Basically there are some steps to follow:
- Make sure you have some basic C++ background. If not, read about the basic techniques at the official C++ manual tutorial.
- Go to the official openCV tutorial page, start with the very first ones and work your way down. It will help you with the very basics, like pixel manipulation, then move to more exciting stuff. Expecially the core module will set you on your way with the basics.
- Use the available books that are published, which guide you step by step through the code of some larger projects, giving you insight in more complex computer vision programming. You have the starter book, which was just released. But there are also some more detailed books, like this one.
And still then, you will only get more experience by hands on training. Most of us have multiple years of programming in OpenCV and still have troubles with some problems. Just keep learning and combining stuff to get to new programming levels.
