algorithms of computer analysis

Research of forest classification methods using high-resolution space images

In this paper a researches of methods for the controlled classification by the spectral distance, Mahalanobis distance and maximum probability to identify forests was conducted. As a materials for the research space images of the satellites Ikonos, QuickBird and test information from ground-based methods were used. As an outcome it was identified that the maximum probability method is the most efficient one for the forest classification issue.