3D Reconstruction Of Polarization Microscopy Image

The three-dimensional reconstruction of microscopic images can model the actual samples corresponding to the real world in three-dimensional,which is helpful for the analysis and utilization of microscopic results.The optical polarization microscopy imaging method can effectively utilize the interaction between polarized light and matter to obtain more abundant sample characterization information.Attempts to three-dimensional reconstruct the polarization microscope images will help to further tap the application potential of polarization microscope images.In this paper,a non-intuitive polarization parameter microscopy imaging system is built,and the vertical displacement of the lens is precisely controlled by piezoelectric ceramic sensors,and tomographic images of gold nanoparticles under different polarization degrees are obtained.Through non-intuitive inversion,microscopic tomographic image of polarization parameters of gold nanoparticles was obtained.Microscopic image of tomographic polarization parameters of nanoparticles.By comparing the comprehensive effects of median filtering and mean filtering algorithms,it was decided to use Laplacian algorithm and median filtering algorithm to preprocess tomographic images,which laid the foundation for the segmentation and registration of polarization microscopy images.Aiming at the multi-objective characteristics of the gold nanoparticle structure,an improved Level Set algorithm is proposed,and the center expansion algorithm is combined with the improved Level Set algorithm to segment the tomographic polarization microscopy images of gold nanoparticles.The experimental results show that the segmentation effect of the center expansion algorithm and the improved Level Set algorithm combined use,which is better than the segmentation effect of the center expansion algorithm and the original Level Set algorithm combined use.Aiming at the relative displacement between different tomographic images caused by mechanical vibration during the acquisition of polarization parameter microscopy images,the SIFT image registration algorithm is used to correct the displacement.The basic principles of the Marching Cubes algorithm for surface rendering are studied.Aiming at the problem of isosurface ambiguity in the 3D reconstruction process,an improved Marching Cubes algorithm based on the asymptotic discriminant method is proposed,and respectively use the original and improved Marching Cubes algorithm to reconstruct the three-dimensionally segemented and registered gold nanoparticle microscopic polarization image.The comparative analysis results of the three-dimensional reconstruction of the two algorithms show that the improved MC algorithm can effectively eliminate the ambiguity when constructing the isosurface,and effectively eliminate the void hole phenomenon of the threedimensional reconstruction surface.