中文摘要：

为了解决利用黄土SEM图像提取孔隙参数的问题,将孔隙区域的提取问题转化为孔隙像素点的分类问题,通过把整个图像的像素点分为孔隙点和骨架点,从而实现孔隙的提取。利用这种方法,对潼关Q3黄土震陷试验前后的微结构图像进行处理,得到了动力作用前后的孔隙分布图,并计算了孔隙面积、面孔隙度、孔隙分形维数等参数,结果表明：震陷试验后大孔隙数量减小,中、小孔隙数量有所增加,震陷主要是由大孔隙破坏所致。

英文摘要：

By applying the SVM theory in a quantitative analysis of loess SEM images,this paper puts forward a new method for calculating soil microstructure parameters,which convert pore zone distinguishment to pore pixel dot assortment.Q3 loess SEM images before and after seismic subsidence test are processed by using the above method and parameters such as pore numbers,pore areas,surface porosity,pore fractal dimension and pore distribution graph are achieved.The results show that after test macropores decrease in areas,mesopores and fine pores increase in numbers,and seismic subsidence is mainly induced by macropore failure.