BGKOctoMap-CRF

This repository performs a CRF semantic voxel map on top of the continuous Bayesian Generalized Kernel Inference occcupancy map, built on top of BGKOctoMap, densecrf and semantic_3d_mapping. The implementation is intended for the replication of the methods on a few datasets.

Visualization with rviz on Kitti seq15:

Performance on Kitti as tested on the BKIOctoMap paper:

Getting Started

Dependencies

We tested BGKOctoMap with ROS Melodic. Dependencies include:

ros-melodic-desktop-full
octomap_ros
openmp

Building with catkin

The repository is set up to work with catkin, so to get started you can clone the repository into your catkin workspace src folder and compile with catkin_make:

my_catkin_workspace/src$ git clone https://github.com/zeroAska/BGKOctoMap-CRF.git
my_catkin_workspace/src$ cd BGKOctoMap-CRF
my_catkin_workspace/src/BGKOctoMap-CRF$ mv dense_crf ../
my_catkin_workspace/src/BGKOctoMap-CRF$ cd ../../
my_catkin_workspace$ source devel/setup.bash
my_catkin_workspace$ catkin_make
my_catkin_workspace$ source devel/setup.bash

Running the Demo

We provide ros launchfiles for kitti seq 05 and kitti seq 15. The format of the dataset follows semantic_3d_mapping. The zip files of kitti seq 15 can be downloaded here. Put the files into data/ folder, e.g. data/data_kitti_15/. The launch files is launch/kitti_node.launch. The config file is in config/datasets/kitti_15.yaml.

If you want to turn on the visualization with rviz, set the visualize flag in the last line of config/datasets/kitti_15.yaml to true. This will increase the memory usage and reduce running speed.

To run the demo on kitti 15:

$ roslaunch la3dm kitti_node.launch

which by default will run the full BGKOctoMap-CRF method.

Relevant Works and Publications

This repository serves as a baseline of the BKIOctoMap. It uses code from BGKOctoMap, densecrf and semantic_3d_mapping.