基于图像分割的煤岩孔隙多尺度分形特征

Multiple-scale fractal features of coal and rock porosity based on image segmentation

根据煤岩孔隙系统的精确描述对流体运移机理的重要性,对来自6个不同矿区的煤岩进行了扫锚电镜观测,提出了基于扫锚电镜图像分割的煤岩孔隙多尺度分形研究方法。采用了图像分割算法中的灰度阈值分割、Laplacian边缘算子及Prewitt边缘算子检测,提取了煤岩扫描电镜(SEM)图片中的大尺度裂隙、边界曲线以及小尺度孔隙特征;采用盒维数法计算了分形维数,研究了多重孔隙介质分形标度律存在的尺度与SEM放大倍数的关系。结果表明:大尺度裂纹分形标度律存在的下限尺度与7 500倍SEM观测尺度相对应;小尺度孔隙分形标度律存在的上限尺度与2 000倍的SEM观测尺度相对应;细观裂纹分维数反映了煤岩渗透性能;边界分维反映了流动边界的迂曲程度相关于流动阻力;微观孔隙分维相关于吸附性能。

According to the importance of the fluid transportation mechanism detail stated by the coal and rock pore system,six coal and rock samples from different mining areas were observed with scanning electron microscope(SEM).The paper proposed a new approach for investigation of multiple scale fractals based on image segmentation.The following segmentation algorithms were employed: the gray level thresholding,the Laplacian edge detector in combination with thresholding,and the Prewitt edge detector to extract,at...