如何从 npy 文件中绘制 3D 点云？

Question

我有一些由包含 3D 点云的 LIDAR 读数创建的 Numpy 二进制文件。我希望能够通过从文件中读取每个点云来绘制自上而下（正交）的视图。我查找了各种 3D 点云库，例如 Open3d、pyntcloud 等，但它们都不适用于 NPY 文件。我怎样才能绘制它们？

我不是在这里要求图书馆推荐。我只是在寻找一个可能的方向，因为我还没有找到一种方法来通过从 NPY 文件中读取点云来绘制点云。

编辑：当我使用np.load() 从其中一个文件中读取数据时，它看起来像这样：

array([[(-0.       ,  0.        ,  0.        , 0.        , 857827240, 1579782324),
    (-0.       ,  0.        ,  0.        , 0.        , 857882120, 1579782324),
    (-0.       ,  0.        ,  0.        , 0.        , 857937680, 1579782324),
    ...,
    (-0.       , -0.        ,  0.        , 0.        , 957653240, 1579782324),
    (-0.       , -0.        ,  0.        , 0.        , 957709120, 1579782324),
    (-0.       , -0.        ,  0.        , 0.        , 957764680, 1579782324)],
   [(15.622366 , -8.086195  ,  5.7023315 , 0.00392157, 857828544, 1579782324),
    (16.292194 , -8.503972  ,  5.8512874 , 0.07843138, 857883424, 1579782324),
    (15.855744 , -8.374023  ,  5.767106  , 0.02352941, 857938984, 1579782324),
    ...,
    (16.500275 , -9.402869  ,  6.0786157 , 0.01568628, 957654544, 1579782324),
    (16.197226 , -9.334285  ,  6.023082  , 0.00392157, 957710424, 1579782324),
    (16.260717 , -9.463429  ,  6.0455737 , 0.00392157, 957765984, 1579782324)],
   [(16.526688 , -8.541684  ,  4.6792016 , 0.00392157, 857829848, 1579782324),
    (15.844723 , -8.292216  ,  4.5818253 , 0.        , 857884728, 1579782324),
    (15.915991 , -8.414634  ,  4.5984206 , 0.00392157, 857940288, 1579782324),
    ...,
    (15.649654 , -8.954793  ,  4.6751213 , 0.01176471, 957655848, 1579782324),
    (17.318968 , -9.951033  ,  4.9357953 , 0.01176471, 957711728, 1579782324),
    (16.125185 , -9.398413  ,  4.7603803 , 0.00392157, 957767288, 1579782324)],
   ...,
   [( 2.5268526, -1.6420269 , -0.24141277, 0.02745098, 857780808, 1579782324),
    ( 2.529189 , -1.6714373 , -0.24518971, 0.03137255, 857836368, 1579782324),
    ( 2.5140662, -1.6922294 , -0.24403782, 0.03137255, 857891248, 1579782324),
    ...,
    ( 1.7650445, -1.4837685 , -0.2509078 , 0.02745098, 957606808, 1579782324),
    ( 1.742465 , -1.5004072 , -0.24779865, 0.02352941, 957662368, 1579782324),
    ( 1.7232444, -1.5187881 , -0.245681  , 0.02745098, 957718248, 1579782324)],
   [(-2.7442074,  0.9481321 ,  1.1273874 , 0.        , 857786024, 1579782324),
    (-2.7466307,  0.94417626,  1.1274364 , 0.        , 857841584, 1579782324),
    (-2.749064 ,  0.94022495,  1.1274853 , 0.        , 857896464, 1579782324),
    ...,
    (-3.4345033,  1.3002251 ,  1.1344001 , 0.        , 957612024, 1579782324),
    (-3.4270716,  1.2909878 ,  1.1304668 , 0.        , 957667584, 1579782324),
    (-3.4362614,  1.2907308 ,  1.1331499 , 0.        , 957723464, 1579782324)],
   [(-3.1056237,  1.1257029 ,  1.1556424 , 0.        , 857782112, 1579782324),
    (-3.1041813,  1.1214051 ,  1.1539782 , 0.        , 857837672, 1579782324),
    (-3.102756 ,  1.1170869 ,  1.1523142 , 0.        , 857892552, 1579782324),
    ...,
    (-3.779868 ,  1.4852207 ,  1.1581781 , 0.        , 957608112, 1579782324),
    (-3.8071766,  1.4963622 ,  1.1718962 , 0.        , 957663672, 1579782324),
    (-3.7931492,  1.4851598 ,  1.163371  , 0.        , 957719552, 1579782324)]],
  dtype=[('x', '<f4'), ('y', '<f4'), ('z', '<f4'), ('intensity', '<f4'), ('t_low', '<u4'), ('t_high', '<u4')])

当我尝试按照@Dorian 的建议进行绘制时：

    x = data[:, 0]
    y = data[:, 1]
    z = data[:, 2]
    fig = plt.figure(figsize=(8, 8))
    ax = fig.add_subplot(111, projection='3d')
    ax.scatter(x, y, z)
    plt.show()

