WASS was designed to be portable to numerous platforms. It was developed and actively tested on Linux but can be used on any recent versions of Mac OSX and Microsoft Windows.
Use the following links to get detailed installation instructions for your system:
Important Note:
Since from version 4.x the OpenCV library dropped the support of the old C API, part of the WASS pipeline can no longer be compiled without "downgrading" to 3.x branch. To be sure of having a tested and working WASS system, please consider installing it as a Docker container.
Advantages of using the containerized version:
No need to manually compile and install all the required dependencies for your system (everything is done by the Dockerfile script)
You'll have a tested and repeatable ecosystem across different platforms
Easily setup in/out/config directories via Docker volumes (bash scripts provided)
If you already have Docker up and running, you can quickly skip to the Docker installation instructions.
WASS relies to the following 3rd party components:
OpenCV Open-Source Computer Vision Library (Version 3.x required)
Boost C++ Libraries
Lapack Linear Algebra PACKage
SBA Sparse Bundle Adjustment C/C++ Library
redis In-memory data structure store
node.js Javascript runtime
Additionally, Matlab is required for some post-processing steps like surface estimation and waves spectrum recovery.
* NOTE: the following procedure was tested on Linux Ubuntu 16.04.01 LTS (Xenial Xerus) but should be valid on any relatively modern Linux distribution providing the appropriate third-party packages.*
Start by installing Boost, Lapack, and all the needed build tools from the official packages provided by Ubuntu. Open a terminal, cd into a directory of your choice end enter the following command:
$ sudo apt-get install git build-essential cmake curl liblapack-dev libblas-dev libboost-all-dev
OpenCV has to be compiled by source since version 3.1 is not provided yet via the official Ubuntu repositories. On the same terminal, enter the following commands:
$ sudo apt-get install ffmpeg libavcodec-dev libavformat-dev
$ git clone https://github.com/opencv/opencv.git --depth 1 -b 3.4.0 --single-branch
$ cd opencv
$ mkdir build
$ cd build
$ cmake ../ -DCMAKE_INSTALL_PREFIX="../dist/" -DCMAKE_BUILD_TYPE="Release"
$ make
$ make install
$ cd ../..
The commands will, in order: install some libraries needed by OpenCV, clone the OpenCV official repository, create a "build" and "dist" directory, configure/build/install the software into opencv/dist subdirectory.
At this point, download the latest WASS source code from the official repository and create a build/ subdirectory:
$ git clone https://github.com/fbergama/wass
$ cd wass
$ git submodule update --init
$ mkdir build
To build WASS, on the same terminal, enter the following:
$ cd build
$ cmake ../src/ -DOpenCV_DIR="../../opencv/build"
$ make
$ make install
$ cd ..
If no error occurred, your terminal should be located in WASS_ROOT and all the wass pipeline executables should be located in dist/bin subfolder.
The last step is to install WASSjs. First, install node.js from packages following the instructions on https://nodejs.org/en/download/package-manager/ (if using Ubuntu, proceed to the "Debian and Ubuntu based Linux distributions").
Then, enter the following commands:
$ cd WASSjs
$ sudo npm install
$ cd ext
$ tar xvfz redis-2.8.19.tar.gz
$ cd redis-2.8.19
$ make
$ cd ../../..
WASS should be installed and configured. You can proceed with testing the pipeline.
We suggest to install Docker Desktop (free registration required) and proceed to the Docker installation instructions.
* NOTE: the following procedure was tested on Mac OS X 10.11.5 (El Capitan) but should be valid on any relatively modern version of Mac OSX matching the required libraries*
The easiest way to install all the required libraries is via Homebrew. If brew is not already installed, follow the instructions on http://brew.sh to install the latest version.
Once brew is installed, open a terminal and enter the following commands
brew update
brew install cmake
brew install boost
brew install webp
brew install homebrew/science/opencv3
to install CMake, Boost and OpenCV in /usr/local/Cellar/.
The next step is to install node.js. Download the "Mac OS X Installer (.pkg)" from the nodejs download page and run the installation package. You can verify that node.js is properly installed with the commands:
node --version
npm --version
that should respond with the currently installed version of both node and npm.
At this point, download the latest wass source code from the offical repository (if you haven't done it yet) in a local directory of your choice. For example, to download wass on your home directory enter the commands:
cd ~
git clone https://github.com/fbergama/wass
cd wass
git submodule update --init
To build the wass pipeline programs, on the same terminal, enter the commands:
mkdir build
cd build
cmake ../src/ -DOpenCV_DIR="/usr/local/opt/opencv3/share/OpenCV"
make
make install
you should verify that everything was build properly by typing:
cd ../dist/bin
./wass_prepare
that should output something like:
wass_prepare v. 1.0_heads/master-0-g5a7e63d
---------------------------------------
Darwin-15.5.0 - AppleClang
wass_prepare arguments:
--workdir arg Workdir name
--calibdir arg Calibration data directory
--c0 arg Cam0 image file
--c1 arg Cam1 image file
To build WASSjs, enter the commands (assuming the current directory being <WASS_ROOT>/dist/bin/ ):
cd ../../WASSjs
npm install
cd ext
tar xvf redis-2.8.19.tar.gz
cd redis-2.8.19
make
cd ../..
wass and WASSjs should be now installed and configured. You can verify if everything is working properly by testing the pipeline.
WASS can be built from sources with the latest Microsoft Visual Studio 2015 but requires a prior building of OpenCV, Boost C++ Libraries and CLAPACK.
For this reason, it's higly recommended to install the Binary Version from the download section and unzip the downloaded package on a directory of your choice. We will refer to that directory as <WASS_ROOT> throughout this documentation.
After that, download the latest nodejs for Windows from http://nodejs.org/en/download, choose the 64bit Windows installer for the "current" (non LTS) version, and proceed with its installation.
You can verify that node is installed properly by opening a command prompt and typing:
node --version
Finally, install the Visual C++ 2015 Redistributable x64.
WASS should be now installed and configured. You can proceed with testing the pipeline.
A complete WASS system can be installed in a Docker container. A set of bash scripts simplify the process of creating and running the appropriate container.
Supposing that Docker is already installed and configured for your system, you can simply follow this steps:
$cd ~
$ git clone https://github.com/fbergama/wass
$ cd wass
$ sudo ./Docker/wass_docker_build.sh
This usually requires root permissions but may vary depending on your Docker installation
Once the image is created, supposing that all the WASS configuration files are located in a folder named <config>, the input data in a folder named <in> and you want the output data to be placed in <out>, just run WASS with:
$ ./Docker/wass_docker_run.sh <config> <in> <out>
And open your browser to http://localhost:8080. Additionally, you can spawn a shell inside the WASS Docker container with the command:
$ ./Docker/wass_docker_shell.sh
and use the pipeline manually.
Files inside the container are not preserved if modified. If you need to make changes to settings.json or worksession.json just copy those files in the <config> directory. Upon bootstrapping, they will be copied to the correct location inside the container.