Novel Approach for Iterative Image Fusion using Fuzzy and Neuro Fuzzy Logic.

Abstract

Image fusion may be defined as an approach of merging two or more input images in to a single fused image which is comparatively more informative than all the input images. Image fusion aims to reduce the redundancy and uncertainty in the output and aims to maximize application related significant information. Fused images are used in diverse applications such as remote sensing, machine vision, biometrics, medical imaging and military domain. In this paper, we carry out iterative image fusion through fuzzy logic and neuro fuzzy logic techniques by fusing satellite images obtained from two or more image sensors to improve visualization capability. We also compare the proposed iterative image fusion approach based on fuzzy logic and neuro fuzzy logic with two more approaches like Principal component analysis (PCA) based fusion and the wavelet transform based image fusion. Fused images obtained from iterative image fusion using fuzzy and neuro fuzzy and other comparative methods are assessed through typical quality evaluation metrics. The proposed iterative fusion based on fuzzy logic approach divulges better values in some metrics whereas iterative image fusion based on neuro fuzzy logic technique provides better values in other metrics. Due to the potentiality of the fuzzy logic and neuro fuzzy logic. The overall experimental results generated from suggested approach substantiate that the utilization of the iterative image fusion using fuzzy logic and neuro fuzzy logic techniques would efficiently improve the quality grade of the fused image with concurrent confinement of spectral and spatial information.