zebrajack / dbot

Depth-Based Bayesian Object Tracking Library

Depth Based Object Tracking Library (dbot)

This library implements two different object tracking algorithms:

• A Particle Filter based Tracker (https://am.is.tuebingen.mpg.de/publications/wuthrich_iicirs_2013)

  This tracker can run on a pure CPU system or optionally run on GPU using CUDA 6.5 or later

   inproceedings{wuthrich-iros-2013,
     title = {Probabilistic Object Tracking Using a Range Camera},
     author = {W{\"u}thrich, M. and Pastor, P. and Kalakrishnan, M. and Bohg, J. and Schaal, S.},
     booktitle = {IEEE/RSJ International Conference on Intelligent Robots and Systems},
     pages = {3195-3202},
     publisher = {IEEE},
     month = nov,
     year = {2013},
     month_numeric = {11}
  }

• A Gaussian Filter based Tracker (https://am.is.tuebingen.mpg.de/publications/jan_icra_2016)

   @inproceedings{jan_ICRA_2016,
     title = {Depth-based Object Tracking Using a Robust Gaussian Filter},
     author = {Issac, Jan and W{\"u}thrich, Manuel and Garcia Cifuentes, Cristina and Bohg, Jeannette and Trimpe, Sebastian and Schaal, Stefan},
     booktitle = {Proceedings of the IEEE International Conference on Robotics and Automation (ICRA) 2016},
     publisher = {IEEE},
     month = may,
     year = {2016},
     url = {http://arxiv.org/abs/1602.06157},
     month_numeric = {5}
   }

All trackers require mesh models of the tracked objects in Wavefront (.obj) format. To get started, we recommend that you follow the instructions at https://github.com/bayesian-object-tracking/getting_started.

Requirements

• Ubuntu 14.04
• C++11 Compiler (gcc-4.7 or later)
• CUDA 6.5 or later (optional)

Dependecies

• Filtering library (fl)
• Eigen 3.2.1 or later

Compiling

The cmake package uses Catkin. If you have installed ROS groovy or later, then you most likely have catkin installed already.

 $ cd $HOME
 $ mkdir -p projects/tracking/src  
 $ cd projects/tracking/src
 $ git clone [email protected]:filtering-library/fl.git
 $ git clone [email protected]:bayesian-object-tracking/dbot.git
 $ cd ..
 $ catkin_make -DCMAKE_BUILD_TYPE=Release -DDBOT_BUILD_GPU=On

If no CUDA enabled device is available, you can deactivate the GPU implementation via

 $ catkin_make -DCMAKE_BUILD_TYPE=Release -DDBOT_BUILD_GPU=Off

How to use dbot

Checkout the ros nodes of each tracker in dbot_ros package for exact usage of the filters.