Singular value decomposition-based anisotropic diffusion for fusion of infrared and visible images. (3rd April 2019)

• Record Type: Journal Article
• Authors: Yin, Shoulin; Zhang, Ye
• Is Part Of: International journal of image and data fusion. Volume 10:Number 2(2019)
• Page Start: 146
• Page End: 163
• Publication Date: 2019-04-03
• Subjects: image fusion -- anisotropic diffusion -- singular-value decomposition -- principal-component analysis -- Fourier transform; Image processing -- Periodicals; Multisensor data fusion -- Periodicals; Multisensor data fusion; Periodicals; 621.36705
• DOI: 10.1080/19479832.2018.1487886
• Languages: English
• ISSNs: 1947-9832

Abstract

ABSTRACT: Image fusion is an important part of image processing because it can collaboratively use image information from different sensors for the same scene. Image fusion outputs the images, which are more suitable for human visual perception or computer processing and analysis; furthermore, it can obviously improve the shortcomings of a single sensor, as well as improve the image definition and packet content, making it conducive to accurately, reliably and comprehensively obtaining information regarding the target or scenarios. To improve the fusion precision of infrared and visible images, a new image fusion method for infrared and visible images is proposed in this paper. The proposed method is partitioned into three parts. First, anisotropic diffusion is used to decompose the source images into base and detail layers. Additionally, singular-value decomposition and principal-component analysis are adopted to calculate the final base and detail layers, respectively. Then, final detail and base layers are fused using the Fourier transform method. Finally, experiments with various infrared and visible images are conducted to evaluate the proposed method. Experimental and contrastive results show that the new algorithm has a good subjective visual effect; in addition, the objective evaluation index is improved significantly in terms of definition, mutual information and entropy.