Multiscale Self-coordination of Bidimensional Empirical Mode Decomposition in Image Fusion
暂无分享,去创建一个
The bidimensional empirical mode decomposition (BEMD) algorithm with high adaptability is more suitable to process multiple image fusion than traditional image fusion. However, the advantages of this algorithm are limited by the end effects problem, multiscale integration problem and number difference of intrinsic mode functions in multiple images decomposition. This study proposes the multiscale self-coordination BEMD algorithm to solve this problem. This algorithm outside extending the feather information with the support vector machine which has a high degree of generalization, then it also overcomes the BEMD end effects problem with conventional mirror extension methods of data processing, . The coordination of the extreme value point of the source image helps solve the problem of multiscale information fusion. Results show that the proposed method is better than the wavelet and NSCT method in retaining the characteristics of the source image information and the details of the mutation information inherited from the source image and in significantly improving the signal-to-noise ratio.
[1] Wenting Huang,et al. Learning Similarity with Probabilistic Latent Semantic Analysis for Image Retrieval , 2015, KSII Trans. Internet Inf. Syst..