To compare the imaging effects between compressed sensing ghost imaging (CGI) and pseudo-reverse ghost imaging (PGI) and explore the effect of morphological weight adaptive on correlated imaging to remove noise. Different images were selected, and simulation experiments were carried out by MATLAB software. The target images were sampled with 64, 256, 512, 1 024, 2 048, and 3 000 rate respectively. The images were reconstructed by correlation imaging, compressed sensing ghost imaging and pseudo-reverse ghost imaging. The two methods to reconstruct the image effect, and then using the peak signal-to-noise ratio (PSNR) and correlation coefficient (CC) as the quantitative indicators, the compressed sensing ghost imaging and pseudo-reverse ghost imaging were used under different usage times. A comparative analysis was conducted. At the same time, in the experiment, morphological weighted adaption was used to remove noise from compressed sensing ghost imaging. pseudo-inverse ghost imaging is better than compressed sensing ghost imaging in the case of low-order sampling. Under high sampling rate, the imaging effect based on compressed sensing ghost imaging is better, but in the actual reconstruction, compressed sensing ghost imaging still has noise, and the morphological weight adaptation can effectively remove the noise in the ghost imaging experiment.