ROS 2 - Humble package for point cloud registration and volumetric map generation using Fast_GICP and Voxeland.
GICP Mapper is a ROS 2 - Humble package for point cloud registration and volumetric mapping. The package uses GICP/VGICP/NDT/NDT-D2D algorithms to estimate the relative transformation between consecutive point clouds and incrementally build a consistent 3D map.
The mapping process is based on Voxeland, providing an efficient voxel-based representation of the environment. The package is designed to provide a modular and configurable framework for experimenting with different registration and mapping configurations.
It also includes a node based on Bonxai that implements a probabilistic occupancy grid map server, storing the point distribution within each voxel.
This repo includes
- Generalized Iterative Closest Point
- Voxelized - Generalized Iterative Closest Point
- Normal Distribution Transform
- NDT - D2D and Probabilistic Map server Node
Install dependencies, alongside Voxeland and Fast_GICP, clone this repo and you’re ready to go.
git clone https://github.com/IsraelNunezDeArenas/gicp_mapper.git