[D455] PointCloud

• MartyG

  • July 16, 2022 05:54
  • Edited

  Hi Nguyenduyduc14  The choice of point cloud generation method can be determined by the needs of your project.  If you require a straightforward point cloud, perhaps with some depth colorization or depth-color alignment, then Python - and in particular the RealSense SDK's pyrealsense2 wrapper - can fulfil that role very well.

   

  If your project requires more complex processing of the point cloud, such as applying advanced filtering like outlier and noise removal, stitching together multiple point clouds into a single combined cloud and converting the cloud to a 3D mesh, then interfacing the RealSense SDK with a specialized point cloud library such as Open3D or PCL will be beneficial.

   

  The choice to interface RealSense with an external library may also be driven by whether that library is being used by other parts of your project.

   

  If your project does not require interfacing with the RealSense SDK to program an application and you just want to use the RealSense camera hardware to generate a cloud with an existing software tool then a wider range of RealSense-compatible point cloud tools could be considered, such as Pyntcloud and CloudCompare.

   

  https://pypi.org/project/pyntcloud/

  https://www.danielgm.net/cc/

   

  The best point cloud method to choose may be the simplest one to implement that can also meet all of your project needs.  I would therefore recommend looking at the pyrealsense2 wrapper first and then exploring other options such as the ones discussed here if you require features that are not available in the wrapper and cannot be created using pyrealsense2 scripting such as your own code or code shared by other RealSense users.

  0

  Comment actions Permalink

Please sign in to leave a comment.