0

I have a university work in C#: to make a 3D approximation from a single scenery photo, at least a crude one, and I am not allowed to use libraries for image segmentation.

At this point, I have a function that segments image into colored segments and assigns color of the segment's root to all points in the segment for better distinction, like so: badly segmented image on puush

Given such segmented image, is it possible to find contour points for these colored segments using EmguCV ? My current aim is to try and use EmguCV to find segment contour points, assign them to lists and then use resulting polygons in SharpGL to make 3D scene.

fwiffo
  • 45
  • 6
  • A simple google search returned this SO QA: http://stackoverflow.com/questions/24970577/finding-contour-points-in-emgucv. There is a lot that comes up. – KDecker Jan 13 '16 at 17:11
  • It's a little difference in a sense that we split black segment from white background. I've seen that question as well. At the moment, I'm thinking of making every segment black (and the others white) one by one in turn, finding its contours and then combine that all into one piece (or use my own little hand-written algorithm I used for initial tests). I was hoping that there was a better way of doing this. – fwiffo Jan 13 '16 at 20:01

0 Answers0