The results of a hybrid machine learning approach for segmentation of methane hydrate-bearing sample.

This rar archive contains the result of applying the hybrid machine learning segmentation algorithm to the reconstructed CT volume obtained during the formation of methane hydrate in a sand sample. Archive content: raw_volume.tiff - CT volume acquired during a dynamic in-situ experiment at the bending magnet beamline 2-BM of the Advanced Photon Source, Argonne National Laboratory. segmented_volume.tiff - segmented CT volume with following phases: 0 - background, 1 - sand grains, 2 - methane gas, 3 - methane gas hydrate, 4 - mixture of hydrate and brine, 5 - NaBr brine. This data is refrenced in our companion paper "Enabling quantitative analysis of time-resolved CT imaging of methane-hydrate formation with a hybrid machine learning approach"本RAR压缩包包含将混合机器学习分割算法应用于砂样甲烷水合物形成过程中获取的重建CT体数据所得的分割结果。压缩包内含以下文件：raw_volume.tiff——阿贡国家实验室（Argonne National Laboratory）先进光子源（Advanced Photon Source）2-BM弯曲磁铁束线开展的原位动态实验中采集的CT体数据；segmented_volume.tiff——已完成分割的CT体数据，其标注相态如下：0代表背景，1代表砂粒，2代表甲烷气体，3代表甲烷水合物，4代表水合物与盐水的混合物，5代表溴化钠（NaBr）盐水。本数据集的相关研究已被引用于配套论文《基于混合机器学习方法实现甲烷水合物形成过程时间分辨CT成像的定量分析》（Enabling quantitative analysis of time-resolved CT imaging of methane-hydrate formation with a hybrid machine learning approach）