Research On Microscope Scene 3D Reconstruction Technology Of Microscope

The 3D reconstruction of microscope scenes is mostly used in high-precision machining in the manufacturing industry and laboratory research.Among the more common reconstruction methods are laser scanning three-dimensional reconstruction technology and binocular vision.However,the laser scanning method is time-consuming,it is easy to ignore the surface texture information of the object,and the equipment is also expensive;the amount of information obtained by the binocular vision method in the microscopic scene is much less than that of the ordinary scene,but the accuracy cannot be guaranteed.In this paper,the 3D reconstruction technology of microscopic scene of electronic video microscopy is taken as the research subject,and the 3D reconstruction method of monocular microscope is studied,so that the 3D reconstruction method can be more widely applied to microscopic scenes.The main works and results of this study are as followsFirst of all,in the deblurring method of defocusing images,the causes of the blurring of microscopic scenes are analyzed in detail.For the main reason of out-of-focus lightness,the image fusion method is selected to obtain a clear all-focus image.The related theory of wavelet analysis,how to implement multi-focus image fusion by wavelet transform and the theory of image fusion based on sparse decomposition are introduced.The fusion efficiency of the original method is improved by combining wavelet transform and sparse decomposition.The experimental results show that the effect is much better than the original sparse decomposition of image fusion theory.Secondly,in the method of defocus image depth estimation,the traditional defocus image depth estimation method is studied,among which there is a geometric constraintbased depth estimation method,a Markov random field depth estimation,and a regularization-based depth method.Secondly,in the method of defocus image depth estimation,the traditional defocus image depth estimation method is studied,among which there is a geometric constraint-based depth estimation method,a Markov random field depth estimation,and a regularization-based depth method.It is estimated that this paper focuses on the depth estimation method based on geometric constraints and improves the iterative methods.The experimental results show that the results are better than the original algorithm.The Markov Random Field Depth Estimation Method is also studied in this paper.Markov Random Field Theory and Depth Estimation Model are introduced.The algorithm is simulated through experiments.The experimental results show that no matter in the effect of depth estimation or the operational efficiency,Markov Random Airport has advantages.The surface reconstruction process based on point cloud data is studied.A filtering method based on double-pass filter and a Delaunay triangulation method based on planar projection and synthesis algorithm are mainly studied.Finally,the 3D sequence of the acquired image sequences are completed.Reconstruction experiments,the experimental results show that the reconstruction surface basically meets the surface profile information of the selected measured object.