Re: Question on despckle relative big speckle in grayscale image

On Aug 26, 3:41 pm, xhm <xhm2...@xxxxxxxxxxx> wrote:
On Aug 26, 2:54 pm, Martin Leese <ple...@xxxxxxxxxxxxxxxxxxxxxxxxxx>
wrote:



xhm wrote:
On Aug 25, 1:36 pm, Martin Leese <ple...@xxxxxxxxxxxxxxxxxxxxxxxxxx>
wrote:
...
If you have not
already done so, try using a structure
element which has the size, shape,
orientation, and fuzziness of the dark spot.
Thanks for the help and confirmation of my problem is challenging,
Martin.

Can you explain a little bit more about the fuzziness of "If you have
not
already done so, try using a structure element which has the size,
shape,
orientation, and fuzziness of the dark spot.". Currently, I only used
a disk structure element with radius of 5. I.e., the se = strel('disk',
5);
I think I can use SE = STREL('arbitrary',NHOOD) creates a flat
structuring element with  the specified neigbhorhood.  NHOOD is a
matrix containing 1's and 0's; to try what you've suggested. But how I
can define the fuzziness, do you mean that I can pad some 0s around or
in the element?

The difficult is also that I need to keep the grayscale image for
later processing. Otherwise, I may try to convert to binary and
discard the small sized blobs.

Forget fuzziness.  I was tired.

 > One more thing to mention. Even if I can create specific structure
 > element for this image and assume it will work (will try), but since I
 > have a lot of images and all the dark small spots are not the same and
 > I don't know what those will be, so this specific structure element
 > will be hard to obtain satisfied results for all images.

If your noise (dark spots) is not consistent
then you are going to achieve poor results
and be left with artifacts.  The underlying
problem is that your "dark small spots" are
not small.  They are actually quite large.

--
Regards,
Martin Leese
E-mail: ple...@xxxxxxxxxxxxxxxxxxxxxxxxxx
Web:http://members.tripod.com/martin_leese/

Thanks for more help, Martin.

I know that my "small" dark spot noises are yet relatively "big",
that's why it is hard as you've confirmed. Now I hope ImageAnalyst can
clarify further on his background modeling method, as his method seems
very similar to mine and promising, and hopefully that will work on my
image too.

xhm

One correction, it should be "his application (noises) seems very
similar to mine" rather than "his method ......".
.

Relevant Pages

  • Re: Question on despckle relative big speckle in grayscale image
    ... orientation, and fuzziness of the dark spot.". ... If your noise (dark spots) is not consistent ...
    (sci.image.processing)
  • Re: Question on despckle relative big speckle in grayscale image
    ... Can you explain a little bit more about the fuzziness of "If you have ... orientation, and fuzziness of the dark spot.". ... If your noise (dark spots) is not consistent ...
    (sci.image.processing)
  • Re: Bad Vision
    ... Its mainly in low light conditions and on a light background the spot ... is dark and on a dark background the spot is white or grey. ... light for the cones to work properly. ...
    (sci.med.vision)
  • Re: Cateye Double Shot as main headlamp?
    ... does throw farther, but as you say, it has an even smaller spot of light ... but the same beam shape rather than making the EL-530 ... So it's fine for riding carefully at low speeds around town, ... I was caught out mountain biking after dark a couple of weeks ago. ...
    (rec.bicycles.tech)
  • Re: Question on despckle relative big speckle in grayscale image
    ... be able to remove the dark spot without ... Can you explain a little bit more about the fuzziness of "If you have ... already done so, try using a structure element which has the size, ...
    (sci.image.processing)
Quantcast