基于数学形态学图像处理的泥质纹层定量表征——以加拿大麦凯河油砂区块为例

Quantitative Characterization of Mud Laminae Based on Mathematical Morphological Image Processing:A Case of Mackay River Oil Sands,Canada

油砂储层内大量发育的毫米至厘米级泥质纹层会显著降低热采开发效益。为明确泥质纹层的发育特征,以加拿大麦凯河油砂区块下白垩统McMurray组上段为研究对象,基于泥质纹层定性研究成果,针对不同类别泥质纹层的岩心照片,利用Matlab软件开展了数学形态学图像处理,在一定误差范围内统计了6类泥质纹层的泥质含量、纹层厚度和发育频数等定量表征参数,有效地将泥质纹层定性研究成果定量化。结果表明:6种泥质纹层的单条纹层厚度变化范围为0.2～6 cm,发育频数为32～95条/m,泥质含量为9%～51%。在形态学处理过程中,图像二值化的分割阈值是影响参数精度的主要因素;对图像进行先腐蚀后膨胀的形态学运算能够在保留关键信息的前提下,去除无效噪点。提出了将数学形态学图像处理技术应用到岩心照片的定量研究中,为充分挖掘岩心照片信息和提高资料处理效率提供了技术支撑。

A large number of mm to cm-scale mud laminae developed in oil sands reservoir will reduce thermal production efficiency.In order to clarify the features of these laminae,a case study of the upper McMurray Formation,Lower Cretaceous,Mackay River oil sands,Canada was conducted.Based on the qualitative division of mud laminae,for the cores photos of different types of muddy laminae,a Mathematical Morphology Image Processing method was carried out by Matlab.Within tolerable error,parameters such as mud content,laminae thickness,and frequency were calculated.The quantitative features of 6 mud layers were characterized effectively.The results show that the thickness of each single layer of 6 mud laminae varies from 0.2 cm to 6 cm,the frequency changes from 32 m-1 to 95 m-1,and the mud content ranges from 9%to 51%.In the images processing,the threshold of binary image is an important factor that affects the accuracy;the morphological process by first corrosion and then expansion can remove noisy of the binary image without losing key information.This study is an attempt to apply mathematical morphological image processing technology into the quantitatively analysis of core images,which could provide a technology basis for digging more information of core images and high efficiency of data processing.