| Traditional imaging methods that can only capture intensity information of the objects no longer meet the demands of further development in many fields,such as biomedicine and micro-nano fabrication.However,the phase information which cannot be collected by traditional imaging methods directly always contains key characteristics of the object,such as three-dimensional profile and refractive index.Therefore,how to retrieve the phase information of the object accurately is a significant research goal.Coherent diffraction imaging(CDI)techniques use phase retrieval algorithms to reconstruct the amplitude and phase information of an object simultaneously from diffraction intensities.The simple system without reference light and complex equipment makes this technique robust and stable.Due to the unique advantages mentioned above,CDI has been successfully applied in biomedicine,crystallography and materials science.However,traditional CDI methods which rely on multi-distance capturing or lateral scanning,have the disadvantages of a large number of intensity acquisitions,slow convergence speed,and low-frequency information losing.To solve such proplems,random scattered light field was introduced into CDI,which can break the symmetry of the optical path.Such method can achieve rapid reconstruction of arbitrary wavefronts by using a small number of recording intensities.Therefore,this article focuses on the CDI technology based on random scattered light field.In this article,we propose two reconstruction algorithms to mitigate the influence of the noise caused by the introduction of random scattered light fields.In addition,a one-shot phase retrieval method based on the random scattered light field is proposed for dynamic target imaging with high convergence speed.The main research works include:1.Firstly,the research background and significance of phase retrieval are introduced.In this part,we have systematically investigated the CDI technique and the corresponding reconstruction algorithms and analyzed the applications of random scattered light fields in imaging systems.We derived the basic process of free space diffraction based on the wave equation and introduced some typical phase retrieval algorithms.2.Secondly,an imaging system for the CDI method based on random scattered light field was built,and some key technical indicators are analyzed,such as sampling rate and resolution.A registration method which aligns the image sensor and the scattered light field device was proposed,which is based on computational holography.To obtain accurate distance parameters,we using the auto-focus method.All of this section is to prepare for the subsequent implementation of the phase retrieval algorithm3.And then,we proposed a parallel reconstruction algorithm for the random scattered light fields CDI,which can enhance the stability of the system while suppressing noise.So the imaging quality is improved effectively.The experiments showed that the algorithm proposed in this paper improves the quality of the reconstructed image by 28.4% from the traditional algorithm with 4 intensities,and 61.0% with 8 intensities.In order to further improve the convergence speed of the parallel algorithm,a series-parallel cascade algorithm is proposed subsequently,which can improve the convergence speed while suppressing noise.Both simulations and experiments were conducted,whose results verified the high performance of the two algorithms.4.Finally,a one-shot phase retrieval method based on random scattered light fields was proposed for dynamic imaging applications.The method uses a random scattered light field and digital in-line holography to obtain an initial reconstruction result.The support of the object plane is obtained by digital image processing from the initial result.Compared with the traditional one-shot phase retrieve method,the proposed method has prominent advantages such as high flexibility,high imaging quality,and fast reconstruction speed. |
PDF Full Text Request