Skip to content

robertogl/ccny_rgbd_tools

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

CCNY RGB-D tools

Ivan Dryanovski
[email protected]

Copyright (C) 2013, City University of New York
CCNY Robotics Lab
http://robotics.ccny.cuny.edu/

Overview

The stack contains a tools for visual odometry and mapping using RGB-D cameras.

This code is at an experimental stage, and licensed under the GPLv3 license.

Installing

From source

Create a directory where you want the package downloaded (ex. ~/ros), and make sure it's added to your$ROS_PACKAGE_PATH.

If you don't have git installed, do so:

sudo apt-get install git-core

Download the stack from our repository:

git clone https://github.com/ccny-ros-pkg/ccny_rgbd_tools.git

Install any dependencies using rosdep.

rosdep install ccny_rgbd_tools

Alternatively, you can manually install the dependencies by

sudo apt-get install libsuitesparse-dev

Compile the stack:

rosmake ccny_rgbd_tools

If you get an error compiling ccny_g2o, it might be because of an incompatible g2o installation. Try removing libg2o:

sudo apt-get remove ros-fuerte-libg2o

Quick usage

Connect your RGB-D camera and launch the Openni device. The openni_launch file will start the driver and the processing nodelets.

roslaunch ccny_openni_launch openni.launch 

For best performace, consider using dynamic reconfigure to change the sampling rate of the rgbd_image_proc nodelet. For example, setting it to to 0.5 will downsample the images by a factor of 2.

Next, launch the visual odometry:

roslaunch ccny_rgbd vo+mapping.launch

Finally, launch rviz.

rosrun rviz rviz

For convenience, you can load the ccny_rgbd/launch/rviz.cfg file.

Configuration

There are many paramters - the first ones you can try changing are:

  • resolution of the OpenNI camera, through dynamic reconfigure. QVGA is recommended, VGA is the default
  • in ccny_rgbd/launch/visual_odometry.launch: feature/GFT/n_features: the number of features to detect in each image. Default is 150, higher numbers (try up to 500) might give more robust tracking)

References

If you use this system in your reasearch, please cite the following paper:

Ivan Dryanovski, Roberto G. Valenti, Jizhong Xiao. Fast Visual Odometry and Mapping from RGB-D Data. 2013 International Conference on Robotics and Automation (ICRA2013).

More info

Coming soon:

Some videos:

About

ccny_rgbd_tools

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published