Spectral Correlation and Spatial High–Low Frequency Information of Hyperspectral Image Super-Resolution Network

Hyperspectral images (HSIs) generally contain tens or even hundreds of spectral segments within a specific frequency range. Due to the limitations and cost imaging sensors, HSIs often trade spatial resolution for finer band resolution. To compensate loss maintain balance between space spectrum, existing algorithms were used obtain excellent results. However, these could not fully mine coupling relationship domain HSIs. In this study, we presented correlation high–low information hyperspectral image super-resolution network (SCSFINet) based on spectrum-guided attention analyzing already obtained from The core our was feature extraction module (SSFM), consisting two key elements: (a) fusion (SGAF) using SGSA/SGCA CFJSF extract spectral–spatial spectral–channel joint attention, (b) high- low-frequency separated multi-level (FSMFF) fusing information. final stage upsampling, proposed channel grouping (CGF) module, which can group channels merge features groups further refine provide details sub-pixel convolution. test three general datasets, compared algorithms, suggested advantage method.