Experimental Investigation of Iterative Pseudoinverse Ghost Imaging

Abstract

An iterative pseudoinverse ghost imaging (IPGI) method is proposed based on iterative denoising and pseudoinverse ghost imaging (PGI). The background noise in the imaging is eliminated in iterations by setting an appropriate threshold. The IPGI method provides a significantly larger enhancement of the peak signal-to-noise ratio (PSNR) than the PGI technique for binary objects. Experiments and data analyses are performed to evaluate the performance of the proposed method. Compared with conventional GI, differential GI, and PGI methods, the proposed method has the highest performance in visual effects and significantly improves the imaging quality. For a certain PSNR, the proposed method provides satisfactory performance in terms of computing time.

Experimental Investigation of Anisotropic Diffusion Applied in Ghost Imaging Reconstruction

“…The object beam is modulated by an unknown object and measured by a bucket detector with no spatial resolution; the reference beam is collected by a detector with spatial resolution. It is an indirect imaging technology in which the two detectors are not directly used for object reconstruction, but the objects can be achieved by the associative calculation of two groups of data the detectors collected [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15]. Therefore, ghost imaging technology can be applied in a variety of complex environments, such as imaging targets in a harsh surrounding, where traditional imaging technology may fail [16][17][18].…”

“…The appearance of ghost imaging with a classical light source has greatly enlarged the application prospect of ghost imaging. Ever since, researchers began focusing on how to improve the imaging quality and speed up the process of application exploration [6][7][8][9][10][11][12][13][14][15][26][27][28]. In the aspect of ghost imaging algorithms, differential ghost imaging (DGI) [6], as well as normalized ghost imaging (NGI) [7], turn attention to modifying the GI algorithm by means of weighting the speckle patterns for each measurement.…”

“…Pseudo-inverse ghost imaging (PGI) replaces the transpose of the matrix constructed by all of the intensity distribution measured by the detector with its pseudo-inverse, thereby providing a new direction of investigation for further study [8,9]. Iterative pseudo-inverse ghost imaging (IPGI) uses the reconstruction results of PGI to construct the noise term, and then obtains better reconstruction results over several iterations [10,11]. In binomial-theorem ghost imaging (BGI) [12], images with low-level noise can be generated by constructing a binomial formula using high-order imaging results that are acquired by reintroducing the reconstruction result back into the imaging formula repeatedly.…”

“…The IPGI algorithm takes the reconstruction result of PGI as the initial image to construct the interference noise term. Theoretically, image quality will keep getting better with successive iterations, so continuous iteration can effectively suppress the noise interference [8][9][10][11]. After several iterations, the noise will be greatly suppressed.…”

“…where the T (0) IPGI = T PGI and the threshold t varies from min {n (x, y)} to max {n (x, y)}. A normalized threshold t = t−min{n(x,y)} max{n(x,y)}−min{n(x,y)} is utilized to simplify the experimental analysis [11,13]. IPGI requires the initial image to be as clear as possible, otherwise the constructed noise interference terms may deviate from the actual noise and more noise will be generated after iterative operations.…”