Detailed development and implementation of:
- Discrete Counting Sensor Model (D-CSM);
- Continuous Counting Sensor Model (C-CSM);
- Discrete Semantic Counting Sensor Model (D-SCSM) &
- Countinuous Semantic Counting Sensor Model (C-SCSM)
- BKI Semantic Mapping
| Model & Info | Map | Variance Map |
|---|---|---|
| D-CSM | ||
| Implemented a 2D counting sensor model (CSM) for occupancy grid mapping in ogm_CSM.py. Visualization of the map is with grid_size = 0.135 m and its associated variance map |  |
 |
| C-CSM | ||
| Implemented a 2D continuous counting sensor model (CSM) in ogm_continuous_CSM.py. Visualization of the map is with grid_size = 0.135 m and its associated variance map |  |
 |
| ....... with grid_size = 0.270 m |  |
 |
| ....... with grid_size = 0.500 m |  |
 |
| D-SCSM | ||
| Implemented a semantic counting sensor model (S-CSM) in ogm_S_CSM.py. Visualization of the map is with grid_size = 0.135 m and variance map of the class with highest probabilities at each grid |  |
 |
| C-SCSM | ||
| Implemented a continuous semantic counting sensor model (S-CSM) in ogm_continuous_S_CSM.py. Visualization of the map is with grid_size = 0.135 m and variance map of the class with highest probabilities at each grid |  |
 |
Build and run the BKI Semantic Mapping on KITTI semantics dataset sequence 04.
Notice the system dependencies:
- Ubuntu system. The mapping algorithm is build on Ubuntu system. It has been tested on Ubuntu 16.04 and Ubuntu 18.04. If you didn’t have an Ubuntu system, you could create a virtual machine.
- ROS system. The mapping algorithm has been tested on ROS kinetic and ROS melodic. You can follow the installation guide for ROS in the documentation.
- A catkin workspace (catkin_ws). You can follow the steps in the tutorial. Read the README in the BKI semantic repository and do the following steps:
- Build the repository with catkin. Run the following commands in your catkin workspace (catkin_ws): • cd src/ • git clone https://github.com/ganlumomo/BKISemanticMapping • cd .. • catkin_make • source catkin_ws/devel/setup.bash
- Download semantic KITTI dataset sequence 04 data from https://drive.google.com/file/d/19Dv1jQqf-VGKS2qvbygFlUzQoSvu17E5/view and uncompress it into the data folder.
- Run the demo with the following command: • roslaunch semantic_bki semantickitti_node.launch
