0

I know there exist many many books about image processing but I need an advise for a particular good one giving practical hints for using the algorithms. I don't need background information about HOW an algorithm works, e.g. HoughTrafo or Canny Filter as I know that already from various books. But I need a good advise on how to use those filters efficiently and in particular on how to set the thresholds etc.

It currently gives me a huge headache on how to chose those values. When I set them to fixed values, they work for one picture and when changing the illumination slightly, the dont work anymore for various reasons. So I wonder on how to dynamically set them from image specific values. I read on SO to e.g. set the canny thresholds to:

low = 0.666*mean(img)
high = 1.333*mean(img)
(http://www.kerrywong.com/2009/05/07/canny-edge-detection-auto-thresholding/)

but somehow I havent had much success with it so far. I'm interested in good advises for books etc in particular but included the special example on how to determine thresholds for canny to make it a valid SO-question :-)

tim
  • 9,896
  • 20
  • 81
  • 137

1 Answers1

1

If you want a practical book, typically the book will be targeted to a specific library.

Here you can find a list of most of the books about the OpenCV library.

Rui Marques
  • 8,567
  • 3
  • 60
  • 91
  • Ah true, that makes sense. Problem is that I'm not yet sure on wether I'd like to use OpenCV and the algorithms of another library might be different internally. But true, for the choice of thresholds e.g. this should fit in all sorts of librarys :-) – tim Mar 22 '14 at 11:36
  • I think some thresholds might be implementation dependent, so your best bet is to read something targeting a library. What are your doubts regarding the choice of a library, what are your requirements? – Rui Marques Mar 22 '14 at 11:38
  • I'd like to do an android app, capturing a single camera image of a rectangle-sized sheet of paper, detecting its borders (Something like Canny/Edge Detection (maybe a thresholding-technique before!?) with a subsequent Hough transformation) and then transform the image perspectively to cut off everything which doesn't belong to the sheet of paper. The algorithm should be nearly identically to the one that CamScanner is using (you know it?)... I think OpenCV might be a bit of overhead for this and not too easy to port/install on android as well. – tim Mar 22 '14 at 11:47
  • 1
    Well I think OpenCV will be ideal for that. They have an Android version, which is not difficult to setup if you follow the install tutorial. You will have Canny, Hough tranforms, and many other things you can use. I do not know about CamScanner but you have questions here on SO asking exactly about that problem (rectangle borders, "warp perspective" = align rectangle corners with full camera image). – Rui Marques Mar 22 '14 at 11:53