Morphological segmentation of remotely sensed forest covers in high spatial resolution images

In this paper novel algorithms developed to automatically segment forest cover types are presented. The digital image analysis followed is mainly based on mathematical morphology operators and exploits the textural features in high spatial resolution images. These input images consist of true colour digital orthophotos and the studied forest covers consist of the main occurrences in Portugal: olive trees, cork oak trees, pine and eucalyptus trees.