The grain identification process in GrainFind is mainly based on the Watershed Algorithm (WA, https://en.wikipedia.org/wiki/Watershed_(image_processing)) that is widely used for the segmentation of image data. The challenge of identifying individual grains in a connected structure can be performed through a segmentation of the structure. The algorithm for the grain identification consists of the following steps:

1. Converting the image into a distance map using the Euclidean Distance Transform (EDT).
2. Identifying local maxima in the distance map and converting them to grain seeds as starting point for the WA.
3. Identifying individual grains by a grain-border determination through the WA.
4. Post-processing of the identified grains.
   1. Handling of grain fragments
   2. Handling of boundary grains

Only the parameterization of the watershed transform algorithm (choosing a minimum grain diameter) and the post-processing (reconnection of grain fragments, boundary grain handling, etc.) require user input. The complexity of the algorithm – such as the EDT - is hidden “under the hood”.

The main steps to run the watershed algorithm for grain identification are:

1. Select the material to be analyzed
2. On the chosen material, the EDT is carried out as a preparation for the watershed transform
3. The watershed transform is conducted based on the EDT.

In many cases, the result of the watershed transform is enough to identify the grains. Otherwise, further steps are required.