我也不知道这玩意主要是干啥用的， 实现如下。
我用剖分的三角形的三个顶点到中心点的距离和作为颜色， 结果显示： 点越密集的地方， 图片上的颜色越深。
from scipy.spatial import Delaunay
import numpy as np
import matplotlib.pyplot as plt
width = 80
height = 40
pointNumber = 50
points = np.zeros((pointNumber, 2))
points[:, 0] = np.random.randint(0, width, pointNumber)
points[:, 1] = np.random.randint(0, height, pointNumber)
tri = Delaunay(points)
center = np.sum(points[tri.simplices], axis=1)/3.0
'''
color = []
for sim in points[tri.simplices]:
x1, y1 = sim[0][0], sim[0][1]
x2, y2 = sim[1][0], sim[1][1]
x3, y3 = sim[2][0], sim[2][1]
s = ((x1-x2)**2+(y1-y2)**2)**0.5 + ((x1-x3)**2+(y1-y3)**2)**0.5 + ((x3-x2)**2+(y3-y2)**2)**0.5
color.append(s)
color = np.array(color)
'''
color = []
for index, sim in enumerate(points[tri.simplices]):
cx, cy = center[index][0], center[index][1]
x1, y1 = sim[0][0], sim[0][1]
x2, y2 = sim[1][0], sim[1][1]
x3, y3 = sim[2][0], sim[2][1]
s = ((x1-cx)**2+(y1-cy)**2)**0.5 + ((cx-x3)**2+(cy-y3)**2)**0.5 + ((cx-x2)**2+(cy-y2)**2)**0.5
color.append(s)
color = np.array(color)
plt.figure(figsize=(20, 10))
plt.tripcolor(points[:, 0], points[:, 1], tri.simplices.copy(), facecolors=color, edgecolors='k')
plt.tick_params(labelbottom='off', labelleft='off', left='off', right='off', bottom='off', top='off')
ax = plt.gca()
plt.scatter(points[:,0],points[:,1], color='r')
#plt.grid()
plt.savefig('Delaunay.png', transparent=True, dpi=600)