Ghost Imaging Studies Based On Compressed Sensing

Ghost imaging is a new kind of imaging technologies developed in recent ten years. Ghost imaging attracts a lot of attentions and shows great potential in practice because of its advantages including the simplicity of system, reliability and high resolution ratio. Limited by the correlation of light fields, ghost imaging needs a large amount of measurements which cost a long time. Combining compressive sensing with ghost imaging can shorten the sampling time and acquire high-quality images.Based on the study of the existing algorithms, the iterative reweighed least square algorithm is applied in compressive ghost imaging. However, due to the complexity of the algorithm, it costs a long reconstruction time. In order to decrease the reconstruction time, the adaptive inverse scale space algorithm with high convergence is adopted. While the reconstruction time is effectively shortened, the adaptive inverse scale space algorithm cannot reconstruct image as good as that obtained with the iterative reweighed least square algorithm. Then we analyzed the possibility of using oblique projector matching to strengthen the reconstruction accuracy. Meanwhile, difference method is applied to our compressive ghost imaging experiments to improve the quality of images by decreasing background noise. We proved the effectiveness of the improvement of the algorithm through simulation experiments.