Correspondence Normalized Correlated Imaging And Image Discrimination Based On Correlated Imaging

With respect to standard imaging, correlated imaging (ghost imaging) offers great potentiality toobtain the imaging of objects located in optically harsh or noisy environment. It can solve theproblems which are difficult using standard imaging techniques. There are some problems neededto slove in traditional correlated imaging, such as, long imaging time and low imaging quality. Thestudy on novel ghost imaging schemes with less imaging time and good performance has animportant theoretical significance and reference value.Based on the principle of correlated imaging and normalized correlated imaging scheme, acorrespondence normalized correlated imaging scheme was proposed in this thesis. The mean valueof several normalized bucket detection values in the signal beam was used as a threshold. Thecorresponding light intensity distribution in the reference beam were selected to recover theimaging when the normalized bucket detection values were greater than the threshold. Thetwo-grayscale ‘double-slit’、‘NUPT’ image and eight-grayscale ‘Lena’ image were selected as theexamples to verify the scheme. The numercial results show that Signal-to-Noise rate of theimagings were improved over20%, and the reconstruction data were decreased50%with thecorrespondence normalized correlated imaging scheme.Based on the proposed scheme, a correspondence normalized compressive correlated imagingscheme was further proposed with compressive sensing technique. In the scheme, the normalizedbucket detection values greater than the threshold in the signal beam were used as the measurementvectors. The corresponding light intensity distribution values in the reference beam were used as themeasurement matrix. Orthogonal matching pursuit algorithm was adopted as the recovery algorithm.The two-grayscale ‘double-slit’、‘NUPT’ image and eight-grayscale ‘Lena’ image were used totestify the proposed scheme. The results show that SNR of the imagings were improved over50%and the reconstruction data were reduced75%with the correspondence normalized compressivecorrelated imaging scheme.Finally, an application of correlated imaging based on entangled photon pairs was discussed innonorthogonal image discrimination. A hologram which contains many nonorthogonal images waspre-established in the reference beam. The different nonorthogonal images could be discriminatednonlocally by making a coincidence counting between the beams from different directions passingthrough the hologram and the beams in the signal beam. The numerical simulations results show that the images placed on the singal beam can be discriminated with different coincidence countingvalues when each two images in the images set had less than one half overlapping. It had enlargedthe applications in image discrimination.