View my account

Merging two point clouds taken from two D435 cameras

Comments

6 comments

  • MartyG

    Hi Amosz  When trying to combine RealSense point clouds into a single cloud, a common recommendation is to perform an affine transform.  Basically, just rotate and move the point clouds, in 3D space, and then once you've done that, you append the point clouds together.

    In the RealSense SDK, the instruction rs2_transform_point_to_point is an affine transform.  The link below provides information about implementing it.

    https://github.com/IntelRealSense/librealsense/issues/5583#issuecomment-570805578 

    0
    Comment actions Permalink
  • Amosz

    Hello MartyG,

    Thanks for the reply.

    I did affine transform in the program.

    Similar to your description, I rotate and translate each point cloud and add them together in the end.

    But I didn't use rs2_transform_point_to_point() function, I used a function from PCL - pcl::transfromPointCloud(),

    which takes a intput point cloud, a output transformed point cloud and a 4x4 transform matrix.

     

     

    0
    Comment actions Permalink
  • MartyG

    Before proceeding further, may I ask whether PCL and Charuco are vital to your project.  I ask because markerless point cloud stitching with 2 cameras observing from different angles can be achieved using ROS, as described in the Intel guide in the link below.

    https://github.com/IntelRealSense/realsense-ros/wiki/Showcase-of-using-2-cameras 

    0
    Comment actions Permalink
  • Amosz

    Thanks for the reply.

    Is there any other solution besides ROS?

    I will try the ROS method as well.

    0
    Comment actions Permalink
  • MartyG

    PCL has the Normal Distributions Transform method of point cloud stitching.

    https://pointclouds.org/documentation/tutorials/normal_distributions_transform.html 

    At a more advanced level, the CONIX Research Center at Carnegie Mellon developed a point cloud stitching system that could stitch clouds from up to 20 RealSense 400 Series cameras over an ethernet network.

    https://github.com/conix-center/pointcloud_stitching 

    0
    Comment actions Permalink
  • Amosz

    Thanks for the help, much appreciated.

    I will dig into those solutions.

    0
    Comment actions Permalink

Please sign in to leave a comment.