Abstract
This paper describes a novel method developed for the segmentation of sedimentary grains in electron microscopy images. The algorithm utilizes the approach of region splitting and merging.
In the splitting stage, the marker-based watershed segmentation is used. In the merging phase, the typical characteristics of grains in electron microscopy images are exploited for proposing special metrics, which are then used during the merging stage to obtain correct grain segmentation.
The metrics are based on the typical intensity changes on the grain borders and the compact shape of grains. The experimental part describes the optimal setting of parameter in the splitting stage and the overall results of the proposed algorithm tested on available database of grains.
The results show that the proposed technique fulfills the requirements of its intended application.