Image Rectification And Phase Analysis Of Four Dimensional Microstructure For Cement Hydration

Cement is the most widely used building material in the world.However,the reproducibility of cement production is poor.Even with the same cement clinker ratio,it is difficult to obtain cement with the same performance under different test conditions.In addition,due to the lack of theoretical knowledge and design aspects of high-performance cement,the output rate of high-performance cement in China is very low.The complex physical and chemical changes in the process of hydration and hardening of cement have made it difficult to precisely control the production process of cement,even the study of cement hydration process is very complicated.In order to design high-performance cementitious materials,scientific and effective methods are urgently needed to study the mechanism of cement hydration.In the study of cement science,the introduction of advanced research tools has made it possible to study the hydration of cement on a microscopic scale.Some new and advanced research tools have been introduced into the study of cement science,so that the relevant researchers' understanding of cement hydration can reach a deeper level.In particular,the introduction of microtomography can demonstrate the evolution of solid structures in the fourdimensional microscopic scale of cement hydration.It can help scientists observe changes in morphology and phase directly from the temporal sequential images.Secondly,the dynamic relationship between the microstructure and physical properties of the cement can be measured and analyzed by the image sequence.Besides,the dynamic image of the cement microstructure provides the original data for guiding thedesign of computer models.However,there are many problems with the original images obtained by microtomography that limit the study of cement hydration.There are also many difficulties in extracting natural laws using observed data.By combining the traditional experimental research methods with advanced research tools,image processing technologies,and computational intelligence technologies,to study the cement hydration mechanism through cement hydration microtomography images will provide a new approach for cement science research,which is of considerable theoretical importance and presents significant application prospect.In this paper,four dimensional microstructure images correction and phase analysis of cement hydration are carried out in the following aspects.(1)Image Registration of Downsampled Images for Cement HydrationIn order to exhibit different kinds and sizes of phases in detail for understanding its hydration mechanism,high resolution three dimensional image visualizing more than 2 Gigapixel or even 500 Gigapixel is required in practice.Therefore,the volume of resources needed for image processing,such as memory and time consumption,is enormous.This paper makes an attempt at image registration of downsampled images for cement hydration based on multi-layer particle swarm optimization mutual information,which succeeds in reducing the complexity of the problem.(2)Rapid Construction of 4D High Quality Microstructural Image for Cement HydrationStudying the microstructural evolution of cement paste during hydration is of considerable significance for understanding its mechanism and designing such material.With the use of microtomography and image registration,the four-dimensional microstructure of cement paste can be captured,thereby assisting material scientists in studying the hydration process in situ.However,the existence of a few flaws,such as large size,isotropy,and homogeneity,remarkably impede the construction of high-quality four-dimensional microstructural image for cement hydration.In this work,a novel partial phase bi-iteration registration method with multifactor multi-layer particle swarm optimization is proposed to rapidly construct fourdimensional microstructural images for cement hydration.Furthermore,a fast bias field correction method based on multiplicative intrinsic component optimization(MICO)and interpolation is also proposed for acceleration.Experimental results manifest the effectiveness of the proposed methods.(3)Three Dimensional Segmentation for Cement Microtomography ImagesThe performance of cement is strongly influenced by its microstructure,among which dynamic microstructure can reveal the formation and development of the cement paste.Therefore,the investigation of dynamic microstructure enables us to understand the cement hydration and try to improve the cement properties.However,the constituents of cement paste are hard to directly segment by human vision due to the fully mixed phases,a lot of noise and low image definition,which influences phase extraction,substance analysis and the study on the change of material composition.This paper studies the three dimensional image segmentation for cement microtomography images using self-organizing map and neighborhood features.The method takes advantage of the neighborhood features and the faulttolerance to missing,confusing,noisy data of self-organizing map.The experimental results manifest that this method perform well.Furthermore,the evolution of cement threedimensional microstructure during hydration is analyzed by the segmented images.(4)Estimating Cement Compressive Strength from Microstructural ImagesThe safety of building facilities is directly affected by the physical properties of cement,among which cement compressive strength plays the most important role in evaluating them.Therefore,the investigation of cement compressive strength is helpful in improving the cement properties.Traditionally,chemical composition,curing condition,and water-cement ratio are used to estimate the strength of cement paste.However,this approach is limited by the extreme complexity of physical changes and chemical reactions during cement hydration.Considering the cement microstructure contains information related to strength microscopically,microtomography,which can image three-dimensional microstructure,provides scientists with another way to study cement compressive strength nondestructively.This study estimates cement compressive strength using microstructure features extracted from microtomography images and gene expression programming.A probabilistic polarized similarity weight tournament selection operator is also proposed to balance the exploration and exploitation.Experimental results corroborate that the obtained relationship possesses higher estimation accuracy,good interpretability and the evolutionary capability performs well.