Applying object - base imagery classification technique to classify forest status based on circular No.34

Abstract

This article presents a result of applying object - base imagery classification technique to determine forest status based on Circular No. 34 for 4 communes: Son Kim 1, Son Kim 2, Son Tay and Tay Son in Huong Son district,ha Tinh province. SPOT5 HRS imagery and eCognition Developer and ArcGIS software were used for classification. Applying multi - resolution segmentation algorithm of eCognition Developer to segment the imagery of study area into 29,974 deference objects with ranging areas from 0.05 to 27.8 hectare. To match the definition of forest, the objects under 0,05 hectare were merge into the contiguous larger object that closest texture characteristic. After merging, the number of objects are 17,465 with 0,5 to 27,8 hectare of area. Based onthe results of classification indices from ground survey, segmented imagery of SPOT5 were classified into 7 statuses of forest: (1) rich ever - green, (2) medium ever - green, (3) poor ever - green, (4) rehabilitation ever - green, (5) mixed wood and bamboo,
(6) plantation and (7) bare land.