Abstract:On the basis of analyzing the cosine light field with determined analytic expression and the pseudo-inverse method, the object is illuminated by a presetting light field with a determined discrete Fourier transform measurement matrix, and the object image is reconstructed by the pseudo-inverse method. The analytic expression of the algorithm of computational ghost imaging based on discrete Fourier transform measurement matrix is deduced theoretically, and compared with the algorithm of compressive computational ghost imaging based on random measurement matrix. The reconstruction process and the reconstruction error are analyzed. On this basis, the simulation is done to verify the theoretical analysis. When the sampling measurement number is similar to the number of object pixel, the rank of discrete Fourier transform matrix is the same as the one of the random measurement matrix, the PSNR of the reconstruction image of FGI algorithm and PGI algorithm are similar, the reconstruction error of the traditional CGI algorithm is lower than that of reconstruction image based on FGI algorithm and PGI algorithm. As the decreasing of the number of sampling measurement, the PSNR of reconstruction image based on FGI algorithm decreases slowly, and the PSNR of reconstruction image based on PGI algorithm and CGI algorithm decreases sharply. The reconstruction time of FGI algorithm is lower than that of other algorithms and is not affected by the number of sampling measurement. The FGI algorithm can effectively filter out the random white noise through a low-pass filter and realize the reconstruction denoising which has a higher denoising capability than that of the CGI algorithm. The FGI algorithm can improve the reconstruction accuracy and the reconstruction speed of computational ghost imaging.

Computational ghost imaging based on array sampling

Ghost imaging based on array sampling

Fast high quality computational ghost imaging based on saliency variable sampling detection

Computational Ghost Imaging Via Adaptive Deep Dictionary Learning.

High-quality computational ghost imaging with multi-scale light fields optimization

Compressive Sensing Ghost Imaging Based On Neighbor Similarity

Low sampling high quality image reconstruction and segmentation based on array network ghost imaging

Real-time physical compression computational ghost imaging based on array spatial light field modulation and deep learning

Computational ghost imaging with deep compressed sensing

A simple and robust single-pixel computational ghost imaging

Mutual Structure Ghost Imaging under Low Sampling

N ov 2 01 8 Multi-resolution Progressive Computational Ghost Imaging

Deep learning based high-resolution incoherent x-ray imaging with a single-pixel detector

Study on the Algorithm of Computational Ghost Imaging Based on Discrete Fourier Transform Measurement Matrix

Multi-resolution Progressive Computational Ghost Imaging

Image quality enhancement in low-light-level ghost imaging using modified compressive sensing method

Demonstration of asynchronous computational ghost imaging through strong scattering media

High Speed Ghost Imaging Based on a Heuristic Algorithm and Deep Learning*

Color computational ghost imaging by deep learning based on simulation data training

High-resolution Sub-Sampling Incoherent X-Ray Imaging with a Single-Pixel Detector

Deep-learning-based Ghost Imaging.