Application of sub-segmentation enhancement in pore detection in soil CT images

Abstract

Computed Tomography imaging is a non-invasive alternative to observe soil structures, mainly the pore space. The porous space corresponds in the image of the soil to an empty or free space in the sense that there is no material present but only fluids and the transport of these depends on the porous spaces in the soil, for this reason it is important to identify the regions that correspond to the pore areas. Due to this, this article presents a methodology based on digital image processing techniques with the objective of segmenting porous spaces in soil images. The methodology consists mainly of two stages. The first is an image contrast enhancement through a nonlinear adaptive transformation function and the second is an image segmentation through a technique known as sub-segmentation enhancement which is based on the Fuzzy Possibilist C clustering algorithm. -Medias (Possibilistic Fuzzy C-Means, PFCM). The results obtained in the segmentation stage are compared with the technique known as sub-segmentation or conventional sub-segmentation, which is also based on the PFCM hybrid algorithm. In this article it is shown that both segmentation techniques are robust, but nevertheless the area of opportunity of the classic sub-segmentation and the improvement process that results in the new sub-segmentation or improvement of the sub-segmentation are also shown. segmentation.