/AgriColMap

Primary LanguageC++

AgriColMap: Aerial-Ground Collaborative 3D Mapping for Precision Farming

This repository contains AgriColMap, an open, research-oriented 3D map registration system for multi-robot in farming scenarios. This software has been tested using the UAV-UGV Collaborative Mapping Dataset distributed within the Flourish Sapienza Datasets collection.

Please also check out our video:

https://www.youtube.com/watch?v=CrfG4v25B8k

Installation with OpenCV > 3.2.0 (with extra modules) and PCL >= 1.7.0

sudo apt-get install libyaml-cpp-dev
## Creating the workspace 
git clone https://github.com/cirpote/AgriColMap.git
cd /agricolmap
git submodule update --init --recursive
mkdir build && cd build
cmake ..
make -j8

Installation with OpenCV > 3.2.0 (with extra modules) without a pre-installed PCL compatible library version

sudo apt-get install libyaml-cpp-dev
## Creating the workspace 
git clone https://github.com/cirpote/AgriColMap.git
cd /agricolmap
sh pcl_configure.sh
git submodule update --init --recursive
mkdir build && cd build
cmake -DBUILD_WITH_PCL=true ..
make -j8

Tutorial

In this tutorial, we briefly show how to use the AgriColMap to register 3D maps gathered by aerial and ground robots. The files you need to download are:

Uncompress the downloaded file into: ${PATH_TO_AGRICOLMAP}/maps/. The "Soybean Dataset" contains UAV and UGV datasets registered in a soybean farm. Other datasets are freely available on Sapienza Collaborative Mapping Datasets.

cd bin
./registration_node ../params/aligner_soybean_params_row3.yaml  10 100 50 2

The 5 parameters are, respectively:

  • the .yaml param file
  • the initial scale error magnitude
  • the translational error magnitude
  • the heading error magnitude
  • the ID number for storing the resulting transform

In this case, we are registering the third row of the soybean dataset with an initial scale error magnitude of 10%, an traslational error magnitude of 2.5 metres, an heading error magnitude of 5 degrees, and an ID of 2.

License

AgriColMap is licensed under the GPL2 License. However, some libraries are available under different license terms. See below.

The following parts are licensed under GPL3:

CPM

The following parts are licensed under GPL2:

cpd

AgriColMap is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the licenses for more details.

Citing

Please cite the following paper when using AgriColMap for your research:

@article{pknsnp_RA-L2019,
  title={{A}gri{C}ol{M}ap: {A}erial-Ground Collaborative {3D} Mapping for Precision Farming},
  author={Potena, Ciro and Khanna, Raghav and Nieto, Juan and Siegwart, Roland and Nardi, Daniele and Pretto, Alberto},
  journal={IEEE Robotics and Automation Letters},
  volume = {4},
  number = {2},
  year={2018}
  pages = {1085--1092},
  doi={10.1109/LRA.2019.2894468}
}