期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2010
卷号：XXXVIII - Part 8
页码：637-642
出版社：Copernicus Publications
摘要：The pistachio (Pistacia.vera) belongs to the Anacardiaceae family. Plants of this family are trees or shrubs (broadleaf) and in total comprising 75 genera and 600 species. This species is found in arid and semi-arid regions, located in the northeast of Iran. Many algorithms have been investigated for tree delineation and identification, in order to assist human operators in the information exploitation of data. Objects of and questions posed by this study are; identification and delineation of the pistachio's tree crown, distinguishing between pistachio trees and other vegetation cover, estimation of tree density by counting trees per hectare, estimation of vegetation indices and the analyses of image segmentation, classification, texture and comparison with numbers of pistachio trees based on ground control data. ALOS satellite data have been used, using multispectral band (AVNIR-2) with 10-meter resolution and Panchromatic band (PRISM) with 2.5-meter resolution. In this paper, the relationship between the true number of trees (tree numbers measured on the ground) in the sub sample plots and introduced methods for tree population measurement has been investigated (considering factors such as tree crown area, shadow area, image classification and image segmentation from panchromatic (PRISM) image for pistachio forests). A sample plot area of about 1600 ha was selected incorporating 15-sub sample plots of 9 ha (300 ×300m) equal to 120 × 120 pixels for this research. A maximum filtering algorithm of 5 × 5 pixels size is used for vegetation indices in the pistachio forest. The results of this paper indicated that NDVI, OSAVI, SAVI (0.5), SAVI (1) and new index TRVI with 5×5 maximum filtering was (R 2 =0.5863, 0.5857, 0.4923, 0.6346 and 0.6414) respectively. The results of tree counting analyses will be showed
