Matching ranges over many dimensions

edit

Hmm, for binary images features2D are probably not the best choice but maybe he has his lanes also in grayscale, then this will work nicely. @JamesHammond: All in all your description of your problem is too vague, formulate your question more comprehensively! And before that - have a look at the docs and this Q&A forum!

Thats just it, you can assign a single matcher to an own created descriptor vector. Its just bits and bytes. Matcher only uses a distance metric.

Definitly!

Hi, I've added a bit of detail to my question. Given the depiction, is 'assign a single matcher to an own created descriptor vector' the way to go?

Basically make a 0 - 1 string of your lanes. Then use a matcher to check the hamming distance between both. This link explains hamming distance a bit more, which is basically an equality measure for input data : http://www.princeton.edu/~achaney/tmve/wiki100k/docs/Hamming_distance.html