Image Processing Reference
In-Depth Information
In the SampleImages folder on this topic's DVD, you can ind the image noise.
png . The photo is a night shot originally saved as a JPEG. The image has been
brightened with the Levels tool. This increased the noise contamination in the
First we will apply the Despeckle filter ( Filters > Enhance > Despeckle ). The
filter gives a soft-focus effect that seems to smudge the “flecks” in the image.
When you first open the filter, the window with the available settings
appears. At the top you find the Median setting, which offers two choices:
Adaptive and Recursive . For now, choose the Adaptive setting, which will adapt
the radius setting to the contents of the image or selection by applying a
histogram from the image. Generally, this setting will lead to a better result
that setting the radius by hand. The Recursive setting automatically applies
the filter several times. Unless you want a distorted effect, I don't recommend
using this setting.
You can also configure the following settings:
Radius: The radius relates to the size of the section the filter computes
with. Keep the radius as small as possible; otherwise, you might destroy
details. Let the program choose the radius.
Black level: A small value (0-20) slightly darkens light pixel noise.
White level: A high value (240-255) preserves bright details. Both settings
eliminate noise contamination that are close to pure white or black. Edit your
image with the filter's default settings. If these don't bring the desired effect,
you can experiment.
The filter will be applied when you click the OK button and the image will
be rendered.
The filter affects the noise.png image only slightly. The noise is reduced
only in the shadows. Even several applications of the filter won't improve the
image. The noise contamination is too high for the filter to have an effect.
However, there are other options to “smoothen” the picture.
The Despeckle filter is also suitable for removing small blemishes such as
dust or scratches that occur from scanned images. In addition, it can remove
the moiré effect that occurs from scanning printed images.
Search WWH ::

Custom Search