Image processing code for characterization of multiphase flow in porous media

This work utilizes microfluidic experiments to gather data captured as snapshots during the experiments. These snapshots provide real-time information and undergo image processing to derive the required data. Image processing involves several steps tailored to the investigations:<br>
<ol>
<li>Making a reference image (mask): This process involves creating a reference image, or mask, to document the initial conditions. For instance, the porous domain is imaged when saturated with one phase to differentiate various areas containing different phases.</li>
<li> Reading and cutting images: Images showing changes in fluid volume fraction are selectively chosen and processed. Each image is read into MATLAB, and the area of interest is extracted.</li>
<li>Image segmentation: Labeling each pixel of the images is done via thresholding and edge detection.</li>
<li>Measuring parameters: Parameters like saturation, interfacial length, area, contact angle, and curvature are measured. These parameters play a crucial role in analyzing the experiments. The interfacial area is calculated through various formulations.</li>
<li>Calculating capillary pressure: Several forms of capillary pressure are calculated using information derived from the experiments. </li>
<li>REV-Scale Quantities: Parameters are upscaled to represent Representative Elementary Volume (REV)-scale values essential for continuum theories. REV-scale capillary pressure is derived from pore-scale values using appropriate averaging techniques.</li>
</ol>

本研究借助微流控实验采集实验过程中拍摄的快照数据。此类快照可提供实时实验信息，后续通过适配本次研究需求的图像处理流程提取所需数据。针对本次研究设计，图像处理包含以下步骤：
1. 制作参考图像（掩膜mask）：本步骤旨在创建参考图像（即掩膜mask）以记录实验初始状态。例如，当多孔区域被单一流体相饱和时进行成像，以此区分含不同流体相的区域。
2. 图像读取与裁剪：筛选出可反映流体体积分数变化的图像进行处理。将每张图像导入MATLAB软件，提取实验感兴趣区域。
3. 图像分割（image segmentation）：通过阈值处理与边缘检测算法，对图像的所有像素进行标记分类。
4. 参数测量：测量饱和度、界面长度、界面面积、接触角及曲率等关键参数，此类参数为实验分析的核心依据。界面面积可通过多种公式进行计算。
5. 毛细管压力计算：基于实验获取的相关信息，计算多种形式的毛细管压力。
6. 代表性单元体积（Representative Elementary Volume, REV）尺度量化：将孔隙尺度的参数向上尺度化，得到连续介质理论所需的REV尺度数值。通过合适的平均方法，可由孔隙尺度的毛细管压力推导得到REV尺度毛细管压力。