D455 depth camera point cloud
I am using realsense D455 camera.
My question is when we want to save a point cloud, is it only based on the depth data, i.e., the stream that is enabled via depth sensor?
Imagine that as a first test in our use-case, we want to have a controlled environment. In such a case, what steps can be taken so that we have point clouds that are of very good quality. Meaning, they capture the surface of the objects in the view very well. At runtime, we would like to use information from scene point clouds to make a correspondence to some object models.
The height of the mount ?
Any other considerations
thanks,
-
Hi Zahid Iqbal If you only need to generate a single frame of point cloud then you could export depth, or depth and color, to a .ply format pointcloud data file. That .ply could then be imported into a 3D modelling program such as MeshLab or Blender for further editing and conversion into a solid mesh object if you wish to do so.
A guide to doing so with the RealSense Viewer tool and the open-source software MeshLab is at the link below.
https://www.andreasjakl.com/capturing-3d-point-cloud-intel-realsense-converting-mesh-meshlab/
If you wish to create your own script to export a pointcloud to .ply file then both C++ and Python languages can do so. However, for a .ply that contains color information, C++ will be the best language to use as Python has problems with exporting color pointcloud data to a .ply.
If the saved pointcloud has to be a continuous stream instead of a single frame then you could record both depth and color streams to a bag file and then use that file as the 'data source' instead of a live camera to generate a point cloud from the file's stored streams as though a live camera was being used.
The minimum height of the mount will be dictated by the minimum depth sensing range of the camera. For the D455 camera model, the camera should be positioned at least 0.5 meters from the object.
Please sign in to leave a comment.
Comments
1 comment