A comparison pixel and object based method for tree extraction by ALOS data: (A case study in juniper forest in the Northeast of IRAN)

Juniperus excelsa subsp. Polycarpos, which Iranians know as the Persian juniper, located in the northeast of Iran. The survey of forest resources is an important task for the management and protection of the forest. Traditionally, such an essential task is heavily dependent on the labor-intensive ground survey. In this paper, we must be able to extract tree density by two methods and compare with ground data. Have been calculated image segmentation with two algorithms, K-nearest Neighbor (KNN) and Support Vector Machine (SVM) as object based method and another method as new method that is pixel based. Subsequently these methods with ground data have been compared. Result showed the simple relationship coefficient regression between them and vegetation indices (Vis) have been evaluated. The results simple regression coefficient between new method and VIS indicate for NDVI, TRVI, OSAVI, SAVI (1), SAVI (0.5), MSAVI and modified of TRVI were (R² = 0.801, 0.83, 0.8422, 0.7008, 0.7339, 0.811 and 0.8328) respectively. The results simple regression coefficient between ground data and VIS for NDVI, TRVI, OSAVI, SAVI (1), SAVI (0.5), MSAVI and modified of TRVI were (0.6325, 0.6287, 0.7469, 0.6424, 0.6397, 0.5878 and 0.7668) respectively. The results simple regression coefficient between ground data and new method was (R² = 0.7812).