/omtb_ml

Machine Learning for OpenManipulator with TurtleBot3(OMTB)

Primary LanguagePythonApache License 2.0Apache-2.0

OMTB_ML

Machine Learning for OpenManipulator with TurtleBot3(OMTB)

Environment

System: Ubuntu 16.04

ROS version: Kinetic

Gazebo version: 7.0 or higher

AutoLabel

This is a tool for collecting coco dataset in Gazebo environment.

  1. Launch camera
roslaunch camera_gazebo empty_world.launch d1:=$(D1) d2:=$(D2) d3:=$(D3) d4:=$(D4) d5:=$(D5)
  1. Create object & take photo
roslaunch auto_label save_pic.launch dir:=$(Your path) file:=$(Your SDF file) model:=$(Model name)
  1. Remove object
roslaunch auto_label delete.launch
  1. Create coco dataset
roslaunch auto_label create_coco.launch dir:=$(Your path)

Semantic Mapping

3D semantic mapping in Gazebo environment

image

  1. Launch robot
roslaunch omtb_gazebo 1tb_room2.launch
  1. Start service
roslaunch semantic_mapping infer.launch
  1. Launch controller
roslaunch omtb_control turtlebot3_key.launch
  1. Start client
roslaunch semantic_mapping mapping.launch