Reconstructs 3d environment with semantics containing 150 classes of ADE20K dataset. Requires color, depth and segmented images as input. "README_SPARK.md" should be refered for more info.
-
Install ROS by following our reference, or the official ROS website.
-
Install system dependencies:
sudo apt-get install python-wstool python-catkin-tools protobuf-compiler autoconf
# Change `melodic` below for your own ROS distro
sudo apt-get install ros-melodic-cmake-modules
Using catkin:
# Setup catkin workspace
mkdir -p ~/catkin_ws/src
cd ~/catkin_ws/
catkin init
catkin config --extend /opt/ros/melodic
catkin config --cmake-args -DCMAKE_BUILD_TYPE=Release
catkin config --merge-devel
# Add workspace to bashrc.
echo 'source ~/catkin_ws/devel/setup.bash' >> ~/.bashrc
# Clone repo
cd ~/catkin_ws/src
git clone https://github.com/pranay731/Kimera-Semantics-on-ADE20K.git
# Install dependencies from rosinstall file using wstool
wstool init # Use unless wstool is already initialized
# For ssh:
wstool merge Kimera-Semantics/kimera/install/kimera_semantics_ssh.rosinstall
# For https:
#wstool merge Kimera-Semantics/kimera/install/kimera_semantics_https.rosinstall
# Download and update all dependencies
wstool update
Finally, compile:
# Compile code
catkin build kimera
# Refresh workspace
source ~/catkin_ws/devel/setup.bash
-
Launch required packages for:
- Camera color and depth input
- Semantic Segmentation input
- Odometry and tf frames
-
Launch file :
- left_cam_info_topic -> color camera_info topic
- left_cam_topic -> color image_raw topic
- left_cam_depth_topic -> depth image_raw topic
- left_cam_segmentation_topic -> segmented color image_raw on ADE20K classes
-
Launch Kimera-Semantics:
roslaunch kimera_semantics_ros kimera_semantics_realsense.launch
- Launch rviz for visualization:
source ~/catkin_ws/devel/setup.bash
rviz -d $(rospack find kimera_semantics_ros)/rviz/kimera_semantics_realsense.rviz
If runing in a Gazebo simulation, use 'kimera_semantics_gazebo' launch and rviz files.