A framelet sparse reconstruction method for pansharpening with guaranteed convergence

Pansharpening refers to the super resolution of a low-resolution multispectral (LR-MS) image in virtue an aligned panchromatic (PAN) image. Such inverse problem mainly requires proper use spatial information from auxiliary PAN In this paper, we suggest nonconvex regularization model for pansharpening via framelet sparse reconstruction, called NC-FSRM, which investigates coefficient similarity among underlying high-resolution MS (HR-MS) and images on domain, then characterizes strong statistical sparsity their error using $ \ell_0 norm. Compared with previous methods, NC-FSRM can more precisely concisely establish relation between HR-MS images. particular, piece-wise smoothness prior former simultaneously be captured without adding additional regularizers. For solving suggested model, further develop efficient proximal alternating minimization (PAM) based algorithm, is theoretically proven converge coordinatewise minimizers under some mild assumptions. Numerical experiments conducted different datasets demonstrate superiority compared other state-of-the-art methods. The source code publicly available at https://github.com/zhongchengwu/code_ncFSRM.