To improve a classification, first assess the accuracy of your results. The accuracy of a classification is measured against a standard that is assumed to be correct. The classification accuracy increases as it approaches the standard.
Once you have assessed the classification accuracy, you can combine classes through a process known as aggregation. Combining classes creates a new aggregate class. A maximum of 255 classes can be reassigned in a single session. Aggregation is often performed on the results of an unsupervised classification. A common approach in unsupervised classification is to generate as many cluster classes as possible. With the benefit of reference data or first-hand knowledge of a scene, you can aggregate the spectral clusters into meaningful thematic classes.
© PCI Geomatics Enterprises, Inc.®, 2026. All rights reserved.