<![CDATA[Crystal Size Distribution (CSD) is a quantitative method used to assess magmatic processes such as cooling rates, crystallization durations, and crystal growth dynamics. Traditionally, mineral segmentation in CSD analysis is performed manually, which is labor-intensive, particularly for large datasets. This study investigates a semi-automatic segmentation approach to improve efficiency in analyzing thin sections of igneous rocks. The segmentation process is divided into two methods: semi-automatic segmentation using SAGA software and manual segmentation for validation. The analysis focuses on phenocrysts (>0.2 mm) and microphenocrysts (0.2–<0.02 mm), with images taken using a polarizing microscope at 4x magnification. While the semi-automatic method showed limitations due to the thin section texture and the inherent characteristics of the OBIA algorithm, it proved effective in estimating magma residence time with an average absolute error of 0.4 years. Additionally, the method demonstrated a mean regression gradient error of 79% for microphenocrysts and 55% for phenocrysts, supporting its application in magma dynamics interpretation. This approach enhances the practicality of CSD analysis, particularly in large datasets, and provides a valuable tool for studying crystallization processes in igneous rocks. However, direct application for interpreting magma dynamic should be done with caution.]]>
Copyrights © 2024
