Multi image fusion based on compressive sensing

Compressive sensing provides a novel framework to acquire and to reconstruct a signal or digital image from sparse measurements acquired at sub-Nyquist/Shannon sampling rate. In this paper, we present an effective image fusion scheme based on a Discrete Cosine Transform (DCT) sampling model for compressive sensing imaging. A sparse sampling model according to the DCT-based spectral energy distribution is proposed. The compressive measurements of multiple input images obtained with the proposed sampling model are fused to a composite measurement by combining their wavelet approximation coefficients and their detail coefficients separately. The combination is done by applying a weighting operation for every sampling location according to the statistical distribution. Furthermore, the fused image is reconstructed from the composite measurement by solving a problem of total variation minimization. Finally, we validate the effectiveness of the algorithm using multiple images.