中文摘要：

砂土具有压硬性、剪胀性、各向异性等力学性质，实质上是受其内部的细观结构的影响。利用试验和数值统计相结合的手段对砂土颗粒进行统计、细观结构进行量化分析，有利于深入研究砂土变形机理。文中基于数字图像中的二值图像处理技术，采用八邻域搜索和聚类方法实现了连通域的搜索和砂土颗粒的统计。结果表明该方法能较好地统计出染色砂土颗粒的个数和颗粒的分布情况，为进一步研究砂土内部结构性质变化与颗粒变化的内在联系奠定基础，对研究不同状态下的砂土微观结构的性质变化具有重要意义。

英文摘要：

Sand,as a kind of typical natural granular material ,its mechanical behaviors,such as compressive hardening, dilatancy,anisotropy,etc, are essentially affected by its internal microstructure .Using the method which unifies the experiments and the numberical statistics to make statistical analysis on the sandy soil pellet and quantitative analysis of meso structure is helpful to the thorough research of sandy soil distortion mechanism .This article based on the digital binary image processing technology , uses eight neighborhood searches and the cluster method to realize the connected domain search and the sandy soil pellet statistics .Results show that it is a good way to count the number of colored sand particles and understand the distribution of particles .It is of great signifi-cance to study the change of properties of .sandy soil microscopic structure under different conditions .