A comparative analysis of image fusion methods using texture

Abstract

Image fusion produces a single image from a set of input images such that the fused image have more complete information useful for human or machine perception. In the proposed paper, authors have used feature based image fusion, where textures of the image are used as feature. Image Texture is a process that can be applied to the pixel of an image in order to generate a measure (feature) related to the texture pattern, to which that pixel and its neighbors belong. Authors have used five different texture feature extraction methods for fusion of multi sensor, multi focal, multi temporal and multi spectral imagery. The methods are: GLCM, Runlength, Statistical, Tamura and Texture Spectrum. The performance of fusion algorithm is measured using a number of nonreference quality assessments metric. A meaningful comparison of results and analysis show the suitability of various texture features for fusion of images from multiple modalities. © 2013 Springer.