A pixel-level entropy-weighted image fusion algorithm based on bidimensional ensemble empirical mode decomposition. (December 2018)

• Record Type: Journal Article
• Authors: Wang, Pei; Fu, Hui; Zhang, Ke
• Is Part Of: International journal of distributed sensor networks. Volume 14:Number 12(2018)
• Publication Date: 2018-12
• Subjects: Bidimensional ensemble empirical mode decomposition -- image fusion -- entropy; Sensor networks -- Periodicals; Intelligent agents (Computer software) -- Periodicals; Multisensor data fusion -- Periodicals; 681.2
• DOI: 10.1177/1550147718818755
• Languages: English
• ISSNs: 1550-1329

Abstract

The bidimensional empirical mode decomposition algorithm is more suitable to handle image fusion than the traditional multi-scale decomposition methods in the image fusion area. Nevertheless, there are several inherent problems of empirical mode decomposition, such as the mode mixing problem or end effects problem. As an improved empirical mode decomposition method, the ensemble empirical mode decomposition improves the empirical mode decomposition, by averaging the modes of all noise-added signals, in order to improve the mode mixing problem. In this article, an adaptive image fusion algorithm based on the representation of bidimensional ensemble empirical mode decomposition is proposed. This novel algorithm decomposes the source image by the bidimensional ensemble empirical mode decomposition algorithm, and a pixel-level weighting fusion method is then presented based on the entropy of intrinsic mode function; the fusion image can thus be obtained by inversing bidimensional ensemble empirical mode decomposition on the composite representation. Based on the quantitative comparison results, the proposed algorithm provides fusion performance to the Laplacian pyramid and wavelet transform methods. In addition, the proposed algorithm has adaptive capabilities and does not need any predetermined filters or wavelet functions.