Transferring/Translating Image coordinates to a Robot

• MartyX Grover

  • August 18, 2021 07:41

  Hi Kim Danilov  The ROS-based MoveIt and the Python compatibility binding created for it may be a good first option to consider.

   

  https://github.com/ros-planning/moveit

  https://github.com/mikeferguson/moveit_python

  https://ros-planning.github.io/moveit_tutorials/doc/hand_eye_calibration/hand_eye_calibration_tutorial.html

  0

  Comment actions Permalink
• Kim Danilov

  • August 18, 2021 08:27

  Thank you. 

  Can those libraries be used on Windows ? or Linux ? 

  Also, could you please give an example of a script that can extract XYZ coordinates from an image/video coming from the Real Sense camera ?

   

   

   

  0

  Comment actions Permalink
• MartyX Grover

  • August 18, 2021 09:14
  • Edited

  You should be able to use it with either Windows or Linux.  MoveIt has a quick-install option on Linux for building from packages, and instructions for building on Windows or Linux from source code.

  https://moveit.ros.org/install/

   

  If you are using an RGB image and wish to obtain XYZ coordinates, you could use the RealSense SDK instruction rs2_project_color_pixel_to_depth_pixel to convert a single specific color pixel to a depth pixel.  Or alternatively, you could align depth to color and then use the instruction rs2_deproject_pixel_to_point to obtain the 3D coordinates.  

  https://github.com/IntelRealSense/librealsense/issues/6239#issuecomment-614261704

  https://github.com/IntelRealSense/librealsense/issues/5603#issuecomment-574019008

  0

  Comment actions Permalink

Please sign in to leave a comment.