Geek asks and answers

Image Recognition Maze

I am trying to do a project with some of my friends, and we came to this:

Let's say I had to decrypt this maze programmatically, how could I continue about this? My first solution when trying to solve the mazes by image recognition, obviously, simply draws an open path, so the end of the paintwork (the arrow at the beginning of the maze exists to provide a recognition method to see “good that start”) indicates an exit.

The problem is that with these filters I cannot draw it, and I have no other idea on how to solve it. So, will there be any way to do this using Open CV? (or any other option, if possible, will be fine too)

I really don’t know how to solve this problem, so if it’s possible, just point me to the option and I will investigate this more.

Many thanks.

+5
source share
4 answers

For simplicity, consider the color space that the channel provides, where this kind of noise is useless. For example, if we take a channel Sfrom HSB, we get an image on the left, which is easily binarized using the Otsu image on the right.

enter image description hereenter image description here

, , . , ( ) . , , , , - .

enter image description hereenter image description here

, , . . , , , .

enter image description here

+16

, . , 2 , . , .

, , , . -, .

, , . , , , A, D .

+1

: . " " . , . , . , , ( ). , U ( , ).

. , , - .

0

I believe that you are talking about "finding the way"

The Wikipedia entry has a simple algorithm for finding paths on the grid: http://en.wikipedia.org/wiki/Pathfinding , but googling should provide some code samples in your language of choice.

-1
source

More articles:


All Articles