = Polygon(
     [[-w/2.0 + boat_x, -l/2.0 + boat_y], [0 + boat_x, l/2.0 + boat_y], [w/2.0 + boat_x, -l/2.0 + boat_y]], 
     closed=True, color='lightsteelblue', transform=t 
    ) 
    ax.add_artist(boat1) 

    # set plot limits 
    ax.set_xlim([boat_x - 10, boat_x + 10]) 
    ax.set_ylim([boat_y - 10, boat_y + 10]) 

    # update plot 
    fig.canvas.update() 

    print("draw plot: {:.3f} ms".format((time.time() - time_debug)*1000), end=', ') 
    time_debug = time.time() 

    print('total plot update rate: {:.3f} ms or {:.1f} fps'.format((time.time()-prev_time)*1000, 1/(time.time()-prev_time))) 
    prev_time = time.time() 

    plt.pause(0.001) 

輸出：

transform: 0.000 ms, clear: 25.002 ms, draw plot: 0.000 ms, total plot update rate: 67.007 ms or 14.9 fps 
transform: 0.000 ms, clear: 31.003 ms, draw plot: 1.000 ms, total plot update rate: 76.008 ms or 13.2 fps 
transform: 0.000 ms, clear: 26.002 ms, draw plot: 1.000 ms, total plot update rate: 73.007 ms or 13.7 fps 

Answer 1

考慮使用FuncAnimation而不是做所有的計算，並在一個大循環繪製。

下面是從matplotlib animation examples樣本：

import numpy as np 
import matplotlib.pyplot as plt 
import matplotlib.animation as animation 

fig, ax = plt.subplots() 
line, = ax.plot(np.random.rand(10)) 
ax.set_ylim(0, 1) 

def update(data): 
    line.set_ydata(data) 
    return line, 

def data_gen(): 
    while True: yield np.random.rand(10) 

ani = animation.FuncAnimation(fig, update, data_gen, interval=100) 
plt.show()