Research Of Space Size Selection Method For Three-dimensional Point Spread Function

Base on the traditional fluorescence microscope,digital confocal microscopy was used to remove the biological cells and tissue sections of microscopic images from the focus of blur,in order to obtain high-quality and high-resolution sample slice microscopic images.Three-dimensional image restoration algorithm is one of the core links of digital confocal microscopy.The time and recovery effect of image restoration process are closely related to the three-dimensional point diffusion function in microscopic optical system.How to choose the appropriate space size of the three-dimensional point spread function deeply impacts the image recovery effect.In the case of accurate estimation,the three-dimensional point spread function space size and density directly determine the recovery effect and recovery time.On the basis of the existing theory,the larger the size and density of three-dimensional point spread function space,the better the recovery effect,but longer the recovery time.Therefore,in the different requirements of how to choose the appropriate size of the three-dimensional point spread function,there is a certain practical significance.This paper studies the point spread function with appropriate radial size and spacing under different restoration requirements.The main research work is as follows:1.The three-dimensional point spread function with different radial size and number of layers is used to restore the blurred image under the three magnifications,and the signal-to-noise ratio is improved as the recovery effect index.From the statistical point of view,the normal distribution map of the signal-to-noise ratio is improved,and then the interval estimation method is introduced.The concept of image restoration quality is proposed by constructing the mathematical model under different restoration requirements.The spatial size of the three-dimensional point spread function based on interval estimation is studied method.The experimental results show that this method has a good feasibility and provides another method for the selection of 3D point spread function of digital confocal microscopy system.2.The method of point spread function selection based on scale invariant feature transformation algorithm is studied.For the reconstructed images obtained at different magnifications,the feature points of each reconstructed image,original image and blurred image are found by the scale invariant feature transformation algorithm,and the reconstructed image is matched with the original image to obtain the matching point.Afterwards the number of matching points and it's increment curve is obtained,corresponding to three-dimensional point spread function layer distance and diameter.According to the curve,combined with the point diffusion function structure,based on scale invariant feature transformation algorithm,the selection of 3D point is proposed.The results of this paper provide a theoretical basis for the selection of 3D point spread function under different restoration requirements.