I have a question about showing files of pointcloud as frames and having bounding boxes drawn per frame, the problem is that it is so slow
vis = o3d.visualization.visualizer()
pcd = o3d.io.read_point_cloud(pcd_path1)
vis.add_geometry(pcd)
for pcdPath in pcdPaths:
pcd = o3d.io.read_point_cloud(pcd_path)
# After Clustering then converting labels into list of boxes
for box in boxes:
vis.add_geometry(box, reset_bounding_box=False)
vis.update_geometry(pcd)
vis.poll_events()
vis.update_renderer()
for box in boxes:
box.clear()
What I observed that add_geometry()
is the line slowing the code from around 0.02sec to 0.5sec per frame.
Seems to be that the method of Non-blocking visualization "visualizer()" is not recommended based on what I have read online but what is the other option if I want to have a run of frames with bounding boxes. Any suggestions?
I found a new solution where I avoided using open3d visualization tools but instead a pyqtgraph widget which requires nothing but the ndarray source