Image Fusion Using Compressive Sensing in Fourier Coefficient Domain

This literature presents an image fusion method based on compressive sensing, implemented specifically in the Fourier coefficient domain. The approach significantly enhances fusion quality while offering distinct advantages for processing large-scale images. By operating within the compressive sensing framework, the method achieves reduced computational time and storage requirements through sparse signal representation and random sampling techniques. The implementation typically involves: Fourier domain transformation for frequency coefficient extraction, compressive sampling using random measurement matrices, and fusion rules applied to sparse coefficients before reconstruction. This methodology demonstrates extensive practical application potential in fields requiring efficient multi-source image integration, making it a promising image fusion technology worthy of further research and implementation. Key algorithmic steps include discrete Fourier transform (DFT) matrix operations, l1-norm optimization for sparse recovery, and inverse transform techniques for image reconstruction.