Document Type : Research Paper
Nowadays, Persian Oak (Quercus brantii Lindl.) Defoliation is one of the main problems, threatening Zagros forest stands. In recent years, the oak stands located in Barm plain (Fars province) also have been facing with a severe defoliation. To protect and to manage the forest stands, employing satellite remote sensing reduces difficulties and expenses of field-based methods for detecting and delineating the defoliation extents. Using current spatial and spectral moderate resolution satellite imageries such as Asters’, by itself, have not resulted satisfactory achievements in detecting the defoliation in single tree level in compare to forest stand level. In the current research, capabilities of high spatial resolution newly-introduced Rapideye Imagery has been investigated in compare to multispectral Aster-L1B imagery in order to detecting the oak single tree defoliation. Plenty of multi-sensorial image processing techniques, supervised classification methods and spectral separability analysis have been employed in this research. The results of accuracy assessment indicate that overall accuracy and kappa coefficient of single tree defoliation map, derived from Aster-L1B imagery, were 75% and 56% respectively. It is while the employed image fusion techniques; particularly multi-sensorial Gram-Schmidt’s has increased the both heuristics up to 94% and 88% in turn. Rapideye imageries, by itself, resulted considerable high accuracy heuristics up to 97% and 94%, seem to be related to its Red Edge capability, specifically designed for detecting themes caused by physiological characteristics of vegetation covers such as defoliation.