Noise Reduction For Digital Images Using Median Filtering Technique
Nowadays, the popularity of digital devices increases. However there are still some limitations in digital technology. One of the limitations is the appearance of additive noise in the images that are acquired using long exposure times. Long exposure times are needed in the case when we need t...
| Main Author: | |
|---|---|
| Format: | Monograph |
| Language: | English |
| Published: |
Universiti Sains Malaysia
2006
|
| Subjects: | |
| Online Access: | http://eprints.usm.my/58745/ http://eprints.usm.my/58745/1/Noise%20Reduction%20For%20Digital%20Images%20Using%20Median%20Filtering%20Technique_Teoh%20Sin%20Hoong.pdf |
| Summary: | Nowadays, the popularity of digital devices increases. However there are still some
limitations in digital technology. One of the limitations is the appearance of additive noise
in the images that are acquired using long exposure times. Long exposure times are needed
in the case when we need to take an image under the conditions that have a low level of
illumination, such as during the night time or in a large room such as in an auditorium.
Forensics images are also included as the images which need a long exposure time. Noise
that is often appears in such conditions is “salt-and-pepper” noise. This noise is called “salt
and pepper” because the granular appearance of individual points of noise in the signal. The
objective of this project is to find the best way to reduce this type of noise in digital images.
From various branched of filter in image processing field, I had chosen median filter as the
technique to reduce noise by using Borland C++ compiler 5.5 to complete the project. This
nonlinear technique can remove single-point noise caused by experimental errors, or other
sampling errors that occur at single points due to bad pixels or other causes. A mean square
error measurement was used for comparison of the results. There are 13 different median
filtering technique used in the project which are square with causal and iterative median
filtering, circle with causal and iterative median filtering, stick with causal and iterative
median filtering, square with causal and non-iterative median filtering, circle with causal
and non-iterative median filtering, stick with causal and non-iterative median filtering,
square with non-causal and iterative median filtering, circle with non-causal and iterative
median filtering, stick with non-causal and iterative median filtering, square with non-causal and non-iterative median filtering, circle with non-causal and non-iterative median
filtering, stick with non-causal and non-iterative median filtering and progressive switching
median filtering. At the end of the project, I suggest the best filter which can used to reduce
“salt and pepper” noise in the image is progressive switching median filtering technique.
Experiments have shown this technique extensively reduces the noise in an image with no
obvious loss of image sharpness. |
|---|