The quantum imaging (ghost) has attracted many attention in recent years. In the ghost imaging scheme, the object
information is captured by the “bucket” detector which has no spatial information. In practical scheme, the CCD array or
the APD are utilized as the “bucket” detector, but the saturation effect of the detector may exist due to the limited
sampling depth of the detector. The two methods are presented to reduce the effect of the saturation for different “bucket”
detectors. The method for CCD array is based on the statistic principle of ghost imaging, and the other is based on the
compressed sensing. After that, we compare the difference between the ghost imaging and compressed sensing in the low
light level condition.
Traditional assessment methods don’t fit for photon-counting image (PCI) system because the output of PCI system is
the number of photon instead of the gray value. Normalized mutual information is presented to assess Photon-counting image system performance and calculated based on improved 2D entropy for PCI system to provide the fidelity of PCI system. Probabilities of target classification error are introduced by Bayesian posterior probability classification method under the different fidelity conditions and the result confirms fidelity values. This method applied in the PCI experiments can reflect image fidelity and system performance well in different experiment conditions and supplies the evidences for assessing the quality for photon-counting image.