| Correlated imaging is a new imaging method which relies on the detection correlation of spatial light field and single-pixel calculation.Because of its novel imaging method,high resolution and good anti-noise performance,it has attracted the attention of researchers in recent years.How to make the correlated imaging technology practical is one of the hot topics in today's scientific research.However,the main factors affecting its practical application are the low imaging quality and imaging speed,which are also the key problems to be solved.Because it uses a single-pixel detector to measure the target scene,once these problems are solved,the correlated imaging technology using second-order correlation calculation will bring a positive impact on human production and life.This paper,a series of theoretical and project studies are carried out to optimize the modulation and reconstruction methods of correlated imaging light field,the relationship between spatial light field and imaging quality is analyzed in detail from different angles,and the corresponding light field optimization scheme is given.Then,from the perspective of high-quality reconstruction method,the imaging method of joint iterative correlated imaging is proposed.In addition,the method of feature extraction using correlation imaging is also discussed in order to carry out the research work related to the practicality of correlated imaging.The main contents are as follows:(1)The light field modulation matrix is optimized.A simple and fast continuous multiresolution correlation imaging method based on Hadamard optimization matrix is proposed,and the factors for achieving multi-resolution correlated imaging are theoretically analyzed.The imaging resolution of this method is proportional to the number of measurements,and the higher the number of measurements,the higher the resolution is.This method can obtain all low-resolution images at the same time when the number of measurements of a highresolution image obtained by traditional imaging methods is equal.This can significantly reduce the reconstruction time and measurement required for multi-resolution images.This approach improves the flexibility of ghost imaging,and can be extended to multi-resolution image-dependent practical applications,such as target tracking and recognition.(2)In view of the practical application requirements of fast multi-resolution imaging,we propose a multi-resolution correlated imaging method based on Hadamard ‘pipeline'coding,which can directly generate two-dimensional Hadamard basis patterns and multiresolution Hadamard optimization sequences,whereby both the memory consumption and the complexity of coding implementation for multi-resolution imaging can be significantly reduced.The commonly used optimization method of Hadamard optimization sequence implementation and time consumption are also discussed.This method provides a new approach for Hadamard sequence optimization and multi-resolution correlated imaging applications.(3)The reconstruction method of high quality correlation imaging with low sampling rate is studied.A high quality joint iteration correlated imaging method based on Landweber regularization and guided filtering is proposed.In this method,the target object is reconstructed by decomposing joint iteration of regularization and filtering,rather than solving the minimization problem during reconstruction.Numerical simulation and experimental results show that this method can achieve better reconstruction effect for binary image and gray image with the same measurement times.Although our method only uses projective Landweber regularization and guided filtering,this method can be extended to other regularization or filtering iterations and is likely to get better results.Furthermore,based on the assumption that the vector superposition matrix of non-locally similar blocks has low rank and sparse singular values,this paper demonstrates theoretically and experimentally a low-rank constrained joint iterative correlation imaging method using projection Landweber regularization and block matching under low sampling rate.Numerical simulation and experimental results show that this method can obtain better imaging quality in terms of peak signal-to-noise ratio,structural similarity and visual observation under low sampling rate.(4)For the practical application of correlated imaging,such as the demand and application of image processing and edge detection in security detection and medical diagnosis,and most edge detection based on correlated imaging systems require special coding or a large amount of measurement time,and cannot directly provide target images.An edge detection method for correlated imaging based on projection Landweber iterative regularization and guided filtering joint iteration is proposed.This method can achieve high quality imaging while obtaining high quality edge features.The numerical simulation and experimental results show that the spatial information and edge information of the target can be successfully recovered from the random speckle pattern without special coding at a relatively low number of measurements,and the edge image quality can be improved significantly.This method improves the applicability of association imaging and meets the practical application fields of simultaneous imaging and edge detection. |
PDF Full Text Request