Filtering classified output

This filtering process removes isolated pixels, or noise, from the classification output. The Majority Filter tool is used to accomplish this task.

In the post-classification workflow, this task is the first in a series of processing steps.

Steps:

1. To locate the tool, click on the Search window button on the Standard toolbar.
2. In the Search window, type majority filter and press Enter.
3. In the search results, click the Majority Filter item. This opens the Majority Filter tool.
4. On the tool dialog box, set the classified image as Input raster. Accept default settings for other parameters.
5. Click OK to run the tool.

The following shows an example of filtering. The first graphic is a raw image from a classification analysis. The second is the same image after filtering has been applied. As you can see, much of the noise is removed as a result.

Image before filtering:

Image after filtering: