jeff1evesque / iscanner Goto Github PK

Since we intend to develop non-UI logic with C++ (Objective-C for front-end), we will develop within a linux virtual machine. This will allow us to separate dependencies from different projects, and easier to deploy in a minimalistic fashion (not installing unneeded packages). Specifically, we will install, and setup Vagrant.

Note: this issue corresponds to jeff1evesque/machine-learning#1393.

After #4 has been resolved, we will implement Jenkins in order to perform automated unit testing.

Note: this issue corresponds to jeff1evesque/machine-learning#204

We will add the /log/ directory to.gitignore`.

We need to add .vagrant/ to the .gitignore.

After #46 has been resolved, we will redirect the following commands in install_opencv.pp:

• cmake
• make
• make install

to its own corresponding log file, in the /log directory.

After #4 has been resolved, we will implement puppet in order to manage various configurations.

Note: this issue corresponds to jeff1evesque/machine-learning#1351

We will define the necessary configurations that will allow port forwarding, such that our Virtual Machine will be accessible outside of VirtualBox.

We need to install the C++ gcc compiler:

sudo apt-get install g++

Note: we will use C++ for non-UI code, which will allow a majority of the code base to have the possibility of being reused for android based devices.

Determine if OpenCV should be installed on the iOS device. If not, then we can send images in real time to a server implementing iris-recognition.

After #6 has been resolved, we will create the subsection Vagrant within the README.md documentation. This will be a subset-specific of the corresponding wiki page from the machine-learning repository.

We will install a minimal gnome UI for the Ubuntu Server.

We will create an implementation of vagrant, therefore, create a vagrantfile.

After #8 has been resolved, and jeff1evesque/machine-learning#1931 has been resolved, we will create the subsection Puppet within the README.md documentation.

We will temporarily allow the VirtualBox GUI to display each time the VM is started.

We will install guest additions within our puppet manifest default.pp. This will Enable USB Controller, and eliminate the 1394 error:

libdc1394 error: Failed to initialize libdc1394

To install OpenCV, we need the following:

Install dependencies

sudo apt-get -y install libopencv-dev build-essential cmake git libgtk2.0-dev pkg-config python-dev python-numpy libdc1394-22 libdc1394-22-dev libjpeg-dev libpng12-dev libtiff4-dev libjasper-dev libavcodec-dev libavformat-dev libswscale-dev libxine-dev libgstreamer0.10-dev libgstreamer-plugins-base0.10-dev libv4l-dev libtbb-dev libqt4-dev libfaac-dev libmp3lame-dev libopencore-amrnb-dev libopencore-amrwb-dev libtheora-dev libvorbis-dev libxvidcore-dev x264 v4l-utils unzip

Download OpenCV

mkdir opencv
cd opencv
wget https://github.com/Itseez/opencv/archive/3.0.0-beta.zip -O opencv-3.0.0-beta.zip
unzip opencv-3.0.0-beta.zip

Install OpenCV

cd opencv-3.0.0-beta
mkdir build
cd build
cmake -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=/usr/local -D WITH_TBB=ON -D BUILD_NEW_PYTHON_SUPPORT=ON -D WITH_V4L=ON -D WITH_QT=ON -D WITH_OPENGL=ON ..
make -j $(nproc)
sudo make install

System Update: to get OpenCV working properly, we need to tell Ubuntu.

sudo /bin/bash -c 'echo "/usr/local/lib" > /etc/ld.so.conf.d/opencv.conf'
sudo ldconfig

• http://rodrigoberriel.com/2014/10/installing-opencv-3-0-0-on-ubuntu-14-04/

Our C++ compiler will compile files into executable files with the .exe file extension. We will ignore all executable files in the main application directory.

We have currently implemented OpenCV version 3.0.0-beta. However, we will update the manifest file install_opencv.pp to build the official OpenCV 3.0:

We need to resolve the issue of being able to mount a USB drive from the host machine, to the guest vagrant virtual machine.

After #28 has been resolved, and jeff1evesque/machine-learning#1937 has been resolved, we will create a subsection Jenkins within the README.md documentation.

We will rename default.pp to install_opencv.pp. Then, we will need to adjust the Vagrantfile correspondingly.

After #30 has been resolved, we will create a corresponding section within the README.md.

We will create a puppet/ directory, and move the manifests/ directory into it. Then, we need to adjust the Vagrantfile, respectively.

Note: we will implement the following directory structure:

• Vagrantfile
• puppet
  • manifests
    • install_opencv.pp
  • modules
    • jenkins
• README.md
• detector.cpp

After #30 has been resolved, we will create a wiki page on OpenCV

We need to install the git package within default.pp:

package {'git': ensure => present, }

which will be responsible for installing git. This file will need to be referenced within the vagrantfile as follows:

Vagrant.configure("2") do |config|
...
  config.vm.provision "puppet" do |puppet|
    puppet.manifests_path = "/manifests"
    puppet.manifest_file = "default.pp"
  end
...
end

Note: by default, the entry point is site.pp for a simple puppet implementation. However, when implementing puppet within vagrant, the entry point can be explicitly defined (as above).

We need to implement the VideoCapture class in detector.cpp. This file will capture video from the default camera.

We need to discuss the following about the gcc compiler:

• g++ treats everything as c++ even when the extensions suggest c code (unless told otherwise), and link c++ standard libraries automatically. gcc treats everything according to known extensions (unless told otherwise) and does not automatically link c++ standard libraries.
• using the -o flag to specify output file name:
  • g++ -o detector.exe detector.cpp pkg-config --cflags --libs opencv
• run executable
  • CMD shell: simply run detector
  • Bash (or Bourne) shell: simply run ./detector
• usage for pkg-config while compiling.
  • pkg-config --cflags opencv prints the list of arguments needed to allow the compiler to find the header files, and make necessary links to OpenCV
  • pkg-config --libs opencv prints the commands needed for linking
  • The following are equivalent:
    • g++ -o detector.exe detector.cpp``pkg-config --cflags --libs opencv
    • g++ -o detector.exe detector.cpp $(pkg-config --cflags --libs opencv)

We will provide a feature that will perform facial recognition on images, similar to the VideoCapture class implementation.

We will describe the intention of using C++ for the back-end logic, is to allow the capacity to reuse code, perhaps for android based devices. For iOS devices, Obj-C would need to be used to develop the user-interface, while for android devices, the backend logic could be extended with additional C++, unless Java is preferred. However, the latter (android topic) is not a concern for this repository.