Live Cell Imaging Based On Digital Holographic Microscop

In the field of microscopic cell research,since cells use the refractive index as an endogenous contrast agent,the light passing through will produce a change in the apparent phase.Quantitative phase imaging(QPI),which can record the phase delay caused by light passing through a sample,provides a better idea for studying high-dimensional measurements of living cells.As a kind of quantitative phase imaging technology,digital holographic microscopy imaging technology has the advantages of high speed,high resolution and low damage.It can encode the three-dimensional complex optical field of the sample into an intensity map(ie,a hologram)through the interference of scattered object light waves and reference light waves,and achieve phase reconstruction through a series of decoding algorithms.Currently,digital holography is widely used in various fields,such as quantitative phase measurement of unstained biological cells,optical metrology of nanostructures,and phase analysis of chemical processes.This paper uses the real-time,high-precision threedimensional reconstruction characteristics of digital holography to carry out quantitative research at the cellular and subcellular scales.This paper is based on the theory of digital holography,a set of transmission type threedimensional real-time monitoring system of living cells is independently built.Starting from the principle of digital holography,it mainly introduces the recording and reproduction of digital holography,and then this paper proposes a method for automatically determining the spectral center based on surface fitting of two-dimensional Hann window function,which changes the discrete spectral matrix into a continuous matrix In order to obtain a more accurate spectral center,the phase recovery accuracy and imaging quality can be improved without reducing the calculation speed.At the same time,on this basis,an automatic frame selection method of +1-order spectrum based on surface fitting is proposed,which can greatly save time while suppressing first-order aberrations,obtain better imaging quality and be more intelligent.Then,several autofocus methods in the process of hologram reconstruction are introduced: weighted spectral analysis method,gray variance method,cumulative edge detection method based on gradient calculation and cumulative edge detection method based on Laplacian filter.In addition,he also tried to introduce deep learning methods into the research of digital holography autofocus,built a deep learning network,and constructed the high-precision defocus phase and focus phase calculated based on the Fourier transform method as the input of the deep learning network.,the trained deep learning network has a good focusing effect on the constructed out-of-focus phase images.Finally,the optical system design and software design of the transmission type live cell three-dimensional real-time microscopic imaging system built in this paper are introduced,and the phase recovery ability,running speed and actual running effect of the system are tested.