Year: 2024
Type: article
Abstract: To address the issue of unremarkable targets and the serious loss of background details that occurs in multispectral image fusion, a multispectral image fusion method based on information decoupling is proposed. In accordance with the U‐shaped network structure, multi‐scale feature fusion is employed to obtain comprehensive contextual features. Subsequently, taking into account the differences between base and texture information in the images, the recovered image is reconstructed by fusing feature information through the information decoupling branches. Finally, the information decoupling is supervised by a loss function. In addition, a series of comparative experiments are conducted on a self‐constructed multispectral dataset and a publicly available dataset. Experimental results demonstrate that our method not only achieves higher fusion accuracy but also has superior visualization results.
Source: IET Image Processing
DOI: 10.1049/ipr2.13296
Authors: Wenqi Yu, Huilin Wang, Jilong Liu, Guan Wang
Language: English
Subfield: Media Technology
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $2,610
Award: G3501960997
OpenAlex ID: W4405644546