garystafford / Object Tracking Blogpost

##Object Tracking Project Use C++ with OpenCV and cvBlob to perform image processing and object tracking, using a webcam. Originally, built to test Raspberry Pi performance.

GitHub Branches

Master

Original code for my blog post, 'Object Tracking on the Raspberry Pi with C++, OpenCV, and cvBlob'

rev05_2014

Changes to original project branch:

• Refactored video blob tracking methods into a separate class.
• Refactored hard-coded color ranges. Now input parameters.
• All methods now accept common struct containing input parameters.
• New window displays low and high RGB color range color sample.
• Added support for Travis-CI. Project now building on each commit to GitHub.
• Project contains bash commands/script to install latest OpenCV, cvBlob, and all necessary dependencies on Ubuntu.
• Project contains bash commands/script to git clone, build, package, and run this project on Ubuntu.
• All installation and configuration tested with fresh Ubuntu image on VirtualBox VM

Installation and Configuration

sudo apt-get update && sudo apt-get -y upgrade

# install openCV
cd ~/
sudo apt-get -y install git
# I had a problem with dependencies.sh finding libfaac-dev in multiverse repository.
# Multiverse repository is commented out by default in sources.list.
# Modify /etc/apt/sources.list. See: http://superuser.com/a/467775
sed -i '/precise multiverse/s/# //g' /etc/apt/sources.list
sed -i '/precise-updates multiverse/s/# //g' /etc/apt/sources.list
#sed -i '/precise-security multiverse/s/# //g' /etc/apt/sources.list
sudo apt-get update

git clone https://github.com/jayrambhia/Install-OpenCV
cd Install-OpenCV/Ubuntu
./opencv_latest.sh | tee ~/opencv_install.log

# install object-tracking-blogpostobject-tracking-blogpost  project
cd ~/
git clone https://github.com/garystafford/object-tracking-blogpost.git

#install cvBlob (uses script within object-tracking-blogpost project)
sh /object-tracking-blogpost/install_cvBlob.sh | tee ~/cvblob_install.log

# optional - fixed problem with app finding cvBlob library when started
echo 'export LD_LIBRARY_PATH=/usr/local/lib:$LD_LIBRARY_PATH' | sudo tee -a ~/.bashrc
bash --login # or exit VM and log back in

# build project (compile object tracking application)
cd ~/object-tracking-blogpost/CppAppOpenCV
make -f nbproject/Makefile-Release.mk QMAKE= SUBPROJECTS= .build-conf

# run object tracking application
cd ~/object-tracking-blogpost/CppAppOpenCV/dist/Release/GNU-Linux-x86
./cppappopencv

# package object tracking application for distribution as .tar (incld. static images)
cd ~/object-tracking-blogpost/CppAppOpenCV
bash nbproject/Package-Release.bash

# use packaged object tracking application
cd ~/object-tracking-blogpost/CppAppOpenCV/dist/Release/GNU-Linux-x86/package/
tar -xvf cppappopencv.tar -C ~/

Color Range Examples

These are the values shown in the blog post and YouTube video

• Red balls
  • Low RGB: 155, 0, 0
  • High RGB: 255, 130, 130
• Blue paper
  • Low RGB: 49, 69, 100
  • High RGB: 134, 163, 216
• Green paper
  • Low RGB: 45, 92, 76
  • High RGB: 70, 155, 124
• Orange Pen
  • Low RGB: 25, 90, 56
  • High RGB: 255, 190, 120

Previews