Fuzzy Analysis of Airborne LiDAR Data for Rainforest Boundary Determination

Abstract

Airborne LiDAR data have advantages over passive remote sensing data in detailed description of vertical forest structure. LiDAR-derived information can potentially be used to solve such problems as forest type classification and forest boundary determination. Forest boundaries were usually represented as sharp lines that attempt to distinguish areas with different forest types. In reality, however, forest boundaries are rarely sharp or crisp, especially in the forest area with multiple canopy layers where species compositions change gradually. Fuzzy analysis offers great potential for characterising the transition zones and determining realistic forest boundaries. This study developed ways of using fuzzy analysis of airborne LiDAR data for determining rainforest boundaries. LiDAR variables were derived and used to define and calculate membership function values for both rainforest and non-rainforest. The confusion index values were then derived to illustrate the transition zones. Finally, the rainforest boundaries were successfully determined in the study area. The results demonstrated the success of proposed method for rainforest boundary determination.