Block-compressed-sensing-based reconstruction algorithm for ghost imaging

A block compressed sensing ghost imaging reconstruction algorithm based on a block sparse Bayesian is proposed to deal with the problem that the large-size target image limited by the memory space cannot be reconstructed with the existing compressed sensing ghost imaging scheme. The large-size target image is divided into several small-sized image blocks of the same size, which is separately subjected to a compressed sensing reconstruction solution. The reconstructed solutions of each image block can be combined into one whole image, resulting in the target image. The simulation results show that the block compressed sensing ghost imaging reconstruction algorithm with a block sparse Bayesian can perform well even when storage requirements are difficult to realize via reconstructing large-scale target images.