-4

I'm working in a method of counting the number of hats in a image.

I slice a binary image vertically in multiple sections, around 250/200, and then the counting of white elements (hats edges) is done in each os these sections. In the end of this process, the value obtained more times will be the result.

My struggle is that, often, the edges got some big gaps due to the binarization process. This way, in many sections, those hats are not counted, which causes imprecisions in the final counting.

Does anyone can suggest a way of surpass this problem? I can post my code if is needed.

Thank you!

This is the original image. Here's the binary image which is divided in multiple vertical sections

Brad Larson
  • 170,088
  • 45
  • 397
  • 571
Thiago Resende
  • 177
  • 1
  • 1
  • 8
  • Binarize the image and use sketelization as explained here http://homepages.inf.ed.ac.uk/rbf/HIPR2/skeleton.htm – flamelite Jan 10 '18 at 06:20

1 Answers1

2

You can :

If the results aren't good, you need to try more complex process as Machine learning

Elodie Dellier
  • 114
  • 6