Impact of image resolution on quantification of mineral abundances and accessible surface areas
Impact of image resolution on quantification of mineral abundances and accessible surface areas
复制标题
图像分辨率对矿物丰度和可及表面积量化的影响
DOI：
10.1016/j.chemgeo.2019.06.004
复制
发表时间：
2019
期刊：
影响因子：
3.9
通讯作者：
L. Beckingham
中科院分区：
文献类型：
--
作者：
F. Qin;L. Beckingham
成像已经成为地质样品表征和参数化反应输运模拟的一种有价值的手段，例如，图像分析可以提供孔隙度、矿物成分和矿物可接近表面积数据。可以使用各种技术和一系列分辨率收集图像，但图像分辨率对测量属性的影响在很大程度上是未知的。在这项工作中，二维图像分辨率对计算的矿物丰度、可达性和有效表面积的影响进行了研究，以密西西比州肯珀县的Paluxy地层为样本。薄片的扫描电镜（SEM）背散射电子（BSE）图像的分辨率为0.3 ~ 6 μm。使用ImageJ和MATLAB开发的算法将图像分割成孔隙和离散的矿物相。通过对分割图像中的孔隙和矿物像元进行计数来计算孔隙度、矿物丰度和矿物可达性，其中可达性矿物视为与孔隙空间相邻的矿物。采集岩心样品的三维x射线计算机断层扫描（CT）图像，对其进行分割和分析，以评估三维连通孔隙度。随机抽取与二维图像总面积相同的长方体，用于计算三维连通表面积。然后将其乘以矿物可达性来计算非粘土矿物的可达性矿物表面积。从不同分辨率的图像中计算出的矿物丰度变化最小。对于0.3 ~ 1 μm的高分辨率图像，矿物可达性一致性较好。在1 ~ 6 μm分辨率范围内，蒙脱石/伊利石的可及度随分辨率的降低而降低，石英的可及度随分辨率的增加而增加。这反过来又导致石英的有效表面积随着分辨率的降低而增加。方解石、菱铁矿和钾长石的含量变化不显著。
Imaging has emerged as a valuable means of geological sample characterization and parameterizing reactive transport simulations where image analysis can provide porosity, mineral composition and mineral accessible surface area data, for example. Images can be collected using a variety of techniques and at a range of resolutions, yet the impact of image resolution on measured properties is largely unknown. In this work, the impact of 2D image resolution on the calculated mineral abundances, accessibilities and effective surface areas are examined for a sample from the Paluxy formation, Kemper County, Mississippi. Scanning electron microscopy (SEM) backscatter electron (BSE) images of thin sections were captured at resolutions ranging from 0.3 μm to 6 μm. Images were segmented into pores and discrete mineral phases using ImageJ and algorithms developed in MATLAB. Porosity, mineral abundances and mineral accessibilities were calculated by counting pore and mineral pixels in the segmented image where accessible minerals were deemed as those adjacent to the pore space. A 3D X-ray computed tomography (CT) image of a core sample was collected, segmented, and analyzed to evaluate the 3D connected porosity. Cuboids with the same total area as the 2D image were randomly sampled and used to calculate the 3D connected surface area. This was then multiplied by mineral accessibility to calculate accessible mineral surface areas for non-clay minerals. Minimum variations were observed for mineral abundances calculated from images with varying resolutions. For high resolution images, 0.3 μm to 1 μm, mineral accessibilities agreed relatively well. For images with resolutions from 1 μm to 6 μm, the calculated accessibility of smectite/illite decreased with decreasing resolution while quartz accessibility increased. This in turn resulted in higher effective surface areas for quartz with decreasing resolution. No significant variations were observed for calcite, siderite and K-feldspar.