Robot vision-based object grasping
Main functionalities
Object grasping from RGB images with the object and target detection tool, executed by a collaborative robot (Franka Emika).
Technical specifications
Object grasping from RGB images with the object and target detection tool, executed by a collaborative robot (Franka Emika).
Inputs and outputs
The annotated image stream can be inspected using rqt_image_view or rviz (select one of the topics /opendr/image_grasp_pose_annotated or /opendr/image_mask_annotated) or rostopic echo /opendr/object_detected This node also publishes a VisionInfo message in /opendr/object_catagories containing the name of a ROS parameter on the server where can be found the correspondance between class ids and class names.
Formats and standards
Linux Ubuntu 20.04 LTS
ROS1: OpenDR v2.0, pytorch 1.9, ros-noetic-moveit, ros-noetic-libfranka, Python 3.8
ROS2: OpenDR v2.0, pytorch 1.9, ros-foxy-moveit, ros-foxy-libfranka, Python 3.8
All dependencies will be installed with the OpenDR toolkit and/or installation instructions are provided
Training material
undefined
Owner (organization)
undefined