HARFA: SCREENSHOTS AND TUTORIALS: SMOOTHMEDIAN |
|
Filtration
SmoothMedian
Linear filters (SmoothBox, SmoothBinomial) are very powerful to remove the Gaussian noise from images. This effect is paid by blurring image. These filters are not effective in the case of binary noise (pepper and salt), because they only spread noise in to the surrounding pixels. To suppress the binary noise, using of rank value filters is very advantageous. One of them is the Median filter. The median filter assigns the changed pixel median of its surrounding. A median is a number, which lies in the middle of data, which are sorted to their size.
![](../images/screenshots/med/schmed.gif)
The scheme of the Median filtration
The HarFA user can alter filtration matrix radius and by this the amount of the smoothing effect. We can enter values from 1 to 15.
![](../images/screenshots/ltest.jpg)
The original image, the median filter with filter matrix of radius 2, 5, 7, 10. (place your mouse on the numbers)
Previous
Next
|
|
|