Tropical Rain Forest Monitoring and Evaluation using Image Segmentation Approach

Abstract

-The tropical deforestation is now being tracked by actors in the nongovernmental, academic, private and government sectors using several different sources of satellite imagery. This paper proposes a satellite image analysis system that can be used for deforestation related examination and monitoring activities in the tropics. In this paper, the extraction of deforestation area based on the image grey level information and the adaptive threshold were obtained by Otsu threshold method. Then, the edge of the image is smoothen and dilated to outline the deforestation area. Finally we calculate the area of deforestation by summation of all the white pixel. Experiments show that the proposed method has acceptable accuracy rate which above 96% when compared with ground truth data.