Visual Servoing Platform  version 3.1.0 under development (2017-07-21)
Tutorial: Installation from source for Linux CentOS

In this tutorial you will learn how to install ViSP from source on CentOS. These steps have been tested for CentOS 7.0 (x86_64) distribution, but should work with any other distribution as well.

Note
Concerning ViSP installation, we provide also other Tutorials.

Install prerequisities

  • gcc 4.4.x or later. This can be installed with:
    $ sudo yum install gcc-c++
  • CMake 2.6 or higher that could be installed with:
    $ sudo yum install cmake

Install 3rd parties

ViSP is interfaced with several 3rd party libraries. The complete list is provided here.

Recommended 3rd parties

We recommend to install the following:

  • OpenCV
    $ sudo yum install opencv-devel
  • libX11 to be able to open a window to display images
    $ sudo yum install libX11-devel
  • lapack to benefit from optimized mathematical capabilities
    $ sudo yum install lapack-devel
  • libdc1394 to grab images from firewire cameras
    $ sudo yum install libdc1394-devel
  • libv4l to grab images from usb or analogic cameras
    $ sudo yum install libv4l-devel
  • libxml2 to be able to configure the model-based trackers from xml files
    $ sudo yum install libxml2-devel

Other optional 3rd parties

We give also the way to install other 3rd party libraries to enable specific capabilities.

  • Coin, to be able to support vrml cad model used by the model-based trackers
    $ sudo yum install Coin2-devel
  • libjpeg and libpng to support jpeg and png images respectively (only useful if OpenCV is not installed)
    $ sudo yum install libjpeg-devel libpng-devel

Install ViSP from source code

Getting ViSP source code

There are different ways to get ViSP source code:

  • You can download the latest release as a zip or a tarball. Once downloaded, uncompress the file using either
    $ tar xvzf visp-x.y.z.tar.gz
    or
    $ unzip visp-x.y.z.zip
  • You can also download a daily snapshot. Once downloaded, uncompress the file using
    $ tar xvzf visp-snapshot-yyyy-mm-dd.tar.gz
  • Or you get the cutting-edge ViSP from GitHub repository using the following command
    $ git clone https://github.com/lagadic/visp.git

We suppose now that ViSP source is in a directory denoted <source_dir>, for example $HOME/visp

Configuring ViSP from source

  • Create first a directory denoted <binary_dir> where you want to build ViSP. This directory will contain generated Makefiles, object files, and output libraries and binaries.
    $ cd $HOME; mkdir visp-build
  • Enter the <binary_dir> and configure the build:
    $ cd $HOME/visp-build
    $ cmake ../visp
    A more versatile way to configure the build is to use ccmake, the CMake GUI:
    $ ccmake ../visp
    The following image shows that this command allows to configure (just by pressing [c] key) the build in a more advanced way where some options could be easily turned On/Off. It allows also to see which are the 3rd parties that will be used. To generate the makefiles, just press [g] key in the ccmake gui.
    img-ccmake-centos-all.png
    Snapshot of the ccmake ../visp command used to configure ViSP.

Building ViSP libraries

  • To build ViSP proceed with:
    $ make -j4
  • To install ViSP proceed with:
    $ sudo make install
    Note
    This stage is optional, since ViSP could be used as a 3rd party without installation.
    The default install location is set to /usr/local. This location could be changed modifying CMAKE_INSTALL_PREFIX var.

Building ViSP documentation

  • To build ViSP documentation, you have first to install Doxygen package:
    $ sudo yum install doxygen graphviz
    Then you can proceed with:
    $ cmake ../visp
    $ make -j4 visp_doc
    The generated documentation is then available in <binary_dir>/doc/html/index.html
  • It is also possible to generate a more complete documentation that includes also all the internal classes. This could be achieved setting CMake var ENABLE_FULL_DOC=ON like:
    $ cmake ../visp -DENABLE_FULL_DOC=ON
    $ make -j4 visp_doc

Install ViSP dataset

Some ViSP examples and tests require data (images, video, models) that are not part of ViSP source code but available in a separate archive named ViSP-images-x.y.z.zip. This archive could be downloaded from http://visp.inria.fr/download page. We provide here after the way to install these data if you want to run ViSP examples.

$ cd $HOME
$ unzip ViSP-images-x.y.z.zip

We suppose now that the data are located in $HOME/ViSP-images.

$ ls $HOME/ViSP-images
Klimt README.md circle ellipse iv mbt mire-2
LICENSE.txt calibration cube ellipse-1 line mire video

Set VISP_INPUT_IMAGE_PATH environment variable to help ViSP examples and tests to find the location of the data set. It's convenient if the environment variables is automatically added to your bash session every time a new shell is launched:

$ echo "export VISP_INPUT_IMAGE_PATH=$HOME" >> ~/.bashrc
$ source ~/.bashrc
Note
For historical reasons VISP_INPUT_IMAGE_PATH should not contain the folder ViSP-images, but the parent folder.

