我有这个问题。我尝试通过scipy.spatial.Delaunay对点云进行三角测量。我用过:
tri = Delaunay(points) # points: np.array() of 3d points
indices = tri.simplices
vertices = points[indices]
但是,这段代码返回了四面体。怎么可能只返回表面三角形?
由于
答案 0 :(得分:7)
要使其像代码形式一样工作,您必须将曲面参数化为2D。例如,在球(r,theta,psi)的情况下,半径是常数(将其丢弃),点由(theta,psi)给出,即2D。
Scipy Delaunay是N维三角剖分,所以如果你给3D点,它会返回3D对象。给它2D点并返回2D对象。
下面是我用于为openSCAD创建多面体的脚本。 U和V是我的参数化(x和y),这些是我给Delaunay的坐标。注意,现在" Delaunay三角剖分属性"仅适用于u,v坐标(角度在uv-space中最大化,而不是xyz -space等)。
该示例是来自http://matplotlib.org/1.3.1/mpl_toolkits/mplot3d/tutorial.html的修改后的副本,该副本最初使用 Triangulation 函数(最终映射到Delaunay?)
import numpy as np
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D
import matplotlib.tri as mtri
from scipy.spatial import Delaunay
# u, v are parameterisation variables
u = np.array([0,0,0.5,1,1])
v = np.array([0,1,0.5,0,1])
x = u
y = v
z = np.array([0,0,1,0,0])
# Triangulate parameter space to determine the triangles
#tri = mtri.Triangulation(u, v)
tri = Delaunay(np.array([u,v]).T)
print 'polyhedron(faces = ['
#for vert in tri.triangles:
for vert in tri.simplices:
print '[%d,%d,%d],' % (vert[0],vert[1],vert[2]),
print '], points = ['
for i in range(x.shape[0]):
print '[%f,%f,%f],' % (x[i], y[i], z[i]),
print ']);'
fig = plt.figure()
ax = fig.add_subplot(1, 1, 1, projection='3d')
# The triangles in parameter space determine which x, y, z points are
# connected by an edge
#ax.plot_trisurf(x, y, z, triangles=tri.triangles, cmap=plt.cm.Spectral)
ax.plot_trisurf(x, y, z, triangles=tri.simplices, cmap=plt.cm.Spectral)
plt.show()
下面是(稍微更结构化的)文本输出:
polyhedron(
faces = [[2,1,0], [3,2,0], [4,2,3], [2,4,1], ],
points = [[0.000000,0.000000,0.000000],
[0.000000,1.000000,0.000000],
[0.500000,0.500000,1.000000],
[1.000000,0.000000,0.000000],
[1.000000,1.000000,0.000000], ]);
答案 1 :(得分:4)
看起来您想要计算点云的convex hull。我想这就是你想要做的事情:
from scipy.spatial import ConvexHull
hull = ConvexHull(points)
indices = hull.simplices
vertices = points[indices]
答案 2 :(得分:0)