This software is build on ROS(Robotic Operating System). Additionaly, it also requires the following libraries.
- OpenCV (3.2)
- cuDNN
- CUDA (recommand version 10.1)
- Python (recommand version 3.5, 3.7)
sudo apt-get update
sudo apt-get install ros-melodic-ublox
sudo apt-get install ros-melodic-novatel-oem7-driver
sudo apt-get install ros-melodic-grid-map
pip3 install empy
pip3 install catkin_pkg
git clone https://github.com/ailab-konkuk/image_landmark_spc_matching_localization.git
cd image_landmark_spc_matching_localization
catkin_make --cmake-args \
-DCMAKE_BUILD_TYPE=Release \
-DPYTHON_EXECUTABLE=/usr/bin/python3 \
-DPYTHON_INCLUDE_DIR=/usr/include/python3.6m \
-DPYTHON_LIBRARY=/usr/lib/x86_64-linux-gnu/libpython3.6m.so
source devel/setup.bash
* If you want to use only point cloud map matching localization, build only 'landmark_matching_localization'
catkin_make -DCMAKE_BUILD_TYPE=Release -DCATKIN_BLACKLIST_PACKAGES='darknet_ros;deeplab_ros'
roslaunch landmark_matching_localization gps_localization.launch
roslaunch landmark_matching_localization map_matching_localization.launch
For using image landmark and semantic point cloud map matching localization, you have to run 3 launch files. (localization, image segmentation AI, image detection AI)
- Image Landmark and Semantic Point Cloud Map Matching Localization
source devel/setup.bash
roslaunch landmark_matching_localization landmark_matching_localizaion.launch
- Image Segmentation AI
source devel/setup.bash
rosrun deeplab_ros inference_deeplab.py
- Image Detection AI
source devel/setup.bash
roslaunch darknet_ros darknet_ros.launch
You can download semantic point cloud map example file at http://gofile.me/5s2Yk/uN3afXdtD
Put this file at src/landmark_matching_localization/map.
You can download weights example file at http://gofile.me/5s2Yk/orOrQn5ND
Put this file at src/darknet_ros/darknet_ros/yolo_network_config/weights.
You can download rosbag example file at
Tunnel Version : http://gofile.me/5s2Yk/pQCSIBYwK
Downtown Version : http://gofile.me/5s2Yk/5S2GyFTkm
rosbag play 15_ccw_in_d_c_2.bag
rosbag play 15_ccw_in_d_c_4.bag