Research On Object Reconstitution Algorithm Of Ghost Imaging

Correlated Imaging, also known as Ghost Imaging, is a newly emerging optical imaging technique in recent years. Unlike the traditional imaging system, correlated imaging uses the high order correlation characteristics of the light field in the imaging process instead of using the reflection or transmission information of the object directly. Correlated imaging has aroused broad attentions with its advantages including separation between detecting and imaging and the ability to image under complicated and difficult circumstances. There are two major researching directions of correlated imaging, enhancing the Signal Noise Ratio(SNR) and investigating new applications. In this paper, two methods which enhance the SNR and an application based on correlated imaging are proposed.Firstly, this paper introduces the history of GI development, the current domestic and overseas reaching situation and a few applications. Meanwhile, the existing problems are analyzed including limited SNR of the reconstruction, large number of measurement times, noise interference on the object arm, etc., and a few frequently-used methods are introduced. Subsequently, the Compressive Sensing(CS) theory which can precisely reconstruct the signal with a small number of measurements is introduced, and a normalized correlated imaging method based on CS, the Compressive Sensing Normalized Ghost Imaging(CSNGI), is proposed. The numerical experiment results show that for grayscale images with rich details the Peak Signal to Noise Ratio(PSNR) of the CSNGI method with equal time of measurements is 6dB higher than the GI method and 2dB higher than the normalized correlated imaging method, and for binary images with less details, the PSNR is 5.2~6.5dB higher than the GI method and 3.4~4.3d B higher than the normalized correlated imaging method. To further improve the PSNR of the reconstruction, we introduces the mathematic pseudo-inverse of matrixes and proposes the Pseudo-inverse Ghost Imaging(PGI). The advantages of PGI compared to GI are computed and proved theoretically, and the simulation results show that PGI can reconstruct images with higher PSNR than GI, DGI and CSGI with only a little longer computation time. Besides, the influence of the error between the two arms and noise on the object arm are also discussed. At last, this paper introduces a Three Dimensional(3D) correlated imaging scheme. Based on the photoacoustic effect, starting with the photoacoustic wave equation and traditional GI algorithm, the formula of the 3D correlated imaging algorithm based on photoacoustic is derived. The optical absorption coefficients of the object are reconstructed on different spherical surfaces circling the transducer, and then assembled as the 3D image of the object.