Open AccessEnhanced Image Inpainting in Remotely Sensed Images by Optimizing NLTV model by Ant Colony Optimization
Author(s) -
Manjinder Singh,
Harpreet Kaur
Publication year - 2016
Publication title -
international journal of electrical and electronics research
Language(s) - English
Resource type - Journals
ISSN - 2347-470X
DOI - 10.37391/ijeer.090306
Subject(s) - ant colony optimization algorithms , artificial intelligence , computer science , regularization (linguistics) , inpainting , total variation denoising , pixel , computer vision , image (mathematics) , variation (astronomy) , similarity (geometry) , pattern recognition (psychology) , algorithm , physics , astrophysics
Filling dead pixels or eliminating unwanted things is typically preferred within the applications of remotely sensed images. In proposed article, a competent image imprinting technique is demonstrated to resolve this drawback, relied nonlocal total variation. Initially remotely sensed images are effected by ill posed inverse problems i.e. image destripping, image de-noising etc. So it is required to use regularization technique to makes these problems well posed i.e. NLTV method, which is the combination of nonlocal operators and total variation model. Actually this method can make use of the good features of non-local operators for textured images and total variation method in edge preserving for color images. To optimize the proposed variation model, an Ant Colony Optimization algorithm is used in order to get similarity with the original image. And evaluate the outcomes of proposed technique with the existing technique i.e. MNLTV optimized by Bregmanized-operator-splitting algorithm which is a prediction based method. The investigation of all outcomes confirms the efficacy of this rule.