Abstract The microscopic identification of rock sections is one of the most basic research methods of petromineralogy, which is important to the research of mineral composition, pore throat characteristics, interstitial filling etc. The traditional manual identification method was time-consuming, and also burdened with errors due to the subjective nature of human judgment, which can bring some uncertainties to the further results. Therefore, this paper proposes a technique for image acquisition and analysis of rock sections, based on sequential images of rock sections in angle domain under polarizing microscope. The analysis includes the following three steps. Firstly, based on the multi-angles extinction feature of rock section, the direction-adaptive correlation coefficient method was used to characterize the edge of the rock sections and segment the rock grains. A series of parameters, including grain size, roundness, convexity, flatness and roughness, were also extracted to characterize the morphological features of rock grains quantitatively. Secondly, the variation characteristics of multi-channel parameters (RGB and Gray channel) with angle domain were analyzed for different minerals and then the mineral composition was accurately identified. On the basis of grain segmentation, the pore structures, including intergranular pores, dissolved pores and throat, were extracted. And the parameters of pore structure were also obtained. Finally, the distribution of the interstitials was characterized based on the change of the extinction characteristics of the interstitials in the pores. Image analysis technique based on rock sections with polarizing microscope provides an effective and convenient method for grain segmentation, mineral composition classification, pore identification and interstitial analysis.

中文翻译：

---

偏光角域岩石断面正交显微图像采集分析技术

摘要 岩石断面显微识别是岩石矿物学最基本的研究方法之一，对矿物成分、孔喉特征、间隙充填等研究具有重要意义。由于人为判断的主观性导致的错误，会给进一步的结果带来一些不确定性。因此，本文提出了一种基于偏光显微镜下角度域岩石断面序列图像的岩石断面图像采集和分析技术。分析包括以下三个步骤。首先，根据岩石断面的多角度消光特征，采用方向自适应相关系数法对岩石断面边缘进行刻画，对岩石颗粒进行分割。还提取了一系列参数，包括粒度、圆度、凸度、平整度和粗糙度，以定量表征岩石颗粒的形态特征。其次，分析了不同矿物的多通道参数（RGB和灰色通道）随角度域的变化特征，从而准确识别了矿物成分。在颗粒分割的基础上，提取出粒间孔、溶孔、喉道等孔隙结构。并得到了孔结构参数。最后，根据孔隙中间隙质消光特性的变化表征间隙质的分布。偏光显微镜下基于岩石断面的图像分析技术为颗粒分割、矿物成分分类、孔隙识别和间隙分析提供了一种有效而便捷的方法。