Adaptive probability filter for removing salt and pepper noises using matlab image processing

ABSTRACT: 

        Salt and pepper noises based on the characteristic of minimum and maximum intensity values of the images, as well as the distribution of noise. The median of noise-free pixels in neighbourhood is used to remove noise. The proposed method is capable of detecting noise more accurately and perform much better than the existing distinguished filters in terms of peak-signal-to-noise ratio, image enhancement factor, and visual representation at all the noise densities.
EXISTING METHOD

           The salt and pepper noises fall into the former type and take the minimum and maximum intensity values randomly over an image. Median filters are profoundly used in removing salt and pepper noises due to their good performance and low computational complexity. While a standard median filter can denoise effectively at low noise densities, it shows poor performance at high noise densities. Some improved versions of the standard median filters have been made by giving more weights to certain selected pixels in neighbourhood such as weighted median filter and centre weighted median filter, as well as adaptive median filters the neighbourhood sizes of which vary adaptively with the noise densities. These filters deal with all pixels uniformly without protecting noise-free pixels.

DISADVANTAGES

Low accuracy.
Computational complexity.
Low PSNR and MSE.

PROPOSED METHOD


A salt and pepper noises over an image are regarded randomly distributed with values 0 for ‘salt’ pixels and 255 for ‘pepper’ pixels; the pepper and the salt noises are assumed to be equally distributed. There is a strong correlation among noise-free pixels in a neighbourhood, so the noise-free pixel with intensity 0 (or 255) is not isolated, and the intensities of its neighbouring pixels are likely to close to 0 (or 255).It is noteworthy that in a white neighbourhood corrupted by salt and pepper noises, the salt noise is assimilated and lost; all the pixels with intensity 255 are taken noise free, while all the pixels with intensity 0 are noise. Similar for the blacks. The proposed APF in terms of peak-signal-to-noise ratio (PSNR), image enhancement factor (IEF), and visual representation at all noise densities.

ADVANTAGES

Better accuracy.
Low Computational complexity.
Better PSNR and MSE.

FLOW DIAGRAM

ALGORITHM DETAILS

Here we are using add noise for both existing and proposed method. In existing method we use denoising by Median filter proposed method for using looping algorithm that is called APF method to denoising and get better performance.

SYSTEM REQUIREMENTS:
HARDWARE REQUIREMENTS
Processor Type : Pentium -IV
Speed : 2.4 GHZ
Ram : 128 MB RAM
Hard disk                             : 20 GB HD

SOFTWARE REQUIREMENTS
Operating System : Windows 7 
Software Programming Package : Matlab R2014a

REFERENCES

Roy, A., Singha, J., Manam, L., et al.: ‘Combination of adaptive vector median filter and weighted mean filter for removal of high-density impulse noise from colour images’, IET Image Process., 2017, 11, (6), pp. 352–361