GIS-based detection of grain boundaries

The recognition of grain boundaries in deformed rocks from images of thin-sections or polished slabs is an essential step in describing and quantifying various fabric elements and strain. However, many of the methods in use today require labor-intensive manual digitization of grain boundary information. Here, we propose an automated, GIS-based method to detect grain boundaries and construct a grain boundary database in which the shape, orientation, and spatial distribution of grains can be quantified and analyzed in a reproducible manner. The proposed method includes a series of operations and functions to identify grain boundaries and construct the grain boundary database. These processes are integrated into a GIS model using ArcGIS ModelBuilder; thus, little or no operator intervention is required to perform the entire analysis. The method was evaluated using thin section images taken from three sandstone samples. The results indicate that the proposed method can correctly identify >70% of grains recognized manually without any intervention and is especially suitable for analyses where large numbers of grains are required.