MGFuse: An Infrared and Visible Image Fusion Algorithm Based on Multiscale Decomposition Optimization and Gradient-Weighted Local Energy

Existing image fusion algorithms have difficulty in effectively preserving valuable target features infrared and visible images, which easily introduces blurry edges unremarkable notable targets during their process. We propose the MGFuse algorithm as a solution to this problem, is novel that utilizes multiscale decomposition optimization gradient-weighted local energy. Initially, non-subsampled shearlet transform (NSST) applied partition both images into several high-frequencies low-frequencies components. Subsequently, acquired low frequencies continue be decomposed via proposed function get base layers texture layers, can optimize quality of preserve fine-grained details, respectively. In addition, we formulated an intrinsic attribute-based energy (IAE) scheme merge two layers. The are extracted by (GE) operator based on structure tensor, employed construct strategy for these parts. At last, parts linearly combined integrated low-frequency layer final using inverse NSST. Numerous experimental observations demonstrate our achieve superior capability than reference nine advanced qualitative quantitative assessment, robustness noisy with different noise levels.