04-12-2012, 12:32 PM
Implementation and Analysis of Image Restoration Techniques
Implementation and Analysis.PDF (Size: 505.69 KB / Downloads: 60)
Abstract
IMAGE restoration is an important issue in high-level
image processing. Images are often degraded during the data
acquisition process. The degradation may involve blurring,
information loss due to sampling, quantization effects, and
various sources of noise. The purpose of image restoration is to
estimate the original image from the degraded data. It is widely
used in various fields of applications, such as medical imaging,
astronomical imaging, remote sensing, microscopy imaging,
photography deblurring, and forensic science, etc. Often the
benefits of improving image quality to the maximum possible
extent for outweigh the cost and complexity of the restoration
algorithms involved. In this paper we are comparing various
image restoration techniques like Richardson-Lucy algorithm,
Wiener filter, Neural Network approach, on the basis of PSNR
(Peak Signal to Noise Ratio).
INTRODUCTION
Images are produced to record or display useful
information. Due to imperfections in the imaging and
capturing process, however, the recorded image invariably
represents a degraded version of the original scene [14]. The
undoing of these imperfections is crucial to many of the
subsequent image processing tasks. There exists a wide range
of different degradations, which are to be taken into account,
for instance noise, geometrical degradations (pincushion
distortion), illumination and color imperfections (under / overexposure,
saturation), and blur [3]. Blurring is a form of
bandwidth reduction of an ideal image owing to the imperfect
image formation process [1]. It can be caused by relative
motion between the camera and the original scene, or by an
optical system that is out of focus. When aerial photographs
are produced for remote sensing purposes, blurs are
introduced by atmospheric turbulence, aberrations in the
optical system, and relative motion between the camera and
the ground [5].
Degradation Model
Capturing an image exactly as it appears in the real world
is very difficult if not impossible. In case of photography or
imaging systems these are caused by the graininess of the
emulsion, motion-blur, and camera focus problems. The result
of all these degradations is that the image is an approximation
of the original. The above mentioned degradation process can
adequately be described by a linear spatial model as shown in
Figure 2.1. The original input is a two-dimensional (2D)
image f(x, y). This image is operated on by the system H and
after the addition of n(x, y). one can obtain the degraded
image g(x,y). Digital image restoration may be viewed as a
process in which we try to obtain an approximation to f(x, y).
given g(x, y). and H and after applying Restoration filters we
obtain restored image f′(x,y) [11].
MOTION ORIENTED ESTIMATION
If we calculate the Fourier spectrum of the motion blurred
images, we can extract good information about the motion
vector. In figure 3 we show the Fourier spectrum of an image
and the Fourier spectrum of the same image blurred with
linear motion. The Fourier spectrum of the motion blurred
image has parallel lines orthogonal (with right angle) with the
motion orientation. So the task of calculating motion
orientation is reduced to the task of calculating the orientation
of these parallel lines. Radon transform [9] is known to be an
efficient method for calculating the orientation of lines in
images. We applied it to find the parallel line orientation. The
result is shown in figure 4.
To determine motion orientation, we should search for the
bright points in figure 4. To determine motion orientation, we
should search for the bright points in figure 4. As we can see
in figure 3-d, the image is blurred with motion orientation of
45±, and we can see one bright point exactly at that angle. For
better distinguishing of these bright points, we find the
maximum response of Radon transform in every angle from 0
to 179 in figure 5.
SIMULATION RESULT
In this section, experimental results are presented which
explored the characteristics of the various restoration
techniques used and tested. The comparative analysis has been
presented on the basis of different percentage of noise for the
standard LENA image which is shown in the Table (1). The
result has been taken by comparing the performance Neural
Network, Lucy, Inverse & Wiener filtering approach on the
basis of PSNR where we are taking different values of SNR
and constant values of angle and length.
CONCLUSION
In this paper, we showed that better restoration based on
neural network with PSNR = 30.1135. as compare to Lucy
Richardson , Inverse filter and Wiener filter.