Fork of the Lucia 2016 Winter School Semantic Object Maps Tutorial.
This repository contains an altered LVR Plain Classifier that is able to get its mesh data from ROS messages and is therefore online runnable. These Messages can be generated by providing a ROS service on /kinfu/mesh_srv which needs to provide a TriangleMeshStamped and a constant GET which routes the request correctly. The TriangleMeshStamped must be filled at least with the faces and vertices array.
The LVR KFusion Node provides this service.
To compile this package you have to have ROS-kinetic, the ROS package mesh_msgs and LVR compiled and installed.
To launch the pipeline you have to:
- Start the furniture classifier node:
cd "YOUR_WORKSPACE"/src/semantic_object_maps/semantic_furniture_classifier/semantic_furniture_classifier_project/
./build/install/semantic_furniture_classifier_project/bin/semantic_furniture_classifier_project com.github.semantic_furniture_classifier.semantic_furniture_classifier_project.SemanticFurnitureClassifierNode
- Start the service providing Node
- Start the lvr_plane_classifier:
rosrun lvr_plane_classifier lvr_plane_classifier_node(This has to be executed in a directory that contains a valid config.xml!)