Tree spatial pattern determination by satellite data, object-based classification and O-ring function (Case study: Sangedeh forest - Sari city)

Determination of the spatial pattern of trees is one of the important structural parameters of the forest for identifying changes in forest communities and it's monitoring in recent years. On the one hand, extracting this structural parameter using satellite data has a significant impact on reducing cost of inventory and makes it easy to evaluate forests of forest managers. In this study, the spatial pattern of trees in the natural forest of Farim of Mazandaran province was determined by using SPOT7 satellite data. Thus, some tree`s characteristics such as species, diameter at breast height, the diameter of canopies, gaps and the position of each tree were measured in 12 plot with area about one ha (100 m x 100 m). On the other hand, after radiometric and geometric corrections of satellite data, the tree`s canopy and gaps were classified, using the object-based method, and tree position per plot also was determined by canopy gravity center. Then, the spatial distribution pattern of the trees and structure, distinguished and extracted by field inventory and satellite data was compared by O-ring function. The results showed that the overall accuracy of canopy and gaps extracted from satellite data by object-based classification was about 91% and kappa coefficient equaled 0.7. The spatial pattern of trees were random except of the small radius and similar to the pattern of trees in-ground inventory. In total, results showed that SPOT-7 data and the object-based classification have an acceptable potential for determining the spatial pattern of forest stands.