Ok some suggestions on where you should start:
- Make sure you master C++ programming. If not, start at the very beginning by looking at the official C++ tutorial and starting guide.
- Make sure you understand the basics of OpenCV, like how images are stored, how to process pixels, how to create a project, ... For that I suggest going to the tutorial page and start by going through the first 2 - 3 sections.
By now you are ready to move to something more difficult and to tackle the challenge of your project. Start by writing out step for step on what you want to accomplish. I can give you some suggestions on where to look:
- For plan extraction in 2D you will need to look for lines in images. Try edge detection techniques combined with for example the HOUGH line detector. It can get you pretty neat results.
- You will have to look for objects. If they have always the same look, you should try the bag of words approach of 2D feature matching. If you have variance in the appearance, try moving to for example cascade classification.
Basically, pushing a project forward is trial and error. Make the best of it! And feel free to ask for problem solutions.
