International Journal of Information and Computation Technology.
ISSN 0974-2239 Volume 4, Number 3 (2014), pp. 247-252
© International Research Publications House
http://www. irphouse.com /ijict.htm
Blurred Image Enhancement Using Contrast Stretching,
Local Edge Detection and Blind Deconvolution
Prasad Nagelli
1
, C. Lokanath Reddy
1
and B.T.R. Naresh Reddy
1
1
Department of Computer Science & Engineering, Vaagdevi Collge of Engineering,
Bollikunta, Warangal, Andhra Pradesh, India.
Abstract
Blurring of image is common problem while taking picture of an
object in motion or due to shooting situations. Various methods have
been proposed to enhance the blurred image. Here contrast stretching
is used for obtaining deblurred image. In the proposed method local
edge detection is applied on original as well as contrast stretched
image. The set of edges obtained from both the images are fussed in
order to get sharper edges. The original image and contrast stretched
image is converted into gray scale image from RGB image before
applying local edge detection to avoid detection of false edges. Since
image distortion information is unknown, so on the obtained fussed
image blind deconvolution is applied to get deblurred image.
Keywords: Blur image enhancemen; local edge detection;contrast
stretching; image fusion.
1. Introduction
Edge detection is one of the common tools for feature detection and feature extraction
of an image (Marr et al, 1980; Martin et al, 2004; Papari et al, 2011). It is a process of
identifying the point in the digital image so that it can be modified in more sharp
output. It has a wide application in the field of image retrieval (Moreno et al, 2009),
object recognition (Olson, 1997) and object tracking (Sullivan, 2002). Theoretically
the output of an ideal edge detection algorithm is an object boundary having
continuous contours. It is very difficult to detect object boundary which become more
complex if image is noisy or blurry (Umgaugh, 2005).
Blurring of image is degradation which can occur in many situations due to
unavoidable image conditions (Banham et al, 1997; Gonzalez, 1992; Tikhhonov, 1977)
.The degradation model can be expressed as