The Region Grow function groups neighboring pixels which are similar into groups. The algorithm starts from an initial seed and examines all pixels within a given neighborhood. If the pixels are similar enough, then they are added to the region. The process is iterated on, until all similar connecting pixels are in the region.
The inputs to this function are the following:
- Raster—The dataset that you want to perform this function on.
- Seed Points— A point feature class, serving as the initial seeds for the algorithm. Each seed (or point) corresponds to one entry in the attribute table, which carries the max growth radius, similarity threshold, and fill value information.
- Max Growth Radius Field—The field in the attribute table that defines the max growth radius, in image's spatial reference unit.
- Similarity Threshold Field—The field in the attribute table that defines the similarity threshold, in the Euclidean distance in spectral space.
- Fill Value Field (Optional)— The field in the attribute table that defines the fill value. In a multiple-band image, all bands will be assigned this value.