Determining the three-dimensional shape and position of transparent objects for the purpose of robotic manipulation
To clone this repo:
git clone https://PersonalAccessToken@github.com/vlatkamihic/Detection-and-manipulation-of-transparent-objects
Paper: https://urn.nsk.hr/urn:nbn:hr:200:906497
ClearGrasp neural network-based approach for 3D object detection and segmentation from RGB-D images
The network is used for 3D transparent object detection which will be used for their manipulation. ClearGrasp network model is updated and changed so it works with RealSense Lidar L515 camera and ROS. Changes made to the ClearGrasp repo are placed into the cleargrasp folder.
RUNNING PROGRAMM:
cd cleargrasp/live_demo/live_demo_ws/
source devel/setup.bash
export PYTHONPATH="$PYTHONPATH:/home/robot/cleargrasp"
source /home/robot/catkin_ws/devel/setup.bash --extend
roslaunch cleargrasp detect.launch