研究目的
To propose a novel image fusion framework based on hybrid image decomposition and sparse representation for the fusion of infrared and visible images, aiming to highlight infrared targets while preserving texture details of visible images.
研究成果
The proposed method effectively highlights infrared targets and preserves texture details of visible images, making the fused image more consistent with human visual perception. Experimental results demonstrate the superiority of the proposed method over current popular image fusion methods.
研究不足
The paper does not explicitly mention the limitations of the proposed method.
1：Experimental Design and Method Selection:
The study employs a hybrid image decomposition method using Gaussian and guided filters to separate texture details, edge, and coarse-scale image information. Sparse representation-based fusion is then applied to the decomposed sub-information.
2：Sample Selection and Data Sources:
Infrared and visible images of the same scene are used as source images for fusion.
3：List of Experimental Equipment and Materials:
Not explicitly mentioned.
4：Experimental Procedures and Operational Workflow:
The source images are decomposed into small-scale texture details, large-scale edge, and coarse-scale image information using Gaussian and guided filters. The fusion of coarse-scale information is determined by significant infrared features in the large-scale edge information. Texture details and edge information are fused using sparse representation-based methods.
5：Data Analysis Methods:
The fused images are evaluated subjectively and objectively using metrics such as QMI, QM, QS, and QCB to assess fusion performance.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
