Open3D Visualize in Google Colab¶
This notebook will think about how to visualize Open3D in Google Colab.
Install Open3D¶
!pip install open3d
Import Open3D and Numpy¶
import numpy as np
import open3d as o3d
Read Point Cloud¶
!wget https://raw.githubusercontent.com/PointCloudLibrary/pcl/master/test/bunny.pcd
--2021-07-29 09:53:05-- https://raw.githubusercontent.com/PointCloudLibrary/pcl/master/test/bunny.pcd
Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.108.133, 185.199.109.133, 185.199.110.133, ...
Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.108.133|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 10878 (11K) [text/plain]
Saving to: ‘bunny.pcd’
bunny.pcd 0%[ ] 0 --.-KB/s
bunny.pcd 100%[===================>] 10.62K --.-KB/s in 0s
2021-07-29 09:53:05 (103 MB/s) - ‘bunny.pcd’ saved [10878/10878]
cloud = o3d.io.read_point_cloud("bunny.pcd")
if cloud.is_empty(): exit()
Estimate Normals¶
cloud.estimate_normals(search_param=o3d.geometry.KDTreeSearchParamHybrid(radius=0.1, max_nn=30))
Visualize using o3d.visualization.draw_geometries()¶
o3d.visualization.draw_geometries() is work fine outside Jupyter.
But, it does not work inside Jupyter and Goolge Colab.
#o3d.visualization.draw_geometries([cloud])
[Open3D WARNING] GLFW Error: X11: The DISPLAY environment variable is missing
[Open3D WARNING] Failed to initialize GLFW
[Open3D WARNING] [DrawGeometries] Failed creating OpenGL window.
Visualize using o3d.visualization.Visualizer()¶
o3d.visualization.Visualizer() is work fine outside Jupyter.
But, it does not work inside Jupyter and Goolge Colab.
#visualizer = o3d.visualization.Visualizer()
#visualizer.create_window(window_name="point cloud", width=1280, height=720, visible=False)
#visualizer.add_geometry(cloud)
#visualizer.run()
Visualize using JVisualizer()¶
Opne3D added experimental support for Jupyter visualization with WebGL. JVisualizer() is work fine inside Jupyter at local.
It is required with Open3D with feature enabled. It does not work in Goolge Colab.
# from open3d import JVisualizer
#visualizer = JVisualizer()
#visualizer.add_geometry(cloud)
#visualizer.show()
Visualize using o3d.visualization.Visualizer() with IPython.display¶
o3d.visualization.Visualizer() with IPython.display is work fine inside Jupyter. It displays image that captured viewer launched outside.
But, it does not work in Goolge Colab.
#import PIL.Image
#from IPython.display import display
#visualizer = o3d.visualization.Visualizer()
#visualizer.create_window(window_name="point cloud", width=1280, height=720, visible=False)
#visualizer.add_geometry(cloud)
#visualizer.update_geometry(cloud)
#visualizer.poll_events()
#image = visualizer.capture_screen_float_buffer()
#visualizer.destroy_window()
#image = (255.0 * np.asarray(image)).astype(np.uint8)
#display(PIL.Image.fromarray(image, "RGB"))
Visualize using Matplotlib¶
In this section, it will visualize point cloud as a scatter plot using matplotlib.
It works fine in inside Jupyter and Goolge Colab. But, it is image.
import matplotlib.pyplot as plt
from mpl_toolkits import mplot3d
points = np.asarray(cloud.points)
colors = None
if cloud.has_colors():
colors = np.asarray(cloud.colors)
elif cloud.has_normals():
colors = (0.5, 0.5, 0.5) + np.asarray(cloud.normals) * 0.5
else:
geometry.paint_uniform_color((1.0, 0.0, 0.0))
colors = np.asarray(geometry.colors)
# Single View
ax = plt.axes(projection='3d')
ax.view_init(90, -90)
ax.axis("off")
ax.scatter(points[:,0], points[:,1], points[:,2], s=1, c=colors)
plt.show()
# Multi View
figsize = plt.rcParams.get('figure.figsize')
fig = plt.figure(figsize=(figsize[0] * 2, figsize[1]))
ax1 = fig.add_subplot(1, 2, 1, projection = '3d')
ax2 = fig.add_subplot(1, 2, 2, projection = '3d')
ax1.axis("off")
ax1.view_init(90, -90) # front view
ax1.scatter(points[:,0], points[:,1], points[:,2], s=1, c=colors)
ax2.axis("off")
ax2.view_init(90 + 90, -90) # top view
ax2.scatter(points[:,0], points[:,1], points[:,2], s=1, c=colors)
plt.show()
Visualize with Plotly¶
In this section, it will visualize point cloud as a scatter plot using plotly. It is able to move viewpoint using mouse.
