Research On Off-axis Digital Holographic Microscopy And Phase Recovery Algorithm

Optical microscopy uses optical principles to amplify objects that cannot be distinguished by the naked eye.Because of its non-destructive,non-contact,and real-time observation,it has become a common analysis method in the field of life sciences.For transparent cells,due to the low contrast,ordinary optical microscopes cannot obtain high contrast,and phase imaging technology shows great advantages.As a phase imaging technology,digital holography(DH)has the unique advantage of quantitative imaging and measurement of the intensity and phase of the object.Similar to the traditional optical imaging method,digital holography is also a microscopy that can only obtain the amplitude of the object.However,digital holography uses the principle of coherent light interference to record the phase information of the light wave in the form of light intensity.The phase information recovery technology can obtain the phase of the light wave and then reconstruct the three-dimensional light field,generate a high-contrast image,and reveal the structure and morphological information of the sample.Digital holography provides a new quantitative method for the imaging of transparent samples such as living cells,which is of great significance for the study of microscopic fields.This paper mainly focuses on the basic theory and key technology research of off-axis digital holographic imaging optical path design,improving phase reconstruction quality,phase distortion compensation,homogeneous sphere sample refractive index and thickness calculation.The main research contents are as follows :1)The theoretical basis of digital holography is introduced,including digital reconstruction algorithm,digital imaging basis,phase retrieval algorithm,phase unwrapping algorithm and the relationship between object thickness,refractive index and phase.Based on Michelson interferometer,an off-axis digital holographic microscopic imaging system is built.The system simplifies the experimental optical path and enhances the anti-interference ability of the optical path.The stability of the optical path is tested by imaging the white blood cell and the USAF1951 resolution board,and the optical path adjustment is completed according to the quality of the reconstructed phase map,which verifies the imaging ability of the system designed in this paper.2)The phase information of pre-amplified off-axis digital holography is obtained by Fourier transform method.However,the introduction of the objective lens and the alignment error of the experimental system components cause the wavefront to have secondary phase distortion and high-order aberrations.To solve this problem,the same compensation method based on RBF neural network is proposed.The algorithm estimates the actual phase of the object by sampling the phase distribution of the expansion,using the RBF network to construct a nonlinear function to minimize the loss function.In the simulation,the global mean square error of the three methods is calculated based on the original model.The result of RBF is 0.0374,PCA is 0.0470,and SCM is 0.3303.The simulation results prove the effectiveness and superiority of the algorithm in off-axis digital holographic aberration elimination.In addition,the effective compensation of background interference and phase distortion is verified by experiments,which is more accurate than the traditional method.3)In digital holography,the phase obtained is the integral of refractive index and thickness.To study the refractive index or thickness separately,another parameter must be known.For spherical samples,the radius is usually determined by using geometric methods,and the scattering of light waves by the sample is thus ignored.In this paper,the scalar scattering theory is introduced,and the Rytov approximation model is established to replace the complete spherical object scattering field.The phase reconstruction of the simulated spherical object under the incident field and the scattering field is carried out.The simulated phase diagram generated by the Rytov approximation model is used to fit the phase diagram of the spherical object.In the fitting process,the genetic algorithm is used to find the radius and refractive index that can minimize the phase difference.The results show that for the simulated spherical structure with a radius of 5μm,when the refractive index difference between the medium and the spherical object is 0.01,the accuracy of the refractive index obtained by this method is 191% higher than that obtained by the geometric method.In the experiment,the mean squared error between the phase map obtained by this method and the quantitative phase map obtained by DHM is only 1/5 of the geometric method.For objects with small phase shift,the refractive index obtained by this method has higher accuracy than the geometric method.4)In this paper,a microsphere-assisted microscale dispersion is proposed.Under oblique white light illumination,the dielectric microspheres on the nanogratings can generate a rainbow focus,which helps the long-wavelength light of the rainbow focus to focus farther away from the microspheres.