Open jevonmao opened 1 year ago
Good afternoon!
I'm investigated behaviour behind this issue.
I've made some code to reproduce the problem:
import open3d as o3d
pcl = o3d.geometry.PointCloud()
points = []
pcl2 = o3d.geometry.PointCloud()
points2 = []
for x in range(0, 100):
for y in range(0, 100):
points.append([x, y, 0])
points2.append([x+50, y+50, 0])
pcl.points = o3d.utility.Vector3dVector(points)
pcl2.points = o3d.utility.Vector3dVector(points2)
o3d.visualization.draw_plotly([pcl,pcl2])
Here you may see that only trace 1 point is visible here:
Here you can also see how draw_plotly function is written: https://github.com/isl-org/Open3D/blob/master/python/open3d/visualization/draw_plotly.py#L175
A common workaround for this kind of overplotting situation is to give the markers an opacity < 1 so that overlapping points look different. Opacity is controlled by the scatter.marker.opacity property.
Unfortunatelly, now the setting of the marker propperties is hidden in the get_point_object.
Thus, in current version of the open3d the solution is not availible, however, I'm now working on a newer version to add this feature.
Hope it helped 🦾
@PurplePegasuss Thank you for the detailed explanation. The ability to adjust opacity to allow visualizing overlaid point clouds would be nice as well. Previously I saw #1086 which also said transparency channel is being worked on.
In my original question I was actually referring to drawing multiple viewports, each with its own 3D coordinate plane. Basically having two or more points clouds side by side, laid out on the same tab. Is this possible with open3D?
Checklist
master
branch).My Question
I am using VSCode to ssh into a remote server, so I can only get visualizations to work with the
draw_plotly
method. However it seems very poorly documented. I want to visualize multipleo3d.geometry.PointCloud()
s on the same window, lying side by side. If I pass in multiple point clouds directly intodraw_plotly
it will overlay them, which is not the desired behavior. Does Open3D currently support this feature?