It works fine in inside Jupyter and Goolge Colab. :thumbsup:
import plotly.graph_objects as go
fig = go.Figure(
data=[
go.Scatter3d(
x=points[:,0], y=points[:,1], z=points[:,2],
mode='markers',
marker=dict(size=1, color=colors)
)
],
layout=dict(
scene=dict(
xaxis=dict(visible=False),
yaxis=dict(visible=False),
zaxis=dict(visible=False)
)
)
)
fig.show()
As well as, You can visualize mesh using plotly in Google Colab.
!wget https://graphics.stanford.edu/~mdfisher/Data/Meshes/bunny.obj
--2021-03-03 04:20:22-- https://graphics.stanford.edu/~mdfisher/Data/Meshes/bunny.obj
Resolving graphics.stanford.edu (graphics.stanford.edu)... 171.67.77.70
Connecting to graphics.stanford.edu (graphics.stanford.edu)|171.67.77.70|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 205917 (201K) [text/plain]
Saving to: ‘bunny.obj’
bunny.obj 100%[===================>] 201.09K 498KB/s in 0.4s
2021-03-03 04:20:22 (498 KB/s) - ‘bunny.obj’ saved [205917/205917]
mesh = o3d.io.read_triangle_mesh("bunny.obj")
if mesh.is_empty(): exit()
if not mesh.has_vertex_normals(): mesh.compute_vertex_normals()
if not mesh.has_triangle_normals(): mesh.compute_triangle_normals()
triangles = np.asarray(mesh.triangles)
vertices = np.asarray(mesh.vertices)
colors = None
if mesh.has_triangle_normals():
colors = (0.5, 0.5, 0.5) + np.asarray(mesh.triangle_normals) * 0.5
colors = tuple(map(tuple, colors))
else:
colors = (1.0, 0.0, 0.0)
fig = go.Figure(
data=[
go.Mesh3d(
x=vertices[:,0],
y=vertices[:,1],
z=vertices[:,2],
i=triangles[:,0],
j=triangles[:,1],
k=triangles[:,2],
facecolor=colors,
opacity=0.50)
],
layout=dict(
scene=dict(
xaxis=dict(visible=False),
yaxis=dict(visible=False),
zaxis=dict(visible=False)
)
)
)
fig.show()
You can use Plotly like o3d.visualization.draw_geometries() for visualize Open3D geometries.
def draw_geometries(geometries):
graph_objects = []
for geometry in geometries:
geometry_type = geometry.get_geometry_type()
if geometry_type == o3d.geometry.Geometry.Type.PointCloud:
points = np.asarray(geometry.points)
colors = None
if geometry.has_colors():
colors = np.asarray(geometry.colors)
elif geometry.has_normals():
colors = (0.5, 0.5, 0.5) + np.asarray(geometry.normals) * 0.5
else:
geometry.paint_uniform_color((1.0, 0.0, 0.0))
colors = np.asarray(geometry.colors)
scatter_3d = go.Scatter3d(x=points[:,0], y=points[:,1], z=points[:,2], mode='markers', marker=dict(size=1, color=colors))
graph_objects.append(scatter_3d)
if geometry_type == o3d.geometry.Geometry.Type.TriangleMesh:
triangles = np.asarray(geometry.triangles)
vertices = np.asarray(geometry.vertices)
colors = None
if geometry.has_triangle_normals():
colors = (0.5, 0.5, 0.5) + np.asarray(geometry.triangle_normals) * 0.5
colors = tuple(map(tuple, colors))
else:
colors = (1.0, 0.0, 0.0)
mesh_3d = go.Mesh3d(x=vertices[:,0], y=vertices[:,1], z=vertices[:,2], i=triangles[:,0], j=triangles[:,1], k=triangles[:,2], facecolor=colors, opacity=0.50)
graph_objects.append(mesh_3d)
fig = go.Figure(
data=graph_objects,
layout=dict(
scene=dict(
xaxis=dict(visible=False),
yaxis=dict(visible=False),
zaxis=dict(visible=False)
)
)
)
fig.show()
o3d.visualization.draw_geometries = draw_geometries # replace function
o3d.visualization.draw_geometries([cloud])
o3d.visualization.draw_geometries([mesh])