under the guidance OfMs.Y.HarshalathaAsst. proffesor,Dept. of E&C SIT,Tumkur
Presented By:Ambuj Kumar 1SI08EC011 Chirag Gupta 1SI08EC031 Navneet Ranjan 1SI08EC069 Neeraj Kumar 1SI08EC070
Objective Modeling of Noise Effects of noise on an image Concepts of random process Basic Degradation model Filtering Techniques used
a)Spatial domain filtering b) Frequency domain filtering
MATLAB implementation Processor implementation ConclusionDepartment of E&C,SIT,TUMKUR2
Objective of restoration To improve the quality of a digital image which has been degraded
Various phenomena which degrade the image Motion Improper focusing of Camera during image acquisition. Atmospheric turbulence Noise
Department of E&C,SIT,TUMKUR
3
Noise is defined as any degradation in an Image caused by external disturbance. The different types of noise effecting an image are given below
Salt and pepper noise Gaussian noise Speckle noise Shot noise Film grain noise Quantization noise
Department of E&C,SIT,TUMKUR
4
Original image
Image with Salt & pepper
Image with Gaussian noise
Department of E&C,SIT,TUMKUR
5
Original
With Gaussian noise
After noise removal6
Department of E&C,SIT,TUMKUR
6
Impulse Noise Reduction Use nonlinear spatial filtering to remove impulsive noise without reducing resolution
Original
With impulse noiseDepartment of E&C,SIT,TUMKUR
After noise removal
7
Any type of noise can be modeled by a Random process. The basic concepts of Random process are as discussed below
Random variable, its PDF and CDF Random process, its joint PDF and CDF Concepts of stationarity WSS process Strict sense stationarity
Department of E&C,SIT,TUMKUR
8
Fig. 3.1 Common image acquisition system (Courtesy: S. C. Park et. al.[4])
Department of E&C,SIT,TUMKUR
9
Spatial filtering requires following steps position the mask over the current pixel. form all products of filter elements. add up all the products. This must be repeated for every pixel in the image.
Basic restoration filters in spatial domain can be broadly categories as below Mean filters Order-static filters
Department of E&C,SIT,TUMKUR
10
Department of E&C,SIT,TUMKUR
11
What do frequencies mean in an image? High frequencies corresponds to pixel value that change rapidly across the image Strong low frequencies corresponds to large scale features in the image
Since noise is a high frequency signal we can use low pass filters to remove noise.
The basic low pass filters are as follows Ideal low pass filter Gaussian low pass filter Butter worth low pass filter
Department of E&C,SIT,TUMKUR
12
Fourier TransformPreprocessing
Filter function H (u, v)
Inverse Fourier Transform
PostProcessing
f(x, y)
g(x, y)
Department of E&C,SIT,TUMKUR
13
Department of E&C,SIT,TUMKUR
14
Department of E&C,SIT,TUMKUR
15
Department of E&C,SIT,TUMKUR
16
