Forest-cover monitoring based on high resolution remote sensing image

Abstract

Abstract:The article put forward a method on forest-cover monitoring based on high resolution remote sensing data of adjacent two year and existing vector data of forest sub-compartment. With the GF-2 remote sensing image and Rapid-Eye images as the data sources, multi-scale segmentation under the forest land boundary control was conducted, and object of deforestation samples were selected. The object-oriented method of KNN and rules-based were used to classify the image and extract the patches of deforestation. Then we get the deforestation vector diagram of early and later periods. The deforestation patches of two periods were compared and analyzed to achieve the deforestation and recovering vector diagram in a year. Finally, the paper proved the feasibility of the method in forest-cover monitoring application by accuracy analysis.