The Porosity Fabric Study Of Sand Based On Image Processing Technology

Sand is typical discrete particle material in meso level. The particle shape, particle spatial arrangement and porosity between particles has a substantial influence on its macroscopic mechanical properties. Irregular distribution of porosity space show the obvious anisotropy, the reasonable description of this features is important to research on sand strength and deformation mechanism. However, on the theoretical and experimental results in the sand microstructure porosity is relatively seldom, and it serious impact on the study of the mesoscopic mechanism. Therefore, based on the theory research of porosity tensor with the principle of stereology and image analysis technology, the theory description, the mesoscopic test and test image of sand porosity is studied in this paper.Based on a description theory of the sand microstructure porosity tensor, a novel determination method of porosity tensor is puts forward. This method use the ratio of porosity component and the mean porosity to define porosity tensor. The trace of tensor is constant equal to one, which has greatly helped to determine porosity tensor. Porosity tensor is described by two independent variation angles. The defined amplitude parameters of porosity on three orthogonal surface by two independent variation angles and space test line, because of the tensor trace is equal to one, there are only two independent variation in the three parameters, so with any two amplitude parameters can describe the porosity orthogonal anisotropy. Along with the variation of porosity distribution, the orthogonal anisotropic natural degradation for transverse isotropic, at this time only an amplitude parameter can describe the transverse isotropic of porosity. The new method overcome many deficiencies of the traditional method, and it has a wider application range and simple method for determination.The meso-experiment of Aeolian sand from Tengger Deser is carried out. The original sand solidified by epoxy resin and curing agent. First the sample in three orthogonal planes was cut. Then, it is to ensure that the microscopic image of sand can be clearly observed by perform the polishing, spraying gold particles on the sample surfaces and other test process. Finally, the meso images of samples on three orthogonal planes is collected with SEM (scanning electron microscope). The collected SEM image has been processed with image processing technology, the process includes median filter, binary image, morphological processing and image segmentation of particles, the image becomes binary image with clear boundary and particle separation, it can be used for the subsequent image analysis.We write Matlab image analysis program for porosity tensor calculated by traditional and proposed method in this paper. The test lines in ring are drew, with which obtain porosity distribution data at different angles on sand sample. The portions of the test line occupied by porosity can be easily obtained by applying a logical ’and’ operator to the test line and the binary image. Using the above method, based on the already had image analysis on sand micro study results, we verify the correctness of the program. The comparison with the test data between traditional method and improved method show the determination method of porosity fabric has a good description effect. Through the image analysis of the Tengger Desert undisturbed sand, we proved the aeolian sand has obvious anisotropy of porosity.