Augmented Reality imageRecognition [closed]

Question

Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.

---

Questions asking us to recommend or find a tool, library or favorite off-site resource are off-topic for Stack Overflow as they tend to attract opinionated answers and spam. Instead, describe the problem and what has been done so far to solve it.

Closed 8 years ago.

Improve this question

I searched some docs for create imageRecognition. But i didn't get correct path to study for beginner. I need to create above image. I stored my image in database. When image matched i need to display URL in cameraView. How to do? Is any open source available

Answer 1

Doing that with some picture is complicated. Usually you use black and white images, QR code like.

Maybe you can search here : http://code.google.com/p/andar/

Answer 2

check opencv features2d module, especially sift, surf and homography methods.

after you find the region of the feature, you can place your AR object onto it using openGL methods.

you may need to search and read tutorials about both openCV and openGL libraries.

Answer 3

Welcome to the world of computer vision - you will encounter much satisfaction and much frustration. There are a few problems you are facing here. Firstly, it isn't a straight image recognition problem because you view the image not necessarily face-on at 90 degrees. I suggest you first detect the orientation of the surface before you, then apply a 3D transform so that the surface is perpendicular to your gaze direction, then do regular 2D image recognition on that. Even with the last step you still have to be able to detect at different scales and orientations.

I don't know specific good techniques for the first two steps, but you can google.

For the last stem, you can search for Viola-Jones detector (2001), which will work for upright orientation. Also just search "image recognition", see what you get. Alternatively, if you want to try writing your own, you can start by calculating a "fourier power spectrum" which is invariant to size and location. Google more for what that means.

For implemenations of a lot of algorithms, google "openCV".