Classification of Remote Sensing Images Based on α-torrent Rough Set

Spectral uncertainty or vagueness caused by spectral confusion between-class and spectral variation within-class leads to the overlap in a large number of features. In these cases, the traditional rough sets can not perform and extract knowledge effectively. To solve this problem, this research introduced α-torrent rough set theory to the field of remote sensing classification, and proposed a classifier based on α-torrent rough set theory. With this classifier classification knowledge can be extracted, allowing certain permissible misclassification rate. The classifier adopted a knowledge ensemble method which can assist classifier to make a decision. The experiments showed that the classification accuracy and knowledge explainable had been greatly improved.