From now, you can try to run ViSP examples and tests. For example you can run displayX example that should open a windows with Klimt painting image and some overlay drawings:

$ cd $HOME/visp-build
$ ./example/device/display/displayX
A click to close the windows...
A click to display a cross...
Cross position: 201, 441
A click to exit the program...
Bye

Tips and tricks

How to uninstall ViSP

After ViSP installation, you can remove installed material using:

$ sudo make uninstall

How to build only ViSP libraries

If you want to build only ViSP modules libraries, nor the examples, tutorials and tests:

$ make -j4 visp_modules

How to build a ViSP specific module

If you want to build a given module and all the dependencies:

$ make -j4 visp_<module_name>

For example to build the model-based tracker module named mbt, run:

$ make -j4 visp_mbt

Which are the targets that could be run with make ?

To know which are the target available with make:

$ make help | grep visp
... visp_tests
... visp_demos
... visp_tutorials
... visp_examples
... visp_modules
... visp_doc
... visp_core
... visp_detection
... visp_gui
... visp_io
... visp_klt
... visp_me
... visp_robot
... visp_sensor
... visp_ar
... visp_blob
... visp_visual_features
... visp_vs
... visp_vision
... visp_mbt
... visp_tt
... visp_tt_mi

Which are the 3rd party libraries that are used in ViSP ?

To see which are the optional 3rd parties that are found during the configuration stage and that will be used by ViSP during the build you can have a look to the text file named ViSP-third-party.txt and located in <binary_dir>. We provide hereafter an example of a possible content of this file:

ViSP third-party libraries
Below you will find the list of third party libraries used to
build ViSP on your computer.
Mathematics:
Gnu Scientific Library : no
Lapack/blas : yes
Simulator:
Ogre simulator : no
\- Ogre3D : no
\- OIS : no
Coin simulator : no
\- Coin3D : no
\- SoWin : no
\- SoXt : no
\- SoQt : no
\- Qt4 : no
\- Qt3 : no
Robots
Afma6 : no
Afma4 : no
Biclops : no
Ptu46 : no
Pioneer : no
Viper S650 : no
Viper S850 : no
Video devices (display)
X11 : yes
GTK : no
OpenCV : yes
GDI : no
Direct3D : no
Framegrabbers
Firewire libdc1394-2.x : yes
Video For Linux Two : yes
DirectShow : no
CMU 1394 Digital Camera SDK : no
OpenCV : yes
Specific devices
Yarp : no
Kinect : no
\-libfreenect : no
\-libusb-1.0 : no
\-pthread : yes
Video and image Read/Write:
FFMPEG : no
libjpeg : no
libpng : no
Misc:
XML2 : yes
pthread : yes
OpenMP : yes
zbar : no
dmtx : no
Documentation:
Doxygen : no
Graphviz dot : no
ViSP built with C++11 features: no

Known issues

libpthread may be hidden by files in //lib64

On CentOS 7.0 with cmake 2.8.11, during cmake configuration you may encounter the following issue:

CMake Warning at src/CMakeLists.txt:80 (add_library):
Cannot generate a safe runtime search path for target visp because files in
some directories may conflict with libraries in implicit directories:
runtime library [libpthread.so] in /usr/lib64 may be hidden by files in:
//lib64
Some of these libraries may not be found correctly.

The problem was that libpthread.so exists in /usr/lib64 and in //lib64. In //lib64 it should be a symbolic link to /usr/lib64.

$ ls -als //lib64
0 lrwxrwxrwx. 1 root root 9 Feb 4 12:16 //lib64 -> usr/lib64
$ ls -als //lib64/libpthread*
140 -rwxr-xr-x. 1 root root 141616 Jan 27 15:13 //lib64/libpthread-2.17.so
4 -rw-r--r--. 1 root root 222 Jan 27 14:42 //lib64/libpthread.so
0 lrwxrwxrwx. 1 root root 18 Feb 4 12:34 //lib64/libpthread.so.0 -> libpthread-2.17.so

The fix consists in removing //lib64/libpthread.so and creating a new symbolic link

$ cd //lib64
$ sudo rm libpthread.so
$ sudo ln -s libpthread-2.17.so libpthread.so
$ ls -als libpthread*
140 -rwxr-xr-x. 1 root root 141616 Jan 27 15:13 libpthread-2.17.so
0 lrwxrwxrwx. 1 root root 18 Feb 4 16:09 libpthread.so -> libpthread-2.17.so
0 lrwxrwxrwx. 1 root root 18 Feb 4 12:34 libpthread.so.0 -> libpthread-2.17.so

vpVideoReader is not able to read mpeg videos

On CentOS 7.0, vpVideoReader is not able to read mpeg videos. If ffmpeg is not installed (this is our case since ffmpeg is not packaged for CentOS 7.0), this class uses OpenCV to read and decode videos. Some examples or tutorials provided in ViSP hang during cv::Capture::open() call. The reason is that OpenCV 2.4.5 cv::Capture seams buggy. This is for example the case if you run:

$ ./example/video/videoReader

A work arround consists in installing a more recent OpenCV version from source.

Next tutorial

You are now ready to see the next Tutorial: How to create and build a CMake project that uses ViSP on Unix or Windows that will show you how to use ViSP as a 3rd party to build your own project.