Research On The Precise Stitching Algorithm Of Large Depth Of Field Images In Computational Microscopy Imaging Syste

In the field of optical microscopy imaging,due to the limitation of the numerical aperture of the objective lens,there are incompatible contradictions between the depth of field and the resolution,and between the field of view and the resolution.The method widely used to solve this kind of problem is the method of depth-of-field expansion and image stitching.They can increase the vertical depth of field and horizontal field of view while maintaining the original resolution.At the same time,most of the samples actually observed in the biomedical field are colorless and transparent weak-phase objects.For a transparent sample,it is difficult for us to directly observe its intensity information.At this time,a quantitative phase imaging method using the phase information of the sample is needed.Therefore,we need to develop an algorithm that can perform accurate and quantitative phase recovery on phase objects,as well as a depth of field expansion and image stitching method suitable for transparent phase object samples.Finally,integrate the above algorithms into a microscopy system,and combine the development trend of multimodality,digitization,and automation of the microscopy system to design a software and hardware combination that integrates multimodal imaging and image splicing and fusion The intelligent microscopy system that integrates analysis functions is of great significance.The main content of this article includes:(1)An accurate quantitative phase recovery algorithm based on spectral fusion and optimized illumination is proposed.This is a differential phase contrast phase recovery method.For complex and thick samples in phase objects,the traditional differential phasecontrast phase recovery method cannot accurately recover both the overview and details of such samples due to the limitation of approximation.Therefore,a quantitative phase recovery method that can have better effects on any sample is required.The algorithm proposed in this paper can combine the advantages of two differential phase contrast methods,break through the limitation of approximation,and achieve accurate recovery of the phase of complex thick samples.At the same time,because of its efficient and flexible algorithm,independent of hardware changes and good real-time performance,it is also very suitable to be mounted on our microscope system.(2)Research,comparation and optimization on the depth of field expansion and image stitching algorithms that are suitable for automatic microscopy systems,and the realization of the depth of field expansion and image stitching for phase objects.In view of today’s new development trend of multi-modality and automation of microscopy systems,this paper selects and optimizes a set of simple and efficient depth-of-field expansion and image stitching algorithms that fit the application scenarios of microscopy systems through research and comparison.At the same time,for phase objects,on the one hand,the precise quantitative phase recovery algorithm in this paper provides strong support for the depth of field expansion and image stitching of phase objects.On the other hand,benefiting from the idea of spectral fusion and multi-scale decomposition,this paper further optimizes the image stitching algorithm of phase objects and improves the efficiency of the algorithm.(3)An automated miniaturized microscopic analysis system integrating multi-modal imaging,depth-of-field expansion,image stitching and fusion and other functions is designed,and the software and hardware are coordinated.In particular,it has unique advantages in phase imaging,expansion of depth of field of phase images,and phase image stitching algorithms.This paper makes a comprehensive research on related algorithms and system design,and finds and proposes some areas worthy of further study and improvement.The future microscope system will have more mature and stable algorithms to improve the horizontal and vertical observable field of view and pay more and more attention to the phase imaging capability of the system.force tool.