基于分形理论探究碳酸盐岩CT图像二值化最佳阈值  

Optimal thresholding in carbonate reservoir CT image binarization based on fractal theory

作　　者：吴国铭 李熙喆[2] 高树生[2] 刘华勋[2] 吴国铨 安为国[2] 

机构地区：[1]中国科学院大学渗流流体力学研究所,河北廊坊065007 [2]中国石油勘探开发研究院廊坊分院,河北廊坊065007 [3]兰州工业学院电气工程学院,甘肃兰州730050

出　　处：《石油地球物理勘探》2017年第5期1025-1032,共8页Oil Geophysical Prospecting

基　　金：国家重大科技专项(2011ZX05013-002)资助

摘　　要：CT扫描构建数字岩心的一个重要环节是灰度图像的二值化。本文基于图像处理分析和分形理论计算孔渗参数,为CT图像二值化提供了更准确、更适用的约束条件。孔洞型碳酸盐岩不同形态和尺度的微观孔隙累积数量与半径分布遵循幂律关系,具有统计意义上的分形特征,且分维值与孔隙度之间存在非线性定量关系。通过自定义四连通像素值梯度突变算法与分形理论相结合的方法,统计分析了孔隙形态、数量等参数并计算了孔隙度,其平均误差小于7.2%;改进"先二值化后边缘识别"的常规方法,运用滤波降噪和边缘识别算子识别标定CT灰度图像微观孔隙,既保证了精度又提高了效率。根据这两种图像处理技术可确定最佳灰度阈值,实现CT图像二值化。该方法处理的二值图像较好地保留了不同尺度微观孔隙结构的形态和分布特征,因此可广泛应用于常规岩心重构及后续的三维数字岩心的构建。Gray image binarization is one key step of digital core built by CT.With the means of imageprocess and analysis and fractal theory,we introduce a porosity calculation as a more accurate and appropriate constraint for CT image binarization.From our study,the cumulative quantity of vugs with different shapes and scales in carbonates has the power law relationship with the radiuses distribution,which means statistical fractal characteristics.Microscopic porosity development degree and heterogeneity can be represented by the fractal dimension,and the fractal dimension has a nonlinear quantitative relationship with porosity.Some parameters such as the number and shapes of vugs are statistically analyzed with the combination of the four direction-connecting pixel gradient mutation algorithm and fractal theory,and the porosity with small error is calculated.An appropriate filtering noise suppression and an edge detection operator can effectively and accurately calibrate microscopic porosity of CT gray images.Characteristics of microscopic porosity morphology and distribution are preserved in binary image that would be helpful for 3Ddigital core reconstruction in the future.

关 键 词：碳酸盐岩 微观孔隙 CT图像 分形 孔隙度 最佳阈值 

分 类 号：P631[天文地球—地质矿产勘探]