Skip to main content
28 events
when toggle format what by license comment
Sep 30, 2013 at 11:57 comment added user5556 i want make a contour of a shape of forme complexe .i want use snake svp give me a code ?
Oct 27, 2012 at 13:13 history edited mchlfchr CC BY-SA 3.0
added 174 characters in body
Oct 27, 2012 at 13:04 history edited mchlfchr CC BY-SA 3.0
added 146 characters in body; Post Made Community Wiki
Oct 26, 2012 at 15:03 answer added mkuse timeline score: 1
Oct 15, 2012 at 8:06 history edited mchlfchr CC BY-SA 3.0
added 137 characters in body; edited title
Oct 15, 2012 at 7:54 history edited mchlfchr CC BY-SA 3.0
removed 1st approach, formatting stuff
Oct 14, 2012 at 17:18 history edited mchlfchr CC BY-SA 3.0
added description on contour detection screenshot (yellow rectangles and blue line)
Oct 14, 2012 at 13:59 comment added mchlfchr @Andrey I put some further pictures into my post. The examples i've posted are the worst cases of this scenario. The glass (Image 2) is getting molten by the way, but I have not pictures at home of that melting process.
Oct 14, 2012 at 13:52 history edited mchlfchr CC BY-SA 3.0
formatting stuff, added further examples
Oct 14, 2012 at 13:33 history edited mchlfchr CC BY-SA 3.0
added new approach, some formatting stuff
Oct 14, 2012 at 13:04 comment added Andrey Rubshtein The second image you added looks totally different. Can you post all possible images?
Oct 13, 2012 at 19:13 history edited mchlfchr CC BY-SA 3.0
added 21 characters in body
Oct 13, 2012 at 18:52 history edited mchlfchr CC BY-SA 3.0
added further images for better explanation of goals
Oct 13, 2012 at 17:23 history edited mchlfchr CC BY-SA 3.0
added 72 characters in body
Oct 13, 2012 at 16:01 answer added Martin Beckett timeline score: 0
Oct 13, 2012 at 15:03 answer added Andrey Rubshtein timeline score: 2
Oct 13, 2012 at 12:26 answer added Yves Daoust timeline score: 1
Oct 13, 2012 at 9:47 history edited mchlfchr CC BY-SA 3.0
added new approach
Oct 13, 2012 at 0:03 history tweeted twitter.com/#!/StackSignals/status/256907878479314944
Oct 12, 2012 at 20:22 history edited mchlfchr CC BY-SA 3.0
added 89 characters in body
Oct 12, 2012 at 19:03 comment added mchlfchr Hi Phil, as far as I am concerned and as I know, there is a series of pictures who are shot before an object gets into the furnace. Kind of calibration is therefore here. I'll talk to the physicist who is responsible for the CCD and the environment on monday. But thanks for the advice, I'll give it a shot!
Oct 12, 2012 at 18:25 comment added PhilMacKay Have you tried some basic image reduction techniques? Such as subtracting a flat-field correction to get rid of that darkening on the top right (en.wikipedia.org/wiki/Flat-field_correction). Also, if the particles are static, they would instantly be removed. Then, you could use any edge detection method you want...
Oct 12, 2012 at 18:09 comment added mchlfchr Hi Rui, I tried adaptive thresholding with this Matlab Extension: Adaptive Thresholding Here are the results for playing around with the window parameters: 1st Try 2nd Try 3rd Try As you can see, with increasing parameters the middle part goes from white to black (what is good, imho) but the rest of the background also turns into black, what is bad.
Oct 12, 2012 at 17:17 comment added Rui Marques Have you tried adaptive threshold? You don't seem to mention it. I think OTSU should work somehow after removing noise but maybe adaptive threshold is better.
Oct 12, 2012 at 16:34 history migrated from stackoverflow.com (revisions)
Oct 12, 2012 at 15:32 answer added oli timeline score: 3
Oct 12, 2012 at 15:12 answer added Dave Durbin timeline score: 0
Oct 12, 2012 at 15:07 history asked mchlfchr CC BY-SA 3.0