Specifying the input channels

Spectral classes are groups of pixels that have uniform or nearly similar brightness values in the different spectral channels of the data. The spectral attributes of the classes used to train the computer are determined by the spectral channels that you choose to include in the training and classification process. You can use Focus to match the information classes you create with a spectral class in the image data.

Note: To determine the optimal combination of bands for spectral discrimination in your training sites you can use the channel select (CHNSEL) algorithm in the Algorithm Librarian.

Once you have set the RGB values for true color, you can define the spectral attributes for the ground cover you want to classify. You must choose a set of data channels that will define the spectral attributes or signatures in your classes.

  1. In the Session Configuration window, click in a cell next to a spectral channel in the Input Channels column. A check mark indicates that the channel is selected.
  2. Select a channel for collecting your training sites.
    Note: You can use an empty channel, a channel where the contents can be overwritten, or you can add a new working channel.
    Note: You can add empty channels to the Session Configuration list and the Files tree.
  3. To add a new channel:
    1. Click the Files tab and right-click the image file folder.
    2. Click New and then click Raster Layer.
    3. In the Add Image Channels window, choose the bit depth and number of new channels you want to add.
    4. Click Add. If you want to close the Add Image Channels window, click Close.

    Once you have added channels, you must select them in the Session Configuration window so that they receive the classification output. If you have not added empty channels, you must select channels that you intend to overwrite.

  4. To select the output channel, click a channel in the Output Channel column, and click OK.

    The details of the configuration setup are now shown in the Maps tree and the reference image opens in the view pane. The Training Site Editor window opens with a blank training channel. The training channel is now write-enabled and ready for you to begin collecting your training sites.

© PCI Geomatics Enterprises, Inc.®, 2026. All rights reserved.