Research On High-Resolution Imaging Method Based On Speckles

When passing through a scattering medium with inhomogeneous refractive index(such as ground glass,paint,biological tissue,etc.),the incident light will scatter away from the original direction,making the imaging system only detect the disordered speckle pattern.Therefore,the scattering medium is generally considered to be an obstacle for imaging.In recent years,researchers have proposed a series of imaging techniques with scattered light.They are collectively referred to as speckle imaging methods.Among them,methods based on transmission matrix and wavefront shaping optimization have the characteristics of intuitive principles and easy integration with other methods.The two methods have been successfully explored in the fields of microscopic imaging,multimode fiber endoscopes,and photoacoustic imaging.However,in real scenes,the insufficient number of calibrated pixels and limited efficiency of optimization algorithms make the imaging process have low-resolution or have slow speed,which makes the two methods difficult to be applied.In response to the above problems,in the field of imaging through scattering samples,this paper introduces digital image processing technologies and proposes a new optimization algorithm to achieve higher imaging resolution and faster imaging speed.The main research contents and innovations are as follows:1)To deal with the problem about low-resolution of image reconstruction by the transmission matrix method,the idea of image super-resolution reconstruction is introduced,and a high-resolution speckle imaging method is proposed.According to the operating objects,the proposed method can be divided into two different types: one is to use the low-resolution image reconstructed by the transmission matrix as the operating object,which can improve the resolution performance directly;the other is using the measured transmission matrix to build a larger one,then the imaging resolution can be improved by using that for image reconstruction.For the first type,we establish the relationship between the measured transmission matrix and the system’s scattering matrix,then define the concept of the down-sampling transmission matrix;the image degradation models base on image reconstruction by the phase conjugate factor(suitable for simple targets)and by the pseudo-inverse factor(suitable for complex targets)are proposed respectively;it is proved that all details of the original object are conserved in its speckle pattern so that the correct low-resolution image can be retrieved;base on that,the final high-resolution image can be retrieved by using the image super-resolution algorithm.For the second type,we use shifted phase masks to calibrate the transmission matrixes,and define that measured as the complementary transmission matrixes of the imaging system;construct a highdimensional transmission matrix based on the principle of multi-frame image super-resolution reconstruction;use the phase conjugate factor to reconstruct the final high-resolution images.In comparison,the first type of the method can use different reconstruction factors for different complexity targets,and can apply a variety of super-resolution algorithms,which is more flexible;the second type can achieve single-frame imaging for stationary objects and thus has better real-time performance.The experimental results prove the feasibility and effectiveness of the proposed method.2)In order to improve the resolution of image transmission based on transmission matrix,the theory of compressed sensing is introduced,and a high-resolution image transmission method is proposed.In this method,the original high-resolution target is encoded as multiframe low-resolution images,during which all information of the original target is preserved and then transformed,therefore the final imaging result has a higher degree of credibility(compared to image super-resolution based methods).The method can be divided into three main steps: encoding,image transmission,and decoding.In essence,for this method,the time resolution is sacrificed in exchange for the spatial resolution.The main feature of this method is the use of compressed sensing theory to perform the image coding process,so it is robust to noise and thus can be better suited for scattering-based imaging scenarios.When projecting the coded signal into the transferred images,a limited phase interval is used to avoid extreme reconstruction errors.Both simulation and experimental results prove the feasibility and effectiveness of this method.3)In order to improve imaging speed in the wavefront shaping optimization,the sparrow search algorithm is modified and introduced,and an imaging method based on wavefront shaping optimization is proposed.This method introduces crossover and mutation operations into the sparrow search algorithm,which solves the drawback of the sparrow search algorithm that is difficult to be applied to high-dimensional optimization problems.The proposed method has the advantages of fewer parameters,fast convergence speed,and simple calculation.That proposed uses a multimode fiber as the scattering medium,by using different feedback functions,focusing and imaging through the fiber are achieved.Compared with the existing optimization algorithms,the proposed method is more robust to environmental noise,has a higher initial growth rate,and can achieve rapid focusing and imaging through the scattering medium.Both simulation and experiments prove the feasibility and effectiveness of this method.In this paper,we thoroughly studied the existing speckle imaging methods.By introducing image processing methods(referring to image super-resolution reconstruction methods and compressed sensing theory)and new optimization algorithms,a new imaging mechanism is proposed and preliminary results are obtained.The above research can provide technical support for using existing speckle imaging systems to achieve high-resolution and high-speed imaging,and can also provide a certain theoretical basis for the design of new imaging methods.