/3DTK-quickstart

Repository that contains 3D point and pose datasets and corresponding ground truth point clouds, using unconventional trajectories such as rolling, pendulum, and descending while rotating.

3DTK quick install and setup

Repository that functions as a collection. It contains links and instructions for the processing of 3D point and pose datasets with corresponding ground truth point clouds. The presented screenshots (see below) were captured from datasets using unconventional trajectories such as rolling, pendulum, and descending while rotating. You can find these and more datasets in the robotics 3D scan repository.

Processing

We use 3DTK, which is a toolbox written in C++ for 3D point clouds that comes e.g. with a fast viewer and video creator. To start, install the following dependencies on your system:

  • cmake
  • cmake-curses-gui
  • boost (libboost)
  • blas (libblas)
  • suitesparse (libsuitesparse)
  • opencv (libopencv)
  • zip (libzip)
  • cgal (libcgal)
  • eigen3 (libeigen3)
  • mesa (libglu1-mesa and mesa-common)
  • newmat (libnewmat)
  • ann (libann)
  • freeglut
  • wxgtk3.0 (optional, if you need wxShow viewer)
  • xvfb (optional, for unit testing on Linux)

In Debian based systems such as Ubuntu, try the following to install them all:

sudo apt install cmake cmake-curses-gui build-essential libboost-all-dev libblas-dev libsuitesparse-dev libopencv-dev libglu1-mesa-dev freeglut3-dev mesa-common-dev libwxgtk3.0-gtk3-dev libzip-dev qtbase5-dev libcgal-dev libeigen3-dev libann-dev libnewmat10-dev xvfb

Install 3DTK as follows:

svn checkout https://svn.code.sf.net/p/slam6d/code/trunk slam6d-code
cd slam6d-code
mkdir .build
make config 

Configure your build flags, such as ROS-, OpenCV-, or 3DConnexion-support, and generate buildfiles ("c" to configure, "g" to generate). Then, install:

make -j<NUMBER_OF_CPUS>

ROS support

If you plan on using the toolkits programms with ROS, enable support using ==make config== and check the ==WITH_ROS== flag. Make sure you have the following dependencies on your system:

  • ROS distribution of your choice (kinetic, noetic, ...)
  • ros--tf2-sensor-msgs
  • orocos-bfl (liborocos-bfl-dev)

Get started using the 3D toolkit

To start, place datasets in the "dat" folder and show them with

bin/show dat/your/dataset -f <FILE_FORMAT> --advanced 

Refer to 3DTKs svn repository for further guidance and wikipage.

Publication

Yet to be published.

Authors: Fabian Arzberger, Jasper Zevering, Dorit Borrmann, Anton Bredenbeck, and Andreas Nüchter

Screenshots

Bild Bild Bild Bild Bild Bild