基于多阈值几何分割算法的膨胀土裂隙特征分析

Analysis of crack characteristics of expansive soil based on multi-threshold segmentation algorithm

为了解膨胀土微细裂隙特征,根据CT扫描技术原理,建立一种多阈值几何分割算法,对膨胀土土体组成成分进行划分,重建膨胀土微观裂隙的三维结构。结果表明,多阈值几何分割算法能够准确分割出CT图像中的不同成分。完整的膨胀土可分为裂隙、潜在裂隙、土颗粒、姜石或铁锰结核4种成分。膨胀土中的裂隙主要为管状和片状分布,裂隙分布呈各向异性。潜在裂隙的灰度值或密度介于裂隙与土颗粒之间。潜在裂隙呈不规则网状分布,几乎遍布整个土壤空间,体积含量较高。

In order to understand the characteristics of expansive soil microstructure,according to the physical principle of CT scanning technology,a multi threshold geometric segmentation algorithm was established,and it was used to divide the soil composition and reconstruct the three-dimensional microstructure of the cracks in expansive soil. It could be seen from the test results that the multi-threshold segmentation algorithm could clearly identify the different components in CT images. The intact expansive soil could be divided into four components:Cracks,potential cracks,soil matrix,and loess-doll or ferromanganese nodules.The cracks in the expansive soil were mainly distributed in tubulars and sheets and its distribution had obvious non-uniformity. The gray value or density of the potential cracks was between the cracks and the soil matrix. The potential cracks were distributed in an irregular network and almost taking up the entire soil space,with a high volume content.