shadimsaleh / cso

Calibration the rigid transformation between the stereo and odometry

calibration_stereo_odom (CSO)

Calibrate the rigid transformation between the stereo and odometry

Authors: Doom

Calibration_stereo_odom (CSO) is a tool to calibrate the rigid body transformation between a stereo camera and robot odometry.

#1. Before installation This tool is compiled under Ubuntu(14.04) and g2o, libviso2, Suitesparse, libpng++, yaml-cpp are required.

C++11 or C++0x Compiler

Use the new thread and chrono functionalities of C++11.

G2O

The G2O(General Graph Optimization) is used to perform non-linear optimizations. g2o. Make sure you have installed g2o into your PC.

libviso2

The visual odometry library, used to calculate the stereo-vo to estimate the stereo's movement. The code has already involved in the ThirdParty folder of CSO. Website: www.cvlibs.net

yaml-cpp

There are some parameters used in this project, we use yaml tool to read them in. This lib have already involved in the ThirdParty folder. yaml-cpp.

Suitesparse, libpng++

use the command line below to install this lib.

$ sudo apt-get install libsuitesparse-dev libpng++-dev

#2. Building CSO Clone the repository into your workspace, e.g CSO:

git clone https://github.com/doomzju/calibration_stereo_odom.git CSO

I provide a script build.sh to build the ThirdParty libraries and calibration_stereo_odom. Please make sure you have installed all required dependencies (see section 1). Execute:

cd CSO
chmod +x build.sh
./build.sh

after that you can find the execute file in the build folder.

#3. Usage and Examples. In this tool, if you want use viso2 to estimate the stereo camera's motion, please set the use_viso parameter to true; if you want not only use graph optimization but also use closed form solution, please set the use_closed_form to true. e.g.:

./CSO [params_folder_name]

#4. Data format: There are some test data in the /data folder of this project.