Watershed

Segmentation > Watershed

Segments the Current Image by first interpreting its grayscale values as topography (ie, basins within level ground). A rainfall simulation is then run such that areas of similar intensity are separated from others, where their basins fill to meet [1].

1. Markers

Helps confine feature outlines near your marked locations. Choose to use the selection in a Memory Image as markers to control watershed. Forces “basins” to be formed at the marker locations prior to the watershed algorithm being run.

2. Resolution

• High: Works better for separating features with high pixel density
• Low: Works better for separating features with low pixel density (more pixelated)

3. Thickness

Thickness of watershed lines

4. Separation

Controls strength of feature detection. Increase slider to create more feature outlines.

Tips

• Powerful algorithm that can be used in combination with Pre-Processing and Find Texture filters to identify feature boundaries
• When trying to separate already segmented features, use the Separate Features tool instead

References

[1] Meyer, Fernand, “Topographic distance and watershed lines,” Signal Processing , Vol. 38, July 1994, pp. 113-125.