我收到以下错误：

TypeError                                 Traceback (most recent call last)
<ipython-input-20-d6d9ea7be681> in <module>
      1 fig = plt.figure(figsize=(8, 8))
      2 ax = fig.add_subplot(111, projection='3d')
----> 3 ax.scatter(x, y, z)
      4 plt.show()

~/anaconda3/envs/pointclouds/lib/python3.8/site-packages/mpl_toolkits/mplot3d/axes3d.py in scatter(self, xs, ys, zs, zdir, s, c, depthshade, *args, **kwargs)
   2325         xs, ys, zs, s, c = cbook.delete_masked_points(xs, ys, zs, s, c)
   2326 
-> 2327         patches = super().scatter(xs, ys, s=s, c=c, *args, **kwargs)
   2328         art3d.patch_collection_2d_to_3d(patches, zs=zs, zdir=zdir,
   2329                                         depthshade=depthshade)

~/anaconda3/envs/pointclouds/lib/python3.8/site-packages/matplotlib/__init__.py in inner(ax, data, *args, **kwargs)
   1597     def inner(ax, *args, data=None, **kwargs):
   1598         if data is None:
-> 1599             return func(ax, *map(sanitize_sequence, args), **kwargs)
   1600 
   1601         bound = new_sig.bind(ax, *args, **kwargs)

~/anaconda3/envs/pointclouds/lib/python3.8/site-packages/matplotlib/axes/_axes.py in scatter(self, x, y, s, c, marker, cmap, norm, vmin, vmax, alpha, linewidths, verts, edgecolors, plotnonfinite, **kwargs)
   4459         else:
   4460             x, y, s, c, colors, edgecolors, linewidths = \
-> 4461                 cbook._combine_masks(
   4462                     x, y, s, c, colors, edgecolors, linewidths)
   4463 

~/anaconda3/envs/pointclouds/lib/python3.8/site-packages/matplotlib/cbook/__init__.py in _combine_masks(*args)
   1122                 x = safe_masked_invalid(x)
   1123                 seqlist[i] = True
-> 1124                 if np.ma.is_masked(x):
   1125                     masks.append(np.ma.getmaskarray(x))
   1126             margs.append(x)  # Possibly modified.

~/anaconda3/envs/pointclouds/lib/python3.8/site-packages/numpy/ma/core.py in is_masked(x)
   6520     if m is nomask:
   6521         return False
-> 6522     elif m.any():
   6523         return True
   6524     return False

~/anaconda3/envs/pointclouds/lib/python3.8/site-packages/numpy/core/_methods.py in _any(a, axis, dtype, out, keepdims)
     43 
     44 def _any(a, axis=None, dtype=None, out=None, keepdims=False):
---> 45     return umr_any(a, axis, dtype, out, keepdims)
     46 
     47 def _all(a, axis=None, dtype=None, out=None, keepdims=False):

TypeError: cannot perform reduce with flexible type

一小部分数据是here。

Answer 1

matplotlib.pyplot 是我个人的首选。

您没有提供任何数据或数据的保存方式，因此我假设点云的点保存在Nx3 维numpy 数组中：

data = np.load('file.npy')
x = data[:, 0]
y = data[:, 1]
z = data[:, 2]

import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import proj3d

fig = plt.figure(figsize=(8, 8))
ax = fig.add_subplot(111, projection='3d')

ax.scatter(x, y, z)
plt.show()

如果您只想拥有 2D（俯视图），请不要使用 3D 投影并忽略您的 z 值：

fig = plt.figure(figsize=(8, 8))
ax = fig.add_subplot(111)

ax.scatter(x, y)
plt.show()

Answer 2

您的数据包含结构化数组，可按如下方式访问：

import numpy as np

data = np.load("LIDAR_0.npy")
# data.dtype is [('x', '<f4'), ('y', '<f4'), ('z', '<f4'), ('intensity', '<f4'), ('t_low', '<u4'), ('t_high', '<u4')]

# coordinates
xs = data["x"]
ys = data["y"]
zs = data["z"]
# attribute
t_low = data["t_low"]

之后，您可以创建类似于 question 的 3D 图：

import matplotlib.pyplot as plt

fig = plt.figure(figsize=(12,7))
ax = fig.add_subplot(projection='3d')
img = ax.scatter(xs, ys, zs, c=t_low, cmap=plt.hot())
fig.colorbar(img)

ax.set_xlabel('X')
ax.set_ylabel('Y')
ax.set_zlabel('Z')

plt.show()

结果：

或者您可以创建一个 2D 图（例如沿 z 轴查看）：

fig = plt.figure(figsize=(12,7))
ax = fig.add_subplot()
img = ax.scatter(xs, ys, c=t_low, cmap=plt.hot())
fig.colorbar(img)

ax.set_xlabel('X')
ax.set_ylabel('Y')

plt.show()

结果：

注意：主要归功于 Dorian，他已经指出了正确的方向。