Supervised Classification - the Maximum Likelihood and Minimum Distance to Means classifiers

This image is the result of the Maximum Likelihood (ML) classifier. The classification is similar to the Minimum Distance to Means (raw) classification (see below). The ML classifier was not so sensitive in separating cloud shadows and aquatic vegetation. Generally it seems that the ML classifier supports frequently appearing classes.
 
 

Back

This image is the result of the Minimum Distance to Means - raw distances (MDMr) classifier. The classification is similar to the Maximum Likelihood classification (see above). The MDMr classifier was also not so sensitive in separating cloud shadows and aquatic vegetation. Generally it seems that the MDM classifier supports frequently appearing classes.

Back

This image is the result of the Minimum Distance to Means - normalized distances (MDMn) classifier. This classifier seems to support rarely appearing classes. The separation of cloud shadows and the Aquatic plant assemblage was well done. But generally, the same general trend in all three methods can be observed. 

Back