Ghost Imaging Edge Detection Scheme For Pure Phase Objects

Pure phase objects are widely used in biology,medicine and industry,but it is difficult to obtain image information because of their weak scattering.Meanwhile,the edge of the image contains a lot of important information about the object,so the edge detection of the pure phase object is of great significance in the fields of medical imaging and pattern recognition.The edge of phase information of pure phase object can be obtained under non-destructive condition by using the weak light imaging characteristic of ghost imaging.In this thesis,the ghost imaging edge detection schemes for pure phase objects are studied.The main works and the results are as follows:(1)With the principle of coherent detection in the computational ghost imaging system,and combing the first order differential operator,a ghost imaging edge detection scheme for pure phase objects is proposed.For each speckle pattern in the ghost imaging system,the new group speckle patterns are designed according to the used edge detection operator template,and then the ghost imaging procedures are performed.The edge information of the pure phase object can be obtained after the reconstruction of ghost imaging.In order to further improve the efficiency of edge detection,the compressed sensing recovery algorithm,TVAL3,is introduced in the reconstruction.The numerical simulation results show that the edge information of the pure phase object can be successfully extracted by the proposed scheme and its compressed sensing improved scheme.The edge detection results are clear and continuous.By comparing to the result using the proposed scheme,the compressed sensing improved scheme can reduce the number of the measurement and so far the number of the speckle patterns,and enhance the edge detection efficiency.(2)With the combination of the classical differential edge detection with the frequency-domain single-pixel imaging,a pure phase object edge detection scheme based on frequency-domain singlepixel imaging is proposed.Based on the sine-cosine structure speckle of the frequency-domain singlepixel imaging,new sine-cosine structure speckles are designed.With the four-step phase-shift method,the Fourier spectrum of edge information for the pure phase object can be obtained.The edge information can be finally achieved by the Fourier transform the obtained spectrum.The numerical simulation results show that the edge of pure phase objects can be obtained under both noisy and noiseless conditions.Comparing with the above proposed edge detection scheme,the scheme has a better performance.At the same time,the scheme has better edge detection results in the same detection time in comparison with the compressed sensing improved scheme.