Recently, I was working on some basic Object Recognition and was having trouble figuring out how to get started with OpenCV2 in Ubuntu. There was a lot of disparate advice and documentation across the web, and after spending a little bit of time stitching it together, came up with these 3 steps to getting started with OpenCV2 development using C++ in Ubuntu.
Step One: Installation
Run the following command to install OpenCV2 and all dependencies:
$ sudo apt-get install libcv2.3 libcvaux2.3 opencv-doc libcv-dev libcvaux-dev
Step Two: Header
Much of the documentation that I came across didn’t account for the distinction between OpenCV and OpenCV2, so the headers often looked like this:
#include <cv.h> #include <highgui.h>
After some Googling, I found the following to be correct:
Step Three: Compiling
Compilation also requires linking the OpenCV2 libraries. There is a lot of different documentation on how to do this, however the following seemed to be the most foolproof to me:
$ g++ hello.cpp -o file `pkg-config opencv --cflags --libs`
Step Four: Profit!
That’s it! Keep in mind, I am fairly new to OpenCV, so this may not be the best or most accurate path to getting started, but it worked for me and I felt I should share it.