A Research Of Speckle Reduction Of Digital Hologram

With the rapid development of digital sensors and optical components,digital holography has become one of the most effective coherent imaging technologies.It has played an essential role in several research fields,such as quantitative phase imaging,optical metrology,3D display technologies,etc.Like many other coherent imaging techniques,digital holography also faces the issue of speckle noise due to the light source's high degree of coherence.In recent decades,the challenge of de-speckling has drawn many scholars and has become an important and attractive research field in digital holography.Despite the massive of the proposed de-speckling methods,many have their limitations,such as the high complexity or the low running efficiency,or signal-damaging.Therefore,removing speckle noise from the hologram's reconstructed image is the bottleneck of the development of digital holography.The paper designs two algorithms: the redundant speckle elimination method and the low dimensional reconstruction method based on the characteristics of digital holography and speckle's statistical property.The language Julia is used to implement algorithms and verify the effectiveness of the methods.The results show that the two types of de-speckling methods have different advantages on different occasions.The redundant speckle elimination has advantages in circumstances where better de-speckling performance is required,while the low dimensional reconstruction method is perfect for applications where better running efficiency is needed.Through theoretical analysis,optical experiments,computer programming,and other effective techniques,the paper focuses on the following contents:(1)The basic principles of digital holography and the types of holograms are introduced.Moreover,the recording and reconstruction principles of lensless Fourier transform digital hologram(LFTDH),are deeply analyzed.(2)The statistical property of speckles is introduced,especially the first-order statistics of the speckle.Then the statistical properties of the sum of speckle patterns are mathematically deduced and analyzed.The results show that the superposition of speckle patterns leads to an image with the speckle noise reduced.(3)To overcome the deficiency of the spatial domain mask method,an improved version named the redundant speckle elimination method,based on the property of hologram and speckle,was proposed to suppress the speckle.The experiment results demonstrate the effectiveness of the proposed method and reveal some properties of the method.(4)Aiming at the common drawback of the mask-based method,which produces diffraction patterns,a new technique called low-dimensional noise reduction is proposed in the paper.The new approach focuses on running efficiency while ensuring well de-speckling strength.The experiment results demonstrate the effectiveness of the proposed method and reveal some properties of the method.