Initializing unsupervised classification

The Unsupervised Classification window allows you to choose the type of algorithm and the parameters you want to use in the classification.

In the Unsupervised Classification window, enable one of the following algorithm options:
- K-Means
- Fuzzy K-Means
- IsoData
In the algorithm Parameters table, click in the appropriate box in the Values column and enter the criteria that you want in the classification.
In the Classifications Options area, enable any of the following check boxes:
- Show report
- Save signatures
- Create PCT
Choose a training site option from the Use bitmap as mask list box. If you choose an option other than None, choose a mask region from the Classify region list box.
Click OK.

Note: It is recommended that you use many clusters (for example, the full 255 allowed in an 8U channel) and then perform aggregation to get the actual number of information classes you want.

Note: To view the unclassified image, drag the unclassified image file layer to the top of the Maps tree above the Classification MetaLayer. The original image appears in the view pane.